ECHA - Guidance of the application of CLP criteria [November 2012]

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012LEGAL NOTICEThis document contains guidance to Regulation (EC) No 1272/2008 (CLP Regulation). However,users are reminded that the text of the CLP Regulation is the only authentic legal reference and thatthe information in this document does not constitute legal advice. The European Chemicals Agencydoes not accept any liability with regard to the contents of this document.ECHA Reference: ECHA-12-G-14-ENDate: 11/2012Language: EN© European Chemicals Agency, 2012.Cover page © European Chemicals AgencyReproduction is authorised provided the source is fully acknowledged in the form “Source:European Chemicals Agency, http://echa.europa.eu/”, and provided written notification is given tothe ECHA Communication Unit (publications@echa.europa.eu).If you have questions or comments in relation to this document please send them (indicating thedocument reference, issue date, chapter and/or page of the document to which your comment refers)using the Guidance feedback form. The feedback form can be accessed via the ECHA Guidancewebsite or directly via the following link:https://comments.echa.europa.eu/comments_cms/FeedbackGuidance.aspxEuropean Chemicals AgencyMailing address: P.O. Box 400, FI-00121 Helsinki, FinlandVisiting address: Annankatu 18, 00120 Helsinki, Finland2

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012DOCUMENT HISTORYVersion Comment Daten.a. First edition August 2009n.a.Version 2Please note that change between the version published in August2009 and that of April 2011 are not recorded in this documenthistory.Revision of the Guidance addressing content in relation to theenvironmental criteria chapters and Annexes following the 2 ndAdaptation to Technical Progress to the CLP Regulation(Commission Regulation (EU) No 286/2011). The ECHASecretariat revised the Guidance Part 4 - Environmental hazardsand Annexes of the guidance document referring to the revisedcriteria for the long-term aquatic hazard for substances andmixtures and added new Part 5 – Additional hazards referring tothe hazard class “hazardous to the ozone layer”. As well, a numberof examples have been included in the respective Parts andAnnexes to illustrate the revisions performed. Further to this, arange of editorial corrections were proposed for Part 1- Generalprinciples for classification and labelling.The update includes the following:- Revision of Part 1, by eliminating and amending out of dateinformation and restructuring the text in order to reflect theGuidance update.- All green boxes in Part 4 that are impacted by the 2 nd ATPwere updated. As the CLP legal text uses commas instead ofdots to define numbers smaller than 1, the green boxes nowshow commas as well.- Revision of Part 4, by providing guidance on the application ofthe new long-term aquatic hazard criteria for substances andmixtures.- Section 4.1.3 Classification of substances hazardous to theaquatic environment and section 4.1.4 Classification ofmixtures hazardous to the aquatic environment weresubstantially revised, for example by addition of newreferences, as well as the new/ revised examples to illustraterelevant topics in the Part 4.- New Part 5 - Additional hazards was added (please note thatPart 5: Labelling was deleted from the Guidance in previousnon recorded versions and covered via a new Guidance onLabelling and Packaging in accordance with Regulation (EC)No 1272/2008 published in April 2011).- Most of the I.3 sub-sections in Annex I – Aquatic toxicity wererevised.- In Annex II – Rapid degradation the terminology wasmodified.- Most of the Annex IV – Metals and Inorganic MetalCompounds was substantially modified and revised, as well asin sub-section IV.7 new examples were added.April 2011April 20123

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Version 3Revision of Guidance Part 3 Health Hazards, relating to specificconcentration limits (SCLs) for 4 hazard classes and the inclusionof a new AnnexThe update includes the following:- Revision of Part 3, by providing guidance on the setting oflower and higher SCLs for 4 health hazard classes in section3.2.2.5 Skin Corrosion/Irritation; section 3.3.2.5 Serious EyeDamage/Eye Irritation; section 3.7.2.5 Reproductive Toxicityand section 3.8.2.6 STOT-SE, in accordance with CLP Article10(7);- Inclusion of a new Annex (Annex VI) providing guidance onsetting SCLs for the reproductive toxicity hazard class basedon potency considerations.November 20124

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012PREFACEThis document is the Guidance on the Application of the CLP Criteria. It is a comprehensivetechnical and scientific document on the application of Regulation (EC) No 1272/2008 on theclassification, labelling and packaging of substances and mixtures (CLP), which will replacethe Dangerous Substances Directive 67/548/EEC (DSD) and the Dangerous PreparationsDirective 1999/45/EC (DPD) in a staggered way. CLP is based on the Globally HarmonisedSystem of Classification and Labelling of Chemicals (GHS) and is implementing theprovisions of the GHS within the EU. The objective of this document is to provide detailedguidance on the application of the CLP criteria for physical, health and environmentalhazards. The guidance is developed to assist primarily manufacturers or importers applyingclassification and labelling criteria and it also includes practical examples. It is also assumedto be the guidance on classification and labelling for Competent Authorities in the MemberStates (MS CA), for the Commission services and the European Chemicals Agency (ECHA).In certain chapters, like for example the ones on carcinogenicity, mutagenicity andreproductive toxicity, the guidance includes to a larger extent scientific advice on how tointerpret different data used for classification. This additional guidance is based onexperience gained within the EU during the application of the classification criteria underDirective 67/548/EEC, and is written for the experts within the respective fields.This guidance document was developed as a REACH Implementation Project (RIP 3.6) at theInstitute for Health and Consumer Products (IHCP) of the Joint Research Centre in Ispra,with support from working groups consisting of experts on classification and labelling fromEU Member States and Industry. The project started in September 2007 and the differentworking groups had meetings and continuous discussions to discuss and develop the guidancetext until spring 2009. Finally all texts were consolidated and edited at the IHCP. RIP 3.6 wasfinancially supported with an administrative arrangement made with Directorate-GeneralEnterprise and Industry. The guidance was handed over to ECHA in summer 2009.At the time of the hand-over, it was clear that further work was necessary in relation to theguidance chapters on health hazards ( for example on setting of specific concentration limits(SCLs)), on the long-term aquatic hazard and in relation to labelling and packaging. Inaddition further drafting work was done in close collaboration with European experts, to takeaccount of a range of guidance aspects 1 following the 2 nd Adaptation to Technical Progress(ATP) to the CLP Regulation (Commission Regulation (EU) No 286/2011 2 ). In relation tolabelling and packaging, a new stand-alone guidance document was prepared (“Guidance onLabelling and Packaging in accordance with Regulation (EC) No 1272/2008”), warrantingthe deletion of Part 5 and of Annex V of the Guidance on the Application of the CLP Criteria.The Guidance on Labelling and Packaging in accordance with Regulation (EC) No1272/2008 is published on ECHA’s guidance website, underhttp://guidance.echa.europa.eu/guidance_en.htm.1 Further guidance on the criteria for respiratory and skin sensitisation, on the aspiration hazard and other humanhealth related points, as well as on the physico- chemical points that were revised following the 2 nd ATP to theCLP Regulation are not part of this update and are planned for a future update of this document in 2013. .2 Commission Regulation (EU) No 286/2011 of 10 March 2011 amending, for the purposes of its adaptation totechnical and scientific progress, Regulation (EC) No 1272/2008 of the European Parliament and of the Councilon classification, labelling and packaging of substances and mixtures.5

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012TABLE OF CONTENTS1 PART 1: GENERAL PRINCIPLES FOR CLASSIFICATION AND LABELLING.........................321.1 INTRODUCTION...............................................................................................................................321.1.1 The objective of the guidance document ................................................................................321.1.2 Background.............................................................................................................................331.1.3 Hazard classification...............................................................................................................331.1.4 Who is responsible for the hazard classification and what is the timetable............................341.1.5 Which substances and mixtures should be classified (the scope)...........................................351.1.6 What information is needed for classification ........................................................................361.1.6.1 Information for the classification of substances......................................................361.1.6.2 Information relevant for the classification of mixtures ...........................................371.1.7 Data evaluation and reaching a decision on classification......................................................371.1.7.1 Classification of substances ....................................................................................371.1.7.2 Influence of impurities, additives or individual constituents on theclassification of a substance ....................................................................................381.1.8 Updating of hazard classifications..........................................................................................381.1.9 The interface between hazard classification and hazard communication...............................381.1.10 The interface between self-classification and harmonised classification, and the list ofharmonised classifications......................................................................................................391.1.11 The Classification and Labelling Inventory (C&L Inventory) ...............................................401.1.12 Relation of classification to other EU legislation ...................................................................401.1.12.1 REACH ...................................................................................................................401.1.12.2 Plant Protection Products and Biocides ..................................................................401.1.12.3 Transport legislation................................................................................................411.2 THE SIGNIFICANCE OF THE TERMS 'FORM OR PHYSICAL STATE’ AND'REASONABLY EXPECTED USE’ WITH RESPECT TO CLASSIFICATIONACCORDING TO CLP.......................................................................................................................411.2.1 'Form or physical state’ and 'reasonably expected use’ ..........................................................411.2.2 The term 'reasonably expected use’ in relation to hazard classification.................................421.2.3 The term ‘form or physical state’ in relation to hazard classification ....................................421.2.3.1 Physical hazards ......................................................................................................421.2.3.2 Human health hazards .............................................................................................431.2.3.3 Environmental hazards............................................................................................431.3 SPECIFIC CASES REQUIRING FURTHER EVALUATION – LACK OFBIOAVAILABILITY..........................................................................................................................441.3.1 Definition................................................................................................................................441.3.2 Bioavailability ........................................................................................................................441.3.2.1 Human health hazards .............................................................................................451.3.2.2 Environmental hazards............................................................................................461.4 USE OF SUBSTANCE CATEGORISATION (READ ACROSS AND GROUPING) AND(Q)SARS FOR CLASSIFICATION AND LABELLING ..................................................................461.4.1 (Q)SAR...................................................................................................................................471.4.2 Grouping.................................................................................................................................481.4.3 Read across.............................................................................................................................481.5 SPECIFIC CONCENTRATION LIMITS AND M-FACTORS .........................................................491.5.1 Specific concentration limits ..................................................................................................491.5.2 Multiplying factors (M-factors)..............................................................................................501.6 MIXTURES.........................................................................................................................................511.6.1 How to classify a mixture.......................................................................................................516

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121.6.2 Classification for physical hazards ........................................................................................531.6.3 Health and environmental hazards..........................................................................................541.6.3.1 Classification derived using data on the mixture itself ...........................................551.6.3.2 Bridging principles..................................................................................................551.6.3.2.1 Dilution .................................................................................................551.6.3.2.2 Batching ................................................................................................561.6.3.2.3 Concentration of highly hazardous mixtures.........................................561.6.3.2.4 Interpolation within one toxicity category ............................................561.6.3.2.5 Substantially similar mixtures...............................................................571.6.3.2.6 Review of classification where the composition of a mixture haschanged .................................................................................................581.6.3.3 Aerosols (some health hazards only) ......................................................................591.6.3.4 Classification based on calculation or concentration thresholds.............................591.6.3.4.1 Classification based on calculation .......................................................591.6.3.4.2 Classification based on concentration thresholds..................................611.6.3.4.3 Additivity of hazards.............................................................................631.6.4 Classification of mixtures in mixtures....................................................................................641.6.4.1 Example: Classification of Mixture A ....................................................................641.6.4.2 Example: Classification of Mixture B.....................................................................671.7 THE APPLICATION OF ANNEX VII...............................................................................................691.7.1 Introduction ............................................................................................................................691.7.2 Use of Annex VII translation tables .......................................................................................701.7.2.1 Applicability of the Annex VII translation tables ...................................................701.7.3 Additional considerations for re-classification due to changes in the classificationcriteria.....................................................................................................................................742 PART 2: PHYSICAL HAZARDS ......................................................................................................772.1 INTRODUCTION...............................................................................................................................772.1.1 General remarks about the prerequisites of classification and testing....................................772.1.2 Safety......................................................................................................................................772.1.3 General conditions for testing.................................................................................................772.1.4 Physical state ..........................................................................................................................782.1.5 Quality ....................................................................................................................................782.2 EXPLOSIVES .....................................................................................................................................792.2.1 Introduction ............................................................................................................................792.2.2 Definitions and general considerations for the classification of explosives ...........................792.2.3 Classification of substances, mixtures or articles as explosives.............................................802.2.3.1 Identification of hazard information .......................................................................802.2.3.2 Screening procedures and waiving of testing..........................................................812.2.3.3 Classification criteria...............................................................................................822.2.3.4 Testing and evaluation of hazard information.........................................................832.2.3.5 Classification procedure and decision logics ..........................................................842.2.3.5.1 Acceptance procedure ...........................................................................852.2.3.5.2 Assignment procedure to a division ......................................................882.2.4 Hazard communication for explosives ...................................................................................932.2.4.1 Pictograms, signal words, hazard statements and precautionary statements...........932.2.4.2 Additional labelling provisions ...............................................................................932.2.5 Re-classification of substances and mixtures classified as explosive according to DSDor already classified for transport ...........................................................................................942.2.5.1 Re-classification of substances and mixtures classified in accordance withDSD.........................................................................................................................942.2.5.2 Relation to transport classification..........................................................................952.2.6 Examples of classification for explosives...............................................................................957

Guidance on the Application of the CLP Criteria Version 3.0 - November 201282.2.6.1 Example of substances and mixtures fulfilling the classification criteria ...............952.2.6.2 Example of substances and mixtures not fulfilling the classification criteria .........962.3 FLAMMABLE GASES ......................................................................................................................992.3.1 Introduction ............................................................................................................................992.3.2 Definitions and general considerations for the classification of flammable gases .................992.3.3 Relation to other physical hazards..........................................................................................992.3.4 Classification of substances and mixtures as flammable gases ..............................................992.3.4.1 Identification of hazard information .......................................................................992.3.4.2 Screening procedures and waiving of testing for gas mixtures...............................992.3.4.3 Classification criteria.............................................................................................1002.3.4.4 Testing and evaluation of hazard information.......................................................1002.3.5 Hazard communication for flammable gases........................................................................1012.3.5.1 Pictograms, signal words, hazard statements and precautionary statements.........1012.3.5.2 Additional labelling provisions .............................................................................1012.3.6 Re-classification of substances and mixtures classified as flammable gases accordingto DSD or already classified for transport ............................................................................1022.3.6.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1022.3.6.2 Relation to transport classification........................................................................1022.3.7 Example of classification for flammable gases ....................................................................1022.4 FLAMMABLE AEROSOLS ............................................................................................................1032.4.1 Introduction ..........................................................................................................................1032.4.2 Definitions and general considerations for the classification of flammable aerosols...........1032.4.3 Classification of flammable aerosols....................................................................................1032.4.3.1 Classification criteria.............................................................................................1032.4.3.2 Testing and evaluation of hazard information.......................................................1042.4.3.3 Decision logic........................................................................................................1042.4.4 Hazard communication for flammable aerosols ...................................................................1072.4.4.1 Pictograms, signal words, hazard statements and precautionary statements.........1072.4.4.2 Additional labelling provisions .............................................................................1082.4.5 Re-classification of flammable aerosols according to DSD .................................................1082.4.6 Examples of classification for flammable aerosols ..............................................................1082.4.6.1 Examples of aerosols fulfilling the classification criteria .....................................1092.4.6.2 Examples of aerosols not fulfilling the classification criteria ...............................1092.5 OXIDISING GASES.........................................................................................................................1092.5.1 Introduction ..........................................................................................................................1092.5.2 Definitions and general considerations for the classification of oxidising gases .................1102.5.3 Classification of substances and mixtures as oxidising gases ..............................................1102.5.3.1 Identification of hazard information .....................................................................1102.5.3.2 Screening procedures and waiving of testing........................................................1102.5.3.3 Classification criteria.............................................................................................1102.5.3.4 Testing and evaluation of hazard information.......................................................1102.5.4 Hazard communication for oxidising gases..........................................................................1112.5.4.1 Pictograms, signal words, hazard statements and precautionary statements.........1112.5.5 Re-classification of substances and mixtures classified as oxidising gases according toDSD or already classified for transport ................................................................................1112.5.5.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1112.5.5.2 Relation to transport classification........................................................................1112.5.6 Examples of classification for oxidising gases.....................................................................1122.6 GASES UNDER PRESSURE...........................................................................................................1122.6.1 Introduction ..........................................................................................................................112

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.6.2 Definitions and general considerations for the classification of gases under pressure.........1122.6.2.1 Definition of “gas” ................................................................................................1122.6.2.2 Definition of “gases under pressure” ....................................................................1132.6.3 Relation to other physical hazards........................................................................................1132.6.4 Classification of substances and mixtures as gases under pressure......................................1132.6.4.1 Identification of hazard information .....................................................................1132.6.4.2 Classification criteria.............................................................................................1132.6.4.3 Testing and evaluation of hazard information.......................................................1142.6.5 Hazard communication for gases under pressure .................................................................1142.6.5.1 Pictograms, signal words, hazard statements and precautionary statements.........1142.6.6 Re-classification of substances and mixtures classified as gases under pressureaccording to DSD or already classified for transport ...........................................................1152.6.6.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1152.6.6.2 Relation to transport classification........................................................................1152.6.7 Examples of classification for gases under pressure ............................................................1152.7 FLAMMABLE LIQUIDS .................................................................................................................1162.7.1 Introduction ..........................................................................................................................1162.7.2 Definitions and general considerations for the classification of flammable liquids.............1162.7.3 Relation to other physical hazards........................................................................................1162.7.4 Classification of substances and mixtures as flammable liquids..........................................1162.7.4.1 Identification of hazard information .....................................................................1162.7.4.2 Screening procedures and waiving of testing........................................................1172.7.4.2.1 Boiling point........................................................................................1172.7.4.2.2 Flash point...........................................................................................1172.7.4.3 Classification criteria.............................................................................................1172.7.4.4 Testing and evaluation of hazard information.......................................................1182.7.4.4.1 Testing.................................................................................................1182.7.4.4.2 Evaluation of hazard information........................................................1192.7.4.5 Decision logic........................................................................................................1192.7.5 Hazard communication for flammable liquids .....................................................................1212.7.5.1 Pictograms, signal words, hazard statements and precautionary statements.........1212.7.5.2 Additional labelling provisions for flammable liquids..........................................1212.7.6 Re-classification of substances classified as flammable liquids according to DSD oralready classified for transport..............................................................................................1222.7.6.1 Re-classification according to DSD ......................................................................1222.7.6.2 Relation to transport classification........................................................................1222.7.7 Examples of classification for flammable liquids.................................................................1222.7.7.1 Examples of substances and mixtures fulfilling the classification criteria............1222.7.7.2 Examples of substances and mixtures not fulfilling the classification criteria......1232.7.8 References ............................................................................................................................1232.8 FLAMMABLE SOLIDS ...................................................................................................................1242.8.1 Introduction ..........................................................................................................................1242.8.2 Definitions and general considerations for the classification of flammable solids...............1242.8.3 Relation to other physical hazards........................................................................................1242.8.4 Classification of substances and mixtures as flammable solids............................................1242.8.4.1 Identification of hazard information .....................................................................1242.8.4.2 Screening procedures and waiving of testing........................................................1242.8.4.3 Classification criteria.............................................................................................1252.8.4.4 Testing and evaluation of hazard information.......................................................1252.8.4.5 Decision logic........................................................................................................1262.8.5 Hazard communication for flammable solids.......................................................................1262.8.5.1 Pictograms, signal words, hazard statements and precautionary statements.........1279

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.8.6 Re-classification of substances and mixtures classified as flammable solids accordingto DSD or already classified for transport ............................................................................1272.8.6.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1272.8.6.2 Relation to transport classification........................................................................1272.8.7 Examples of classification for flammable solids ..................................................................1282.8.7.1 Example of substances and mixtures fulfilling the classifiction criteria...............1282.8.7.2 Examples of substances and mixtures not fulfilling the classification criteria......1282.9 SELF-REACTIVE SUBSTANCES ..................................................................................................1282.9.1 Introduction ..........................................................................................................................1282.9.2 Definitions and general considerations for the classification of self-reactives.....................1292.9.3 Classification of substances and mixtures as self-reactive ...................................................1292.9.3.1 Identification of hazard information .....................................................................1292.9.3.2 Classification criteria.............................................................................................1302.9.3.3 Testing and evaluation of hazard information.......................................................1312.9.3.3.1 Thermal stability tests and temperature control ..................................1312.9.3.3.2 Additional testing ................................................................................1322.9.3.3.3 Additional classification considerations..............................................1332.9.3.4 Decision logic........................................................................................................1332.9.4 Hazard communication for self-reactives.............................................................................1352.9.4.1 Pictograms, signal words, hazard statements and precautionary statements.........1352.9.5 Re-classification of substances and mixtures classified as self-reactives according toDSD or already classified for transport ................................................................................1352.9.5.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1352.9.5.2 Relation to transport classification........................................................................1362.9.6 Examples of classification for self-reactives ........................................................................1362.9.6.1 Examples of substances and mixtures fulfilling the classification criteria............1362.10 PYROPHORIC LIQUIDS AND SOLIDS ........................................................................................1402.10.1 Introduction ..........................................................................................................................1402.10.2 Definitions and general considerations for the classification pyrophoric liquids andsolids.....................................................................................................................................1402.10.3 Relation to other physical hazards........................................................................................1402.10.4 Classification of substances and mixtures as pyrophoric liquids and solids ........................1412.10.4.1 Identification of hazard information .....................................................................1412.10.4.2 Screening procedures and waiving of testing........................................................1412.10.4.3 Classification criteria.............................................................................................1412.10.4.4 Testing and evaluation of hazard information.......................................................1422.10.4.5 Decision logic........................................................................................................1422.10.5 Hazard communication for pyrophoric liquids and solids....................................................1442.10.5.1 Pictograms, signal words, hazard statements and precautionary statements.........1442.10.6 Re-classification of substances and mixtures classified as pyrophoric liquids and solidsaccording to DSD or already classified for transport ...........................................................1442.10.6.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1442.10.6.2 Relation to transport classification........................................................................1452.10.7 Examples of classification for pyrophoric liquids and solids ...............................................1452.10.7.1 Examples of substances and mixtures fulfilling the classification criteria............1452.10.7.2 Examples of substances and mixtures not fulfilling the classification criteria......1462.10.8 References ............................................................................................................................1472.11 SELF-HEATING SUBSTANCES AND MIXTURES .....................................................................1472.11.1 Introduction ..........................................................................................................................14710

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.11.2 Definitions and general considerations for the classification of self-heating substancesand mixtures .........................................................................................................................1472.11.3 Relation to other physical hazards........................................................................................1472.11.4 Classification of self-heating substances and mixtures ........................................................1472.11.4.1 Identification of hazard information .....................................................................1472.11.4.2 Screening procedures and waiving of testing........................................................1482.11.4.3 Classification criteria.............................................................................................1482.11.4.4 Testing and evaluation of hazard information.......................................................1492.11.4.4.1 General remarks ..................................................................................1492.11.4.4.2 Sample preparation..............................................................................1492.11.4.4.3 Criteria and evaluation ........................................................................1492.11.4.5 Decision logic........................................................................................................1502.11.4.6 Exemption .............................................................................................................1522.11.5 Hazard communication for self-heating substances and mixtures........................................1532.11.5.1 Pictograms, signal words, hazard statements and precautionary statements.........1532.11.6 Re-classification of substances and mixtures classified according to DSD or alreadyclassified for transport ..........................................................................................................1532.11.6.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1532.11.6.2 Relation to transport classification........................................................................1542.11.7 Examples of classification for self-heating substances and mixtures...................................1542.11.7.1 Examples of substances and mixtures fulfilling the classification criteria............1542.11.7.2 Examples of substances and mixtures not fulfilling the classification criteria......1542.11.8 References ............................................................................................................................1552.12 SUBSTANCES AND MIXTURES WHICH, IN CONTACT WITH WATER, EMITFLAMMABLE GASES ....................................................................................................................1552.12.1 Introduction ..........................................................................................................................1552.12.2 Definitions and general considerations for the classification of substances andmixtures which, in contact with water, emit flammable gases .............................................1562.12.3 Classification of substances and mixtures which, in contact with water, emitflammable gases ...................................................................................................................1562.12.3.1 Identification of hazard information .....................................................................1562.12.3.2 Screening procedures and waiving of testing........................................................1572.12.3.3 Classification criteria.............................................................................................1572.12.3.4 Testing and evaluation of hazard information.......................................................1582.12.3.4.1 Testing procedure................................................................................1582.12.3.4.2 Evaluation of hazard information........................................................1592.12.3.5 Decision logic........................................................................................................1592.12.4 Hazard communication for substances and mixtures which, in contact with water, emitflammable gases ...................................................................................................................1612.12.4.1 Pictograms, signal words, hazard statements and precautionary statementsfor substances and mixtures ..................................................................................1612.12.4.2 Additional labelling provisions .............................................................................1612.12.5 Re-classification of substances and mixtures which, in contact with water, emitflammable gases according to DSD or already classified for transport................................1622.12.5.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1622.12.5.1.1 Differences in classification and labelling ..........................................1622.12.5.1.2 Differences in the test procedures .......................................................1622.12.5.2 Relation to transport classification........................................................................1632.12.6 Examples of classification for substances and mixtures which, in contact with water,emit flammable gases ...........................................................................................................1632.12.6.1 Example of a substance fulfilling the classification criteria..................................1632.12.6.2 Example of a substance not fulfilling the classification criteria............................1642.12.7 References ............................................................................................................................16411

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.13 OXIDISING LIQUIDS AND OXIDISING SOLIDS .......................................................................1642.13.1 Introduction ..........................................................................................................................1642.13.2 Definitions and general considerations for the classification of oxidising liquids andoxidising solids.....................................................................................................................1652.13.3 Classification of substances and mixtures as oxidising liquids and oxidising solids ...........1662.13.3.1 Identification of hazard information .....................................................................1662.13.3.1.1 Non-testing data ..................................................................................1662.13.3.2 Classification criteria.............................................................................................1672.13.3.2.1 General ................................................................................................1672.13.3.2.2 Oxidising liquids .................................................................................1672.13.3.2.3 Oxidising solids...................................................................................1682.13.3.3 Testing and evaluation of hazard information.......................................................1682.13.3.4 Decision logic........................................................................................................1692.13.3.4.1 Decision logic 2.13 for oxidising liquids ............................................1702.13.3.4.2 Decision logic 2.14 for oxidising solids..............................................171Hazard communication for oxidising liquids and oxidising solids .......................171Hazard communication for oxidising liquids and oxidising solids .......................1722.13.3.5 Pictograms, signal words, hazard statements and precautionary statements.........1722.13.4 Re-classification of substances and mixtures classified as oxidising liquids andoxidising solids according to DSD or already classified for transport .................................1722.13.4.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1722.13.4.1.1 Liquids.................................................................................................1722.13.4.1.2 Solids...................................................................................................1732.13.4.2 Relation to transport classification........................................................................1732.13.5 Examples of classification for oxidising liquids and oxidising solids..................................1732.13.5.1 Examples of substances and mixtures fulfilling the classification criteria............1732.13.5.1.1 Liquids.................................................................................................1732.13.5.1.2 Solids...................................................................................................1732.13.5.2 Examples of substances and mixtures not fulfilling the classification criteriafor 1742.13.5.2.1 Liquids.................................................................................................1742.13.5.2.2 Solids...................................................................................................1742.13.6 Reference..............................................................................................................................1742.14 ORGANIC PEROXIDES..................................................................................................................1742.14.1 Introduction ..........................................................................................................................1742.14.2 Definitions and general considerations for the classification of organic peroxides .............1742.14.3 Relation to other physical hazards........................................................................................1752.14.4 Classification of substances and mixtures as organic peroxides ..........................................1752.14.4.1 Identification of hazard information .....................................................................1752.14.4.2 Classification criteria.............................................................................................1752.14.4.3 Testing and evaluation of hazard information.......................................................1772.14.4.3.1 Thermal stability tests and temperature control ..................................1772.14.4.3.2 Additional testing ................................................................................1782.14.4.3.3 Additional classification considerations..............................................1782.14.4.4 Decision logic........................................................................................................1782.14.5 Hazard communication for organic peroxides......................................................................1802.14.5.1 Pictograms, signal words, hazard statements and precautionary statements.........1802.14.5.2 Additional labelling provisions for organic peroxides ..........................................1812.14.6 Re-classification of substances and mixtures classified as organic peroxides accordingto DSD or already classified according to transport.............................................................1812.14.6.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1812.14.6.2 Relation to transport classification........................................................................1812.14.7 Examples of classification for organic peroxides.................................................................18112

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.14.7.1 Examples of substances and mixtures fulfilling the classification criteria............1812.14.8 Additional remarks ...............................................................................................................183Explosive properties ..........................................................................................................................1832.15 CORROSIVE TO METALS .............................................................................................................1842.15.1 Introduction ..........................................................................................................................1842.15.2 Definitions and general considerations for the classification of substances andmixtures corrosive to metals.................................................................................................1852.15.3 Classification of substances and mixtures as corrosive to metals.........................................1852.15.3.1 Identification of hazard information .....................................................................1852.15.3.2 Screening procedures and waiving of testing........................................................1862.15.3.3 Classification criteria.............................................................................................1872.15.3.4 Testing and evaluation of hazard information.......................................................1872.15.3.4.1 General considerations ........................................................................1872.15.3.4.2 Additional notes on best practice for testing .......................................1882.15.4 Hazard communication for substances and mixtures corrosive to metals ............................1902.15.4.1 Pictograms, signal words, hazard statements and precautionary statements.........1902.15.5 Re-classification of substances and mixtures classified as corrosive to metalsaccording to DSD .................................................................................................................1912.15.5.1 Re-classification of substances and mixtures classified in accordance withDSD.......................................................................................................................1912.15.5.2 Relation to transport classification........................................................................1912.15.6 Examples of classification for substances and mixtures corrosive to metals .......................1912.15.6.1 Example of metal specimen plates after exposure to a corrosive mixture ............1922.15.7 References ............................................................................................................................1923 HEALTH HAZARDS .......................................................................................................................1933.1 ACUTE TOXICITY..........................................................................................................................1933.1.1 Definitions and general considerations for acute toxicity ....................................................1933.1.2 Classification of substances for acute toxicity......................................................................1943.1.2.1 Identification of hazard information .....................................................................1943.1.2.1.1 Identification of human data ...............................................................1943.1.2.1.2 Identification of non-human data ........................................................1943.1.2.2 Classification criteria.............................................................................................1953.1.2.3 Evaluation of hazard information..........................................................................1973.1.2.3.1 Evaluation of human data....................................................................1973.1.2.3.2 Evaluation of non-human data ............................................................1973.1.2.3.3 Weight of evidence..............................................................................2003.1.2.4 Decision on classification......................................................................................2003.1.2.5 Setting of specific concentration limits.................................................................2003.1.2.6 Decision logic........................................................................................................2003.1.3 Classification of mixtures for acute toxicity.........................................................................2023.1.3.1 General considerations for classification ..............................................................2023.1.3.2 Identification of hazard information .....................................................................2023.1.3.3 Classification criteria.............................................................................................2023.1.3.3.1 When data are available for the complete mixture..............................2023.1.3.3.2 When data are not available for the complete mixture: bridgingprinciples .............................................................................................2033.1.3.3.3 When data are available for all components or only for somecomponents .........................................................................................2033.1.3.3.4 When data are not available for all components .................................2053.1.3.3.5 Components that should be taken into account for the purpose ofclassification........................................................................................2083.1.3.4 Generic concentration limits for substances triggering classification ofmixtures.................................................................................................................20813

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.3.5 Decision on classification......................................................................................2093.1.3.6 Decision logic........................................................................................................2103.1.4 Hazard communication in form of labelling for acute toxicity.............................................2113.1.4.1 Pictograms, signal words, hazard statements and precautionary statements.........2113.1.4.2 Additional labelling provisions .............................................................................2133.1.5 Re-classification of substances and mixtures classified for acute toxicity according toDSD and DPD ......................................................................................................................2133.1.5.1 Is direct “translation” of classification and labelling possible?.............................2133.1.5.2 Re-evaluation of data ............................................................................................2143.1.6 Examples of classification for acute toxicity........................................................................2143.1.6.1 Examples of substances fulfilling the criteria for classification............................2143.1.6.1.1 Example 1: Methanol ..........................................................................2143.1.6.1.2 Example 2: N,N-Dimethylaniline........................................................2143.1.6.1.3 Example 3............................................................................................2153.1.6.1.4 Example 4............................................................................................2163.1.6.1.5 Example 5............................................................................................2163.1.6.1.6 Example 6............................................................................................2163.1.6.1.7 Example 7: 2,3-Dichloropropene ........................................................2173.1.6.1.8 Example 8............................................................................................2173.1.6.1.9 Example 9............................................................................................2183.1.6.2 Examples of substances not fulfilling the criteria for classification......................2183.1.6.2.1 Example 10..........................................................................................2183.1.6.3 Examples of mixtures fulfilling the criteria for classification...............................2193.1.6.3.1 Example 11..........................................................................................2193.1.6.3.2 Example 12 a.......................................................................................2203.1.6.4 Examples of mixtures not fulfilling the criteria for classification.........................2213.1.6.4.1 Example 12 b.......................................................................................2213.1.7 References ............................................................................................................................2223.2 SKIN CORROSION/IRRITATION..................................................................................................2223.2.1 Definitions for classification for skin corrosion/irritation ....................................................2223.2.2 Classification of substances for skin corrosion/irritation .....................................................2223.2.2.1 Identification of hazard information .....................................................................2223.2.2.1.1 Identification of human data ...............................................................2223.2.2.1.2 Identification of non human data.........................................................2223.2.2.1.2.1 Consideration of physico-chemical properties .................2233.2.2.1.2.2 Non-testing methods: (Q)SARs and expert systems ........2233.2.2.1.2.3 Testing-methods: pH and acid/alkaline reserve................2233.2.2.1.2.4 Testing methods: in vitro methods ...................................2233.2.2.1.2.5 Testing methods: In vivo data...........................................2253.2.2.2 Classification criteria.............................................................................................2253.2.2.3 Evaluation of hazard information..........................................................................2263.2.2.3.1 Evaluation of human data....................................................................2273.2.2.3.2 Evaluation of non human data.............................................................2273.2.2.3.2.1 In vitro data.......................................................................2273.2.2.3.2.2 In vivo data.......................................................................2273.2.2.3.3 Weight of evidence..............................................................................2283.2.2.4 Decision on classification......................................................................................2293.2.2.5 Setting of specific concentration limits.................................................................2293.2.2.6 Decision logic for classification of substances......................................................2303.2.3 Classification of mixtures for skin corrosion/irritation.........................................................2333.2.3.1 Identification of hazard information .....................................................................2333.2.3.2 Classification criteria.............................................................................................2333.2.3.2.1 When data are available for the complete mixture..............................2333.2.3.2.1.1 Mixtures with extreme pH................................................23314

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.2.3.2.2 When data are not available for the complete mixture: bridgingprinciples .............................................................................................2343.2.3.2.3 When data are available for all components or only for somecomponents .........................................................................................2353.2.3.2.3.1 Components that should be taken into account for thepurpose of classification.........................................................2353.2.3.2.3.2 The additivity approach is applicable...............................2353.2.3.2.3.3 The additivity approach is not applicable.........................2363.2.3.3 Generic concentration limits for substances triggering classification ofmixtures.................................................................................................................2363.2.3.3.1 When the additivity approach is applicable ........................................2363.2.3.3.2 When the additivity approach is not applicable ..................................2373.2.3.4 Decision logic for classification of mixtures.........................................................2373.2.4 Hazard communication in form of labelling for skin corrosion/irritation ............................2393.2.4.1 Pictograms, signal words, hazard statements and precautionary statements.........2393.2.4.2 Additional labelling provisions .............................................................................2403.2.5 Re-classification of substances and mixtures classified for skin corrosion/irritationaccording to DSD and DPD..................................................................................................2413.2.5.1 Is direct “translation” of classification and labelling possible?.............................2413.2.5.2 Re-evaluation of data ............................................................................................2413.2.6 Examples of classification for skin corrosion/irritation........................................................2413.2.6.1 Examples of substances fulfilling the criteria for classification............................2413.2.6.1.1 Example 1: Standard test according to OECD TG 404 with threeanimals ................................................................................................2413.2.6.1.2 Example 2: Test carried out with one animal with a test substancewhich is suspected as corrosive...........................................................2423.2.6.1.3 Example 3a: Test carried out with more than three animals................2423.2.6.1.4 Example 3b: Test carried out with more than three animals................2433.2.6.2 Examples of mixtures fulfilling the criteria for classification...............................2433.2.6.2.1 Example 4............................................................................................2433.2.6.2.2 Example 5............................................................................................2443.2.6.3 Examples of mixtures not fulfilling the criteria for classification.........................2443.2.6.3.1 Example 6............................................................................................2443.2.7 References ............................................................................................................................2453.3 SERIOUS EYE DAMAGE/EYE IRRITATION ..............................................................................2463.3.1 Definitions for classification for serious eye damage/eye irritation .....................................2463.3.2 Classification of substances for serious eye damage/eye irritation ......................................2463.3.2.1 Identification of hazard information .....................................................................2463.3.2.1.1 Identification of human data ...............................................................2463.3.2.1.2 Identification of non human data.........................................................2463.3.2.1.2.1 Consideration of physico-chemical properties .................2463.3.2.1.2.2 Non-testing methods: (Q)SARs and expert systems ........2463.3.2.1.2.3 Testing-methods: pH and the acid/alkaline reserve..........2473.3.2.1.2.4 Testing methods: in vitro methods ...................................2473.3.2.1.2.5 Testing methods: In vivo data...........................................2473.3.2.2 Classification criteria.............................................................................................2483.3.2.3 Evaluation of hazard information..........................................................................2493.3.2.3.1 Evaluation of human data....................................................................2493.3.2.3.2 Evaluation of non-human data ............................................................2493.3.2.3.2.1 In-vitro data ......................................................................2493.3.2.3.2.2 In-vivo data.......................................................................2493.3.2.3.3 Weight of evidence..............................................................................2513.3.2.4 Decision on classification......................................................................................2523.3.2.5 Setting of specific concentration limits.................................................................2523.3.2.6 Decision logic........................................................................................................25415

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012163.3.3 Classification of mixtures for serious eye damage/eye irritation..........................................2553.3.3.1 Identification of hazard information .....................................................................2553.3.3.1.1 Identification of existing human data..................................................2563.3.3.2 Classification criteria.............................................................................................2563.3.3.2.1 When data are available for the complete mixture..............................2563.3.3.2.1.1 Mixtures with extreme pH................................................2573.3.3.2.2 When data are not available for the complete mixture: bridgingprinciples .............................................................................................2583.3.3.2.3 When data are available for all components or only for somecomponents of the mixture..................................................................2583.3.3.2.3.1 Components that should be taken into account for thepurpose of classification.........................................................2583.3.3.2.3.2 The additivity approach is applicable...............................2583.3.3.2.3.3 The additivity approach is not applicable.........................2593.3.3.3 Generic concentration limits for substances triggering classification ofmixtures.................................................................................................................2593.3.3.3.1 When the additivity approach is applicable ........................................2603.3.3.3.2 When the additivity approach is not applicable ..................................2603.3.3.4 Decision logic........................................................................................................2603.3.4 Hazard communication in form of labelling for serious eye damage/eye irritation .............2633.3.4.1 Pictograms, signal words, hazard statements and precautionary statements.........2633.3.5 Re-classification of substances and mixtures classified for serious eye damage/eyeirritation according to DSD and DPD...................................................................................2633.3.5.1 Is direct “translation” of classification and labelling possible?.............................2633.3.5.2 Re-evaluation of data ............................................................................................2643.3.6 Examples of classification for serious eye damage/eye irritation.........................................2643.3.6.1 Examples of substances fulfilling the criteria for classification............................2643.3.6.1.1 Example 1: Standard test according to OECD TG 405 with threeanimals ................................................................................................2643.3.6.1.2 Example 2: Test carried out with more than 3 rabbits.........................266Cornea: 2663.3.6.2 Examples of mixtures fulfilling the criteria for classification...............................2673.3.6.2.1 Example 3: Application of the additivity approach for mixturescontaining ingredients without SCLs ..................................................2673.3.6.2.2 Example 4: Application of the additivity approach for mixturescontaining ingredients which may have SCLs ....................................2683.3.6.2.3 Example 5: Application of the additivity approach for mixturescontaining ingredients which may have SCLs ....................................2683.3.7 References ............................................................................................................................2693.4 RESPIRATORY OR SKIN SENSITISATION ................................................................................2693.4.1 Definitions and general considerations for respiratory or skin sensitisation ........................2693.4.2 Classification of substances for respiratory or skin sensitisation .........................................2703.4.2.1 Identification of hazard information .....................................................................2703.4.2.1.1 Identification of human data ...............................................................2703.4.2.1.2 Identification of non human data.........................................................2703.4.2.2 Classification criteria for substances.....................................................................2713.4.2.3 Evaluation of hazard information..........................................................................2713.4.2.3.1 Human data on respiratory sensitisation .............................................2713.4.2.3.2 Human data on skin sensitisation ........................................................2713.4.2.3.3 Non human data on respiratory sensitisation ......................................2713.4.2.3.4 Non human data on skin sensitisation .................................................2723.4.2.3.4.1 Mouse Local Lymph Node Assay (LLNA, OECD TG429).........................................................................................2723.4.2.3.4.2 Guinea Pig Maximisation Test (GPMT, OECD TG406).........................................................................................273

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.4.2.3.4.3 Buehler occluded patch test (OECD TG 406) ..................2743.4.2.3.4.4 Non-compliant skin sensitisation tests .............................2743.4.2.3.4.5 Animal test methods conducted for purposes otherthan sensitisation ....................................................................2743.4.2.3.5 Weight of evidence..............................................................................2743.4.2.4 Decision on classification......................................................................................2753.4.2.5 Setting of specific concentration limits.................................................................2753.4.2.6 Decision logic for classification of substances......................................................2773.4.3 Classification of mixtures for respiratory or skin sensitisation ............................................2793.4.3.1 General considerations for classification ..............................................................2793.4.3.2 Identification of hazard information for skin sensitisation....................................2793.4.3.3 Classification criteria.............................................................................................2793.4.3.3.1 When data are available for all components or only for somecomponents .........................................................................................2793.4.3.3.2 When data are available for the complete mixture..............................2803.4.3.3.3 When data are not available for the complete mixture: BridgingPrinciples.............................................................................................2813.4.3.4 Decision logic for classification of mixtures.........................................................2823.4.4 Hazard communication for respiratory or skin sensitisation ................................................2843.4.4.1 Pictograms, signal words, hazard statements and precautionary statements.........2843.4.4.2 Additional labelling provisions .............................................................................2843.4.5 Re-classification of substances and mixtures classified for respiratory or skinsensitisation according to DSD and DPD.............................................................................2853.4.5.1 Is direct “translation” of classification and labelling possible?.............................2853.4.5.2 Re-evaluation of the skin sensitisation data ..........................................................2853.4.6 Examples of classification for skin sensitisation ..................................................................2853.4.6.1 Example of substance fulfilling the criteria for classification for skinsensitisation ...........................................................................................................2853.4.6.1.1 Example 1............................................................................................2853.4.6.1.2 Example 2............................................................................................2853.4.6.1.3 Example 3............................................................................................2853.4.6.1.4 Example 4............................................................................................2853.4.6.2 Example of substances or mixtures not fulfilling the criteria for classificationfor skin sensitisation..............................................................................................2863.4.6.2.1 Example 5............................................................................................2863.4.6.2.2 Example 6............................................................................................2863.4.7 References ............................................................................................................................2863.5 GERM CELL MUTAGENICITY.....................................................................................................2873.5.1 Definitions and general considerations for classification for germ cell mutagenicity..........2873.5.2 Classification of substances for germ cell mutagenicity ......................................................2883.5.2.1 Identification of hazard information .....................................................................2883.5.2.1.1 Identification of human data ...............................................................2883.5.2.1.2 Identification of non human data.........................................................2883.5.2.2 Classification criteria for substances.....................................................................2893.5.2.3 Evaluation of hazard information..........................................................................2903.5.2.3.1 Evaluation of human data....................................................................2903.5.2.3.2 Evaluation of non human data.............................................................2903.5.2.4 Decision on classification......................................................................................290Classification as a Category 1B mutagen..........................................................................................2903.5.2.5 Setting of specific concentration limits.................................................................2923.5.2.6 Decision logic for substances................................................................................2933.5.3 Classification of mixtures for germ cell mutagenicity..........................................................2943.5.3.1 Classification criteria for mixtures........................................................................2943.5.3.1.1 When data are available for the complete mixture..............................29417

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012183.5.3.1.2 When data are not available for the complete mixture: bridgingprinciples .............................................................................................2943.5.3.2 Generic concentration limits for substances triggering classification ofmixtures.................................................................................................................2943.5.3.3 Decision logic for mixtures...................................................................................2953.5.4 Hazard communication in form of labelling for germ cell mutagenicity .............................2973.5.4.1 Pictograms, signal words, hazard statements and precautionary statements.........2973.5.4.2 Additional labelling provisions .............................................................................2973.5.5 Re-classification of substances classified for germ cell mutagenicity according to DSDand DPD ...............................................................................................................................2983.6 CARCINOGENICITY ......................................................................................................................2983.6.1 Definitions and general considerations for classification for carcinogenicity......................2983.6.2 Classification of substances for carcinogenicity...................................................................2993.6.2.1 Identification of hazard information .....................................................................2993.6.2.2 Classification criteria for substances.....................................................................2993.6.2.3 Evaluation of hazard information..........................................................................3003.6.2.3.1 Specific considerations for classification............................................3013.6.2.3.2 Additional considerations for classification ........................................302(f) Whether responses are in a single species or several species.......................................................3063.6.2.3.3 Consideration of mutagenicity ............................................................3103.6.2.3.4 Non testing data...................................................................................3103.6.2.4 Decision on classification......................................................................................3113.6.2.5 Setting of specific concentration limits.................................................................3113.6.2.6 Decision logic for classification of substances......................................................3133.6.3 Classification of mixtures for carcinogenicity......................................................................3143.6.3.1 Classification criteria for mixtures........................................................................3143.6.3.1.1 When data are available for all ingredients or only for someingredients ...........................................................................................3143.6.3.1.2 When data are available for the complete mixture..............................3143.6.3.1.3 When data are not available for the complete mixture: bridgingprinciples .............................................................................................3153.6.3.2 Decision logic for classification of mixtures.........................................................3163.6.4 Hazard communication in form of labelling for carcinogenicity..........................................3173.6.4.1 Pictograms, signal words, hazard statements and precautionary statements.........3173.6.4.2 Additional labelling provisions .............................................................................3173.6.5 Re-classification of substances and mixtures classified for carcinogenicity accordingto DSD and DPD ..................................................................................................................3183.6.5.1 Is direct “translation” of classification and labelling possible?.............................3183.6.5.2 Some additional considerations for re-classification.............................................3183.6.6 Examples of classification for carcinogenicity.....................................................................3183.6.7 References ............................................................................................................................3183.7 REPRODUCTIVE TOXICITY.........................................................................................................3213.7.1 Definitions and general considerations for reproductive toxicity.........................................3213.7.1.1 Special considerations on effects on or via lactation ............................................3223.7.2 Classification of substances for reproductive toxicity..........................................................3233.7.2.1 Identification of hazard information .....................................................................3233.7.2.1.1 Identification of human data ...............................................................3233.7.2.1.2 Identification of non human data.........................................................3233.7.2.2 Classification criteria.............................................................................................3233.7.2.2.1 Classification in the presence of parental toxicity...............................3243.7.2.2.1.1 Effects to be considered in the presence of markedsystemic effects ......................................................................3243.7.2.2.1.2 Relevance of specific effects in the parent .......................3253.7.2.2.2 Substances causing effects on or via lactation ....................................326

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.2.3 Evaluation of hazard information..........................................................................3273.7.2.3.1 Use of data from standard repeat dose tests ........................................3283.7.2.3.2 Study design ........................................................................................3283.7.2.3.3 Evaluation of evidence relating to effects on or via lactation .............3283.7.2.4 Decision on classification......................................................................................3303.7.2.5 Setting of specific concentration limits.................................................................3303.7.2.5.3.1 DETERMINATION IN PRACTICE................................................................................3323.7.2.5.3.2 QUANTAL OR NON-PARAMETRIC DATA ................................................................332TABLE 3.7.2.5.1 EXAMPLE OF THE CALCULATION OF THE ED 10 ....................................................3333.7.2.5.3.3 CONTINUOUS OR PARAMETRIC DATA ...................................................................333TABLE 3.7.2.5.2 EXAMPLE ON THE CALCULATION OF THE ED 10 ....................................................3333.7.2.5.3.4 DATA COMBINING INCIDENCE AND MAGNITUDE..............................................334NON-ORAL STUDIES..................................................................................................................................334HUMAN DATA.............................................................................................................................................3353.7.2.5.5 MODIFYING FACTORS..................................................................................................3353.7.2.5.5.1 TYPE OF EFFECT / SEVERITY.....................................................................................3363.7.2.5.5.2 DATA AVAILABILITY ....................................................................................................3363.7.2.5.5.3 DOSE-RESPONSE RELATIONSHIP .............................................................................337THE ED 10 WILL IN MOST CASES PROBABLY BE IN THE SAME RANGE AS THE NOAELAND LOAEL. HOWEVER, IN CASES OF A SHALLOW DOSE EFFECT RELATIONSHIPCURVE, THE LOAEL MAY SOMETIMES BE CLEARLY BELOW THE ED 10 . IN SUCHSITUATIONS, IF A SUBSTANCE WOULD FALL INTO A LOWER POTENCY GROUP BASEDON THE ED 10 BUT INTO A HIGHER POTENCY GROUP BASED ON THE LOAEL THEN THEHIGHER POTENCY GROUP SHOULD BE USED FOR THAT SUBSTANCE........................................3373.7.2.5.5.4 MODE OR MECHANISM OF ACTION.........................................................................3373.7.2.5.5.5 TOXICOKINETICS ..........................................................................................................3373.7.2.5.5.6 BIO-ACCUMULATION OF SUBSTANCES..................................................................338TABLE 3.7.2.5.5 SCLS FOR SUBSTANCES IN EACH POTENCY GROUP ANDCLASSIFICATION CATEGORY.................................................................................................................3393.7.2.6 Decision logic........................................................................................................3503.7.3 Classification of mixtures for reproductive toxicity.............................................................3513.7.3.1 Classification criteria.............................................................................................3513.7.3.1.1 When data are available for the individual ingredients.......................3513.7.3.1.2 When data are available for the complete mixture..............................3523.7.3.1.3 When data are not available for the complete mixture: bridgingprinciples .............................................................................................3523.7.3.2 Decision logic........................................................................................................3533.7.4 Hazard communication in form of labelling for reproductive toxicity.................................3553.7.4.1 Pictograms, signal words, hazard statements and precautionary statements.........3553.7.4.2 Additional labelling provisions .............................................................................35719

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012203.7.5 Re-classification of substances and mixtures classified for reproductive toxicityaccording to DSD and DPD..................................................................................................3573.7.5.1 Is direct “translation” of classification and labelling possible?.............................3573.8 SPECIFIC TARGET ORGAN TOXICITY – SINGLE EXPOSURE (STOT-SE)...........................3583.8.1 Definitions and general considerations for STOT-SE ..........................................................3583.8.2 Classification of substances for STOT-SE ...........................................................................3593.8.2.1 Identification of hazard information .....................................................................3593.8.2.1.1 Identification of human data ...............................................................3593.8.2.1.2 Identification of non human data.........................................................3593.8.2.2 Classification criteria for Categories 1 and 2 ........................................................3613.8.2.2.1 Guidance values ..................................................................................3623.8.2.3 Classification criteria for Category 3: Transient target organ effects ...................3623.8.2.4 Evaluation of hazard information on STOT-SE for substances ............................3643.8.2.4.1 Evaluation of human data....................................................................3643.8.2.4.2 Evaluation of non human data.............................................................3663.8.2.4.3 Evaluation of non-testing and in vitro data .........................................3683.8.2.4.4 Conversions.........................................................................................3683.8.2.4.5 Weight of evidence..............................................................................3683.8.2.5 Decision on classification of substances ...............................................................3693.8.2.6 Setting of specific concentration limits for STOT-SE ..........................................3693.8.2.7 Decision logic........................................................................................................3723.8.3 Classification of mixtures for STOT-SE ..............................................................................3743.8.3.1 Identification of hazard information .....................................................................3743.8.3.2 Classification criteria for mixtures........................................................................3743.8.3.2.1 When data are available for the complete mixture..............................3743.8.3.2.2 When data are not available for the complete mixture: bridgingprinciples .............................................................................................3743.8.3.2.3 When data are available for all components or only for somecomponents of the mixture..................................................................3743.8.3.2.4 Components of a mixture that should be taken into account forthe purpose of classification................................................................3753.8.3.3 Generic concentration limits for substances triggering classification ofmixtures for STOT-SE ..........................................................................................375Categories 1 and 2 ................................................................................................................3753.8.3.4 Decision logic for mixtures...................................................................................3763.8.4 Hazard communication in form of labelling for STOT-SE ..................................................3803.8.4.1 Pictograms, signal words, hazard statements and precautionary statements.........3803.8.4.2 Additional labelling provisions .............................................................................3813.8.5 Re-classification of substances and mixtures classified for STOT-SE according toDSD and DPD ......................................................................................................................3813.8.5.1 Is direct “translation” of Classification and Labelling possible for STOT-SEsubstances?............................................................................................................3813.8.5.2 Re-evaluation of the STOT-SE data......................................................................3823.8.6 Examples of classification for STOT-SE .............................................................................3833.8.6.1 Examples of substances fulfilling the criteria for classification............................3833.8.6.1.1 Example 1: Methanol ..........................................................................3833.8.6.1.2 Example 2: Tricresyl phosphate..........................................................3843.8.6.1.3 Example 3: Sulfur dioxide...................................................................3843.8.6.1.4 Example 4: Toluene ............................................................................3843.8.6.2 Examples of substances not fulfilling the criteria for classification......................3853.8.6.2.1 Example 5: ABC .................................................................................3853.8.6.2.2 Example 6: N,N-Dimethylaniline........................................................3853.9 SPECIFIC TARGET ORGAN TOXICITY – REPEATED EXPOSURE (STOT-RE) ....................3863.9.1 Definitions and general considerations for STOT-RE..........................................................386

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9.2 Classification of substances for STOT-RE...........................................................................3873.9.2.1 Identification of hazard information .....................................................................3873.9.2.1.1 Identification of human data ...............................................................3873.9.2.1.2 Identification of non human data.........................................................3873.9.2.2 Classification criteria for substances.....................................................................3883.9.2.3 Evaluation of hazard information..........................................................................3913.9.2.3.1 Evaluation of human data....................................................................3913.9.2.3.2 Evaluation of non human data.............................................................392Figure 3.9.2.3.2 Comparison between the NOAEL and the ED versus the guidance values ........3923.9.2.3.3 Conversions.........................................................................................394Table 3.9.2.4.2(a) Food conversion...................................................................................................3943.9.2.3.4 Weight of evidence..............................................................................3953.9.2.4 Decision on classification......................................................................................3953.9.2.5 Additional considerations......................................................................................3973.9.2.5.1 Irritating/corrosive substances.............................................................3973.9.2.5.2 Hematotoxicity....................................................................................3973.9.2.5.3 Mechanisms not relevant to humans (CLP Annex I, 3.9.2.8.1. (e))....3993.9.2.5.4 Adaptive responses (CLP Annex I, 3.9.2.8.1. (d)) ..............................4003.9.2.5.5 Post-observation periods in 28 day and 90 day studies.......................4013.9.2.6 Setting of specific concentration limits.................................................................4013.9.2.7 Decision logic for classification of substances......................................................4033.9.3 Classification of mixtures for STOT-RE..............................................................................4043.9.3.1 Identification of hazard information .....................................................................4043.9.3.2 Classification criteria for mixtures........................................................................4043.9.3.3 When data are available for the complete mixture................................................4043.9.3.3.1 When data are not available for the complete mixture: bridgingprinciples .............................................................................................4043.9.3.3.2 When data are available for all components or only for somecomponents of the mixture..................................................................4043.9.3.3.3 Components of a mixture that should be taken into account forthe purpose of classification................................................................4053.9.3.4 Generic concentration limits for substances triggering classification ofmixtures.................................................................................................................4053.9.3.5 Decision logic for mixtures...................................................................................4063.9.4 Hazard communication in form of labelling for STOT RE ..................................................4073.9.4.1 Pictograms, signal words, hazard statements and precautionary statements.........4073.9.4.2 Additional labelling provisions .............................................................................4083.9.5 Re-classification of substances and mixtures classified for STOT-RE according toDSD and DPD ......................................................................................................................4083.9.5.1 Is direct “translation” of classification and labelling possible for STOT-REsubstances?............................................................................................................4083.9.5.2 Re-evaluation of the STOT-RE data .....................................................................4093.9.6 Examples of classification for STOT-RE .............................................................................4093.9.6.1 Examples of substances fulfilling the criteria for classification............................4093.9.6.1.1 Example 1: Hydroxylamine / Hydroxylamonium salts (CAS no.7803-49-8)...........................................................................................4093.9.6.1.2 Example 2: But-2-yn-1,4-diol (EC No 203-788-6; CAS No 110-65-6) ....................................................................................................4113.9.6.1.3 Example 3: XYZ .................................................................................4133.9.6.2 Examples of substances not fulfilling the criteria for classification......................4153.9.6.2.1 Example 4: MCCPs (Medium Chain Chlorinated Paraffins) =Alkanes, C 14-17 , Chloro- (EC No 287-477-0; CAS No 85535-85-9) .........................................................................................................4153.9.6.3 Examples of mixtures fulfilling the criteria for classification...............................4163.9.6.3.1 Example 5: ..........................................................................................4163.9.6.3.2 Example 6............................................................................................41621

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9.6.3.3 Example 7............................................................................................4173.9.6.3.4 Example 8............................................................................................4173.9.6.4 Example of mixtures not fulfilling the criteria for classification ..........................4183.9.6.4.1 Example 9............................................................................................4183.9.7 References ............................................................................................................................4184 PART 4: ENVIRONMENTAL HAZARDS .....................................................................................4194.1 HAZARDOUS TO THE AQUATIC ENVIRONMENT ..................................................................4194.1.1 Introduction ..........................................................................................................................4194.1.2 Scope ....................................................................................................................................4194.1.3 Classification of substances hazardous to the aquatic environment .....................................4204.1.3.1 Information applicable for classification of substances hazardous to theaquatic environment ..............................................................................................4204.1.3.1.1 Substance properties used for classification........................................4204.1.3.1.2 Information and data availability ........................................................4204.1.3.2 Evaluation of available information......................................................................4214.1.3.2.1 General considerations ........................................................................4214.1.3.2.2 Substances difficult to test...................................................................4214.1.3.2.3 Interpretation of data for aquatic toxicity, degradation andbioaccumulation ..................................................................................4234.1.3.2.3.1 Aquatic toxicity ................................................................4234.1.3.2.3.2 Degradation ......................................................................4244.1.3.2.3.3 Bioaccumulation...............................................................4274.1.3.2.4 Using weight of evidence in evaluations in the context of C&L ........4294.1.3.2.4.1 General aspects of weight of evidence .............................4294.1.3.2.4.2 Guidance on WoE for data deficient substances ..............4294.1.3.2.4.3 Guidance on WoE for substances for which morethan one valid piece of data is available for a given dataelement ...................................................................................4304.1.3.2.4.4 Outliers .............................................................................4304.1.3.2.4.5 Weight of evidence in degradation...................................4314.1.3.2.4.6 Weight of evidence in bioaccumulation ...........................4314.1.3.3 Classification categories and criteria.....................................................................4314.1.3.3.1 Outline of the core classification system.............................................4314.1.3.3.2 The “safety net”...................................................................................4364.1.3.3.3 Setting an M-factor for highly toxic substances..................................4374.1.3.4 Decision on classification: examples for substances.............................................4384.1.3.4.1 Example A: Hydrophilic substance, straightforward classificationbased on acute and chronic toxicity data.............................................4404.1.3.4.2 Example B: Hydrophilic substance, straightforward classificationbased on acute data, no chronic data available....................................4434.1.3.4.3 Example C: Moderately water soluble substance, straightforwardclassification based on acute data, chronic data available for twotrophic levels only; combined set of QSAR data and experimentaldata ......................................................................................................4464.1.3.4.4 Example D: Substance with several toxicity data for a trophiclevel .....................................................................................................4494.1.3.4.5 Example E: “Safety net” classification category Chronic 4................4524.1.3.4.6 Example F: Substance difficult to test, toxicity above level ofwater solubility....................................................................................4544.1.4 Classification of mixtures hazardous to the aquatic environment ........................................4574.1.4.1 General considerations for classification of mixtures hazardous to the aquaticenvironment...........................................................................................................4574.1.4.2 Information requirements......................................................................................45922

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4.3 Classification criteria for mixtures hazardous to the aquatic environmentbased on test data on the mixture as a whole ........................................................4604.1.4.4 When experimental aquatic toxicity data are not available for the completemixture: bridging principles..................................................................................4614.1.4.5 When hazard data (information on toxicity or classification) are available forall the components of the mixture .........................................................................4624.1.4.6 When hazard data (information on toxicity or classification) are available foronly some components of the mixture ..................................................................4664.1.4.7 Decision on classification: examples for mixtures................................................4664.1.4.7.1 Example A: When classification data are available for some or allcomponents of a mixture: straightforward application of thesummation method ..............................................................................4684.1.4.7.2 Example B1: When toxicity data on the mixture as a whole isavailable for all three trophic levels: classification based on testdata for the mixture .............................................................................4704.1.4.7.3 Example B2: When information on the classification of thecomponents is available and toxicity data on the mixture as awhole is available for some, but not all three trophic levels: useof the summation method....................................................................4724.1.4.7.4 Example C: When no data are available on the mixture as a wholeand its components, but test data are available on a similar testedmixture: use of the bridging principles – dilution with water .............4744.1.4.7.5 Example D: When test data are available for some, but not allcomponents of the mixture: use of the additivity formula and ofthe summation method ........................................................................4764.1.5 Metal and metal compounds.................................................................................................4804.1.6 Hazard communication for hazards to the aquatic environment ..........................................4804.1.7 Re-classification of substances and mixtures classified as hazardous to the aquaticenvironment according to DSD/DPD ...................................................................................4824.1.8 References ............................................................................................................................483PART 5: ADDITIONAL HAZARDS............................................................................................................4845.1 HAZARDOUS TO THE OZONE LAYER.......................................................................................484ANNEXES .....................................................................................................................................................485IANNEX I: AQUATIC TOXICITY...................................................................................................485I.1 Introduction ..........................................................................................................................485I.2 Description of tests ...............................................................................................................485I.2.1 Fish tests................................................................................................................486I.2.1.1 Acute testing........................................................................................486I.2.1.2 Chronic testing ....................................................................................486I.2.2 Tests with Crustaceae............................................................................................486I.2.2.1 Acute testing........................................................................................486I.2.2.2 Chronic testing ....................................................................................486I.2.3 Algae / other aquatic plant tests ............................................................................487I.2.3.1 Tests with algae...................................................................................487I.2.3.2 Tests with aquatic macrophytes ..........................................................487I.3 Aquatic toxicity concepts .....................................................................................................487I.3.1 Acute toxicity ........................................................................................................487I.3.2 Chronic toxicity.....................................................................................................488I.3.3 Exposure regimes ..................................................................................................489I.3.4 Test media for algae and Lemna ...........................................................................489I.3.5 Use of substance categorisation (read across and grouping) and (Q)SARs forclassification and labelling....................................................................................49023

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012I.4 Substances which are difficult to test ...................................................................................490I.4.1 Unstable substances...............................................................................................490I.4.2 Poorly soluble substances......................................................................................491I.4.3 Other factors contributing to concentration loss ...................................................492I.4.4 Perturbation of the test media................................................................................492I.4.5 Complex substances ..............................................................................................493I.5 References ............................................................................................................................493II ANNEX II: RAPID DEGRADATION .............................................................................................494II.1 Introduction ..........................................................................................................................494II.2 Interpretation of degradability data ......................................................................................494II.2.1 Ready biodegradability .........................................................................................495II.2.1.1 Concentration of test substance...........................................................495II.2.1.2 Time window.......................................................................................495II.2.2 BOD 5 /COD............................................................................................................496II.2.3 Other convincing scientific evidence ....................................................................496II.2.3.1 Aquatic simulation tests ......................................................................496II.2.3.2 Field investigations .............................................................................497II.2.3.3 Monitoring data...................................................................................497II.2.3.4 Inherent and Enhanced Ready Biodegradability tests.........................497II.2.3.5 Sewage treatment plant simulation tests .............................................497II.2.3.6 Soil and sediment degradation data.....................................................498II.2.3.7 Anaerobic degradation data.................................................................498II.2.3.8 Hydrolysis ...........................................................................................498II.2.3.9 Photochemical degradation .................................................................498II.2.3.10 Estimation of degradation ...................................................................499II.2.3.11 Volatilisation .......................................................................................499II.2.4 No degradation data available...............................................................................499II.3 General interpretation problems ...........................................................................................499II.3.1 Complex substances ..............................................................................................499II.3.2 Availability of the substance.................................................................................500II.3.3 Test duration less than 28 days .............................................................................500II.3.4 Primary biodegradation.........................................................................................500II.3.5 Conflicting results from screening tests ................................................................501II.3.6 Variation in simulation test data............................................................................501II.4 Decision scheme ...................................................................................................................502II.5 References ............................................................................................................................503III24ANNEX III: BIOACCUMULATION...............................................................................................504III.1 Introduction ..........................................................................................................................504III.2 Interpretation of bioconcentration data.................................................................................504III.2.1 Bioconcentration factor (BCF)..............................................................................505III.2.1.1 BCF in different test species ...............................................................506III.2.1.2 Use of radio-labelled substances .........................................................506III.2.2 Octanol-water-partitioning coefficient (K ow ) ........................................................507III.2.2.1 Experimental determination of K ow .....................................................507III.2.2.2 Use of QSARs for determination of log K ow .......................................507III.3 Chemical classes that need special attention with respect to BCF and K ow values ..............508III.3.1 Substances difficult to test.....................................................................................508III.3.2 Poorly soluble and complex substances ................................................................509III.3.3 High molecular weight substances........................................................................509III.3.4 Surface-active substances (surfactants).................................................................509III.3.4.1 Octanol-water-partition coefficient (K ow ) ...........................................510III.4 Conflicting data and lack of data..........................................................................................510III.4.1 Conflicting BCF data ............................................................................................510III.4.2 Conflicting log K ow data........................................................................................510

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012III.4.3 Expert judgement ..................................................................................................510III.5 Decision scheme ...................................................................................................................511III.6 References ............................................................................................................................511IV ANNEX IV: METALS AND INORGANIC METAL COMPOUNDS..................................................513IV.1 Introduction ..............................................................................................................................513IV.2 Application of aquatic toxicity data and solubility data for classification ...............................515IV.2.1 Interpretation of aquatic toxicity data.......................................................................515IV.2.2 Interpretation of solubility data ................................................................................517IV.2.2.1 Assessment of existing data ...................................................................517IV.2.2.2 Screening T/D test for assessing solubility of metal compounds...........518IV.2.2.3 Full T/D test for assessing solubility of metals and metalcompounds ..........................................................................................518IV.2.3 Comparison of aquatic toxicity data and solubility data ..........................................518IV.3 Assessment of environmental transformation ..........................................................................519IV.4 Bioaccumulation.......................................................................................................................519IV.5 Classification strategies for metals and metal compounds.......................................................520IV.5.1 Introduction ..............................................................................................................520IV.5.2 Classification strategies for metals...........................................................................521IV.5.2.1 Classification strategy for determining acute aquatic hazard formetals ..................................................................................................521IV.5.2.2 Classification strategy for determining long-term aquatic hazard formetals ..................................................................................................522IV.5.2.2.1 Approach based on available chronic toxicity referencedata .........................................................................................522IV.5.2.2.2 The surrogate approach.........................................................523IV.5.3 Classification strategies for metal compounds .........................................................525IV.5.3.1 Classification strategies for determining acute aquatic hazard formetal compounds.................................................................................526IV.5.3.2 Classification strategy for determining long-term aquatic hazard formetal compounds.................................................................................527IV.5.3.2.1 Approach based on available chronic toxicity referencedata .........................................................................................527IV.5.3.2.2 The surrogate approach.........................................................529IV.5.4 Setting M-factors for metals and inorganic metal compounds.................................532IV.5.5 Particle size and surface area.....................................................................................533IV.5.6 Classification of mixtures of metals and metal compounds .....................................534IV.5.6.1 Classification of alloys and complex metal containing materials ..........535IV.6 References ................................................................................................................................535IV.7 Decision on classification: examples for metals and metal compounds ..................................536Example A: Soluble metal compound with acute and chronic toxicity dataand no evidence of rapid environmental transformation (Me 2(SO4) 2 ). ...............................................................................................538Example B: Poorly soluble metal compound with acute and chronic toxicitydata, transformation/dissolution data at 7 days (low loading rate)and at 28 days (only low and medium loading rates) and noevidence of rapid environmental transformation.................................541Example C: Metal in powder and massive form with acute and chronictoxicity data and Transformation/Dissolution data at 7 days (low,medium and high loading rates) and at 28 days (only the highloading rate) and no evidence of rapid environmentaltransformation .....................................................................................545Explanatory note to Example C - Critical Surface Area (CSA)approach .................................................................................55025

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Example D: Hazard classification of a soluble metal salt: the case of rapidenvironmental transformation through speciation in the watercolumn.................................................................................................552VVIANNEX V: COLLECTION OF INTERNET LINKS FOR THE USERS OF THEGUIDANCE ......................................................................................................................................555ANNEX VI: BACKGROUND DOCUMENT TO THE GUIDANCE FOR SETTINGSPECIFIC CONCENTRATION LIMITS FOR SUBSTANCES CLASSIFIED FORREPRODUCTIVE TOXICITY ACCORDING TO REGULATION (EC) NO 1272/2008..............5561 Executive summary ..............................................................................................................5562 Introduction ..........................................................................................................................5572.1 General description of the classification system for reprotoxic substancesand mixtures ..........................................................................................................5572.2 Description of the process for the development of a method to set SCLs forreproductive toxic substances................................................................................5582.3 Considering potency in setting specific concentration limits for varioushealth hazards........................................................................................................5592.4 Parameters for potency for reproductive toxicity..................................................5603 Modifying factors .................................................................................................................5644 Non-modifying factors .........................................................................................................5654. 1 Species and strains ................................................................................................5654. 2 Systemic or maternal toxicity................................................................................5654. 3 Mutagenicity .........................................................................................................5654.4 Volatility ...............................................................................................................5655 Potency groups and specific concentration limits.................................................................5665.1 Justification of the proposed potency boundaries and specific concentrationlimits......................................................................................................................5665.2 Assigning SCLs.....................................................................................................5726 References ............................................................................................................................57226

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012LIST OF ABBREVIATIONSADNADRANEASTMATEBAMBCFBCOPBfRBfR DSSBMFBPbwC&LCAcATpECLPCNSCSACSRAccord européen relatif au transport international des marchandises dangereuses parvoie de navigation intérieure (European Agreement concerning the InternationalCarriage of Dangerous Goods by Inland Waterways) 3Accord européen relatif au transport international des marchandises dangereuses parroute (European Agreement concerning the International Carriage of Dangerous Goodsby Road) 4Ammonium Nitrate EmulsionAmerican Society for the Testing of MaterialsAcute Toxicity EstimateBundesanstalt für Materialforschung und -prüfung (Federal Institute for MaterialsResearch and Testing)Bioconcentration FactorBovine Corneal Opacity and Permeability testGerman Federal Institute for Risk AssessmentDecision support system by the German Federal Institute for Risk AssessmentBiomagnification factorBoiling pointBody weightClassification and LabellingCompetent AuthorityConverted Acute Toxicity point EstimateRegulation (EC) No 1272/2008 on classification, labelling and packaging ofsubstances and mixtures 5Central Nervous SystemChemical Safety AssessmentChemical Safety ReportDINStandard of the German Institute for StandardisationDNA Deoxyribonucleic AcidDOC Dissolved Organic CarbonDPD Directive 1999/45/EC on the classification and labelling of Dangerous Preparations 6DSD Directive 67/548/EEC on the classification and labelling of Dangerous Substances 73 European Agreement concerning the International Carriage of Dangerous Goods by Inland Waterways,concluded at Geneva on 26 May 2000, as amended4 European Agreement concerning the International Carriage of Dangerous Goods by Road, concluded at Genevaon 30 September 1957, as amended5 Regulation (EC) No 1272/2008 of the European Parliament and Council of 16 December 2008 onclassification, labelling and packaging of substances and mixtures, amending and repealing Directives67/548/EEC and 1999/45/EC and amending Regulation (EC) No 1907/2006 [OJ L 353, 31.12.2008, p. 1]6 Directive 1999/45/EC of the European Parliament and of the Council of 31 May 1999 concerning theapproximation of the laws, regulations and administrative provisions of the Member States relating to theclassification, packaging and labelling of dangerous preparations [OJ L 200, 30.7.1999, p. 1]27

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012EC3ECBECHAECVAMEDEffective Concentration inducting a stimulation index of 3 in the LLNA testEuropean Chemicals BureauThe formerly known European Chemicals Bureau (ECB) was part of the Institute forHealth and Consumer Protection (IHCP), which is one of the seven scientific institutesin the European Commission's Joint Research Centre (JRC). Its mission was to providescientific and technical support to the conception, development, implementation andmonitoring of EU policies on chemicals and consumer products.(http://ecb.jrc.ec.europa.eu/)European Chemicals Agency, Helsinki (http://echa.europa.eu/home_en.asp)European Centre for the Validation of Alternative Methods (http://ecvam.jrc.it/)Effective DoseERVEcotoxicity Reference ValueESAC ECVAM Scientific Advisory Committee (http://ecvam.jrc.it/)f/FFemaleFPFlash pointGCLGeneral Concentration LimitsGHS Globally Harmonised System of Classification and Labelling of Chemicals 8GJICGLPGnRHGPMTGVHbHET-CAMHSHSMHtIARCIATA(DGR)IBCICAO TIICEIECIMDGGap junction intercellular communicationGood Laboratory PracticeGonadotropin-releasing hormoneGuinea Pig Maximisation TestGuidance ValueHaemoglobinHen's Egg Test on Chorio-allantoic MembraneHazard statementHuman skin modelHematocritInternational Agency for Research on Cancer (http://www.iarc.fr/)International Air Transport Association (Dangerous Goods Regulations Manual)Intermediate Bulk ContainerInternational Civil Aviation Organization (Technical Instructions for the SafeTransport of Dangerous Goods by Air)Isolated Chicken EyeInternational Electrotechnical Commission (http://www.iec.ch/)International Maritime Dangerous Goods Code7 Council Directive 67/548/EEC of 27 June 1967 on the approximation of laws, regulations and administrativeprovisions relating to the classification, packaging and labelling of dangerous substances [OJ 196, 16.8.1967, p.1]8 Globally Harmonised System of Classification and Labelling of Chemicals (GHS), Second revised edition,United Nations New York and Geneva, 200728

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012INS Guidance on Identification and Naming of Substances under REACH, ECHA, 2007(http://guidance.echa.europa.eu/docs/guidance_document/substance_id_en.pdf)IPCS International Programme on Chemical Safety (joint programme of WHO, ILO andUNEP)IR/CSAIREISOGuidance on Information Requirements and Chemical Safety Assessment, ECHA,2008(http://guidance.echa.europa.eu/docs/guidance_document/information_requirements_en.htm)Isolated Rabbit EyeInternational Standards OrganisationITDG Directive 2008/68 on the Inland Transport of Dangerous Goods 9ITSLD 50 /LC 50LLNALO (A) EL/CLVETm/MMetHBMetHbMPMSCAMTDMWn.a.NCNENO(A)ECNO(A)ELODSODPOECDOECD TGIntegrated Testing StrategyMedian (50%) lethal dose/concentrationLocal Lymph Node AssayLowest Observed (Adverse) Effect Level/ConcentrationLow Volume Eye TestMaleMethaemoglobinaemiaMethaemoglobinMelting PointMember State Competent AuthorityMaximal Tolerated DoseMolecular weightNot availableNo ClassificationNarcotic effect(s)No Observed (Adverse) Effect ConcentrationNo Observed (Adverse) Effect LevelOzone Depleting SubstancesOzone Depleting PotentialOrganisation for Economic Co-operation and DevelopmentOECD Test GuidelineThe OECD Guidelines for the Testing of Chemicals are a collection of the mostrelevant internationally agreed test methods used by government, industry andindependent laboratories to determine the safety of chemicals and chemical mixtures,including pesticides and industrial chemicals. All Test Guidelines are available at the9 Directive 2008/68/EC of the European Parliament and of the Council of 24 September 2008 on the inlandtransport of dangerous goods, implementing the European Agreement concerning the International Carriageof Dangerous Goods by Road (ADR), the Regulations concerning the International Carriage of DangerousGoods by Rail (RID) and the European Agreement concerning the International Carriage of Dangerous Goodsby Inland Waterways (ADN) [OJ L 260, 30.9.2008, p. 13]29

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012OPP statement(or PS)PB/PKPCPPARαPS (or Pstatement)(Q)SARREACHRIDRIPRTDGRTISADTSCEGHS (orUNSCEGHS)SCETDG (orUNSCETDG)SCLSDSSIFTSSDSTOT-SESTOT-RESVCT25OECD homepage:http://www.oecd.org/document/40/0,3343,en_2649_34377_37051368_1_1_1_1,00.htmlOxidising PowerPrecautionary statementPhysiologically-based pharmacokineticPhysico-chemicalPeroxisome proliferator-activated receptor-alphaPrecautionary statement(Quantitative) Structure Activity RelationshipRegulation (EC) No 1907/2006 of the European Parliament and of the Councilconcerning the Registration, Evaluation, Authorisation and Restriction of Chemicals 10Règlement concernant le transport international ferroviaire de marchandisesdangereuses (Regulations concerning the International Carriage of Dangerous Goodsby Rail) 11REACH Implementation ProjectRegulations on the Transport of Dangerous Goods. Generic term that covers all modaltransport regulations (ADR, RID, ADN, IMDG and ITDG)Respiratory tract irritationSelf-Accelerating Decomposition TemperatureSub-Committee of Experts on the Globally Harmonised System(http://www.unece.org/trans/danger/publi/ghs/ghs_welcome_e.html)Sub-Committee of Experts on the Transport of Dangerous Goods(http://www.unece.org/trans/danger/danger.htm)Specific Concentration LimitSafety Data SheetSkin integrity function testSpecies Sensitivity DistributionSpecific Target Organ Toxicity - Single ExposureSpecific Target Organ Toxicity - Repeated ExposureSaturated Vapour ConcentrationThe daily dose (in mg/kg bodyweight/day) inducing a tumour incidence of25 % upon lifetime exposure10 Regulation (EC) No 1907/2006 of the European Parliament and of the Council concerning the Registration,Evaluation, Authorisation and Restriction of Chemicals (REACH), establishing a European Chemicals Agency,amending Directive 1999/45/EC and repealing Council Regulation (EEC) No 793/93 and CommissionRegulation (EC) No 1488/94 as well as Council Directive 76/769/EEC and omission of Directives 91/155/EEC,93/67/EEC, 93/105/EC and 2000/21/EC. [OJ L 396, 30.12.2006 p.1.] [Corrigendum: OJ L 136, 29.5.2007 p.3]11 Regulations concerning the International Carriage of Dangerous Goods by Rail, appearing as Appendix C tothe Convention concerning International Carriage by Rail (COTIF) concluded at Vilnius on 3 June 1999, asamended30

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012T95 Inhalation chamber equilibrium (attained at the time t95)T/DT/DpTERTGTGDTMTest MethodsRegulationTOPKATUDPUDPGUGTUNUN-MTCUNSCEGHS(or SCEGHS)UNSCETDG(or SCETDG)US-FHSAVDIUVCBVPWAFWoEWSFTransformation/DissolutionTransformation/Dissolution ProtocolTranscutaneous electrical resistanceTest GuidelineTechnical Guidance DocumentTest Method as listed in the Test Methods RegulationRegulation (EC) No 440/2008 laying down test methods pursuant to the REACHRegulation 12Mathematical (Q)SAR model for prediction of skin corrosion/irritationUridine 5'-diphosphateUridine diphosphate glucuronylUDP-glucuronyltransferaseUnited NationsUnited Nations (2003). Manual of Tests and Criteria. ST/SG/AC.10/11/Rev. 4, asamended: Fourth revised edition of the Manual of Tests and Criteria, containingcriteria, test methods and procedures to be used for classification of dangerous goodsaccording to the provisions of Parts 2 and 3 of the United Nations Recommendationson the Transport of Dangerous Goods, Model Regulations, as well as of chemicalspresenting physical hazards according to the Globally Harmonized System ofClassification and Labelling of Chemicals(http://www.unece.org/trans/danger/publi/manual/manual_e.html).United Nations SubCommittee of Experts on the Globally Harmonised System(http://www.unece.org/trans/danger/publi/ghs/ghs_welcome_e.html)United Nations SubCommittee of Experts on the Transport of Dangerous Goods(http://www.unece.org/trans/danger/danger.htm)United States Federal Hazardous Substance Act - 40 Code of Federal Regulations1500.41Verein Deutscher Ingenieure (The Association of German Engineers)Substances of unknown or variable composition, complex reaction products orbiological materialsVapour PressureWater Accommodated FractionWeight of EvidenceWater soluble fractionIn this document text cited from Regulation (EC) No 1272/2008 is indicated in green boxes.12 Council Regulation (EC) No 440/2008 of 30 May 2008 laying down test methods pursuant to Regulation (EC)No 1907/2006 of the European Parliament and of the Council on the Registration, Evaluation, Authorisation andRestriction of Chemicals (REACH) [OJ L 142, 31.5.2008, p. 1] [Corrigendum: OJ L 143, 3.6.2008, p. 55]31

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121 PART 1: GENERAL PRINCIPLES FOR CLASSIFICATION ANDLABELLING1.1 INTRODUCTION1.1.1 The objective of the guidance documentThis document is a comprehensive technical and scientific guidance on the application ofRegulation (EC) No 1272/2008 on the classification, labelling and packaging of substancesand mixtures 13 , hereafter referred to as CLP.CLP amends the Dangerous Substance Directive 67/548/EEC 14 (DSD), the DangerousPreparations Directive 1999/45/EC 15 (DPD) and Regulation (EC) No 1907/2006 16 (REACH),and will replace DSD and DPD from 1 June 2015 (CLP Article 61). CLP is based on the 3 rdrevision of the United Nations’ Globally Harmonised System of Classification and Labellingof Chemicals (UN GHS) and is implementing the provisions of the GHS within the EU,without lowering the protection of human health and the environment, compared to theclassification, labelling and packaging system in DSD and DPD.A core principle of CLP is “self-classification” of a substance or mixture by the manufacturer,importer or downstream user (CLP Article 4(3) and Recital 17), which involves identificationof its hazards followed by classification as a result of the comparison of the hazardinformation with the criteria in CLP. This guidance will enable industry to self-classifychemicals and to provide appropriate hazard communication information to the targetpopulations potentially exposed. For substances of particular concern (carcinogens, mutagens,substances toxic for reproduction (CMRs) and respiratory sensitisers) or for other substanceswhere EU-wide action is needed, CLP sets out a system for formal harmonisation ofclassifications at EU level.Given that many provisions under REACH are linked to classification, the implementation ofREACH and CLP is interlinked and should be planned and applied in tandem. Further adviceon the implementation of CLP is available in the Agency 17 ’s Introductory Guidance on theCLP Regulation, available at ECHA website(http://echa.europa.eu/documents/10162/13562/clp_introductory_en.pdf ).The objective of this document is to provide detailed guidance on the application of the CLPcriteria for physical, health and environmental hazards.13 Regulation (EC) No 1272/2008 of the European Parliament and of the Council on classification, labelling andpackaging of substances and mixtures, amending and repealing Directives 67/548/EEC and 1999/45/EC, andamending Regulation (EC) No 1907/2006 [ OJ L 353, 31.12.2008, p. 1]14 Council Directive 67/548/EEC relating to the classification, packaging and labelling of dangerous substances,as amended [OJ 196, 16.8.1967, p. 1]15Directive 1999/45/EC as of 30 July 2002 of the European Parliament and of the Councilrelating to the classification, packaging and labelling of dangerous preparation, as amended[OJ L 200, 30.7.1999, p.1]16 Regulation (EC) No 1907/2006 of the European Parliament and of the Council concerning the Registration,Evaluation, Authorisation and Restriction of Chemicals (REACH), establishing a European Chemicals Agency,amending Directive 1999/45/EC and repealing Council Regulation (EEC) No 793/93 and CommissionRegulation (EC) No 1488/94 as well as Council Directive 76/769/EEC and omission of Directives 91/155/EEC,93/67/EEC, 93/105/EC and 2000/21/EC. [OJ L 396, 30.12.2006 p.1.] [Corrigendum: OJ L 136, 29.5.2007 p.3]17 'the Agency' means the European Chemicals Agency established by Regulation (EC) No 1907/2006(REACH).32

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121.1.2 BackgroundThe aim of classification and labelling is to identify the hazardous properties of a substance ora mixture by applying specific criteria to the available hazard data (classification), and then toprovide any appropriate hazard labelling and information on safety measures.The EU has had a comprehensive system for the classification and labelling of dangeroussubstances and mixtures for over 40 years, mainly DSD and DPD. In addition, the Safety DataSheet (SDS) Directive 91/155/EEC 18 required suppliers to provide more detailed informationfor professional users. These directives contributed to a single market in chemicals in the EU,based on a high level of protection of human health and the environment.The GHS was developed worldwide to minimise differences between systems of differentjurisdictions for classification and labelling of substances and mixtures. The GHS aims tocontribute towards global efforts to provide protection from hazardous effects of chemicalsand to facilitate trade.The GHS criteria for classifying hazardous substances were developed taking into accountexisting systems for hazard classification, such as the EU supply and use system, theCanadian and US Pesticide systems, GESAMP 19 hazard evaluation procedure, IMO 20 Schemefor Marine Pollutants, the European Road and Rail Transport Scheme (RID/ADR), and theUS Land Transport. These systems include supply and subsequent use of chemicals, the seatransport of chemical substances as well as transport of chemical substances by road and rail.The harmonised criteria are therefore intended to identify hazardous chemicals in a commonway for use throughout all these systems.The GHS provides a basis for an internationally uniform information system on hazardoussubstances and mixtures. It provides harmonised criteria for classification and hazardcommunication measures for different target audiences, including consumers, workers andemergency responders, and in transport. It follows a “building block” approach to enablejurisdictions to adopt the system according to the needs of their law and the various targetaudiences.The GHS was agreed by the UN Committee of Experts on the Transport of Dangerous Goodsand the Globally Harmonized System of Classification and Labelling of Chemicals(CETDG/GHS). It was formally approved by the UN Economic and Social Council (UNECOSOC) in July 2003 and published further in 2003 after a decade of negotiations. It isupdated biannually.1.1.3 Hazard classificationHazard classification is a process involving identification of the physical, health andenvironmental hazards of a substance or a mixture, followed by comparison of those hazards(including degree of hazard) with defined criteria in order to arrive at a classification of thesubstance or mixture. Under CLP, a manufacturer, importer or downstream user will apply thefollowing three steps to arrive at a self-classification of a substance or a mixture:18Council Directive 91/155/EEC relating to defining and laying down the detailedarrangements for the system of specific information relating to dangerous preparations anddangerous substances, as amended [OJ L 076, 22.03.1991, p. 35], repealed and replaced byRegulation (EC) No 1907/2006 as of 1 June 2007.19 Group of Experts on the Scientific Aspects of Marine Environmental Protection20 International Maritime Organisation33

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012 identification and examination of relevant available information regarding thepotential hazards of a substance or mixture; comparison of the information (data) with the classification criteria; and decision on whether the substance or mixture shall be classified as hazardous inrelation to the hazard classes or differentiations provided in CLP Annex I, and thedegree of hazard, where appropriate.Preliminary information on identification and review of relevant data is provided in section1.1.6 of this guidance document, while further guidance is provided in Part B of the ECHAGuidance document on Information Requirements and Chemical Safety Assessment(Chapters R.2 to R.4, IR/CSA), available on the ECHA Website(http://echa.europa.eu/web/guest/guidance-documents/guidance-on-information-requirementsand-chemical-safety-assessment).Classification according to CLP is based on intrinsic hazards, i.e. the basic properties of asubstance as determined in standard tests or by other means designed to identify hazards. AsCLP is hazard-based, it does not take exposure into consideration in arriving at either aclassification or appropriate labelling, unless for specific exceptions when a chemical can beconsidered as not being biologically available, such as the derogation not to label a metal inthe massive form.1.1.4 Who is responsible for the hazard classification and what is the timetableCLP and REACH places the responsibility for hazard classification and related provisionssuch as packaging, hazard communication and SDS on the suppliers of substances andmixtures.From 1 December 2010 to 1 June 2015 (CLP Article 61):Substances shall be classified in accordance with both DSD Directive and CLP Regulationin order to allow these classifications to be used in the classifications of mixtures.Classification and labelling information in accordance with both systems shall be includedin SDS (see the Guidance on the compilation of Safety Data Sheets, available on theAgency’s website). Labelling and packaging shall be in accordance with CLP Regulation.Mixtures shall be classified, labelled and packaged in accordance with DPD. They mayalso be classified, labelled and packaged in accordance with CLP. In that case they shallnot be labelled and packaged according to DPD. When a mixture is classified, labelled andpackaged according to CLP, classification and labelling information according to bothsystems shall be provided in SDS (see the Guidance on the compilation of Safety DataSheets, available on the Agency’s website).From 1 June 2015 (CLP Articles 60 and 61):Both substances and mixtures shall be classified, labelled and packaged in accordance withCLP. DSD and DPD are repealed from 1 June 2015 and classification according to thesedirectives is not allowed.However, substances classified, labelled and packaged in accordance with DSD and alreadyplaced on the market (“on the shelves”) before 1 December 2010, and mixtures classified,labelled and packaged in accordance with DPD and already placed on the market (“on theshelves”) before 1 June 2015, do not have to be relabelled and repackaged in accordance withCLP until 1 December 2012 and 1 June 2017, respectively.34

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121.1.5 Which substances and mixtures should be classified (the scope)Substances and mixtures placed on the market fall within the scope of classification underCLP and should be evaluated in order to reach a decision as to whether they should beclassified or not. Substances are also subject to classification where they are subject toregistration or notification under REACH, even if they are not placed on the market.However, a number of substances and mixtures are exempted from the requirements of theCLP Regulation as a whole (CLP Article 1):– radioactive substances and mixtures (Directive 96/29/Euroatom 21 );– substances and mixtures which are subject to customs supervision, provided that they donot undergo any treatment or processing, and which are in temporary storage, or in a freezone or free warehouse with a view to re-exportation, or in transit;– non-isolated intermediates;– substances and mixtures used in scientific experimentation, analysis or chemical research,provided they are not placed on the market and they are used under controlled conditionsin accordance with EU workplace and environmental legislation;– waste, as defined in Directive 2006/12/EC 22 ; and– certain substances or mixtures in the finished state, intended for the final user:• medicinal products, as defined in Directive 2001/83/EC 23 ,• veterinary medicinal products, as defined in Directive 2001/82/EC 24 ,• cosmetic products, as defined in Directive 76/768/EEC 25 ,• medical devices as defined in Directive 90/385/EEC 26 (active implantablemedical devices) and 93/42/EEC 27 (medical devices in general), which areinvasive or used in direct physical contact with the human body, and in vitrodiagnostic medical devices (Directive 98/79/EC 28 ), and• food or feeding stuffs as defined in Regulation 178/2002 29 , including whenthey are used as food additives within the scope of Directive 89/107/EEC 30 , as21 Council Directive 96/29/Euratom of 13 May 1996 laying down basic safety standards for the protection of thehealth of workers and the general public against the dangers arising from ionizing radiation [OJ L 159,29.6.1996, p. 1]22 Directive 2006/12/EC of the European Parliament and of the Council of 5 April 2006 on waste [OJ L 114,27.4.2006, p. 9]23 Directive 2001/83/EC of the European Parliament and of the Council of 6 November 2001 on the Communitycode relating to medicinal products for human use [OJ L 311, 28.11.2001, p. 67]24 Directive 2001/82/EC of the European Parliament and of the Council of 6 November 2001 on the Communitycode relating to veterinary medicinal products [OJ L 311, 28.11.2001, p. 1]25 Council Directive 76/768/EEC of 27 July 1976 on the approximation of the laws of the Member States relatingto cosmetic products [OJ L 262, 27.9.1976, p. 169]26 Council Directive 90/385/EEC of 20 June 1990 on the approximation of the laws of the Member Statesrelating to active implantable medical devices [OJ L 189, 20.7.1990, p. 17]27 Council Directive 93/42/EEC of 14 June 1993 concerning medical devices [OJ L 169, 12.7.1993, p. 1]28 Directive 98/79/EC of the European Parliament and of the Council of 27 October 1998 on in vitro diagnosticmedical devices [OJ L 331, 7.12.1998, p. 1]29 Regulation (EC) No 178/2002 of the European Parliament and of the Council of 28 January 2002 laying downthe general principles and requirements of food law, establishing the European Food Safety Authority and layingdown procedures in matters of food safety [OJ L 31, 1.2.2002, p. 1]35

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012a flavouring in foodstuffs within the scope of Directive 88/388/EEC andDecision 1999/217/EC 31 , as an additive in feeding stuffs within the scope ofRegulation (EC) 1831/2003 32 , and in animal nutrition within the scope ofDirective 82/471/EEC 33 .In addition, Member States may exempt certain substances or mixtures in specific caseswhere necessary for the purpose of national defence.Although CLP does not apply to the transport of dangerous goods by air, sea, road, rail orinland waterways (CLP Article 1(6)), the criteria for classification are normally intended to bethe same in the two systems. Thus, a substance or mixture classified in a hazard class which iscommon to both CLP and the transport legislation will normally be classified the same in bothsystems. However, the transport classifications do not include all of the GHS categories, sothe absence of a transport classification does not mean the substance or mixture should not beclassified under CLP.1.1.6 What information is needed for classification1.1.6.1 Information for the classification of substancesThe classification of a substance is based on the relevant information available on itshazardous properties. This information can include experimental data generated in tests forphysical hazards, toxicological and ecotoxicological tests, historical human data such asaccident records or epidemiological studies, or information generated in in vitro tests,(Quantitative) Structure Activity Relationships ((Q)SAR), “read across”, or categoryapproaches.CLP does not require new testing for the purpose of classification for health or environmentalhazards; testing for physical hazards is required unless adequate and reliable information isalready available (CLP Article 8(2)). However, a substance or mixture placed on the marketfor research and development (R&D) purposes may have been manufactured or imported inquantities that are too small to perform physical hazard testing. In these cases it would not beproportionate to request the respective manufacturer, importer or downstream user to performthe tests required in Part 2 of Annex I to CLP.Although data may be provided through the application of REACH, it should be recognisedthat the data set required by REACH (particularly at lower tonnages) will not necessarilyenable the comparison with the criteria for all hazard classes. Information may also beavailable from other EU legislation for which there are specific requirements for test data tobe generated, such as legislation on plant protection products (Regulation (EC) No1107/2009 34 and Directive 91/414/EEC 35 ) and on biocidal products (Directive 98/8/EC 36 ), or30 Council Directive 89/107/EEC of 21 December 1988 on the approximation of the laws of the Member Statesconcerning food additives authorized for use in foodstuffs intended for human consumption [OJ L 40, 11.2.1989,p. 27]31 1999/217/EC: Commission Decision of 23 February 1999 adopting a register of flavouring substances used inor on foodstuffs drawn up in application of Regulation (EC) No 2232/96 of the European Parliament and of theCouncil of 28 October 1996 [OJ L 84, 27.3.1999, p. 1]32 Regulation (EC) No 1831/2003 of the European Parliament and of the Council of 22 September 2003 onadditives for use in animal nutrition [OJ L 268, 18.10.2003, p. 29]33 Council Directive 82/471/EEC of 30 June 1982 concerning certain products used in animal nutrition [OJ L213, 21.7.1982, p. 8]34 Regulation (EC) No 1107/2009 of the European Parliament and of the Council of 21 October 2009 concerning the placingof plant protection products on the market repeals Council Directives 79/117/EEC and 91/414/EEC with effect from 14 June2011. However Article 80 of Regulation (EC) No 1107/2009 specifies that directive 91/414/EEC shall continue to apply withrespect to active substances included in Annex I to that Directive for certain transitional periods.36

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012from various non-EU programmes. Finally, the supplier may decide to conduct new testing inorder to fill data gaps, provided that he has exhausted all other means of generatinginformation. Testing on animals must be avoided wherever possible and alternative methods(including in vitro testing, the use of (Q)SARs, read-across and/or category approaches) mustalways be considered first, provided they are scientifically validated, sufficiently adequate andreliable.If, for the purpose of CLP, it is required or decided to generate new data, certain test methodsand quality conditions must be met. Studies must be conducted in accordance with the EU testmethods (Regulation 440/2008) 37 or other international test methods validated according tointernational procedures such as those of the OECD. For physical hazards new tests shall becarried out (at least from January 2014) in compliance with relevant recognised qualitysystem or by laboratories complying with a relevant recognised standard, and for health andenvironmental hazards in compliance with the principles of Good Laboratory Practice (GLP).Animal tests must comply with the Directive 86/609/EEC 38 . Tests on non-human primates areprohibited for the purposes of CLP. Tests on humans shall not be performed for the purposeof CLP. However, existing data obtained from other sources, such as accident records andepidemiological and clinical studies, can be used.1.1.6.2 Information relevant for the classification of mixturesFor mixtures, classification for physical hazards should normally be based on the results oftests carried out on the mixtures themselves.When considering health and environmental hazards, the classification should preferably bebased on available information (including test data) on the mixture itself, except whenclassifying for e.g. CMR effects or for the evaluation in relation to the bioaccumulation anddegradation properties within the ‘hazardous to the aquatic environment’ hazard class referredto in sections 4.1.2.8 and 4.1.2.9 of Annex I to CLP. In these cases classification of themixtures shall be based on the information on the substances.If no in vivo test data are available on a mixture, such data should normally not be generated;rather, all available information on the ingredients of the mixture should be used to derive aclassification. Only when the manufacturer, importer or downstream user has exhausted allother means of generating information, new tests may be performed.Annex I to CLP specifies “bridging principles” which enables suppliers to derive health orenvironmental classifications of their mixtures based on available data on similar testedmixtures and on the ingredient substances. It also provides specific rules for the classificationof mixtures based on the classification of the individual substances in the mixture.1.1.7 Data evaluation and reaching a decision on classification1.1.7.1 Classification of substances35 Council Directive 91/414/EEC of 15 July 1991 concerning the placing of plant protection products on themarket, as amended [OJ L 230, 19.8.91, p. 1]36 Directive 98/8/EC of the European Parliament and of the Council of 16 February 1998 concerning the placingof biocidal products on the market, as amended [OJ L 123, 24.4.98, p. 1]37 Council Regulation (EC) No 440/2008 laying down test methods pursuant to Regulation (EC) No 1907/2006of the European Parliament and of the Council concerning the Registration, Evaluation, Authorisation andRestriction of Chemicals (REACH)[OJ L 142, 31.5.2008, p. 1]38 Directive 86/609/EEC regarding the protection of animals used for experimental and other scientific purposes,[OJ L 358, 18.12.1986, p. 1]37

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012After the available information has been assembled, a systematic evaluation of thisinformation is necessary in order to derive a classification. The information must be comparedwith the criteria for classification for each hazard class or differentiation within the hazardclass. Differentiation is a distinction depending on the route of exposure or the nature of theeffects. A decision should be made as to whether the substance meets the criteria forclassification. When this is the case; the classifier should assign one or more hazardcategories for each relevant hazard class or differentiation. The substance is then assigned theappropriate hazard communication elements.In some cases the classification decision may be straightforward, requiring only an evaluationof whether the substance gave a positive or negative result in a specific test that can bedirectly compared with the classification criteria. In other cases, scientific judgements must bemade (e.g. on dose/response relationships, equivocal results and non-standardised tests).Expert judgement may therefore be needed to decide whether the results of a particular testmeet the criteria laid down in Annex I.1.1.7.2 Influence of impurities, additives or individual constituents on theclassification of a substanceSubstances may contain impurities, additives, or other constituents while still meeting thesubstance definition in CLP. This applies to both mono-constituent, multi-constituent (e.g.reaction masses) and UVCB substances. The classification of such impurities, additives orindividual constituents may influence the classification of the substance, in addition to theother hazardous properties.1.1.8 Updating of hazard classificationsUpdating of classifications may be necessary, if new information is obtained or if the criteriain CLP are amended. When manufacturers, importers or downstream users become aware ofnew information or an amendment to CLP or when a change is introduced in a mixture, theymust reconsider the classification of the substance or mixture (but note that a downstreamuser can rely on the classification from his supplier, provided he shares the new informationwith that supplier to allow him to meet the requirements).1.1.9 The interface between hazard classification and hazard communicationIn addition to SDS, CLP provides an integrated system of hazard communication elements(hazard pictograms, signal words, hazard statements and precautionary statements) on thelabel. Provision of this information to the end user is obligatory, irrespective of conditions ofuse and risk. While the Chemical Safety Assessment (CSA) on a particular substanceperformed for the purpose of REACH may indicate "safe use", a situation resulting inunforeseen exposure may occur, such as in an accident. In such a situation, workers, managersand emergency personnel will need information on the hazard profile of the substance, whichwill be provided by the label and the SDS. These sources of information will also provideuseful information to the worker on the safe handling of the chemical.It is recognised that the hazard communication needs of the various end users may differ.Consumers are primarily dependent on the label of a substance or a mixture as a source ofhazard and precautionary information, while the requirement for provision of an SDS isprimarily applicable to professional users. Thus, the label facilitates communication of keyhazard information and additional safety advice (precautionary statements) to consumers of asubstance or a mixture.38

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121.1.10 The interface between self-classification and harmonised classification, andthe list of harmonised classificationsCLP places emphasis on self-classification by industry of the substances or mixtures theysupply. In some cases, substances are subject to harmonised classification at EU level, whilemixtures must always be self-classified, except for pesticidal and biocidal products where theMember State competent authorities (MSCAs) decide on the classification as part of thenational authorisation scheme (CLP Article 36(2)).If a substance has a harmonised classification as provided in Annex VI to CLP, thisclassification must always be used by a manufacturer, importer or downstream user, exceptfor so-called minimum classifications listed in Table 3.1 that may be amended in accordancewith section 1.2.1 of Annex VI. Where some but not all hazard classes or differentiationswithin hazard classes have been harmonised, the remainder should to be self-classified tocomplete the classification.Harmonised classification normally applies to those properties of the highest concern (CMRand respiratory sensitisation) and may also apply for other properties if there is a need for aEU-level action. Decisions on harmonised classification are taken by the EuropeanCommission through comitology (CLP Article 37(5)), following a proposal submitted to theAgency and an opinion of the Agency's Risk Assessment Committee (RAC) (CLP Article37(4)).Substances regulated under the Biocidal Products Directive 98/8/EC 39 or under the PlantProtection Products Regulation (EC) No 1107/2009 will normally be subject to harmonisedclassification and labelling for all hazardous properties. These proposals for harmonisedclassification and labelling are prepared by MSCAs only (CLP Article 36(2)). However, ingeneral proposals for harmonised classification for a particular substance to be added toAnnex VI to CLP can be made by both MSCAs and by manufacturers, importers anddownstream users (CLP Article 37). Only MSCAs can propose a revision of an existingharmonised classification and labelling (CLP Article 37(6)).Harmonised classification and labelling of a substance provides for a high level of protectionof health and the environment, and provides legal clarity for suppliers of the same substanceof high concern (i.e. manufacturers of substances, importers of substances or mixtures,producers of specific articles, downstream users (including manufacturers of mixtures) anddistributors).Part 3 of Annex VI to CLP contains the list of harmonised classifications. All harmonisedclassifications previously adopted under DSD and listed in Annex I to DSD were carried overto the list of harmonised classifications in Annex VI to CLP, table 3.2, also including theNotes assigned to the entries as referred to in the DSD. This was done to maintain the samelevel of protection under CLP as under DSD. The harmonisation of classification ofsubstances is a continuous work building on all efforts already done within the EU so far toevaluate hazards of substances that caused concern.Under DSD, as a rule all hazards for which data were available were evaluated for asubstance. While it was in general the objective to obtain a complete (harmonised)classification, some substances (such as complex coal- and oil-derived substances) wereexempted. Under CLP, the harmonised classification and labelling of substances shall benormally partial and cover only hazard classes of particular concern listed in Article 36(1)CLP (i.e. respiratory sensitisation, germ cell mutagenicity, carcinogenicity and reproductive39 Directive 98/8/EC of the European Parliament and of the Council of 16 February 1998 concerning the placingof biocidal products on the market, as amended [OJ L 123, 24.4.98, p. 1]39

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012toxicity). A substance that is used as an active substance in the meaning of Directives91/414/EEC and 98/8/EC shall normally be subjected to harmonised classification andlabelling (Article 36(2) CLP). Where a substance fulfils the criteria for hazard classes otherthan those referred in Article 36(1) CLP and does not fall under Directives 91/414/EEC and98/8/EC, a harmonised classification and labelling may be added to Annex VI to CLP on acase-by-case basis, if justification is provided demonstrating the need for action at EU level(Article 36(3) CLP). This means that self-classification should be done for non-harmonisedhazard classes, according to CLP Article 4(3) and CLP Recital 17.1.1.11 The Classification and Labelling Inventory (C&L Inventory)Manufacturers and importers are required to notify the Agency of the classification andlabelling of hazardous substance(s) placed on the market and of substances which are placedon the market and subject to registration in accordance with the REACH Regulation. TheAgency will then include the information in a classification and labelling inventory in form ofa database. Substances placed on the market on or after 1 December 2010 require notificationwithin one month after their placing on the market. There is no need to notify the substance ifthe same information has already been submitted as part of a registration under REACH bythe same actor, as the classification and labelling, when part of the registration package, willautomatically be added to the C&L Inventory (CLP Article 40(1)). Further guidance on whatshould be included in a notification and how to do it is available on the ECHA websitehttp://echa.europa.eu/web/guest/regulations/clp/cl-inventory/notification-to-the-cl-inventory .The Agency shall make certain information from the C&L Inventory publicly available on itswebsite, including the substance name, the classification, labelling and any relevant specificconcentration limit or M-factor(s). It will be indicated if there is a harmonised classificationfor the entry, or if it is an agreed entry between manufacturers or importers. While multiplenotifications of the same substance may be made by different manufacturers or importers,with the potential for differences in the classifications notified, over time this should providethe stimulus for suppliers to liaise in order to agree on a single entry.1.1.12 Relation of classification to other EU legislationA network of EU legislation relies on classification in one way or the other (see section 23 ofthe Introductory Guidance on the CLP Regulation for a detailed list of the laws concerned).This downstream legislation includes laws protecting consumers and workers, as well as ruleson biocides, pesticides and waste. Therefore, the consequences of classification are greaterthan just a hazard label or an SDS in that it also has a direct effect on the management ofassociated risks.1.1.12.1 REACHClassification plays a key role in REACH; it must be included in the registration dossier for asubstance and it triggers certain provisions such as the performance of an exposureassessment and risk characterisation as part of the CSA and the obligation to provide an SDS.Classification of a substance as mutagenic, carcinogenic or toxic to reproduction (CMR) mayalso lead to restrictions and the need to apply for authorisations ((EC) No 1907/2006).1.1.12.2 Plant Protection Products and BiocidesActive substances as well as any plant protection or biocidal products containing them shallbe classified in accordance with the CLP Regulation by the applicable deadlines. On the otherhand, and pursuant to Recital 47 of the CLP Regulation, Directive 91/414/EEC on plantprotection products and Directive 98/8/EC on biocidal products “should remain fully40

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012applicable to any product within their scope.” For example, there are separate provisions forlabelling and for updating labels for such substances and mixtures in these acts, and theirsuppliers must apply these provisions instead of the CLP rules, see e.g. CLP Article 30(3).It should be noted that with effect from 14 June 2011, Directive 91/414/EEC has beenrepealed by Regulation (EC) 1107/2009. This means that references to the repealed Directiveshall now be construed as references to the new Regulation. Nevertheless, Article 80 of thenew Regulation specifies that Directive 91/414/EEC shall continue to apply with respect toactive substances included in Annex I to that Directive for certain transitional periods.Furthermore, it specifies that products labelled in accordance with Article 16 of Directive91/414/EEC may continue to be placed on the market until 14 June 2015.In relation to classification, the new Regulation brings about some changes, e.g. certainclassifications (e. g. CMR, Cat. 1 A and 1B) may now preclude approval of the respectivesubstance as an active substance, safener, or synergist in plant protection products.1.1.12.3 Transport legislationMany of the GHS criteria (by hazard class) are already implemented through the UN ModelRegulations for Transport of Dangerous Goods and related legal instruments (ADR, RID,ADN, IMDG Code and ICAO TI).Available transport classifications can be a source of information for the classification andlabelling of substances and mixtures under CLP, especially for physical hazards, see alsosection 1.7 of this document.1.2 THE SIGNIFICANCE OF THE TERMS 'FORM OR PHYSICAL STATE’AND 'REASONABLY EXPECTED USE’ WITH RESPECT TOCLASSIFICATION ACCORDING TO CLP1.2.1 'Form or physical state’ and 'reasonably expected use’CLP refers to the terms 'form or physical state’ and 'reasonably expected use’ in the followingArticles:Article 5 (1)The information shall relate to the forms or physical states in which the substance is placed on themarket and in which it can reasonably be expected to be used.Article 6 (1)The information shall relate to the forms or physical states in which the mixture is placed on themarket and, when relevant, in which it can reasonably be expected to be used.Article 8 (6)Tests that are carried out for the purposes of this Regulation shall be carried out on the substance oron the mixture in the form(s) or physical state(s) in which the substance or mixture is placed on themarket and in which it can reasonably be expected to be used.The object of hazard classification is to identify the intrinsic physical, health andenvironmental hazards of substances and mixtures taking into account all uses that can bereasonably expected.In this context, the intention of the UN GHS should be kept in mind:41

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012“1.3.2.2.1 The GHS uses the term “hazard classification” to indicate that only the intrinsichazardous properties of substances or mixtures are considered.1.3.2.2.2 Hazard classification incorporates … identification of relevant data regarding thehazards of a substance or mixture …”The following guidance is intended to clarify the references to 'reasonably expected use' and'form or physical state' in this context.1.2.2 The term 'reasonably expected use’ in relation to hazard classificationHazard classification is based on intrinsic properties of the substance and does not take intoaccount exposure. Reasonably expected use summarises all physical forms and states of asubstance or mixture that may occur during intended use or reasonably foreseeable conditionsof misuse.Reasonably expected use of a substance is as follows: Any process, including production, handling, maintenance, storage, transport ordisposal. All technical operations/manufacturing activities like e.g. spraying, filing, and sawing Any putative consumer contact through e.g. do-it-yourself or household chemicals. All professional and non-professional uses including reasonably foreseeable misuse,but not abuse such as criminal or suicidal uses.Reasonably expected use is also related to any consumer disposal or any work in which asubstance or mixture is used, or intended to be used irrespective of its present limited use oruse pattern. Thus, use should not be mixed up with usage category.1.2.3 The term ‘form or physical state’ in relation to hazard classificationDepending on different prerequisites, form or physical state is taken into account differentlyin the practice of testing and classification for physical, health, and environmental hazardswhich is described in the following paragraphs.1.2.3.1 Physical hazardsDifferent forms or physical states of a substance or mixture may result in different physicalproperties and hazards with possible consequences for the hazard classification of a substanceor mixture. Putative forms comprise properties such as crystal structure, particle size,homogeneity (e.g. emulsions) and texture (e.g. viscosity or tablet form). Examples of physicalstate factors are: surface treatment (e.g. coating), state of aggregation, moisture content,residual solvent, activation or stabilisation.The classification of a substance or mixture relates to the tested form and physical state. If theform and / or physical state is changed it has to be evaluated whether this might affect theclassification and whether re-testing is necessary. For example, a hazardous phase separationmay occur due to a temperature change under conditions of storage, or a solid substance maybe molten to bring it into the liquid phase (e.g. for pumping).General considerationsThe form of a substance or mixture as placed on the market might be such that it is notpossible to test it in this form, e.g. if it is in the form of tablets or pellets. In suchcircumstances, the physical hazards of the substance or mixture shall be considered for42

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012classification especially if they are friable and produce secondary effects due to abrasion orcrushing during supply and use. If phase separation does occur, the hazardous properties ofthe most hazardous phase of the substance or mixture shall be communicated.The test sample should in any case be representative for the substance or mixture placed onthe market. This is especially important in case of small 'batch' production. Mixtures mightfor example contain inert components which, if they are over-represented in the test sample,will lead to incorrect hazard classification.Specific requirements of certain test methodsSome test methods for the classification of physical hazards have specific requirementsregarding the form / particle size of the sample to be tested. In these cases, the specificrequirements of the test methods prevail. Examples of tests which have specific requirementsregarding the form/particle size of the sample to be tested include those used to determine theclassification of explosives and of substances which in contact with water emit flammablegases.In other test methods, there are no specific requirements regarding the particle size but it isstated explicitly that the particle size may have a significant effect on the test result.Therefore, these properties should be mentioned in the test report (i.e. testing of oxidisingsolids). Moreover, particle size is crucial for several other classes such as explosives,flammable solids, self-reactive substances, pyrophoric solids, self-heating substances, solidorganic peroxides and substances which, in contact with water, emit flammable gases.1.2.3.2 Human health hazardsAlso for human health, different forms (e.g. particle sizes, coating) or physical states mayresult in different hazardous properties of a substance or mixture in use. However, due to testcomplexity, not every form or physical state can be tested for each health hazard. In general,testing should be performed on the smallest available particle size and the default approach isto test for different routes of exposure (oral, dermal, inhalation). Again, due to testcomplexity, mostly the data for only one exposure route are available.In general, the assumption is made that the testing conditions of valid animal assays reflectthe hazards to man and these data shall be used for classification. Moreover, it is assumed thatclassification for human health hazards takes into account all the potential hazards which arelikely to be faced for all forms or physical states in which the substance is placed on themarket and can reasonably be expected to be used. It is assumed that it comprises putativeaccidental exposures. This approach generally, but not necessarily comprehensively, coversthe whole range of intrinsic properties of a substance or mixture: in some cases, substances ormixtures have to be transformed into specific forms not mirroring ‘real-life’ exposures inorder that an animal test can be performed. As a consequence, the results of such tests mayhave to be evaluated taking into account any limitations due to the fact that the specific formof the tested substance or mixture does not or not perfectly represent that to which humanexposure may occur during intended, known, or reasonably expected use. Such evaluation hasto be performed according to the state of the scientific and technical knowledge. The burdenof proof is on the person placing a substance or mixture on the market.1.2.3.3 Environmental hazardsThe environmental hazard classification is principally concerned with the aquaticenvironment and the basis of the identification of hazard is the aquatic toxicity of thesubstance or mixture, and information on the degradation and bioaccumulation behaviour.43

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The system of classification is designed to ensure that a single classification applies to asubstance. In general it takes no account of the specific form since this can vary and is notintrinsic to the substance. The form in which the substance is placed on the market is takeninto account when deciding what label to apply and various derogations from labelling exist,e.g. the metals in the massive form. In the massive form the hazard may not be present andthe substance need not be labelled. The SDS will, however, indicate the classification andintrinsic hazardous properties to warn the user that subsequent transformation of the substancemay produce the hazardous form.For aquatic hazard classification, organic substances are generally tested in the dissolvedform. Exceptions to this approach include complex, multi-component substances and metalsand their compounds. Examples of alternative approaches include the use of WaterAccommodated Fractions (WAF) for complex, multi-component substances where thetoxicity cut-off is related to the loading, and a test strategy for metals and their compounds inwhich the specific form (i.e. particle size) used for testing is standardised and forms orphysical states are not further taken into account.1.3 SPECIFIC CASES REQUIRING FURTHER EVALUATION – LACK OFBIOAVAILABILITY1.3.1 DefinitionBioavailability is the rate and extent to which a substance can be taken up by an organism andis available for metabolism or interaction with biologically significant receptors.Bioavailability (biological availability) involves both release from a medium (if present) andabsorption by an organism (IPCS 2004).1.3.2 Bioavailability44Article 12Specific cases requiring further evaluationWhere, as a result of the evaluation carried out pursuant to Article 9, the following properties oreffects are identified, manufacturers, importers and downstream users shall take them into accountfor the purposes of classification:[…](b) conclusive scientific experimental data show that the substance or mixture is notbiologically available and those data have been ascertained to be adequate and reliable;[…]In general, bioavailability is not explicitly evaluated in hazard classification – the observationof systemic toxicity implicitly demonstrates a degree of bioavailability. On the other hand,when no toxicity is demonstrated in a test, this may be a result of either lack of intrinsictoxicity of the substance or lack of bioavailability in the test system employed. Nevertheless,as indicated in Article 12 (b) of CLP there may be cases where a specific evaluation ofbioavailability is warranted.In general terms, for a substance or mixture to have an effect on a biological or environmentalsystem, there must be some degree of bioavailability. Therefore, it follows that a substance ormixture need not be classified when it can be shown by conclusive experimental data frominternationally acceptable test methods, e.g. from Council Regulation (EC) No 440/2008, thatthe substance or mixture is not biologically available (UN GHS 1.3.2.4.5.1). A non

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012bioavailable substance may, however, react with the media to transform to soluble availableforms. The rate and extent at which this process, known as “transformation” for the purposesof the classification guidance, takes place can vary extensively between different substances,and can be an important factor in determining the appropriate hazard category (see Annex IV,section IV.1of this document).When considering the non-bioavailability of a mixture, the evaluation should be based on datafor all relevant ingredients of the mixture. Further, one should consider potential interactionof the ingredients that could influence the bioavailability of the mixture as such or one of itscomponents.Bioavailability considerations are only relevant with respect to classification for health and orenvironmental hazards and not for physical hazards.1.3.2.1 Human health hazardsThe assumption is that all substances and mixtures are considered to be bioavailable to someextent. However, there are a few specific cases in which bioavailability may have an influenceon hazard classification. For instance in the case of some metals and polymers, the nature ofthe physical form (metals in solid form) and the molecular size (polymers are very largemolecules), or their physico-chemical properties may limit absorption. Where a supplierproposes derogation from hazard classification on the basis of bioavailability, he has toprovide adequate and robust data to support the conclusion of lack of bioavailability. It ispossible that a substance is bioavailable by one route but not another (e.g. absorbed followinginhalation but not absorbed through the skin). In such cases the lack of bioavailability mayderogate classification for the relevant route.Information on relative bioavailability (e.g. relative amounts of absorption) within a relatedgroup/category of chemicals can be of some use in classification. It is possible thatconsideration of bioavailability data in a semi-quantitative manner would lead to theclassification for the same hazard class but in a different category on the grounds that theextent of bioavailability would be reflected in the relative potency. In general, a prediction oflower bioavailability must be supported by robust evidence and a weight of evidencedetermination using expert judgment shall be applied.Information on bioavailability is usually obtained from adequate, reliable, and conclusivetoxicokinetic studies for all relevant routes of exposure and all relevant forms or physicalstates where the substance and/or metabolite(s) of the substance have been quantified in bodyfluids and/or target organs. It should be noted that concluding that there is lack of or reducedbioavailability has a high burden of evidence and needs to be supported by robust data andexpert evaluation.Bioavailability of a substance or a mixture is normally assumed if there are in vitro studiesavailable which show the solubility of a substance or mixture in body fluids or artificialsimulated body fluids. Furthermore, conclusions on bioavailability of a substance or a mixturemay be based on considerations of the physical properties of a substance or derived fromStructural Activity Relationships (SAR). In certain exceptional circumstances it may bepossible that a substance on its own or in a mixture can be considered to be non-bioavailable,based on either appropriate in vitro data, e.g. from skin absorption models, SARconsiderations or considering the physical properties of a substance, if the respectiverequirements described above have been taken into account in an adequate analysis.45

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121.3.2.2 Environmental hazardsThe hazard classification for the aquatic environment is based on the three elements aquatictoxicity, bioaccumulation and degradation. The measurement of toxicity to aquatic organismsand its use within a hazard classification system introduces a number of compoundingproblems. The substance is not dosed directly into the organism but rather into water in whichthe organism lives. While this reflects more accurately the manner in which the organism willreceive the dose in the environment, it does not allow the direct control of the dose which isan important part of much mammalian toxicity testing. The dose is limited by thebioavailability of the substance, the maximum dose being determined by the level of watersolubility.It is usually assumed that toxic effects are only measured following exposure to the dissolvedfraction, i.e. organisms are exposed to substances dissolved in water. It is assumed that thesubstances will either be absorbed by the organisms through passive diffusion or taken upactively by a specific mechanism. Bioavailability may, therefore, vary between differentorganisms. In the case of bioaccumulation, oral exposure could also be considered forsubstances with high Log K ow . Further guidance of the impact of bioavailability caused by thesize of the molecule and how this is considered for aquatic hazard classification can be foundin Annex III to this document.In general, there are no specific environmental test methods developed to measure biologicalavailability of substances or mixtures. This aspect is built into the testing methodology fortoxicity and if adverse effects are identified the substance should be classified accordingly.Substances which lack bioavailability would not be absorbed by the exposed organisms andtherefore due to lack of toxic effects these substances would not be classified, unless they areknown to degrade or transform to hazardous products. For example see the strategy for metalsclassification in Annex IV to this document.1.4 USE OF SUBSTANCE CATEGORISATION (READ ACROSS ANDGROUPING) AND (Q)SARS FOR CLASSIFICATION AND LABELLINGArticle 5(1) Manufacturers, importers and downstream users of a substance shall identify the relevantavailable information for the purposes of determining whether the substance entails a physical, healthor environmental hazard as set out in Annex I, and, in particular, the following:[…](c) any other information generated in accordance with section 1 of Annex XI to Regulation (EC) No1907/2006;Article 6(1) Manufacturers, importers and downstream users of a mixture shall identify the relevantavailable information on the mixture itself or the substances contained in it for the purposes ofdetermining whether the mixture entails a physical, health or environmental hazard as set out in AnnexI, and, in particular, the following:[…](c) any other information generated in accordance with section 1 of Annex XI to Regulation (EC) No1907/2006 for the mixture itself or the substances contained in it;Section 1 of Annex XI to REACH provides a list of data that can be used instead of testingwhen standard data are missing. This Annex specifies the conditions under which results of(Q)SARs, read across and grouping may be used for the classification of substances. It statesthat results of (Q)SARs may be used instead of testing when the (Q)SAR models have beenscientifically validated, “the substance falls within the applicability domain”, the "results areadequate for the purpose of classification and labelling" and “adequate and reliable46

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012documentation of the applied method is provided”. Results generated by read across andgrouping may according to the same principles be used for classification and labelling if theyare "adequate for classification and labelling", “have adequate and reliable coverage of thekey parameters addressed in the corresponding test method”, “cover an exposure durationcomparable to or longer than the corresponding test method”, and “adequate and reliabledocumentation of the applied method” is provided. A weight of evidence approach has to beused where the criteria cannot be applied directly to the available data according to CLPArticle 9(3). This approach is further worked out in CLP Annex I, 1.1.1.No specific guidance is given in REACH, Annex XI on when a result obtained with one of themethods is “adequate for the purpose of classification and labelling”. However, it is importantto note that most of the criteria for classification are directly related to specific test methods.Thus, the adequacy of results of (Q)SARs, read across and grouping should be evaluatedagainst the criteria taking into account that normally the individual method attempts toestimate the same hazard as the criterion. Nevertheless, when grouping, read across and(Q)SARs are being used alone or as a part of the basis for classification, it is normallynecessary to do so employing weight of evidence and expert judgement to decide on theclassification.CLP Annex I, 1.1.1.3 refers to the consideration of the category approach which encompassesgrouping and read-across and (Q)SAR results to help in the weight of evidence determinationof the classification category.Annex 1: 1.1.1.3. A weight of evidence determination means that all available information bearing onthe determination of hazard is considered together, such as the results of suitable in vitro tests, relevantanimal data, information from the application of the category approach (grouping, read-across),(Q)SAR results, human experience such as occupational data and data from accident databases,epidemiological and clinical studies and well-documented case reports and observations. The qualityand consistency of the data shall be given appropriate weight. Information on substances or mixturesrelated to the substance or mixture being classified shall be considered as appropriate, as well as site ofaction and mechanism or mode of action study results. Both positive and negative results shall beassembled together in a single weight of evidence determination.IR/CSA, Chapter R.6 provides extensive advice on the use of (Q)SARs and grouping ofsubstances including guidance on read across, for developing the data set for hazardevaluation. Guidance on the use of (Q)SAR and grouping for specific hazard classes is givenin IR/CSA, Chapter R.7.In general, read across, grouping and use of (Q)SARs as the sole information elements toobtain data on basic physical-chemical properties is not recommended, since reliable datashould normally be available or is easily obtainable through testing. However, there mayoccasionally be practical problems with testing of substances for physical-chemicalproperties, especially for UVCBs where the properties may be dependent on the variablecomposition. Therefore, the appropriateness of using read across, categorisation and (Q)SARsfor physical-chemical assessment should be considered on a case by case basis.Given the availability of extensive guidance only a brief overview of each approach ispresented below. For classification of mixtures see section 1.6 of this document.1.4.1 (Q)SARStructure Activity Relationships and Quantitative Structure Activity Relationships,collectively referred to as (Q)SARs, are defined in IR/CSA, Chapter R.6.1.1 as theoreticalmodels that can be used to predict in a qualitative or quantitative manner the physicochemical,biological (e.g. toxicological) or environmental fate properties of compounds fromknowledge of their chemical structure.47

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012It should be noted that the use of (Q)SAR results requires the user to be sufficiently skilled tounderstand the applicability of the selected (Q)SAR and to interpret the results in terms ofreliability and adequacy for the purpose of classification and labelling.Extensive guidance on the use of (Q)SARs for hazard identification is given in IR/CSA,Chapter R.6.1. Guidance on the use of (Q)SARs for classification and labelling according toDSD is also given in IR/CSA, Chapter R.6.1.4.2. This guidance is directly applicable to CLP.It should be noted that the (Q)SAR approach is not directly applicable to inorganicsubstances.1.4.2 GroupingGuidance on grouping of substances for the purpose of hazard evaluation is given in IR/CSA,Chapter R.6.2. Annex XI to REACH opens the possibility of evaluating substances not on aone-by-one basis, but by grouping substances in categories. A substance category is a groupof substances whose physico-chemical, human health, environmental and/or environmentalfate properties are expected to be similar or to follow a regular pattern as a result of structuralsimilarity.The use of grouping for hazard evaluation in the category approach means that not everysubstance needs to be tested for every hazard. Read across by interpolation can be used to filldata gaps, as well as trend analysis and (Q)SAR, and in addition the overall data for thatcategory must prove adequate to support the hazard assessment.Classification of all substances within an initially considered category may be inappropriateas substances may fall into more than one hazard classification category. Experience hasshown that, an effect can be present for some but not all members of an initially consideredcategory. One example is the glycol ethers, where some members of the category showreproductive toxicity whilst other members do not. In other cases, the category may show aconsistent trend where the resulting potencies lead to different classifications (IR/CSA,Chapter R.6.2.1.2). In such cases it is proposed to use sub-categories for the different hazardclasses where each sub-category receives the most appropriate classification.1.4.3 Read acrossRead across is the use of hazard specific information for one substance (“source”) to predictthe same hazard for another substance (“target”), which is considered to have similar physicochemical,environmental fate and/or (eco)toxicological properties. This can be based onstructural similarity (e.g. (Q)SAR) of a parent substance or its transformation products, andtheir bioavailability, bioaccessiblity, or known physico-chemical properties such as watersolubility. In principle, read across can be applied to characterise physico-chemical properties,environmental fate, human health effects and ecotoxicity. For certain substances without testdata the formation of common significant metabolites or information with those of testedsubstances or information from precursors may be valuable information (IR/CSA, ChapterR.6.2.5.2 and OECD 2004). For any hazard class, read across may be performed in aqualitative or quantitative manner. Extensive guidance on the use of read across is given inIR/CSA, Chapter R.6.2.2.1.Specific guidance for certain types of substances such as reaction products and multiconstituentsubstances, complex substances, isomers, metals and metal compounds and otherinorganic compounds is given in IR/CSA, Chapter R.6.2.5. This is because the concept ofsubstance categories has traditionally been widely used for hazard classification and to someextent also for risk assessment.48

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121.5 SPECIFIC CONCENTRATION LIMITS AND M-FACTORS1.5.1 Specific concentration limitsArticle 10(1) Specific concentration limits and generic concentration limits are limits assigned to asubstance indicating a threshold at or above which the presence of that substance in anothersubstance or in a mixture as an identified impurity, additive or individual constituent leads to theclassification of the substance or mixture as hazardous.Specific concentration limits shall be set by the manufacturer, importer or downstream user whereadequate and reliable scientific information shows that the hazard of a substance is evident when thesubstance is present at a level below the concentrations set for any hazard class in Part 2 of Annex Ior below the generic concentration limits set for any hazard class in Parts 3, 4 and 5 of Annex I.In exceptional circumstances specific concentration limits may be set by the manufacturer, importeror downstream user where he has adequate, reliable and conclusive scientific information that ahazard of a substance classified as hazardous is not evident at a level above the concentrations setfor the relevant hazard class in Part 2 of Annex I or above the generic concentration limits set for therelevant hazard class in Parts 3, 4 and 5 of that Annex.Article 10(3) Notwithstanding paragraph 1, specific concentration limits shall not be set forharmonised hazard classes or differentiations for substances included in Part 3 of Annex VI.The specific concentration limit (SCL) concept allows a fine tuning of the contribution ofcertain hazardous substances to the classification of mixtures based on the potency of thesubstances, as well as a classification of other substances containing these substances asimpurities, additives or individual constituents. The SCL concept is only applicable to healthhazards. For physical hazards, classification shall be established on the basis of test data forthe respective mixture, where applicable.The procedure of derivation of SCLs is different for every health hazard class and thereforeguidance on how to set SCLs is provided in the respective sections of this document.Guidance on setting of SCLs is supplied in the respective chapters of the different healthhazard classes. A general overview on the applicability of SCLs and guidance availability forsetting SCLs for health hazards is given in this chapter.An overview of guidance available is also illustrated by Table 1.5.1 below.SCLs should take precedence over the generic concentration limits (GCLs) given in therelevant health hazard sections of Annex I to CLP. In case specific concentration limits havebeen set in Annex VI to CLP, these must be applied. Moreover, suppliers may not set ownSCLs for harmonised classifications in Annex VI to CLP.SCLs should be available in the C&L Inventory, and established in accordance with CLP.Table 1.5.1 Possibilities for setting SCL for health hazards as addressed in relevant sections of theguidance.Hazard classCategoryLowerSCLthan GCLHigher SCLsthan GCL (inexceptionalcircumstances)GuidanceAcute toxicity all not applicable not applicable not necessarySkin corrosion/irritationall yes yes available in section 3.249

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Serious eyedamage/eye irritationRespiratorysensitisationall yes yes available in section 3.31 yes noSkin sensitisation 1 yes yesto be provided insection 3.4 40to be provided insection 3.4 (see above)Germ cellmutagenicityall no no currently not possibleCarcinogenicity all yes yes available in section 3.6Reproductiveavailable in section 3.7all yes yestoxicityand in Annex VISTOT-SE 1 yes no available in section 3.82 no no see section 3.83 yes yes available in section 3.8STOT-RE 1 yes no available in section 3.92 no no see section 3.9Aspiration hazard 1 not appropriate not appropriate not necessary1.5.2 Multiplying factors (M-factors)Article 10(2) M-factors for substances classified as hazardous for the aquatic environment, acutecategory 1 or chronic category 1, shall be established by manufacturers, importers and downstreamusers.Article 10(4) Notwithstanding paragraph 2, M-factors shall not be set for harmonised hazard classesor differentiations for substances included in Part 3 of Annex VI for which an M-factor is given inthat Part.However, where an M-factor is not given in Part 3 of Annex VI for substances classified ashazardous to the aquatic environment, acute category 1 or chronic category 1, an M-factor based onavailable data for the substance shall be set by the manufacturer, importer or downstream user.When a mixture including the substance is classified by the manufacturer, importer or downstreamuser using the summation method, this M-factor shall be used.For the hazard class “Hazardous to the Aquatic Environment”, SCLs are not applicable.Instead the M-factors concept is used.The M-factors are used in application of summation method for classification of mixturescontaining substances that are classified as very toxic. The concept of M-factors has beenestablished to give an increased weight to very toxic substances when classifying mixtures.M-factors are only applicable to the concentration of a substance classified as hazardous tothe aquatic environment (categories Acute 1 and Chronic 1) and are used to derive by thesummation method the classification of a mixture in which the substance is present. They are,however, substance-specific and it is important that they are being established already whenclassifying substances.For further guidance in how to establish the M-factor see Section 4.1.3.3.3 of this document.40 Guidance on the setting of SCLs relating to the revised criteria for respiratory and skin sensitization that arebased on the 2 nd ATP to the CLP Regulation is planned for a future update of this guidance document.50

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012M-factors should have been established in accordance with Article 10 of CLP and beavailable in the C&L Inventory.For the harmonised classifications in Annex VI to CLP, M-factors shall be set by themanufacturer, importer or downstream user in case there is no M-factor provided, inaccordance with CLP Article 10(4).1.6 MIXTURES1.6.1 How to classify a mixtureThe classification of mixtures under CLP is for the same hazards as for substances. As ageneral rule and as is the case with substances, available data on the mixture as a wholeshould primarily be used to determine classification where applicable. If this cannot be done,further approaches to mixture classification may be applied.It is important to choose the most appropriate method to determine the classification for amixture for each hazard class, differentiation or category. The method will depend on whetherthe mixture is being assessed for physical, health or environmental hazards and on the typeand quality of information that is available (see also section 1.2.3 of this document on form orphysical state).It is important to get a clear picture on which substances and mixtures are contained in amixture. Basic information on substances would include the substance identity, itsclassification and any applied SCLs or M-factors, and concentration in the mixture and, whererelevant, details of any impurities and additives including their identity, classification andconcentration. Where an ingredient in a mixture is itself a mixture, it is necessary to getinformation on the ingredient substances of that mixture together with their concentrations,classifications and any applied SCLs or M-factors.Useful sources for such information are the SDS from the supplier of the substance or themixture, and the C&L Inventory provided by ECHA, which also includes the harmonisedclassifications of substances listed in Annex VI to CLP.REACH: Article 31(3)The supplier shall provide the recipient at his request with a safety data sheet compiled in accordancewith Annex II, where a mixture does not meet the criteria for classification as dangerous inaccordance with Articles 5, 6 and 7 of Directive 1999/45/EC, but contains:(a) in an individual concentration of ≥ 1 % by weight for non-gaseous mixtures and ≥ 0,2 % byvolume for gaseous mixtures at least one substance posing human health or environmental hazards; or(b) in an individual concentration of ≥ 0,1 % by weight for non-gaseous mixtures at least onesubstance that is persistent, bio-accumulative and toxic or very persistent and very bio-accumulativein accordance with the criteria set out in Annex XIII or has been included for reasons other than thosereferred to in point (a) in the list established in accordance with Article 59(1); or(c) a substance for which there are Community workplace exposure limits.NOTE: Article 31(3) is amended from 1 June 2015 by CLP Article 59 (2)(b) to read as follows:The supplier shall provide the recipient at his request with a safety data sheet compiled in accordancewith Annex II, where a mixture does not meet the criteria for classification as hazardous inaccordance with Titles I and II of Regulation (EC) No 1272/2008, but contains:(a) in an individual concentration of ≥ 1 % by weight for non-gaseous mixtures and ≥ 0,2 % byvolume for gaseous mixtures at least one substance posing human health or environmental51

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012hazards; or(b) in an individual concentration of ≥ 0,1 % by weight for non-gaseous mixtures at least onesubstance that is carcinogenic category 2 or toxic to reproduction category 1A, 1B and 2, skinsensitiser category 1, respiratory sensitiser category 1, or has effects on or via lactation or ispersistent, bioaccumulative and toxic (PBT) in accordance with the criteria set out in Annex XIII orvery persistent and very bioaccumulative (vPvB) in accordance with the criteria set out in Annex XIIIor has been included for reasons other than those referred to in point (a) in the list established inaccordance with Article 59(1); or(c) a substance for which there are Community workplace exposure limits.Further dialogue with the supplier may be necessary to obtain additional information. Forexample on compositional information for the mixture supplied.The classification of mixtures follows the sequence displayed in Figure 1.6.1, for eachhazard class independently:52

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 1.6.1 How to classify a mixtureThere is a mixture to classifyAll available information should begatheredAre available test data for the mixturesufficient for classification?(CLP Article 9 (2)-(3))(For physical hazards: considerwhether new testing needs to beperformed. Consult the criteria.)YESClassify the mixture for the relevant hazardNOIs there data available onsimilar tested mixtures andindividual hazardousingredients?YESIs it possible to applyany of the bridgingprinciples?YESClassify themixture for therelevant hazardNONOAre hazard data available forall or some ingredients?YESNOUnable to classify the mixture – go back to ingredientsuppliers to obtain additional informationUse the known or derived hazarddata on the individual ingredients toclassify the mixture for the relevanthazard, using the other methods ineach section of CLP Annex I, Part 3and Part 4Note: The principles for using expert judgement and weight of evidence determination (CLP Article 9(3) and (4)) and Annex I, section 1.1.1.) should be taken into account.1.6.2 Classification for physical hazards53

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The majority of the physical hazards of mixtures should be determined through testing basedon the methods or standards referred to in CLP Annex I, Part 2. In few cases, such as hazardclass “Flammable liquids”, the classification of mixtures can also be derived through acalculation, see CLP Annex I, 2.6.4.2 and 2.6.4.3.The test methods can be found for example in the UN Manual of Tests and Criteria, see thewebsite http://www.unece.org/trans/danger/publi/manual/manual_e.html , which is normallyused to classify substances and mixtures for transport. In cases where test results areavailable, based on other methods or standards, then these data may still be used, providedthey are adequate for the purpose of hazard determination. To conclude on the adequacy theresults should be checked by the expert involved to ensure that there is sufficientdocumentation to assess the suitability of the test used, and whether the test was carried outusing an acceptable level of quality assurance.Please note that in practice the physical hazards of a substance or mixture may differ fromthose shown by tests, e.g. in case of certain ammonium-nitrate-based compounds (explosive /oxidising properties) and certain halogenated hydrocarbons (flammable properties). Suchexperience must be taken into account for the purpose of classification (CLP Article 12(a)).The information available or generated must be checked to determine if it is directlycomparable to the respective hazard criteria and if it is, then it can be used to derive theclassification immediately. Where the criteria cannot be directly applied to the available data,expert judgement should be used for the evaluation of the available information in a weight ofevidence determination (CLP Article 9(3) and CLP Annex I, 1.1.1.).1.6.3 Health and environmental hazardsFor the purpose of classification for health or environmental hazards, check whether or notthere is information: on the mixture itself; on similar tested mixtures and ingredient substances; or on the classification of ingredient substances and their concentrations in the mixture.As pointed out in the introduction to this chapter, the supplier should be contacted if it isconsidered that the information on the substances or mixtures supplied is not sufficient forclassification purposes.The information available on the hazard under consideration, will determine if the mixtureshould be classified using the approaches below in the following sequence (CLP Article 9):(a)Classification derived using data on the mixture itself (see section 1.6.3.1 of thisdocument), by applying the substance criteria of Annex I to CLP;(b) Classification based on the application of bridging principles (see section 1.6.3.2of this document), which make use of test data on similar tested mixtures andingredient substances; and(c)Classification based on calculation or on concentration thresholds, including SCLsand M-factors.54

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121.6.3.1 Classification derived using data on the mixture itselfClassification derived using data on the mixture itself, by applying the substance criteria ofAnnex I to CLP, is applicable in many cases. Exceptions are: CMR hazards (see CLP Article6(3)), bioaccumulation and biodegradation properties and the evaluation within the‘hazardous to the aquatic environment’ hazard class referred to in sections 4.1.2.8 and 4.1.2.9of Annex I to CLP (see CLP Article 6(4)).Article 6 (3)For the evaluation of mixtures pursuant to Chapter 2 of this Title in relation to the ‘germ cellmutagenicity’, ‘carcinogenicity’ and ‘reproductive toxicity’ hazard classes referred to in sections3.5.3.1, 3.6.3.1 and 3.7.3.1 of Annex I, the manufacturer, importer or downstream user shall onlyuse the relevant available information referred to in paragraph 1 for the substances in the mixture.Further, in cases where the available test data on the mixture itself demonstrate germ cellmutagenic, carcinogenic or toxic to reproduction effects which have not been identified from theinformation on the individual substances, those data shall also be taken into account.Article 6(4)For the evaluation of mixtures pursuant to Chapter 2 of this Title in relation to the ‘biodegradationand bioaccumulation’ properties within the ‘hazardous to the aquatic environment’ hazard classreferred to in sections 4.1.2.8 and 4.1.2.9 of Annex I, the manufacturer, importer or downstreamuser shall only use the relevant available information referred to in paragraph 1 for the substancesin the mixture.Where the criteria cannot be directly applied to the available data, expert judgement should beused for the evaluation of the available information in a weight of evidence determination(CLP Article 9(3) and CLP Annex I, 1.1.1).1.6.3.2 Bridging principlesIn the case of a classification for health or environmental hazards, information on the mixtureitself may not always be available. However, where there are sufficient data on similar testedmixtures and individual hazardous ingredient substances, CLP allows bridging principles tobe used to classify the mixture (CLP Annex I, 1.1.3). To apply these bridging principlescertain conditions should be considered for their application which are summarised below.Not all of the bridging principles as described in sections 1.6.3.2.1-1.6.3.2.5 of this documentneed to be applied when assessing a particular health or environmental hazard. It is necessaryto consult Annex I of CLP, Part 3 for health hazards and Part 4 for environmental hazards,before undertaking any of these assessments.In case it is not possible to classify the mixture by applying bridging principles and a weightof evidence determination using expert judgement, then the mixture should be classified usingthe other methods described in CLP Annex I, Parts 3 and 4.1.6.3.2.1 DilutionWhere the tested mixture is diluted with a substance (diluent) that has an equivalent or lowerhazard category than the least hazardous original ingredient substance, then it can be assumedthat the respective hazard of the new mixture is equivalent to that of the original testedmixture. The application of dilution for determining the classification of a mixture isillustrated by Figure 1.6.3.2.1.55

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 1.6.3.2.1 Application of the bridging principle: dilution for determining the acute toxicityclassification of a mixtureDiluent B(classification known)Mixture A(tested)Mixture C(A+B)(not tested)Example: Mixture A, which has been classified as acute toxic category 2 based on test data, issubsequently diluted with diluent B to form mixture C. If diluent B has an equivalent or loweracute toxicity classification than the least acutely toxic ingredient in mixture A and is notexpected to affect the hazard classification of other ingredients, then mixture C may be alsoclassified as acutely toxic category 2. However, this approach may over-classify mixture C,thus the supplier may choose to apply the additivity formula described in CLP Annex I,3.1.3.6 (see Section 1.6.3.4.1 of this document).Note that also the diluent of the tested mixture is considered a relevant ingredient.Consider using this particular bridging principle also when, for example,- diluting an irritant mixture with water,- diluting an irritant mixture with a non-classified ingredient, or- diluting a corrosive mixture with a non-classified or irritant ingredient.In case a mixture is diluted with another mixture, see section 1.6.4 of this document.Within the ‘hazardous to the aquatic environment’ hazard class, if a mixture is formed bydiluting another classified mixture or substance with water or other totally non-toxic material,the toxicity of the mixture can also be calculated from the original mixture or substance (seesection 4.1.3.4.3 of Annex I to CLP and mixture example C in section 4.1.4.7 of thisdocument).1.6.3.2.2 BatchingWhere a batch of a mixture is produced under a controlled process, then it can be assumedthat the hazards of each new batch are equivalent to those of previous batches. This methodmust not be used where there is reason to believe that the composition may vary significantly,affecting the hazard classification.1.6.3.2.3 Concentration of highly hazardous mixturesWhere a tested mixture is already classified in the highest hazard category or sub-category, anuntested mixture which contains a higher concentration of those ingredient substances that arein that category or sub-category should also be classified in the highest hazard category orsub-category (CLP Annex I, 1.1.3.3).1.6.3.2.4 Interpolation within one toxicity category56

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Assume there are three mixtures (A, B and C) which contain identical hazardous components.If mixtures A and B have been tested and are in the same hazard category, and mixture C isnot tested and has concentrations of those hazardous components intermediate to theconcentrations in mixtures A and B, then mixture C is assumed to be in the same hazardcategory as A and B. The application of interpolation for determining the classification of amixture is illustrated by Figure 1.6.3.2.4. (CLP Annex I, 1.1.3.4).Figure 1.6.3.2.4 Application of the bridging principle: interpolation for determining the aquatic acutehazard classification of a mixture90% 30%10%70%Mixture A(Aquatic Acute 1)Mixture B(Aquatic Acute 1)60%30% ≤ conc. ≤ 90%40%10% ≤ conc. ≤ 70%Mixture C(Interpolate as Aquatic Acute 1)1.6.3.2.5 Substantially similar mixturesTwo mixtures contain an identical ingredient at the same concentration. Each of the twomixtures contains an additional ingredient which is not identical with each other; howeverthey are present in equivalent concentrations and the hazard category of these two ingredientsis the same and neither of them is expected to affect the hazard classification of the other. Ifone of the mixtures is classified based on test data it may be assumed that the hazard categoryof the other mixture is the same. The application of substantially similar mixtures fordetermining the classification of a mixture is illustrated by Figure 1.6.3.2.5. (CLP Annex I,1.1.3.5).Figure 1.6.3.2.5 Application of the bridging principle: substantially similar mixtures for determiningthe skin irritation classification of a mixtureIngredient A10% Ingredient B90%Ingredient B90%Ingredient C10%Mixture P(tested)(Skin Irrit. 2)Mixture Q(not tested)57

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Example: If the Ingredient C has the same hazard category and the same potency as IngredientA, then Mixture Q can be classified as Skin Irrit. 2 like Mixture P. Potency may be expressedby, for example, differences in the specific concentration limits of Ingredients A and C. Thismethod should not be applied where the irritancy of Ingredient C differs from that ofIngredient A.1.6.3.2.6 Review of classification where the composition of a mixture has changedArticle 15(2) Where the manufacturer, importer or downstream user introduces a change to a mixturethat has been classified as hazardous, that manufacturer, importer or downstream user shall carry out anew evaluation in accordance with this Chapter where the change is either of the following:(a) a change in the composition of the initial concentration of one or more of the hazardousconstituents in concentrations at or above the limits in Table 1.2 of Part 1 of Annex I;(b) […]Annex I: 1.1.3.6 Review of classification where the composition of a mixture has changedThe following variations in initial concentration are defined for the application of Article 15(2)(a):58Table 1.2Bridging Principle for changes in the composition of a mixtureInitial concentration range of the constituentPermitted variation in initial concentration of theconstituent≤ 2,5 % ± 30 %2,5 < C ≤ 10 % ± 20 %10 < C ≤ 25 % ± 10 %25 < C ≤ 100 % ± 5 %NOTE: The guidance below explaining Table 1.2 in the green box relates to a change in thecomposition of mixtures already classified as hazardous. A change in the composition of nonhazardousmixtures may result in concentration thresholds being reached and a need toclassify the changed mixture as hazardous. Where the manufacturer, importer or downstreamuser introduces a change to a mixture not classified for a specific hazard, that manufacturer,importer or downstream user must therefore always carry out a new evaluation for that hazardin accordance with Chapter 2 of Title II to CLP (see Article 15(1) of CLP).Where a manufacturer, importer or downstream user introduces a change in the compositionof the initial concentration of one or more of the hazardous constituents of a mixtureclassified as hazardous, that manufacturer, importer or downstream user shall carry out a newevaluation where the change in concentrations is at or above the limits in Table 1.2 of Part 1of Annex I to CLP.However, where the variations of the initial concentrations of the constituents lie within thepermitted variation, manufacturer, importer or downstream user does not need to carry out anew evaluation and may use the current classification of the mixture.The following example is to illustrate what is meant by the permitted variations in Table 1.2.Example: Mixture A is classified as hazardous based on the initial concentration of two

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012hazardous constituents, substance A and substance B. The initial concentrations in the mixtureof substance A and substance B are 2 % and 12 %, respectively. The permitted variationaccording to table 1.2 is for substance A ± 30 % of the initial concentration and for substanceB ± 10 % of the initial concentration. This means that the concentration in the mixture mayfor substance A vary between 1.4 % and 2.6 % and for substance B between 10.8 % and13.2 %, without having to carry out a new evaluation in accordance with Chapter 2 of Title IIto CLP:Substance A: 2 ±0.3 = ±0.6 1.4 – 2.6Substance B: 12 ±0.1 = ±1.2 10.8 – 13.21.6.3.3 Aerosols (some health hazards only)A mixture in aerosol form is considered to have the same classification as the non-aerosolisedform of a mixture, provided that the propellant used does not affect these hazards uponspraying and data demonstrating that the aerosolised form is not more hazardous than the nonaerosolisedform is available (see CLP Annex I, 1.1.3.7.).1.6.3.4 Classification based on calculation or concentration thresholdsIn most cases, test data on the mixture itself will not be available for a mixture, thereforebridging principles and weight of evidence determination using expert judgement for all ofthe necessary health and environmental hazard assessments may not be applied. In thesecases, classification must be based on calculation or on concentration thresholds referring tothe classified substances present in the mixture.In the case where one or more mixtures are added to another mixture, the same requirementapplies: it is necessary to know all ingredient substances, their hazard classifications and theirconcentrations to be able to derive a correct hazard classification of the final mixture. Forfurther details see section 1.6.4 of this document.1.6.3.4.1 Classification based on calculationThe calculation methods set out under the different chapters of Annex I to CLP mostly differfrom those applied under DPD. More detailed guidance on the selection of the mostappropriate method is provided in the specific section for each hazard class.An example is the hazard class acute toxicity where a calculation formula is used which isbased on acute toxicity estimates and concentrations, and a modified formula for determiningthe classification of a mixture containing substances of unknown acute toxicity.Annex I: 3.1.3.6.1.[…]The ATE of the mixture is determined by calculation from the ATE values for all relevant ingredientsaccording to the following formula for Oral, Dermal or Inhalation Toxicity:where:100ATEC i = concentration of ingredient i ( % w/w or % v/v)i = the individual ingredient from 1 to nmix nCiATEi59

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012n = the number of ingredientsATE i = Acute Toxicity Estimate of ingredient i.Annex I: 3.1.3.6.2.3. If the total concentration of the ingredient(s) with unknown acute toxicity is ≤ 10% then the formula presented in section 3.1.3.6.1 shall be used. If the total concentration of theingredient(s) with unknown toxicity is > 10 %, the formula presented in section 3.1.3.6.1 shall becorrected to adjust for the total percentage of the unknown ingredient(s) as follows:100 (Cunknownif 10%) ATEmixnCiATEFor more information on the CLP calculation formulae for this hazard, please see section3.1.3.3.3 of this document.Another example is provided by hazard class “hazardous to the aquatic environment”, namelythe additivity formula:i60

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: 4.1.3.5.2. Mixtures can be made of a combination of both components that are classified (asAcute Category 1 and/or Chronic Category 1, 2, 3 or 4) and others for which adequate toxicity testdata are available. When adequate toxicity data are available for more than one component in themixture, the combined toxicity of those components is calculated using the following additivityformulas(a) and (b), depending on the nature of the toxicity data:(a) Based on acute toxicity:where: Ci L(E)C50mC i = concentration of component i (weight percentage)L(E)C 50i = (mg/l) LC 50 or EC 50 for component iη = number of componentsηCiL(E)CL(E)C 50m = L(E)C 50 of the part of the mixture with test dataThe calculated toxicity may be used to assign that portion of the mixture an acute hazard categorywhich is then subsequently used in applying the summation method;(b) Based on chronic aquatic toxicity:50iCiCjCi Eq NOECmnNOECinC0,1x NOECjjWhere:C i = concentration of component i (weight percentage) covering the rapidly degradable componentsC j = concentration of component i (weight percentage) covering the non-rapidly degradablecomponentsNOEC i = NOEC (or other recognised measures for chronic toxicity) for component i covering therapidly degradable components, in mg/l;NOEC j = NOEC (or other recognised measures for chronic toxicity) for component i covering thenon-rapidly degradable components, in mg/l;n = number of components, and I and j are running from 1 ton;EqNOEC m = Equivalent NOEC of the part of the mixture with test data;[…]NOTE: To make full use of this approach requires access to the whole aquatic toxicity dataset and the necessary knowledge to select the best and most appropriate data. CLP has limitedthe use of the additivity formulae to those circumstances where the substance hazard categoryis not known, although the acute and/or chronic toxicity data are available.For more information on the CLP calculation formulae for this hazard please see section4.1.4.3 of this document.1.6.3.4.2 Classification based on concentration thresholdsGeneric concentration thresholds61

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012For some hazard classes or differentiations, classification based on concentration thresholdsmay be applicable. CLP distinguishes between two different kinds of generic concentrationthresholds:- Generic cut-off values: these values are the minimum concentrations for a substance tobe taken into account for classification purposes. These substances are also referred toas relevant ingredients in some hazard classes (see sections 3.1, 3.2 and 3.3). When aclassified substance is present in a concentration above the generic cut-off value itcontributes to the mixture classification even if it does not trigger classification of themixture directly. The generic cut-off values are defined for some hazard classes andcategories only and are listed in Table 1.1 of Annex I to CLP;- Generic concentration limits: these values are the minimum concentrations for asubstance which trigger the classification of a mixture if exceeded by the individualconcentration or the sum of concentrations of relevant substances (where the individualsubstance concentrations can be ‘added’ to each other in a straight forward way); theyare set out in parts 2-5 of Annex I for those hazard classes where they apply.Generic concentration thresholds are generic for a hazard class, differentiation or category.The difference between a generic cut-off value and a generic concentration limit (GCL) isdemonstrated through the example of the skin irritation hazard: while Table 1.1 of Annex I toCLP defines the generic cut-off value to be 1 % a skin irritant substance which is present in amixture would trigger classification of the mixture as skin irritant if it were present above orequal to the concentration limit of 10 % in the mixture, see Table 3.2.3 of Annex I to CLP.However, at 1 % and below 10 %, it may still contribute to the classification of the mixtureas skin irritant, since the concentration would be taken into account if other skincorrosive/irritant substances are present in the mixture below the relevant genericconcentration limits. In some cases, classification as provided by the summation in CLPAnnex I, Table 3.2.3 may be applicable, i.e.:(10 Skin Corrosive Categories 1A, 1B, 1C) + Skin Irritant Category 2 should be ≥ 10 %Specific concentration thresholdsIn contrast to generic thresholds, “Specific Concentration Limits” (SCLs) and/or specific cutoffvalues may be established for substances:1. SCLs are described in section 1.5.1 of this document and where they have beenestablished they are included in Tables 3.1 and 3.2 of Annex VI to CLP and/or in theC&L Inventory (CLP Article 42). For “hazardous to the aquatic environment” theMultiplying factors (M-factors) concept 41 is used instead of SCLs, see section 1.5.2 ofthis guidance. SCLs and M-factors included in Tables 3.1 and 3.2 must be used whereapplicable and, for classifications not included in Annex VI, SCLs and M-factorsincluded in the C&L Inventory shall be used where applicable unless justifiedotherwise.2. Cut-off values that may be different from the generic values and that are to be used inspecific cases are given in 1.1.2.2.2(a) and (b) of Annex I to CLP. For exampleconcerning aquatic hazard, for a substance with an established M-factor, the cut-off41 M-factors are used to derive, by means of the summation method, the classification of a mixture in which thesubstance is present for which the M-factor has been established. For further guidance on how to establish anduse M-factors see sections 4.1.3.3.2 and 4.1.4.5, respectively.for which the M-factor has been established is present. For further guidance in how to establish and use theMfactor see Sections 4.1.3.3.2 and 4.1.4.5 respectively62

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012value is always the generic cut-off value divided by the M-factor; hence, (0.1/M) % (see1.1.2.2.2(b) and 4.1.3.1 of Annex I to CLP).Specific concentration thresholds take precedence over generic thresholds. In Annex I to DSDalso generic concentration limits were listed in case SCLs were described to a certain entry.However in Tables 3.1 and 3.2 of Annex VI to CLP, these were deleted because under CLP,SCLs and M-factors can be set by the manufacturer or importer and they would then still takeprecedence to the generic thresholds, why those cannot be defined for specific entries.1.6.3.4.3 Additivity of hazardsFor some hazard classes additivity concepts are not applicable. In these cases, if the mixturecontains two substances each below the GCLs defined for that hazard class or differentiation,even if the sum is above this limit, the mixture will not be classified, as far as no lower SCLhas been set.Non-additivity is applied for the following hazard classes:(a) skin and respiratory sensitisers;(b) germ cell mutagenicity;(c) carcinogenicity;(d) reproductive toxicity;(e) specific target organ toxicity, single and repeated exposure, categories 1 and 2;(f) aspiration hazard (plus consideration of viscosity of the final mixture);(g) skin corrosion/irritation in some special cases (see CLP Annex I, 3.2.3.3.4); and(h) serious eye damage/eye irritation in some special cases (see CLP Annex I, 3.3.3.3.4).For example, where there are two ingredient substances classified for specific target organtoxicity - repeated exposure in Category 1 present in the mixture, but none of them is presentat or above 10 % or below 1 %, then the mixture will not be classified in Category 1 but willbe Category 2 (even if the sum would be greater than 10 %, because the additivity concept isnot applicable).Additivity is used for the following hazard classes or differentiations:(a) skin corrosion/irritation (besides the cases mentioned in CLP Annex I, 3.2.3.3.4);(b) serious eye damage/eye irritation (besides the cases mentioned in CLP Annex I,3.3.3.3.4);(c) specific target organ toxicity, single exposure Category 3 (respiratory tract irritation);(d) specific target organ toxicity, single exposure Category 3 (narcotic effects); and(e) acute and long-term aquatic hazards.In these cases, if the sum of the concentrations of one or several classified substances in themixture equals or exceeds the GCL set out for this hazard class/category, the mixture must beclassified for that hazard. For substances that have an SCL or M-factor(s), these should betaken into account when applying the summation methods.An example is provided for the hazard class serious eye damage /eye irritation: In case thereare only substances classified as eye irritation Category 2 present in a mixture, then their summust be equal to or exceed the generic concentration limit of 10 % in order for the mixture tobe classified in Category 2 as well. Note that only relevant substances should be summed up63

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012and contribute to mixture classification. Further guidance on the application of SCLs whenusing the summation method to derive skin corrosion / irritation or serious eye damage/eyeirritation hazards can be found in sections 3.2 and 3.3 of this document.1.6.4 Classification of mixtures in mixturesFor physical hazards, an adequate hazard classification is generally derived by testing. Todetermine the classification of a mixture for health or environmental hazards using theadditivity or summation methods, information on all the constituent substances, includingtheir individual hazard classification and concentration, is generally required. In the casewhere one or more mixtures are added to another mixture, the same requirement applies: it isgenerally necessary to know all ingredient substances, their hazard classifications and theirconcentrations to be able to derive a correct hazard classification of the final mixture. It isgenerally not possible to derive the correct hazard classification for the final mixture by usingonly the hazard classification(s) of the mixtures that were combined to make it with oneexception. The exception is that in case the acute toxicity estimate (ATE) of a mixture isknown (either actual or derived), this value can be used to derive a correct classification foracute toxicity if this mixture is added to another mixture.Thus, it is very important that suppliers of mixtures communicate the necessary informationlisted above on constituent substances (including their individual hazard classification andconcentration) down the supply chain, for instance in the SDS, to enable a correctclassification to be established by downstream users formulating new mixtures from theirproducts. However, the information provided in the SDS may not be sufficient, for examplewhere only a concentration range is quoted for a particular substance or where the mixturecontains other substances classified as hazardous but which are present below theconcentration for declaration in the SDS. Thus further dialogue with the supplier of themixture may be necessary to obtain additional information on the constituent substances toensure correct classification and labelling of the new mixture.In situations, where tested mixtures are added to other tested or untested mixtures, anadequate hazard classification can only be derived by taking account of both the test data aswell as the knowledge on all substances, their hazard classifications, and their concentrationsin these mixtures. Such an approach is a case-by-case analysis and requires expert judgement.1.6.4.1 Example: Classification of Mixture ANote that the example only addresses health hazards. For compositional details see Table1.6.4.1(a) and Table 1.6.4.1(b) below.No test data are available on Mixture A so it is not possible to apply bridging principles due tolack of data on similar tested mixtures. Therefore it is necessary to identify the ingredients inMixture A (including their % w/w and classification).Mixture A does not contain any ingredients classified as a respiratory sensitiser, CMR, STOTor aspiration hazard. Therefore it is possible to conclude that Mixture A will not be classifiedas hazardous for these particular hazard classes.Acute toxicityAs indicated in CLP Annex I, 3.1.3.3(b), there are two options to calculate acute toxicity ofMixture A: (i) treat the 'fragrance mixture' as an ingredient when calculating the ATE forMixture A, or (ii) break the 'fragrance mixture' down into its component ingredients and onlytake over the relevant ingredients (CLP Annex I, 3.1.3.3(a) and 3.1.3.6.1) into the calculationfor the ATE of Mixture A.64

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Following option (i) it is first necessary to calculate ATE mix of the 'fragrance mixture' (see1.6.4.1(b)) taking into account 'FM component 1' and 'FM component 2' (other componentscan be excluded as their LD 50 values are > 2000 mg/kg):100ATEmixnCiATEiATEmixn100CiATEiATEmix10035.2 17.01230 500 1597mg/kgThe ATE mix for the 'fragrance mixture' can then be included in the calculation of the ATE mixfor Mixture A:ATE mix1008.0 5.01800 1597 13300mg/kgFollowing option (ii) it is only necessary to include 'FM component 1' from the 'fragrancemixture' (present in Mixture A at 1.76 %), as 'FM component 2' is present in a concentration 2000 mg/kg thus mixture Ais not classified as hazardous for acute toxicity by the oral route.N.B. If an acute oral toxicity test (i.e. an actual LD 50 value) was available for the fragrancemixture, then this should be used in the calculation for the ATE of Mixture A.Skin corrosion/irritationWork out the actual levels of the 'fragrance mixture' ingredients in Mixture A and carry outthe summation method (CLP Annex I, Table 3.2.3) using the relevant ingredients.Mixture A does not contain any ingredient classified as Skin Corr. 1A, B or C. ThereforeMixture A is not classified as Skin Corr. 1A, B or C.The 'fragrance mixture' contains ingredients classified as Skin Irrit. 2, but these are all presentin Mixture A at concentrations < 1 % and can be disregarded (CLP Annex I, Table 1.1).Mixture A does also contain 8 % of the 'anionic surfactant' classified as Skin Irrit. 2, but asthe concentration of the 'anionic surfactant' < 10 %, Mixture A is not classified as Skin Irrit. 2.Serious eye damage/eye irritationWork out the actual levels of the 'fragrance mixture' ingredients in Mixture A and carry outthe summation method (CLP Annex I, Table 3.3.3) using the relevant ingredients:'65

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Mixture A contains 8 % of an ingredient classified as Eye Dam. 1, thus Mixture A must alsobe classified as Eye Dam. 1 (the relevant ingredient is present in a concentration > 3 %). The'fragrance mixture' also contains an ingredient classified as Eye Dam. 1, but this is present inMixture A at a concentration < 1 % and can disregarded.Skin sensitisationThe 'fragrance mixture' contains four ingredients classified as skin sensitisers but their actuallevels in Mixture A are < 1 % thus Mixture A is not classified as a skin sensitiser. However,the four skin sensitiser ingredients are present above 0.1 %, thus additional labellinginformation (CLP Annex II, 2.8) would be required on the label for Mixture A.Table 1.6.4.1(a) Ingredients in Mixture AIngredient % w/w Oral LD 50 (rat) ClassificationAnionic surfactant 8.00 1800 mg/kg Acute Tox. 4 (oral)Eye Dam. 1Skin Irrit. 2Thickening agent 0.80 > 5000 mg/kg Not classifiedDye 0.05 > 5000 mg/kg Not classifiedFragrance mixture(see list of ingredients below)5.00 not tested Acute Tox. 4 (inhalation, oral)Skin Sens. 1Eye Dam. 1Skin Irrit. 2Aquatic Chronic 2Water 86.15 Not classifiedTotal: 100.00Table 1.6.4.1(b) Ingredient ' Fragrance mixture'Ingredient % w/w % in Mixture A Oral LD 50 (rat) ClassificationFM component 1Acute Tox. 435.20 1.76 1230 mg/kg(inhalation, oral)FM component 2FM component 317.00 0.85not available(use cATpE 500)16.00 0.8 3600 mg/kgAcute Tox. 4 (oral)Skin Sens. 1Skin Sens. 1Skin Irrit. 2FM component 4 13.40 0.67 3100 mg/kg Skin Sens. 1FM component 5FM component 67.00 0.35 > 2000 mg/kg6.00 0.3 4400 mg/kgEye Dam. 1Aquatic Chronic 2Flam. Liq. 3Skin Sens. 1Skin Irrit. 2Aquatic Chronic 166

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012FM component 7 2.80 0.14 > 5000 mg/kg Not classifiedFM component 8 2.60 0.13 > 5000 mg/kg Aquatic Chronic 1Total: 100.00 5.001.6.4.2 Example: Classification of Mixture BNote that the example only addresses health hazards. For compositional details see Table1.6.4.2(a) and Table 1.6.4.2(b) below.No test data are available on Mixture B so it is not possible to apply bridging principles due tolack of data on similar tested mixtures. Therefore it is necessary to identify the ingredients inMixture B (including their % w/w and classification).Mixture B does not contain any ingredients classified as a skin sensitiser, CMR or aspirationhazard. Therefore it is possible to conclude that Mixture A will not be classified as hazardousfor these particular hazard classes.Acute toxicityAs indicated in CLP Annex I, 3.1.3.3(b), there are two options to calculate acute toxicity ofMixture B: (i) treat the 'base powder' as an ingredient when calculating the ATE for MixtureB, or (ii) break the 'base powder' down into its component ingredients and only take over therelevant ingredients (CLP Annex I, 3.1.3.3(a) and 3.1.3.6.1) into the calculation for the ATE ofMixture B.Following option (i) it is first necessary to calculate the ATE mix of the 'base powder' takinginto account the non-ionic surfactant (other components can be excluded as LD 50 values are >2000 mg/kg):100ATEmixnCiATEiATEmixn100CiATEi100ATEmix 2778mg/kg18.0 500 The ATE mix for the 'base powder' can then be used for the calculation of the ATE mix forMixture B:ATE mix20.0277810018.0 7708.01800 2860mg/kgFollowing option (ii) it is only necessary to include the non-ionic surfactant from the 'basepowder' (present in Mixture B at 3.6%). Other ingredients in the 'base powder' can beexcluded as LD 50 > 2000 mg/kg for all of them. The calculation of the ATE mix for Mixture Bapplying option (ii):67

Guidance on the Application of the CLP Criteria Version 3.0 - November 201268ATE mix3.650010018.0 7708.01800 2860mg/kgBoth options indicate that the calculated ATE mix of Mixture B is > 2000 mg/kg. ThereforeMixture B is not classified as hazardous for acute toxicity by the oral route.N.B. If an acute oral toxicity test (i.e. an actual LD 50 value) was available for the 'basepowder' then this should be used in the calculation for the ATE of Mixture B.Skin corrosion/irritationWork out the actual levels of the 'base powder' ingredients in Mixture B and carry out thesummation method (CLP Annex I, Table 3.2.3) using the relevant ingredients:Mixture B does not contain any ingredients classified as Skin Corr. 1A, B or C thus MixtureB is not classified as Skin Corr. 1A, B or C.Mixture B does however contain 23 % ingredients classified as Skin Irrit. 2 (11% silicates,8% anionic surfactant and 4% anionic surfactant from the 'base powder'), as the content ofclassified ingredients are > 10% also Mixture B is classified as Skin Irrit. 2.Serious eye damage/eye irritationWork out the actual levels of the 'base powder' ingredients in Mixture B and carry out thesummation method (CLP Annex I, Table 3.3.3) using the relevant ingredients:Mixture B contains 40.6 % ingredients classified as Eye Dam.1 (18% oxygen bleach, 11%silicates, 8 % anionic surfactant and 3.6 % non-ionic surfactant), thus Mixture B is alsoclassified as Eye Dam.1.Respiratory sensitisationMixture B contains 0.7% of the ingredient 'enzymes' classified for respiratory sensitisation.However this is below the concentration triggering classification (CLP Annex I, Table 3.4.3)thus Mixture B is not classified as a respiratory sensitiser. However ingredient 'enzymes'trigger additional labelling information (CLP Annex II, 2.8).STOTMixture B does not contain any ingredients classified as STOT RE or STOT SE 1 or 2, but itcontains 11% of an ingredient classified as STOT SE 3 (respiratory tract irritation). Thegeneric concentration limit is 20 % for extrapolating the classification as STOT SE 3 from aningredient to the mixture (CLP Annex I, 3.8.3.4.5.), thus Mixture B does not triggerclassification as STOT SE 3 (respiratory tract irritation).Table 1.6.4.2(a) Ingredients in Mixture BIngredient % w/w Oral LD 50 (rat) ClassificationBase powderEye Dam.120.00 not tested(see list of ingredients below)Skin Irrit. 2Ox. Sol. 1Oxygen bleach 18.00 770 mg/kg Acute Tox. 4 (oral)Eye Dam. 1Silicates 11.00 3400 mg/kgEye Dam. 1Skin Irrit. 2STOT SE 3 (respiratory

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012tract irritation)Carbonate 7.00 4090 mg/kg Eye Irrit. 2Inorganic processing aid 11.30 > 5000 mg/kg Not classifiedBuilder 16.00 > 5000 mg/kg Not classifiedAnionic surfactant 8.00 1800 mg/kgAcute Tox. 4 (oral)Eye Dam. 1Skin Irrit. 2Bleach activator 5.00 > 5000 mg/kg Not classifiedEnzymes 0.70 > 2000 mg/kg Resp. Sens. 1Polycarboxylate 3.00 > 5000 mg/kg Not classifiedTotal: 100.00Table 1.6.4.2(b) Ingredient ' base powder 'Ingredient % w/w % in Mixture B Oral LD 50 (rat) ClassificationNon-ionic surfactant 18.00 3.6 500 mg/kgAcute Tox. 4 (oral)Eye Dam. 1Aquatic Acute 1Anionic surfactant 20.00 4.0 > 2000 mg/kgSkin Irrit. 2Eye Irrit. 2Builder 50.00 10.0 > 5000 mg/kg Not classifiedCarbonate 8.00 1.6 4090 mg/kg Eye Irrit. 2Inorganic processingaid4.00 0.8 > 5000 mg/kg Not classifiedTotal: 100.00 20.001.7 THE APPLICATION OF ANNEX VII1.7.1 IntroductionIn order to assist industry, especially small and medium enterprises (SMEs) to implementCLP, Annex VII to CLP contains translation tables to translate a classification derived inaccordance with DSD or DPD into a CLP classification.Article 61(5) Where a substance or mixture has been classified in accordance with Directive67/548/EEC or 1999/45/EC before 1 December 2010 or 1 June 2015 respectively, manufacturers,importers and downstream users may amend the classification of the substance or mixture using theconversion table in Annex VII to this Regulation.Note: Article 61 uses the term “conversion table” and Annex VII uses the term “translationtable”. These terms have the same meaning i.e. the tables in Annex VII that relateclassifications according to DSD or DPD to a classification according to CLP.Although conceptually similar, the coverage of CLP and the DSD or DPD is different. Insome places, there is a good relationship between the category of danger and corresponding69

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012R-phrases and hazard categories and corresponding hazard statements but in others, therelationship is less well defined. Additionally CLP introduces new hazard classes reflectinghazards that were not covered or only partly covered by DSD and DPD.While the tables in Annex VII explicitly point out where no translation is possible or whereminimum classification can be applied, they do not identify cases where CLP hazard classesor categories, not covered by the DPD and DSD, are required under CLP. In the particularcase of “no classification” under DPD, the table should not be used as there is no reasonableindication about a potential translation outcome.This guidance will help classifiers to identify where translations contained in the tables ofAnnex VII to CLP may not be precise and also help classifiers to use existing transportclassifications to fill some of the gaps.1.7.2 Use of Annex VII translation tablesAnnex VII Translation table from classification under Directive 67/548/EEC to classification underthis RegulationThis Annex includes a table to assist translation of a classification made for a substance or a mixtureunder Directive 67/548/EEC or Directive 1999/45/EC, respectively, into the correspondingclassification under this Regulation. Whenever data for the substance or mixture are available, anevaluation and classification shall be done in accordance with Articles 9 to13 of this Regulation.When classifying in accordance with CLP, the use of the tables contained in Annex VII isoptional. They can only be used to translate an existing classification provided that:- the substance was classified according to the DSD before 1 st December 2010 or themixture was classified according to the DPD before 1 st June 2015; and- there is no data (scientific or technical information) for the substance or mixtureavailable for an individual hazard class.When data for the substance or mixture is available for a hazard class, the substance ormixture must be classified in accordance with CLP criteria; the Annex VII tables must not beused. In practice, this could lead to an approach for a substance/mixture where some hazardclasses are re-classified using the Annex VII translation tables and other hazard classes are reclassifiedin accordance with CLP criteria.1.7.2.1 Applicability of the Annex VII translation tablesAs mentioned in section 1.7.1 of this document, the Annex VII translation tables do notalways give a direct translation. For certain hazard classes, including acute toxicity and STOTrepeated exposure, there is a recommended minimum classification in CLP, Annex VII Table1.1. This minimum classification should only be used if no additional hazard information isavailable (see also CLP Annex VI, 1.2.1).Table 1.7.2.1(a) of this document identifies where the use of the Annex VII translation tablesfor substances and mixtures requiring classification under DSD or DPD, may lead to aclassification that differs from one produced using the CLP criteria.In addition to the differences indicated in Table 1.7.2.1(a), attention is drawn to the fact thatfor some hazards the DPD generic concentration limits, to be applied for mixtures, werelowered under CLP. Lower generic concentration limits were set for skin corrosion (R34 andR35), severe eye damage and eye irritation (R41 and R36), skin irritancy (R38) andreproductive toxicity (R60, R61, R62 and R63). Where mixtures containing substances withrisk phrases R34 or R41 have been classified on basis of the hazards of individual ingredients,70

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012the use of the translation table will lead to an under-classification of the mixture. Therefore,for mixtures with these R-phrases, the use of the translation tables may not be appropriate andre-classification may be done by using the existing data.It is recommended that classifiers carefully consider the implications of these differencesbefore choosing to use the translation tables. Possible consequences from downstreamlegislation or Responsible Care issues need to be considered e.g. if the use of the translationtables increased the severity of the classification compared to using the CLP criteria, thiscould trigger additional duties under the Seveso Directive or national explosives legislation.Similarly a CLP hazard might not be identified by using the translation table which wouldhave been identified if the CLP criteria had been used, leading to risks or company/productimage and reputation issues.Table 1.7.2.1(b) contains additional translations, using the transport classification that can beused in addition to the translations in Annex VII to improve the quality of the translatedclassifications. However these translations also have certain restrictions on their applicability. The transport classification of named substances or mixtures may be based onexperience or certain events that are specific to transport The transport classification of named substances or mixtures in the transportregulations have not been systematically reviewed after the transport regulations wereadapted to take into account the GHS criteria in particular classes 3 and 6.1. In generalthe transport classification of named substances or mixtures should be used withcaution. The transport regulations include the concept of precedence of hazards. CLP does notapply a precedence of hazards and therefore substances or mixtures might need to beclassified in additional hazard classes under CLP which are not reflected in thetransport classification or are only considered as so-called subsidiary risks. There isusually insufficient information on subsidiary risks to allow a translation to CLPclassification to be made. Sometimes special provisions are linked to the entries in the Dangerous Goods Listwhich have to be met in order to be classified in the respective class for transport. Inthese cases the classification for the purposes of supply and use might be different.Sometimes one substance even has two different entries with two differentclassifications where one of the classifications is linked to one or more specialprovisions.If the translation table is used to re-classify a substance or mixture, the new classificationremains valid until either new data or change in composition requires the classification to bereviewed.In deciding whether or not to use the translation table and the additional guidance containedin this document, a classifier should balance the speed and ease of its use against theconsequences of the limitations. This judgment will be specific to each situation. Thisguidance will identify for which hazard classes the use of the translation table will give adifferent outcome from the direct application of the CLP criteria, and will explain why this isthe case. Where possible, the use of an available transport classification as additionalinformation is also described. This will help a classifier to make an informed decision aboutwhether to use the translation tables and additional information contained in this guidance orto re-classify using the CLP criteria.71

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Table 1.7.2.1(a) Hazard classes where reclassification using the translation tables gives a differentoutcome compared to reclassification using CLP criteriaClassifications underDSD or DPDE, R2E, R3Potentialoutcomestranslation1) Explosive.2) Organic peroxide3) Flammable solid4) Oxidising solid5) Self-reactive6) No classificationCommentsChange of classification criteria and method;individual treatmentSee Table 1.7.2.1(b) for additional information usingtransport classificationsO, R8 (liquid) Oxidising liquid All liquid substances or mixtures classified O,R8 areclassified as oxidising liquids under CLP.See Table 1.7.2.1(b) for additional information usingtransport classificationsO, R8 (solid) Oxidising solid The test methods for oxidising solids in 67/548/EECand CLP are different. Most solids classified O, R8 arealso classified as oxidising solids under CLP.F, R11 (solid) 1) Flammable solid1a) Possibly selfheatingin addition2) Self-reactiveF, R15 Substance or mixturewhich, in contact withwater, emit(s)flammable gas(es)See Table 1.7.2.1(b) for additional information usingtransport classificationsSolid substances or mixtures classified F, R11 may beclassified as flammable solids or self reactives underCLP. If classified as flammable solids, they mayadditionally be classified as self-heating.See Table 1.7.2.1(b) for additional information usingtransport classificationsSee Table 1.7.2.1(b) for additional information usingtransport classificationsTable 1.7.2.1(b) Additional information using transport classifications(Note that within transport, the term "substances" covers also mixtures in CLP terms)Transport classificationTransport Packing group,class division, type,group or codeand(sub)division(ifapplicable)PhysicalstateCLP-classificationHazard classHazardcategory,division,typegrouporRemarks72

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Class 1 Division 1.1Class 2 -GasesDivision 1.2Division 1.3Division 1.4Division 1.5Division 1.61 CompressedgasLiquid orsolidGaseousExplosives Division 1.1Division 1.2Division 1.3Division 1.4Division 1.5Division 1.6Compressedgas2 Liquefied gas. Gaseous Liquefiedgas.3 Refrigeratedliquefied gasGaseous4 Dissolved gas Gaseous5 Aerosoldispensers, class2.1Notrelevant(Articles)Gases underpressureFlammableaerosolsFlammable gases Gaseous FlammablegasesOxidising gases Gaseous OxidisinggasesClass 3 Packing group 1 Liquid FlammableliquidPacking group 2 Liquid FlammableliquidPacking group 3 Liquid FlammableliquidClass 4.1 Types B-F Solid orliquidClass 4.1(only readilycombustiblesolids)Class 4.1(only readilycombustiblesolids)Class 4.2PyrophoricSelf-reactivesubstancesPacking group II Solid FlammablesolidsPacking group III Solid FlammablesolidsPacking group ILiquidPyrophoricliquidsRefrigeratedliquefied gasDissolvedgasCategory 1Category 2Category 1Category 1Category 1Category 2Category 3Types B-FCategory 1Category 2Category 1Matching criteria.However, ifexplosives are unpackedor repacked,they have to beassigned to division1.1 unless the hazardis shown tocorrespond to one ofthe other divisions.This translation onlyapplies to the form inwhich the gas istransported. If it isused in a differentform, then theclassification has tobe amendedThe transportclassification does notdifferentiate betweenCategory 1 and 2flammable aerosolsCategory 2 flammablegases cannot beidentified using thetransport criteria73

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012SolidPyrophoricsolidsClass 4.2 Packing group II Solid Self-heatingsubstances andmixturesClass 4.2 Packing group III Solid Self-heatingsubstances andmixturesClass 4.3Class 5.1Class 5.1Packing group IPacking group IIPacking group IIIPacking group IPacking group IIPacking group IIIPacking group IPacking group IIPacking group IIILiquid orsolidSolidLiquidClass 5.2 Types B-F Solid orliquidClass 8 Packing group III Liquid orsolidSubstanceswhich incontact withwater emitflammablegasesOxidisingsolidOxidisingliquidOrganicperoxidesCorrosive tometalsCategory 1Category 1Category 2Category 1Category 2Category 3Category 1Category 2Category 3Category 1Category 2Category 3Types B-FCategory 1Applies only when thesubstance or mixtureis not classified C;R35 or C;R341.7.3 Additional considerations for re-classification due to changes in theclassification criteriaDue to changes in the classification criteria, and lowering of several GCLs for mixtures, CLPmay trigger classification for certain hazards which were not required by DPD or DSD.Table 1.7.3 (c) below identifies when a substance or mixture, that does not requireclassification and labelling according to DSD or DPD, may require classification andlabelling according to CLP.Table 1.7.3(c) Examples when classification may not be required under DSD and DPD, but may berequired under CLP74

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Non-classificationsunder DSD or DPDAdditionalhazards underCLPCommentsNon-classified explosives Explosive Certain explosives, not classified as E, R2 or E, R3,which are manufactured with the view to producinga practical explosive or pyrotechnical effect will beclassified as explosive under CLP.Self-reactiveor mixturesFlammable aerosolssubstancesGases under pressureSelf-heatingor mixturessubstancesSubstances or mixturesthat are corrosive tometals, but not corrosiveto skinMixtures containingsubstances with nonadditiveeffects for skincorrosion/irritation andeye damage/irritationMixtures containing 1-5% of R34 substances (andthus not classified)Self-reactivesubstanceFlammableaerosolGas underpressureSelf-heatingsubstance ormixtureCorrosive tometal1) Skincorrosive/seriouseye damage(Category 1)2) Skin/eyeirritant (Category2)Skin IrritantCategory 2See Table 1.7.2.1(b) for additional informationusing transport classificationsSelf-reactive substances or mixtures may not beidentified under the DSD.See Table 1.7.2.1(b) for additional informationusing transport classificationsFlammable aerosols are not explicitly identifiedunder DSD or DPD.See Table 1.7.2.1(b) for additional informationusing transport classificationsGases under pressure will not be identified as no Rphrase for gases under pressure currently exists. Theassignment of the correct group of a gas underpressure (compressed, liquefied or dissolved)depends on the physical state in which the gas ispackaged or handled. It therefore has to be assignedindividually. Note that the transport classificationmay be different.Self-heating substances or mixtures will not beidentified as no R phrase for self-heating substancesor mixtures currently exists. See Table 1.7.2.1(b)for additional information using transportclassificationsSubstances or mixtures that are corrosive to metals,but not corrosive to skin, will not be identified as noR phrase for corrosive to metals currently exists.See Table 1.7.2.1(b) for additional informationusing transport classificationsThe concept of non-additive effects for skincorrosion/irritation and eye damage/irritation is notexplicitly considered in the current Directives (seeCLP Annex I, Tables 3.2.4 and 3.3.4).The generic concentration limit is 1 % in the CLPbut the corresponding limit is 5 % in the DPD.75

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Mixtures containing 10 –20 % of R38 substances(and thus not classified)Mixtures containing 1-3% of R41 or R34substances (and thus notclassified)Mixtures containing 3-5% of R41 or R34substances (and thus notclassified)Mixtures containing 10 –20 % of R36 substances(and thus not classified)Mixtures containing 3 – 5% of R62 or R63substances (and thus notclassified)Mixtures containing 0.3-0.5 % of R60 or R61substances (and thus notclassified)1) Skin irritantCategory 21) Eye irritantCategory 21) Serious eyedamage Category11) Eye irritantCategory 21) Reproductivetoxicant,Category 21) Reproductivetoxicant Category1A/1BThe generic concentration limit is 10% in the CLPbut the corresponding limit is 20% in the DPD.The lower generic concentration limit is 1% in theCLP but the corresponding limit is 5% in the DPD.The generic concentration limit is 3 % in the CLPbut the corresponding limit is 10 % in the DPD.The generic concentration limit is 10 % in the CLPbut the corresponding limit is 20 % in the DPD.The generic concentration limit is 3 % in the CLPbut the corresponding limit is 5 % in the DPD.The generic concentration limit is 0.3 % in the CLPbut the corresponding limit is 0.5 % in the DPD.76

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122 PART 2: PHYSICAL HAZARDS 422.1 INTRODUCTION2.1.1 General remarks about the prerequisites of classification and testingThe purpose of this chapter is to give some general guidance with respect to the generation oftest data for physical hazards and their interpretation. The intention of CLP is to identifyhazards of chemical substances and mixtures and to provide a systematic approach – usingclassification - to communicate them based on harmonized criteria. The classification processinvolves three steps:1. Gathering of relevant information regarding the hazards of a substance or mixture(articles 5 – 8);2. Evaluation of hazard information to ascertain the hazards associated with thesubstance or mixture (articles 9 ff); and3. A decision on whether the substance or mixture will be classified as a hazardoussubstance or mixture and the degree of hazard, where appropriate, by comparison ofthe data with agreed hazard classification criteria (article 13).Generally for both, substances and mixtures, testing is required to determine physical hazardsincluding the physico-chemical properties necessary for the respective classification unlessalternative methods are specifically permitted. Before undertaking testing of substances,enquiries should be made to ascertain the availability of data, e.g. flash points, on thesubstance.2.1.2 SafetyIn most cases, the classification is based on test data which are determined in a laboratory.Special care is required when new or unknown substances or mixtures are tested. If possible,preliminary tests should be carried out before larger quantities are handled. Appendix 6 of theUN Manual of Tests and Criteria (UN-MTC) ('Screening procedures') allows gatheringvaluable information about physico-chemical properties based on small-scale tests. Furtheraspects of safety are given in the general introduction, Section 1.4 of the UN-MTC or withinthe individual test procedures.2.1.3 General conditions for testingSamples offered for testing must in all aspects be representative of the substance or mixture tobe classified. Therefore, it is helpful to characterise or specify the sample for the purposes ofdocumentation (i.e. batch number, production code etc.). Further characterisation (i.e.analysis) is highly recommended in cases where the presence of diluents, activators,stabilisers or moisture may influence the outcome of the test.42 The guidance provided in this chapter is based on the classification criteria from the original version of theCLP Regulation (EC) No 1272/2008. This chapter is currently being updated based on the 2 nd ATP to the CLPRegulation and planned for a future update of this document in 2013.77

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In some cases, additional parameters (physical condition, particle size, density, crystalstructure) may influence the test result. Where relevant, the test should be performed on thesubstance or mixture in the appropriate physical form where changes in that form mayinfluence the outcome of the test (see also Articles 5 and 6 of CLP and Section 1.2 on formand physical state).2.1.4 Physical stateThe physical state determines which hazard classes should be considered for testing. Thedefinitions for gases, liquids and solids are given in Annex I, Part 1 of CLP:Annex I: Part 1, 1.0.Gas means a substance which:Definitions(i) at 50 °C has a vapour pressure greater than 300 kPa (absolute); or(ii) is completely gaseous at 20 °C at a standard pressure of 101.3 kPa;Liquid means a substance or mixture which:(i) at 50 °C has a vapour pressure of not more than 300 kPa (3 bar);(ii) is not completely gaseous at 20 °C and at a standard pressure of 101,3 kPa; and(iii) which has a melting point or initial melting point of 20 °C or less at a standard pressure of 101,3kPa;Solid means a substance or mixture which does not meet the definitions of liquid or gas.In some cases (i.e. viscous substances or mixtures), a specific melting point cannot bedetermined. Such substance or mixture shall be regarded as a liquid if either the result of theASTM D 4359-90 test (standard test method for determining whether a material is a liquid ora solid) indicates ‘liquid’ or the result of the test for determining fluidity (penetrometer test)prescribed in Section 2.3.4 of Annex A of ADR indicates ’not pasty’.2.1.5 QualityThe determination of data should be based on the methods named in Annex I, Part 2 of CLP.For most hazard classes (except gases and liquids) in Annex I, Part 2 there is reference madeto the UN-MTC which gives very detailed descriptions of the test methods. For gases andliquids there are references to international standards. Whenever possible, the methods usedshould be validated. Any deviation from the test procedure or standard should be documentedand, if necessary, justified.The reliability of all test results used for the classification of dangerous substances isimportant and therefore their transparency and comparability must be ensured.For these purposes, CLP requires in Article 8 the following:Article 8 (5)Where new tests for physical hazards are carried out for the purposes of this Regulation, they shall becarried out, at the latest from 1 January 2014, in compliance with a relevant recognised quality systemor by laboratories complying with a relevant recognised standard.Even though the quality requirement does not become immediately effective, it is highlyrecommended to do so if reasonably possibly. In general, the following alternative strategiescan be pursued:78

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121. compliance with the principles of good laboratory practice (GLP) (as formerlyrequired by the DSD)2. application of EN ISO/IEC 17025 "General requirements for the competence oftesting and calibration laboratories" as a relevant recognised standard.3. other internationally recognised standards of comparable scope.Any testing organisation that carries out physical hazard tests for classification purposes cantherefore choose how to fulfill the quality requirements of CLP.2.2 EXPLOSIVES2.2.1 IntroductionThe classification of substances, mixtures and articles in the class of explosives and furtherallocation to a division is a very complex procedure. Reference to Part I of the UN RTDGManual of Testing and Criteria (MTC) and related expertise are necessary.The GHS classification system is almost entirely adopted of the UN Recommendations on theTransport of Dangerous Goods, which is very appropriate for transport and also storage ofpackaged explosives.The explosive properties of substances and mixtures regarding their stability and sensitivityare only investigated within test series 1, 2 and 3 during the acceptance procedure.Subsequent tests for the assignment to the divisions 1.1, 1.2, 1.3 and 1.4 (test series 6) arecarried out with the packaged substance / mixture or articles. The type of packaging maysignificantly influence the test outcome.For unpacked or repacked explosive substances and mixtures there are some deficiencies inthe hazard communication of the GHS, especially for substances and mixtures, which areprovisionally accepted in the class of explosives but later are rejected from this class due totheir packaging in the assignment procedure. These substances and mixtures have explosiveproperties but there might be no hazard communication about these properties due to thesubsequent classification in a hazard class other than the class of explosives. The example formusk xylene (see Section 2.2.6.2) clarifies this issue. The results of test series 6 for muskxylene in the specified packaging lead to the exclusion of this substance from the hazard classof explosives. But musk xylene on its own (unpacked) shows explosive properties due toheating under confinement (Koenen test). Also repacking of the substance in a packagingother than tested can result in a completely different outcome of test series 6.This issue is not sufficiently clarified under GHS, but should be kept in mind by everyoneapplying CLP.Some R-phrases which are not yet covered by hazard classes in the GHS are added assupplemental hazard statements in Annex II part 1 of CLP. The following EU hazardstatements are important in connection with explosive properties:EUH001 “Explosive when dry”EUH044 “Risk of explosion if heated under confinement”For more information on additional labelling provisions, see Section 2.2.4.2.2.2 Definitions and general considerations for the classification of explosives79

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The following definition is given in CLP for the class of explosives.Annex I: 2.1.1.1. The class of explosives comprises(a)(b)(c)Explosive substances and mixtures;Explosive articles, except devices containing explosive substances or mixtures in such quantityor of such a character that their inadvertent or accidental ignition or initiation shall not cause anyeffect external to the device either by projection, fire, smoke, heat or loud noise; andSubstance, mixtures and articles not mentioned under (a) and (b) which are manufactured withthe view to producing a practical, explosive or pyrotechnic effect.Additional remark related to 2.1.1.1 (a) (reference to UN RTDG, Model Regulations, Volume1):A substance or mixture which is not itself an explosive but which can form an explosiveatmosphere of gas, vapour or dust is not included in this class.A substance or mixture with explosive properties, but where the predominant hazard iscovered by another class (e.g. organic peroxides, self-reactive substances and mixtures), is notincluded in the class of explosives.In addition the following definitions apply for explosives:Annex I: 2.1.1.2. An explosive substance or mixture is a solid or liquid substance or mixture ofsubstances which is in itself capable by chemical reaction of producing gas at such a temperature andpressure and at such a speed as to cause damage to the surroundings.Pyrotechnic substances are included even when they do not evolve gases.A pyrotechnic substance or mixture is a substance or mixture of substances designed to produce aneffect by heat, light, sound, gas or smoke or a combination of these as the result of non-detonative selfsustainingexothermic chemical reactions.An unstable explosive is an explosive which is thermally unstable and/or too sensitive for normalhandling, transport and use.An explosive article is an article containing one or more explosive substances or mixtures.A pyrotechnic article is an article containing one or more pyrotechnic substances or mixtures.An intentional explosive is a substance, mixture or article which is manufactured with a view toproduce a practical explosive or pyrotechnic effect.2.2.3 Classification of substances, mixtures or articles as explosives2.2.3.1 Identification of hazard informationInformation on the following types of hazards is relevant for the evaluation of substances,mixtures and articles for the class of explosives: sensitivity to shock effects of heating and ignition under confinement thermal stability sensitiveness to impact and friction80

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012 mass explosion hazard projection hazard fire and radiant heat hazard2.2.3.2 Screening procedures and waiving of testingThe screening procedure is described in: CLP, Annex I, Part 2, paragraphs 2.1.4.2 and 2.1.4.3 Appendix 6 of the UN Recommendations on the Transport of Dangerous Goods,Manual of Tests and Criteria Technical Guidance Document on the Information Requirements for REACH, Part 2EWG 1-7, REACH Implementation Project (RIP) 3.3 Phase 2, chapter 7.1.11.3The screening procedure may be used for new substances which are suspected of havingexplosive properties. It should not be used for substances manufactured with the intention ofproducing a practical explosive or pyrotechnic effect.Explosive properties are associated with the presence of certain chemical groups in amolecule which can react to produce very rapid increases in temperature or pressure. Thescreening procedure is aimed at identifying the presence of such reactive groups and thepotential for rapid energy release.If the screening procedure identifies the material to be a potential explosive or if the substanceis a mixture containing any known explosives, the classification (acceptance) procedure forthe class of explosives (see Section 2.2.3.5.1) should be applied. If the exothermicdecomposition energy of organic materials is less than 800 J/g, neither a Series 1 type (a)propagation of detonation test nor a Series 2 type (a) test of sensitivity to detonative shock isrequired.A substance or mixture shall not be classified as explosive:(a) When there are no chemical groups associated with explosive properties present in themolecule. Examples of groups which may indicate explosive properties are: C-C unsaturation (e.g. acetylenes, acetylides, 1, 2-dienes), C-Metal, N-Metal (e.g. Grignard reagents, organo-lithium compounds), Contiguous nitrogen atoms (e.g. azides, aliphatic azo compounds, diazonium salts,hydrazines, sulphonylhydrazides) Contiguous oxygen atoms (e.g. peroxides, ozonides) N-O (e.g. hydroxyl amines, nitrates, nitro compounds, nitroso compounds, N-oxides,1,2-oxazoles) N-halogen (e.g. chloramines, fluoroamines) O-halogen (e.g. chlorates, perchlorates, iodosyl compounds)or(b) When the substance or mixture contains chemical groups associated with explosiveproperties which include oxygen and the calculated oxygen balance is less than -200.The oxygen balance is calculated for the chemical reaction:81

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012C H O y zx y z x O24 2 y x OCO 2 H2 2 using the formula:2x y2z 1600molecular weightor(c) When the organic substance or a homogenous mixture of organic substances containschemical groups associated with explosive properties but the exothermic decompositionenergy is less than 500 J/g and the onset of exothermic decomposition is below 500 ºC. (Thetemperature limit is to prevent the procedure being applied to a large number of organicmaterials which are not explosive but which will decompose slowly above 500 ºC to releasemore than 500 J/g.) The exothermic decomposition energy may be determined using asuitable calorimetric technique.or(d) For mixtures of inorganic oxidising substances with organic material(s), theconcentration of the inorganic oxidising substance is: less than 15 % by mass, if the oxidising substance is assigned to Categories 1 or 2; less than 30 % by mass, if the oxidising substance is assigned to Category 3.822.2.3.3 Classification criteriaThe criteria for the classification of explosives are given in the following tables.Annex I: 2.1.2.1. Substances, mixtures and articles of this class are classified as an unstableexplosive on the basis of the flowchart in Figure 2.1.2. The test methods are described in Part I of theUN Recommendations on the Transport of Dangerous Goods, Manual of Tests and Criteria2.1.2.2. Substances, mixtures and articles of this class, which are not classified as an unstableexplosive, shall be assigned to one of the following six divisions depending on the type of hazardthey present:(a) Division 1.1 Substances, mixtures and articles which have a mass explosion hazard (amass explosion is one which affects almost the entire quantity present virtuallyinstantaneously).(b) Division 1.2 Substances, mixtures and articles which have a projection hazard but not amass explosion hazard.(c) Division 1.3 Substances, mixtures and articles which have a fire hazard and either a minorblast hazard or a minor projection hazard or both, but not a mass explosion hazard:(i)(ii)both.combustion of which gives rise to considerable radiant heat; orwhich burn one after another, producing minor blast or projection effects or(d) Division 1.4 Substances, mixtures and articles which present no significant hazard:Substances, mixtures and articles which present only a small hazard in the event ofignition or initiation. The effects are largely confined to the package and noprojection of fragments of appreciable size or range is to be expected. An externalfire shall not cause virtually instantaneous explosion of almost the entire contents ofthe package.(e) Division 1.5 Very insensitive substances or mixtures which have a mass explosion

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012hazard:Substances and mixtures which have a mass explosion hazard but are so insensitivethat there is very little probability of initiation or of transition from burning todetonation under normal conditions.(f) Division 1.6 Extremely insensitive articles which do not have a mass explosion hazard:Articles which contain only extremely insensitive detonating substances or mixturesand which demonstrate a negligible probability of accidental initiation orpropagation.2.1.2.3. Explosives, which are not classified as an unstable explosive, shall be classified in one of thesix divisions referred to in section 2.1.2.2 based on Test Series 2 to 8 in Part I of the UNRecommendations on the Transport of Dangerous Goods, Manual of Tests and Criteria according tothe results of the tests laid down in Table 2.1.1:Table 2.1.1Criteria for explosivesCategoryUnstable explosives orexplosives of Divisions 1.1to 1.6CriteriaFor explosives of Divisions 1.1 to 1.6, the following are the core set oftests that need to be performed:Explosibility: according to UN Test Series 2 (section 12 of the UNRecommendations on the Transport of Dangerous Goods, Manual ofTests and Criteria). Intentional explosives 43 shall not be subject to UNTest Series 2.Sensitiveness: according to UN Test Series 3 (section 13 of the UNRecommendations on the Transport of Dangerous Goods, Manual ofTests and Criteria).Thermal stability: according to UN Test 3(c) (sub-section 13.6.1 of theUN Recommendations on the Transport of Dangerous Goods, Manualof Tests and Criteria).Further tests are necessary to allocate the correct Division.Certain physical hazards (due to explosive properties) are altered by dilution, as is the case fordesensitized explosives, by inclusion in a mixture or article, packaging or other factors.Explosive substances and mixtures wetted with water or alcohols, or diluted with othersubstances to suppress their explosive properties, may be treated differently in terms ofclassification and other hazard classes may apply, according to their physical properties.Where the test is conducted in the package form and the packaging is changed, a further testshall be conducted where it is considered that the change in packaging will affect the outcomeof the test.Classification tests should be performed on the substance or mixture as presented and used.If the same chemical is to be presented in a physical form different from that which was testedand which is considered likely to materially alter its performance in a classification test, thesubstance or mixture must also be tested in the new form.2.2.3.4 Testing and evaluation of hazard information43This comprises substances, mixtures and articles which are manufactured with a view to producing apractical, explosive or pyrotechnic effect.83

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Where test data are available, these shall be evaluated against the set criteria for classificationand labelling.When the screening procedure indicates that a substance or mixture may possess explosiveproperties, a cautious approach when performing the tests is necessary to ensure safehandling.For information on the test procedures see the following Section 2.2.3.5 where the individualtest series are described in context with the respective decision logic.The test procedures for the classification of explosives are described in detail in the Part I ofthe UN-MTC.2.2.3.5 Classification procedure and decision logicsAny substance, mixture or article having or suspected of having explosives characteristicsshall be considered for classification in the hazard class of explosives. Substances, mixturesand articles classified in this hazard class shall be assigned to the appropriate division or asunstable explosive.The classification is divided into two stages, the acceptance procedure and the assignmentprocedure.In the acceptance procedure, the potential of a substance, mixture or article to explode shouldbe ascertained and its stability and sensitivity shown to be acceptable. If the substance,mixture or article is not characterised as unstable explosive and is provisionally accepted intothe class of explosives, it is then necessary to ascertain the correct division by the assignmentprocedure. The further subdivision into compatibility groups A to S is described in detail inthe UN-RTDG, section 2.1.1. The compatibility groups and their recommended combinationidentify types of explosives which are deemed to be compatible, e.g. for combined storage ortransportation and can therefore be used to distinguish technical requirements (especially) inthese sectors.The tests for acceptance and the further tests to determine the correct division are groupedinto eight test series. Classification procedures, test methods and criteria are described indetail in Part I of the UN-MTC.NOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.84

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: Figure 2.1.2Procedure for provisional acceptance of a substance, mixture or article in the class of explosives(Class 1 for transport)2.2.3.5.1 Acceptance procedure85

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The acceptance procedure is used to determine whether or not a substance, mixture or articleis a candidate for the class of explosives or is an unstable explosive.The test methods used for deciding on provisional acceptance into the class of explosives aregrouped into four series, numbered 1 to 4 (see CLP Annex I, Figure 2.1.2).The numbering of test series 1 to 4 relates to the sequence of assessing the results rather thanthe order in which the tests are conducted. It may be important for the safety ofexperimenters that certain tests, using small amounts of material, be conducted firstbefore proceeding to experiment with larger quantities. To start the testing procedure withtest series 3 is highly recommended, because these tests involve relatively small sample sizes,which reduces the risk to test personnel.Test series 1Within test series 1 the question "Is it an explosive substance / mixture?" is answered on thebasis of international definitions of an explosive substance and the results of three types ofseries 1 test enable to assess possible explosive effects. The question is answered "yes" if a"+" is obtained in any of the three types of test. If the answer is “no”, the substance / mixtureis rejected from this class; it is not an explosive.The three types of test used are (recommended test is indicated within brackets):Type 1 (a): a shock test with defined booster and confinement to determine the abilityof the substance to propagate a detonation (UN Gap test):Type 1 (b): a test to determine the effect of heating under confinement (Koenen test);andType 1 (c): a test to determine the effect of ignition under confinement (time/pressuretest).Test series 2Series 2 tests are used to answer the question "Is the substance / mixture too insensitive foracceptance into this Class?". In general, the basic apparatus and method used is the same asthat for Test Series 1 but with less stringent criteria, e.g. in the case of gap tests, the gap usedis greater than zero. The question is answered "no" if a "+" is obtained in any of the threetypes of test. If the answer is “yes”, the substance / mixture is rejected from this class; it is notan explosive.If the substance / mixture is excluded at this point without performing test series 3, noinformation about the thermal stability and the sensitivity to mechanical stimuli (impact,friction) of the substance or mixture will be available. Also for this reason a generalperformance of test series 3 is highly recommended.The following three types of test are used (recommended test is indicated within brackets):Type 2 (a): a shock test with defined initiation system and confinement to determinesensitivity to shock (UN Gap test);Type 2 (b): a test to determine the effect of heating under confinement (Koenen test);andType 2 (c): a test to determine the effect of ignition under confinement (time/pressuretest).86

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012If the substance is manufactured with a view to producing a practical explosive or pyrotechniceffect, it is unnecessary to conduct Test Series 1 and 2.Test series 3Test series 3 is used to answer the questions "Is the substance / mixture thermally stable?" and"Is the substance / mixture too dangerous in the form in which it was tested?" This involvestests for determining the sensitiveness of the substance to mechanical stimuli (impact andfriction), and to heat and flame.The following four types of tests are used (recommended test is indicated within brackets):Type 3 (a): a falling weight test to determine sensitiveness to impact (BAMFallhammer);Type 3 (b): a friction, or impacted friction, test to determine sensitiveness to friction(BAM friction apparatus);Type 3 (c): an elevated temperature test to determine thermal stability (thermalstability test at 75 °C); andType 3 (d): an ignition test to determine the response of a substance to fire (small scaleburning test)The first question is answered "no" if a "+" is obtained in test type 3(c) and the substance /mixture is considered as thermally unstable and is classified as an unstable explosive.The second question is answered "yes" if a "+" is obtained in any of the test types 3(a), 3(b) or3(d). If a "+" is obtained, the substance / mixture may be encapsulated or otherwisedesensitized or packaged to reduce its sensitiveness to external stimuli or is classified as anunstable explosive.Test series 4Series 4 tests are intended to answer the question "Is the article, packaged article or packagedsubstance too dangerous?". Conditions which may occur during supply and use include high/low temperature and high relative humidity, vibration, bumping and dropping.The two types of test to be carried out are:Type 4 (a): a test of thermal stability for articles; andType 4 (b): a test to determine the hazard from dropping.The question is answered "Yes" if a "+" is obtained in either test type 4 (a) or 4 (b) and thearticle is classified as an unstable explosive.It is important to note that a substance / mixture which fails test series 2 may still, if properlypackaged, leave the class of explosives provided that the product is not designed to have anexplosive effect and does not exhibit any explosive hazard in test series 6 of the assignmentprocedure (see example for musk xylene). Such an exclusion from the class of explosives isrestricted to the specific type and size of package tested.Especially for substances / mixtures, which have explosive properties according to test series1 and 2 but can leave the class of explosives after test series 6 due to proper packaging, it isnecessary to communicate these properties in the Safety Data Sheet (SDS). Furthermore, the87

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012results from test types 3 (a) and 3 (b) should be documented in the SDS when they meet thecriteria of the EU test method A 14 in Council Regulation (EC) No 440/2008.2.2.3.5.2 Assignment procedure to a divisionThe assignment procedure to one of six divisions, depending on the type of hazard theypresent, applies to all substances, mixtures and/or articles that are candidates for class ofexplosives. A substance or article should be assigned to the division which corresponds to theresults of the tests to which the substance, mixture or article, as offered for supply and use,has been subjected. Other test results, and data assembled from accidents which haveoccurred, may also be taken into account.The test methods used for assignment to a division are grouped into three series - numbered 5to 7 - designed to provide the information necessary to answer the questions in Figure 2.1.3 inCLP.NOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.88

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: Figure 2.1.3Procedure for assignment to a division in the class of explosives (Class 1 for transport)89

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Test series 5The results from three types of series 5 tests are used to answer the question "Is it a veryinsensitive explosive substance with a mass explosion hazard?"The test types are (recommended test is indicated within brackets):Type 5 (a): a shock test to determine the sensitivity to intense mechanical stimulus(cap sensitivity test ) ;Type 5 (b): thermal tests to determine the tendency for transition from deflagration todetonation (USA DDT test); andType 5 (c): a test to determine if a substance, when in large quantities, explodes whensubjected to a large fire.The question is answered "No" if a "+" is obtained in any of the three test types. A candidatefor Division 1.5 should pass one test of each type.Test series 6The results from three types of series 6 tests are used to determine which division, amongstDivisions 1.1, 1.2, 1.3 and 1.4, corresponds most closely to the behaviour of a product if aload is involved in a fire resulting from internal or external sources or an explosion frominternal sources. The results are also necessary to assess whether a product can be assigned toCompatibility Group S of Division 1.4 and whether or not it should be excluded from thisclass. Test series 6 should be applied to packages of explosive substances and articles in thecondition and form in which they are offered for supply and use.The three types of test are (recommended test is indicated within brackets):Type 6 (a): a test on a single package to determine if there is mass explosion of thecontents (single package test);Type 6 (b): a test on packages of an explosive substance or explosive articles, or nonpackagedexplosive articles, to determine whether an explosion ispropagated from one package to another or from a non-packaged article toanother (stack test); andType 6 (c): a test on packages of an explosive substance or explosive articles, or nonpackagedexplosive articles, to determine whether there is a massexplosion or a hazard from dangerous projections, radiant heat and/orviolent burning or any other dangerous effect when involved in a fire(bonfire test).Test types 6 (a), 6 (b) and 6 (c) are performed in alphabetical order. However, it is not alwaysnecessary to conduct tests of all types. Test type 6 (a) may be waived if explosive articles arecarried without packaging or when the package contains only one article. Test type 6 (b) maybe waived if in each type 6 (a) test:- The exterior of the package is undamaged by internal detonation and/or ignition;or- The contents of the package fail to explode, or explode so feebly as wouldexclude propagation of the explosive effect from one package to another in testtype 6(b).90

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Test type 6(c) may be waived if, in a type 6(b) test, there is practically instantaneousexplosion of virtually the total contents of the stack. In such cases the product is assigned toDivision 1.1.If a substance gives a "—" result (no propagation of detonation) in the Series 1 type (a) test,the 6(a) test with a detonator may be waived.If a substance gives a "—" result (no or slow deflagration) in a Series 2 type (c) test, the 6 (a)test with an igniter may be waived.Test series 7The question "Is it an extremely insensitive explosive article?" is answered by series 7 testsand any candidate for Division 1.6 should pass one of each of the ten types of test comprisingthe series. The first six types of test (7(a)-7(f)) are used to establish if a substance is anExtremely Insensitive Detonating Substance (EIDS) and the remaining four types of test(7(g), 7(h), 7(j) and 7 (k)) are used to determine if an article containing an EIDS may beassigned to Division 1.6.Test series 7 aims at military explosives and is generally not relevant for explosives for civiluse. Therefore the individual tests are not described here. If needed, they can be found in theUN Manual of Test and Criteria, Part I, Section 17.Test series 8The question "Is the substance a candidate for "ammonium nitrate emulsion or suspension orgel, intermediate for blasting explosives (ANE)?" (CLP Annex I, Figure 2.1.4) is answered byseries 8 tests and any candidate should pass each of the three tests comprising the series. Thethree test types are (recommended test is indicated within brackets):Type 8 (a): a test to determine the thermal stability (Thermal Stability Test for ANE);Type 8 (b): a shock test to determine sensitivity to intense shock (ANE gap test); andType 8 (c): a test to determine the effect of heating under confinement (Koenen test).Test series 8 should be used to establish whether an ammonium nitrate emulsion orsuspension or gel, intermediate for blasting explosives (ANE) shall be classified as anoxidising liquid or solid. Substances failing any of the tests shall be classified as explosives(Division 1.1. or 1.5) or as an unstable explosive in accordance with CLP Annex I, Figure2.1.4.91

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 2.1.4Procedure for classification of ammonium nitrate emulsions, suspensions or gels92

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.2.4 Hazard communication for explosives2.2.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: Table 2.1.2Label elements for explosivesClassificationUnstableDivision 1.1 Division 1.2 Division 1.3 Division 1.4 DivisionDivisionExplosive1.51.6GHSPictogramsSignal Word Danger Danger Danger Danger Warning Danger No signalwordHazardStatementPrecautionaryStatementPreventionPrecautionaryStatementResponsePrecautionaryStatementStoragePrecautionaryStatementDisposalH200:UnstableExplosiveP201P202P281P372P373P380H201:Explosive;massexplosionhazardP210P230P240P250P280P370+P380P372P373H202:Explosive;severeprojectionhazardP210P230P240P250P280P370+P380P372P373H203:Explosive;fire, blast orprojectionhazardP210P230P240P250P280P370+P380P372P373H204: Fire orprojectionhazardP210P240P250P280P370+P380P372P373H205:Maymassexplodein fireP210P230P240P250P280P370+P380P372P373NohazardstatementNoprecautionarystatementNoprecautionarystatementP401 P401 P401 P401 P401 P401 No precautionarystatementP501 P501 P501 P501 P501 P501 NoprecautionarystatementThe wording of the Precautionary Statements is found in CLP Annex IV, Part 2.2.2.4.2 Additional labelling provisionsAccording to CLP Annex I, 2.1.3, unpackaged explosives or explosives repacked inpackaging other than the original or similar packaging shall have the following labelelements:Annex I: 2.1.3.(a)Hazard communicationthe pictogram: exploding bomb;93

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(b)(c)the signal word: “Danger”; andthe hazard statement: 'explosive; mass explosion hazard'unless the hazard is shown to correspond to one of the hazard categories in Table 2.1.2, in which casethe corresponding symbol, signal word and/or the hazard statement shall be assigned.Supplemental hazard statements shall be included in the section for supplemental informationon the label and are defined in Annex II of CLP as follows:Annex II: 1.1.1. EUH001 “Explosive when dry”For explosive substances and mixtures as referred to in chapter 2.1 of part 2 of Annex I, placed on themarket wetted with water or alcohols or diluted with other substances to suppress their explosivespropertiesAnnex II: 1.1. 6 EUH044 “Risk of explosion if heated under confinement”For substances and mixtures not in themselves classified as explosive in accordance with section 2.1of part 2 of Annex I, but which may nevertheless display explosive properties in practice if heatedunder sufficient confinement. In particular, substances which decompose explosively if heated in asteel drum do not show this effect if heated in less-strong containers.2.2.5 Re-classification of substances and mixtures classified as explosive accordingto DSD or already classified for transport2.2.5.1 Re-classification of substances and mixtures classified inaccordance with DSDA direct “translation” from classification according to DSD or DPD to the GHS classificationaccording to CLP is not possible.A lot of substances and mixtures which are labelled with the symbol “E” and the risk phrasesR2 or R3 in accordance with DSD will be classified as "explosive" under CLP and therespective division can be derived from the transport classification. However, there are alsomany substances and mixtures which will be classified in other hazard classes, such asorganic peroxides, self-reactives, flammable solids (e.g. musk xylene) or oxidising solids (e.g.troclosone).In DSD explosive properties are determined by the EU test method A.14 as described inRegulation (EC) No 440/ 2008 (former Annex V to DSD). The EU test method A.14 is basedon the sensitivity of substances and mixtures to thermal and mechanical stimuli.The EU test method A.14 comprises three parts: thermal sensitivity test to determine the effect of heating under confinement (Koenentest) mechanical sensitivity test to determine the sensitivity to impact mechanical sensitivity test to determine the sensitivity to frictionThe criterion linked of the thermal sensitivity test to determine the effect of heating underconfinement (Koenen test) is equal in EU test method A.14 and GHS test series 2.At first glance, the tests for mechanical sensitivity to impact and friction also seem to be thesame in EU test method A.14 and GHS test series 3. However, the questions to be answeredby these tests and the criteria are different. The results of tests 3(a) and 3(b) in the GHS lead94

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012to the decision whether a substance/mixture is too sensitive to mechanical stimuli. For thispurpose, lower limits are stated. On the other hand, upper limits as defined in the A.14 testmethod result in the decision whether a substance/mixture has an explosive hazard.Some additional remarks are necessary regarding differences between the above mentionedsupplemental hazard statements EUH001 and EUH044 and their respective R-phrases R1 andR44. These differences originate from the different systematic approach of classifyingexplosives and explosive properties, respectively.EUH001 can be assigned only to such explosive substances and mixtures of the hazard classof explosives which are properly desensitised to fulfil the criteria of the future hazard class“Desensitised Explosives” (and therefore do not meet the criteria of the hazard class ofexplosives).Risk phrase R1 shall be assigned to all explosive substances and preparations (evaluated bytest method A14 and which have E; R2 or R3 in the undiluted state) put on the market insolution or in a wetted form.The criteria for the assignment of R44 according to DSD, Annex VI are defined as follows:“For substances and preparations not in themselves classified as explosive in accordance withsection 2.2.1 above but which may nevertheless display explosive properties in practice ifheated under sufficient confinement. For example, certain substances which woulddecompose explosively if heated in a steel drum do not show this effect if heated in lessstrongcontainers.”It has to be mentioned that there is a slight difference between the criteria for EUH044 in CLPand the criteria for R44 in DSD, Annex VI. According to the DSD, Annex VI labelling withR44 is possible for substances and mixtures which do not fulfil the criteria for R3 or R2(evaluated by test method A.14). According to CLP labelling with EUH044 is possible forsubstances and mixtures which are not classified as explosive.2.2.5.2 Relation to transport classificationNormally, the transport classification can be translated one-to-one into the CLP classificationfor explosives, which are packaged in authorised transport packaging.For the use of other packaging’s or for unpacked substances the additional labellingprovisions (see Section 2.2.4.2) have to be observed or re-testing is necessary.2.2.6 Examples of classification for explosivesExamples are given below for the classification of substances; however, these are also validfor the classification of mixtures.2.2.6.1 Example of substances and mixtures fulfilling theclassification criteriaA) RESULTS FROM APPLICATION OF THE ACCEPTANCE PROCEDURE0. General data:0.1 Name of the substance / mixture Hexanitrostilbene1. Is the substance a candidate forammonium nitrate emulsion,suspension or gel, intermediate forblasting explosive ANE?No95

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122. Is the substance manufactured withthe view to producing a practicalexplosive or pyrotechnic effect?3. Test Series 3Yes3.1 Thermal stability: 75 °C/48 hour test (test3(c))3.2 Impact sensitivity: BAM Fallhammer test(test 3(a)(ii))3.3 Friction sensitivity: BAM friction test (test3(b)(i))4. Is the substance thermally stable? Yes5. Is the substance too dangerous in theform in which it was tested?6. Conclusion: PROVISIONALLYACCEPT INTO THISCLASS10.1 Exit: Apply the assignmentprocedureNoResult: "—",thermally stableResult: Limitingimpact energy 5 JResult:Limiting load> 240 N"—", not toodangerous in formtested"—", not toodangerous in formtestedB) RESULTS FROM APPLICATION OF THE ASSIGNMENT PROCEDURE1. Is the substance a candidate forDivision 1.5?2. Test Series 6NoResult: Package thesubstance2.1 Effect of initiation in the package: Test 6(a) with detonator Result: detonation,crater2.2 Effect of propagation: Type 6(b) withdetonator2.4 Effect of fire engulfment: Test 6(c) may bewaived because of theresult of 6(b) test.3. Is the result a mass explosion? Yes4. Conclusion: Assignment toDivision 1.1Result: detonation ofthe whole stack ofpackages, crater2.2.6.2 Example of substances and mixtures not fulfilling theclassification criteriaThis example is taken from the UN Manual of Tests and Criteria, Part I, Section 10.5.2,Figure 10.5.A) RESULTS FROM APPLICATION OF THE ACCEPTANCE PROCEDURE96

Guidance on the Application of the CLP Criteria Version 3.0 - November 20120. General data:0.1 Name of the substance 5-tert-butyl-2,4,6-trinitro-m-xylene(musk xylene)1. Is the substance a candidate forammonium nitrate emulsion,suspension or gel, intermediate forblasting explosive ANE?2. Is the substance manufactured withthe view to producing a practicalexplosive or pyrotechnic effect?3. Test Series 1NoNo3.1 Propagation of Detonation: UN gap test (test 1(a)) Result:"+",propagationdetonation3.2 Effect of heating underconfinement:3.3 Effect of ignition underconfinement:4. Is it an explosive substance? Yes5. Test Series 2Koenen test (test 1(b))Time/pressure test (test1(c)(i))Result: Limitingdiameter 12.0 mmofResult: "—", noeffect on ignitionunder confinementFragmentationtype "F" "+",shows someexplosive effectson heating underconfinement5.1 Sensitivity to shock: UN gap test (test 2(a)) Result: "—", notsensitive to shock5.2 Effect of heating underconfinement:5.3 Effect of ignition underconfinement:6. Is the substance too insensitive foracceptance into this class?Koenen test (test 2(b))Time/pressure test (test2(c)(i))Conclusion: Substance to beconsidered for this class7. Test Series 37.1 Thermal stability: 75 °C/48 hour test (test3(c))NoResult:Limitingdiameter 12.0 mmResult: "—", noeffect on ignitionunder confinementResult: "—",thermally stable7.2 Impact sensitivity: BAM Fallhammer test Result: LimitingFragmentationtype "F" "+",violent effect onheating underconfinement.97

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(test 3(a)(ii))7.3 Friction sensitivity: BAM friction test (test3(b)(i))8. Is the substance thermally stable? Yes9. Is the substance too dangerous in theform in which it was tested?10. Conclusion: PROVISIONALLYACCEPT INTO THISCLASS10.1 Exit Apply the assignmentprocedureNoimpact energy 25 J",not too dangerous inform tested.Result: Limiting load> 360 N"—", not toodangerous inform testedB) RESULTS FROM APPLICATION OF THE ASSIGNMENT PROCEDURE1. Is the substance a candidate forDivision 1.5?2. Test Series 6NoResult: Package thesubstance2.1 Effect of initiation in the package: Test 6(a) with detonator Result: Onlylocaliseddecompositionaround detonator2.2 Effect of ignition in the package: Test 6(a) with igniter Result: Onlylocaliseddecompositionaround igniter2.3 Effect of propagation: Type 6(b) test notrequired as no effectoutside packagebetween packages in6(a) test2.4 Effect of fire engulfment: Test 6 Result: Only slowburning with blacksmoke occurred.3. Is the result a mass explosion? No4. Is the major hazard that fromdangerous projections?5. Is the major hazard radiant heatand/or violent burning but with nodangerous blast or projection hazard?6. Is there nevertheless a small hazardin the event of ignition or initiation?7. Is the substance manufactured withthe view to producing a practicalexplosive or pyrotechnic effect?NoNoNoNoNo significantreactionNo significantreactionNo effects whichwould hinder firefighting98

Guidance on the Application of the CLP Criteria Version 3.0 - November 20128. Conclusion: NOT ANEXPLOSIVE8.1 Exit Consider for anotherclass (e.g. flammablesolid)2.3 FLAMMABLE GASES2.3.1 IntroductionThe requirements in Chapter 2.2 “Flammable Gases” of Annex I of CLP are identical to thosein Chapter 2.2 of GHS.In addition, the DSD identifies R6 flammable gases that are unstable under certain conditions.2.3.2 Definitions and general considerations for the classification of flammablegasesAnnex I: 2.2.1.DefinitionsFlammable gas means a gas or gas mixture having a flammable range with air at 20°C and a standardpressure of 101.3 kPaThe flammability range of a flammable gas is defined between the “lower flammability limit”(LFL) in air and the “upper flammability limit” (UFL) in air. In technical literature, the terms“lower explosion limit” (LEL) and “upper explosion limit” (UEL) are often used instead ofthe LFL and UFL, respectively.2.3.3 Relation to other physical hazardsFor flammable gases that are packaged in aerosols dispensers, see Section 2.4 Flammableaerosols.2.3.4 Classification of substances and mixtures as flammable gases2.3.4.1 Identification of hazard informationMany gases are classified in Annex VI of CLP and more gases are classified in the RTDG.For gases that are not classified in Annex VI nor in the RTDG, there is ample scientificliterature giving the flammability range for most gases (e.g. IEC 79-20 “Data for flammablegases and vapours, relating to the use of electrical apparatus” – under revision or thedatabank Chemsafe at http://www.dechema.de/en/chemsafe.html).In the case a gas or gas mixture needs to be tested for flammability, a recognised internationalstandard shall be used such as EN 1839:2003, Determination of explosion limits of gases andvapours or ISO 10156: 1996 Gases and gas mixtures – Determination of fire potential andoxidising ability for the selection of cylinder valves outlets (under revision).2.3.4.2 Screening procedures and waiving of testing for gasmixturesThere are thousands of gas mixtures on the market and there are a limited number of testreports for the flammability of gas mixtures in the scientific literature. Tests to determine theflammability range are time consuming and expensive for gas mixtures that are made on99

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012demand. In most of the cases, the formulator of the gas mixture will use a calculation methodas described in ISO 10156 (see Section 2.3.4.4) to determine if the mixture is flammable ornot.2.3.4.3 Classification criteriaCategory12Annex I, 2.2.2. Table 2.2.1Criteria for flammable gasesCriteriaGases, which at 20°C and a standard pressure of 101.3 kPa:(a) are ignitable when in a mixture of 13% or less by volume in air; or(b) have a flammable range with air of at least 12 percentage points regardless of thelower flammable limit.Gases, other than those of Category 1, which, at 20°C and a standard pressure of101.3 kPa, have a flammable range while mixed in air.2.3.4.4 Testing and evaluation of hazard informationThe calculation method described in ISO 10156 uses the criterion that a gas mixture isconsidered non-flammable in air if:ni1A'Tic i 1Equation 2.3.4.4 (a)where:AiA' Equation 2.3.4.4 (b)ini1Aipk 1KkBkand where:100A'iis the equivalent content in mole% of the i:th flammable gas in the mixtureTciis the maximum content in mole% of the flammable gas i which, when mixed withnitrogen, is not flammable in airAiis the molar fraction in mole% of the i:th flammable gas in the mixtureBkis the molar fraction in mole% of the k:th inert gas in the mixtureKkis the coefficient of equivaleny of the inert gas k compared to nitrogenn is the total number of flammable gases in the mixturep is the total number of inert gases in the mixtureThe principle of the calculation method is the following: Where a gas mixture contains aninert diluent other than nitrogen, the volume of this diluent is adjusted to the equivalentvolume of nitrogen using the equivalency coefficient for the inert gas K . From this theequivalent contents A'iare then derived through Equation 2.3.4.4(b), which should be viewedas the corresponding concentration of the flammable gases if nitrogen was the only inert gask

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012present in the mixture. In Equation 2.3.4.4(a) the equivalent contents are then compared to theconstants Tci, which have been experimentally found using nitrogen as the (only) inert gas.It should be noted that ISO 10156 uses molar fractions in some of its equations. For mostgases under normal (i.e. non-extreme) conditions, however, the volume fraction can beassumed to be equal to the molar fraction, which is the same as assuming ideal gas behaviourfor all gases in the mixture. Furthermore, although normally a fraction is a number rangingfrom 0 to 1, in this case it is easier to express it as percentage, i.e. the fraction multiplied by100.The calculation method described in ISO 10156 determines only if the mixture is flammableor not. It does not determine a flammability range and therefore the calculation method cannotdetermine if the mixture is flammable Category 1 or Category 2. Therefore, to be on the safeside, mixtures determined to be flammable according the calculation method are classified“Flammable gas; Category 1”. If, however, there is a need to distinguish between Category 1and 2, the lower and the upper explosion limits have to be determined by using a suitable testmethod (e.g. EN 1839 or ISO 10156).For mixtures containing both flammable and oxidising components, special calculationmethods are described in ISO 10156.2.3.5 Hazard communication for flammable gases2.3.5.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 2.2.3. Table 2.2.2Label elements for flammable gasesClassification Category 1 Category 2GHS PictogramNo pictogramSignal word Danger WarningHazard statement H220: Extremely flammable gas H221: Flammable gasPrecautionary StatementPreventionPrecautionary StatementResponseP210P377P381P210P377P381Precautionary Statement Storage P403 P403Precautionary StatementDisposalThe wording of the Precautionary Statements is found in CLP Annex IV, Part 2.2.3.5.2 Additional labelling provisions101

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Some flammable gases are unstable under certain conditions. This is identified by theallocation of the phrase EUH006.Annex II, 1.1.2. EUH006 — ‘Explosive with or without contact with air’For substances and mixtures which are unstable at ambient temperatures, such as acetylene.So far, EUH006 has been allocated to two flammable gases (acetylene, ethylene oxide) inAnnex VI to CLP. The test method and criteria to allocate this hazard statement are underdevelopment at the UN Sub-Committee of experts on the GHS.2.3.6 Re-classification of substances and mixtures classified as flammable gasesaccording to DSD or already classified for transport2.3.6.1 Re-classification of substances and mixtures classified inaccordance with DSDBecause DSD has no sub-categories for flammable gases, there is no direct translationpossible between the classification for flammability of the gases according to DSD and thetwo categories for flammable gases according to CLP.Flammable gases that are listed in Annex I of DSD with “F+; R12” have all been reclassifiedin accordance with the criteria above and identified with either “Flam.Gas1; H220” or“Flam.Gas 2; H221” in Annex VI of CLP.Flammable gases that are not listed in Annex VI should be reclassified according to the newcriteria.2.3.6.2 Relation to transport classificationThe criteria for Category 1 correspond to the criteria that have been in use for classifying“Flammable Gases” in the RTDG. Consequently all gases listed as flammable in the RTDGshall be classified as “Flam.Gas 1; H220”.2.3.7 Example of classification for flammable gasesExample of a classification using the calculation method of ISO 10156Example mixture: 2 % (H 2 ) + 6 % (CH 4 ) + 27 % (Ar) + 65 % (He)Calculation steps:Step 1: Assign the gases and state their molar fractions, assuming the molar fractions areequal to the volume fractions (ideal gas behaviour for all gases).H 2 is flammable gas 1, yielding A1= 2 mole %CH 4 is flammable gas 2, yielding A2= 6 mole %Ar is inert gas 1, yielding B1= 27 mole %He is inert gas 2, yielding B2= 65 mole %n =2 since there are two flammable gases in the mixturep =2 since there are two inert gases in the mixtureStep 2: Look up the values of TciandKiin ISO 10156.102

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Tc1= 5.7 mole %Tc2=14.3 mole %K1= 0.5K2= 0.5Step 3: Calculate the equivalent gas contents A'ifor the flammable gases according toEquation 2.3.4.4(b).2A = 3.7 mole %' 12 60.527 0.5656A = 11.1 mole %' 22 60.527 0.565Step 4: Calculate the flammability of the gas mixture according to Equation 2.3.4.4(a).2i1A'TiciA'T1c1A'T2c23.7 11.1 = 1.435.7 14.3Step 5: Compare the outcome to the criterion in Equation 2.3.4.4(a).Since 1.43 > 1, this particular gas mixture is considered to be flammable.2.4 FLAMMABLE AEROSOLS2.4.1 IntroductionThe criteria for flammable aerosols are found in Annex I, Section 2.3 of CLP and in theAerosol Dispensers Directive 75/324/EEC.2.4.2 Definitions and general considerations for the classification of flammableaerosolsAnnex I: 2.3.1. Aerosols, this means aerosol dispensers, are any non-refillable receptacles made ofmetal, glass or plastics and containing a gas compressed, liquefied or dissolved under pressure, with orwithout a liquid, paste or powder, and fitted with a release device allowing the contents to be ejected assolid or liquid particles in suspension in a gas, as a foam, paste or powder or in a liquid state or in agaseous state.2.4.3 Classification of flammable aerosols2.4.3.1 Classification criteriaAnnex I, 2.3.2.1. Aerosols shall be considered for classification as flammable in accordance with 2.3.2.2if they contain any component which is classified as flammable according to the criteria contained in thispart i.e.:– Liquids with a flash point ≤ 93°C, which includes flammable liquids according to section 2.6 of thisAnnex– Flammable gases (see 2.2);– Flammable solids (see 2.7)103

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012NoteFlammable components do not cover pyrophoric, self-heating or water-reactive substances and mixturesbecause such components are never used as aerosol contents.2.3.2.2. A flammable aerosol shall be classified in one of the two categories for this Class on the basis ofits components, of its chemical heat of combustion and, if applicable, of the results of the foam test (forfoam aerosols) and of the ignition distance test and enclosed space test (for spray aerosols) in accordancewith Figure 2.3.1 and the UN Recommendations on the Transport of Dangerous Goods, the Manual ofTests and Criteria, part III, chapters 31.4, 31.5 and 31.6.Note: Flammable aerosols do not fall additionally within the scope of sections 2.2 (flammable gases),2.6 (flammable liquids) or 2.7 (flammable solids) of Annex I of CLP. Depending on their contents,aerosol dispensers may additionally fall within the scope of other hazard classes (e.g. health andenvironmental hazard classes).The following definitions can be found in the Aerosol Dispensers Directive 75/324/EEC:Non-flammable aerosol: The aerosol is not classified in the hazard class for flammableaerosols if it contains 1% or less flammable components and the chemical heat of combustionis less than 20 kJ/g.Extremely flammable aerosol: The aerosol is classified as extremely flammable aerosol(Category 1) in the hazard class for flammable aerosols if it contains 85% or more flammablecomponents and the chemical heat of combustion exceeds or is equal to 30 kJ/g.Other aerosols: All other aerosols will be submitted to appropriate flammability classificationprocedures in order to select the appropriate Category 1 or 2 or to decide not to classify theaerosol.If the aerosols are not submitted to the flammability classification procedures, then they shallbe automatically classified as ‘extremely flammable’, as specified in Directive 75/324/EEC.Under Directive 75/324/EEC, flammability classification for aerosols refers to ‘extremelyflammable’ and ‘flammable’. This corresponds to the terms ‘Category 1’ and ‘Category 2’which are used in CLP.The chemical heat of combustion will be determined in accordance with CLP Annex I, 2.3.4.1or with point 1.10 of the Annex to Directive 75/324/EEC.2.4.3.2 Testing and evaluation of hazard informationResults from the ignition distance test, the enclosed space test and the foam flammability testmay be used for the classification for flammable aerosols. These test methods also describedunder point 6.3 of the Annex to Directive 75/324/EEC and are therefore available in all EUlanguages.If the evaluation according to the appropriate criteria (see previous sections) shows that theclassification criteria are fulfilled, the aerosol will be classified in one of the two categories.2.4.3.3 Decision logicNOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.104

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: Figure 2.3.1Figure 2.3.1 (a) for flammable aerosolsAEROSOLDoes it contain ≤ 1% flammable components anddoes it have a heat of combustion < 20 kJ/g?YESNOT CLASSIFIEDNOCategory 1Does it contain ≥ 85% flammable components anddoes it have a heat of combustion ≥ 30 kJ/g?NOYESDangerFor spray aerosols, go to decision logic 2.3.1 (b);For foam aerosols, got to decision logic 2.3.1 (c).105

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 2.3.1 (b) for spray aerosolsSPRAY AEROSOLCategory 1In the ignition distance test, does ignition occur at adistance ≥ 75 cm?NODoes it have a heat of combustion < 20 kJ/g?YESYESNODangerCategory 2WarningIn the ignition distance test, does ignition occur at adistance ≥ 15 cm?YESCategory 2NOIn the enclosed space ignition test; is: (a) the timeequivalent ≤ 300 s/m³or (b) the deflagration density≤ 300 g/m³?NONOT CLASSIFIEDYESWarningCategory 2Warning106

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 2.3.1 (c) for foam aerosolsFOAM AEROSOLCategory 1In the foam test, is: (a) the flame height ≥ 20 cm and theflame duration ≥ 2 s; or (b) the flame height ≥ 4 cm and theflame duration ≥ 7 s?YESDangerNOCategory 2In the foam test; is the flame height ≥ 4 cm and the flameduration ≥ 2 s?NOYESWarningNOT CLASSIFIED2.4.4 Hazard communication for flammable aerosols2.4.4.1 Pictograms, signal words, hazard statements andprecautionary statementsCLASSIFICATIONAnnex I: 2.3.3. Table 2.3.2Label elements for flammable aerosolsCategory 1 Category 2GHS PictogramsSignal word Danger WarningHazard statementPrecautionary StatementPreventionH222: Extremely flammableaerosolP210P211P251H223: Flammable aerosolP210P211P251Precautionary Statement107

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012ResponsePrecautionary StatementStoragePrecautionary StatementDisposalP410 + P412P410 + P412The wording of the Precautionary Statements is found in CLP Annex IV, Part 2.2.4.4.2 Additional labelling provisionsDirective 75/324/EEC imposes additional labelling requirements on all aerosols, flammableor not, including those which are not within the scope of CLP.For example:- Where an aerosol dispenser contains flammable components but is not classified asflammable, the quantity of flammable material contained in the aerosol dispenser must bestated clearly on the label, in the form of the following wording: “X% by mass of thecontents are flammable”.2.4.5 Re-classification of flammable aerosols according to DSDIn DSD no hazard class 'flammable aerosols' is defined. In CLP flammable aerosols is a newdistinct hazard class. DSD made, in its Annex VI, point 1.7, reference to the flammabilitycriteria of the Aerosol Dispensers Directive 75/324/EEC. No previous classification is usefulto complete classification under CLP.Until 2008, the Aerosol Dispensers Directive 75/324/EEC followed a very conservativeapproach as in principle all aerosols with flammable contents (according to the criteria laiddown for the categories “extremely flammable”, “highly flammable” and “flammable” andlisted in Annex VI to DSD) had to be considered as flammable in the strictest categoryconcerned, regardless of the amount of flammable content.Only where the person responsible for the marketing of aerosol dispensers was in possessionof test results or other data showing that although those aerosol dispensers had flammablecontents they did not present any risk of ignition under normal or reasonably foreseeableconditions of use, he could on his own responsibility decide not to apply the labellingprovisions for flammable aerosols. Only very few aerosol products could benefit from thatexemption. Flammability testing of aerosols was the exception, as it was easily foreseeablethat the vast majority of aerosol products could not fulfil these strict conditions anyway. Dueto the widespread use of propellants which are classified as ‘extremely flammable’, the vastmajority of aerosols were labelled as ‘extremely flammable’.At UN level this conservative philosophy of the former Aerosol Dispensers Directive wasacknowledged and a provision was introduced stating that aerosols not submitted to theflammability classification procedures shall be classified (as ‘extremely flammable’) inCategory 1.Following the CLP criteria for flammable aerosols, the vast majority of aerosols will thereforecontinue to be classified as ‘extremely flammable’ in Category 1, without the need to performsuperfluous testing.2.4.6 Examples of classification for flammable aerosols108

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012For reasons of simplification the active materials chosen in the examples have beenconsidered as non combustible materials (H c = 0 kJ/g). However this is not the case inpractice.2.4.6.1 Examples of aerosols fulfilling the classification criteriaDeodorant:Composition:Butane/propane: 70% (flammable components, H c = 43.5 kJ/g)Ethanol:25% (flammable components, H c = 24.7 kJ/g)Others:5% (non-flammable components, H c = 0 kJ/g)This spray aerosol contains 95% of flammable components, and its chemical heat ofcombustion equals 36.6 kJ/g (= 0.70 * 43.5 + 0.25 * 24.7).This aerosol is classified in Category 1.Air freshener (wet):Composition:Butane/propane: 30% (flammable components, H c = 43.5 kJ/g)Others:70% (non-flammable components, H c = 0 kJ/g)This spray aerosol contains 30% of flammable components and its chemical heat ofcombustion equals 13.1 kJ/g.In the ignition distance test, the ignition occurs at less than 75 cm but more than 15 cm.This aerosol is classified in Category 2.2.4.6.2 Examples of aerosols not fulfilling the classificationcriteriaShaving foam:Composition:Butane/propane: 4% (flammable components, H c = 43.5 kJ/g)Others:96% (non-flammable components, H c = 0 kJ/g)This foam aerosol contains 4% of flammable components and its chemical heat of combustionequals 1.7 kJ/g.In the foam test, the flame height is less than 4 cm and the flame duration less than 2 s.This aerosol is not classified as flammable aerosol.However the quantity of flammable components must be stated clearly on the label: “4% bymass of the contents are flammable”.2.5 OXIDISING GASES2.5.1 Introduction109

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The requirements in Chapter 2.4 “Oxidising Gases” of Annex I of CLP are identical to thosein chapter 2.4 of the GHS.2.5.2 Definitions and general considerations for the classification of oxidising gasesAnnex I: 2.4.1. Oxidising gas means any gas or gas mixture which may, generally by providing oxygen,cause or contribute to the combustion of other material more than air does.2.5.3 Classification of substances and mixtures as oxidising gases2.5.3.1 Identification of hazard informationThere are not many gases that are oxidising. Most oxidising gases are identified as such in theRTDG and in ISO 10156-2: 2005 Gas cylinders - Gases and gas mixtures: - Part 2:Determination of oxidizing ability of toxic and corrosive gases and gas mixtures.2.5.3.2 Screening procedures and waiving of testingThere are thousands of gas mixtures containing oxidising gases on the market and there arevery few test reports on oxidising potential of gas mixtures in the scientific literature. Testsaccording to ISO 10156-2 in order to determine the oxidising potential are time consumingand expensive for gas mixtures that are made on demand. In most of the cases, the formulatorof the gas mixture will use a calculation method as described in ISO 10156: 1996 Gases andgas mixtures – Determination of fire potential and oxidising ability for the selection ofcylinder valves outlets (under revision) or ISO 10156-2 to determine if the mixture isoxidising or not.2.5.3.3 Classification criteriaCategory1Annex I: 2.4.2. Table 2.4.1Criteria for oxidising gasesCriteriaAny gas which may, generally by providing oxygen, cause or contribute to the combustionof other material more than air does.The criteria “more than air does” is further defined in the Note as “having an oxidising powergreater than 23.5% as determined by a method specified in the last revision of ISO 10156 andISO 10156-2”.2.5.3.4 Testing and evaluation of hazard informationThe classification method described in ISO 10156:1996 and ISO 10156-2:2005 uses thecriterion that a gas mixture should be considered as more oxidising than air if the “Oxidising¨Power (OP)” of the gas mixture is higher than 0.235 (23.5%).The OP is calculated as follows:Where:110x iOP nxni1x Cpii1 k1iis the molar fraction in mole% of the i:th oxidising gas in the mixtureiKkBk

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012C iK kB knpis the coefficient of oxygen equivalency of the i:th oxidising gas in themixtureis the coefficient of equivalency of the inert gas k compared to nitrogenis the molar fraction in mole % of the k:th inert gas in the mixtureis the total number of oxidising gases in the mixtureis the total number of inert gases in the mixtureFor mixtures containing both flammable and oxidising components, special calculationmethods are described in ISO 10156.2.5.4 Hazard communication for oxidising gases2.5.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 2.4.3. Table 2.4.2Label elements for oxidising gasesClassification Category 1GHS PictogramSignal wordHazard statementPrecautionary Statement PreventionPrecautionary Statement ResponsePrecautionary Statement StoragePrecautionary Statement DisposalDangerH270: May cause or intensify fire; oxidiserP220P244P370 + P376P403The wording of the Precautionary Statements is found in CLP Annex IV, Part 2.2.5.5 Re-classification of substances and mixtures classified as oxidising gasesaccording to DSD or already classified for transport2.5.5.1 Re-classification of substances and mixtures classified inaccordance with DSDOxidising gases that were listed in Annex I of DSD with O; R8 have all been identified with“Ox. Gas 1; H270" in the Annex VI of CLP. Steps have been taken to align the classificationof oxidising gases both in the transport regulations (e.g. for chlorine) and in CLP (e.g.nitrogen dioxide, chlorine).2.5.5.2 Relation to transport classification111

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Most oxidising gases are classified as such with subsidiary risk 5.1 in the RTDG.Consequently all gases listed as oxidising in the RTDG shall be classified as “Ox. Gas 1”.2.5.6 Examples of classification for oxidising gasesExample of a classification using the calculation method of ISO 10156Example Mixture: 9 % (O 2 ) + 16 % (N 2 O) + 75 % (N 2 )Calculation stepsStep 1: Ascertain the coefficient of oxygen equivalency (Ci) for the oxidising gases in themixture and the nitrogen equivalency factors (K k ) for the non-flammable, non-oxidisinggases.C i (N 2 O)= 0.6 (nitrous oxide)C i (O)= 1 (oxygen)K k (N 2 )= 1 (nitrogen)Step 2: Calculate if the Oxidising Power (OP) of the gas mixtureOP nxni1x Cpii1 k 1iiKkBk0,09 10,160,6 0,1860,09 0,16 0,75118.6 < 23.5, therefore the mixture is not considered as an oxidising gas.Important note:The example is only given to illustrate the principles of the calculation method described inISO 10156:1996 and ISO 10156-2:2005. For the actual classification of gas mixtures, themost recent version of the ISO standards shall be used where all C i values for oxidising gasescan be found.2.6 GASES UNDER PRESSURE2.6.1 IntroductionThe requirements in Chapter 2.5 “Gases under pressure” of Annex I of CLP are identical tothose in chapter 2.5 of GHS. The hazard Class “Gases under pressure” corresponds to thedanger class 2 “Gases” in the RTDG.2.6.2 Definitions and general considerations for the classification of gases underpressure1122.6.2.1 Definition of “gas”Annex I: 1.0. Gas means a substance which (i) at 50 °C has a vapour pressure greater than 300 kPa(absolute); or (ii) is completely gaseous at 20 °C at a standard pressure of 101.3 kPa;This definition means that pure substances are considered as gases when their boiling point(BP) is not higher than 20°C. Substances with a boiling point higher than 20°C are “liquids”except those few that develop a vapour pressure higher than 300 kPa at 50°C; these liquids areconsidered as “gases” because of the hazard of pressure when packaged.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Hydrogen fluoride (HF) with a BP of 19.4°C is a borderline line case that has always beenclassified as a liquid.2.6.2.2 Definition of “gases under pressure”Annex I: 2.5.1.1. Gases under pressure are gases or gas mixtures which are contained in a receptacleat a pressure of 200 kPa (gauge) or more, or which are liquefied or liquefied and refrigerated.They comprise compressed gases, liquefied gases, dissolved gases and refrigerated liquefied gases.This definition means in practice that compressed gases or dissolved gases that are packagedat a pressure less than 200 kPa are not classified for this hazard.Dissolved gases packaged at a pressure less than 200 kPa (gauge) are liquids and should beclassified as such if they have other hazardous properties, e.g. flammable liquids.Also, liquids packaged under a layer of inert gas (e.g. nitrogen or helium) remain to beclassified as liquids and not as “gases under pressure”.2.6.3 Relation to other physical hazardsGases under pressure need also to be classified for the hazard classes 'flammable gases' and'oxidising gases' where relevant.2.6.4 Classification of substances and mixtures as gases under pressure2.6.4.1 Identification of hazard informationMany gases are identified as such in the RTDG and many flammable gases and someoxidising gases are identified as gases in Annex VI of CLP. The RTDG identify further if thegas can be packaged as a “compressed gas”, “liquefied gas”, “refrigerated liquefied gas” and“dissolved gas”. When the gas is not listed in the RTDG and in case of doubt, the followingphysical characteristics are necessary to classify a pure substance as a gas: The boiling point The vapour pressure at 50°C ;For those pure substances that meet the definition of a gas (see Section 2.6.2), the criticaltemperature is also necessary.The following references generally provide good quality data on boiling points, vapourpressure and the critical temperature of pure substances (see Section 2.7.8 for full references):(a) CRC Handbook of Chemistry and Physics (CRC, 2005)(b) The Merck Index (Merck, 2001)(c) ChemFinder (ChemFinder, database)(d) CHEMSAFE (contains evaluated/recommended data) (CHEMSAFE, database)(e) Safety Characteristic Data (contains evaluated/recommended data) (Brandes, 2008)2.6.4.2 Classification criteriaGroupAnnex I: 2.5.2. Table 2.5.2Criteria for gases under pressureCriteria113

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Compressed gasLiquefied gasRefrigeratedliquefied gasDissolved gasA gas which when packaged under pressure is entirely gaseous at-50°C; including all gases with a critical temperature -50°C.A gas which when packaged under pressure, is partially liquid at temperaturesabove -50°C. A distinction is made between:i) High pressure liquefied gas: a gas with a critical temperature between -50°Cand +65°C; andii)Low pressure liquefied gas: a gas with a critical temperature above +65°C.A gas which when packaged is made partially liquid because of its lowtemperature.A gas which when packaged under pressure is dissolved in a liquid phase solvent.2.6.4.3 Testing and evaluation of hazard informationThe critical temperature of pure gases is well defined and can be found in technical literature,e.g. EN 13096 “Transportable gas cylinders — Conditions for filling gases into receptacles— Single component gases”.For gas mixtures, the classification is based on the “pseudo-critical temperature” which can beestimated as the mole weighted average of the components’ critical temperatures.Pseudo Critical Temperature = xiCwhere x i is the component in molar fraction and C Tk is the Critical Temperature of thecomponent in Kelvin.2.6.5 Hazard communication for gases under pressure2.6.5.1 Pictograms, signal words, hazard statements andprecautionary statementsniTkClassificationAnnex I: 2.5.3. Table 2.5.2Label elements for gases under pressureCompressed gasGHS PictogramsSignal wordHazard statementPrecautionary StatementPreventionPrecautionary StatementResponsePrecautionary StatementStorageWarningH280: Contains gas under pressure; may explode if heatedP410 + P403114Precautionary Statement

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012DisposalThe wording of the Precautionary Statements is found in CLP Annex IV, Part 2.Packages of gases labelled for transport do not need to bear the relevant GHS Pictograms forclassification for “gases under pressure” (Article 33 (3)).2.6.6 Re-classification of substances and mixtures classified as gases underpressure according to DSD or already classified for transport2.6.6.1 Re-classification of substances and mixtures classified inaccordance with DSDThe hazard class “gases under pressure” is a new hazard class that was not considered inDSD.Gases that are classified in Annex VI of CLP have been identified with the indication“Press.Gas” in the Classification column but without the indication of the group and thecorresponding hazard statement (H280 or H281). The group depends on the physical state inwhich the gas is packaged and therefore has to be assigned case-by-case (see note U in Part 1of Annex VI).2.6.6.2 Relation to transport classificationMore gases are classified in the RTDG (ADR/RID/ADN) with an indication of the physicalstate in the Classification Code that can be used to identify the group of “gases underpressure” according to CLP: 1 = compressed gas (e.g. Argon, compressed: Classification code: 1A) 2 = liquefied gas (e.g. Butane: Classification code: 2F) 3 = refrigerated liquefied gas (e.g Oxygen, refrigerated liquid: 3O) 4 = dissolved gas (e.g. Acetylene, dissolved: 4F)2.6.7 Examples of classification for gases under pressureExample mixture: 9%(O 2 ) + 16%(N 2 O) + 75%(N 2 )Calculation steps:Step 1: Ascertain the critical temperatures in Kelvin for the gases in the mixture:Oxygen (O 2 ): Temp.Crit.= -118.4°C= 154.75 KNitrous Oxide (N 2 O): Temp.Crit.= +36.4°C= 309.55 KNitrogen (N 2 ): Temp.Crit.= -147°C= 126.15 KStep 2: Calculate the pseudo-critical temperature:0.09 154.75 K + 0.16 309.55 K + 0.75 126.15 K= 158.7 Kelvin = - 115.08 °CThe pseudo-critical temperature is lower than -50°C, therefore the mixture is a “compressedgas”115

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.7 FLAMMABLE LIQUIDS2.7.1 IntroductionFlammable liquids with a flashpoint not more than 60°C are classified in accordance withCLP into one of three categories according to their boiling point and flashpoint.The threshold limits for the categories differ from the respective threshold limits of DSD forflammable liquids (see Section 2.7.6.1).They are however identical to the threshold limits of packing group 1, 2 and 3 whenclassifying “flammable liquids” according to the RTDG.Substances or mixtures which do not show a flashpoint but do have an explosion range ormay become flammable in use have to be marked with EUH018.2.7.2 Definitions and general considerations for the classification of flammableliquidsAnnex I: 2.6.1. Flammable liquid means a liquid having a flashpoint of not more than 60°CThe flashpoint is the lowest temperature of the liquid, corrected to a barometric pressure of101.3 kPa, at which application of a test flame causes the vapour of the liquid to ignitemomentarily and a flame to propagate across the surface of the liquid under the specifiedconditions of test. This means, the lower explosion limit is exceeded at the flashpoint.2.7.3 Relation to other physical hazardsFor flammable liquids that are packaged in aerosols dispensers, see Section 2.4.3 Flammableaerosols.2.7.4 Classification of substances and mixtures as flammable liquids2.7.4.1 Identification of hazard informationFor the decision if a substance or mixture is a liquid see Section 2.1.4.For the classification of a substance or mixture as a flammable liquid, data on the flash pointand on the boiling point (or the initial boiling point) are needed. For experimentaldetermination of the flash point information on the viscosity of the liquid is needed, in orderto select a suitable method. Furthermore, in order to make use of the derogation forclassification in Category 3 according to Annex I Section 2.6.4.5 of CLP (see Section2.7.4.1.3), information on sustained combustibility is necessary.Experimentally determined data or data taken from reliable data sources are to be preferredover calculated ones. See also IR/CSA, Section R7.1.3 (boiling point), R7.1.9 (flashpoint).The following references generally provide good quality data on boiling points (a,b,c,d,e) andflashpoint (c,d,e) of pure substances may be found in:(a) CRC Handbook of Chemistry and Physics (CRC, 2005)(b) The Merck Index (Merck, 2001)(c) ChemFinder (ChemFinder, database)116

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(d) CHEMSAFE (contains evaluated/recommended data) (CHEMSAFE, database)(e) Safety Characteristic Data (contains evaluated/recommended data) (Brandes, 2008)Special care is required when viscous substances or mixtures are tested or when halogenatedcompounds are present (see Section 2.7.4.4.1).2.7.4.2 Screening procedures and waiving of testing2.7.4.2.1 Boiling pointNormally calculation methods based on increments give satisfying results for pure substancesand mixtures. With respect to the interesting figure for flammable liquids (35°C) only thatmethod with a mean absolute error lower than 5 °C could be recommended for screening.2.7.4.2.2 Flash pointCalculation should work for pure liquids, neglecting impurities, if the vapour pressure curveand lower explosion limit are accurately known. For mixtures, calculation of the flashpoint issometimes not reliable and at this time, it is not possible to predict what reliance can beplaced on a calculated value. Calculation can be used as a screening test for mixtures, and aflashpoint need not be determined experimentally if the calculated value using the methodcited in CLP Annex I, 2.6.4.3 is 5 °C greater than the relevant classification criterion.However, the restrictions outlined in the CLP Annex I, 2.6.4.2 should be taken account of.Calculation based on structural similarity or properties is often only applicable to a narrowlydefined set of substances. For mixtures they are not yet applicable.Therefore for both flashpoint and boiling point experimental determination is recommended.2.7.4.3 Classification criteriaA flammable liquid has to be classified in one of the 3 categories of this class.Annex I: 2.6.2. Table 2.6.1CategoryLabel elements for flammable liquidsCriteria1 Flash point < 23°C and initial boiling point ≤ 35°C2 Flash point < 23°C and initial boiling point > 35°C3Flash point ≥ 23°C and ≤ 60°C 11 For the purpose of this Regulation gas oils, diesel and light heating oils having a flash point between> 55°C and ≤ 75°C may be regarded as Category 3Furthermore,Annex I: 2.6.4.5. Liquids with a flash point of more than 35 °C may be regarded as non-flammableliquids if negative results have been obtained in the sustained combustibility test L.2 of the UNRecommendations on the Transport of Dangerous Goods, Manual of Tests and Criteria.The sustained combustibility test L.2 can be found in the UN-MTC, Part III, section 32.5.2.Gas oils, diesel and light heating oils in the flashpoint range of 55-75°C may be regarded as awhole as diesel and these hydrocarbon mixtures have varying flashpoints in that range due toseasonal requirements (EN 590). If they are regarded as a whole for CLP they have to beregarded as Category 3. This states however no preliminary decision with respect todownstream Regulations and legislation.117

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.7.4.4 Testing and evaluation of hazard informationThe assignment to the respective hazard category will determine the technical means to betaken to avoid dangerous events. In combination with other endpoints like explosion limits orauto ignition temperature this can lead to clear restrictions in the conditions of use. Therelevant data are to be communicated via the CSR and SDS (see IR/CSA Parts F and G,respectively).2.7.4.4.1 TestingSuitable methods are listed in CLP Annex I, Table 2.6.3.In case of substances with a high decomposition potential, a method using small amounts ofliquid (e.g. EN ISO 3679: 2004 Determination of flash point - Rapid equilibrium closed cupmethod) is recommended to reduce the amount of substance under test.The method to be used has to be chosen taking into account the properties of the liquid(viscosity, halogenated compounds present) and the scope of the standard.For classification purposes it is recommended to use the mean of at least two test runs. One ofthese runs may be automated. In case of a deviation between manual and automateddetermination beyond the tolerance limits of the method, the lower value should be taken or atleast the result the determination should be repeated with manual observation. If theexperimentally determined flashpoint is found to be within ± 2 °C a threshold limit whenusing a non-equilibrium method, it is recommended to repeat the determination with anequilibrium method.If in doubt, or if no flashpoint is found up to 60 °C and the conditions laid down in EUH018,EUH209 and EUH209A are met, (presence of (partly) halogenated compounds, possibility toa loss of volatile flammable or non-flammable components) determination of explosion limitsaccording to EN 1839:2003, Determination of explosion limits of gases and vapours or ISO10156: 1996 Gases and gas mixtures – Determination of fire potential and oxidising abilityfor the selection of cylinder valves outlets (under revision) or determination of explosionpoints according to DIN EN 15794: 2008, Determination of explosion points of flammableliquids, is recommended to decide on labelling with EUH018, EUH209 or EUH209A.SubstancesFor pure non-halogenated substances, the flashpoint is usually found 80 °C to 130 °C belowthe boiling point. Special care has to be taken when a sample contains impurities with a lowerboiling point than the main compound. Even if their concentration is below 0.5%, especiallyif their boiling point is substantially lower, they may have a strong effect on the test result.Impurities with a higher boiling point will normally have no effect on the flashpoint.Within the respective scope, every standard is applicable.MixturesThe flashpoint may be lower than the lowest flashpoint of the components and non-volatilecomponents may influence the flashpoint.Equilibrium methods are advised if the boiling points of the components of the mixture covera wide range of temperatures or their concentrations are very different. They are also advisedin case of viscous mixtures (alternatively: test methods with low heating rates (1 °C per min)using a stirrer).In case of viscous mixtures or if an inerting substance is present at low concentrations and thisis a highly volatile compound, the ignitability of the mixture may depend on the temperature118

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012at which the tests are started. When an inerting substance is present temperature ranges mayexist where the vapour phase is inerted and other temperature ranges where it is not.Halogenated compoundsThe difference between boiling point and flashpoint may be lower than with non-halogenatedcompounds.It is highly recommended to run the tests under careful control with manual observation.Test results may be very difficult to reproduce. In such cases, classification should be basedon the lowest value found (flash or burning inside or outside the cup) or on the value obtainedduring the screening run if in the main trial performed in accordance with the standard, noflash could be found.2.7.4.4.2 Evaluation of hazard informationExperimentally derived boiling points are to be preferred over calculated ones because of theerror of most of the QSAR methods.Flashpoints determined by testing or from the mentioned internationally recognised qualifiedliterature are to be preferred over those derived by calculation because of the error of most ofthe QSAR methods respectively their limited application range.If in literature different flashpoints are found for the same substance the one found asevaluated/recommended has to be preferred.If in literature different flashpoints are found for the same substance where none is found asevaluated/recommended the lower one has to be preferred because of safety reasons or anexperimental determination should be carried out.According to the criteria either Category 1, 2 or 3, including the relevant hazard statement andsignal word, have to be assigned (see Section 2.7.5). In case the criteria for EUH018,EUH209 or EUH209A are met, the liquid has to be labelled with either one of thesesupplemental hazard statements as well. In the majority of cases EUH018 covers EUH209and EUH209A.2.7.4.5 Decision logicNOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.119

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012This decision logic is amended to include EUH phrases 018, 209 and 209A.The substance or mixture is a liquidFlash point ≤ 60°CyesnoGas oil, diesel, light heating oil withflash point up to 75°CYesnoHalogenated substance,mixture containinghalogenated, volatile ornon volatile flammablesubstancesnoMay be regarded asCategory 3yesNot subject of hazardclass ‘flammable liquid’Flash point < 23°CnoWarningExplosive vapour/airmixture possible(EN 1839, EN 15794)noyesFlash point > 35°CyesnoYesEUH209,EUH209A,EUH018Sustained combustibilityyesnonoNo need to be classifiedas ‘flammable liquid’Category 2Boiling point ≤ 35°CnoDangeryesCategory 1yesDanger120

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.7.5 Hazard communication for flammable liquids2.7.5.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 2.6.3. Table 2.6.2Label elements for flammable liquidsClassification Category 1 Category 2 Category 3GHS PictogramsSignal word Danger Danger WarningHazard statementH224: Extremelyflammable liquid andvapourH225: Highlyflammable liquid andvapourH226: Flammable liquidand vapourPrecautionaryStatementPreventionP210P233P240P241P242P243P280P210P233P240P241P242P243P280P210P233P240P241P242P243P280PrecautionaryStatement ResponseP303 + P361 + P353P370 + P378P303 + P361 + P353P370 + P378P303 + P361 + P353P370 + P378PrecautionaryStatement DisposalP501 P501 P501The wording of the Precautionary Statements is found in CLP Annex IV, Part 2.2.7.5.2 Additional labelling provisions for flammable liquidsAnnex II: 1.1.4. EUH018 - 'In use, may form flammable/explosive vapour-air mixture'For substances and mixtures not classified as flammable themselves, which may formflammable/explosive vapour-air mixtures. For substances this might be the case for halogenatedhydrocarbons and for mixtures this might be the case due to a volatile flammable component or due tothe loss of a volatile non-flammable component.Annex II: 2.9. Liquid mixtures containing halogenated hydrocarbonsFor liquid mixtures which show no flashpoint or a flashpoint higher than 60 ˚C but not more than 93˚C and contain a halogenated hydrocarbon and more than 5 % highly flammable or flammablesubstances, the label on the packaging shall bear one of the following statements, depending onwhether the substances referred to above are highly flammable or flammable:EUH209 — ‘Can become highly flammable in use’ orEUH209A — ‘Can become flammable in use’121

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Note: EUH209 and EUH209A are limited to special types of mixtures whereas EUH018covers a wider range of mixtures. In the majority of cases EUH018 covers EUH209 andEUH209A.2.7.6 Re-classification of substances classified as flammable liquids according toDSD or already classified for transport2.7.6.1 Re-classification according to DSDDirect translation is only partly possible, see Figure 2.7.6.1. For substances and mixtureswhich are R11 or R10 according to DSD the flashpoint and boiling point as well as thesustained combustibility (R10) data have to be re-evaluated. Re-determination may benecessary if only the flashpoint range is available.Figure 2.7.6.1: Comparison of the DSD and the CLP classificationFlashpoint605523210DSD classificationCLP classificationR10 Cat. 3R11Cat. 1 Cat. 2R12≤35 > 35 ≤35 > 35Boiling point / initial boiling point in °C2.7.6.2 Relation to transport classificationIn transport class 3 corresponds to hazard class ‘flammable liquid’. Except UN 1203 directtranslation of the packing groups into categories is possible if class 3 is the main risk. If class3 is a subsidiary risk no general one-to-one translation is possible.2.7.7 Examples of classification for flammable liquids2.7.7.1 Examples of substances and mixtures fulfilling theclassification criteriaExample 1Mixture of: Butylacetate + 1-Methoxy-2-propylacetate + Xylene + Methylisobutylketone(24 mol% + 5 mol% + 69 mol% + 2 mol%)Initial boiling point (calculated): 130 °CFlash point (calculated): 22 °Ccalculated flashpoint is within 5 °C to the limiting value of 23°C flash point has to be measured.Dyn. Viscosity at 20 °C (DIN 53019):8 mPas122

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Flash point (EN ISO 3679): 25.0 °C According to boiling point and measured flashpoint result: Category 3Example 2Mixture of: Hydrocarbons and dichloromethane(70 vol % + 30 vol%)Initial Boiling point (calculated): 52 °CFlash point:no flashpoint according to a standard Because the hydrocarbon part of the mixture has a flashpoint by itself (-12 °C) the question "is anexplosive vapour/air mixture possible (EN 1839, DIN EN 15794) or can it become highly flammable /flammable during use?" has to be answered.Answer: Yes an explosion range exists, yes it can become highly flammable during use. According to the answer, the mixture has to be labelled with EUH018 or EUH209Note: In that case EUH018 covers EUH2092.7.7.2 Examples of substances and mixtures not fulfilling theclassification criteriaExample 3Aqueous formulation of aliphatic polyurethane resinBoiling point (EC 440/2008 A.2): 92 °CDyn. Viscosity at 20 °C (DIN 53019):1938 mPasSample is highly viscous, use low heating rate for flashpoint determination (1 °C /min).Flash point (EN ISO 13736): 42.5 °CSustained combustibility test (UN L.2)at 60.5 °C:Sustained combustibility test (UN L.2)at 75 °C:combustion not sustainedcombustion not sustained According to the flashpoint result: Category 3May however not be regarded as Category 3 because it did not sustain combustion.2.7.8 ReferencesBrandes, E. and Möller, W.: Safety Characteristic Data, Volume 1, Flammable gases andliquids, nw-Verlag, 2008ChemFinder (database): http://chemfinder.cambridgesoft.comHEMSAFE (database): http://www.dechema.de/en/chemsafe.htmlCRC (2005) CRC Handbook of Chemistry and Physics 86 th Edition. Editor in Chief, D. Lide.CRC Press, Taylor and Francis, Boca Raton, FLMerck (2001) Merck Index 13th Edition. Edited by S Budavari et al. Merck & Co, Inc, USA123

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.8 FLAMMABLE SOLIDS2.8.1 IntroductionSolid substances and mixtures are classified as flammable according to their burningbehaviour.2.8.2 Definitions and general considerations for the classification of flammablesolidsAnnex I: 2.7.1.1. A flammable solid means a solid which is readily combustible, or may cause orcontribute to fire through friction.Readily combustible solids are powdered, granular, or pasty substances or mixtures which aredangerous if they can be easily ignited by brief contact with an ignition source, such as a burningmatch, and if the flame spreads rapidly.2.8.3 Relation to other physical hazardsExplosives, organic peroxides, self-reactive substances and mixtures as well as pyrophoric oroxidising solids should not be considered for classification as flammable solids sinceflammability is an intrinsic hazard in these classes.However, flammable solids can present other physical hazards at the same time, i.e. theymight be self-heating or corrosive or emit flammable gases in contact with water.For flammable solids that are packaged in aerosols dispensers, see Section 2.4, Flammableaerosols.2.8.4 Classification of substances and mixtures as flammable solids1242.8.4.1 Identification of hazard informationIn many cases, a simple screening test (see Section 2.8.4.4) can be used to determine whethera solid should be classified as flammable.For the classification of a substance or mixture as a flammable solid data on the followingproperties are needed: Melting point Information on water reactivity Information on flash point if solids containing flammable liquidsMany organic solid substances or mixtures fulfil the criteria to be classified as flammablesolids. For inorganic solids, the classification as flammable is rather rare.2.8.4.2 Screening procedures and waiving of testingIn general, a possible classification as a flammable solid should be considered for any solidorganic substance or mixture containing such material. For inorganic material, testing may bewaived in cases where the substance is commonly known to be not flammable (i.e. stable saltsor metal oxides) or where a flammability hazard can be excluded by any other scientificreasoning.The test method as described in sub-section 33.2.1.4.3.1 in the UN-MTC should be appliedfor screening purposes. Alternatively, for determination of explosion characteristics, the

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012burning index as obtained from the Grewer Oven test (VDI guideline 2263, part 1, 1990, Testmethods for the Determination of the Safety Characteristics of Dusts) may be used. If aburning index of 3 or less is found, the substance should not be classified as a flammable solidand no further testing is required. However, if smouldering or a flame is observed, the full testmust be carried out.2.8.4.3 Classification criteriaThe classification criteria are fully in accordance with the GHS system.Annex I: 2.7.2.1. Powdered, granular or pasty substances or mixtures (except powders of metals ormetal alloys – see 2.7.2.2) shall be classified as readily combustible solids when the time of burningof one or more of the test runs, performed in accordance with the test method described in Part III,sub-section 33.2.1, of the UN Recommendations on the Transport of Dangerous Goods, Manual ofTests and Criteria, is less than 45 seconds or the rate of burning is more than 2,2 mm/s.2.7.2.2. Powders of metals or metal alloys shall be classified as flammable solids when they can beignited and the reaction spreads over the whole length of the sample in 10 minutes or less.2.7.2.3. A flammable solid shall be classified in one of the two categories for this class using MethodN.1 as described in 33.2.1 of the UN Recommendations on the Transport of Dangerous Goods,Manual of Tests and Criteria.Table 2.7.1Criteria for flammable solidsCategoryBurning rate testCriteria1Substances and mixtures other than metal powders:(a) wetted zone does not stop fire and(b) burning time < 45 seconds or burning rate > 2.2 mm/sMetal powders:burning time 5 minutesBurning rate test2Substances and mixtures other than metal powders:(a) wetted zone stops the fire for at least 4 minutes and(b) burning time < 45 seconds or burning rate > 2.2 mm/sNoteMetal powders:burning time > 5 minutes and 10 minutesThe test shall be performed on the substance or mixture in its physical form as presented. If, forexample, for the purposes of supply or transport, the same chemical is to be presented in a physicalform different from that which was tested and which is considered likely to materially alter itsperformance in a classification test, the substance shall also be tested in the new form.2.8.4.4 Testing and evaluation of hazard informationFor safety reasons, it is advisable to test for explosive and self-reactive properties first and torule out pyrophoric behaviour before performing this test. The classification test is describedin sub-section 33.2.1.4.3.2 of the UN-MTC. The sample should be tested in its commerciallyrelevant form. Special care has to be taken that the sample forms an unbroken strip or powder125

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012train in the test mould. Large pieces that do not fit into the mould should be gently crushed.For pasty or sticking substances it may be helpful to line the mould with a thin plastic foilwhich is withdrawn after having formed the train. Classification is based upon the fastestburning rate / shortest burning time obtained in six test runs, unless a positive result isobserved earlier. For substances and mixtures other than metal powders, the category isassigned depending on whether the wetted zone is able to stop the flame.2.8.4.5 Decision logicNOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.Decision logic for Flammable solids (Decision logic 2.7 of GHS Revision 2):The substance/mixture is a solidScreening testNegativeNot classifiedPositiveBurning rate test:(a) For substances or mixtures other than metal powders:Burning time < 45 s or burning rate > 2.2 mm/s?(b) Metal powders: Burning time 10 min?NoNot classifiedYesCategory 1(a) For substances or mixtures other than metal powders:Does the wetted zone stop propagation of the flame?(b) Metal powders: Burning time > 5 min?NoDangerYesCategory 2Warning2.8.5 Hazard communication for flammable solids126

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.8.5.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 2.7.3. Table 2.7.2Label elements for flammable solidsClassification Category 1 Category 2GHS PictogramsSignal Word Danger WarningHazard Statement H228: Flammable Solid H228: Flammable SolidPrecautionary StatementPreventionP210P240P241P280P210P240P241P280Precautionary StatementP370 + P378ResponsePrecautionary StatementStoragePrecautionary StatementDisposalThe wording of the Precautionary Statements is found in CLP Annex IV, Part 2.P370 + P3782.8.6 Re-classification of substances and mixtures classified as flammable solidsaccording to DSD or already classified for transport2.8.6.1 Re-classification of substances and mixtures classified inaccordance with DSDIn most cases, solid substances and mixtures classified as “F; R11” according to DSD willtranslate into a flammable solid in CLP. However, such a translation is not unambiguous, andeach case should be carefully checked. A substance or mixture classified as “F; R11” mightbe self-reactive or even – in rare cases – explosive according to CLP. Factors like chemicalstructure, energy content and decomposition onset as obtained from a Differential ScanningCalorimetry measurement should be taken into account where an unambiguous decisioncannot be taken.Once the classification as flammable solid is established, the assignment of the correctcategory remains difficult. In case of any doubt, a conservative approach should be taken, andCategory 2 should be assigned only if the decision can be reasonably justified.2.8.6.2 Relation to transport classificationIf a transport classification is available, the following translation applies. It should be kept inmind that transport classification is based on prioritisation of hazards (see ADR, section2.1.3.5.3) and that flammable solids have a relatively low rank in the precedence of hazards.Therefore, the translation from transport classification to CLP using the table below should beonly done if a transport classification as shown is explicitly available. The conclusion that asubstance or mixture not classified as flammable solid for transport should not be classified asa flammable solid according to CLP is, in general, not correct.127

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.8.7 Examples of classification for flammable solids2.8.7.1 Example of substances and mixtures fulfilling theclassifiction criteriaThe following example shows a classification based on test data:Test substance: „Flammalene“ (organic material, solid):Screening test (VDI 2263, part 1): Burning index: 5 (burning with an open flame or emissionof sparks)Conclusion: Substance is candidate for classification as a flammable solid, further testingrequired.UN N.1 test (Test method for readily combustible solids):Burning times for a distance of 100 mm (6 runs): 44 s; 40 s; 49 s; 45 s; 37 s; 41 s.Shortest burning time is less than 45 s; substance is a flammable solid.Wetted zone stops the fire, no reignition.Conclusion: Classify as flammable solid, Category 2.2.8.7.2 Examples of substances and mixtures not fulfilling theclassification criteriaMany inorganic salts and oxides are not flammable such as NaCl, NaBr, KI, FeO, MnO etc.Urea or phthalic acid anhydride are examples of organic substances that would not beclassified as flammable solids.2.9 SELF-REACTIVE SUBSTANCES2.9.1 IntroductionIn general, substances classified as self-reactive substances can decompose stronglyexothermically when 50 kg are exposed to temperatures of 75 °C or lower depending on theSelf-Accelerating Decomposition Temperature (SADT) of the substance or mixture.Self-reactive substances and mixtures display a very wide range of properties. The mosthazardous type is TYPE A of self-reactive substances and mixtures that are too dangerous totransport commercially though they can be stored safely with appropriate precautions. At theother end of the scale this classification includes substances that only decompose slowly attemperatures well above the normal storage and transport temperatures (e.g. 75 °C).The decomposition of self-reactive substances can be initiated by heat, contact with catalyticimpurities (e.g. acids, heavy-metal compounds, and bases), friction or impact. The rate ofdecomposition increases with temperature and varies with the substance. Decomposition,particularly if no ignition occurs, may result in the evolution of toxic gases or vapours. Forcertain self-reactive substances, the temperature shall be controlled during storage andhandling. Some self-reactive substances may decompose explosively, particularly if confined.This characteristic may be modified by the addition of diluents or by the use of appropriatepackaging. Some self-reactive substances burn vigorously. Self-reactive substances are, forexample, some compounds of the types listed below:(a) Aliphatic azo compounds (-C-N=N-C-);(b) Organic azides (-C-N 3 );128

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(c) Diazonium salts (-CN + 2 Z - );(d) N-nitroso compounds (-N-N=O); and(e) Aromatic sulfohydrasides (-SO 2 -NH-NH 2 ).This list is not exhaustive and substances with other reactive groups, combination of groupsand some mixtures of substances may have similar properties. Additional guidance onsubstances, which may have self-reactive properties, is given in Appendix 6, section 5.1 of theUN-MTC.Additional hazardous properties, resulting in subsidiary labelling, are indicated in the list ofalready classified self-reactive substances incorporated included in the UN RTDG, Section2.4.2.3.2.3.Neither the burning properties nor the sensitivity to impact and friction form part of theclassification procedure for self-reactive substances in CLP. These properties may be ofimportance in safe handling of self-reactive substances (see additional tests in Section2.9.3.3.2).Commercial self-reactive substances are commonly formulated by dilution with solid andliquid substances with which they are compatible.2.9.2 Definitions and general considerations for the classification of self-reactivesIn CLP the following definition is given for self-reactive substances:Annex I: 2.8.1.1. Self-Reactive substances or mixtures are thermally unstable liquid or solidsubstances or mixtures liable to undergo a strongly exothermic decomposition even withoutparticipation of oxygen (air). This definition excludes substances and mixtures classified according tothis Part as explosives, organic peroxides or as oxidising.2.8.1.2. A self-reactive substance or mixture is regarded as possessing explosive properties when inlaboratory testing the formulation is liable to detonate, to deflagrate rapidly or to show a violent effectwhen heated under confinement.General considerationsAnnex I, 2.8.3.Hazard communicationType G has no hazard communication elements assigned but shall be considered for propertiesbelonging to other hazard classes.2.9.3 Classification of substances and mixtures as self-reactive2.9.3.1 Identification of hazard informationThe classification of a self-reactive substance in one of the seven categories “Types A to G” isdependent on its detonation, explosive thermal explosion and deflagrating properties, itsresponse to heating, the concentration and the type of diluent added to desensitize thesubstance. Specifications of acceptable diluents that can be used safely are given in the UNRTDG, Section 2.4.2.3.5.The classification of a self-reactive substance as Type A, B or C is also dependent on the typeof packaging in which the substance is tested as it affects the degree of confinement to whichthe substance is subjected. This has to be considered in when handling of the substance;stronger packaging may result in more violent reactions when the substance decomposes. Thisis why it is important that storage and transport is done in packaging, allowed for the type of129

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012self-reactive substance, that conforms the requirements of the UN-packaging or IBCinstruction (P520/IBC520) or tank instruction (T23).The traditional aspects of explosive properties, such as detonation, deflagration and thermalexplosion, are incorporated in the decision logic Figure 2.8.1 of CLP (see Section 2.9.3.4).Consequently, the determination of explosive property properties determination as prescribedin the hazard class explosives needs not to be conducted for self-reactive substances.1302.9.3.2 Classification criteriaAccording CLP, substances and mixtures should be considered for classification in thishazard class, unless:Annex I: 2.8.2.1.(a) They are explosives, according to the criteria given in 2.1;(b) They are oxidising liquids or solids, according to the criteria given in 2.13 or 2.14, exceptthat mixtures of oxidising substances, which contain 5% or more of combustible organicsubstances shall be classified as self-reactive substances according to the proceduredefined in 2.8.2.2;(c) They are organic peroxides, according to the criteria given in 2.15;(d) Their heat of decomposition is less than 300 J/g; or(e) Their self-accelerating decomposition temperature (SADT) is greater than 75°C for a50 kg package (See United Nations Manual of Tests and Criteria, sub-sections 28.1, 28.2,28.3 and Table 28.3.)2.8.2.2. Mixtures of oxidising substances, meeting the criteria for classification as oxidisingsubstances, which contain 5% or more of combustible organic substances and which do not meet thecriteria mentioned in (a), (c), (d) or (e) in 2.8.2.1, shall be subjected to the self-reactive substancesclassification procedure;Such a mixture showing the properties of a self-reactive substance type B to F (see 2.8.2.3) shall beclassified as a self-reactive substance.In addition to the above, substances and mixtures should be considered for classification inthis hazard class unless:(f) There are no chemical groups present in the molecule associated with explosive or selfreactiveproperties; examples of such groups are given in Tables A6.1 and A6.2 in the UN RTDG,Manual of Tests and Criteria, Appendix 6.In the CLP decision logic (see Section 2.9.3.4), classification of self-reactive substances isbased on performance based testing in both small scale tests and, where necessary, somelarger scale tests with the substance in its packaging. The concept of “intrinsic properties” is,therefore, not necessarily, applicable to this hazard class.Self-reactive substances are classified in one of the seven categories of “Types A to G”according to the classification criteria given in Section 2.8.2.3 of Annex I, of CLP. Theclassification principles are given in the decision logic in Figure 2.8.1 of CLP (see Section2.9.3.4) and the test series A to H, as described in the Part II of the UN-MTC, should beperformed.Annex I: 2.8.2.3. Self-reactive substances and mixtures shall be classified in one of the sevencategories of ‘types A to G’ for this class, according to the following principles:(a) any self-reactive substance or mixture which can detonate or deflagrate rapidly, as packaged, shall

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012be defined as self-reactive substance TYPE A;(b) any self-reactive substance or mixture possessing explosive properties and which, as packaged,neither detonates nor deflagrates rapidly, but is liable to undergo a thermal explosion in that packageshall be defined as self-reactive substance TYPE B;(c) any self-reactive substance or mixture possessing explosive properties when the substance ormixture as packaged cannot detonate or deflagrate rapidly or undergo a thermal explosion shall bedefined as self-reactive substance TYPE C;(d) any self-reactive substance or mixture which in laboratory testing:(i) detonates partially, does not deflagrate rapidly and shows no violent effect when heatedunder confinement; or(ii) does not detonate at all, deflagrates slowly and shows no violent effect when heatedunder confinement; or(iii) does not detonate or deflagrate at all and shows a medium effect when heated underconfinement;shall be defined as self-reactive substance TYPE D;(e) any self-reactive substance or mixture which, in laboratory testing, neither detonates nordeflagrates at all and shows low or no effect when heated under confinement shall be defined as selfreactivesubstance TYPE E;(f) any self-reactive substance or mixture which, in laboratory testing, neither detonates in thecavitated state nor deflagrates at all and shows only a low or no effect when heated under confinementas well as low or no explosive power shall be defined as self-reactive substance TYPE F;(g) any self-reactive substance or mixture which, in laboratory testing, neither detonates in thecavitated state nor deflagrates at all and shows no effect when heated under confinement nor anyexplosive power, provided that it is thermally stable (SADT is 60 o C to 75 o C for a 50 kg package),and, for liquid mixtures, a diluent having a boiling point not less than 150 o C is used fordesensitisation shall be defined as self-reactive substance TYPE G. If the mixture is not thermallystable or a diluent having a boiling point less than 150 o C is used for desensitisation, the mixture shallbe defined as self-reactive substance TYPE F.Where the test is conducted in the package form and the packaging is changed, a further test shall beconducted where it is considered that the change in packaging will affect the outcome of the test.A list of currently classified self-reactive substances is included in the UN RTDG, Section2.4.2.3.2.3.2.9.3.3 Testing and evaluation of hazard information2.9.3.3.1 Thermal stability tests and temperature controlIn addition to the classification tests given in decision logic Figure 2.8.1 of CLP, the thermalstability of the self-reactive substances has to be assessed in order to determine the Self-Accelerating Decomposition Temperature (SADT).The SADT is defined as the lowest temperature at which self-accelerating decomposition mayoccur with a substance in the packaging as used in transport, handling and storage. The SADTis a measure of the combined effect of the ambient temperature, decomposition kinetics,package size and the heat transfer properties of the substance and its packaging.There is no relation between the SADT of a self-reactive substance and its classification inone of the seven categories “Types A to G”. The SADT is used to derive safe handling,storage and transport temperatures (control temperature) and alarm temperature (emergencytemperature).Depending on its SADT a self-reactive substance needs temperature control and the rules asgiven in CLP Annex I, 2.8.2.4, consist of the following two elements:131

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121) Criteria for temperature controlSelf-reactive substances need to be subjected to temperature control when the SADT is≤ 55° C.2) Derivation of control and emergency temperatures:Type of receptacle SADT a) Control temperature EmergencytemperatureSingle packagings and 20 °C or less20 °C below SADT 10 °C below SADTIBC’sover 20 °C to 35 °Cover 35 °C15 °C below SADT10 °C below SADT10 °C below SADT5 °C below SADTTanks < 50 °C 10 °C below SADT 5 °C below SADTa) i.e. the SADT of the substance as packaged for transport, handling and storage.It should be emphasized that the SADT is dependent on the nature of the self-reactivesubstance itself, together with the volume and heat-loss characteristics of the packaging orvessel in which the substance is handled. The temperature at which self-acceleratingdecomposition occurs falls: as the size of the packaging or vessel increases; and with increasing efficiency of the insulation on the package or vessel.The SADT is only valid for the substance as tested and when handled properly. Mixing theself-reactive substance with other chemicals, or contact with incompatible materials(including incompatible packaging or vessel material) may reduce the thermal stability due tocatalytic decomposition, and lower the SADT. This may increase the risk of decompositionand has to be avoided.2.9.3.3.2 Additional testingThe sensitivity of self-reactive substances to impact (solids and liquids) and friction (solidsonly) may be of importance for the safe handling of the substances, in the event that thesesubstances have pronounced explosive properties (e.g. rapid deflagration and/or violentheating under confinement). Test methods to determine these properties are described in testseries 3 of the UN-MTC. This information should be part of the hazard communication insafety data sheets.The flashpoint for liquid self-reactive substances is only relevant in the temperature rangewhere the product is thermally stable. Above the SADT of the product flashpointdetermination is not relevant because decomposition products are evolved.Note: In case a flashpoint determination seams reasonable (expected flashpoint below theSADT) a test method using small amount of sample is recommended. In case the self-reactivesubstance is diluted or dissolved, the diluent may determine the flashpoint.Although there are currently no dedicated storage guidelines for self-reactive substances(although in some countries under development), often the regulations for organic peroxidesare referred to. For storage classification the burning rate is commonly used, see Section 2.14on organic peroxides.The determination of the auto ignition temperature is not relevant for self-reactive substances,because the vapours decompose during the execution of the test. Available test methods arefor non-decomposing vapour phases. Auto ignition of self-reactive substance vapours when132

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012they decompose, can never be excluded. This information should be part of the hazardcommunication in safety data sheets.Also self-ignition temperature determination (test applicable for solids) is not relevant. Thethermal stability of self-reactive substances is quantitatively given by the SADT test.2.9.3.3.3 Additional classification considerationsDetermination of explosive property properties is incorporated in the classification decisionlogic. Flammability is not incorporated in the decision flow chart.Currently, the following properties are not incorporated in CLP: mechanical sensitivity i.e. impact and friction sensitivity (for handling purposes); burning tests (for storage purposes); and flammability aspects.In addition to the GHS criteria CLP mentions that:Annex I: 2.8.2.2Where the test is conducted in the package form and the packaging is changed, a further test shall beconducted where it is considered that the change in packaging will affect the outcome of the test.2.9.3.4 Decision logicThe following decision logic for self-reactive substances is applicable according to CLP.NOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.133

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: Figure 2.8.1Self reactive substances and mixtures134

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.9.4 Hazard communication for self-reactives2.9.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAccording to CLP the following label elements shall be used for substances or mixturesmeeting the criteria for this hazard class:Annex I: 2.8.3. Table 2.8.1Label elements for self-reactive substances and mixturesClassification Type A Type B Type C & D Type E & F Type GGHS pictogramsSignal words Danger Danger Danger WarningHazard Statement H240:PrecautionarystatementPreventionPrecautionarystatementResponsePrecautionarystatementStoragePrecautionarystatementDisposalHeating maycause anexplosionP210P220P234P280P370 + P378P370 + P380+ P375P403 + P235P411P420H241:Heating maycause a fire orexplosionP210P220P234P280P370 + P378P370 + P380+ P375P403 + P235P411P420H242:Heating maycause a fireP210P220P234P280P370 + P378P403 + P235P411P420H242:Heatingmay cause afireP210P220P234P280P370 + P378P403 + P235P411P420P501 P501 P501 P501The wording of the Precautionary Statements is found in CLP Annex IV, Part 2.There areno labelelementsallocated tothis hazardcategory2.9.5 Re-classification of substances and mixtures classified as self-reactivesaccording to DSD or already classified for transport2.9.5.1 Re-classification of substances and mixtures classified inaccordance with DSD135

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In DSD no hazard class “self-reactive substances” is defined. In CLP self-reactive substancesare a distinct hazard class. Self-reactivity is not a single “intrinsic property”; self-reactivesubstances are a group of substances that can release a certain amount of decompositionenergy and which may be thermally unstable (see Section 2.9.1).In DSD explosive properties and flammability are determined separately by the tests A14 (forexplosive properties) and A9 or A10 (for flammable properties), as published in the CouncilRegulation (EC) No 440/2008.Substances earlier listed under other hazard categories may now under CLP fulfil the criteriaof a self-reactive substance. For the correct assignment of an individual self-reactivesubstance, the classification criteria as given in Section 2.9.3.2 should be applied. Ifnecessary, expert advice should be sought.Consequently the translation table in Annex VII to CLP is not applicable to this hazard class.2.9.5.2 Relation to transport classificationA list of already classified self-reactive substances is included in RTDG, Section 2.4.2.3.2.3.This table includes self-reactive substances type B-type F.2.9.6 Examples of classification for self-reactives1362.9.6.1 Examples of substances and mixtures fulfilling theclassification criteriaSubstance to be classified: NPMolecular formula: n.a.According to GHS 2.8.2.1, the substance has: an energy content of 1452 kJ/kg; and a SADT of 45 °C;and consequently it has to be considered for classification in the hazard class Self-ReactiveSubstances.Test results and classification according to CLP decision logic 2.8.1 for Self-ReactiveSubstances and the UN Recommendations on the transport of Dangerous Goods, Manual ofTests and Criteria, Part II, is as follows:Classification test results1. Name of the Self-Reactive Substance : NP2. General data2.1. Composition : NP, technically pure2.2. Molecular formula : n.a.2.3. Physical form : solid, fine powder2.4. Colour : brown2.5. Density (apparent) : 460 kg/m 33. Detonation (test series A)Box 1 of the decision logic : Does the peroxide propagate a detonation?3.1. Method : UN Test A.1: BAM 50/60 steel tube test3.2. Sample conditions : technically pure substance3.3. Observations : fragmented part of the tube: 12, 18cm

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.4. Result : No3.6. Exit : 1.34. Deflagration (test series C)Box 5 of the decision logic : Does the peroxide propagate a deflagration?4.1. Method 1 : Time/pressure test (test C.1)4.1.1. Sample conditions : ambient temperature4.1.2. Observations : 498, 966, 3395 ms4.1.3. Result : Yes, slowly4.2. Method 2 : Deflagration test (test C.2)4.2.1. Sample conditions : temperature: 20 °C4.2.2. Observations : deflagration rate: 0.90, 0.87 mm/s4.2.3. Result : Yes, slowly4.3. Final result : Yes, slowly4.4. Exit : 5.25. Heating under confinement (test series E)Box 8 of the decision logic:What is the effect of heating it under definedconfinement?5.1. Method 1 : Koenen test (test E.1)5.1.1. Sample conditions : -5.1.2. Observations : limiting diameter: < 1.0 mmfragmentation type "A"5.1.3. Result : Low5.2. Method 2 : Dutch pressure vessel test(test E.2)5.2.1. Sample conditions : -5.2.2. Observations : limiting diameter:

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Temperature control : Needed based on SADT (45 °C, in package)Control temperature* : 35°C (in package)Emergency temperature* : 40°C (in package)*see UN-TDG, manual of tests and criteria, table 28.2Additional remarks(1) Control and emergency temperatureThe Control and Emergency temperatures are based on the SADT as determined by UN testH.4. The Dewar vessel used in the UN H.4 test was representative for the substance handledin packages. For handling of the substance in larger quantities (IBCs/tanks/vessels etc.) and/orin better (thermally) insulated containers under more thermal insulated conditions, the SADThas to be determined for that quantity with the given that degree of insulation factor. Fromthat SADT the Control and Emergency temperatures can be derived (see also section 2.3)(2) Explosive propertiesThe explosive properties do not have to be determined according to Annex I, Chapter 2.1 forexplosives, because this is incorporated in the decision logic. Substance may have explosiveproperties when handled under more confined conditions of greater confinement.138

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 2.9.6.1: Decision logic for self reactive substance example: NP, technically pureSUBSTANCE/MIXTURE2.1 Yes6.1 YesBox 2Test BCan itdetonate as packaged?3.1Yes, rapidlyBox 6Test DDoes itdeflagrate rapidlyin package?7.1ViolentBox 10Test GCan itexplode as packagedfor transport?10.1 Yes2.2 No6.2 No7.2 Medium7.3 Low7.4 NoCan it propagatea deflagration10.2 No1.1 YesBox 3Test C?3.2 Yes, slowly3.3 NoBox 7Test EWhat isthe effect of heatingunder confinement?Box 1Test A4.1Yes, rapidly4.2 Yes, slowly4.3 No8.1ViolentDoes it propagatea detonation?1.2 Partial8.2 Medium8.3 Low8.4 NoBox 4Test CCan it propagatea deflagration?Box 8Test EWhat isthe effect of heatingunder confinement?1.3 No5.2 Yes, slowlyBox 5Test CCan it propagate5.1 a deflagrationYes, rapidly ? 5.3 No9.1Violent9.2 MediumBox 9Test EWhat isthe effect of heatingunder confinement?Box 11Packagedinthan 400 kg/450 l or to beconsidered forexemption?11.2 No9.3 Low9.4 No11.1 YesBox 12Test FWhat is its12.1explosive powerNot low ?12.2 Low12.3 NoneWhat is the effectof heating it under defined13.1 Low confinement?13.2 NoneBox 13Test EType A Type B Type C Type D Type E Type F Type G139

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.10 PYROPHORIC LIQUIDS AND SOLIDS2.10.1 IntroductionPyrophoricity, i.e. the ability to spontaneously ignite in air, is the result of a reaction of asubstance or mixture with the oxygen in the air. The reaction is exothermic and has theparticularity that it starts spontaneously, i.e. without the aid of a supplied spark, flame, heat orother energy source. Another way of saying this is that the auto-ignition temperature for apyrophoric substance is lower than room (ambient) temperature.Organo-metals and organo-metalloids may be suspected of being pyrophores, as well as theirderivatives. Also organo-phosphines and their derivatives, hydrides and their derivatives,haloacetylene derivatives, and complex acetylides may show pyrophoricity (Urben, 1995).Furthermore, powders or fine particles of metals could be pyrophoric. However, althoughmany solid metallic substances, like e.g. aluminium, would be suspected of being pyrophoricwhen considering their general reactivity, they form a protective oxide-coat upon reactionwith air. This thin coat of metal oxide prevents the metal from reacting further, and hencesuch substances may not show pyrophoric behaviour in reality.There are also pyrophoric substances that do not belong to the above mentioned groups ofchemicals, i.e. the list above is not exhaustive. Since pyrophoric substances ignitespontaneously in air, pyrophoricity is a very dangerous property. In case of doubt it shouldtherefore be thoroughly investigated whether a given substance or mixture is pyrophoric.More information on pyrophoric substances can e.g. be found in Bretherick’s Handbook ofReactive Chemical Hazards (Urben, 1995).2.10.2 Definitions and general considerations for the classification pyrophoricliquids and solidsThe definitions in CLP for pyrophoric liquids and pyrophoric solids are as follows:Annex I: 2.9.1. Pyrophoric liquid means a liquid substance or mixture which, even in smallquantities, is liable to ignite within five minutes after coming into contact with air.2.10.1. Pyrophoric solid means a solid substance or mixture which, even in small quantities, isliable to ignite within five minutes after coming into contact with air.Special consideration on particle size for solidsThe finer the particle size of a solid substance or mixture, the greater the area exposed to airwill be, and since pyrophoricity is a reaction with the oxygen in air, the particle size willgreatly influence the ability to spontaneously ignite. Hence it is very important thatpyrophoric properties for solids are investigated on the substance/mixture as it is actuallypresented (including how it can reasonably be expected to be used, see Article 8(6) of CLP).This is indicated by the Note cited in CLP Annex I, 2.10.2.1.Annex I; 2.10.2.1. Note: The test shall be performed on the substance or mixture in its physical formas presented. If for example, for the purposes of supply or transport, the same chemical is to bepresented in a physical form different from that which was tested and which is considered likely tomaterially alter its performance in a classification test, the substance shall also be tested in the newform.2.10.3 Relation to other physical hazardsPyrophoric substances will react spontaneously with air already in small amounts and more orless instantaneously (within minutes). This differentiates them from self-heating substances,which also react spontaneously with air but only when in larger amounts and after an140

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012extended period of time (hours or days). A substance that is not classified as a PyrophoricLiquid or Pyrophoric Solid may thus belong to the hazard class Self-heating Substances andMixtures, and should be considered for classification in that hazard class.Pyrophoricity may be expected for certain reactive metals and some of their compounds (e.g.hydrides and other organo-metal compounds). Many of these substances will also reactvigorously with water under the production of flammable gases. Such substances may thus beclassified in the hazard class Substances and Mixtures which in Contact with Water EmitFlammable Gases, as well as in the hazard class Pyrophoric Solids or Pyrophoric Liquids. Itshould be noted in this context that water-reactive substances may also to some extent reactwith the humidity in air, although such a reaction is seldom vigorous. A substance thatspontaneously ignites in air in accordance with the test procedures is to be consideredpyrophoric, regardless of the reaction mechanism.Solids not classified as pyrophoric may still be able to burn rapidly if subjected to enoughinitiating energy, such as the flame from a gas burner, to start the reaction. Therefore theymay be subject to classification in the hazard class flammable solids, i.e. they may be 'readilycombustible solids'.Liquids not classified as pyrophoric but that can burn may belong to the hazard classflammable liquids depending on their flash point and ability to sustain combustion.2.10.4 Classification of substances and mixtures as pyrophoric liquids and solids2.10.4.1 Identification of hazard informationSince the tests to determine pyrophoricity are simple and require no special equipment, seeSection 2.10.4.4 below, there is in general no reason to go to data sources instead ofperforming tests. Furthermore, the possibilities of waiving tests are ample both for knownpyrophores and for substances and mixtures known not to be pyrophoric, see Section 2.10.4.2below. If information anyway is taken from literature or other data sources, it is of utmostimportance that the correct physical form is considered, see Section 2.1.4. Naturally, all datasources should be carefully evaluated with regard to reliability and scientific validity.2.10.4.2 Screening procedures and waiving of testingIn case a substance or mixture is known from practical handling to be pyrophoric no testing isnecessary. Such liquids and solids are classified as pyrophoric without testing. This wouldalso be the case if the substance or mixture spontaneously ignites upon opening of thereceptacle when trying to perform the tests for classification.According to the additional classification considerations in CLP Annex I, 2.9.4 and 2.10.4,the classification procedure for pyrophoric solids or liquids need not be applied whenexperience in manufacture or handling shows that the substance or mixture does not ignitespontaneously on coming into contact with air at normal temperatures (i.e. the substance isknown to be stable at room temperature for prolonged periods of time (days)).2.10.4.3 Classification criteriaSections 2.9.2.1 and 2.10.2.1 of Annex I of CLP specify the classification criteria:Annex I: 2.9.2. Table 2.9.1Criteria for pyrophoric liquidsCategoryCriteria141

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121 The liquid ignites within 5 min when added to an inert carrier and exposed to air,or it ignites or chars a filter paper on contact with air within 5 min.Annex I: 2.10.2. Table 2.10.1Criteria for pyrophoric solidsCategoryCriteria1 The solid ignites within 5 minutes of coming into contact with air.2.10.4.4 Testing and evaluation of hazard informationIn Section 2.9.2.1 and 2.10.2.1 of Annex I of CLP reference to the test-methods are made:Annex I: 2.9.2.1. A pyrophoric liquid shall be classified in a single category for this class using testN.3 in part III, sub-section 33.3.1.5 of the UN Recommendations on the Transport of DangerousGoods, Manual of Tests and Criteria.2.10.2.1. A pyrophoric solid shall be classified in a single category for this class using test N.2 inpart III, sub-section 33.3.1.4 of the UN Recommendations on the Transport of Dangerous Goods,Manual of Tests and Criteria.The N.2 and N.3 tests for pyrophoricity are quite simple and are sufficiently described in Part3 Section 33 of the UN-MTC. No special equipment is needed. Essentially the substance ormixture is exposed to air to see if it ignites. For liquids which do not spontaneously ignitewhen poured, the surface in contact with air is increased using a filter paper. Ignition orcharring of the filter paper is regarded as a positive response in the test, i.e. such a liquid isconsidered to be pyrophoric.It is important that samples for testing of pyrophoric properties are carefully packed andsealed. Furthermore, the material offered for testing should be freshly prepared, since thereactive properties may diminish due to aging or agglomeration. Whenever experiments are tobe done one should be careful – a pyrophoric substance may well ignite already upon openingthe receptacle!It should be noted that the mechanism of oxidation is, in general, very complex, and that thehumidity of air might influence the rate of reaction. It is known that certain metals will notreact in dry air, whereas in the presence of moisture the reaction is almost instantaneous(often even trace amounts of moisture are sufficient). Therefore a false negative may resultwhen performing the tests in an extremely dry environment, and this condition must beavoided when performing the tests for classification for pyrophoricity.2.10.4.5 Decision logicNOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.Decision logic for pyrophoric liquids (taken from GHS Rev. 2):142

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The substance/mixture is a liquidCategory 1Does it ignite within 5 min when poured into a porcelain cup filledwith diatomaceous earth or silica gel?YesNoDangerCategory 1Does it ignite or char a filter paper within 5 min?YesNoDangerNot classifiedDecision logic for pyrophoric solids (taken from GHS Rev. 2):The substance/mixture is a solidCategory 1Does it ignite within 5 min after exposure to air?YesDangerNoNot classified143

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.10.5 Hazard communication for pyrophoric liquids and solids2.10.5.1 Pictograms, signal words, hazard statements andprecautionary statementsThe hazard communication is the same for liquids and for solids, except for one of theprecautionary statements (P335 for solids and P302 for liquids):Annex I: 2.9.3 & 2.10.3GHS PictogramLabel elements for pyrophoric liquids (Table 2.9.2) and solids (Table 2.10.2)Classification Category 1, liquids Category 1, solidsSignal word Danger DangerHazard statementH250: Catches firespontaneously if exposed to airH250: Catches firespontaneously if exposed to airPrecautionary StatementPreventionP210P222P210P222P280P302 + P334P370 + P378P280P335 + P334P370 +P378Precautionary StatementResponsePrecautionary StatementStorageP422P422Precautionary StatementDisposalThe wording of the Precautionary Statements is found in CLP Annex IV, Part 2.2.10.6 Re-classification of substances and mixtures classified as pyrophoric liquidsand solids according to DSD or already classified for transport2.10.6.1 Re-classification of substances and mixtures classified inaccordance with DSDAccording to the DSD, the A.13 test in EC-Regulation 440/2008 is used to characterise thepyrophoric properties of solids and liquids. Substances or mixtures reacting positively in theA.13-test are assigned the risk phrase R17 – 'Spontaneously flammable in air'.The test methods used to determine pyrophoric properties in CLP are methods N.2 (for solids)and N.3 (for liquids) as described in Part 3 Section 33 of the UN-MTC. These tests methodsare identical to the A.13-test used in the DSD, apart from details in the environmentalconditions. The A.13-test specifies a temperature of circa 20C, but does not specify the airhumidity. In the N.2 and N.3 tests on the other hand, specific environmental conditions areonly given for the filter paper test (25 ± 2°C and relative humidity 50 ± 5 %).A small difference in temperature or humidity could possibly result in a slight change inreaction rate or a delayed effect, but unless extreme environmental parameters have beenapplied (such as an extremely low air humidity or extreme temperatures) this is unlikely tohave any effect on the outcome as far as classification is concerned. Therefore the N.2 and144

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012N.3 test methods can be regarded the same as the A.13 test method as described in CouncilRegulation (EC) No 440/2008 for both solids and liquids in virtually all cases.The CLP hazard classes pyrophoric solids and pyrophoric liquids each contain only a singlehazard category (Category 1), and the classification criteria for this category are identical tothat for assignment of the R17 risk phrase for both solids and liquids. So in virtually all cases,substances and mixtures that have been assigned the risk phrase R17 on the basis of the resultof the A.13-test fall into Category 1 of the hazard class Pyrophoric Solids if they are solid andCategory 1 of the hazard class Pyrophoric Liquids if they are liquid. Normally no re-testing isthus required. The straight translation from R17 is also reflected in Annex VII to CLP.2.10.6.2 Relation to transport classificationThe tests N.2 and N.3 that are used for classification for pyrophoricity according to CLP arealso those used for classification in the subdivision pyrophoric substances in Class 4.2(Substances liable to spontaneous combustion) according to the RTDG. The criteria forCategory 1 according to CLP (which is the only category for pyrophoric liquids andpyrophoric solids) and for packing group I in Class 4.2 according to the ADR are also exactlythe same. Furthermore, all pyrophoric substances and mixtures are assigned to packing groupI, which is also used exclusively for pyrophoric substances and mixtures.Therefore, any substance or mixture assigned to Class 4.2 packing group I according to ADRwill be classified in Category 1 of the hazard classes pyrophoric liquids or pyrophoric solidsaccording to CLP. Naturally, if the substance or mixture is a liquid it belongs to pyrophoricliquids, and if it is a solid it belongs to pyrophoric solids.2.10.7 Examples of classification for pyrophoric liquids and solidsPlease note that the substance names in this chapter are fictitious.2.10.7.1 Examples of substances and mixtures fulfilling theclassification criteriaExample 1:Name: PyroferilPhysical state: SolidPyrophoric properties: Pyroferil is known to self-ignite upon contact with air atambient conditions.Classification: Pyrophoric solid Category 1Example 2:Name: ZorapyrolePhysical state: SolidPyrophoric properties: Unknown, therefore the N.2-test of the UN RTDG – Manual ofTests and Criteria was applied.Test result: When poured from one meter height according to the test procedure,Zorapyrole self-inited after two minutes already in the first trial.Classification: Pyrophoric solid Category 1Example 3:Name: Pyrpherdine145

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Physical state: LiquidPyrophoric properties: Unknown, therefore the N.3-test of the UNRTDG – Manual ofTests and Criteria was applied. However, when opening the receptacle in order toperform the test, Pyrpherdine self-ignited.Classifiction: Pyrophoric liquid Category 1Example 4:Name: QulipyrPhysical state: LiquidPyrophoric properties: Unknown, therefore the N.3-test of the UN-MTC was applied.Test result: When poured according to the test procedure, nothing happened. Theprocedure was repeated six times, each time giving a negative result (i.e. no ignition).Therefore Qulipyr was supplied to a filter paper in accordance with the test method. Inthe second trial the filter paper was charred within five minutes.Classification: Pyrophoric liquid Category 12.10.7.2 Examples of substances and mixtures not fulfilling theclassification criteriaExample 1:Name: NonopyrPhysical state: SolidPyrophoric properties: Nonopyr has been handled extensively in air and has neverself-ignited. From the chemical structure no pyrophoricity is expected.Classification: Not a pyrophoric solidExample 2:Name: PyronotPhysical state: SolidPyrophoric properties: Unknown, therefore test N.3 of the UN-MTC was applied.Test result: When poured from one meter height according to the test procedure noignition occurred within five minutes. The procedure was repeated six times and eachtime the result was negative.Classification: Not a pyrophoric solidExample 3:Name NotpyratalPhysical state: LiquidPyrophoric properties: Unknown, therefore test N.3 of the UN-MTC was applied.Test result: When poured according to the test procedure nothing happened in either ofsix trials. Therefore Notpyratal was supplied to a filter paper in accordance with thetest method, whereupon no ignition or charring occurred in either of three trials.Classification: Not a pyrophoric liquid146

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.10.8 ReferencesUrben, P.G. (ed), Bretheric’s Handbook of Reactive Chemical Hazards, 5:th ed., Butterworth-Heinemann, Oxford (1995)2.11 SELF-HEATING SUBSTANCES AND MIXTURES2.11.1 IntroductionSelf-heating is the result of an exothermic reaction of a substance or mixture with the oxygenin the air. Initially, the reaction rate may be very small. However, when the heat producedcannot be removed rapidly enough (i.e. heat accumulation), the substance or mixture will selfheat,with the possible consequence of self-ignition. The phenomenon can occur only where alarge surface of substance or mixture is in contact with air or oxygen (for example, piles ofpowders, crystals, splinters, any other rough surface etc.). The initiation occurs usually at ornear the centre of the substance pile with the available air in the interspace between theparticles.2.11.2 Definitions and general considerations for the classification of self-heatingsubstances and mixturesThe definitions in CLP for self-heating substances and mixtures are as follows:Annex I: 2.11.1.1. A self-heating substance or mixture is a liquid or solid substance or mixture, otherthan a pyrophoric liquid or solid, which, by reaction with air and without energy supply, is liable to selfheat;this substance or mixture differs from a pyrophoric liquid or solid in that it will ignite only when inlarge amounts (kilograms) and after long periods of time (hours or days).2.11.1.2. Self-heating of substances or mixtures, leading to spontaneous combustion, is caused byreaction of the substance or mixture with oxygen (in the air) and the heat developed not being conductedaway rapidly enough to the surroundings. Spontaneous combustion occurs when the rate of heatproduction exceeds the rate of heat loss and the auto-ignition temperature is reached.2.11.3 Relation to other physical hazardsPyrophoric solids and liquids should not be considered for classification as self-heatingsubstances and mixtures.2.11.4 Classification of self-heating substances and mixtures2.11.4.1 Identification of hazard informationSelf-heating is a very complex phenomenon which is influenced by many parameters (someof them being volume, temperature, particle shape and size, heat conductivity and bulkdensity). Therefore, self-heating behaviour cannot be predicted from any theoretical model. Insome cases, properties might even differ between producers of seemingly very similarsubstances or mixtures. Differences in self-heating behaviour are especially to be anticipatedwhere surface treatment occurs in the production process. Hence, all data sources should becarefully evaluated with regard to reliability and scientific validity.It is of utmost importance that in compliance with Articles 5 and 6 of CLP authentic andrepresentative material in the correct form and physical state be used for testing. In manycases, a simple screening test (see Section 2.11.4.2) can be used to determine whether selfheatingoccurs or not.147

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.11.4.2 Screening procedures and waiving of testingAnnex I: 2.11.4.2. The classification procedure for self-heating substances or mixtures need not beapplied if the results of a screening test can be adequately correlated with the classification test and anappropriate safety margin is applied. Examples of screening tests are:(a)(b)The Grewer Oven test (VDI guideline 2263, part 1, 1990, Test methods for the De-terminationof the Safety Characteristics of Dusts) with an onset temperature 80 K above the referencetemperature for a volume of 1 l;The Bulk Powder Screening Test (Gibson, N. Harper, D.J. Rogers, R. Evaluation of the fireand explosion risks in drying powders, Plant Operations Progress, 4 (3), 181-189, 1985) withan onset temperature 60 K above the reference temperature for a volume of 1 l.Test method A.16 as described in Council Regulation (EC) No 440/2008 checks for selfheatingproperties. However, the method used is generally inappropriate for a soundassessment, and the findings do not lead to a classification. Therefore, special care must betaken if results from A.16 testing are interpreted towards a CLP classification for self-heatingsubstances and mixtures.In general, liquids are not classified as self-heating since the phenomenon applies only tosolids (i.e. the surface for reaction with air is not large enough) and the test method is notapplicable to liquids. However, if liquids are absorbed on a large surface (e.g. on powderparticles), a self-heating hazard should be considered.Substances with a low melting point (< 160 °C) should not be considered for classification inthis class since the melting process is endothermic and the substance-air surface is drasticallyreduced. However, this criterion is only applicable if the substance or mixture is completelymolten up to this temperature.2.11.4.3 Classification criteriaA self-heating substance or mixture shall be classified in one of the two categories for thisclass if, in a test performed in accordance with test method N.4 in Part III, subsection33.3.1.6 of the UN-MTC, the result meets the criteria according to following table:Annex I: Table 2.11.1Criteria for self-heating substances and mixturesCategory12CriteriaA positive result is obtained in a test using a 25 mm sample cube at 140°C(a) a positive result is obtained in a test using a 100 mm sample cube at 140°Cand a negative result is obtained in a test using a 25 mm cube sample at 140°Cand the substance or mixture is to be packed in packages with a volume ofmore than 3 m 3 ; or(b) a positive result is obtained in a test using a 100 mm sample cube at 140°Cand a negative result is obtained in a test using a 25 mm cube sample at140°C, a positive result is obtained in a test using a 100 mm cube sample at120°C and the substance or mixture is to be packed in packages with a volumeof more than 450 litres; or(c) a positive result is obtained in a test using a 100 mm sample cube at 140°Cand a negative result is obtained in a test using a 25 mm cube sample at 140°Cand a positive result is obtained in a test using a 100 mm cube sample at100°C.148

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.11.2.3. Substances and mixtures with a temperature of spontaneous combustion higher than 50°Cfor a volume of 27 m³ shall not be classified as a self-heating substance or mixture.2.11.2.4. Substances and mixtures with a spontaneous ignition temperature higher than 50°C for avolume of 450 litres shall not be assigned to Category 1 of this class.2.11.2.2. Note: The test shall be performed on the substance or mixture in its physical form aspresented. If, for example, for the purposes of supply or transport, the same chemical is to bepresented in a physical form different from that which was tested and which is considered likely tomaterially alter its performance in a classification test, the substance shall also be tested in the newform.2.11.4.4 Testing and evaluation of hazard informationA self-heating substance or mixture shall be classified in one of the two categories for thisclass using test method N.4 in Part III, sub-section 33.3.1.6 of the UN-MTC.2.11.4.4.1 General remarksIf self-heating behaviour cannot be ruled out by a screening test, further testing becomesnecessary. UN test method N.4 as described in the latest version of the UN-MTC should beused.Explosive substances should not be tested according to this method. For safety reasons, it isadvisable to test for explosive and self-reactive properties and to rule out pyrophoric behaviorbefore performing this test. The oven should be equipped with an appropriate pressure-reliefdevice in case an energetic decomposition is triggered by a temperature rise. For samplescontaining flammable solvents explosion protection measures have to be taken.The tests may be performed in any order. It is suggested to start with the 25 mm sample cubeat 140 °C. If a positive result is obtained, the substance or mixture shall be classified as a selfheatingsubstance or mixture, Category 1, and no further testing is necessary.The test procedure need not be applied if the substance or mixture is completely molten at 160°C.2.11.4.4.2 Sample preparationThe sample (powder or granular) in its commercial form should be used. The material shouldnot be milled or ground. It should be filled to the brim of the sample container and thecontainer tapped several times. If the sample settles, more is added. If the sample is heaped itshould be levelled to the brim. The sample container is placed in the oven as described in theUN Manual.2.11.4.4.3 Criteria and evaluationA positive result is obtained if spontaneous ignition occurs or if the temperature of the sampleexceeds the oven temperature by 60 K. The testing time is 24 hours. The time count startswhen the temperature in the centre of the sample has reached a value of 2 K below the oventemperature. This is especially important when the sample contains solvents which evaporateunder the test conditions or when larger test volumes are used for extrapolation purposes (seebelow).Before starting test series UN N.4, the decomposition behaviour of the sample should beknown. In general, it is sufficient to perform a screening with Differential ScanningCalorimetry. Special care with respect to the interpretation of the test data is necessary whenexothermic decomposition may occur at the test temperatures. In such cases, a test under aninert atmosphere (i.e. nitrogen) should be run to determine the temperature rise due to149

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012decomposition. Careful flushing is essential since otherwise much air may be retainedbetween the crystals of the sample in the container.2.11.4.5 Decision logicNOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.150

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012SUBSTANCE/MIXTUREAnnex I: Figure 2.11.1.Self-heating substances and mixturesDoes it undergo dangerous self-heating when tested in a100 mm sample cube at 140°C?NONOTCLASSIFIEDYESCategory 1Does it undergo dangerous self-heating when tested in a25 mm sample cube at 140°C?YESDangerNOCategory 2Is it packaged in more than 3 m 3 ?NOYESWarningDoes it undergo dangerous self-heating when tested in a100 mm sample cube at 120°C?NONOTCLASSIFIEDYESIs it packaged in more than 450 litres volume?YESCategory 2NOWarningDoes it undergo dangerous self-heating when tested in a100 mm sample cube at 100°C? YESCategory 2NOWarningNOT CLASSIFIED151

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.11.4.6 ExemptionThe following exemptions apply (see Section 1.11.4.3):- Substances and mixtures with a temperature of spontaneous combustion higher than50°C for a volume of 27 m³ shall not be classified as a self-heating substance ormixture.- Substances and mixtures with a spontaneous ignition temperature higher than 50°C fora volume of 450 litres shall not be assigned to Category 1 of this class.However, the UN-MTC does not provide any guidance how these values should bedetermined. The UN test regime is based on the silent assumption of a cubic sample shape.For the extrapolation to larger volumes, an improved model has to be used. According toGrewer, plotting (Grewer, 1994) the logarithm of the volume to surface ratio (log (V/A))versus the reciprocal temperature gives good results without knowledge of the Frank-Kamenetzskii (Frank-Kamenetzskii, 1969) shape factor.The critical temperature for a volume of 450 l or 27 m³ can be found by extrapolation of thecritical temperature in a log (V/A) vs. 1/T plot (see Figure 2.11.4.6):Figure 2.11.4.6 Extrapolation towards large volumesTemperature in °Clog(V/A) in m-0,4-0,6-0,8-1,0-1,2-1,4-1,6-1,8-2,0-2,2-2,4-2,6-2,8240220200180Note: sample gave positiveresult at 140 °C / 1 liter160140-3,00,0019 0,0020 0,0021 0,0022 0,0023 0,0024 0,0025 0,0026 0,0027 0,0028 0,0029 0,0030 0,0031reciprocal abs. Temperature in 1/K120Extrapolation to 27 m³Linear regression lines100908070positive resultnegative result605027 m310 m31 m3100 L1,6 L400 mL110 mL16 mLSample volumeThe test setup is essentially the same as in test N.4 of the UN-MTC but now the sample sizeand possibly the shape are systematically varied. The criteria of Section 2.11.4.3 apply aswell.The critical temperature must be determined for at least four different volumes covering atleast two decades and with a volume not smaller than 16 ml. If possible, larger volumes152

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012should be also tested. The borderline temperature should be determined as precisely aspossible. For small volumes (< 1 litre), the temperature rise due to self-heating may beconsiderably less than 60 K; in this case a noticeable temperature rise is interpreted as apositive result.A conservative approach is required for the evaluation. The uncertainty of measurement mustbe taken into account. The extrapolation shall be based on a linear regression of the negativeand positive borderline data sets in the log (V/A) vs. 1/T diagram. The maximum permissibledifference between a positive and a negative result should be 5 K. An exemption may beclaimed if the more conservative endpoint for the particular volume is well beyond 50 °C (i.e.55 °C or higher).2.11.5 Hazard communication for self-heating substances and mixtures2.11.5.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 2.11.3. Table 2.11.2Label elements for self-heating substances and mixturesClassification Category 1 Category 2GHS PictogramsSignal Word Danger WarningHazard StatementH251: Self-heating; may catch H252: Self-heating in largefirequantities; may catch firePrecautionary StatementPreventionP235 + P410P280P235 + P410P280Precautionary StatementResponsePrecautionary StatementStorageP407P413P420Precautionary StatementDisposalThe wording of the Precautionary Statements is found in CLP Annex IV, Part 2.P407P413P4202.11.6 Re-classification of substances and mixtures classified according to DSD oralready classified for transport2.11.6.1 Re-classification of substances and mixtures classified inaccordance with DSDThe DSD used the A.16 test to determine the “relative self-ignition temperature for solids”.However, the method used is generally inappropriate for a sound assessment and has neverhad any relevance for classification. The A.16 test determines the oven temperature at whichthe sample temperature reaches 400 °C by self-ignition, and this criterion cannot be correlatedwith the CLP classification. Therefore, special care must be taken if results from A.16 testing153

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012are interpreted towards a CLP classification for self-heating substances and mixtures. In somecases, the original test data may be used as a screening test and may be interpreted in analogyto the Grewer Oven Test (VDI guideline 2263, part 1, 1990, Test methods for theDetermination of the Safety Characteristics of Dusts).1542.11.6.2 Relation to transport classificationIn transport, Division 4.2 – substances liable to spontaneous combustion – comprises thefollowing entries(a) Pyrophoric substances(b) Self-heating substancesWhereas pyrophoric substances in transport are assigned to packing group I, self-heatingsubstances are assigned to packing groups II and III. In cases where a substance (or mixture)is classified in Division 4.2, packing group II or III, the translation into the CLP system isstraightforward.It should be kept in mind that transport classification is based on prioritisation of hazards (seeADR, section 2.1.3.5.3) and that self-heating substances have a relatively low rank in theprecedence of hazards. Therefore, the translation from transport classification to CLP usingthe above table should be only done if a transport classification as shown is explicitlyavailable. The conclusion that a substance or mixture not classified as self-heating fortransport should not be classified as self-heating substance or mixture according to CLP is, ingeneral, not correct.2.11.7 Examples of classification for self-heating substances and mixtures2.11.7.1 Examples of substances and mixtures fulfilling theclassification criteria Many organometallic compounds, especially substances or mixtures containing transitionmetals Many organic substances or mixtures; the tendency to self-heat increases with decreasingparticle size Many metals, especially catalysts2.11.7.2 Examples of substances and mixtures not fulfilling theclassification criteriaIn general, liquids show no self-heating behaviour unless absorbed on a large surface.Scientific backgroundA basic model for the thermal explosion of solids was first developed by Frank-Kamenetzskii.It is based on the assumption that only the heat loss by thermal conduction is relevant for thephenomenon. In this case, the critical criterion for a thermal runaway reaction can bedescribed as a linear relationship between the reciprocal absolute temperature and thelogarithm of volume.The classification scheme of the UN for self-heating substances and mixtures is based oncharcoal as a reference system. The critical temperature for a 1 litre cube of charcoal is 140°C and for a cube of 27 m³ 50 °C. When a parallel line is drawn in the 1/T vs. logarithm ofvolume diagram from the reference points 1 litre / 120 °C and 1 litre / 100 °C, the

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012corresponding volumes for a critical temperature of 50 °C are found to be 3 m³ and 450 l,respectively (see Figure 2.11.7.2). The black dotted line in Figure 2.11.7.2 separates Category1 from Category 2.However, the slope of the line in the 1/T vs. volume diagram depends on the individualactivation energy of the substance or mixture, and therefore it may vary within certain limits.It must be born in mind that this test regime has been developed to facilitate classification andthat it may not suffice to solve safety issues in storage.Figure 2.11.7.2 Volume dependency of the critical temperature for charcoal0,002Grewer test2202000,00220,0024recipr.Temp.0,00260,00280,003Cat. 1Yellow dots symbolizeUN Manual test pointsCharcoal reference line(UN Manual T + C)Exemption < 450 ltrExemption < 3 m³Not self-heating18016014012010090807060Temperature (°C)Cat. 2500,00320,001 0,01 0,1 1 10 100 1000 10000 100000Volume (liter)2.11.8 ReferencesGrewer, T., Thermal hazards of chemical reactions, Elsevier, Amsterdam – London - NewYork – Tokyo 1994.Frank-Kamenetzski, D.A., Diffusion and heat transfer in chemical kinetics, 2nd edition,Plenum Press, New York, London 1969.2.12 SUBSTANCES AND MIXTURES WHICH, IN CONTACT WITH WATER,EMIT FLAMMABLE GASES2.12.1 IntroductionDepending on the chemical structure and/or the physical state (e.g. particle size) substances ormixtures may be able to react with water (even water damp / air humidity) under normalambient temperature conditions. Sometimes this reaction can be violent and/or withsignificant generation of heat. Especially if gases are evolved this reaction may become verydangerous during use. In addition, it is important to know whether a substance emits155

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012flammable gases after contact with water because special precautions are necessary especiallywith regard to explosion protection.Examples are demonstrated in the following table.Table 2.12.1 Examples of hazards, depending on the property of the emitted gas, when substances andmixtures are in contact with waterType ofemitted gasGas(in general)Flammable gasExample of the hazard• Heating up of the substance• Splashing of the substance and thus e.g. contact withskin etc. or additional risk during fire fighting• Pressure rise and bursting of e.g. the packaging, tank• Ignition• Flash of fireCLP ReferenceAnnex II, 1.1.3:Supplemental hazardinformation:EUH014*Annex I, 2.12:H260/H261Toxic gas • Damage to health: intoxication (acute) Annex II, 1.2.1:Supplemental hazardinformation:EUH029** For supplemental hazard information: see Section 2.12.4.2For substances and mixtures which, in contact with water, emit flammable gases the generalclassification principles of GHS and CLP are widely comparable.2.12.2 Definitions and general considerations for the classification of substances andmixtures which, in contact with water, emit flammable gasesThe following definition is given in CLP for substances and mixtures which, in contact withwater, emit flammable gases (CLP Annex I, 2.12).Annex I: 2.12.1. Substances or mixtures which, in contact with water, emit flammable gases meanssolid or liquid substances or mixtures which, by interaction with water, are liable to becomespontaneously flammable or to give off flammable gases in dangerous quantities.2.12.3 Classification of substances and mixtures which, in contact with water, emitflammable gases1562.12.3.1 Identification of hazard informationFor the classification of substances and mixtures which, in contact with water, emitflammable gases the following data are needed, if applicable: Chemical structure Water solubility Chemical identity and flammability of the emitted gas Pyrophoric properties of the tested substance or mixture Particle size in case of solids Friability in case of solids Hazard properties in general

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012 Information concerning the experience in production or handlingInformation about the chemical structure is used to check whether the substance or mixturecontains metals and/or metalloids (see the following chapter 2.1.1 on non-testing data).The water solubility is used to decide whether the substance or mixture is soluble in water toform a stable mixture. This may also be decided based on information concerning experiencein handling or use, e.g. the substance is manufactured with water or washed with water (seeSection 2.12.3.4.1).The chemical identity of the emitted gas is used to decide whether the evolved gas isflammable or not. If the chemical identity of the emitted gas is unknown, the gas shall betested for flammability (see Section 2.3).In case of pyrophoric substances and mixtures the test UN N.5 of the UN-MTC, Part III,section 33.4.1.4 shall be executed under nitrogen atmosphere. Therefore, in regard to CLPdata about pyrophoric properties are needed.Melting point, boiling point and information about viscosity are necessary to identify thephysical state of the substance or mixture. Even though the UN N.5 test can be applied toboth, solids and liquids, these data are necessary to decide whether information concerningthe friability (for solids) in accordance with the test method is necessary.The particle size and the friability of a solid substance or mixture are crucial parameters forthe classification of substances and mixtures which, in contact with water, emit flammablegases. These parameters have a significant effect on the test result. Thus specific requirementsregarding the particle size and the friability are prescribed in the test method UN N.5. Forfurther details regarding the test procedure see Section 2.12.3.4.1.The following references generally provide good quality data on physical hazards (see Section2.7.8 for full references):(a) Bretherick’s Handbook of Reactive Chemical Hazards (Urben, 1999)(b) ChemFinder (ChemFinder, database)(c) CHEMSAFE (contains evaluated/recommended data) (CHEMSAFE, database)(d) CRC Handbook of Chemistry and Physics (CRC, 2005)(e) GESTIS-database on hazardous substances (GESTIS database)(f) The Merck Index (Merck, 2001)2.12.3.2 Screening procedures and waiving of testingFor the majority of substances, flammability as a result of contact with water is not a typicalproperty and testing can be waived based on a consideration of the structure and experiencesin handling and use.Annex I: 2.12.4.1. The classification procedure for this class need not be applied if:a) The chemical structure of the substance or mixture does not contain metals or metalloids; orb) Experience in handling and use shows that the substance or mixture does not react with water, e.g.the substance is manufactured with water or washed with water; orc) The substance or mixture is known to be soluble in water to form a stable mixture.2.12.3.3 Classification criteria157

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Category158123Annex I: Table 2.12.1Criteria for substances or mixtures which in contact with water emit flammable gasCriteriaAny substance or mixture which reacts vigorously with water at ambient temperaturesand demonstrates generally a tendency for the gas produced to ignite spontaneously, orwhich reacts readily with water at ambient temperatures such that the rate of evolution offlammable gas is equal to or greater than 10 litres per kilogram of substance over any oneminute.Any substance or mixture which reacts readily with water at ambient temperatures suchthat the maximum rate of evolution of flammable gas is equal to or greater than 20 litresper kilogram of substance per hour, and which does not meet the criteria for Category 1.Any substance or mixture which reacts slowly with water at ambient temperatures suchthat the maximum rate of evolution of flammable gas is equal to or greater than 1 litreper kilogram of substance per hour, and which does not meet the criteria for Categories 1and 2.Note: The test shall be performed on the substance or mixture in its physical form as presented. If forexample, for the purposes of supply or transport, the same chemical is to be presented in a physicalform different from that which was tested and which is considered likely to materially alter itsperformance in a classification test, the substance must also be tested in the new form.2.12.2.2. A substance or mixture shall be classified as a substance or mixture which in contact withwater emits flammable gases if spontaneous ignition takes place in any step of the test procedure.2.12.3.4 Testing and evaluation of hazard information2.12.3.4.1 Testing procedureCare shall be taken during testing as the emitted gas might be toxic as well.The testing procedure for substances and mixtures which in contact with water emitflammable gases is sensitive to a number of influencing factors and therefore should becarried out by experienced personnel. Some of these factors are described in the following:1. Apparatus / measuring techniqueIn test method UN N.5 no special laboratory apparatus / measuring technique to determinegas evolving flow is required and no reference material is prescribed. As demonstrated in thepast by a round robin test, the gas evolution rate measured by different apparatuses may varyin a wide range. Therefore in order to avoid measuring and classification errors adequatequality control measures are necessary to validate the results and should be noted in the testreport.2. Particle size and/or friabilityThe particle size of a solid has a significant effect on the test result. Therefore, if for solids thepercentage of powder with a particle size of less than 500 µm constitutes more than 1 % of thetotal mass, or if the substance is friable, then the complete sample shall be ground to a powderbefore testing to consider a possible reduction in particle size during handling and transport.In other cases, a grinding procedure may not be applicable and/or the sample cannot beground completely to a particle size of less than 500 µm (e.g. metal granules).Information on these pre-treatments and the respective procedures, the particle size and thefriability has to be mentioned in the test report.3. Atmospheric parameters

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Variations of the atmospheric parameters (mainly air pressure and temperature) during the testhave a considerable influence on the test result. Therefore the substance or mixture shall betested at 20 °C, i.e. make sure that the test apparatus is acclimatised to 20 °C.On the other hand it is difficult to regulate and stabilise the air pressure during the testing. Tocharacterise this influencing factor and to avoid false positive results, an additional “blanktest” is strictly recommended. The results of the blank test should be noted in the test report.4. Test with demineralised (distilled) waterThe UN N.5 test is performed with demineralised (distilled) water. In practice, contact withwater can be to water in the liquid state (fresh water, sea water) or humid air, respectively.Note that the reactivity and thus the gas evolution rate observed in practice may differ fromthe gas evolution rate value measured by demineralised water. This circumstance should betaken into account when handling substances which in contact with water emit flammablegases.5. Stirring procedures during the testStirring of the sample/water mixture during the test may have a considerable effect on the testresult (e.g. significant increase or decrease of the gas evolution rate). Therefore, thesample/water mixture should not be stirred continuously during the test, e.g. by an automaticmagnetic stirrer, even if the test sample has hydrophobic properties and the moistening of thesample becomes impossible.6. Spontaneous ignitionThis term means spontaneous ignition of the evolved gas in the air but without contact to anadditional ignition source, i.e. without the flame of the gas burner.2.12.3.4.2 Evaluation of hazard informationIn order to evaluate test results the evaluator person shall have sufficient experience in theapplication of the test methods and in the disturbing / influencing factors as described above.The evaluation of data comprises two steps Evaluation of all available data and Identification of the study or studies giving rise to the highest concern (key studies).The criterion for the gas evolution rate for assignment to Category 2 or 3 amounts to 20 or 1litre per kilogram of substance per hour, respectively, but for Category 1 the relevant criterionis 10 litres per kilogram of substance over any one minute period. This has to be consideredwhile testing and for correct evaluation of the test results.The assignment to the respective hazard class / category will further determine the technicalmeans to be taken to avoid dangerous events which, in combination with other endpoints suchas i) explosion limits, ii) flash points (applicable only for liquids) or iii) self-ignitiontemperature, can lead to clear restrictions in the conditions of use.2.12.3.5 Decision logicThe decision logic and guidance which follow, are not part of the harmonized classificationsystem, but have been provided here as additional guidance.NOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.159

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 2.12.3.5 Decision logic for substances and mixtures which, in contact with water, emitflammable gases (Taken from GHS, Revision 2)Substance/mixtureIn contact with water, does it react slowly at ambient temperaturessuch that the maximum rate of evolution of flammable gas is ≥ 1 litreper kg of substance per hour?NoNot classifiedYesIn contact with water, does the substance react vigorously withwater at ambient temperatures and demonstrate generally atendency for the gas produced to ignite spontaneously, or does itreact readily with water at ambient temperatures such that the rateof evolution of flammable gas is ≥ 10 litres per kg of substance overany one minute?YesCategory 1DangerNoIn contact with water, does it react readily with water at ambienttemperatures such that the maximum rate of evolution offlammable gas is ≥ 20 litres per kg of substance per hour?YesCategory 2DangerNoCategory 3Warning160

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.12.4 Hazard communication for substances and mixtures which, in contact withwater, emit flammable gases2.12.4.1 Pictograms, signal words, hazard statements andprecautionary statements for substances and mixturesAnnex I: 2.12.3. Table 2.12.2Label elements for substances or mixtures which in contact with water emit flammable gasesClassification Category 1 Category 2 Category 3GHS PictogramsSignal word Danger Danger WarningHazard statement H260:PrecautionaryStatement PreventionPrecautionaryStatement ResponsePrecautionaryStatement StoragePrecautionaryStatement DisposalIn contact with waterreleases flammablegases which may ignitespontaneouslyP223P231 + P232P280P335 + P334P370 + P378H261:In contact with waterreleases flammablegasesP223P231 + P232P280335 + P334P370 + P378H261:In contact with waterreleases flammablegasesP231 + P232P280P370 + P378P402 + P404 P402 + P404 P402 + P404P501 P501 P501The wording of the Precautionary Statements is found in CLP Annex IV, Part 2.2.12.4.2 Additional labelling provisionsAnnex II of CLP provides the following additional labelling provisions for water-reactivesubstances. These statements shall be assigned in accordance with CLP, Article 25 (1), tosubstances and mixtures classified for physical, health or environmental hazards. There are nocriteria or test methods provided for these EUH statements.Annex II: 1.1.3. EUH014 – 'Reacts violently with water'For substances and mixtures which react violently with water, such as acetyl chloride, alkali metals,titanium tetrachloride.Annex II: 1.2.1. EUH029 - 'Contact with water liberates toxic gas'For substances and mixtures which in contact with water or damp air, evolve gases classified foracute toxicity in category 1, 2 or 3 in potentially dangerous amounts, such as aluminium phosphide,161

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012phosphorus pentasulphide.2.12.5 Re-classification of substances and mixtures which, in contact with water,emit flammable gases according to DSD or already classified for transport2.12.5.1 Re-classification of substances and mixtures classified inaccordance with DSD2.12.5.1.1 Differences in classification and labellingThe differences between the classification principles of CLP and DSD are relevant and adirect translation is not possible in all cases.All substances and mixtures with F, R15 are classified as substances and mixtures which, incontact with water, emit flammable gases under CLP. The assignment of the correct categorycan be done in accordance with the transport classification on the basis of the UN N.5 testresults.Substances and mixtures with F; R15 where spontaneous ignition was observed in any step ofthe test procedure EC A.12 are classified in Category 1. In all other cases a re-evaluation ofthe EC A.12 test report data is not giving all relevant information (due to the missing value ofthe gas evolution rate of the minute intervals) and the assignment of the correct Category 1, 2or 3 can be done only on the basis of the UN N.5 test results (see section 2.12.3.4.2).Attention! According to DSD in case of pyrophoric substances and mixtures (F; R17) the testEC A.12 was not to be performed (see instruction of test method A.12 as described in inCouncil Regulation (EC) No 440/2008) and no additional classification with R15 wasrequired. On the other hand, the CLP (and GHS) stipulate an additional classification as asubstance and mixture which, in contact with water, emit flammable gases, even forpyrophoric substances or mixtures (F; R17). In case of pyrophoric substances and mixturesthe UN N.5 test shall be executed under nitrogen atmosphere (see Table 2.12.5.1.2).Therefore, for pyrophoric substances or mixtures a direct translation with respect to theirreaction with water is not possible.2.12.5.1.2 Differences in the test proceduresThere are relevant methodological differences between the test method EC A.12 of theCouncil Regulation (EC) No 440/2008 and the UN Test N.5 as described in Part III, subsection33.4.1 of the UN-MTC. The main differences are listed in the following table.Table 2.12.5.1.2 Differences between the method EC A.12 and UN N.5Parameter EC A.12 CLP and UN N.5Testing of pyrophoricsubstances and mixturesGas evolution rateintervalnot required yes, required under nitrogenatmosphere1-hour interval:1-minute-interval:yes, requirednot required1-hour interval: yes, required1-minute-interval: yes, required(with respect to Category 1)Amount of sample 10 g …enough (up to a maximum massof 25 g) to produce between 100 mland 200 ml of gasAmount of water 10 to 20 ml no instructionDivision into categories no yes, Category 1, 2, or 3162

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012If no spontaneous ignition of the evolved gas was observed in step 1 to 3 of the procedure,then the gas evolution rate must be determined. In contrast to test UN N.5 the method ECA.12 does not require to determine the gas evolution rate at the 1-minute interval. Thus, aclassification based on an A.12 test report is not possible if the gas evolution rate is greaterthan 1 litre per kilogram of substance per hour. In this case the assignment of the correctcategory shall be done in accordance with the transport classification on the basis of the UNN.5 test results.For substances and mixtures with F; R15 a re-evaluation of the test will lead to a classificationin Category 1 if a spontaneous ignition was observed in any step of the test procedure ECA.12. In all other cases a re-evaluation of the EC A.12 test report data is not giving allrelevant information (due to the missing value of the gas evolution rate of the minuteintervals) and the assignment of the correct Category 1, 2 or 3 can be done only on the basisof the UN N.5 test results (see chapter 4.1).Attention! Special care is required in those cases where the gas evolution rate depends on therelative amounts of sample and water. However, the requirements for the sample and wateramounts are different in the test methods EC A.12 and UN N.5. Thus, significant differencesbetween the test results of different methods and/or amounts may occur.For these reasons, the person responsible for classification shall have sufficient experience inthe differences of both test methods.In addition, it has to be mentioned that the lower criteria of both methods are different:According to CLP (and GHS) the criterion for the gas evolution rate is:"equal to or greater than 1 litre per kilogram of substance per hour".According to EC test method A.12 and the UN Test Manual it is "greater than 1 litre perkilogram of substance per hour"2.12.5.2 Relation to transport classificationSubstances which are classified as class 4.3 for transport or are labelled with 4.3 in addition toclass 4.2, class 8 or class 6.1 are classified as substances and mixtures which, in contact withwater, emit flammable gases under CLP.2.12.6 Examples of classification for substances and mixtures which, in contact withwater, emit flammable gases2.12.6.1 Example of a substance fulfilling the classification criteriaSubstances and mixtures which, in contact with water, emit flammable gases may belong tomany different classes of substances, for example, alkali metals, alkyl aluminium derivatives,alkyl metals, metal hydrides, metal phosphides, certain metal powders. A comprehensive listcan be found in Bretherick’s Handbook of Reactive Chemical Hazards (Urben P (editor 2007)and Urben P, 1999).Example: Pyrophoric substance fulfilling the criteria for CLP classificationSubstance: Magnesium alkyls (Index No. 012-001-00-4)Chemical structure:Flammable gas:Gas evolution rate:Spontaneous ignition:R 2 MgHydrogennot applicablenot possible due to the nitrogen atmosphere during the UN N.5 test163

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012EU classification: F; R14-17Transport classification: -Reference:Former Annex I to DSD and Annex VI to CLP CLP Classification: H260 Water-react. 1H250 Pyr. Sol. 1EUH0142.12.6.2 Example of a substance not fulfilling the classificationcriteriaExample: Manganese ethylene bis (dithiocarbamate) complex with zinc salt 88% (Mancozeb)Gas evolution rate:Spontaneous ignition:0 litre per kilogram of substance per hour.not applicableTransport classification: not class 4.3Reference: Method UN N.5, Table 33.4.1.4.5, United Nations (2003) CLP Classification:2.12.7 Referencesnot classified as substances and mixtures which, in contact with water,emit flammable gasesChemFinder (database): http://chemfinder.cambridgesoft.comCHEMSAFE (database): http://www.dechema.de/en/chemsafe.htmlCRC (2005) CRC Handbook of Chemistry and Physics 86 th Edition. Editor in Chief, D. Lide.CRC Press, Taylor and Francis, Boca Raton, FLGESTIS-database on hazardous substances:http://www.dguv.de/bgia/en/gestis/stoffdb/index.jspMerck (2001) Merck Index 13th Edition. Edited by S Budavari et al. Merck & Co, Inc, USAUrben P (editor 2007) Bretherick's Handbook of Reactive Chemical Hazards, Volumes 1-2(7th Edition). ElsevierUrben P, Bretherick L (1999) Bretherick’s Handbook of Chemical Reactive Hazards,Volumes 1-2 (6th Edition). Butterworth Heinemann, London2.13 OXIDISING LIQUIDS AND OXIDISING SOLIDS2.13.1 IntroductionThe hazard classes “oxidising liquids” (CLP Annex I, 2.13) and “oxidising solids” (CLPAnnex I, 2.14) comprise substances and mixtures whose hazard is characterised by the factthat, in contact with other materials, they are able to cause or contribute to the combustion ofthose materials. The other materials do not necessarily have to belong to a certain hazard classin order to be able to be affected by the presence of oxidising materials. For example, when164

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012coming into contact with an oxidising material, a solid that is not classified into the hazardcategory “flammable solids” (CLP Annex I, 2.7) may upon ignition behave like a flammablesolid. This is for example the case when a solid material (e.g. wood) is soaked with anoxidising liquid or when a liquid fuel (e.g. gas oil) mixes with an oxidising solid. Certaincombinations of combustible materials and oxidising materials may even result inspontaneous combustion, thermal instability or form an explosive mixture, this means thatthey may have explosive properties or may be regarded as self-reactive substances.The oxidising properties of a solid depend on its particle size. Smaller particles enable a moreintimate contact between the solid oxidiser and a combustible solid. The smaller the particlesize, the higher the oxidising capability of the solid. As a consequence, it may happen thatlarge particles of a certain solid are considered to be non-hazardous, while small particles ofthe same solid need to be classified into the hazard class of oxidising solids.Although widely known as “oxidising materials”, their hazard and behaviour might be betterunderstood by considering them to be “fire enhancing substances”.The hazards communication of oxidising liquids and oxidising solids intends to communicatethe property that it may cause fire or explosion or that it may intensify fire.Apart from the combustion hazard, the production of toxic and/or irritating fumes may causean additional hazard. For example, when nitrates are involved in a fire, nitrous fumes may beformed.The classification procedure and criteria for oxidising substances is not applicable for organicperoxides. Under DSD organic peroxides were considered to be oxidising substances becauseof the presence of the –O–O– bond. The majority of the organic peroxides do not possessoxidising properties; their main hazards are reactivity and flammability. Under CLP organicperoxides are comprised in a separate hazard class (CLP Annex I, 2.15) and they must not beconsidered according to the procedures described for oxidising solids and oxidising liquids.The testing procedure and criteria for oxidising substances do not work properly forammonium nitrate, ammonium nitrate compounds, ammonium nitrate based fertilizers andammonium nitrate emulsions, suspensions or gels.Ammonium nitrate is not an oxidising substance, but by default it may be classified is anoxidising substance. The classification of ammonium nitrate or ammonium nitrate compoundsis based on their composition and not on test results according to test O.1 or O.2 of the UN-MTC, Part III, sections 34.4.1 and 34.4.2, respectively. The procedure for the classification ofammonium nitrate emulsions, suspensions or gels is comprised in test series 8 of the UN-MTC.For classification and labelling of materials containing ammonium nitrate, expert judgementshould be sought.2.13.2 Definitions and general considerations for the classification of oxidisingliquids and oxidising solidsThe CLP text comprises the following definitions for oxidising liquids and oxidising solids.Annex I: 2.13.1 & 2.14.1DefinitionsAn oxidising liquid or solid means a liquid or solid substance or mixture which, while in itself notnecessarily combustible, may, generally by yielding oxygen, cause, or contribute to, the combustionof other material.165

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.13.3 Classification of substances and mixtures as oxidising liquids and oxidisingsolids1662.13.3.1 Identification of hazard informationOxidising liquids and oxidising solids may cause, or contribute to, the combustion of othermaterial. Although the definition states that they generally do this by yielding oxygen,halogens can behave in a similar way. Therefore, any substance or mixture containing oxygenand/or halogen atoms should in principle be considered for inclusion into the hazardcategories oxidising liquids or oxidising solids. This does not necessarily mean that everysubstance or mixture containing oxygen and/or halogen atoms should be subjected to the fulltesting procedure. Possibilities to waive testing are outlined in the next paragraph as well asparagraph 2.3.2.13.3.1.1 Non-testing dataExperience in the handling and use of substances or mixtures which shows them to beoxidising is an important additional factor in considering classification as oxidising solid oroxidising liquid. In the event of divergence between test results and known experience,judgement based on known experience should take precedence over test results.Before submitting a substance or a mixture to the full test procedure, an evaluation of itschemical structure may be very useful as it may prevent unnecessary testing. The personapplying this procedure should have sufficient experience in testing and in theoreticalevaluation of hazardous substances. The following text provides a guideline for the theoreticalevaluation of potential oxidising properties on basis of its composition and chemical structure.In case of doubt, the full test shall be performed.For organic substances or mixtures the classification procedure for these hazard classes neednot to be applied if:(a) The substance or mixture does not contain oxygen, fluorine or chlorine; or(b) The substance or mixture contains oxygen, fluorine or chlorine and these elementsare chemically bonded only to carbon or hydrogen.For inorganic substances or mixtures, the classification procedure for these hazard classesneed not be applied if they do not contain oxygen or halogen.On basis of this theoretical evaluation only a distinction can be made between “potentiallyoxidising” (i.e. further testing required) and “non-oxidising” (i.e. no further testing for thishazard category required). It is not possible to assign a hazard category on basis of atheoretical evaluation.Any substance or mixture that complies with the above evaluation criteria can be safelyregarded to have no oxidising properties and, hence, needs not to be tested and needs not to beregarded as an oxidising liquid or an oxidising solid. However, such a substance or mixturemay still possess other hazardous properties that require classification into another hazardclass.In case a mixture of an oxidising material and a non-hazardous inert material is offered forclassification, the following should be taken into account.An inert material by definition does not contribute to the oxidising capability of the oxidisingmaterial. Hence, the mixture can never be classified into a more severe hazard category.If an oxidising material is mixed with an inert material, the oxidising capability of the mixturedoes not linearly decrease with decreasing content of oxidising substance. The relationship is

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012more or less logarithmic and depends on the characteristics of the oxidising material. Forinstance, a mixture containing 50% of a strong oxidiser and 50% of an inert material mayretain 90% of the oxidising capability of the original oxidising component. Non-testingclassification of mixtures based solely on test data for the original oxidising substance shouldtherefore be done with extreme care and only, if sufficient experience in testing existsThe determination of the oxidising properties of an aqueous solution of solid oxidisingsubstances and the classification as an oxidising preparation is not necessary provided that thetotal concentration of all solid oxidisers in the aqueous solution is less then or equal to20%(w/w).2.13.3.2 Classification criteria2.13.3.2.1 GeneralThe testing procedures for oxidising liquids and oxidising solids are based on the capability ofan oxidising material to enhance the combustion of a combustible material. Therefore,oxidising solids and oxidising liquids that are submitted for classification testing are mixedwith a combustible material. In principle, dried fibrous cellulose is used as a combustiblematerial. The mixture of the potentially oxidising material and cellulose is then ignited and itsbehaviour is observed and compared to the behaviour of reference materials.For liquids the mixture with cellulose is ignited under confinement in an autoclave and thepressure rise rate that is caused by the ignition and the subsequent reaction is recorded. Thepressure rise rate is compared to that of three reference materials. The higher the pressure riserate, the stronger the oxidising capability of the liquid tested.For solids the mixture with cellulose is ignited at atmospheric conditions and the timenecessary for the combustion reaction to consume the mixture is recorded. The faster thecombustion rate, the stronger the oxidising capability of the solid tested.The classification criteria as included in CLP are copied in sections 2.2.2 and 2.2.3.2.13.3.2.2 Oxidising liquidsAnnex I: 2.13.2.1. An oxidising liquid shall be classified in one of the three categories for this classusing test O.2 in Part III, sub-section 34.4.2 of the UN Recommendations on the Transport of DangerousGoods, Manual of Tests and Criteria in accordance with Table 2.13.1:Category123Table 2.13.1Criteria for oxidising liquidsCriteriaAny substance or mixture which, in the 1:1 mixture, by mass, of substance (or mixture)and cellulose tested, spontaneously ignites; or the mean pressure rise time of a 1:1 mixture,by mass, of substance (or mixture) and cellulose is less than that of a 1:1 mixture, by mass,of 50% perchloric acid and cellulose.Any substance or mixture which, in the 1:1 mixture, by mass, of substance (or mixture)and cellulose tested, exhibits a mean pressure rise time less than or equal to the meanpressure rise time of a 1:1 mixture, by mass, of 40% aqueous sodium chlorate solution andcellulose; and the criteria for Category 1 are not met.Any substance or mixture which, in the 1:1 mixture, by mass, of substance (or mixture)and cellulose tested, exhibits a mean pressure rise time less than or equal to the meanpressure rise time of a 1:1 mixture, by mass, of 65% aqueous nitric acid and cellulose; and167

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012168the criteria for Category 1 and 2 are not met.2.13.3.2.3 Oxidising solidsAnnex I: 2.14.2.1. An oxidising solid shall be classified in one of the three categories for this class usingtest O.1 in Part III, sub section 34.4.1 of the UN Recommendations on the Transport of DangerousGoods, Manual of Tests and Criteria, in accordance with Table 2.14.1:CategoryTable 2.14.1Criteria for oxidising solidsCriteria1 Any substance or mixture which, in the 4:1 or 1:1 sample-to-cellulose ratio (by mass)tested, exhibits a mean burning time less than the mean burning time of a 3:2 mixture, bymass, of potassium bromate and cellulose.2 Any substance or mixture which, in the 4:1 or 1:1 sample-to-cellulose ratio (by mass)tested, exhibits a mean burning time equal to or less than the mean burning time of a 2:3mixture (by mass) of potassium bromate and the criteria for Category 1 are not met.3 Any substance or mixture which, in the 4:1 or 1:1 sample-to-cellulose ratio (by mass)tested, exhibits a mean burning time equal to are less than the mean burning time of a 3:7mixture (by mass) of potassium bromate and cellulose and the criteria for Categories 1 and2 are not met.Note 1:Some oxidising solids also present explosion hazards under certain conditions (when stored in largequantities). Some types of ammonium nitrate may give rise to an explosion hazard under extremeconditions and the 'Resistance to detonation test' (BC Code, Annex 3, Test 5) can be used to assess thishazard. Appropriate information shall be made available in the SDS.Note 2:The test shall be performed on the substance or mixture in its physical form as presented. If for example,for the purposes of supply or transport, the same chemical is to be presented in a physical form differentfrom that which was tested and which is considered likely to materially alter its performance in aclassification test, the substance shall also be tested in the new form.Note 1 may also apply to other oxidising ammonium salts. Experience indicates that theconditions required for ammonium nitrate to present an explosion hazard involve acombination of factors: storage in large volumes (multiple tonnes) and either contamination ofthe material (e.g. with metals, acids, organics) or excessive heat (e.g. under conditions offire). The resistance to detonation (RTD) test is extensively described in Regulatioin2003/2003/EC for ammonium nitrate.Testing and evaluation of hazard information see Section 2.13.3.3 regarding the application ofnon-testing data. See Urben, 2007 for additional information regarding the use of non-testingdata.Testing can be waived in the following cases where the tests are not applicable: gases,explosive or highly flammable substances, organic peroxides.2.13.3.3 Testing and evaluation of hazard informationThe test methods for oxidising liquids and oxidising solids are designed to give a finaldecision regarding their classification. Apart from testing, also experience in the handling anduse of substances or mixtures which shows them to be oxidising is an important additionalfactor in considering classification in these hazard classes. In the event of divergence betweentest results and known experience, judgement based on known experience should take

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012precedence over test results. However, on basis of experience only a substance or mixtureshall never be classified into a category of a less severe hazard.If the evaluation according to the appropriate criteria shows that the classification criteria arefulfilled, one or more hazard categories and the corresponding hazard statements shall beassigned (see Section 2.13.3.2).2.13.3.4 Decision logicClassification of oxidising liquids and oxidising solids is done according to decision logics2.13 and 2.14 as included in the GHS.NOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.169

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.13.3.4.1 Decision logic 2.13 for oxidising liquidsThe substance/mixtures is a liquidDoes it, in the 1:1 mixture, by mass, of substance (ormixture) and cellulose tested, exhibits a pressure rise ≥2070kPa gauge?NONot classifiedYESDoes it, in the 1:1 mixture, by mass, of substance (ormixture) and cellulose tested, exhibit a mean pressurerise time less than or equal to the mean pressure rise timeof a 1:1 mixture, by mass, of 65% aqueous nitric acid andcellulose?NONot classifiedYESDoes it, in the 1:1 mixture, by mass, of substance (ormixture) and cellulose tested, exhibit a mean pressurerise time less than or equal to the mean pressure rise timeof a 1:1 mixture, by mass, of 40% aqueous sodiumchlorate and cellulose?NOCategory 3WarningYESDoes it, in the 1:1 mixture, by mass, of substance (ormixture) and cellulose tested, spontaneously ignite orexhibit a mean pressure rise time less than that of a 1:1mixture, by mass, of 50% perchloric acid and cellulose?NOCategory 2WarningYESCategory 1Danger170

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.13.3.4.2 Decision logic 2.14 for oxidising solidsThe substance/mixtures is a solidDoes it, in the 4:1 or 1:1 sample-to-cellulose, by mass,tested ignite or burn?NONot classifiedYESDoes it, in the 4:1 or 1:1 sample-to-cellulose, by mass,tested exhibit a mean burning time less than or equeal tothe mean burning time of a 3:7 mixture, by mass, bypotassium bromate and cullulose?NONot classifiedYESDoes it, in the 4:1 or 1:1 sample-to-cellulose, by mass,tested exhibit a mean burning time less than or equeal tothe mean burning time of a 2:3 mixture, by mass, bypotassium bromate and cullulose?NOCategory 3WarningYESDoes it, in the 4:1 or 1:1 sample-to-cellulose, by mass,tested exhibit a mean burning time less than the meanburning time of a 3:2 mixture, by mass, by potassiumbromate and cullulose?NOCategory 2WarningYES Category 1Danger171

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Hazard communication for oxidising liquids and oxidising solids2.13.3.5 Pictograms, signal words, hazard statements andprecautionary statementsThe symbols and hazard statements are designed to indicate that oxidising materials maycause or contribute to fire or explosion and therefore in principle should be separated fromcombustible materials.According to CLP, the same label elements must be used for liquid and solid oxidisingsubstances and mixtures (Tables 2.13.2 and 2.14.2 of CLP are equal).SymbolAnnex I, 2.13.3. & 2.14.3. Tables 2.13.2 (liquids) & 2.14.2 (solids)Label elements for oxidising liquids and solidsCategory 1 Category 2 Category 3Signal word Danger Danger WarningHazard statementPrecautionaryStatement PreventionPrecautionaryStatement ResponsePrecautionaryStatement StorageH271: May cause fireor explosion; strongoxidiserP210P220P221P280P283P306 + P360P371 + P380 + P375P370 + P378H272: May intensifyfire; oxidiserP210P220P221P280P370 + P378The wording of the Precautionary Statements is found in CLP Annex IV, Part 2.H272: May intensifyfire; oxidiserP210P220P221P280P370 + P3782.13.4 Re-classification of substances and mixtures classified as oxidising liquidsand oxidising solids according to DSD or already classified for transport2.13.4.1 Re-classification of substances and mixtures classified inaccordance with DSD2.13.4.1.1 LiquidsSubstances that have been classified as oxidising liquids according to DSD can be reclassifiedto the GHS classification according to CLP. The test method that was used underDSD was included as Test Guideline A.21 in Council Regulation (EC) No 440/2008. Theprinciple and criteria of this method are equivalent to those of the GHS method. However,previously it was only used to indicate whether or not the material had oxidising properties.Classification into hazard categories was possible but not necessary.172

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Since the cut-off limit is equivalent, substances that were classified as oxidising liquidsaccording to DSD do also meet the CLP criteria. If no hazard category was assignedpreviously, re-testing will be necessary.2.13.4.1.2 SolidsSubstances that have been classified as oxidising solids according to DSD cannot be reclassifiedstraightforward to the CLP classification. The test method that was used under DSDwas included as Test Guideline A.17 of Council Regulation (EC) No 440/2008 for oxidisingsolids. Although the principle of this method was quite similar to that of the GHS method, theset-up and the criteria of the test method were very different. Moreover, the previous methodwas only designed to indicate whether or not the material had oxidising properties. Noclassification into hazard categories was possible.In general, substances that were classified as oxidising solids according to DSD will also meetthe criteria for GHS classification according to CLP. If no hazard category was assignedpreviously, re-testing will be necessary.2.13.4.2 Relation to transport classificationSubstances or mixtures which are classified as class 5.1 for transport are classified asoxidising liquids or solids under CLP.2.13.5 Examples of classification for oxidising liquids and oxidising solids2.13.5.1 Examples of substances and mixtures fulfilling theclassification criteriaThe list of substances and mixtures fulfilling the criteria for classification is only presentedfor information purposes. This list is not exhaustive.2.13.5.1.1 LiquidsFerric nitrate, saturated aqueous solutionLithium perchlorate, saturated aqueous solutionMagnesium perchlorate, saturated aqueous solutionPerchloric acid, 55%Sodium nitrate, 45% aqueous solution2.13.5.1.2 SolidsCalcium nitrate, anhydrousChromium trioxidePotassium nitritePotassium perchloratePotassium permanganateSodium chlorateSodium nitriteSodium nitrateStrontium nitrate, anhydrous173

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.13.5.2 Examples of substances and mixtures not fulfilling theclassification criteria for2.13.5.2.1 LiquidsNickel nitrate, saturated aqueous solutionPotassium nitrate, 30% aqueous solutionSilver nitrate, saturated aqueous solution2.13.5.2.2 SolidsCalcium nitrate, tetrahydrateCobalt nitrate, hexahydrate2.13.6 ReferenceUrben, P.G., Bretherick’s Handbook of Reactive Chemical Hazards, Seventh Edition, 2007,Academic Press, Elsevier, Amsterdam.2.14 ORGANIC PEROXIDES2.14.1 IntroductionThe hazard class “Organic Peroxides” is unique in the respect that it is the only category towhich chemicals are assigned on the basis of their chemical structure. Organic peroxidescannot be seen as an “intrinsic property”; it is a family of chemical substances which mayhave various properties. However, the type of peroxide is determined by testing.2.14.2 Definitions and general considerations for the classification of organicperoxidesAnnex I: 2.15.1.DefinitionOrganic peroxides means liquid or solid organic substances which contain the bivalent –O-O-structureand may be considered derivatives of hydrogen peroxide, where one or both of the hydrogen atomshave been replaced by organic radicals. The term organic peroxide includes organic peroxide mixtures(formulations) containing at least one organic peroxide. Organic peroxides are thermally unstablesubstances or mixtures, which can undergo exothermic self-accelerating decomposition. In addition,they can have one or more of the following properties:(i) be liable to explosive decomposition;(ii) burn rapidly;(iii) be sensitive to impact or friction;(iv) react dangerously with other substances.2.15.1.2. An organic peroxide is regarded as possessing explosive properties when in laboratorytesting the mixture (formulation) is liable to detonate, to deflagrate rapidly or to show a violent effectwhen heated under confinement.174

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In CLP, the following definition is given for organic peroxides.2.14.3 Relation to other physical hazardsIn addition to the definition (CLP Annex I, 2.15.1), organic peroxides may:(v) Be flammable.(vi) Emit flammable gas when heated,In general, organic peroxides do not have or have only weak oxidising properties.The additional (subsidiary) labelling, as indicated in the list of classified organic peroxidesincluded in the RTDG section 2.5.3.2.4, represents the additional hazardous properties.Neither the burning properties nor the sensitivity to impact and friction form part of theclassification procedure for organic peroxides in CLP. These properties may be of importancefor the safe handling of organic peroxides (see Section 2.14.4.3.2, additional testing).In addition, the hazard statement for flammable properties for liquid organic peroxides shouldbe based on the appropriate category for flammable liquids, as long as the flashpoint isrelevant, (see Section 2.14.4.3.2). The translation table in Annex VII to CLP can be used forthis.2.14.4 Classification of substances and mixtures as organic peroxides2.14.4.1 Identification of hazard informationThe classification of an organic peroxide in one of the seven categories “Types A to G” isdependent on its detonation, thermal explosion and deflagrating properties, its response toheating, the concentration and the type of diluent added to desensitize the substance.Specifications of acceptable diluents that can be used safely are given in the UNRecommendations on the Transport of Dangerous Goods, 2.5.3.5. The classification of anorganic peroxide as Type A, B or C is dependent on the type of packaging in which thesubstance is tested as it affects the degree of confinement to which the substance is subjected.This has to be considered when handling the substance; stronger packaging may result inmore violent reactions when the substance decomposes. This is why it is important thatstorage and transport is done in packaging, allowed for the type of organic peroxide, thatconforms the requirements of the UN-packaging or IBC instruction (P520/IBC520) or tankinstruction (T23).The traditional aspects of explosive properties, such as detonation, deflagration and thermalexplosion, are incorporated in the decision logic Figure 2.15.1 of CLP. Consequently,explosive property determination as prescribed for the hazard class ‘explosives’ needs not tobe conducted for organic peroxides.A list of currently classified organic peroxides is included in the RTDG section 2.5.3.2.4.2.14.4.2 Classification criteriaIn CLP, organic peroxides are not classified as oxidisers but they are a distinct hazard class.Annex I: 2.15.2.1. Any organic peroxide shall be considered for classification in this class, unless itcontains:a) not more than 1,0 % available oxygen from the organic peroxides when containing not more than175

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121,0 % hydrogen peroxide; orb) not more than 0,5% available oxygen from the organic peroxides when containing more than 1,0 %but not more than 7,0 % hydrogen peroxide.Determination of the explosive properties is incorporated in the classification decision logic.Flammability is not incorporated into the decision flow chart (see Section 2.14.4.4).In CLP decision logic Annex I, Figure 2.15.1, classification of organic peroxides is based onperformance based testing both small scale tests and, where necessary, some larger scale testwith the substance in its packaging. The concept of “intrinsic properties” is, therefore, notapplicable to this hazard class.Organic peroxides are classified into one of the seven categories of “Types A to G” accordingto the classification criteria of CLP. The classification principles are given in decision logicFigure 2.15.1 of CLP and the test series A to H, as described in the Part II of the UN-MTC,should be performed.Annex I: 2.15.2.2. Organic peroxides shall be classified in one of the seven categories of ‘Types Ato G’ for this class, according to the following principles:(a) any organic peroxide which, as packaged, can detonate or deflagrate rapidly shall bedefined as organic peroxide TYPE A;(b) any organic peroxide possessing explosive properties and which, as packaged, neitherdetonates nor deflagrates rapidly, but is liable to undergo a thermal explosion in thatpackage shall be defined as organic peroxide TYPE B;(c) any organic peroxide possessing explosive properties when the substance or mixture aspackaged cannot detonate or deflagrate rapidly or undergo a thermal explosion shall bedefined as organic peroxide TYPE C;(d) any organic peroxide which in laboratory testing:(i) detonates partially, does not deflagrate rapidly and shows no violent effect whenheated under confinement; or(ii) does not detonate at all, deflagrates slowly and shows no violent effect whenheated under confinement; or(iii) does not detonate or deflagrate at all and shows a medium effect when heatedunder confinement;shall be defined as organic peroxide TYPE D;(e) any organic peroxide which, in laboratory testing, neither detonates nor deflagrates at alland shows low or no effect when heated under confinement shall be defined as organicperoxide TYPE E;(f) any organic peroxide which, in laboratory testing, neither detonates in the cavitated statenor deflagrates at all and shows only a low or no effect when heated under confinement aswell as low or no explosive power shall be defined as organic peroxide TYPE F;(g) any organic peroxide which, in laboratory testing, neither detonates in the cavitated statenor deflagrates at all and shows no effect when heated under confinement nor any explosivepower, provided that it is thermally stable, i.e. the SADT is 60 o C or higher for a 50 kgpackage 44 , and, for liquid mixtures, a diluent having a boiling point of not less than 150 o C isused for desensitisation, shall be defined as organic peroxide TYPE G. If the organicperoxide is not thermally stable or a diluent having a boiling point less than 150 o C is usedfor desensitisation, the organic peroxide shall be defined as organic peroxide TYPE F.Where the test is conducted in the package form and the packaging is changed, a further test shall beconducted where it is considered that the change in packaging will affect the outcome of the test.44See UN Recommendations on the Transport of Dangerous Goods, Manual of Tests and Criteria, subsections28.1, 28.2, 28.3 and Table 28.3.176

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012A list of currently classified organic peroxides is included in the UN RTDG, Section2.5.3.2.4.2.14.4.3 Testing and evaluation of hazard information2.14.4.3.1 Thermal stability tests and temperature controlIn addition to the classification tests given in decision logic Figure 2.15.1 of CLP, the thermalstability of the organic peroxide has to be assessed in order to determine the Self-AcceleratingDecomposition Temperature (SADT). For the determination of the SADT, the testing methodin UN-MTC, Part II, section 28, may be used.The SADT is defined as the lowest temperature at which self-accelerating decomposition mayoccur with a substance in the packaging as used in transport, handling and storage. The SADTis a measure of the combined effect of the ambient temperature, decomposition kinetics,package size and the heat transfer properties of the substance and its packaging.There is no relation between the SADT of an organic peroxide and its classification in one ofthe seven categories “Types A to G”. The SADT is used to derive safe handling, storage andtransport temperatures (control temperature) and alarm temperature (emergency temperature).Depending on its SADT an organic peroxide needs temperature control and the rules as givenin CLP Annex I, 2.15.2.3, are the following two elements:1) Criteria for temperature controlThe following organic peroxides need to be subjected to temperature control:(a) Organic peroxide types B and C with a SADT ≤ 50° C;(b) Organic peroxide type D showing a medium effect when heated underconfinement with a SADT ≤ 50° C or showing a low or no effect when heatedunder confinement with a SADT ≤ 45° C; and(c) Organic peroxide types E and F with a SADT ≤ 45° C.2) Derivation of control and emergency temperatures:Type of receptacle SADT a) Control temperature Emergency temperatureSingle packagings andIBC’s20 °C or lessover 20 °C to 35 °Cover 35 °C20 °C below SADT15 °C below SADT10 °C below SADT10 °C below SADT10 °C below SADT5 °C below SADTTanks < 50 °C 10 °C below SADT 5 °C below SADTa) i.e. the SADT of the substance as packaged for transport, handling and storage.It should be emphasized that the SADT is dependent on the nature of the organic peroxideitself, together with the volume and heat-loss characteristics of the packaging or vessel inwhich the substance is handled. The temperature at which self-accelerating decompositionoccurs falls: as the size of the packaging or vessel increases; and with increasing efficiency of the insulation on the package or vessel.The SADT is only valid for the substance as tested and when handled properly. Mixing theorganic peroxide with other chemicals, or contact with incompatible materials (includingincompatible packaging or vessel material) may reduce the thermal stability due to catalytic177

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012decomposition, and lower the SADT. This may increase the risk of decomposition and has tobe avoided.2.14.4.3.2 Additional testingThe sensitivity of organic peroxides to impact (solids and liquids) and friction (solids only)may be of importance for the safe handling of the substances, in the event that thesesubstances have pronounced explosive properties (e.g. rapid deflagration and/or violentheating under confinement). Test methods to determine these properties are described in testseries 3 of the UN-MTC. This information should be part of the hazard communication insafety data sheets.In national storage guidelines burning rate is commonly used for classification andconsequential storage requirements. Test methods are incorporated in these national storageregulations.The flashpoint for liquid organic peroxides is only relevant in the temperature range where theproduct is thermally stable. Above the SADT of the product flashpoint determination is notrelevant because decomposition products are evolved.Note: In case a flashpoint determination seams reasonable (expected flashpoint below theSADT) a test method using small amount of sample is recommended. In case the organicperoxide is diluted or dissolved, the diluent may determine the flashpointThe determination of the auto ignition temperature is not relevant for organic peroxides,because the vapours decompose during the execution of the test. Available test methods arefor non-decomposing vapour phases. Auto ignition of organic peroxide vapours when theydecompose, can never be excluded. This information should be part of the hazardcommunication in safety data sheets.Also self-ignition temperature determination (test applicable for solids) is not relevant. Thethermal stability of organic peroxides is quantitatively given by the SADT test.2.14.4.3.3 Additional classification considerationsCurrently the following properties are not incorporated in CLP: mechanical sensitivity i.e. impact and friction sensitivity (for handling purposes); burning tests (for storage purposes); and flammability aspects (definition of label and relevance of flashpoint).Furthermore:2.15.4.2. Mixtures of already classified organic peroxides may be classified as the same type oforganic peroxide as that of the most dangerous component. However, as two stable components canform a thermally less stable mixture, the SADT of the mixture shall be determined.Note: The sum of the individual parts can be more hazardous than the individual components.Formulated commercial organic peroxides are classified according to their SADT.2.14.4.4 Decision logicThe following decision logic for organic peroxides is applicable according to CLP.NOTE: The person responsible for classification should study the criteria for classificationbefore and during use of the decision logics.178

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: Figure 2.15.1Organic Peroxides179

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.14.5 Hazard communication for organic peroxides2.14.5.1 Pictograms, signal words, hazard statements andprecautionary statementsAccording to CLP the following label elements shall be used for substances or mixturesmeeting the criteria for this hazard class:Annex I: Table 2.15.1Label elements for organic peroxidesClassification Type A Type B Type C & D Type E & F Type GGHS pictogramsSignal words Danger Danger Danger WarningHazardStatementH240:Heating maycause anexplosionH241:Heating maycause a fireor explosionH242:Heating maycause a fireH242:Heating maycause a fireThere are nolabelelementsallocated tothis hazardcategoryPrecautionarystatementPrevention210220234P210P220P234P210P220P234P210P220P234280P280P280P280PrecautionarystatementResponsePrecautionarystatementStorageP411 + P235P410P420P410P420P410P420P410P420PrecautionarystatementP501 P501 P501 P501DisposalThe wording of the Precautionary Statements is found in CLP Annex IV, Part 2.Although CLP does not provide any precautionary statements for response, it is recommendedto consider the same statements as for the hazard class self-reactive substances.Precautionary statementP370 + P378P370 + P378P370 + P378P370 + P378ResponseP370 + P380P370 + P380+ P375+ P375180

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.14.5.2 Additional labelling provisions for organic peroxidesAdditional hazardous properties, resulting in additional (subsidiary) labelling, are indicated inthe list of classified organic peroxides included in the RTDG, section 2.5.3.2.4.2.14.6 Re-classification of substances and mixtures classified as organicperoxides according to DSD or already classified according to transport2.14.6.1 Re-classification of substances and mixtures classified inaccordance with DSDAccording to the DSD, organic peroxides are classified as oxidising substances or mixtures,on basis of their structure and composition. In CLP organic peroxides are NOT classified asoxidisers but they are a distinct hazard class. The classification procedure is described inSection 2.14.4. Under DSD, explosive properties and flammability were determinedseparately by the EU tests A14 (for explosive properties) and A9 (for flammable properties).Annex VII to CLP provides direct translations from DSD to CLP classifications. For organicperoxides a translation from symbol O and R-phrases, to a dedicated organic peroxide TypeA-F is not possible. For the correct assignment of an individual organic peroxide substance ormixture, to the relevant Type B-F, the tables of the UN RTDG, 2.5.3.2.4, IBC 520 and T23can be used. For Type A organic peroxides the classification principles of CLP should beapplied.2.14.6.2 Relation to transport classificationA list of currently classified organic peroxides is included in the UN RTDG, Section2.5.3.2.4. This table includes organic peroxides Type B-Type F (and some formulations TypeG, so-called exempted organic peroxides).An exceptional case in this respect is a peroxyacetic acid formulation, as currently classifiedin the RTDG under UN 3149, with the following description: HYDROGEN PEROXIDEAND PEROXYACETIC ACID MIXTURE with acid(s), water and not more than 5%peroxyacetic acid, STABILISED. In the classification procedure for organic peroxides, seedecision logic in Section 2.14.4.4, this formulation will be assigned to organic peroxide TypeG, and consequently no label elements are allocated. In view of the above, this formulationcan be classified, also in accordance with CLP, as an oxidising liquid, Category 2.2.14.7 Examples of classification for organic peroxides2.14.7.1 Examples of substances and mixtures fulfilling theclassification criteriaSubstance to be classified: BEMolecular formula: n.a.According to GHS 2.15.2.1, the substance has an active oxygen content of 7.40 % and thushas to be considered for classification in the hazard class organic peroxides.Test results and classification according to CLP decision logic 2.15.1 for organic peroxidesand the UN-MTC, Part II, is as follows:Classification test results1. Name of the organic peroxide : BE2. General data181

Guidance on the Application of the CLP Criteria Version 3.0 - November 20122.1. Composition : BE, technically pure (97%)2.2. Molecular formula: n.a.2.3. Active oxygen content: 7.18 %2.4. Physical form: liquid2.5. Colour: colourless2.6. Density (apparent): 900 kg/m 33. Detonation (test series A)Box 1 of the decision logic: Does the peroxide propagate a detonation?3.1. Method: UN Test A.1: BAM 50/60 steel tube test3.2. Sample conditions: peroxide assay 97 %3.3. Observations: fragmented part of the tube: 18 cm3.4. Result: No3.6. Exit: 1.34. Deflagration (test series C)Box 5 of the decision logic: Does the peroxide propagate a deflagration?4.1. Method 1: Time/pressure test (test C.1)4.1.1. Sample conditions: ambient temperature4.1.2. Observations: 4000 ms4.1.3. Result: Yes, slowly4.2. Method 2: Deflagration test (test C.2)4.2.1. Sample conditions: temperature: 25 °C4.2.2. Observations: deflagration rate: 0.74 mm/s4.2.3. Result: Yes, slowly4.3. Final result: Yes, slowly4.4. Exit: 5.25. Heating under confinement (test series E)Box 8 of the decision logic: What is the effect of heating it under defined confinement?5.1. Method 1: Koenen test (test E.1)5.1.1. Sample conditions: -5.1.2. Observations: limiting diameter: 2.0 mmfragmentation type "F"5.1.3. Result: Violent5.2. Method 2: Dutch pressure vessel test(test E.2)5.2.1. Sample conditions: -5.2.2. Observations: limiting diameter: 6.0 mm (with 10 g)5.2.3. Result: Medium182

Guidance on the Application of the CLP Criteria Version 3.0 - November 20125.3. Final result: Violent5.4. Exit: 8.16. Explosion test in package (test series G)Box 10 of the decision logic: What is the effect of heating it under defined confinement?6.1. Method: Thermal explosion test in package (test G.1)6.2. Sample conditions: 30 litre packaging,6.3. Observations: no fragmentation (N.F.)6.4. Result: No6.5. Exit: 10.27. Thermal stability (outside of the decision logic)7.1. Method: Heat accumulation storage test (test H.4)7.2. Sample conditions: mass 380 g. Half life time of cooling of Dewar vessel with400 ml DMP:10.0 hrs.(representing substance in package)7.3. Observations: self-accelerating decomposition at 35 °Cno self-accelerating decomposition at 30 °C7.4. Result: SADT 35 °C8. General remarks: The decision logic is given in figure 19. Final classificationHazard hazard class:organic peroxide, Type C, liquid, temperature controlledLabel:Flame over circleSignal word: DangerHazard statement:Heating may cause a fireTemperature control:Needed based on SADT (35 °C, in package)Control temperature*:20°C (in package)Emergency temperature* :25°C (in package)*see UN-TDG, manual of tests and criteria, table 28.22.14.8 Additional remarksControl and emergency temperatureThe Control and Emergency temperatures are based on the SADT as determined by UN testH.4. The Dewar vessel used in the UN H.4 test was representative for the substance handledin packages. For handling the substance in larger quantities (IBCs/tanks/vessels etc.) and/or in(thermally) insulated containers, the SADT has to be determined for that quantity with thatdegree of insulation. From that SADT the Control and Emergency temperatures can bederived (see also Section 2.14.4.3.1)Explosive propertiesThe explosive properties do not have to be determined according to CLP Annex I, 2.1 forexplosives, because this is incorporated in the decision logic, see also Section 2.14.4.4.183

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Substance may have explosive properties when handled under greater confinement than isafforded by the packaging in which it was tested in UN test G.1 (see 6 of classification testresults).The appropriate precautionary statement should be assigned to indicate the hazard of thermalexplosion under confined conditions.Because the substance shows propagation [marked [mass] explosive] properties in test seriesE (see, 5 of the classification test results) the sensitivity to impact and friction (friction onlyfor solids) are of importance for safe handling (see Section 2.14.4.3.2). Impact sensitivityaccording to UN test series 3, test 3 (a) (ii), BAM Fallhammer of the substances is 20 J. Theappropriate precautionary statement should be assigned to indicate the hazard of impactsensitivity.Burning propertiesTogether with the classification of the organic peroxide, the burning properties are ofimportance for storage classification (see Section 2.14.4.3.2). For example the burningproperties as determined by the test method described in the storage guidelines, currently inplace in France, Germany, Netherlands and Sweden, is 7.0 kg/min/m². Based on this figureand the classification as organic peroxide type C, the storage classification can be assigned inthose countries.FlashpointThe substance thermally decomposes before the temperature at which the vapour can beignited is reached (see Section 2.14.4.3.2).2.15 CORROSIVE TO METALS2.15.1 IntroductionThe hazard class “corrosive to metals” (CLP AnnexI, 2.16) is a physico-chemical propertythat is new in the EU classification scheme and appears for the first time in CLP. So far, onlythe health hazard “corrosivity to skin” was considered in the classification scheme. To someextent, both properties relate to each other and, in the context of transport of dangerous goods,have been considered for classification in class 8, despite the different nature of the hazard(material damage versus living tissue damage).A substance or a mixture that is corrosive to metal under normal conditions is a substance or amixture liable to undergo an irreversible electrochemical reaction with metals that leads tosignificant damage or, in some cases, even to full destruction of the metallic components. The“corrosive to metal” property is a quite complex property, since it is a substance (or mixture)related as well as a material (metal) related property. This means a corrosive substance ormixture leads to corroded material (metal), according to a number of external conditions.From the material side, many types of corrosion processes may occur, according toconfigurations, liquid or fluid media inducing the corrosion process, nature of metal, potentialpassivation occuring by oxide formation during corrosion.From the substance or mixture side, many parameters may influence the corrosion propertiesof a substance or mixture, such as the nature of the substance or the pH. From anelectochemistry point of view, corrosion conditions are often studied using Pourbaixdiagrams, which plot the electrochemical potential (in Volt) that develops according toelectrical charges transfer versus the pH-value. Such a diagram is shown for the case of ironand applies only for carbon steel corrosion (Jones, 1996).184

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 2.15.1 Potential pH (also called Pourbaix) diagram for iron in water at 25°C, indicatingstable form of the Fe element and implicitly, corrosion domainsFor the purposes of CLP, corrosion to metal will only be considered, by pure convention, forsubstances that are liable to attack carbon steel or aluminum, two of the most common metalsthat may come in contact with chemical substances (containment material, reactor material).The classification scheme applied here shall not be considered as a material (metal)classification method for metals regarding resistance to corrosion. By no means steel oraluminium specimens that are treated to resist to corrosion, shall be selected for testing.2.15.2 Definitions and general considerations for the classification of substances andmixtures corrosive to metalsCLP comprises the following definition for substances and mixtures that are corrosive tometal.Annex I: 2.16.1.DefinitionA substance or a mixture that is corrosive to metals means a substance or a mixture which by chemical actionwill materially damage, or even destroy, metals.2.15.3 Classification of substances and mixtures as corrosive to metals2.15.3.1 Identification of hazard informationImportance of the physical state of the test substance or mixtureThere is no reference in the definition (CLP Annex I, 2.16.1) to the physical state of thesubstances or mixtures that needs consideration for potential classification in this hazardclass. According to the test method to be employed for considering classification under thishazard class, we may state at least that gases are out of the scope of the “corrosive to metal”hazard class. The hazard related to the formation of corrosive gases is not currently coveredby the criteria for physical hazards in CLP. Corrosivity of gases (to metal) is assumed not torepresent an actual physico-chemical hazard, and is therefore not applicable here.According to the classification criteria (CLP Annex I, 2.16.2.1) only substances and mixturesfor which the application of the test “C.1”described in part III, section 37 of the UN-MTC (4 th185

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012revised edition) is relevant and needs to be considered. This means that non soluble solids areexcluded, while “liquids and solids that may become liquids (during transport)”, as mentionedin this reference text, have to be considered for such a classification.The wording “solids that may become liquids” was developed for TDGs classificationpurposes, and needs further explanation.Solids may become liquids by melting (due to increase in temperature). Solids having amelting point lower than 55°C (which is the test temperature required in test C.1) must thenbe taken into consideration. The other physical way to transform a solid into liquid is bydissolution in water or another solvent. Classification of solid substances that may becomeliquids by dissolution is subject to further expert judgement, and may need adaptation of theclassification criteria or test protocol (see Section 2.15.3.4.2). Interaction with liquids maycome from air moisture or unintentional contact with water. Other solvent traces may resultfrom the extraction process during manufacturing and these may induce corrosion in practice.Substances and mixtures in a liquid state must be tested without any modification beforetesting, by using the C.1 test protocol. For other cases (solids that may become liquids),appropriate testing procedures require further work by the Committees of experts in charge ofdeveloping and updating the GHS at UN level. It needs to be further specified how suchsubstances or mixtures shall be prepared (transformed into liquids) to be able to determinetheir corrosivity to metals. As an example, it is thought that the quantity of solvent (water orany other solvent) to liquefy the test substance before testing would greatly influence resultsof the C.1 test and may not necessarily represent the real life situation of a product duringtransport, handling or use.Non-testing dataFollowing parameters are helpful to evaluate corrosive properties before testing:- Melting points for solids,- Chemical nature of the substances and mixtures under evaluation (e.g. strong acids),- pH values (liquids),Literature may also provide information on widely used substances and liquids “compatibilitytables”, taking account of the corrosiveness of the products that may serve to decide whethertesting must be conducted before assigning the “corrosive to metals” hazard class, on basis ofexpert judgement.The following substances and mixtures should be considered for classification in this class: Substances and mixtures having acidic or basic functional groups; Substances or mixtures containing halogen; Substances able to form complexes with metals and mixtures containing such substances.1862.15.3.2 Screening procedures and waiving of testingExperience may have proven the corrosivity of given substances and mixtures. In such caseno more testing is needed (see examples in Section 2.15.6).Generally extreme pH-values point to a higher likelihood that the substance is corrosive.However, it can not lead to immediate classification in the hazard class "corrosive to metals".As a proof of that, Figure 2.15.1 shows that immunity zones (where steel does not corrode)still exist on the full spectrum of pH values as far as carbon steel is concerned.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Corrosivity is so complex that the evaluation of a mixture cannot be extrapolated from similarbehaviour of constituents of a mixture. However, if one significant component of a mixture iscorrosive to metals the mixture is likely to be corrosive to metals as well. Testing the actualmixture is therefore highly recommended. As already mentioned, solids are currently difficultto test according to the current CLP requirements, as the C.1 test has obviously been designedfor liquids.Where an initial test on either steel or aluminium indicates the substance or mixture beingtested is corrosive, the follow up test on the other metal is not required.2.15.3.3 Classification criteriaSubstances and mixtures of hazard class 'corrosive to metals' are classified in a single hazardcategory on the basis of the outcome of the UN Test C.1 (UN-MTC, part III, section 37,paragraph 37.4).CategoryAnnex I, 2.16.2. Table 2.16.1Criteria for substances and mixtures corrosive to metalsCriteria1 Corrosion rate on steel or aluminium surfaces exceeding 6.25 mm per year at a test temperature of55°C2.15.3.4 Testing and evaluation of hazard information2.15.3.4.1 General considerationsIt is important to point out that the criteria of corrosion rate will never be applied in anabsolute way, but by extrapolating the measured rate of corrosion over the test period to theannual assumed correlating corrosion rate. This exercise has to take account of the fact thatthe corrosion rate is not necessarily constant over time. Expert judgement may be required toconsolidate the optimum test duration and to ascertain test results. However, the possibility ofincreasing the testing period from minimum one week to four weeks as well as the use of twodifferent metals in the test protocol C.1 act as barriers against erroneous classification.Whatever the result of the classification may be, the classification as "corrosive to metals"relates to steel and/or aluminium only and does not provide information with regard to thecorrosivity potential to other metals than those tested.Two types of corrosion phenomena need to be distinguished for classification of substancesand mixtures in this hazard class, although not reported in CLP: the uniform corrosion attackand the localised corrosion (e.g. pitting corrosion, shallow pit corrosion).Table 1 (Section 37.4.1.4.1 of the UN Manual of test and criteria) translates the correspondingminimum mass loss rates leading to classify the test substance as corrosive to metals forstandard metal specimens (2 mm of thickness), according to time of exposure, for reasons ofuniform corrosion process. In case of use of metal plates of a thickness that differs from thespecified 2 mm (see comments in 2.4.2), the values in tables 1 and 2 need adjustments due tothe fact that the corrosion process depends on the surface of specimen.Table 2.15.3.4.1(a): Minimum mass loss of specimens after different exposure times(corresponding to the criterion of 6.25 mm/year)Exposure timeMass loss7 days 13.5%187

Guidance on the Application of the CLP Criteria Version 3.0 - November 201214 days 26.5%21 days 39.2%28 days 51.5%Table 2 (Section 37.4.1.4.2 of the UN-MTC) indicates the criteria leading to classification ofthe test substance as corrosive to metals for standard metal specimens, according to time ofexposure, for reasons of localised corrosion process.Table 2.15.3.4.1(a): Minimum intrusion depths after exposure times(corresponding to the the criterion of localized corrosion of 6.25 mm/year)Exposure time Min. intrusion depth7 days 120m14 days 240m21 days 360m28 days 480mIt is not mentioned explicitly in the text that localised corrosion as well as uniform corrosionhas also be taken into account. However, localised corrosion, that is entirely part of test C.1protocol, has actually to be taken into account. In addition, although the type of corrosion isnot reflected in the classification result, this valuable information should be given in the SDS.2.15.3.4.2 Additional notes on best practice for testingCompetence required for testingThe overall evaluation of appropriate data for considering the corrosion properties of asubstance or a mixture and in particular for testing it according to the mentioned criteria forthis hazard class, requires certain qualifications and experience. Expertise is often needed forthis hazard class, which relates to a complex and multi-faceted hazardous phenomenon.Selection of metal specimensCLP refers to two types of metals (carbon steel and aluminium) meeting accuratespecifications (technical characteristics of metal sheets and plate thickness). Thicker metalsheets, such as cast materials, of which the thickness is reduced by any form of mechanicaltreatment, may never be used. Mechanical reduction of sheet (metal) thickness could inducecorrosion enhanced process due to cross section heterogeneity in metal grain and impurities. Itis far better to use slightly different specifications of metal in the correct thickness or slightlydifferent specimen plate thicknesses. It is recognised that it will not always be easy to obtainmetal specimens with the profile as described above.Regarding the type of aluminium or steel to be used for this test see UN Manual of Tests andCriteria, sub-section 37.4.1.2.Minimum corrosive media volumeIn order to prevent any limitation on the corrosion process due to full consumption of thecorrosive media before the end of the testing period, a minimum volume of substance (1.5 L,according to the UN Manual) has to be used. (Note: volume/surface ratio of 10 mL/cm² isstated in DIN 50905, similar in ASTM G31–72.)Adjustment of the test temperatureCorrosion processes are temperature dependent. In the context of CLP, the property“corrosive to metals” is assessed through testing metal specimens at a specified temperatureof 55°C 1°C. In practice, it may be difficult with standard testing equipment to stay within188

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012the temperature window (55°C 1°C) of the gas phase, all over the test period. In such case,the test can be performed conservatively at a slightly higher temperature and somewhat loweraccuracy (e.g. 57°C 3°C).Selecting the appropriate test durationThe evaluation of the criterion of 6.25 mm/year is generally based on a test duration notexceeding 1 month. There is, however, the option to stop the test procedure already after 1week (see table 1). For the decision on test duration, the non linear behaviour of the corrosionprocess must be taken due account of. In borderline cases a non-appropriate test duration mayresult in either false positive or false negative results.Specimen cleaningAttention must be paid to the correct cleaning of the corroded residue before measurement ofthe corrosion characteristics. In case of adhesive corroded layer, the same cleaning processneeds to be carried out on a non corroded sample to verify if the cleaning procedure is notsignificantly abrasive. For further information see UN-MTC, sub-section 37.4.1.3.Testing soluble solidsAs said in Section 2.15.3.1, for solids that may become liquids through dissolution in water orin a solvent, the adequate testing procedure is more complex (not explicitly describe in theC.1 test protocol). In no case will simple dilution of the solid substance in any quantity ofwater lead to satisfactory testing of the substance for corrosion to metals. It is recommendedto perform the test with solutions containing extreme concentrations of the solid substance ormixture in water (very diluted e.g. 0.1% or saturated).For the specific case where the corrosion potential is linked to the presence of solvent traces(other than water), expert judgement is needed to determine if further testing must beperformed (where the solid is put in interaction with the metallic part considered).Example of equipment relevant for the performance test C.1189

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 2.15.3.4.2: Example of testing equipment available on the market to perform test C.12.15.4 Hazard communication for substances and mixtures corrosive to metals2.15.4.1 Pictograms, signal words, hazard statements andprecautionary statementsTable 2.16.2 of CLP Annex I provides the label elements for hazard class 'corrosive tometals'. The hazard statement H290, using the wording “may”, reflects that classificationunder this hazard class does not cover all metals (testing only considers carbon steel andaluminium). Thus we may find examples of substances that are classified in this hazard class‘corrosive to metals’ but will not induce corrosive action on other more corrosive resistantmetals than those serving as reference materials (e.g. platinium).Label elements shall be used for substances and mixtures meeting the criteria forclassification in this hazard class in accordance with Table 2.16.2.Annex I: 2.16.3. Table 2.16.2Label elements for substances and mixtures corrosive to metalsClassification Category 1GHS Pictogram190

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Signal WordHazard StatementPrecautionary Statement, PreventionPrecautionary Statement, ResponsePrecautionary Statement, StorageWarningH290: May be corrosive to metalsP234P390P406Precautionary Statement, DisposalThe wording of the Precautionary Statements is found in CLP Annex IV, Part 2.2.15.5 Re-classification of substances and mixtures classified as corrosive to metalsaccording to DSD2.15.5.1 Re-classification of substances and mixtures classified inaccordance with DSDThe hazard class ‘corrosive to metals’ was not included in the DSD. Therefore, reclassificationis not applicable.Only the substances and mixtures presenting the health related property “Skin corrosive” wereincluded (Symbol C with R-phrases 34 or 35). These substances generally present asignificant potential for the “corrosive to metals” property and should be considered fortesting.2.15.5.2 Relation to transport classificationValuable information can be obtained from TDG regulation. Transport class 8 coverssubstances that are classified for corrosivity to skin, metals or both. Existing test resultsobtained in the context of transport may be applied since the C.1 test serves as reference fortesting in both classification systems.2.15.6 Examples of classification for substances and mixtures corrosive to metalsThe following table lists some examples of substances and mixtures that should be classifiedor not in class 2.16 (according to known test C.1 results) in comparison with predicted resultsfor skin corrosion hazard.Table 2.15.6: Examples of classified and non classified substances and mixtures in class 2.16Note: “corroded” means corrosion attack in the sense of test C.1;“non corroded” means corrosion resistant in the sense of test C.1;“positive or negative” are results from skin corrosion.Substance or mixture Steel Aluminium CLP Annex I, 2.16classificationHydrofluoric acid> 70% (UN1790)Highly concentrated nitric acid(97%) (UN2031)NotcorrodedNotcorrodedSkin (for comparison)Corroded Classified PositiveCorroded Classified Positive191

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012HNO 3 red fuming (UN2032)Hydrochloric acid (diluted)(UN1789)NaOH solutions (UN1824)NotcorrodedNot corroded Not classified PositiveCorroded Corroded Classified NegativeNotcorrodedCorroded Classified Positive2.15.6.1 Example of metal specimen plates after exposure to acorrosive mixtureFigure 2.15.6.1: Example of corroded metal plates after testing according to C.1 test for a classifiedmixturePlate locatedin the liquidphasePlate locatedin theinterfacePlate located in thevapour phaseThis example shows that the corrosion may develop at different rates according to theaccurate position of the specimen related to the corroding mixture (sunk in the liquid, placedin the gas phase above liquid or at the liquid/gas interface).2.15.7 ReferencesASTM G31-72(2004) Standard Practice for Laboratory Immersion Corrosion Testing ofMetals.Jones, D.A., Principles and Prevention of Corrosion, 2nd edition, 1996, Prentice Hall, UpperSaddle River, NJ. ISBN 0-13-359993-0 Page 50-52.DIN 50905-1: 2007, Corrosion of metals - Corrosion testing - Part 1: General guidance(Korrosion der Metalle - Korrosionsuntersuchungen - Teil 1: Grundsätze).192

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123 HEALTH HAZARDS 453.1 ACUTE TOXICITY3.1.1 Definitions and general considerations for acute toxicityAnnex I: 3.1.1.1. Acute toxicity means those adverse effects occurring following oral or dermaladministration of a single dose of a substance or a mixture, or multiple doses given within 24 hours,or an inhalation exposure of 4 hours.Acute toxicity relates to effects occurring after a single or relatively brief exposure to asubstance. The definition in CLP reflects the fact that the evidence for acute toxicity is usuallyobtained from animal testing. In particular, acute toxicity is usually characterised in terms oflethality and exposure times are based around those used in experimental protocols. However,classification for acute toxicity can also be based on human evidence which shows lethalityfollowing human exposure.There are two hazard classes for acute toxicity – “Acute toxicity” and “STOT-SE”. These areindependent of each other and both may be assigned to a substance or a mixture if therespective criteria are met. However, care should be taken not to assign each class for thesame effect, essentially giving a “double classification”, even where the criteria for bothclasses are fulfilled. In such a case the most appropriate class should be assigned.Acute toxicity classification is generally assigned on the basis of evident lethality (e.g. anLD 50 /LC 50 value), or, where the potential to cause lethality can be concluded from evidenttoxicity (e.g. from the fixed dose procedure). STOT-SE should be considered where there isclear evidence of toxicity to a specific organ, especially when it is observed in the absence oflethality (see Chapter 3.8).For more details see IR/CSA, Section R.7.4.1.1.Annex I: 3.1.1.2. The hazard class Acute Toxicity is differentiated into:– Acute oral toxicity;– Acute dermal toxicity;– Acute inhalation toxicity.The classification shall be considered for each route of exposure, using the appropriateapproach as described in Section 3.1.2.2. If different hazard categories are assigned, the moresevere hazard category will be used for the classification for acute toxicity, with theappropriate pictogram and signal word. For each relevant route of exposure, the hazardstatement will correspond to the classification of this specific route.45 The guidance provided in this chapter is based on the classification criteria from the original version of theCLP Regulation (EC) No 1272/2008. This chapter is currently being updated based on the 2 nd ATP to the CLPRegulation and planned for a future update of this document in 2013.193

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.2 Classification of substances for acute toxicity3.1.2.1 Identification of hazard information3.1.2.1.1 Identification of human dataRelevant information with respect to acute toxicity may be available from case reports,epidemiological studies, medical surveillance and reporting schemes and national poisoncentres. Human data to be considered for acute toxicity should report severe effects aftersingle exposure or exposure of less than 24h, but data on severe effects after a few exposuresover a few days can also be considered on a case by case basis.For more details see IR/CSA, Section R.7.4.3.2.3.1.2.1.2 Identification of non-human dataNon-testing data:Physicochemical dataPhysico-chemical properties, such as pH, physical state, form, solubility, vapour pressure andparticle size, can be important parameters in evaluating toxicity studies and in determining themost appropriate classification. This is especially valid with respect to inhalation wherephysical form and particle size can have a significant impact on toxicity (see Section3.1.2.3.2).(Q)SAR models, expert systems and grouping methods“Non-testing data can be provided by the following approaches: a) structure-activityrelationships (SARs) and quantitative structure-activity relationships (QSARs), collectivelycalled (Q)SARs; b) expert systems incorporating (Q)SARs and/or expert rules; and c)grouping methods (read-across and categories. These approaches can be used to assess acutetoxicity if they provide relevant and reliable (adequate) data for the chemical of interest. .…Compared with some endpoints, there are relatively few (Q)SAR models and expert systemscapable of predicting acute toxicity.” (IR/CSA, Section R.7.4.3.1).Testing data:In vitro dataThere are currently no in vitro tests that have been officially adopted by the EU or OECD forassessment of acute toxicity (IR/CSA, Section R.7.4.3.1). Any available studies should beassessed by using expert judgement.Animal dataA number of different types of studies have been used to investigate acute toxicity. Olderstandard studies were designed to determine lethality and estimate the LD 50 /LC 50 . In contrast,contemporary study protocols, such as the fixed dose procedure, use signs of overt (“evident”)toxicity rather than lethality as indications of acute toxicity. These studies are generallyconducted using preferred species, i.e. the rat for acute oral and inhalation toxicity studies,and in addition rabbit for dermal toxicity studies.The animal studies are listed in IR/CSA, Section R.7.4.3.1.194

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.2.2 Classification criteriaAnnex I: 3.1.2.1. Substances can be allocated to one of four toxicity categories based on acutetoxicity by the oral, dermal or inhalation route according to the numeric criteria shown inTable 3.1.1. Acute toxicity values are expressed as (approximate) LD50 (oral, dermal) or LC50(inhalation) values or as acute toxicity estimates (ATE). Explanatory notes are shown followingTable 3.1.1.Table 3.1.1Acute toxicity hazard categories and acute toxicity estimates (ATE) defining the respectivecategoriesExposure Route Category 1 Category 2 Category 3 Category 4Oral (mg/kgbodyweight)See Note (a)Dermal (mg/kgbodyweight)See Note (a)Gases (ppmV 1 )see: Note (a)Note (b)Vapours (mg/l)see: Note (a)Note (b)Note (c)Dusts and Mists(mg/l)see: Note (a)Note (b)ATE ≤ 5 5 < ATE ≤ 50 50 < ATE ≤ 300 300 < ATE≤ 2000ATE ≤ 50 50 < ATE ≤ 200 200 < ATE≤ 1000ATE ≤ 100 100 < ATE ≤ 500 500 < ATE≤ 25001000 < ATE≤ 20002500 < ATE≤ 20000ATE ≤ 0.5 0.5 < ATE ≤ 2.0 2.0 < ATE ≤ 10.0 10.0 < ATE≤ 20.0ATE ≤ 0.05 0.05 < ATE ≤ 0.5 0.5 < ATE ≤ 1.0 1.0 < ATE ≤ 5.01 Gas concentrations are expressed in parts per million per volume (ppmV)195

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Notes to Table 3.1.1:(a) The acute toxicity estimate (ATE) for the classification of a substance or ingredient in amixture is derived using:- the LD50/LC50 where available,- the appropriate conversion value from Table 3.1.2 that relates to the results of a range test, or- the appropriate conversion value from Table 3.1.2 that relates to a classification category.(b) Generic concentration limits for inhalation toxicity in the table are based on 4 hour testingexposures. Conversion of existing inhalation toxicity data which have been generated using a 1hour exposure can be carried out by dividing by a factor of 2 for gases and vapours and 4 for dustsand mists.(c) For some substances or mixtures the test atmosphere will not just be a vapour but willconsist of a mixture of liquid and vapour phases. For other substances or mixtures the testatmosphere may consist of a vapour which is near the gaseous phase. In these latter cases,classification shall be based on ppmV as follows: Category 1 (100 ppmV), Category 2 (500 ppmV),Category 3 (2500 ppmV), Category 4 (20 000 ppmV).The terms 'dust', 'mist' and 'vapour' are defined as follows:- Dust: solid particles of a substance or mixture suspended in a gas (usually air);- Mist: liquid droplets of a substance or mixture suspended in a gas (usually air);- Vapour: the gaseous form of a substance or mixture released from its liquid or solid state.Dust is generally formed by mechanical processes. Mist is generally formed by condensation ofsupersaturated vapours or by physical shearing of liquids. Dusts and mists generally have sizesranging from less than 1 to about 100 µm.Comment to Table 3.1.1, Note (b):The classification criteria for acute inhalation toxicity relate to a 4-hour experimentalexposure period. Where LC 50 values have been obtained in studies using exposure durationsshorter or longer than 4 hours values these may be adjusted to a 4-hour equivalent usingHaber’s law (C n·t=k) for direct comparison with the criteria. The value of n, which is specificto individual substances, should be chosen using expert judgement. If an appropriate value ofn is not available in the literature then it may sometimes be derived from the availablemortality data using probits (i.e.the inverse cumulative distribution functions associated withthe standard normal distribution). Alternatively, some default values are recommended(IR/CSA, Section R.7.4.4.1).Particular care should be taken when using Haber’s law to assess inhalation data onsubstances which are corrosive or locally active. In all cases, Haber’s law should only be usedin conjunction with expert judgement.It is noted that the statements in IR/CSA, Section R.7.4.4.1, with respect to Haber’s law arenot consistent with those of CLP. However the CLP approach must be used for classificationand labelling.Comment to Table 3.1.1, Note (c):The term “aerosol” is commonly used for “dust and mists”.196

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.2.3 Evaluation of hazard information3.1.2.3.1 Evaluation of human dataThe evaluation of human data often becomes difficult due to various limitations frequentlyfound with the types of studies and data highlighted in Section 3.1.2.1.1. These includeuncertainties relating to exposure assessment (i.e. unreliable information on the amount ofsubstance the subjects were exposed to) and uncertain exposure to other substances. As such,human data needs careful expert evaluation to properly judge the reliability of the findings. Itshould be acknowledged that human data often do not provide sufficiently robust evidence ontheir own to support classification. They may however contribute to a weight of evidenceassessment with other available information such as data from animal studies.The classification for acute toxicity is based primarily on the dose/concentration that causesmortality (the Acute Toxicity Estimate, ATE), which is then related to the numerical values inthe classification criteria according to CLP Annex I, Table 3.1.1 (See Section 3.1.2.2) forsubstances or for use in the additivity formula in CLP Annex I, 3.1.3.6.1 and 3.1.3.6.2.3 formixtures (See Section 3.1.3.3). The ATE is usually obtained from animal studies but inprinciple suitable human data can also be used if available. Where human data are availablethey should be used to estimate the ATE which can be used directly for classification asdescribed above.The minimum dose/concentration or range shown or expected to cause mortality after a singlehuman exposure can be used to derive the human ATE directly, without any adjustments oruncertainty factors. See Example 1 (methanol).If there are no exact/quantitative lethal dose data the procedure described in CLP Annex I,3.1.3.6.2.1.(b) (See Section 3.1.3.3.4) would have to be followed using Table 3.1.2, (SeeSection 3.1.3.3.3) with an assessment of the available information on a semi-quantitative orqualitative basis.Expert judgement is needed in a total weight of evidence approach taking relevance,reliability, and adequacy of the information into account. See Example 2 (N,Ndimethylaniline).If the available human data alone are too limited to support a classification they may stillprovide supporting evidence in the overall weight of evidence assessment.3.1.2.3.2 Evaluation of non-human dataAnnex I: 3.1.2.2. Specific considerations for classification of substances as acutely toxic3.1.2.2.1. The preferred test species for evaluation of acute toxicity by the oral and inhalation routesis the rat, while the rat or rabbit are preferred for evaluation of acute dermal toxicity. Whenexperimental data for acute toxicity are available in several animal species, scientific judgementshall be used in selecting the most appropriate LD 50 value from among valid, well-performed tests.Evaluation of non-testing and in vitro data:Results of (Q)SAR, grouping and read-across may be used instead of testing, and substanceswill be classified and labelled on this basis if the method fulfils the criteria described inAnnex XI of REACH. See also IR/CSA, Section R.7.4.4.1.ATE – establishing:- Basis LD 50 /LC 50 : An available LD 50 /LC 50 is an ATE at first stage.197

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012- Results from a range test: According to CLP Annex I, Table 3.1.2 results from range tests(i.e. doses/exposure concentrations that cause acute toxicity in the range of numeric criteriavalues) can be assigned to the four different categories of acute toxicity for each possibleroute of exposure (centre column). Further, Table 3.1.2 allows allocating a single value, theconverted acute toxicity point estimate (cATpE), to each experimentally obtained acutetoxicity range estimate or classification category (right column), see Note (a) to Table 3.1.1.This cATpE can be used in the additivity formulae (Annex I, 3.1.3.6.1 and 3.1.3.6.2.3) tocalculate the acute toxicity of mixtures.- In case of multiple LD 50 /LC 50 s or from several species:Where several experimentally determined ATE values (i.e. LD 50 , LC 50 values or ATE derivedfrom studies using signs of non-lethal toxicity) are available, expert judgement needs to beused to choose the most appropriate value for classification purposes. Each study needs to beassessed for its suitability in terms of study quality and reliability, and also for its relevance tothe substance in question in terms of technical specification and physical form. Studies notconsidered suitable on reliability or other grounds should not be used for classification.In general, classification is based on the lowest ATE value available i.e. the lowest ATE inthe most sensitive appropriate species tested. However, expert judgement may allow anotherATE value to be used in preference, provided this can be supported by a robust justification.If there is information available to inform on species relevance, then the studies conducted inthe species most relevant for humans should normally be given precedence over the studies inother species. If there is a wide range of ATE values from the same species, it may beinformative to consider the studies collectively, to understand possible reasons for thedifferent results obtained. This would include consideration of factors such as the animalstrains used, the experimental protocols, the purity of the substance and form/phase in whichit was tested (e.g. the particle size distribution of any aerosols or dusts tested), as well asexposure mode and numerous technical factors in inhalation studies. This assessment may aidselection of the most appropriate study on which to base the classification.If there are different LD 50 values from tests using different vehicles e.g. water vs. corn oil orneat substance vs. corn oil), generally the lowest valid value would be the basis forclassification. It is not considered appropriate to combine or average the available ATEvalues. The studies may not be equivalent (in terms of experimental design such as protocol,purity of material tested, species of animal used etc) making such a collation or combinationunsound.If there is a study available with a post-observation period of only ca 7 days (instead of the 14days according to the OECD guidelines) and there are effects still observed at the end of thestudy, the resulting LD 50 might be misleading. A long persistency of effects may be indicativeof cumulative toxicity, sometimes coinciding with flat dose-response relationships, sometimeswith species differences. Such information should be included in the weight of evidenceconsideration.Annex I: 3.1.2.3. Specific considerations for classification of substances as acutely toxic by theinhalation route3.1.2.3.1 Units for inhalation toxicity are a function of the form of the inhaled material. Values fordusts and mists are expressed in mg/l. Values for gases are expressed in ppmV. Acknowledging thedifficulties in testing vapours, some of which consist of mixtures of liquid and vapour phases, thetable provides values in units of mg/l. However, for those vapours which are near the gaseousphase, classification shall be based on ppmV.Conversions:198

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Differentiation between vapour and mist will be made on the basis of the saturated vapourconcentration (SVC) for a volatile substance, which can be calculated by the followingequation:SVC [mg/l] = 0.0412 x MW x vapour pressure in hPa at 20°C.The conversion from mg/l to ppm assuming an ambient pressure of 1 at = 101.3 kPa and 25°Cis: ppm=0.0245 mg/l x 1/MW.An LC 50 well below the SVC will be considered for classification according to the criteria forvapours; whereas an LC 50 close to or above the SVC will be considered for classificationaccording to the criteria for mists (see also Draft OECD TG 39).Considerations with respect to physical forms or states / bioavailability:Article 9(5) When evaluating the available information for the purposes of classification, themanufacturers, importers and downstream users shall consider the forms or physical states in whichthe substance or mixture is placed on the market and in which it can reasonably be expected to beused.For further details see Sections 1.2 and 1.3.Special considerations concerning aerosols:The test guidelines for acute inhalation toxicity with aerosols require rodents to be exposed toan aerosol containing primarily respirable particles (with a Mass Median AerodynamicDiameter (MMAD) of 1 – 4 µm), so that particles can reach all regions of the respiratorytract. The use of such fine aerosols helps to avoid partial overloading of extra-thoracicairways in obligate nasal breathing species like rats. Results from studies in which substanceswith particle size with a MMAD > 4 µm have been tested can generally not be used forclassification, but expert judgement is needed in cases where there are indications of hightoxicity.The use of highly respirable aerosols is ideal to fully investigate the potential inhalationhazard of the substance. However, it is acknowledged that these exposures may notnecessarily reflect realistic conditions. For instance, solid materials are often micronised to ahighly respirable form for testing, but in practice exposures will be to a dust of much lowerrespirability. Similarly, pastes or highly viscous materials with low vapour pressure needstrong measures to be taken to generate airborne particulates of sufficiently high respirability,whereas for other materials this may occur spontaneously. In such situations, specificproblems may arise with respect to classification and labelling, as these substances are testedin a form (i.e. specific particle size distribution) that is different from all the forms in whichthese substances are placed on the market and in which they can reasonably be expected to beused.A scientific concept has been developed as a basis for relating the conditions of acuteinhalation tests to those occurring in real-life, in order to derive an adequate hazardclassification. This concept is applicable only to substances or mixtures which are proven tocause acute toxicity through local effects and do not cause systemic toxicity (Pauluhn, 2008).Corrosive substancesIt is presumed that corrosive substances (and mixtures) will cause toxicity by inhalationexposure. In cases where no acute inhalation test has been performed special considerationshould be given to the need to communicate this potential hazard.199

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Corrosive substances (and mixtures) may be acutely toxic after inhalation to a varying degreeand by different modes of action. Therefore, it is not possible to estimate the acute inhalationtoxicity from the corrosivity data alone.There are special provisions for hazard communication of acutely toxic substances by acorrosive effect, see Section 3.1.4.2.3.1.2.3.3 Weight of evidenceIn cases where there is sufficient human evidence that meets the criteria given in Section3.1.2.2 then this will normally lead to classification for acute toxicity, irrespective of otherinformation available.If there are human data indicating no classification but there are also non-human dataindicating classification then the classification is based on the non-human data unless it isshown that the human data cover the exposure range of the non-human data or that the nonhumandata are not relevant for humans. If the human and non-human data both indicate noclassification then classification is not required.If there are no human data then the classification is based on the non-human data.For the role and application of expert judgement and weight of evidence determination, seeCLP Annex I, 1.1.1.3.1.2.4 Decision on classificationThe classification has to be performed with respect to all routes of exposure (oral, dermal,inhalation) on the basis of all adequate and reliable available information.3.1.2.5 Setting of specific concentration limitsSpecific concentration limits are not applicable for acute toxicity classification. Rather, therelative potency of substances is implicitly taken into account in the additivity formula (seeSection 3.1.3.3.3). For this reason specific concentration limits for acute toxicity will notappear in CLP Annex VI, Table 3.1 or in the classification and labelling inventory (CLPArticle 42).3.1.2.6 Decision logicThe decision logic below is provided as additional guidance. It is strongly recommended thatthe person responsible for classification is fully familiar with the criteria for acute toxicityclassification before using the decision logic.For a complete classification of a substance, the decision logic must be worked out for eachroute of exposure for which data and/or information is available. For example, if a certainsubstance is classified in Category 1 based on an oral LD 50 5 mg/kg bodyweight (the answerwas 'Yes' in box 2 for item (a)), it is still necessary to go back to box 2 in the decision logicand complete the classification for the dermal (b) and inhalation (c)-(e) route of exposure,when data is available for one or both of these routes of exposure. In case there are data for allthree routes of exposure, the classification for acute toxicity of the substance will include thethree differentiations of the hazard class, which might end up in three different categories.The route of exposure will then be specified in the corresponding hazard statement.200

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Are there data and/or information toevaluate acute toxicity?NOClassification not possibleYESAccording to the criteria in CLP Annex I, 3.1.2 to 3.1.3.4, doesit have an:(a) Oral LD 50 5 mg/kg bodyweight; or(b) Dermal LD 50 50 mg/kg bodyweight; or(c) Inhalation (gas) LC 50 100 ppm; or(d) Inhalation (vapour) LC 50 0.5 mg/l ; or(e) Inhalation (dust/mist) LC50 0.05 mg/l?YESCategory 1DangerNOAccording to the criteria in CLP Annex I, 3.1.2 to 3.1.3.4, doesAccording to the criteria in CLP Annex I, 3.1.2 to 3.1.3.4, does ithave it have an: an:(a)Oral LD 50 50 >5 >5 but but 50 50 mg/kg mg/kg bodyweight; bodyweight; or or(b) Dermal LD 50 >50 but 200 mg/kg bodyweight; or(b) Dermal LD 50 >50 but 200 mg/kg bodyweight; or(c) Inhalation (gas) LC 50 >100 but < 500 ppm; or(c) (d)Inhalation (gas) (vapour) LC 50 >100 LC 50 but > 0.5 < 500 but ppm; < 2.0 or mg/l; or(d) (e)Inhalation (vapour) (dust/mist) LC 50LC >05but< 50 >0.05 20mg/l;or but 0.5 mg/l?YESCategory 2DangerNOAccording to the criteria in CLP Annex I, 3.1.2 to 3.1.3.4, doesit have an:(a) Oral LD 50 >50 but ≤ 300 mg/kg bodyweight; or(b) Dermal LD 50 > 200 but ≤ 1000 mg/kg bodyweight; or(c) Inhalation (gas) LC 50 >500 but ≤ 2500 ppm; or(d) Inhalation (vapour) LC 50 >2 but ≤ 10.0 mg/l; or(e) Inhalation (dust/mist) LC 50 >0.5 but ≤ 1.0 mg/l?YESCategory 3DangerNOAccording to the criteria in CLP Annex I, 3.1.2 to 3.1.3.4, doesit have an:(a) Oral LD 50 >300 but ≤ 2000 mg/kg bodyweight; or(b) Dermal LD 50 >1000 but ≤ 2000 mg/kg bodyweight; or(c) Inhalation (gas) LC 50 >2500 but ≤ 20000 ppm; or(d) Inhalation (vapour) LC 50 >10 but ≤ 20 mg/l; or(e) Inhalation (dust/mist) LC 50 >1 but ≤ 5 mg/l?YESCategory 4WarningNONo classification201

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.3 Classification of mixtures for acute toxicity3.1.3.1 General considerations for classificationAnnex I: 3.1.3.1. The criteria for classification of substances for acute toxicity as outlined insection 3.1.2 are based on lethal dose data (tested or derived). For mixtures, it is necessary to obtainor derive information that allows the criteria to be applied to the mixture for the purpose ofclassification. The approach to classification for acute toxicity is tiered, and is dependent upon theamount of information available for the mixture itself and for its ingredients.The procedure for classifying mixtures is a tiered i.e. a stepwise approach based on ahierarchy principle and depending on the type and amount of available data/information. Ifvalid test data are available for the whole mixture they have precedence. If no such data exist,the so called bridging principles have to be applied if possible. If the bridging principles arenot applicable an assessment on the basis of ingredient information will be applied (seeSections 3.1.3.3.3, 3.1.3.3.4 and 3.1.3.5).3.1.3.2 Identification of hazard informationWhere toxicological information from human evidence and animal studies is available on amixture, this should be used to derive the appropriate classification. Such information may beavailable from the mixture manufacturer. Where such information on the mixture itself is notavailable, information on similar mixtures and/or the component substances in the mixturemust be used, as described in Section 3.1.3.3.Alternatively, the hazard information on all individual components in the mixture could beidentified as described in Section 3.1.2.2.3.1.3.3 Classification criteriaAnnex I: 3.1.3.2. For acute toxicity each route of exposure shall be considered for the classificationof mixtures, but only one route of exposure is needed as long as this route is followed (estimated ortested) for all ingredients. If the acute toxicity is determined for more than one route of exposure,the more severe hazard category will be used for classification. All available information shall beconsidered and all relevant routes of exposure shall be identified for hazard communication.The classification shall be considered for each route of exposure, using the appropriateapproach as described in Section 3.1.2.3. If different hazard categories are assigned, the moresevere hazard category will be used for the classification for acute toxicity, with theappropriate pictogram and signal word. For each relevant route of exposure, the hazardstatement will correspond to the classification of this specific route.3.1.3.3.1 When data are available for the complete mixtureAnnex I: 3.1.3.4.1. Where the mixture itself has been tested to determine its acute toxicity, it shallbe classified according to the same criteria as those used for substances, presented in Table 3.1.1.In general, where a mixture has been tested those data should be used to support classificationaccording to the same criteria as used for substances. However, there should be someconsideration of whether the test is appropriate. For instance, if the mixture contains asubstance for which the test species is not considered appropriate (for instance a mixturecontaining methanol tested in rats which are not sensitive to methanol toxicity), then theappropriateness of these data for classification should be considered using expert judgement.202

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012With respect to the classification of mixtures in the form of dust for acute inhalation toxicity,the particle size can affect the toxicity and the resulting classification should take this intoaccount (see Section 3.1.2.3.2).3.1.3.3.2 When data are not available for the complete mixture: bridging principlesAnnex I: 3.1.3.5.1. Where the mixture itself has not been tested to determine its acute toxicity, butthere are sufficient data on the individual ingredients and similar tested mixtures to adequatelycharacterise the hazards of the mixture, these data shall be used in accordance with the bridgingrules set out in section 1.1.3.3.1.3.3.3 When data are available for all components or only for some componentsAnnex I: 3.1.3.6. Classification of mixtures based on ingredients of the mixture (Additivityformula)3.1.3.6.1. Data available for all ingredientsIn order to ensure that classification of the mixture is accurate, and that the calculation need only beperformed once for all systems, sectors, and categories, the acute toxicity estimate (ATE) ofingredients shall be considered as follows:(a)(b)(c)Include ingredients with a known acute toxicity, which fall into any of the acute toxicitycategories shown in Table 3.1.1;Ignore ingredients that are presumed not acutely toxic (e.g., water, sugar);Ignore ingredients if the oral limit test does not show acute toxicity at 2000 mg/kgbodyweight.Ingredients that fall within the scope of this paragraph are considered to be ingredients with aknown acute toxicity estimate (ATE).The ATE of the mixture is determined by calculation from the ATE values for all relevantingredients according to the following formula below for Oral, Dermal or Inhalation Toxicity:where:100ATEmix nCiATEC i = concentration of ingredient i (% w/w or % v/v)i = the individual ingredient from 1 to nn = the number of ingredientsATE i = Acute Toxicity Estimate of ingredient i.The additivity formula cannot be used directly for mixtures containing substances tested forinhalation toxicity as vapours and others as dust, because it is unclear when the numericvalues for vapours or dusts must be used. Therefore for acute inhalation toxicity the additivityformula should be used separately for each relevant physical form (i.e. gas, vapour and/ordust/mist), using the appropriate categories in Table 3.1.1. In case of different outcomes, themost severe classification applies.Annex I: Table 3.1.2Conversion from experimentally obtained acute toxicity range values (or acute toxicityhazard categories) to acute toxicity point estimates for classification for the respective routesof exposureExposure routesClassification category or experimentallyiConverted acutetoxicity point estimate203

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Oral(mg/kg bodyweight)Dermal(mg/kg bodyweight)Gases(ppmV)Vapours(mg/l)Dust/mist(mg/l)obtained acute toxicity range estimate (see Note 1)0 < Category 1 55 < Category 2 5050 < Category 3 300300 < Category 4 20000 < Category 1 5050 < Category 2 200200 < Category 3 10001000 < Category 4 20000 < Category 1 100100 < Category 2 500500 < Category 3 25002500 < Category 4 200000 < Category 1 0.50.5 < Category 2 22.0 < Category 3 10.010.0 < Category 4 20.00< Category 1 0.050.05 < Category 2 0.50.5 < Category 3 1.01.0 < Category 4 5.00.5510050055030011001010070045000.050.53110.005Note 1:These values are designed to be used in the calculation of the ATE for classification of a mixturebased on its components and do not represent test results.Some converted Acute Toxicity point Estimates (cATpEs) are equal to the upper limit of thenext lower category, for example the cATpE of oral Category 2 (5 mg/kg) is equal to theupper limit of oral Category 1 (also 5 mg/kg).This can lead to a problem when using the cATpE values for calculating the acute toxicity ofmixtures. For instance, using the cATpEs for a mixture containing only substances classifiedin Category 2 actually results in a Category 1 classification for the mixture. Similarly, amixture containing substances classified as Category 3 for dust/mist results in a Category 2classification. Clearly these outcomes are incorrect and are an unintended side-effect of theapproach. To address this problem the following proposal has been endorsed at UN SCE-GHS: “If the acute toxicity range values (or acute toxicity hazard classification categories) forall ingredients of a mixture are within the same range or category, then the mixture should beclassified in that category.” Applying this to the cases highlighted above, these mixtureswould be classified in Category 2 and 3, respectively.0.050.51.5204

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: 3.1.3.3.(b) where a classified mixture is used as an ingredient of another mixture, theactual or derived acute toxicity estimate (ATE) for that mixture may be used, when calculating theclassification of the new mixture using the formulas in section 3.1.3.6.1 and paragraph 3.1.3.6.2.3.It is important that the downstream user has sufficient information in order to enable him toperform a correct classification of mixtures.3.1.3.3.4 When data are not available for all componentsAnnex I: 3.1.3.6.2.1. Where an ATE is not available for an individual ingredient of the mixture, butavailable information such as that listed below can provide a derived conversion value such asthose laid out in Table 3.1.2, the formula in paragraph 3.1.3.6.1 shall be applied.This includes evaluation of:(a)(b)(c)(d)_______________extrapolation between oral, dermal and inhalation acute toxicity estimates ( 1 ). Such anevaluation could require appropriate pharmacodynamic and pharmacokinetic data;evidence from human exposure that indicates toxic effects but does not provide lethal dosedata;evidence from any other toxicity tests/assays available on the substance that indicates toxicacute effects but does not necessarily provide lethal dose data; ordata from closely analogous substances using structure/activity relationships.( 1 ) For ingredients with acute toxicity estimates available for other than the most appropriate exposure route,values may be extrapolated from the available exposure route(s) to the most appropriate route. Dermal andinhalation route data are not always required for ingredients. However, in case data requirements for specificingredients include acute toxicity estimates for the dermal and inhalation route, the values to be used in theformula need to be from the required exposure route.Derivation of ATEs from available information:When ingredients have a known acute toxicity (LC 50 or LD 50 values), this value has to beused in the additivity formula. However, for many substances, acute toxicity data will not beavailable for all exposure routes.CLP allows for two ways of deriving acute toxicity conversion values. One option is to usethe converted acute toxicity point estimates supplied in Annex I, Table 3.1.2. The otheroption, expert judgement would recommend in substantiated cases the use of the directlyderived ATE values.a) Route-to-route extrapolation (CLP Annex I, 3.1.3.6.2.1.(a)):Route-to-route extrapolation is defined as the prediction of the total amount of a substanceadministered by one route that would produce the same systemic toxic response as thatobtained by a given amount of a substance administered by another route. Thus, route-torouteextrapolation is only applicable for the evaluation of systemic effects. It is notappropriate to assess direct local effects.This extrapolation is possible if certain conditions are met, which substantiate the assumptionthat an internal dose causing a systemic effect at the target is related to an externaldose/concentration; preferably the absorption can be quantified. Therefore information on thephysico-chemical and biokinetic properties should be available and assessed in order to allowsuch a conclusion and performing an extrapolation across routes. In the absence of anyinformation on absorption, 100% absorption has to be presumed as a worst case for thedermal and inhalation route. Extrapolating from the oral route to other routes, the assumptionof absorption of 100% for the oral route is, however, not a worst case. Absorption of less than205

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012100% by the oral route will lead to lower ATEs. Another important factor is the local andsystemic metabolic pathways; in particular it must be assured that no route-specificmetabolism/degradation of substance occurs.If extrapolating from oral data, the influence of first-pass metabolism in thestomach/intestines and the liver should be considered, especially if the substance is detoxified.Such first pass metabolism is unlikely to occur to any significant extent by the dermal orinhalation routes, and so this would lead to an underestimate of toxicity by these routes. Thusif based on kinetic or (Q)SAR data a specific first-pass effect is excluded, oral data may beused for extrapolation purposes.For an extrapolation to the dermal route, information on the potential skin penetration may bederived from the chemical structure (polar vs. nonpolar structure elements, Log P ow ,molecular weight) if kinetic data are not available which would allow a quantitativecomparison. When no such information is available 100% dermal absorption should bepresumed.Similarly for an extrapolation to the inhalation route if there is no quantitative information onabsorption then 100% absorption should be presumed. Inhalation volatility is an importantfactor which on one hand may increase the exposure, but on the other hand may reduceabsorption due to higher exhalation rates. The solubility (in water and non-polar solvents) hasto be considered, as well as particle size, which plays a particularly important role ininhalation toxicity.Route-to-route extrapolation is not always appropriate. For example where there is asubstantial difference in absorption between oral and inhalation uptake (e.g. poorly solubleparticles, substances that decompose within the gastro intestinal-tract), or where the substancecauses local effects, the toxicity by different routes may be significantly different, and routeto-routeextrapolation may not be appropriate (ECETOC TR 86, 2003).i: Extrapolation oral inhalation:If the mentioned conditions are met an extrapolation from oral data would be performed asfollows:Incorporated dose = concentration x respiratory volume x exposure time1 mg/kg bw = 0.0052 mg/l/4husing a respiratory volume for a 250 g rat of 0.20 l/min and 100 % absorption and postulating100% deposition and absorption (IR/CSA, Chapter R7C, Table R.7.12-10).Valid information that the deposition and/or absorption rate for the extrapolated route is lowerwould allow a higher equivalent derived ATE (see Section 3.1.6.1.9 Example 9).ii: Extrapolation oral dermalIf based on kinetic or SAR data a high penetration rate can be assumed and a specific firstpass-effect is excluded, oral and dermal toxicity might be regarded as equivalent. This israrely the case.Solids themselves may have a very low absorption rate, but if diluted in an appropriatesolvent there may be appreciable absorption. Thus depending on the kinetic and physicochemicalproperties and kind of mixture varying ATEs will result. An example for thesedifferences is butyn-1,4-diol which dermally applied as solid shows no mortality in rats at5000 mg/kg, whereas the aqueous solution giving LD 50 of 659 and 1,240 mg/kg, the oralLD 50 are in the range of 200 mg/kg.206

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012For more details on inter-route extrapolation see IR/CSA, Section R.7C.12.1.5. Example 9and 10 illustrate this approach.b) Evidence from human exposure:Human evidence can be used to derive an appropriate ATE to use in the additivity approachfor mixtures (CLP Annex I, 3.1.3.6.1 and 3.1.3.6.2.3). Therefore it is necessary to extrapolatefrom adequate and reliable data and taking the potency (i.e. the magnitude of the lethal dosereported) of the effects in humans into account. Thus an equivalent ATE may be derived onthe basis of valid human toxicity data (minimum dose/concentration) and used directly in theadditivity formulae (see Section 3.1.6.1.1 Example 1). The alternative to the derivation of anequivalent ATE is the allocation to a category. The category should be justified by semiquantitativeor qualitative data and a subsequent derivation of a converted ATE (cATpE)according to CLP Annex I, Table 3.1.2 and subsequently use in the formulae (see Section3.1.6.1.2 Example 2). See also Section 3.1.2.3.1 for more details.c) Evidence from other toxicity tests:Information from other types of studies can sometimes be useful in deriving an acute toxicityclassification. (see Section 3.1.2.2). These studies will not usually provide an LD 50 /ATE valuethat can be used directly for classification, but they may provide enough information to allowan estimate of acute toxicity to be made, which would be sufficient to support a decision onclassification.Example:Available information: In a range finding study with respect to repeated dose toxicity dailyoral doses of 1000 mg/kg over 5 days prove to be neither lethal nor cause serious symptomsin rats at the end of the observation period of 14 days.Conclusion: the LD 50 =ATE is >2000 mg/kg since 2 doses following (within roughly) 24 h arenot lethal (see Section 3.1.2.2). Thus this ingredient can be ignored in the additivityprocedure.d) Use of (Q)SAR:LD 50 /LC 50 values predicted by a highly reliable model (see Section 2.1.2) may be usedaccording to Note (a) to Annex I, Table 3.1.1 directly as LD 50 /LC 50 =ATE in the additivityformula CLP Annex I, 3.1.3.6.1. If the assessment using (Q)SARs gives a more general resulta cATpE acc. to Table 3.1.2 may be derived. It has to be emphasised that these approachesgenerally require substantial technical information, and expert judgement, to reliably estimateacute toxicity.Annex I: 3.1.3.6.2.2. In the event that an ingredient without any useable information at all is usedin a mixture at a concentration of 1% or greater, it is concluded that the mixture cannot beattributed a definitive acute toxicity estimate. In this situation the mixture shall be classified basedon the known ingredients only, with the additional statement that x percent of the mixture consistsof ingredient(s) of unknown toxicity.Further guidance on how to apply this provision is given in Section 3.1.3.3.5.Annex I: 3.1.3.6.2.3. If the total concentration of the ingredient(s) with unknown acute toxicity is ≤10 % then the formula presented in section 3.1.3.6.1 shall be used. If the total concentration of theingredient(s) with unknown toxicity is > 10 %, the formula presented in section 3.1.3.6.1 shall becorrected to adjust for the total percentage of the unknown ingredient(s) as follows:207

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012208100 Cumknownif 10% ATEmixnCiATE3.1.3.3.5 Components that should be taken into account for the purpose ofclassificationAnnex I: 3.1.3.3.(a) the ‘relevant ingredients’ of a mixture are those which are present inconcentrations of 1 % (w/w for solids, liquids, dusts, mists and vapours and v/v for gases) orgreater, unless there is a reason to suspect that an ingredient present at a concentration of less than1 % is still relevant for classifying the mixture for acute toxicity (see Table 1.1).When a mixture contains a “relevant” ingredient (i.e. constituting ≥ 1%; Annex I, 3.1.3.3 (a))for which there is inadequate acute toxicity data then the mixture must be classified on thebasis of the ingredients with known toxicity, with an additional statement to indicate that themixture contains ingredients of unknown toxicity (CLP Annex I, 3.1.3.6.2.2). Thedetermination of the classification depends on what proportion of the mixture such ingredientsof unknown toxicity constitute. If these ingredients constitute ≤10% of the total mixture, theadditivity formula in 3.6.1.1 may be used. However, in cases where these ingredientsconstitute over 10%, a modified additivity formula, which adjusts for the presence of asignificant proportion of ingredients of unknown toxicity, is used. This reflects the greateruncertainty as to the true toxicity of the mixture (CLP Annex I, 3.3.3.2.4).Hazard classAcute Toxicity:Annex I: Table 1.1Generic cut-off valuesiGeneric cut-off values to be taken into accountCategory 1-3 0,1 %Category 4 1 %Note: Generic cut-off values are in weight percentages except for gaseous mixtures where they arein volume percentage.As indicated in CLP Annex I, Table 1.1, when components are present in low concentrationsthey do not need to be taken into account when determining the classification of the mixture,according to the approaches detailed in CLP Annex I, 3.1.3.6.1 and 3.1.6.2.3 (see Section3.1.6.3.1 Example 11). Accordingly, all components classified in Categories 1-3 at aconcentration

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012experience or information. These validated data will then be used in the additivity formula inAnnex I, 3.1.3.6.1 as ATEs or cATpEs (Annex I, Table 3.1.2).3.1.3.5 Decision on classificationThe assessment on classification has to be performed with respect to the relevant routes ofexposure (oral, dermal, inhalation) on the basis of all adequate reliable data. If a classificationis warranted in different categories for different routes, then the mixture has to be classified inthe more severe category, the other routes fulfilling the criteria for a classification are takencare by allocating the corresponding hazard statement(s) and appropriate precautionarystatement(s). If for example, a mixture fulfils the criteria for oral toxicity Category 3 and forinhalation Category 2, then the mixture will be classified in Category 2, the correspondinghazard statements for both inhalation Category 2 and oral Category 3 will be assigned.209

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.3.6 Decision logicThe decision logic is provided as additional guidance. It is strongly recommended that theperson responsible for classification, study the criteria for classification before and during useof the decision logic.Does the mixture as a whole have data/informationto evaluate acute toxicity?NOYESClassify in appropriatecategory according to CLPAnnex I, Table 3.1.1toxicity?Can bridging principles be applied?YESClassify inappropriatecategoryNOIs acute toxicity data available for allingredients of mixture?NOYESApply the acute toxicityestimate calculation todetermine the ATE of themixtureIs it possible to estimate missing ATE(s)of the ingredient(s), i.e. can conversionvalue(s) be derived?NOIs the total concentration of theingredient(s) with unknown acuteiiYESYESwhere:100ATEmixnCiATEC i = concentration ofingredient ii = the individual ingredientfrom 1 to nn = the number of ingredientsiATE mix toDecisionlogic inshould beNOApply the acute toxicity estimate calculation (i.e. when the totalconcentration of ingredients with unknown acute toxicity is> 10%):100 Cumknownif 10% Ci ATEn ATEmixiATE mix toDecision logicin r (category210

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.4 Hazard communication in form of labelling for acute toxicity3.1.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: Table 3.1.3Acute toxicity label elementsClassification Category 1 Category 2 Category 3 Category 4GHS PictogramsSignal Word Danger Danger Danger WarningHazard Statement:H300: Fatal ifH300: Fatal ifH301: Toxic ifH302: Harmful– Oralswallowedswallowedswallowedif swallowed– Dermal H310: Fatal inH310: Fatal inH311: Toxic inH312: Harmfulcontact withcontact withcontact within contact withskinskinskinskin– InhalationH330: Fatal ifH330: Fatal ifH331: Toxic ifH332: Harmful(see Note 1)inhaledinhaledinhaledif inhaledPrecautionary StatementP264P264P264P264Prevention (oral)P270P270P270P270Precautionary StatementP301 + P310P301 + P310P301 + P310P301 + P312Response (oral)P321P321P321P330P330P330P330Precautionary StatementStorage (oral)Precautionary StatementDisposal (oral)P405 P405 P405P501 P501 P501 P501Precautionary StatementP262P262P280P280Prevention (dermal)P264P264P270P270P280P280211

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Precautionary StatementP302 + P350P302 + P350P302 + P352P302 + P352Response (dermal)P310P310P312P312P322P322P322P322P361P361P361P363P363P363P363Precautionary StatementStorage (dermal)Precautionary StatementDisposal (dermal)P405 P405 P405P501 P501 P501 P501Precautionary StatementP260P260P261P261Prevention (inhalation)P271P271P271P271P284P284Precautionary StatementP304 + P340P304 + P340P304 + P340P304 + P340Response (inhalation)P310P310P311P312P320P320P321Precautionary StatementP403 + P233P403 + P233P403 + P233Storage (inhalation)P405P405P405Precautionary StatementDisposal (inhalation)P501 P501 P501Note 1In addition to classification for inhalation toxicity, if data are available that indicates that themechanism of toxicity is corrosivity, the substance or mixture shall also be labelled as EUH071:‘corrosive to the respiratory tract’ — see advice at 3.1.2.3.3. In addition to an appropriate acutetoxicity pictogram, a corrosivity pictogram (used for skin and eye corrosivity) may be addedtogether with the statement ‘corrosive to the respiratory tract’.Note 2In the event that an ingredient without any useable information at all is used in a mixture at aconcentration of 1 % or greater, the mixture shall be labelled with the additional statement that ‘xpercent of the mixture consists of ingredient(s) of unknown toxicity’ — see advice at 3.1.3.6.2.2.EUH071 can also be applied to inhaled corrosive substances not tested for acute inhalationtoxicity according to CLP Annex II, Section 1.2.6If a mixture fulfils the classification criteria with respect to different routes the classificationwill be based on the more severe one. This and other routes have then to be addressed withthe respective hazard statements to CLP Annex I, Table 3.1.3.212

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.4.2 Additional labelling provisionsAnnex I: 3.1.3.6.2.2. In the event that an ingredient without any useable information at all is usedin a mixture at a concentration of 1 % or greater, it is concluded that the mixture cannot beattributed a definitive acute toxicity estimate. In this situation the mixture shall be classified basedon the known ingredients only, with the additional statement that x percent of the mixture consistsof ingredient(s) of unknown toxicity.Though there is no standardised statement with respect to the requirement of CLP Annex I,3.1.3.6.2.2 the following statement would be appropriate, specifying that the information gaprefers only to acute toxicity: “This mixture contains x % of ingredients of unknown acute(…….*) toxicity (* to be specified on a case by case basis if appropriate: oral, dermal,inhalation)”, to be included in the section for supplemental information on the label.Corrosivity:Annex I: 3.1.2.3.3. In addition to classification for inhalation toxicity, if data are available thatindicates that the mechanism of toxicity was corrosivity, the substance or mixture shall also belabelled as ‘corrosive to the respiratory tract’ (see note 1 in 3.1.4.1). Corrosion of the respiratorytract is defined by destruction of the respiratory tract tissue after a single, limited period ofexposure analogous to skin corrosion; this includes destruction of the mucosa. The corrosivityevaluation can be based on expert judgment using such evidence as: human and animal experience,existing (in vitro) data, pH values, information from similar substances or any other pertinent data.In addition to the application of the classification for acute inhalation toxicity, the mixtureshall also be labelled as EUH071 where data are available which indicate that the mode oftoxic action was corrosivity (see Note 1 to Table 3.1.3). Such information can be derived fromdata which warrant classification as corrosive according to the hazard skin corrosion/irritation(see Chapter 3.2). In this case the substance or mixture has to be classified and labelled forskin corrosion with the pictogram for corrosivity, GHS05, hazard statement H314 and alsolabelling with EUH071 (for criteria, see CLP Annex II) is required.Corrosive mixtures may be acutely toxic after inhalation to a varying degree, although this isonly occasionally proved by testing. In case no acute inhalation study is available for acorrosive mixture, it is strongly recommended to apply the precautionary statement P260: Donot breathe dust/fume/gas/mist/vapours/spray.Toxic by eye contact:In cases where a substance or mixture has shown clear signs of severe systemic toxicity ormortality in an eye irritation study a supplemental labelling phrase EUH070 “Toxic by eyecontact“ is required. This additional labelling, based on relevant data, is independent of anyclassification in an acute toxicity category.3.1.5 Re-classification of substances and mixtures classified for acute toxicityaccording to DSD and DPD3.1.5.1 Is direct “translation” of classification and labellingpossible?The CLP allows a minimum classification of substances and mixtures classified according toDSD and DPD, by use of a translation table in Annex VII (Table 1.1) into the correspondingclassification under CLP. For more details see Chapter 1.7 on the application of Annex VII.213

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.5.2 Re-evaluation of dataIf there is new information which might be relevant with respect to classification a reevaluationhas to be performed. Classified gases should be re-evaluated because the guidancevalues changed from general guidance values in mg/l for aerosols, vapours and gases to aspecific guidance value for gases in ppm. Often the values for classification are higheraccording to CLP compared to DSD which may require a re-evaluation on a case by casebasis.3.1.6 Examples of classification for acute toxicityRemark: The classification proposals for the examples refer only to Acute Toxicity.3.1.6.1 Examples of substances fulfilling the criteria forclassification3.1.6.1.1 Example 1: MethanolApplicationAvailableinformationRemarksUse of adequate and reliable human data allowing derivation of an equivalentATE according to CLP Annex I, Table 3.1.1. Animal data not appropriate.Test Data Classification RationaleAnimal data:Oral LD 50 rat ≥ 5000 mg/kgHuman experience:Methanol is known to causelethal intoxications in humans(mostly via ingestion) inrelatively low doses:”…minimal lethal dose in theabsence of medical treatmentis between 300 and 1000mg/kg” (IPCS, EnvironmentalHealth Criteria 196, Methanol,WHO, 1997)Classificationnot possibleCategory 3The rat is known to beinsensitive to the toxicity ofmethanol and is thus notconsidered to be a good modelfor human effects (differenteffect/mode of action)The minimum lethal dosereported of 300 mg/kg is usedas equivalent ATE; accordingto Table 3.1.1 the resultingclassification is Category 3 inTable 3.1.1Test data in rats from mixtures containing methanol should not be used directly inadditivity formula.3.1.6.1.2 Example 2: N,N-DimethylanilineApplicationAvailableinformationUse of qualitative human data and of SAR information with extrapolation to anATE (CLP Annex I, 3.1.3.6.2.1(b) and Table 3.1.2. Animal data are notappropriate.Test Data Classification RationaleAnimal data:Acute dermal toxicity: LD 50values > 1690 mg/kg bwrabbit.Category 4214

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012RemarksHuman experience:Broad human experience,reported in many case reports,demonstrating death fromMetHB following relativelylow oral/dermal/inhalationexposure to aromatic aminessuch as N,N-dimethylaniline.For N,N-Dimethyl -aniline itself no exact humantoxicity values are available.Category 3(oral, dermal,inhalation)The extensive and consistenthuman experience isconsidered to be sufficientlyrobust by expert judgement tobe used for classification intoCategory 3. The rabbit LD 50suggests lower sensitivity toMetHB formation thanhumans which is consistentwith what is known fromother rabbit tests withsubstances known to induceMetHB in humans. The rabbitdata are therefore notconsidered to be adequate foracute toxicity classification.Therefore the human data onthis and structurallyrelated substances are used togive a converted AcuteToxicity point Estimate(cATpE) according to Table3.1.2 for Category 3; e.g.cATpE dermal = 300mg/kgbw, which is thenfalling in a higher categorythan the rabbit data.3.1.6.1.3 Example 3ApplicationAvailableinformationNo exact LD 50 value available. Expert judgement needed.Test Data Classification RationaleCorrosive volatile liquid.Animal data:In a GLP-compliant acuteoral toxicity study in rats, thefollowing results wereobserved:At a test dose of 200 mg/kgbw: no mortality, onlytransient symptoms and nonecropsy findings.At a test dose of 500 mg/kg:100% mortality, symptoms:poor general state; necropsyfindings: hyperemia instomach (due to localirritation /corrosivity), noother organs affectedCategory 4Since at a dose of 200 mg/kgbw no mortality and only slighttransient symptoms withoutnecropsy findings wereobserved, and at 500 mg/kg bwthe high amount/concentrationof the corrosive substancecaused serious effect only atthe site of action and mortality,based on expert judgement itcan be assumed that the likelyLD 50 is > 300 mg/kg bw.Therefore, the Acute ToxicityEstimate (ATE) value forclassification purpose isbetween 300 and 500 mg/kgbw, corresponding to Category4 classification for acutetoxicity.215

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012RemarksLabelling: C (pictogram optional)Additional Hazard statement: EUH071 Corrosive to the respiratory tract3.1.6.1.4 Example 4ApplicationUse of non-standard-guideline test data.Test Data Classification RationaleAvailableinformationAnimal data:A study to evaluate the acutedermal (percutaneous) toxicitywas performed in rabbits. Thefollowing test data resultswere reported:Category 2Rationale for classification:Since the dermal LD 50 isabove 50 mg/kg bw and lessthan 200 mg/kg bw, Category2 classification is warranted(see Table 3.1.2)- At the dose level of 50mg/kg bw: no mortality wasobserved- At 200 mg/kg bw: 100%mortalityTherefore, LD 50 was estimatedto be between 50mg/kg bwand 200mg/kg bwRemarks3.1.6.1.5 Example 5ApplicationUse of Table 3.1.1 and experimentally obtained LC 50 valueTest Data Classification RationaleAvailableinformationA gasAnimal data:A GLP-compliant test foracute inhalation toxicity(gaseous form) wasperformed in accordance withtest guideline 403 in rats. Thefollowing LC 50 wascalculated: LC 50 : 4500ppm/4hCategory 4Rationale for classification:LC 50 = 4500 ppm isconsidered an Acute ToxicityEstimate (ATE) forclassification purposes;according to the classificationcriteria for acute inhalationtoxicity for gases (Table3.1.1), this value correspondsto Category 4. ThereforeCategory 4 Acute InhalationToxicity classification iswarranted.Remarks3.1.6.1.6 Example 6ApplicationTime extrapolation; Note (b) in Table 3.1.1; Haber’s lawTest Data Classification RationaleAvailable Solid substance Category 3 The classification criteria foracute inhalation toxicity in216

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012informationRemarksAnimal data:The acute inhalation toxicitywas studied in rats in a GLPcompliantstudy performed inprinciple according to testguideline 403, but withrespect for transport only with1-h exposure. The LC 50 (1-h)of3 mg/l was calculated.Table 3.1.1 refer to a 4hexposure time; therefore toclassify a substance, existinginhalation toxicity datagenerated from 1-hourexposure should be convertedaccordingly: LC 50 values with1h have to be converted bydividing by 4 (Haber’srule/law, dusts and mists)LC 50 (4-h) = (LC 50 (1-h) : 4) =(3mg/l : 4) = 0.75 mg/l, thusCategory 3 classification iswarranted according to Table3.1.1.3.1.6.1.7 Example 7: 2,3-DichloropropeneApplicationAvailableinformationRemarksDiscrimination from STOT-SETest Data Classification RationaleAnimal data:- Oral LD 50 , rat 250-320mg/kg (assumption: resultsfrom different tests; lowest LD50 is valid)- Inhalation LC 50 rat 2.3mg/l/4h (vapour)Observations: extensive liverand kidney damage followingoral and inhalation exposure tolethal doses ( insufficientinformation)Category 3 oraland Category 3inhalationThe substance is classified for acute toxicity and not for STOT-SE, since theobserved organ toxicity is clearly the cause of the lethality.3.1.6.1.8 Example 8ApplicationAvailableinformationRoute-to-route extrapolation: oral to inhalation (Section 3.1.3.3.4). Expertjudgement.Test DataAnimal data:LD 50 oral rat: 250 mg/kg bw(Category 3)a) No specific kineticinformationExtrapolatedinhalationATE/CATpE0.5 mg/l/4h(cATpE)Rationalea) Using the extrapolationformula 1mg/kgbw = 0.0052217

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Remarksb) Robust kinetic informationallows the conclusion thatonly 50% is absorbed due toan exhalation rate of 50 %.2.6 mg/l/4h(ATE)mg/l/4h:250 x 0.0052 mg/l/4h = 1.3mg/l/4h Category 2according to Table 3.1.2b)Based on the 50% inhalationabsorption rate the equivalentATE would be 2.6 (2 x 1.3) Category 3 according to Table3.1.2Robust kinetic and other information would allow the use of directly derivedATEs in the additivity formulae by expert judgement3.1.6.1.9 Example 9ApplicationAvailableinformationRemarksRoute-to-route extrapolation: oral to dermal (Section 3.1.3.3.4). Expert judgementTest DataAnimal data:LD 50 rat oral: 270 mg/kg bw;100 % oral absorptionassumeda) Assumed dermal absorptionrate: 100%b) Dermal absorption ratebased on robust kinetic/SARinformation: 25%ExtrapolateddermalATE/cATpE300 mg/kgLD 50 dermal1080 mg/kgRationalea) Based on the assumption of100 % dermal absorption theconverted dermal ATE will bederived by using Table 3.1.2 forCategory 3 300 mg kg/bw ascATpE.b) Since dermal absorption isonly 25%, the dermal ATE hasto be accordingly increased 4x270 mg/kg bw = 1080 mg/kgbw. This is regarded as anequivalent ATE which can bedirectly used in the additivityformulae.Robust kinetic and other information would allow the use of directly derived ATEsin the additivity formulae by expert judgement3.1.6.2 Examples of substances not fulfilling the criteria forclassification3.1.6.2.1 Example 10ApplicationAvailableinformationAvailable data are of different quality. Expert judgement. WoETest Data Classification RationaleA liquidAnimal data:Three studies for acuteNoclassificationWith 3 different availablevalues a validity check provedthat the study with LC 50 = 19mg/l is not fully valid in218

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Remarksinhalation toxicity (vapour) inrats are described. Two studieswere performed in accordancewith test guideline 403 andwere GLP-compliant. Onestudy has deficiencies withrespect to study methodologyand description of studyperformance anddocumentation of the testresults; no GLP-compliance.The LC 50 were as follows:– LC 50 : 19 mg/l/4h (no GLP)– LC 50 : 23 mg/l/4h (TG 403,GLP)– LC 50 : 28 mg/l/4h (TG 403,GLP)contrast to the two others;thus in a weight of evidenceapproach it is concluded thatthe LC 50 = ATE > 20 mg/l/4h.The criteria for Category 4 arenot fulfilled.3.1.6.3 Examples of mixtures fulfilling the criteria forclassification3.1.6.3.1 Example 11ApplicationAvailableinformationApplication of the “Relevant ingredient” (Annex I, 3.1.3.3 (a)) and “Genericcut-off values to be taken into account” concepts (Annex I, Table 1.1) formixtures with data gaps using the equation in Annex I, 3.1.3.6.2.3.Test DataAnimal data(oral rat):Classification(ingredient)RationaleIngredient 1(4%)Ingredient 2(92%)Ingredient 3 (3%)Ingredient 4(0.9%)Ingredient 5(0.2%)LD 50 : 125mg/kgNoavailabledataLD 50 : 1500mg/kgNoavailabledataOral Category 3-Oral Category 4LD 50 : 10 mg/kg Oral Category 2-Apply the equation in Annex I,3.1.3.6.2.3:100 (100 92ATE mixC unknownif 10%) ATE4125mix315000.210= 0.032 0.002 0.02 0. 054ATEmix = 148 mg/kg Category 3nCiATERemarks Rationale for classification of the mixture in Category 3:1. Classification via application of substance criteria is not possible since acutetoxicity test data was not provide for the complete mixture (Annex I, 3.1.3.4).2. Classification via the application of bridging principles is not possible sincedata on a similar mixture was not provided (Annex I, 3.1.3.5.1).i219

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123. Classification based on ingredient data for the mixture can be considered(Annex I, 3.1.3.6).4. Applying the “relevant ingredients” concept from Annex I, 3.1.3.3 a) meansthat Ingredient 4 is excluded from the ATE mix calculation since itsconcentration is < 1%. The same reasoning cannot apply to Ingredient 5,though its concentration is below the “relevant ingredients” threshold of 1% butit is higher than the cut-off value of 0.1% for a Category 2 ingredient in AnnexI, Table 1.1.5. The total concentration of ingredients with unknown acute toxicity (i.e.,Ingredient 2) is 92%; therefore, the ATE mix equation in Annex I, 3.1.3.6.2.3must be used. This calculation corrects for relevant ingredients with unknownacute toxicity above 10% of the mixture.6. Ingredients 1, 3 and 5 are included in the ATE mix calculation because theyhave data that fall within a CLP acute toxicity category, Annex I, 3.1.3.6.1 (a).7. Applying the guidance in Note (a) to Table 3.1.1 results in using the actualLD 50 data for Ingredients 1, 3 & 5 in the ATE mix calculation since data isavailable.Additional Labelling: “The mixture contains 92% of ingredients of unknownacute oral toxicity”3.1.6.3.2 Example 12 aApplicationAvailableinformationDifferent phases in inhalation exposure. ExtrapolationTest Data Classification RationaleUse /exposure as aerosol(mist)Animal data (rat):LC 50 (mg/l/4h)Ingredient 1solid (6%)Ingredient 2solid (11%)Ingredient 3solid (10%)Ingredient 4liquid (40 %)Ingredient 5(33%)220Category 4 Conv. ATE (mg/l/4h) =1.5 mg/1/4h0.6 Category 3 ATE = LC 506 (dust) - Neglected, since not classifiedin any acute category.11 (vapour) Category 4 Conv. ATE (mg/l/4h) = 1.5mg/1/4h, assuming identicalcategory for vapour and mistby expert judgement- Water; neglectedRemarks Classification: Category 4No test data available for the whole mixture.Bridging principles not applicable since no test data on similar mixturesavailable.Classification therefore based on ingredients.Use additivity formula in Annex I, 3.1.3.6.1, as information is available for all

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012ingredients.100/ATE mix = 6/1.5+11/0.6+0+40/1.5+0 = 49 ATE mix = 2.04 mg/l/4h Category 4Conclusion: The mixture Example 12a) has to be classified formally in Category 4 withrespect to inhalation toxicity. It is notable that this classification is only derived from thecalculation for the aerosol phase, not for the vapour phase.3.1.6.4 Examples of mixtures not fulfilling the criteria forclassification3.1.6.4.1 Example 12 bApplicationAvailableinformationIngredient 1solid (6%)Ingredient 2solid (11%)Ingredient 3solid (10%)Ingredient 4liquid (40%)Ingredient 5(33%)RemarksDifferent phases in inhalation exposure. ExtrapolationTest Data Classification RationaleUse / exposure as vapourAnimal data (rat):LC 50 (mg/l/4h)Category 4A solid with no sublimation,therefore not present in thevapour phase; neglected.0.6 (dust) Category 3 As Ingredient 16 (dust - Neglected, since not classifiedin any acute category.11 (vapour) Category 4 ATE = LC 50- Water; not relevantClassification: NCInhalation is appropriate route since one hazardous ingredient withappreciable vapour pressure.No test data on the whole mixture.Bridging principles not applicable since no test data on similar mixturesavailable.Classification is therefore based on ingredients.Use additivity formula in Annex I, 3.1.3.6.1 as information is available for allingredients.There is no contributions from ingredients 1 and 2 in the formula since thediluted solid ingredients do not sublime, and thus are not present in the vapourphase; ingredient 3 is in addition not classified in any acute toxicity category.Ingredient 5 does not show acute toxicity.100/ATEmix = 0+0+0+40/11+0 = 3.64ATEmix = 27.527.5 mg/l/4h is above the upper generic concentration limit for vapour NC221

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.1.7 ReferencesDraft OECD TG 39. Draft guidance document on acute inhalation toxicity testing, OECD, 28November 2008.ECETOC TR 86, 2003: European centre for ecotoxicology and Toxicology of Chemicals,Brussels, Belgium, Technical report N°86.Pauluhn, J. (2008) Inhalation toxicology: methodological and regulatory challenges. ExpToxicol Pathol. 60(2-3):111-24.3.2 SKIN CORROSION/IRRITATION3.2.1 Definitions for classification for skin corrosion/irritationAnnex I: 3.2.1.1. Skin Corrosion means the production of irreversible damage to the skin; namely,visible necrosis through the epidermis and into the dermis, following the application of a testsubstance for up to 4 hours. Corrosive reactions are typified by ulcers, bleeding, bloody scabs, and,by the end of observation at 14 days, by discolouration due to blanching of the skin, complete areasof alopecia, and scars. Histopathology shall be considered to evaluate questionable lesions.Skin Irritation means the production of reversible damage to the skin following the application of atest substance for up to 4 hours.3.2.2 Classification of substances for skin corrosion/irritation2223.2.2.1 Identification of hazard information3.2.2.1.1 Identification of human dataCLP Article 7.3 specifies that testing on humans is not allowed for the purposes of CLP;however it does acknowledge that existing data obtained from other sources can be used forclassification purposes.Human data may be retrieved from a number of sources, e.g. epidemiological studies, clinicalstudies, well-documented case reports, poison information units and accident databases oroccupational experience.In this context the quality and relevance of existing human data for hazard assessment shouldbe critically reviewed. There may be a significant level of uncertainty in human data due topoor reporting and lack of specific information on exposure. Diagnosis confirmed by expertphysicians may be missing. Confounding factors may not have been accounted for. Smallgroup sizes may flaw the statistical strength of evidence. Many other factors may compromisethe validity of human data. In clinical studies the selection of individuals for the test and thecontrol groups must be carefully considered. A critical review of the value of human studies isprovided in IR/CSA Section R.4.3.3 and more specific considerations for skincorrosion/irritation are given in IR/CSA Section R.7.2.4.2.Data indicates that human skin is, in most cases, less sensitive than rabbits (ECETOC, 2002).3.2.2.1.2 Identification of non human dataNon human data include physico-chemical properties, results from (Q)SARs and expertsystems, and results from in vitro and in vivo tests. Available skin corrosion/irritation

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012information on substances may include existing data generated by the test methods in the TestMethods Regulation or by methods based on internationally recognised scientific principles.Several of the following non-testing methods and in vitro methods have been validatedagainst the DSD criteria but not against CLP criteria for classification. As the criteria differslightly between DSD and CLP, it should be checked whether the method is sufficientlyvalidated for classification according to CLP.3.2.2.1.2.1 Consideration of physico-chemical propertiesSubstances with oxidising properties can give rise to highly exothermic reactions in contactwith other substances and human tissue. High temperatures thus generated maydamage/destroy biological materials. This applies, for example, to organic peroxides, whichcan be assumed to be skin irritants, unless evidence suggests otherwise (IR/CSA SectionR.7.2.3.1).For a hydro peroxide classification as Skin Corrosive Category 1B should be considered,whereas Skin Irritation Category 2 should be considered for peroxides. Appropriate evidencemust be provided in order to consider non-classification of substances with oxidisingproperties.3.2.2.1.2.2 Non-testing methods: (Q)SARs and expert systemsNon-testing methods such as (Q)SARs and expert systems may be considered on a case-bycasebasis. (Q)SAR systems that also account for skin effects are for example TOPKAT,TerraQSAR, and the BfR-DSS. These systems go beyond the structural similarityconsiderations encompassing also other parameters such as topology, geometry and surfaceproperties. For full guidance consult IR/CSA Sections R.6 and R.7.2.3.1.The BfR-DSS has been recommended in IR/CSA Section R.7.2.4 since there is no othermodel that sufficiently describes the absence of effects. The BfR rules to predict skinirritation and corrosion have been integrated in the internet tool “toxtree”,http://ecb.jrc.ec.europa.eu/qsar.Conclusion on no classification can be made if the (Q)SAR or expert system has been shownto adequately predict the absence of the classified effect (IR/CSA Figure R.7.2-2, footnote f).Since a formal adoption procedure for those non-testing methods is not foreseen and noformal validation process is in place, appropriate documentation is very important. In order toachieve acceptance under REACH the documentation must conform the so-called QSARModel Reporting Format (QMRF). For more details consult the IR/CSA Section R.6.1.3.2.2.1.2.3 Testing-methods: pH and acid/alkaline reserveAnnex I: 3.2.2.2. Likewise, pH extremes like ≤ 2 and ≥ 11,5 may indicate the potential to causeskin effects, especially when buffering capacity is known, although the correlation is not perfect.Generally, such substances are expected to produce significant effects on the skin. If considerationof alkali/acid reserve suggests the substance may not be corrosive despite the low or high pH value,then further testing shall be carried out to confirm this, preferably by use of an appropriatevalidated in vitro test.The acid/alkaline reserve is a measure of the buffering capacity of chemicals. For details ofthe methodology, see Young et al, 1988, and Young and How, 1994.3.2.2.1.2.4 Testing methods: in vitro methodsTable R.7.2-2 in IR/CSA lists the status of validation and regulatory acceptance for in vitrotest methods for skin corrosion and skin irritation.223

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In vitro methods for skin corrosionIn recent years, the OECD has accepted new guidelines for in vitro skin corrosion tests asalternatives for the standard in vivo rabbit skin test (OECD TG 404). Accepted in vitro testsfor skin corrosivity are found in the Test Methods Regulation (TM) and in OECD TestGuidelines (TG):The transcutaneous electrical resistance (TER; using rat skin) test (TM B.40; OECD TG 430)Human skin model (HSM) tests (TM B.40 bis; OECD TG 431)The in vitro membrane barrier test method (OECD TG 435)Positive in vitro results do not generally require further testing and can be used forclassification. Negative in vitro corrosivity responses must be subject to further evaluation.Whereas the TER test and the human skin models at present only allow a classification intoSkin Corrosion Category 1A, the membrane barrier test allows for the differentiation into thethree Categories 1A, 1B and 1C. The applicability domain of the three tests outlined here(TER-, HSM- and membrane barrier test) with regard to the alkalinity and acidity of the testedsubstance should be carefully considered to decide which data are most appropriate for theactual substance.The TER and the HSM assays have been validated for the classification of skin corrosion. Theresults of this validation are well founded, because the CLP criteria for skin corrosion areidentical with the ones referred to in the past validation study.The membrane barrier method has been endorsed as a scientifically validated test for a limitedrange of substances - mainly acids, bases and their derivatives (ECVAM, 2000).In vitro methods for skin irritationThree in vitro skin irritation test methods based on reconstructed human epidermis (RHE)technology are currently under review by the OECD for regulatory acceptance as test methodsable to reliably distinguish non-irritants from irritant substances using one single irritantcategory. The three assays are the EpiSkin TM , the modified EpiDerm TM and the SkinEthicRHE TM test method. The EpiSkin and EpiDerm assays have undergone formal ECVAMvalidation from 2003 – 2007 (Spielmann et al, 2007). In 2007 the EpiSkin was consideredvalid by ESAC as a full replacement test (ECVAM/ESAC, 2007). Originally validated for usein a testing strategy for the identification of positives only (ECVAM/ESAC, 2007), theEpiDerm test methods protocol was subsequently modified. In November 2008, also themodified EpiDerm and the SkinEthic assay were found reliable and relevant test methodscapable of distinguishing non-irritants from irritants and may therefore fully replace thetraditional skin irritation test (ECVAM/ESAC, 2008). It should be noted that conclusions onthe applicability domain of the three methods rest mainly on the optimisation and validationdata set. All three methods are valid for the classification of substances for skin irritancyaccording to CLP criteria (ECVAM/ESAC, 2009).The Skin integrity function test (SIFT) is also listed in IR/CSA, Table R.7.2-2. This test hasonly undergone prevalidation so far and the applicability domain is limited to surfactants.Positive data from SIFT may be used in a weight of evidence approach to considerclassification for irritation, while negative data are not conclusive for a non - classification.Other suitable in vitro methodsPositive data from other suitable in vitro methods may be used in a weight of evidenceapproach to determine classification as irritant, while negative data are not conclusive for anon-classification. In this context 'suitable' means sufficiently well developed according tointernationally agreed development criteria (see REACH Annex XI, section 1.4).224

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.2.2.1.2.5 Testing methods: In vivo dataThe in vivo test in rabbits according to TM B.4 (OECD TG 404) is the standard test for thehazard assessment and classification required under the REACH Annex VIII provisions (10tons per year and more). However it should be noted that according to REACH (Annexes VIIto X) in vivo testing of corrosive substances at concentration/dose levels causing corrosivityshall be avoided.Until 1987 the OECD standard protocol used occlusive patching for the application of the testsubstance, which resulted in more rigorous test conditions compared to the semi-occlusivepatching used today. Especially in borderline cases of classification the method of applicationshould be accounted for in the evaluation of effects.Studies performed according to the USA Federal Hazardous Substances Act (US-FHSA) maybe used for classification purposes although they deviate in their study protocol from theOECD TG 404. They do not include a 48-hour observation time and involve a 24-hour testmaterial exposure followed by observations at 24 hour and 72 hours. Moreover, the testmaterial is patched both on abraded and on intact skin of six rabbits. Studies usually areterminated after 72 hours. In case of no or minimal responses persisting until the 72 hourstime points it is feasible to use such data for classification by calculating the mean values forerythema and oedema on the basis of only the 24 and 72 hours time points. Calculation ofmean scores should normally be restricted to the results obtained from intact skin. In case ofpronounced responses at the 72 hours time point an expert judgement is needed as to whetherthe data is appropriate for classification.Data on skin effects on animals may be available from tests that were conducted for otherprimary purposes than the investigation of skin corrosion / irritation. Such information may begained from acute or repeated dose dermal toxicity studies on rabbits or rats (TM B.3, OECDTG 402; TM B.9, OECD TG 410), guinea pig skin sensitisation studies (TM B.6, OECDguideline 406) and from irritation studies in hairless mice.Annex I: 3.2.2.6. Corrosion3.2.2.2 Classification criteria3.2.2.6.1. On the basis of the results of animal testing a substance is classified as corrosive, asshown in Table 3.2.1. A corrosive substance is a substance that produces destruction of skin tissue,namely, visible necrosis through the epidermis and into the dermis, in at least 1 tested animal afterexposure up to a 4 hour duration. Corrosive reactions are typified by ulcers, bleeding, bloody scabsand, by the end of observation at 14 days, by discoloration due to blanching of the skin, completeareas of alopecia and scars. Histopathology shall be considered to discern questionable lesions.3.2.2.6.2. Three subcategories are provided within the corrosive category: subcategory 1A – whereresponses are noted following up to 3 minutes exposure and up to 1 hour observation; subcategory1B – where responses are described following exposure between 3 minutes and 1 hour andobservations up to 14 days; and subcategory 1C – where responses occur after exposures between 1hour and 4 hours and observations up to 14 days.3.2.2.6.3. The use of human data is discussed in paragraphs 3.2.2.1 and 3.2.2.4 and also inparagraphs 1.1.1.3, 1.1.1.4 and 1.1.1.5.Table 3.2.1Skin Corrosive category and subcategoriesCorrosive in 1 of 3 animals*Corrosive subcategory Exposure Observation225

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Category 1: Corrosive 1A 3 minutes 1 hour3.2.2.7. Irritation2261B > 3 minutes - 1 hour 14 days1C > 1 hour - 4 hours 14 days3.2.2.7.1. Using the results of animal testing a single irritant category (Category 2) is presented inTable 3.2.2. The use of human data is discussed in paragraphs 3.2.2.1 and 3.2.2.4 and also inparagraphs 1.1.1.3, 1.1.1.4 and 1.1.1.5. The major criterion for the irritant category is that at least 2of 3 tested animals have a mean score of ≥ 2,3 - ≤ 4,0.CategoryCategory 2:IrritantTable 3.2.2Skin irritation categoryCriteria(1) Mean value of ≥ 2,3 - ≤ 4,0 for erythema/eschar or for oedema in at least 2of 3 tested animals from gradings at 24, 48 and 72 hours after patch removalor, if reactions are delayed, from grades on 3 consecutive days after theonset of skin reactions; or(2) Inflammation that persists to the end of the observation period normally 14days in at least 2 animals, particularly taking into account alopecia (limitedarea), hyperkeratosis, hyperplasia, and scaling; or(3) In some cases where there is pronounced variability of response amonganimals, with very definite positive effects related to chemical exposure in asingle animal but less than the criteria above.3.2.2.8. Comments on responses obtained in skin irritation tests in animals3.2.2.8.1. Animal irritant responses within a test can be quite variable, as they are with corrosion.The major criterion for classification of a substance as irritant to skin, as shown in paragraph3.2.2.7.1, is the mean value of the scores for either erythema/eschar or oedema calculated in at least2 of 3 tested animals. A separate irritant criterion accommodates cases when there is a significantirritant response but less than the mean score criterion for a positive test. For example, a testmaterial might be designated as an irritant if at least 1 of 3 tested animals shows a very elevatedmean score throughout the study, including lesions persisting at the end of an observation period ofnormally 14 days. Other responses could also fulfil this criterion. However, it should be ascertainedthat the responses are the result of chemical exposure.3.2.2.8.2. Reversibility of skin lesions is another consideration in evaluating irritant responses.When inflammation persists to the end of the observation period in 2 or more test animals, takinginto consideration alopecia (limited area), hyperkeratosis, hyperplasia and scaling, then a materialshall be considered to be an irritant.* Note: In Table 3.2.1 it should read "Corrosive in 1 of 3 animals". There is a misprint in the BG, CS, ET, EL,EN, LV, PT, and RO versions of CLP published in the Official Journal 31.12.2008.Annex I: 3.2.2.4.…3.2.2.3 Evaluation of hazard informationAlthough information might be gained from the evaluation of single parameters within a tier (seeparagraph 3.2.2.5), e.g. caustic alkalis with extreme pH shall be considered as skin corrosives, thereis merit in considering the totality of existing information and making an overall weight of evidencedetermination. This is especially true when there is information available on some but not allparameters. Generally, primary emphasis shall be placed upon existing human experience and data,followed by animal experience and testing data, followed by other sources of information, but case-

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012by-case determinations are necessary.3.2.2.5. A tiered approach to the evaluation of initial information shall be considered, whereapplicable, recognising that all elements may not be relevant in certain cases.3.2.2.3.1 Evaluation of human dataThe usefulness of human data for classification purposes will depend on the extent to whichthe effect, and its magnitude, can be reliably attributed to the substance of interest. Furtherguidance on evaluation of human data for skin corrosion/irritation can be found in IR/CSASection R.7.2.4.2.The criteria in Annex I, Table 3.2.2 are not applicable to human data.3.2.2.3.2 Evaluation of non human data3.2.2.3.2.1 In vitro dataIn evaluation of data from in vitro tests the applicability domain has to be taken into account.The in vitro membrane barrier test method e.g. is mainly applicable for acids and bases and isnot applicable for solutions with pH values between 4.5 and 8.3.2.2.3.2.2 In vivo dataTests in albino rabbits (OECD TG 404)Evaluation criteria for local effects on the skin are severity of the damage and reversibility.For the severity of damage the responses are evaluated according to the Draize score rankingfrom “0” (no response”) up to “4” (severe response”). Evaluation takes place separately forerythema and oedema.Reversibility of skin lesions is the other decisive factor in evaluating responses in the animaltest. The criteria are fulfilled if, for– corrosion– the full thickness of the skin is destroyed resulting in ulcers, bleeding, bloody scabsdiscoloration, complete areas of alopecia and scars. In questionable cases a pathologistshould be consulted. One animal showing this response at the end of the observation periodis sufficient for the classification as corrosive.– irritation– a limited degree of alopecia, hyperkeratosis, hyperplasia and scaling occurs. Twoanimals showing this response are sufficient for the classification as irritant.– very elevated mean scores throughout the study are revealed, including lesionspersisting at the end of an observation period of normally 14 days. One animal showingthis response throughout and at the end of the observation period is sufficient for theclassification as irritant (In cases of suspected corrosives, existing test data may only beavailable for one animal due to testing restrictions, see Example 2.).With regard to severity the main criterion for classification of a substance as irritant to skin, isthe mean score per animal for either erythema/eschar or oedema. During the observationperiod following the removal of the patch each animal is scored on erythema and oedema. Foreach of the three test animals the average scores for three consecutive days (usually 24, 48and 72 hours) are calculated separately for oedema and erythema. If 2/3 animals exceed thecut-off-values defined in the CLP, the classification has to be done accordingly.227

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012With regard to reversibility the test report must prove that these effects are transient i.e. theaffected sites are repaired within the observation period of the test (see Example 1).Non-classification as corrosive can be only justified, if the test was performed with at leastthree animals and the test results were negative for all three animals.Tests that have been conducted with more than three animalsCurrent guidelines foresee a sequential testing of rabbits until a response is confirmed.Typically, up to 3 rabbits may be used. The basis for a positive response is the individualrabbit value averaged over days 1, 2, and 3. The mean score for each individual animal is usedas a criterion for classification. The Skin Irritant Category 2 is used if at least 2 of 3 animalsshow a mean score of 2.3 or above. Other test methods, however, have been using up to 6rabbits. This is also the case for the studies performed according to the US-FSHA.For existing test data with more than three animals, specific provisions need to be applied. Forthe sake of flexibility basically two approaches can be accepted for evaluation: the overall average over all animals will be used (see Example 3a). This has been commonpractice under the DSD. According to the second approach the average score is determined per animal (seeExample 3b). In this case Skin Irritant Category 2 is assigned if 4 of 6 rabbits show amean score of 2.3 or above. Likewise, if the test was performed with 4 or 5 animals, for atleast 3 individuals the mean score must exceed the value of 2.3 to classify as Skin IrritantCategory 2.The more stringent result has to be used if the evaluation according to the method shownunder Example 3a is different to that under Example 3b.Other dermal tests in animalsRelevant data may also be available from animal studies that were conducted for otherprimary purposes than the investigation of skin corrosion/irritation. However, due to thedifferent protocols and the interspecies differences in sensitivity, the use of such data ingeneral needs to be evaluated on a case-by-case basis. These are considered significant if theeffects seen are comparable to those described above. For further guidance how to evaluatedata from studies on dermal toxicity or skin sensitisation, see IR/CSA Figure R.7.2-2footnotes d) and e), respectively.3.2.2.3.3 Weight of evidenceWhere the criteria cannot be applied directly to available identified information, a weight ofthe evidence determination using expert judgement shall be applied in accordance with CLPArticle 9(3).A weight of the evidence determination means that all available and scientifically justifiedinformation bearing on the determination of hazard is considered together, such as physicochemicalparameters (e.g., pH, reserve alkalinity/acidity), information from the application ofthe category approach (grouping, read-across), (Q)SAR results, the results of suitable in vitrotests, relevant animal data, skin irritation information/data on other similar mixtures, humanexperience such as occupational data and data from accident databases, epidemiological andclinical studies and well-documented case reports and observations. The quality andconsistency of the data shall be given appropriate weight. Both positive and negative resultsshall be assembled together in a single weight of evidence determination.228

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Evaluation must be performed on a case-by-case basis and with expert judgement. However,normally positive results that are adequate for classification should not be overruled bynegative findings.Annex I: 1.1.1.4. For the purpose of classification for health hazards (Part 3) established hazardouseffects seen in appropriate animal studies or from human experience that are consistent with thecriteria for classification shall normally justify classification. Where evidence is available fromboth humans and animals and there is a conflict between the findings, the quality and reliability ofthe evidence from both sources shall be evaluated in order to resolve the question of classification.Generally, adequate, reliable and representative data on humans (including epidemiological studies,scientifically valid case studies as specified in this Annex or statistically backed experience) shallhave precedence over other data. However, even well-designed and conducted epidemiologicalstudies may lack a sufficient number of subjects to detect relatively rare but still significant effects,to assess potentially confounding factors. Therefore, positive results from well-conducted animalstudies are not necessarily negated by the lack of positive human experience but require anassessment of the robustness, quality and statistical power of both the human and animal data.For further guidance, if both human and animal data are available, see IR/CSA SectionR.7.2.3.2.3.2.2.4 Decision on classificationWhere the substance is classified as a skin corrosive but the data used for classification doesnot allow differentiation between the skin corrosion subcategories 1A/1B/1C, then thesubstance should be assigned skin corrosive Category 1.3.2.2.5 Setting of specific concentration limitsArticle 10(1) Specific concentration limits and generic concentration limits are limits assigned to asubstance indicating a threshold at or above which the presence of that substance in anothersubstance or in a mixture as an identified impurity, additive or individual constituent leads to theclassification of the substance or mixture as hazardous.Specific concentration limits shall be set by the manufacturer, importer or downstream user whereadequate and reliable scientific information shows that the hazard of a substance is evident when thesubstance is present at a level below the concentrations set for any hazard class in Part 2 of Annex Ior below the generic concentration limits set for any hazard class in Parts 3, 4 and 5 of Annex I.It is more difficult to prove the absence of a hazardous property; the legal text states that:Article 10(1)In exceptional circumstances specific concentration limits may be set by the manufacturer, importeror downstream user where he has adequate, reliable and conclusive scientific information that ahazard of a substance classified as hazardous is not evident at a level above the concentrations setfor the relevant hazard class in Part 2 of Annex I or above the generic concentration limits set for therelevant hazard class in Parts 3, 4 and 5 of that Annex.A specific concentration limit (SCL) set in accordance with the above mentioned provisionsshall take precedence over the generic concentration limit (GCL) set out in Tables 3.2.3 and3.2.4 of Annex I to CLP (Article 10(6)). Furthermore, an SCL is substance-specific andshould be applicable to all mixtures containing the substance, instead of any GCL thatotherwise would apply to a mixture containing the substance.What type of information may be the basis for setting a specific concentration limit?229

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Existing human data may in certain cases (especially if dose-response information isavailable) indicate that the threshold for the irritation hazard in humans for a substance in amixture, would be higher or lower than the GCL. A careful evaluation of the usefulness andthe validity of such human data, as well as their representativeness and predictive value(IR/CSA, sections R.4.3.3. and R.7.2.4.2), should be performed. As pointed out in 1.1.1.4(Annex I to CLP), positive results from well-conducted animal studies are not necessarilynegated by the lack of positive human experience but require an assessment of robustness,quality and a degree of statistical certainty of both the human and animal data.The aim of the standard test method for “Acute Dermal Irritation/Corrosion” TM B.4/OECDTG 404 46 is to identify potential skin corrosion or irritation. The test material is generallyadministered undiluted, thus, no dose-response relationship can be obtained from anindividual test.However, if there are adequate, reliable, relevant and conclusive existing data from otheralready performed animal studies with a sufficient number of animals tested to ensure a highdegree of certainty, and with information on dose-response relationships, such data may beconsidered for setting a lower or, in exceptional cases, a higher SCL on a case-by-case basis.It should be noted that any additional animal testing to set SCLs (of dilutions) of substancesalready classified as a skin corrosive or skin irritant, is strongly discouraged and may onlytake place on a case-by-case basis if there are no alternatives providing adequate reliabilityand quality of data (see CLP Articles 7(1) and 8(1)). The possibilities to use in vitro testmethods are being explored as a basis for setting SCLs, but an accepted common approach isnot yet available. Thus, at the present point in time, it is not possible to provide guidance forthe use of in vitro methods for the purpose of setting SCLs. However, this does not excludethat a method to set SCLs based on in vitro tests could be developed in the future, as theyprovide a promising option for SCL setting.An SCL should apply to any mixture containing the substance instead of the GCL (thatotherwise would apply to the mixture containing the substance). Thus, if the SCL is based ondata derived from tests with dilutions of the substance in a specific solvent, it has to beconsidered that the derived concentration should be applicable to all mixtures for which theSCL should apply.Annex VI Part 3 (Table 3.2) to CLP includes examples of substances for which a higher orlower SCL was set under Directive 67/548/EEC (old DSD system).3.2.2.6 Decision logic for classification of substancesThe decision logic, which is based on IR/CSA Figure R.7.2-2 is revised to meet CLPrequirements. It is strongly recommended that the person responsible for classification, studythe criteria for classification, as well as the guidance above, before and during use of thedecision logic.Step1aIs the substance an organic hydro peroxideor an organic peroxide? YES NOConsider to classify as– corrosive (Skin Corr. 1B) if thesubstance is a hydro peroxide, or– irritating (Skin Irrit. 2) if the substance46 TO NOTE: In OECD TG 404 test substance refers to the test material, test article or test item. The termsubstance may be used differently from the REACH/CLP definition.230

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012is a peroxide.ORProvide evidence for the contrary andproceed to step 1b1b Is the pH of the substance 2 or 11.5?YES NO1cAre there other physical or chemicalproperties that indicate that the substance isirritating / corrosive?YES NO2 Are there adequate existing human datawhich provide evidence that the substance iscorrosive or irritant?YES NOConsider to classify as corrosive.Where classification is based uponconsideration of pH alone (i.e.buffering capacity is not known), SkinCorr. 1A should be applied.Where consideration of alkali/acidreserve suggests that the substance isnot corrosive, this has to be confirmed(preferably by use of an appropriate invitro test). Proceed to step 1cUse this information for weight of evidence(WoE) determination (step 7).Proceed to step 2Classify accordingly.3 Are there data from existing studies onirritation and corrosion in laboratoryanimals, which provide sound conclusiveevidence that the substance is a corrosive,irritant or non-irritant?YES NOClassify accordingly (either Skin Corr.1A/1B/1C or Skin Irrit. 2 or noclassification).4a4bHas the substance proven to be a corrosive,irritant or non-irritant in a suitable acutedermal toxicity test?YES NOHas the substance proven to be a corrosive oran irritant in sensitisation studies or afterrepeated exposure?YES NOIf test conditions are consistent with OECDTG 404, classify accordingly (Skin Corr.1A/ 1B/1C or Skin Irrit. 2 or noclassification)If test conditions are not consistent withOECD TG 404, use this information in theWoE determination (step 7) and proceed tostep 4bClassification cannot be considereddirectly. Use this information for WoEdetermination (step 7).Proceed to step 5a231

Guidance on the Application of the CLP Criteria Version 3.0 - November 20125a5b6a6b6cAre there structurally related substances(suitable “read-across” or grouping), whichare classified as corrosive (Skin Cat. 1) onthe skin, or do suitable (Q)SAR methodsindicate corrosive potential of the substance?YES NOAre there structurally related substances(suitable “read-across” or grouping), whichare classified as irritant on the skin (SkinCat. 2), or do suitable (Q)SAR methodsindicate the presence of irritating potential ofthe substance?YES NOHas the substance demonstrated corrosiveproperties in an OECD adopted in vitro test?YES NOAre there acceptable data from a validated invitro test (adopted by OECD or not), whichprovide evidence that the substance is anirritant or non-irritant?YES NOAre there data from a suitable in vitro test,which provide sound conclusive evidencethat the substance is an irritant? YES NOConsider to classify as Skin Corr. 1.Proceed to step 5bConsider to classify as Skin Irrit. 2.Proceed to step 6aClassify as corrosive. If discriminationbetween Skin Corr. 1A/1B/1C is notpossible, Skin Corr. 1 must be chosen.Consider to classify accordingly (Skin Irrit.2 or no classification).Proceed to step 6cConsider to classify as Skin Irrit. 2Proceed to step 77 Taking all existing and relevant data (steps1-6) into account, is there sufficientinformation to make a decision onclassification?YES NOUnable to classify substance for skincorrosion/irritation232Classify accordingly (Skin Corr. 1A orSkin Corr. 1B or Skin Corr. 1C or SkinIrrit. 2 or no classification)Decision to undertake generation of newtest data should be made in compliancewith REACH and Article 8 of CLP.It is recommended that IR/CSA R.7.2.6should also be considered.

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.2.3 Classification of mixtures for skin corrosion/irritation3.2.3.1 Identification of hazard informationThe procedure for classifying mixtures is a tiered, i.e. a stepwise, approach based on ahierarchy principle and depending on the type and amount of available data/informationstarting from evaluating existing human data on the mixture, followed by a thoroughexamination of the existing in vivo data, physico-chemical properties, and finally in vitro dataavailable on the mixture. For mixtures that have been on the market for a long time, humandata and experience may exist that may provide useful information on the skin irritationpotential of the respective mixtures. See Section 3.2.2.1.1 for further information on theidentification of human data.If valid test data are available for the whole mixture they have precedence. If no such dataexist, the so called bridging principles have to be applied if possible. If the bridging principlesare not applicable an assessment on the basis of data for the components of the mixture willbe applied.Where it is decided to base the classification of a mixture upon consideration of pH alone,Skin corrosion Category 1A should be applied. In this case no further retrieval of informationon the mixture itself is needed.3.2.3.2 Classification criteria3.2.3.2.1 When data are available for the complete mixtureAnnex I: 3.2.3.1.1. The mixture will be classified using the criteria for substances, and taking intoaccount the testing and evaluation strategies to develop data for these hazard classes.3.2.3.1.2. Unlike other hazard classes, there are alternative tests available for skin corrosivity ofcertain types of substances and mixtures that can give an accurate result for classification purposes,as well as being simple and relatively inexpensive to perform. When considering testing of themixture, classifiers are encouraged to use a tiered weight of evidence strategy as included in thecriteria for classification of substances for skin corrosion and irritation (paragraph 3.2.2.5), to helpensure an accurate classification as well as avoid unnecessary animal testing. A mixture isconsidered corrosive to skin (Skin Category 1) if it has a pH of 2 or less or a pH of 11.5 or greater.If consideration of alkali/acid reserve suggests the substance or mixture may not be corrosivedespite the low or high pH value, then further testing shall be carried out to confirm this, preferablyby use of an appropriate validated in vitro test.There are a range of available in vitro test systems that have been validated for theirsuitability in assessing skin corrosion/irritation potential of substances. Some but not all testsystems have been validated for mixtures and not all available in vitro test systems workequally well for all types of mixtures. Prior to testing a mixture in a specific in vitro assay forclassification purposes, it has to be assured that the respective test has been previously shownto be suitable for the prediction of skin corrosion/irritation properties for the type of mixtureto be evaluated.3.2.3.2.1.1 Mixtures with extreme pHAs a general rule, mixtures with a pH of ≤ 2 or ≥ 11.5 should be considered as corrosive.However, assessment of the buffering capacity of the mixture indicated by its acid or alkalireserve should be considered. If the additional consideration of the acid/alkaline reserveaccording to Young et al. (1987, 1994) suggests that classification for corrosion or evenirritation may not be warranted, then further in vitro testing to confirm final (or no)233

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012classification shall be carried out. The consideration of acid/alkali reserve should not be usedalone to exonerate mixtures from classification.Where the mixture has an extreme pH value but the only corrosive/irritant ingredient presentin the mixture is an acid or base with an assigned SCL (either in CLP Annex VI or set bysupplier), then the mixture should be classified according to the SCL. In this instance, pH ofthe mixture should not be considered a second time since it would have already been takeninto account when deriving the SCL for the substance.If this is not the case, then the steps to be taken into consideration when classifying a mixturewith pH 2 or 11.5 are described in the following decision logic:Mixture without in vivo data on skin corrosion or relevant data from similar tested mixtures,pH is 2 or 11.5Does the acid alkaline reserve indicate that themixture may not be corrosive? NO YESIs the mixture tested in an OECD adopted invitro test for skin corrosion?NO YESDoes the mixture demonstrate corrosiveproperties in an OECD adopted in vitro test?YES NOApply methods in Annex I, sections 3.2.3.3.2(Table 3.2.3) / 3.2.3.3.4 (Table 3.2.4) (When validated in vitro skin irritation testmethods are available, these may be used togenerate data to classify the mixture instead ofusing the summation method.)Classify as corrosive, Skin Corr. Cat. 1A.Classify as corrosive, Skin Corr. Cat. 1A.Classify as corrosive. If discrimination betweenSkin Corr. 1A/1B/1C is not possible, Skin Corr. 1must be chosen.Classify accordingly.The mixture must be classified as Skin corrosion Category 1 should the supplier decide not tocarry out the required confirmatory testing.It is also important to note that the pH-acid/alkali reserve to change classification fromcorrosive to irritant or from irritant to not classified assumes that the potential corrosivity orirritancy is due to the effect of the ionic entities. When this is not the case, especially whenthe mixture contains non-ionic (non-ionisable) substances themselves classified as corrosiveor irritant, then the pH-reserve method cannot be a basis for modifying the classification butshould be considered in a weight of evidence analysis.3.2.3.2.2 When data are not available for the complete mixture: bridging principlesAnnex I: 3.2.3.2.1. Where the mixture itself has not been tested to determine its skinirritation/corrosion hazards, but there are sufficient data on the individual ingredients and similartested mixtures to adequately characterise the hazards of the mixture, these data shall be used inaccordance with the bridging rules set out in section 1.1.3.In order to apply bridging principles, there needs to be sufficient data on similar testedmixtures as well as the components of the mixture.234

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012When the available identified information is inappropriate for the application of the bridgingprinciples then the mixture should be classified using the methods described in Section1.6.3.2.3.2.3.2.3 When data are available for all components or only for some components3.2.3.2.3.1 Components that should be taken into account for the purpose ofclassificationAnnex I: 3.2.3.3.1. …..Assumption: the 'relevant ingredients' of a mixture are those which arepresent in concentrations of 1% (w/w for solids, liquids, dusts, mists and vapours and v/v for gases)or greater, unless there is a presumption (e.g., in the case of corrosive ingredients) that an ingredientpresent at a concentration of less than 1% can still be relevant for classifying the mixture for skinirritation/corrosion.3.2.3.2.3.2 The additivity approach is applicableAnnex I: 3.2.3.3.2. In general, the approach to classification of mixtures as irritant or corrosive toskin when data are available on the components, but not on the mixture as a whole, is based on thetheory of additivity, such that each corrosive or irritant component contributes to the overall irritant orcorrosive properties of the mixture in proportion to its potency and concentration. A weighting factorof 10 is used for corrosive components when they are present at a concentration below the genericconcentration limit for classification with Category 1, but are at a concentration that will contribute tothe classification of the mixture as an irritant. The mixture is classified as corrosive or irritant whenthe sum of the concentrations of such components exceeds a concentration limit.3.2.3.3.3. Table 3.2.3 provides the generic concentration limits to be used to determine if the mixtureis considered to be an irritant or a corrosive to the skin.When the supplier is unable to derive the classification using either data on the mixture itselfor bridging principles, he must determine the skin corrosion/irritation properties of themixture using data on the individual ingredients. The supplier must ascertain whether theadditivity approach is applicable, the first step in the process being to identify all theingredients in the mixture (i.e. their name, chemical type, concentration level, hazardclassification and any SCLs) and the pH of the mixture. In addition to for example surfactantinteraction, neutralisation of acids/bases could also occur in a mixture, which also makes itimportant to consider effects of the entire mixture (i.e. pH and the acid/alkaline reserve) ratherthan considering contributions of individual ingredients. Additivity may not apply where themixture contains substances mentioned in Annex I, 3.2.3.3.4, see Section 3.2.3.2.3.3.Application of SCLs when applying the additivity approachThe generic concentration limits (GCLs) are specified in Annex I, Table 3.2.3. However,according to CLP Article 10(5) SCLs take precedence over GCLs. Thus, if a given substancehas a SCL, then this limit has to be taken into account when applying the summation(additivity) method for skin corrosion/irritation (see Examples 5 and 6).In cases where additivity applies for skin corrosion/irritation to a mixture with two or moresubstances some of which may have SCLs assigned, then the following formula should beused:The mixture is classified for skin corrosion/irritation if theSum of (ConcA / clA) + (ConcB / clB) + ….+ (ConcZ / clZ) is 1Where ConcA = the concentration of substance A in the mixture;235

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012clA = the concentration limit (either specific or generic) for substance A;ConcB = the concentration of substance B in the mixture;clB = the concentration limit (either specific or generic) for substance B; etc.This approach is similar to that used in the DPD where a substance SCL replaces the defaultlimits in the conventional method equations.3.2.3.2.3.3 The additivity approach is not applicableAnnex I: 3.2.3.3.4.1. Particular care must be taken when classifying certain types of mixturescontaining substances such as acids and bases, inorganic salts, aldehydes, phenols, and surfactants.The approach explained in paragraphs 3.2.3.3.1 and 3.2.3.3.2 may not be applicable given thatmany of such substances are corrosive or irritant at concentrations < 1%.3.2.3.3.4.2. For mixtures containing strong acids or bases the pH shall be used as a classificationcriterion (see paragraph 3.2.3.1.2) since pH is a better indicator of corrosion than the concentrationlimits of Table 3.2.3.3.2.3.3.4.3. A mixture containing ingredients that are corrosive or irritant to the skin and that cannotbe classified on the basis of the additivity approach (Table 3.2.3), due to chemical characteristicsthat make this approach unworkable, shall be classified as Skin Corrosive Category 1A, 1B or 1C ifit contains ≥ 1% of an ingredient classified in Category 1A, 1B or 1C respectively or as Category 2when it contains ≥ 3% of an irritant ingredient. Classification of mixtures with ingredients forwhich the approach in Table 3.2.3 does not apply is summarised in Table 3.2.4.3.2.3.3.5. On occasion, reliable data may show that the skin corrosion/irritation hazard of aningredient will not be evident when present at a level above the generic concentration limitsmentioned in Tables 3.2.3 and 3.2.4. In these cases the mixture shall be classified according to thatdata (see also Articles 10 and 11). On other occasions, when it is expected that the skincorrosion/irritation hazard of an ingredient is not evident when present at a level above the genericconcentration limits mentioned in Tables 3.2.3 and 3.2.4, testing of the mixture shall be considered.In those cases the tiered weight of evidence strategy shall be applied, as described in paragraph3.2.2.5.3.2.3.3.6. If there are data showing that (an) ingredient(s) is/are corrosive or irritant at aconcentration of < 1 % (corrosive) or < 3 % (irritant), the mixture shall be classified accordingly.2363.2.3.3 Generic concentration limits for substances triggeringclassification of mixtures3.2.3.3.1 When the additivity approach is applicableAnnex I: Table 3.2.3Generic concentration limits of ingredients classified for skin corrosive/irritant hazard(Category 1 or 2)that trigger classification of the mixture as corrosive/irritant to skinSum of ingredients classified as:Skin corrosive Categories 1A, 1B,1CConcentration triggering classification of a mixture as:Skin CorrosiveCategory 1(see note below)Skin IrritantCategory 2 5% 1% but < 5%Skin irritant Category 2 10%(10 x Skin corrosive Category 1A, 10%

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121B, 1C) + Skin irritant Category 2NoteThe sum of all ingredients of a mixture classified as Skin Corrosive Category 1A, 1B or 1Crespectively, shall each be ≥ 5% respectively in order to classify the mixture as either SkinCorrosive Category 1A, 1B or 1C. If the sum of the Skin Corrosive Category 1A ingredients is

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121bAre there other physical or chemical propertiesthat indicate that the mixture iscorrosive/irritating?YES NOUse this information for WoE analysis(step 6).Proceed to step 22 Is there adequate existing human experiencewhich provides evidence that the mixture iscorrosive or irritant?YES NO3 Are there data from existing studies onirritation and corrosion in laboratory animals,which provide sound conclusive evidence thatthe mixture is corrosive, irritant or non-irritant?YES NOClassify accordingly (Skin Corr. 1 or SkinIrrit. 2).Classify accordingly (Skin Corr. 1A orSkin Corr. 1B or Skin Corr. 1C or SkinIrrit. 2 or no classification).4a4b5a5b5c238Has the mixture proven to be a corrosive,irritant or non-irritant in a suitable acute dermaltoxicity test?YES NOHas the mixture proven to be a corrosive or anirritant in sensitisation studies or after repeatedexposure?YES NOHas the mixture demonstrated corrosiveproperties in an OECD adopted in vitro test?YES NOAre there acceptable data from a validated invitro test (adopted by OECD or not), whichprovide evidence that the mixture is an irritantor non-irritant?YES NOAre there data from a suitable in vitro test,which provide sound conclusive evidence thatthe mixture is an irritant?YES – If test conditions are consistent withOECD TG 404, classify accordingly(Skin Corr. 1A/1B/1C or Skin Irrit. 2or no classification).– If test conditions are not consistentwith OECD TG 404, use thisinformation in the WoE determination(step 6) and proceed to step 4bClassification cannot be considereddirectly. Use this information for WoEdetermination (step 6).Proceed to step 5aClassify as corrosive. If discriminationbetween Skin Corr. 1A/1B/1C is notpossible, Skin Corr. 1 must be chosen.Consider to classify accordingly (SkinIrrit. 2 or no classification).Proceed to step 5cConsider to classify as Skin Irrit. 2.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012NO6 Taking all existing and relevant data (steps 1-5)into account including potentialsynergistic/antagonistic effects andbioavailability, is there sufficient information tomake a decision on classification? YES NOProceed to step 6Classify accordingly (Skin Corr. 1A orSkin Corr. 1B or Skin Corr. 1C or SkinIrrit. 2 or no classification)2. When data are not available for the complete mixture: bridging principles7aAre existing sufficient skin corrosion/irritationdata available on similar tested mixtures and onthe individual ingredients?NO YESProceed to step 87b Can bridging principles be applied? YES NOClassify in appropriate category (SkinCorr. 1A or Skin Corr. 1B or Skin Corr.1C or Skin Irrit. 2 or no classification)3. When data are available for all components or only for some components of the mixture8a Is pH of the mixture 2 or 11.5? YES 8bNOIs there any indication that the additivityprinciple does not apply?YES NOFollow decision logic in Section3.2.3.2.1.1 and classify accordingly.Annex I, section. 3.2.3.3.4 and Table3.2.4 may apply. Take into accountrelevant ingredients (Annex I, 3.2.3.3.1.and SCLs as appropriate.Classify in appropriate category (SkinCorr. 1A/1B/1C or Skin Irrit. 2 or noclassification)Annex I, section 3.2.3.3.2 and Table 3.2.3applies. Take into account relevant ingredients(Annex I, 3.2.3.3.1. and SCLs as appropriate.Classify in appropriate category (Skin Corr.1A/1B/1C or Skin Irrit. 2 or no classification)Where the mixture is classified ascorrosive but the data used forclassification does not allowdifferentiation between the skin corrosionsubcategories 1A/1B/1C, then the mixtureshould be assigned Skin corrosionCategory 1.3.2.4 Hazard communication in form of labelling for skin corrosion/irritation3.2.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 3.2.4.1. Label elements shall be used for substances or mixtures meeting the criteria forclassification in this hazard class in accordance with Table 3.2.5.239

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Table3.2.5Label elements for skin corrosion/irritationClassification Category 1A / 1B / 1C Category 2GHS PictogramsSignal Word Danger WarningHazard StatementPrecautionary StatementPreventionH314: Causes severe skinburns and eye damageP260P264P280H315: Causes skin irritationP264P280Precautionary StatementResponsePrecautionary StatementStoragePrecautionary StatementDisposalP301 + P330 + P331P303 + P361 + P353P363P304 + P340P310P321P305 + P351 + P338P405P501P302 + P352P321P332 + P313P3623.2.4.2 Additional labelling provisionsAnnex II: 1.2.6. EUH071 — Corrosive to the respiratory tractFor substances and mixtures in addition to classification for inhalation toxicity, if data are availablethat indicate that the mechanism of toxicity is corrosivity, in accordance with section 3.1.2.3.3 andNote 1 of Table 3.1.3 in Annex I.For substances and mixtures in addition to classification for skin corrosivity, if no acute inhalationtest data are available and which may be inhaled.Corrosive substances (and mixtures) may be acutely toxic after inhalation to a varying degree,which is only occasionally proved by testing. In case no acute inhalation study is available fora corrosive substance (or mixture) and such substance (or mixture) may be inhaled, a hazardof respiratory tract corrosion may exist. As a consequence, substances and mixtures have tobe supplementary labelled with EUH071. Moreover, in such a case it is stronglyrecommended to apply the precautionary statement P260: “Do not breathedust/fume/gas/mist/vapours/spray.”Annex II: 1.2.4. EUH066 — Repeated exposure may cause skin dryness or crackingFor substances and mixtures which may cause concern as a result of skin dryness, flaking orcracking but which do not meet the criteria for skin irritancy in section 3.2 of Annex I, based oneither:— practical observations; or— relevant evidence concerning their predicted effects on the skin.240

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.2.5 Re-classification of substances and mixtures classified for skincorrosion/irritation according to DSD and DPD3.2.5.1 Is direct “translation” of classification and labellingpossible?A direct translation as indicated in the translation table in Annex VII to CLP is generallypossible. Translation from classification according to DSD or DPD to the classificationaccording to CLP is as follows:– C; R35 is translated into Skin Corr. 1A; H314. The criteria in CLP and in DSD areidentical.– C; R34 is translated into Skin Corr. 1B; H 314 with the following note:Annex VII: Table 1.1Note 2It is recommended to classify in Category 1B even if it also could be possible that 1C could beapplicable for certain cases. Going back to original data, may not result in a possibility todistinguish between Category 1B or 1C, since the exposure period has normally been up to 4 hoursaccording to Regulation (EC) No 440/2008. However, for the future, when data are derived fromtests following a sequential approach as foreseen in the Regulation (EC) No 440/2008, Category 1Cshould be considered.– Xi; R38 is translated into Skin Irrit. 2; H315. The criteria in CLP and DSD are almostidentical.It should be noted that where mixtures containing substances with risk phrase R34 have beenclassified on basis of the hazards of individual ingredients, the use of the translation table maylead to an under-classification of the mixture. This is because the general concentration limits,to be applied for mixtures, are lowered under CLP compared to DPD. For mixtures containingsubstances with this classification the use of the translation table may therefore not beappropriate and re-classification done by using the existing data would be more correct. Formore details see Chapter 1.7.3.2.5.2 Re-evaluation of dataIf there is new information which might be relevant with respect to classification a reevaluationhas to be performed.3.2.6 Examples of classification for skin corrosion/irritation3.2.6.1 Examples of substances fulfilling the criteria forclassification3.2.6.1.1 Example 1: Standard test according to OECD TG 404 with three animalsIn a guideline test according to OECD TG 404 the test substance was applied for three minutesand 1 hour. No scars or other irreversible effects were found. The scoring results obtained after 4hours application time are listed in the following table:AnimalNr.Degree of erythema after…[observation time]Degree of oedema after …[observationtime] 24/48/72 h2.3 ?1h 24h 48h72h7d 14d 1h 24h 48h72h7d 14d ErythemaOedema241

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121 3 3 3 2 0 1 2 2 2 0 Yes No 24/48/72 h =2.7 24/48/72 h =2.0=>”positiveResponder”2 3 3 3 3 0 1 2 2 1 0 Yes No 24/48/72 h =3 24/48/72 h =1.7=>”positiveResponder”3 1 1 1 0 0 1 1 1 1 0 No No 24/48/72 h =0,66 24/48/72 h = 1Classification: Skin Irritant Category 2The classification is made on basis of 2/3 "positive responder" exceeding 2.3 mean score forerythema.3.2.6.1.2 Example 2: Test carried out with one animal with a test substance which issuspected as corrosiveDue to the unprecedented structure the biological effects of the substance cannot beanticipated. Therefore, the test according to OECD TG 404 was started with one animal onlyin line with testing restrictions. Exposure times were 3 min and 1h. The followingscores/effects were observed:ExposuretimeDegree of erythema after……[observation time]Degree of oedema after……[observation time]Visiblenecrosis,irreversibleskin damage1h 24h 48h 72h ... 1h 24h 48h 72h ... After 14d3 min 0 0 0 0 0 0 0 0 No1h 0 1 2 3 0 2 2 3 YesClassification: Skin Corrosion Category 1BRationale for the classification is destruction of the tissue within 1 hour exposure.3.2.6.1.3 Example 3a: Test carried out with more than three animalsA substance was tested on acute skin irritation / corrosion according to OECD TG 404.Contact time was 4 hours. No effects were seen after a contact time of 3 min and one hour.The following scores were obtained:AnimalNrDegree of erythema after …[observationtime]Degree of oedema after …[observationtime]1h 24h 48h 72h 7d 14d 1h 24h 48h 72h 7d 14d1 3 3 2 2 1 0 2 3 2 2 1 02 3 2 2 2 1 0 2 2 2 2 1 0242

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123 2 2 1 1 1 0 2 2 2 2 1 04 2 2 1 1 1 0 2 2 2 2 1 0Evaluation was made based on the arithmetic mean of all animals.The arithmetic mean after 24/48/72 hours for erythema M E = 21:12 = 1.8; and for oedema M O= 25:12 = 2.1. Both values are below 2.3, i.e. no classification warranted for skin irritation.3.2.6.1.4 Example 3b: Test carried out with more than three animalsA substance was tested on acute skin irritation / corrosion according to OECD TG 404. Contacttime was 4 hours. No effects were seen after a contact time of 3 min and one hour. The followingscores were obtained after a contact time of 4 hours:Observation time1h 24h 48h 72h 7d 14d 1h 24h 48h 72h 7d 14d Pos responderAnimalNrErythema Oedema ErythemaOedema1 3 3 2 2 1 0 2 3 2 2 1 0 Yes Yes2 3 2 2 2 1 0 2 2 2 2 1 0 No No3 2 2 1 1 1 0 2 2 2 2 1 0 No No4 2 2 1 1 1 0 2 2 2 2 1 0 No NoEvaluation was made based on the average score per animal.Only 1/4 of the animals reached the cut-off value of 2.3, i.e. only animal No 1 is a positiveresponder. No classification is warranted with regard to skin irritation.3.2.6.2 Examples of mixtures fulfilling the criteria forclassificationWhere the mixture is made up of ingredients with no assigned SCLs, then the appropriatesummation(s) and generic concentration limits from CLP Annex I, Table 3.2.3 should beused.3.2.6.2.1 Example 4Ingredient Skin corrosion /irritation classificationConcentration(% w/w)Surfactant A Skin Cat 2 1,8 Not assignedSubstance B Not classified 0,5Substance C Skin Cat 2 5,4 Not assignedSubstance D Not classified 4Acid Skin Cat 1A 2 Not assignedWater Not classified 86.3pH of the mixtureis is 9.0 – 10.0, thus extreme pH provisions do not apply. The mixturecontains a surfactant and an acid but neither are corrosive/irritant below 1% (as identified bythe absence of SCLs in either CLP Annex VI or the Classification and Labelling Inventory).Additivity is considered to apply.SCL243

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Substance B, substance D and water can be disregarded as they are not classified for skincorrosion/irritation.The mixture contains 2% acid, the only ingredient classified as Skin Corr. Cat 1. As this isbelow the 5% GCL, the mixture is not classified Skin Corr. Cat. 1 but is classified Skin Irrit.Cat. 2 ( 1% < 5%).3.2.6.2.2 Example 5244IngredientSkin corrosion / irritationclassificationConcentration(% w/w)Surfactant A Skin Cat 2 3,8 Not assignedSubstance B Not classified 0,5Base E Skin Cat 1B 5,4 C ≥ 10 %: Skin Cat 1B5 % ≤ C < 10 %: Skin Cat 2Substance D Not classified 4Substance F Skin Cat 1B 2 Not assignedWater Not classified 84.3pH of the mixture is 10.5 – 11.0, thus extreme pH provisions do not apply. The mixturecontains a surfactant and a base but none are corrosive/irritant below 1% (as identified byabsence of specific concentration limits in either CLP Annex VI or the Classification andLabelling Inventory). Additivity is considered to apply.Substance B, substances D and water can be disregarded as they are not classified for skincorrosion/irritation.SCLs are neither assigned to substance F nor surfactant A, thus GCLs apply for theseingredients. SCLs are assigned to Base E (see section 3.2.3.2.3.2 under Application of SCLswhen applying the additivity approach).Skin Cat 1:(% substance F/GCL) + (% base E/SCL) = (2/5) + (5.4/10) = 0.94 < 1, thus mixture is notclassified as Skin Corr. Cat 1Skin Cat 2:(% substance F/GCL) + (% base E/SCL) + (% surfactant A/GCL) = (2/1) + (5.4/5) + (3.8/10)= 3.46 which is > 1, thus the mixture is classified Skin Irrit. Cat. 23.2.6.3 Examples of mixtures not fulfilling the criteria forclassification3.2.6.3.1 Example 6IngredientSkin corrosion / irritationclassificationConcentration(% w/w)Surfactant C Skin Cat 2 0,4 Not assignedSurfactant G Skin Cat 2 3.0 Not assignedSurfactant A Skin Cat 2 0,7 Not assignedSubstance H Skin Cat 1A 3,0 C ≥ 70 %: Skin Cat 1A50 % ≤ C < 70 %: Skin Cat 1BSCLSCL

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Substance D Not classified 2Water Not classified 90.935 % ≤ C < 50 %: Skin Cat 2pH of the mixture is: 2.5 – 3.0, thus extreme pH provisions do not apply. The mixturecontains three surfactants but none are corrosive/irritant below 1% (as identified by theabsence of specific concentration limits in either CLP Annex VI or the Classification andLabelling Inventory) Additivity is considered to apply.Substance D and water can be disregarded as they are not classified for skincorrosion/irritation. Also surfactant C and surfactant A can be disregarded as both are presentbelow 1%.A SCL is not assigned to surfactant G, thus GCL apply for this ingredient.Skin Cat 1:The mixture contains 3% substance H, the only ingredient classified as Skin Corr. Cat. 1. Asthis is below the 50% SCL for substance H, the mixture is not classified as Skin Corr. Cat. 1.Skin Cat 2:(% substance H/SCL) + (% surfactant G/GCL) = (3/35) + (3/10) = 0.39 which is < 1, thus themixture is not classified Skin Irrit. Cat. 2.3.2.7 ReferencesECETOC (2002), Use of human data in hazard classification for irritation and sensitisation,Monograph No 32, Brussels ISSN 0773-6374-32ECVAM/ESAC (2007) Statement on the validity of in-vitro tests for skin irritation. Online:http://ecvam.jrc.it/ECVAM/ESAC (2008) Statement on the validity of in-vitro tests for skin irritation. Online:http://ecvam.jrc.it/ECVAM/ESAC (2009) Statement on the performance under UN GHS of three in-vitro assaysfor skin irritation testing and the adaptation of the reference chemicals and defined accuracyvalues of the ECVAM skin irritation performance standards. Online: http://ecvam.jrc.it/Spielmann, H., Hoffmann, S., Liebsch, M., Botham, P., Fentem, J., Eskes, C., Roguet, R.,Cotovió, J., Cole, T., Worth, A., Heylings, J., Jones, P., Robles, C., Kandárová, H., Gamer,A., Remmele, M., Curren, R., Raabe, H., Cockshott, A., Gerner, I. and Zuang, V. (2007) TheECVAM International Validation Study on In Vitro Tests for Acute Skin Irritation: Report onthe Validity of the EPISKIN and EpiDerm Assays and on the Skin Integrity Function Test.ATLA 35, 559-601.Young J.R., How M.J., Walker A.P., Worth W.M.H. (1988): Classification as corrosive orirritant to skin of preparations containing acidic or alkaline substances, without test onanimals. Toxicology in Vitro 2, 19-26.Young J.R., How M.J. (1994): Product classification as corrosive or irritant by measuring pHand acid / alkali reserve. In Alternative Methods in Toxicology vol. 10 - In Vitro SkinToxicology: Irritation, Phototoxicity, Sensitization, eds. A.Rougier, A.M. Goldberg and H.IMaibach, Mary Ann Liebert, Inc. 23-27.245

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.3 SERIOUS EYE DAMAGE/EYE IRRITATIONIt should be noted that if a substance or mixture is classified as Skin corrosive Category 1then serious damage to eyes is implicit and there is no need to proceed with classification foreye effects.3.3.1 Definitions for classification for serious eye damage/eye irritationAnnex I: 3.3.1.1. Serious eye damage means the production of tissue damage in the eye, or seriousphysical decay of vision, following application of a test substance to the anterior surface of the eye,which is not fully reversible within 21 days of application.Eye irritation means the production of changes in the eye following the application of test substanceto the anterior surface of the eye, which are fully reversible within 21 days of application.3.3.2 Classification of substances for serious eye damage/eye irritation2463.3.2.1 Identification of hazard information3.3.2.1.1 Identification of human dataExisting data on eye effects in humans may include well-documented epidemiological studies,clinical studies, case reports, and data from poison information units and accident databasesor occupational experience. Their quality and relevance for hazard assessment should bethoroughly reviewed. A critical review of the value of human studies is provided in IR/CSASection R.4.3.3 and more specific considerations for eye damage/irritation are given inIR/CSA Section R.7.2.4.2.3.3.2.1.2 Identification of non human dataAvailable serious eye damage/eye irritation information on substances may include existingdata generated by the test methods in the Test Methods Regulation or by methods based oninternationally recognised scientific principles.Several of the following non-testing and in vitro methods have been validated against theDSD criteria but not against the CLP criteria for classification. Therefore it should be checkedwhether the method is sufficiently validated for classification according to CLP.3.3.2.1.2.1 Consideration of physico-chemical propertiesSubstances with oxidising properties can give rise to highly exothermic reactions in contactwith other substances and human tissue. High temperatures thus generated maydamage/destroy biological materials. This applies, for example, to organic peroxides, whichcan be assumed to be eye irritants, unless evidence suggests otherwise (IR/CSA SectionR.7.2.3.1).For a hydro peroxide classification as Eye Damage Category 1 should be considered, whereasEye Irritation Category 2 should be considered for peroxides. Appropriate evidence must beprovided in order to consider non-classification of substances with oxidising properties.3.3.2.1.2.2 Non-testing methods: (Q)SARs and expert systemsNon-testing methods such as (Q)SARs and expert systems may be considered on a case-bycasebasis. (Q)SARs are in general not very specific for eye irritancy. In many cases rules areused in a similar manner to those used for skin irritation and corrosion. (Q)SAR systems thatalso account for eye effects are for example TOPKAT, Derek for Windows, and SICRET. For

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012full guidance, consult the IR/CSA Section R.6 (“QSAR and grouping of chemicals”), inwhich also the many shortcomings of the existing systems are discussed.Since a formal adoption procedure for those non-testing methods is not foreseen and noformal validation process is in place, appropriate documentation is crucial. In order to achieveacceptance under REACH, the documentation must conform to the so-called QSAR ModelReporting Format (QMRF). For more details consult the IR/CSA Section R.6.1.3.3.2.1.2.3 Testing-methods: pH and the acid/alkaline reserveAnnex I: 3.3.2.3. ….Likewise, pH extremes like ≤ 2 and ≥ 11,5 may produce serious eye damage,especially when associated with significant buffering capacity. Such substances are expected toproduce significant effects on the eyes. Possible skin corrosion has to be evaluated prior toconsideration of serious eye damage/eye irritation in order to avoid testing for local effects on eyeswith skin corrosive substances…Substances can be predicted to be corrosive, if the pH is 2 or 11.5. Where extreme pH isthe only basis for classification as serious eye damage, it is important to take intoconsideration the acid/alkaline reserve, a measure of the buffering capacity (Young et al,1988, and Young and How, 1994). However, lack of buffering capacity should not be usedalone to exonerate from classification as corrosive.If pH is < 3.2 or > 8.6, then consider the substance for severe eye damage/eye irritation(IR/CSA Section R.7.2.4.1). Further information and/or reasoning is needed to concludewhether the substance is causing severe eye damage or eye irritation. This model is notrecommended for the stand-alone discrimination between eye irritants and non-irritants.However, it could be used in the context of a tiered testing strategy to identify eye irritants(due to its very low false positive rate) but not for non-irritants (due to its relatively high falsenegative rate).3.3.2.1.2.4 Testing methods: in vitro methodsThere are no OECD adopted in-vitro/ex-vivo tests for serious eye damage/eye irritation atpresent. However, there is regulatory acceptance in the EU that a substance can be considereda severe eye irritant (Serious eye damage Category 1) based on positive results in the IsolatedChicken Eye (ICE) test, the Bovine Corneal Opacity and Permeability (BCOP) test, theIsolated Rabbit Eye (IRE) test or the Hen's Egg Test on Chorio-allantoic Membrane (HET-CAM) test. Negative in vitro corrosivity responses in these tests must be followed by furthertesting (IR/CSA Section R.7.2.4.1)There are no in vitro tests with regulatory acceptance for eye irritation at present, but the twohuman corneal epithelium models, EpiOcular and SkinEthic , have been submitted toECVAM for validation.3.3.2.1.2.5 Testing methods: In vivo dataTesting for eye irritation would not be carried out on substances known or predicted to becorrosive to skin. Such substances are automatically considered to be severely damaging tothe eye. A parallel classification with serious eye damage in addition to skin corrosion is notrequired.The in vivo test in rabbits according to OECD TG 405 (B.5 in the Test Methods Regulation) isthe standard test for the hazard assessment under the REACH.The Low Volume Eye Test (LVET; Griffith et al 1980) is a modification of the standardOECD TG 405 test method, the differences being:247

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012- the test material is placed directly on the cornea instead of introducing it in theconjunctival sac inside the lower lid;- a reduction in the volume of test material applied (0.01 ml (or corresponding weight forsolids) compared with the standard 0.1 ml).Data from the LVET should be considered but must be carefully evaluated. The applicabilitydomain up to now is limited to detergent and cleaning products. It is stated that positive dataare a trigger for appropriate classification, but that negative data are not conclusive for a nonclassification(IR/CSA R.7.2.4.1). However, they should be considered in a weight ofevidence determination.3.3.2.2 Classification criteriaAnnex I: 3.3.2.6. Irreversible effects on the eye/serious damage to eyes (Category 1)3.3.2.6.1. Substances that have the potential to seriously damage the eyes are classified in Category1 (irreversible effects on the eye). Substances are classified in this hazard category on the basis ofthe results of animal testing, in accordance with the criteria listed in Table 3.3.1. Theseobservations include animals with grade 4 cornea lesions and other severe reactions (e.g.,destruction of cornea) observed at any time during the test, as well as persistent corneal opacity,discoloration of the cornea by a dye substance, adhesion, pannus, and interference with the functionof the iris or other effects that impair sight. In this context, persistent lesions are considered thosewhich are not fully reversible within an observation period of normally 21 days. Substances are alsoclassified in Category 1 if they fulfil the criteria of corneal opacity ≥ 3 or iritis > 1,5 detected in aDraize eye test with rabbits, recognising that such severe lesions usually do not reverse within a 21days observation period.Table 3.3.1Category for irreversible eye effectsCategoryIrreversibleeffects on theeye(Category 1)CriteriaIf, when applied to the eye of an animal, a substance produces:– at least in one animal effects on the cornea, iris or conjunctiva that are notexpected to reverse or have not fully reversed within an observation period ofnormally 21 days;and/or– at least in 2 of 3 tested animals, a positive response of:– corneal opacity ≥ 3 and/or– iritis > 1.5calculated as the mean scores following grading at 24, 48 and 72 hours afterinstallation of the test material.Annex I: 3.3.2.7. Reversible effects on the eye (Category 2)3.3.2.7.1. Substances that have the potential to induce reversible eye irritation are classified inCategory 2 (irritating to eyes).Table 3.3 2Category for reversible eye effects248CategoryIrritating to eyes(Category 2)Criteriaif, when applied to the eye of an animal, a substanceproduces:– at least in 2 of 3 tested animals, a positive response of:– corneal opacity ≥ 1 and/or– iritis ≥ 1, and/or– conjunctival redness ≥ 2 and/or

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012– conjunctival oedema (chemosis) ≥ 2– calculated as the mean scores following grading at 24, 48and 72 hours after installation of the test material, andwhich fully reverses within an observation period of 21days3.3.2.7.2. For those substances where there is pronounced variability among animal responses, thisinformation shall be taken into account in determining the classificationThe classification criteria apply to the results of the OECD TG 405 and to the results of theLVET. Negative data from the LVET are not conclusive for non-classification, but should beconsidered in a weight of evidence determination.3.3.2.3 Evaluation of hazard informationAnnex I: 3.3.2.5. A tiered approach to the evaluation of initial information shall be consideredwhere applicable, while recognising that all elements may not be relevant in certain cases.3.3.2.4. ….Although information may be gained from the evaluation of single parameters within atier (e.g. caustic alkalis with extreme pH shall be considered as local corrosives), the totality ofexisting information shall be considered in making an overall weight of evidence determination,particularly when there is information available on some but not all parameters. Generally, primaryemphasis shall be placed upon expert judgement, considering human experience with the substance,followed by the outcome of skin irritation testing and of well-validated alternative methods.3.3.2.3.1 Evaluation of human dataQuality data on substance-induced eye irritation in humans are likely to be rare. Where humandata are available, the usefulness of such data for classification purposes will depend on theextent to which the effect, and its magnitude, can be reliably attributed to the substance ofinterest. The quality and relevance of such data for hazard assessment should be criticallyreviewed.If a substance is diagnostically confirmed by a physician to be the cause for decay in visionwith the effects not being transient but persistent this should lead to the most serious eyeclassification, i.e. Eye Damage Category 1.Further information on the evaluation of human data for eye irritation can be found in IR/CSASection R7.2.4.2.3.3.2.3.2 Evaluation of non-human dataThe results of the non-testing methods fulfilling the criteria of REACH Annex XI paragraphs1.3 and 1.5 should be used instead of testing or as part of the weight of evidence approach.3.3.2.3.2.1 In-vitro dataOnly positive results in the BCOP, ICE, IRE and HET-CAM in vitro assays can be used forclassification as severe eye irritants. Negative results are not conclusive for a nonclassification.There are currently no validated in vitro eye irritation test methods available. However, tworeconstituted human tissue models (the EpiOcular TM and SkinEthic TM HCE models) areundergoing formal validation.3.3.2.3.2.2 In-vivo dataTests in albino rabbits (OECD TG 405)Evaluation criteria for local effects on the eye are severity of the damage and reversibility.249

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012For the severity of damage the degree of inflammation is assessed. Responses are gradedaccording to the grading of ocular lesions in OECD TG 405.Evaluation takes place separately for cornea, iris and conjunctiva (erythema and swelling). Ifthe scoring meets the criteria in Annex I, Tables 3.3.1 / 3.3.2, the substances are classified asCategory 1 for serious eye damage or Category 2 for eye irritation, respectively.Reversibility of eye lesions is the other decisive factor in evaluating responses in the animaltest. If the effects are not transient within the observation time of 21 days but cause persistentdamage, they are considered irreversible and the test substance needs to be classified intoCategory 1. In the case of studies with a shorter observation period with irreversible effects,classification based on expert judgement should be considered.With regard to reversibility the test report must prove that these effects are transient, i.e. theaffected sites are repaired within the observation period of the test (see Example 1).Evaluation of reversibility or irreversibility of the observed effects does not need to exceed 21days after instillation for the purpose of classification.According to OECD TG 405, in cases of suspected serious eye damage, the test is started withone animal only. If effects in this animal are irreversible until the end of the observationperiod, sufficient information is available to classify the substance for serious eye damage.For a decision on no classification for serious eye damage and/or irritation or for a decision onclassification as irritant, two additional animals have to be tested.For each of the three test animals the average scores for three consecutive days (usually 24, 48and 72 hours) are calculated separately for the cornea, iris and conjunctiva (erythema andswelling). If the mean scores for 2 out of 3 animals exceed the values in Tables 3.3.1 / 3.3.2,classification has to be assigned accordingly.Tests that have been conducted with more than three animalsOlder test methods, however, have been using up to six rabbits. The CLP does not providecriteria for the evaluation of such studies. The current US EPA/UN Recommendation may beconsidered (see Example 2):In case of 6 rabbits the following applies:Classification as serious eye damage – Category 1 if at least in one animal effects on the cornea, irisor conjunctiva that are not expected to reverse or have not fully reversed within an observation periodof normally 21 days; and/orat least 4 out of 6 rabbits show a mean score of 3 for the cornea and/or 1.5 for the irisClassification as eye irritation – Category 2 if at least 4 out of 6 rabbits show a mean score of 1 for the cornea and/or 1 for the iris and/or 2 conjunctival erythema and/or 2 conjunctival swellingIn case of 5 rabbits the following applies:Classification as serious eye damage – Category 1 if at least in one animal effects on the cornea, irisor conjunctiva that are not expected to reverse or have not fully reversed within an observation periodof normally 21 days; and/or250

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012at least 3 out of 5 rabbits show a mean score of 3 for the cornea and/or 1.5 for the irisClassification as eye irritation – Category 2 if at least 3 out of 5 rabbits show a mean score of 1 for the cornea and/or 1 for the iris and/or 2 conjunctival erythema and/or 2 conjunctival swellingIn case of 4 rabbits the following applies:Classification as serious eye damage – Category 1 if at least in one animal effects on the cornea, irisor conjunctiva that are not expected to reverse or have not fully reversed within an observation periodof normally 21 days; and/orat least 3 out of 4 rabbits show a mean score of 3 for the cornea and/or 1.5 for the irisClassification as eye irritation – Category 2 if at least 3 out of 4 rabbits show a mean score of 1 for the cornea and/or 1 for the iris and/or 2 conjunctival erythema and/or 2 conjunctival swellingIn this case the irritant categories 1 and 2 are used if 4 of 6 rabbits show a mean score asoutlined in the criteria. Likewise, if the test was performed with 4 or 5 animals, for at least 3individuals the mean score must exceed the values laid down in the classification criteria. Asingle animal showing irreversible or otherwise serious effects consistent with corrosion willnecessitate classification as serious eye damage Category 1 irrespective of the number ofanimals used in the test.Other animal testsThe LVET uses the same scoring system as for results from the OECD TG 405, but data fromthe test is not .conclusive for a non-classification. However, they can be included in a weightof evidence determination.Note that in case there are test data that originate from non-OECD tests and scoring has notbeen performed according to the Draize system, the values in Annex I, Tables 3.3.1 / 3.3.2 areno longer applicable for classification purposes. However these data from non-OECD testsshould be considered in a weight of evidence determination.3.3.2.3.3 Weight of evidenceWhere the criteria cannot be applied directly to available identified information, a weight ofthe evidence determination using expert judgement shall be applied in accordance with CLPArticle 9(3).251

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012A weight of the evidence determination means that all available and scientifically justifiedinformation bearing on the determination of hazard is considered together, such as humanexperience (including occupational data and data from accident databases, epidemiologicaland clinical studies, and well-documented case reports and observations), relevant animaldata, skin irritation information/data, physico-chemical parameters (e.g., pH, reservealkalinity/acidity), the results of suitable in vitro tests, information from the application of thecategory approach (grouping, read-across), QSAR results. The quality and consistency of thedata shall be given appropriate weight. Both positive and negative results shall be assembledtogether in a single weight of evidence determination. Evaluation must be performed on acase-by-case basis and with expert judgement. However, normally positive results that areadequate for classification should not be overruled by negative findings.Annex I: 1.1.1.4. For the purpose of classification for health hazards (Part 3) established hazardouseffects seen in appropriate animal studies or from human experience that are consistent with thecriteria for classification shall normally justify classification. Where evidence is available fromboth humans and animals and there is a conflict between the findings, the quality and reliability ofthe evidence from both sources shall be evaluated in order to resolve the question of classification.Generally, adequate, reliable and representative data on humans (including epidemiological studies,scientifically valid case studies as specified in this Annex or statistically backed experience) shallhave precedence over other data. However, even well-designed and conducted epidemiologicalstudies may lack a sufficient number of subjects to detect relatively rare but still significant effects,to assess potentially confounding factors. Therefore, positive results from well-conducted animalstudies are not necessarily negated by the lack of positive human experience but require anassessment of the robustness, quality and statistical power of both the human animal data.For further guidance, if both human and animal data are available, see IR/CSA SectionR.7.2.3.2.3.3.2.4 Decision on classificationA skin corrosive substance is considered to also cause serious eye damage which is indicatedin the hazard statement for skin corrosion (H 314: Causes severe skin burns and eye damage).Thus, in case a substance has to be classified for skin corrosion an additional classificationwith H318 “Causes serious eye damage” is not indicated.3.3.2.5 Setting of specific concentration limitsArticle 10(1) Specific concentration limits and generic concentration limits are limits assigned to asubstance indicating a threshold at or above which the presence of that substance in anothersubstance or in a mixture as an identified impurity, additive or individual constituent leads to theclassification of the substance or mixture as hazardous.Specific concentration limits shall be set by the manufacturer, importer or downstream user whereadequate and reliable scientific information shows that the hazard of a substance is evident when thesubstance is present at a level below the concentrations set for any hazard class in Part 2 of Annex Ior below the generic concentration limits set for any hazard class in Parts 3, 4 and 5 of Annex I.It is more difficult to prove the absence of a hazardous property, the legal text states that:Article 10(1)In exceptional circumstances specific concentration limits may be set by the manufacturer, importeror downstream user where he has adequate, reliable and conclusive scientific information that ahazard of a substance classified as hazardous is not evident at a level above the concentrations setfor the relevant hazard class in Part 2 of Annex I or above the generic concentration limits set for the252

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012relevant hazard class in Parts 3, 4 and 5 of that Annex.A specific concentration limit (SCL) set in accordance with the above mentioned provisionsshall take precedence over the generic concentration limit (GCL) set out in Tables 3.2.3 and3.2.4 of Annex I to CLP (Article 10(6)). Furthermore, an SCL is substance-specific andshould be applicable to all mixtures containing the substance, instead of any GCL thatotherwise would apply to a mixture containing the substance.What type of information may be the basis for setting a specific concentration limit?Existing human data may in certain cases (especially if dose-response information isavailable) indicate that the threshold for the irritation hazard in humans for a substance in amixture, would be higher or lower than the GCL. A careful evaluation of the usefulness andthe validity of such human data as well as their representativeness and predictive value(IR/CSA, sections R.4.3.3. and R.7.2.4.2) should be performed. As pointed out in Section1.1.1.4 of Annex I, CLP, positive results from well-conducted animal studies are notnecessarily negated by the lack of positive human experience but require an assessment ofrobustness, quality and a degree of statistical certainty of both the human and animal data.The aim of the standard test method for “Acute Eye Irritation/Corrosion” TM B.5/OECD TG405 47 is to identify potential serious eye damage or eye irritation. The test material isgenerally administered undiluted. Thus, no dose-response relationship can be obtained froman individual test.However, if there are adequate, reliable, relevant and conclusive existing data from otheralready performed animal studies with a sufficient number of animals tested to ensure a highdegree of certainty, and with information of dose-response relationships, such data may beconsidered for setting a lower or, in exceptional cases, a higher SCL on a case-by-case basis.It should be noted that any additional animal testing to set SCLs (of dilutions) of substancesalready classified as causing serious eye damage or as an eye irritant, is strongly discouragedand may only take place on a case-by-case basis if there are no alternatives providingadequate reliability and quality of data (see CLP Articles 7(1) and 8(1)). The possibilities touse in vitro test methods as a basis for setting SCLs have not yet been explored and therefore,at the present point in time, it is not possible to provide guidance for the use of in vitromethods for the purpose of setting SCLs. However, this does not exclude that a method to setSCLs based on in vitro tests could be developed in the future, and these tests may provide apromising option for SCL setting.An SCL should apply to any mixture containing the substance instead of the GCL (thatotherwise would apply to the mixture containing the substance). Thus, if the SCL is based ondata derived from tests with dilutions of the substance in a specific solvent, it has to beconsidered that the derived concentration, should be applicable to all mixtures for which theSCL should apply.Annex VI Part 3 (Table 3.2) to CLP Regulation includes examples of substances for which ahigher or lower SCL was set under Directive 67/548/EEC (old Dangerous SubstancesDirective (DSD) system).47 TO NOTE: In OECD TG 404 the term test substance refers to the test material, test article or test item. Theterm substance may be used differently from the REACH/CLP definition.253

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.3.2.6 Decision logicThe decision logic which is based on IR/CSA Figure R.7.2-3 is revised to meet CLPrequirements. It is strongly recommended that the person responsible for classification, studythe criteria for classification before and during use of the decision logic.Step0 Is the substance classified as a skincorrosive?NO1aYES Is the substance an organic hydro peroxide oran organic peroxide?YES NO1b Is the pH of the substance 2 or 11.5?YES NO1cAre there other physical or chemicalproperties that indicate that the substance hasthe potential to cause serious eye damage oris irritating to the eye?YES NO2 Is there adequate existing human experiencewhich provides evidence that the substancehas the potential to cause serious eye damageor is irritating to the eye? YES NO3 Are there data from existing studies on eyeirritation in laboratory animals, whichprovide sound conclusive evidence that thesubstance has the potential to cause seriouseye damage, is an eye irritant or non-irritant?When classified as Skin Corr. 1, the risk ofsevere damage to eyes is consideredimplicit.No need to proceed.– Consider to classify asserious eye damage (Eye Dam. 1) if thesubstance is a hydro peroxide, or– eye irritating (Eye Irrit. 2) if thesubstance is a peroxide.ORProvide evidence for the contrary andproceed to step 1b– Where classification is based uponconsideration of pH alone (i.e. bufferingcapacity not known), Eye Dam. 1should be applied. When assigned SkinCorr. 1, the risk of severe damage toeyes is considered implicit.– Where consideration of thealkali/alkaline reserve suggests that thesubstance is not corrosive, this has to beconfirmed (preferably by use of anappropriate in vitro test). Proceed tostep 1cUse this information for weight ofevidence (WoE) determination (step 6).Proceed to step 2Classify accordingly (Eye Dam. 1 or EyeIrrit. 2).Classify accordingly (Eye Dam. 1 or EyeIrrit. 2 or no classification).254

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012NOYES 4 Are there structurally related substances(suitable “read-across” or grouping), whichare classified as serious eye damage or eyeirritant, or do valid QSAR methods indicatethe presence/absence of serious eyedamage/eye irritation potential of thesubstance?YES NO5a5bAre there data from a validated in vitro test(adopted by OECD or not), which provideevidence that the substance is an eye irritantor non-irritant?YES NOAre there acceptable data from a suitable invitro test, which provide evidence that thesubstance is a severe eye irritant? YES NO6 Taking all existing and relevant data) intoaccount, is there sufficient information tomake a decision on classification? YES NOUnable to classify substance for serious eyedamage/eye irritationConsider to classify accordingly (EyeDam. 1 or Eye Irrit. 2). If discriminationbetween Eye Dam. 1 and Eye Irrit. 2 is notpossible, Eye Dam. 1 must be chosen.Proceed to step 5aConsider to classify accordingly (EyeDam. 1 or Eye Irrit. 2 or no classification).If discrimination between Eye Dam. 1 andEye Irrit. 2 is not possible, Eye Cat. 1 mustbe chosen.Proceed to step 5bConsider to classify as Eye Dam. 1.Proceed to step 6Classify accordingly (Eye Dam. 1 or EyeIrrit. 2 or no classification).Decision to undertake generation of newtest data should be made in compliancewith REACH and Article 8 of the CLP.It is recommended that ECHA guidanceR.7.2.6 should also be considered.3.3.3 Classification of mixtures for serious eye damage/eye irritation3.3.3.1 Identification of hazard informationThe procedure for classifying mixtures is a tiered i.e. a stepwise approach based on ahierarchy principle and depending on the type and amount of available data/informationstarting from evaluating existing human data on the mixture, followed by a thoroughexamination of the existing in vivo data, physico-chemical properties, and finally in vitro dataavailable on the mixture. If valid test data are available for the whole mixture they haveprecedence. If no such data exist, the so called bridging principles have to be applied ifpossible. If the bridging principles are not applicable an assessment on the basis of data forthe components of the mixture will be applied.255

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Where it is decided to base the classification of a mixture upon consideration of pH alone,Eye Damage Category 1 should be applied. In this case no further retrieval of information onthe mixture itself is needed.3.3.3.1.1 Identification of existing human dataFor mixtures that have been on the market for a long time, some human data and experiencemay exist that could provide useful information on the eye irritation potential of the respectivemixtures. However, lack of data on effects in humans may be due to, for example, poorreporting or adequate preventive measures. Therefore, lack of data cannot be taken asevidence of the mixture being non-hazardous. See Section 3.3.2.1.1 for further information onthe identification of human data.3.3.3.2 Classification criteria3.3.3.2.1 When data are available for the complete mixtureAnnex I: 3.3.3.1.1. The mixture will be classified using the criteria for substances, and taking intoaccount the testing and evaluation strategies used to develop data for these hazard classes.Unlike other hazard classes, there are alternative tests available for skin corrosivity of certain typesof mixtures that give an accurate result for classification purposes, as well as being simple andrelatively inexpensive to perform. When considering testing of the mixture classifiers areencouraged to use a tiered weight of evidence strategy as included in the criteria for classificationof substances for skin corrosion and serious eye damage and eye irritation to help ensure anaccurate classification, as well as avoid unnecessary animal testing. A mixture is considered tocause serious eye damage (Category 1) if it has a pH ≤ 2.0 or ≥ 11.5. If consideration of alkali/acidreserve suggests the mixture may not have the potential to cause serious eye damage despite thelow or high pH value, then further testing needs to be carried out to confirm this, preferably by useof an appropriate validated in vitro test.Where the criteria cannot be applied directly to available identified information, a weight ofthe evidence determination using expert judgement shall be applied in accordance with CLPArticle 9(3). A weight of the evidence determination means that all available andscientifically justified information bearing on the determination of hazard is consideredtogether, such as physico-chemical parameters, the results of suitable in vitro tests, relevantanimal data, and human experience. The quality and consistency of the data shall be givenappropriate weight. Both positive and negative results shall be assembled together in a singleweight of evidence determination.The integration of all information to come to a final hazard assessment based on weight ofevidence in general requires in-depth toxicological expertise.There are a number of available in vitro test systems that currently being validated for theirsuitability in assessing serious eye damage/eye irritation potential of substances and mixtures.When validated in vitro eye irritation test methods are available in the future the results fromsuch tests can be used for classification. Then these results can also be used to classify themixture. However, not all available in vitro test systems work equally well for all types ofmixtures. Prior to testing a mixture in a specific in vitro assay for classification purposes, ithas to be assured that the respective test has been previously shown to be suitable for theprediction of serious eye damage/eye irritation properties for the type of mixture to beevaluated.256

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.3.3.2.1.1 Mixtures with extreme pHWhere the mixture has an extreme pH value but the only corrosive/irritant ingredient presentin the mixture is an acid or base with an assigned SCL (either CLP Annex VI or set bysupplier), then the mixture should be classified accordingly. In this instance, pH of themixture should not be considered a second time since it would have already been taken intoaccount when deriving the SCL for the substance.If this is not the case, then the steps to be taken into consideration when classifying a mixturewith pH 2 or 11.5 are described in the following decision logic:Mixture not classified as Skin Corr. 1 and without in vivo data on serious eye damage/eye irritationor relevant data from similar tested mixtures.pH is 2 or 11.5Does the acid/alkaline reserve indicate that the mixturemay not be corrosive?NO YESIs the mixture tested for serious eye damagingproperties in an accepted in vitro test? NO YESDoes the mixture demonstrate serious eye damagingproperties in an accepted in vitro test?YES NOApply methods in Annex I, 3.3.3.3.2 (Table 3.3.3) /3.3.3.3.4 (Table 3.3.4) (When validated in vitro eye irritation test methods areavailable, these may be used to generate data toclassify the mixture instead of using the summationmethod.)Classify as serious eye damaging, EyeDam. 1.Classify as serious eye damaging, EyeDam. 1.Classify as serious eye damaging, EyeDam. 1.Classify accordingly.If consideration of extreme pH and acid/alkaline reserve indicates the mixture may not havethe potential to cause serious eye damage, then the supplier should carry out further testing toconfirm this (Annex I, Section 3.3.3.2.1). The mixture must be classified as Serious eyedamage Category 1 if the supplier decide not to carry out the required confirmatory testing.If further testing confirms that the mixture should not be classified for serious eye damageeffects, then the supplier should assess the mixture for eye irritation either using in vitro eyeirritation test methods when available or the summation method.It must be note that the pH-acid/alkali reserve method assumes that the potential corrosivity orirritancy is due to the effect of the ionic entities. When this is not the case, especially whenthe mixture contains non-ionic (non-ionisable) substances themselves classified as corrosiveor irritant, then the pH-reserve method cannot be a basis for modifying the classification.Where the mixture has an extreme pH value and contains some other corrosive/irritantingredients (some of which may have SCLs assigned) in addition to an acid or base with orwithout an assigned SCL, then the mixture shall follow the procedure described in thedecision logic.257

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.3.3.2.2 When data are not available for the complete mixture: bridging principlesAnnex I: 3.3.3.2.1. Where the mixture itself has not been tested to determine its skin corrosivity orpotential to cause serious eye damage or irritation, but there are sufficient data on the individualingredients and similar tested mixtures to adequately characterise the hazards of the mixture, thesedata shall be used in accordance with the bridging rules set out in section 1.1.3.In order to apply bridging principles, there needs to be sufficient data on similar testedmixtures as well as the components of the mixture.When the available identified information is inappropriate for the application of the bridgingprinciples then the mixture should be classified using the methods described in Section1.1.3.2.3.3.3.3.2.3 When data are available for all components or only for some components ofthe mixture3.3.3.2.3.1 Components that should be taken into account for the purpose ofclassificationAnnex I: 3.3.3.3.1. ….. Assumption: The 'relevant ingredients' of a mixture are those which arepresent in concentrations of 1% (w/w for solids, liquids, dusts, mists and vapours and v/v for gases) orgreater, unless there is a presumption (e.g. in the case of corrosive ingredients) that an ingredientpresent at a concentration of less than 1% is still relevant for classifying the mixture for eyeirritation/serious eye damage.3.3.3.2.3.2 The additivity approach is applicableAnnex I: 3.3.3.3.2. In general, the approach to classification of mixtures as eye irritant or seriouslydamaging to the eye when data are available on the components, but not on the mixture as a whole,is based on the theory of additivity, such that each corrosive or irritant component contributes to theoverall irritant or corrosive properties of the mixture in proportion to its potency and concentration.A weighting factor of 10 is used for corrosive components when they are present at a concentrationbelow the generic concentration limit for classification in Category 1, but are at a concentration thatwill contribute to the classification of the mixture as an irritant. The mixture is classified asseriously damaging to the eye or eye irritant when the sum of the concentrations of suchcomponents exceeds a concentration limit.3.3.3.3.3. Table 3.3.3 provides the generic concentration limits to be used to determine if themixture shall be classified as irritant or as seriously damaging to the eye.When the supplier is unable to derive the classification using either data on the mixture itselfor bridging principles, he must determine the serious eye damage/ eye irritation properties ofhis mixture using data on the individual ingredients. The supplier must ascertain whether theadditivity approach is applicable, the first step in the process being to identify all theingredients in the mixture (i.e. their name, chemical type, concentration level, hazardclassification and any SCLs) and the pH of the mixture. In addition, for example surfactantinteraction or neutralisation of acids/bases could occur in a mixture, which makes it importantto consider not only the contribution of individual ingredients but also the effects of the entiremixtureAdditivity may not apply where the mixture contains substances mentioned in Annex I,3.3.3.3.4.1 which may be corrosive/irritant at concentrations below 1%, see Section3.3.3.2.3.3.Application of SCLs when applying the additivity approach258

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The generic concentration limits are specified in Table 3.3.3. However, Article 10.5 indicatesthat specific concentration limits (SCLs) take precedence over generic concentration limits.Thus, if a given substance has a SCL, then this specific concentration limit has to be takeninto account when applying the summation (additivity) method for serious eye damage/eyeirritation (see Examples 4 and 5).In cases where additivity applies for serious eye damage/eye irritation to a mixture with twoor more substances some of which may have SCLs assigned, then the following formulashould be used:The mixture is classified for serious eye damage/eye irritation if theSum of (ConcA / clA) + (ConcB / clB) + ….+ (ConcZ / clZ) is 1Where ConcA = the concentration of substance A in the mixture;clA = the concentration limit (either specific or generic) of substance A;ConcB = the concentration of substance B in the mixture;clB = the concentration limit (either specific or generic) of substance B; etc.This approach is similar to that used in the DPD where a substance SCL can replace thedefault limits in the conventional method equations.3.3.3.2.3.3 The additivity approach is not applicableAnnex I; 3.3.3.3.4.1. Particular care must be taken when classifying certain types of mixturescontaining substances such as acids and bases, inorganic salts, aldehydes, phenols, and surfactants.The approach explained in paragraphs 3.3.3.3.1 and 3.3.3.3.2 might not work given that many ofsuch substances are corrosive or irritant at concentrations < 1 %.3.3.3.3.4.2. For mixtures containing strong acids or bases the pH shall be used as classificationcriteria (see paragraph 3.3.2.3) since pH will be a better indicator of serious eye damage than thegeneric concentration limits of Table 3.3.3.3.3.3.3.4.3. A mixture containing corrosive or irritant ingredients that cannot be classified based onthe additivity approach (Table 3.3.3), due to chemical characteristics that make this approachunworkable, shall be classified as Category 1 for effects on the eye if it contains ≥ 1 % of acorrosive ingredient and as Category 2 when it contains ≥ 3 % of an irritant ingredient.Classification of mixtures with ingredients for which the approach in Table 3.3.3 does not apply issummarised in Table 3.3.4.3.3.3.3.5. On occasion, reliable data may show that the reversible/irreversible eye effects of aningredient will not be evident when present at a level above the generic concentration limitsmentioned in Tables 3.3.3 and 3.3.4. In these cases the mixture shall be classified according tothose data. On other occasions, when it is expected that the skin corrosion/irritation hazards or thereversible/irreversible eye effects of an ingredient will not be evident when present at a level abovethe generic concentration limits mentioned in Tables 3.3.3 and 3.3.4, testing of the mixture shall beconsidered. In those cases, the tiered weight of evidence strategy shall be applied.3.3.3.3.6. If there are data showing that (an) ingredient(s) may be corrosive or irritant at aconcentration of < 1 % (corrosive) or < 3 % (irritant), the mixture shall be classified accordingly.3.3.3.3 Generic concentration limits for substances triggeringclassification of mixtures259

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.3.3.3.1 When the additivity approach is applicableAnnex I: Table 3.3.3Generic concentration limits of ingredients of a mixture classified as Skin corrosive Category1 and/or eye Category 1 or 2 for effects on the eye that trigger classification of the mixture foreffects on the eye (Category 1 or 2)Sum of ingredients classified as:Eye effects Category 1 or Skincorrosive Category 1A, 1B, 1CConcentration triggering classification of a mixture as:Irreversible Eye EffectsReversible Eye EffectsCategory 1 Category 2 3 % 1 % but < 3 %Eye Effects Category 2 10 %(10 x Eye Effects Category 1) +Eye effects Category 2Skin Corrosive Category 1A, 1B,1C + Eye effects Category 110 x (Skin corrosive Category1A, 1B, 1C + Eye EffectsCategory 1) + Eye EffectsCategory 23.3.3.3.2 When the additivity approach is not applicable 10 % 3 % 1 % but < 3 % 10 %Annex I: Table 3.3.4Generic concentration limits of ingredients of a mixture for which the additivity approachdoes not apply, that trigger classification of the mixture as hazardous to the eyeIngredient Concentration Mixture classified as: EyeAcid with pH ≤ 2 ≥ 1% Category 1Base with pH ≥ 11,5 ≥ 1% Category 1Other corrosive (Categories 1)ingredients for which additivity doesnot applyOther irritant (Category 2) ingredientsfor which additivity does not apply,including acids and bases≥ 1% Category 1≥ 3% Category 2There are ongoing discussions at UN level whether 'Other irritant (Category 2) ingredients' inTable 3.3.4 (last row) include skin and eye irritants or only eye irritants.3.3.3.4 Decision logicThe decision logic which is based on IR/CSA Figure R.7.2-3 is revised to meet CLPrequirements. It is strongly recommended that the person responsible for classification, studythe criteria for classification before and during use of the decision logic.260

Guidance on the Application of the CLP Criteria Version 3.0 - November 20121. When data are available for the complete mixture0 Is the mixture classified as a skin corrosive?YES NO1a Is the pH of the mixture 2 or 11.5?YES NO1bAre there other physical or chemicalproperties that indicate that the mixture hasthe potential to cause serious eye damage or isirritating to the eye?YES NO2 Are there adequate existing human experiencedata which provide evidence that the mixturehas the potential to cause serious eye damageor is irritating to the eye?YES NO3 Are there data from existing studies on eyeirritation in laboratory animals, which providesound conclusive evidence that the mixturehas the potential to cause serious eye damage,is an eye irritant or non-irritant? YES NO4a4bAre there data from a validated in vitro or exvivo test (adopted by OECD or not), whichprovide evidence that the mixture is an eyeirritant or non-irritant?YES NOAre there acceptable data from a suitable invitro test, which provide evidence that themixture is an irritant to the eye? YES NOWhen assigned Skin Corr. 1, the risk ofsevere damage to eyes is consideredimplicit.No need to proceed.– Where classification is based uponconsideration of pH alone (i.e. bufferingcapacity not known), Eye Dam. 1should be applied. When assigned SkinCorr. 1, the risk of severe damage toeyes is considered implicit.– Where consideration of the acid/alkalinereserve suggests that the substance isnot corrosive, this has to be confirmed(preferably by use of an appropriate invitro test). Proceed to step 1b.Use this information for weight of evidence(WoE) determination (step 6).Proceed to step 2.Classify accordingly (Eye Dam. 1 or SkinIrrit. 2).Classify accordingly (Eye Dam. 1 or EyeIrrit. 2 or no classification).Consider to classify accordingly (Eye Dam.1 or Eye Irrit. 2 or no classification).If discrimination between Eye Dam. 1 andEye Irrit. 2 is not possible, Eye Dam. 1must be chosen.Proceed to step 4bConsider to classify accordingly (Eye Dam.1 or Eye Irrit. 2). If discrimination betweenEye Dam. 1 and Eye Irrit. 2 is not possible,Eye Dam. 1 must be chosen.261

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012 Proceed to step 55 Taking all existing and relevant data (steps 1-4) into account including potentialsynergistic/antagonistic effects andbioavailability, is there sufficient informationto make a decision on classification? YES NOClassify accordingly (Eye Dam. 1 or EyeIrrit. 2 or no classification)2. When data are not available for the complete mixture: bridging principles6a6bAre existing eye irritation data available onsimilar tested mixtures and on the individualingredients?NO YESCan bridging principles be applied? YES NOProceed to step 7aClassify in appropriate category (Eye Dam.1 or Eye Irrit. 2 or no classification)3. When data are available for all components or only for some components of the mixture7a Is pH of the mixture 2 or 11.5? YES 7bNOIs there any indication that the additivityprinciple does not apply?YES NOSection. 3.3.3.3.2 and Table 3.3.3 applies.Take relevant ingredients (Annex I, 3.2.3.3.1)and SCLs into account, as appropriate.Classify in appropriate category (Eye Dam. 1or Eye Irrit. 2 or no classification).Follow decision logic in Section 3.3.3.2.1.1and classify accordingly.Section 3.3.3.3.4 and Table 3.3.4 mayapply.Take relevant ingredients (Annex I,3.2.3.3.1) and SCLs into account, asappropriate.Classify in appropriate category (Eye Dam.1 or Eye Irrit. 2 or no classification)262

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.3.4 Hazard communication in form of labelling for serious eye damage/eyeirritation3.3.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I; 3.3.4.1 Label elements shall be used for substances or mixtures meeting the criteria forclassification in this hazard class in accordance with Table 3.3.5.Table3.3.5Label elements for serious eye damage/eye irritationClassification Category 1 Category 2GHS PictogramsSignal Word Danger WarningHazard StatementPrecautionary StatementPreventionPrecautionary StatementResponsePrecautionary StatementStoragePrecautionary StatementDisposalH318: Causes serious eyedamageP280P305 + P351 + P338P310H319: Causes serious eyeirritationP264P280P305 + P351 + P338P337 + P313A skin corrosive mixture is considered to also cause serious eye damage which is indicated inthe hazard statement for skin corrosion, H 314: Causes severe skin burns and eye damage.Thus, in case a mixture has to be classified for skin corrosion an additional classification withH318: Causes serious eye damage is not indicated.3.3.5 Re-classification of substances and mixtures classified for serious eyedamage/eye irritation according to DSD and DPD3.3.5.1 Is direct “translation” of classification and labellingpossible?A direct translation as indicated in the translation table in Annex VII to CLP is generallypossible. However, an evaluation and classification must be carried out in accordance withCLP Articles 9 – 13 when data for the mixture are available. Translation from classificationaccording to DSD to the classification according to CLP is as follows:– Xi; R41 is translated into Eye Dam. 1; H318. The criteria in DSD are completely coveredby the criteria in CLP.– Xi; R36 is translated into Eye Irrit. 2; H 319. The criteria in DSD are completely coveredby the criteria in CLP.263

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012It should be noted that CLP eye irritation Category 2 will include more substances which arecurrently not classified under the DSD, but with values of cornea opacity >1 and 2 and < 2.5, will be classified as eye irritants under CLP.It should be noted that where mixtures containing substances with risk phrase R41 have beenclassified on basis of the hazards of individual ingredients, the use of the translation table maylead to an under-classification of the mixture. This is because the general concentration limits,to be applied for mixtures, are lowered under CLP compared to DPD. For mixtures containingsubstances with this classification the use of the translation table may therefore not beappropriate and re-classification done by using the existing data would be more correct. Formore details see Chapter 1.7.3.3.5.2 Re-evaluation of dataIf there is new information which might be relevant with respect to classification a reevaluationhas to be performed.3.3.6 Examples of classification for serious eye damage/eye irritation3.3.6.1 Examples of substances fulfilling the criteria forclassification3.3.6.1.1 Example 1: Standard test according to OECD TG 405 with three animalsIn a study according to OECD 405 the test substance was applied on the eyes of three rabbits.The scoring results obtained are listed in the following table:Cornea:AnimalNr123Evaluation after …Positive responder? Score …1 hr 24 hrs 48 hrs 72 hrs 21 days 1 30 2 2 2 0 24/48/72 h animal 1 is 2 Yes No2 2 2 2 0 24/48/72 h animal 2 is 2 Yes No2 2 1 1 0 24/48/72 h animal 3 is 1.3 Yes NoIris:264AnimalNr1Evaluation after …Effects are reversiblePositive responder? Score …1 hr 24 hrs 48 hrs 72 hrs 21 days 1 1,50 1 1 1 0 24/48/72 h animal 1 is 1 Yes No2 1 1 1 1 0

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123 24/48/72 h animal 2 is 1 Yes No1 1 1 1 0 24/48/72 h animal 3 is 1 Yes NoEffects are reversibleConjunctiva – Erythema:Animal #Evaluation after …Positive responder? Score …1 hr 24 hrs 48 hrs 72 hrs 21 days 21 2 2 2 2 0 24/48/72 h animal 1 is 2Yes2 1 1 1 1 0 24/48/72 h animal 2 is 1No3 1 1 1 1 0 24/48/72 h animal 3 is 1NoEffects are reversibleConjunctiva – Swelling:Animal #Evaluation after …Positive responder? Score …1 hr 24 hrs 48 hrs 72 hrs 21 days 21 0 3 3 3 0 24/48/72 h animal 1 is 3Yes2 2 2 2 1 0 24/48/72 h animal 2 is 1.7No3 2 3 2 2 0 24/48/72 h animal 3 is 2.3YesEffects are reversibleClassification according to CLP: Eye irritant Category 2Rationale: Cornea and Conjunctiva ”positive responder” 2: 2/3 animalsIris ”positive responder” 1: 3/3 animals265

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.3.6.1.2 Example 2: Test carried out with more than 3 rabbitsCornea:Animal No.1234Iris:Positive responder?Evaluation after … Score …1h 24h 48h 72h 7d 14d 21d 3 11 2 3 3 1 1 0 24/48/72h = 2.7no yes1 2 2 3 1 1 0 24/48/72h = 2.3no yes1 2 3 3 2 1 0 24/48/72h = 2.7no yes1 2 4 4 2 1 0 24/48/72h = 3.3yes yesEffects are reversibleAnimal No.1234Positive responder?Evaluation after … Score …1h 24h 48h 72h 7d 14d 21d 1.5 10 0 0 0 0 0 0 24/48/72h = 0no no0 0 0 0 0 0 0 24/48/72h = 0no no0 1 1 1 1 0 0 24/48/72h = 1no yes0 0 0 0 0 0 0 24/48/72h = 0no noConjunctiva – Erythema:Animal No.266Evaluation after …Effects are reversible1h 24h 48h 72h 7d 14d 21d 2Positive responder? Score …

Guidance on the Application of the CLP Criteria Version 3.0 - November 201212342 2 2 1 1 1 0 24/48/72h = 1.72 2 2 1 1 0 0 24/48/72h = 1.72 2 2 1 1 1 1 24/48/72h = 1.72 2 2 1 0 0 0 24/48/72h = 1.7nonononoConjunctiva – Swelling:Animal No.1234Evaluation after …Effects are NON-reversible1h 24h 48h 72h 7d 14d 21d 22 2 2 1 1 1 0 24/48/72h = 1.72 2 1 1 1 0 0 24/48/72h = 1.32 2 2 1 1 1 1 24/48/72h = 1.72 2 2 1 1 1 1 24/48/72h = 1.7Positive responder?nononono Score …Classification according to CLP: Serious eye damage Category 1Rationale: Conjunctiva with irreversible effectsEffects are NON-reversible3.3.6.2 Examples of mixtures fulfilling the criteria forclassification3.3.6.2.1 Example 3: Application of the additivity approach for mixtures containingingredients without SCLsWhere the mixture is made up of ingredients with no assigned SCLs, then the appropriatesummation(s) from Table 3.3.3 should be used.Ingredient Skin / eye classification Concentration(% w/w)Surfactant A Eye Cat 1 1.8 Not assignedSubstance B Eye Cat 2 0.5 Not assignedSubstance C Eye Cat 1 5.4 Not assignedSCL267

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Substance D Not classified 4.0Acid E Skin Cat 1A 2.0 Not assignedWater Not classified 86.3pH of the mixture is 9.0 – 10.0, thus extreme pH provisions do not apply. The mixturecontains a surfactant and an acid but neither are corrosive/irritant below 1% (as identified bythe absence of specific concentration limits in either CLP Annex VI or the Classification andLabelling Inventory). Additivity is considered to apply.Substance D and water can be disregarded as they are not classified for serious eyedamage/eye irritation. Substance B can also be disregarded as present below 1%.Mixture contains 7.2% Eye Cat 1 ingredients as well as 2% acid E so the summation {Skincorrosion Cat 1A, 1B, 1C + Eye Cat 1} applies and is > 3%, thus mixture is classified EyeCat 1.3.3.6.2.2 Example 4: Application of the additivity approach for mixtures containingingredients which may have SCLs268Ingredient Skin / eye classification Concentration(% w/w)Surfactant A Eye Cat 1 2.0 Not assignedSubstance B Eye Cat 2 0.5 Not assignedSubstance C Skin Cat 1B 5.4 C ≥ 10 %: Skin Cat 1B5 % ≤ C < 10 %: Eye Cat 2Substance D Not classified 4.0Substance E Skin Cat 1B 2.0 Not assignedWater Not classified 86.1pH of the mixture is 10.5 – 11.0, thus extreme pH provisions do not apply. The mixturecontains a surfactant, an acid and a base but none are corrosive/irritant below 1% (asidentified by the absence of specific concentration limits in either CLP Annex VI or theClassification and Labelling Inventory). Additivity is considered to apply.Substance D and water can be disregarded as they are not classified for serious eyedamage/eye irritation. Substance B can also be disregarded as present below 1%.SCLs are not assigned to substance E or surfactant A, thus generic concentration limits (GCL)apply for these ingredientsEye Cat 1(% surfactant A / GCL) + (% Substance C / SCL) + (% Substance E / GCL) = (2/3) + (5.4/10)+ (2/3) = 1.9 > 1 thus mixture is classified Eye Cat 13.3.6.2.3 Example 5: Application of the additivity approach for mixtures containingingredients which may have SCLsIngredientSerious eye damage/ eyeirritation classificationConcentration(% w/w)Surfactant B Eye Cat 1 0.7 Not assignedSubstance C Eye Cat 2 74.9 Not assignedSCLSCLSubstance D Eye Cat 1 8.5 C ≥ 25 %: Eye Cat 1

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Substance E Not classified 15.910 % ≤ C < 25 %: Eye Cat 2pH of the mixture is 10.0 – 10.5 (10% solution), thus extreme pH provisions do not apply.The mixture contains a surfactant which is not corrosive/irritant below 1% (as identified bythe absence of specific concentration limits in either CLP Annex VI or the Classification andLabelling Inventory). Additivity is considered to apply.Substance E can be disregarded as it is not classified for serious eye damage/eye irritation.Surfactant B can also be disregarded as present below 1%.SCLs are not assigned to substance C, thus GCL apply for this ingredientEye Cat 1Mixture contains 8.5% substance D, the only ‘relevant’ ingredient classified as Eye Cat 1. Asthis is below the 25% SCL for substance D, the mixture is not classified Eye Cat 1Eye Cat 2(%substance D/ SCL) + (%substance C / GCL) = (8.5/10) + (74.9/10) which is > 1 thusmixture is classified Eye Cat 23.3.7 ReferencesGriffith J.F., Nixon G.A., Bruce R.D., Reer P.J., Bannan E.A. (1980): Dose-response studieswith chemical irritants in the albino rabbit eye as a basis for selecting optimum testingconditions for predicting hazard to the human eye. Toxicol Appl Pharmacol 55, 501-513.Young J.R., How M.J., Walker A.P., Worth W.M.H. (1988): Classification as corrosive orirritant to skin of preparations containing acidic or alkaline substances, without test onanimals. Toxicology in Vitro 2, 19-26.Young J.R., How M.J. (1994), Product classification as corrosive or irritant by measuring pHand acid / alkali reserve. In Alternative Methods in Toxicology vol. 10 - In Vitro SkinToxicology: Irritation, Phototoxicity, Sensitization, eds. A.Rougier, A.M. Goldberg and H.IMaibach, Mary Ann Liebert, Inc. 23-27.3.4 RESPIRATORY OR SKIN SENSITISATION3.4.1 Definitions and general considerations for respiratory or skin sensitisationAnnex I: 3.4.1.1. Respiratory sensitiser means a substance that will lead to hypersensitivity of theairways following inhalation of the substance.3.4.1.2. Skin sensitiser means a substance that will lead to an allergic response following skincontact.In terms of prevention it might be important to note that respiratory sensitisation may beinduced not only by inhalation but also by skin contact.269

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: 3.4.1.3. For the purpose of section 3.4, sensitisation includes two phases: the first phaseis induction of specialised immunological memory in an individual by exposure to an allergen.The second phase is elicitation, i.e. production of a cell-mediated or antibody-mediated allergicresponse by exposure of a sensitised individual to an allergen.3.4.1.4. For respiratory sensitisation, the pattern of induction followed by elicitation phases isshared in common with skin sensitisation. For skin sensitisation, an induction phase is required inwhich the immune system learns to react; clinical symptoms can then arise when subsequentexposure is sufficient to elicit a visible skin reaction (elicitation phase). As a consequence,predictive tests usually follow this pattern in which there is an induction phase, the response towhich is measured by a standardised elicitation phase, typically involving a patch test. The locallymph node assay is the exception, directly measuring the induction response. Evidence of skinsensitisation in humans normally is assessed by a diagnostic patch test.3.4.1.5. Usually, for both skin and respiratory sensitisation, lower levels are necessary forelicitation than are required for induction. Provisions for alerting sensitised individuals to thepresence of a particular sensitiser in a mixture can be found at section 3.4.4.3.4.1.6. The hazard class Respiratory or Skin Sensitisation is differentiated into:- Respiratory Sensitisation;- Skin Sensitisation.3.4.2 Classification of substances for respiratory or skin sensitisation3.4.2.1 Identification of hazard informationThere are no formally recognised and validated animal tests for respiratory sensitisation.However there may be data from human observation indicating respiratory sensitisation inexposed populations.With respect to identification of relevant information for skin sensitisation see IR/CSA,Section R.7.3.3.3.4.2.1.1 Identification of human dataRelevant information with respect to respiratory or skin sensitisation may be available fromcase reports, epidemiological studies, medical surveillance and reporting schemes. For moredetails see IR/CSA, Section R.7.3.3.2.3.4.2.1.2 Identification of non human dataAt present no validated non-testing systems exist to predict skin sensitising potential. Thechemical structure of a molecule, when similar to that of known sensitisers, may form part ofthe weight of evidence for classification (see also IR/CSA, Section R.7.3.3).The subject of in vitro testing for skin sensitisation has also been dealt with in IR/CSA,Section R.7.3.3. At present no validated in vitro methods exist to identify the sensitisingpotential of a chemical.There are three animal test methods used to evaluate skin sensitisation for substances: themouse local lymph node assay (LLNA), the guinea pig maximisation test (GPMT) and theBuehler occluded patch test. They are further described in IR/CSA, Section R.7.3.3, and in thecontext of classification in Section 3.4.2.3.4.270

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: 3.4.2.1. Respiratory sensitisers3.4.2.2 Classification criteria for substancesSubstances shall be classified as respiratory sensitisers (Category 1) in accordance with the criteriain Table 3.4.1:Table 3.4.1CategoryCategory 13.4.2.2. Skin SensitisersHazard category for respiratory sensitisersCriteriaSubstances shall be classified as respiratory sensitisers (Category 1) in accordancewith the following criteria:(a) if there is evidence in humans that the substance can lead to specificrespiratory hypersensitivity and /or(b) if there are positive results from an appropriate animal test.3.4.2.2.1. Substances shall be classified as skin sensitisers (Category 1) in accordance with criteriain Table 3.4.2:Table 3.4.2Hazard category for skin sensitisersCategoryCategory 1CriteriaSubstances shall be classified as skin sensitisers (Category 1) in accordance withthe following criteria:(i) if there is evidence in humans that the substance can lead to sensitisation byskin contact in a substantial number of persons, or(ii) if there are positive results from an appropriate animal test (see specificcriteria in paragraph 3.4.2.2.4.1).3.4.2.3 Evaluation of hazard information3.4.2.3.1 Human data on respiratory sensitisationSubstances shall be classified as respiratory sensitisers if there is evidence in humans that thesubstance can lead to specific respiratory hypersensitivity.3.4.2.3.2 Human data on skin sensitisationAnnex I: 3.4.2.2.2.1. For classification of a substance as a skin sensitiser, evidence shall includeany or all of the following:(a) positive data from patch testing, normally obtained in more than one dermatology clinic;(b) epidemiological studies showing allergic contact dermatitis caused by the substance;Situations in which a high proportion of those exposed exhibit characteristic symptoms are tobe looked at with special concern, even if the number of cases is small;…(d) positive data from experimental studies on humans (see Article 7(3));(e) well documented episodes of allergic contact dermatitis, normally obtained in more than onedermatology clinic.3.4.2.3.3 Non human data on respiratory sensitisationNo formally recognised and validated animal tests currently exist for respiratory sensitisation.271

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.4.2.3.4 Non human data on skin sensitisationCurrently CLP allows classification of skin sensitisers in only one hazard category. Since it ispossible to refine the evaluation of skin sensitisers on the basis of the potency of thesensitising effect, this guidance advises how to evaluate the potency on the basis of therecommended test methods. The potency considerations may be used as a basis for settingspecific concentration limits (see Section 3.4.2.5) and it is also concluded in the GHS that thispotency consideration will be included as further classification criteria in the future.There are currently three recognised and officially accepted animal test methods for skinsensitisation defined by OECD Test Guidelines. These are the Mouse Local lymph NodeAssay (LLNA), Guinea Pig Maximisation Test by Magnusson & Kligman (GPMT) and theBuehler occluded patch test in the guinea pig. The mouse and guinea pig methods differfundamentally with respect to the endpoints used; whereas the mouse LLNA measures theresponses provoked during the induction of sensitisation, the two guinea pig tests measurechallenge induced elicitation reactions in previously sensitised animals. For new testing ofsubstances the LLNA is now the method of first choice. In the exceptional circumstance thatthe LLNA is not appropriate, one of the alternative tests may be used (Buehler or GPMT), butjustification shall be provided (see IR/CSA, Section R.7.3.2.1).Test results from the LLNA, GPMT and the Buehler test could be used directly forclassification. They may also be used for potency evaluation.A sensitising potential of a substance is identified if a significant effect has been obtained inan acceptable in vivo test. A significant skin sensitising effect in each of the three recognisedanimal tests is defined as follows:272Table 3.4.2.3.4: Definition of significant skin sensitising effectTestMouse local lymph node assay (LLNA) Stimulation Index ≥ 3Guinea pig maximisation test (GPMT)Buehler occluded patch testResultRedness in ≥ 30% of the test animalsRedness in ≥ 15% of the test animalsA substance may be classified a skin sensitiser on the basis of a positive test result in one ofthe above described animal tests. A positive result obtained by another not officiallyrecognised test method may also justify classification as a skin sensitiser, but can normallynot overrule a negative result obtained in one of the three recognised, above described animaltests. A new animal study should not be conducted in an attempt to negate a clearly positiveresponse in a not officially recognised test method particularly where there is other supportingevidence that the substance is a skin sensitiser.3.4.2.3.4.1 Mouse Local Lymph Node Assay (LLNA, OECD TG 429)The LLNA is used both for determination of skin sensitising potential (hazard identification)and for determination of relative skin sensitisation potency (hazard characterisation). In bothinstances the metric is cellular proliferation induced in draining lymph nodes followingtopical exposure to a chemical, lymph node cell proliferation being causally andquantitatively correlated with the acquisition of skin sensitisation (Basketter et al. 2002a,2002b). A correlation has been demonstrated between the concentration of chemical requiredfor the acquisition of skin sensitisation in humans according to historical predictive data andskin sensitisation potency as measured in the mouse LLNA (Schneider and Akkan 2004,Basketter et al. 2005b). Potency is measured as a function of derived EC3-values. The EC3-value is the amount of test chemical (% concentration, molar value or dose per unit area)required to elicit a stimulation index of 3 in the standard LLNA (Kimber et al. 2003). An

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012inverse relationship exists between EC3-value and potency meaning that extremely potentsensitisers have extremely low EC3-values. The relevance of potency derives from anappreciation that skin sensitisers vary by up to four or five orders of magnitude with respect tothe minimum concentration required inducing skin sensitisation. Potency is graded on thebasis of these minimum concentrations each grade reflecting a concentration range ofapproximately one order of magnitude.The following scheme could be used for determination of potency categories for sensitisers.However, classification into potency categories is currently not a requirement in theclassification of sensitisers.Table 3.4.2.3.4.1: Skin Sensitisation Potency in the Mouse Local Lymph Node AssayEC3-value (% w/v)≤ 0.2PotencyExtreme> 0.2 - ≤ 2 Strong> 2 ModeratePotency may be considered when setting a specific concentration limit for a substance inmixtures (see Section 3.4.2.5).3.4.2.3.4.2 Guinea Pig Maximisation Test (GPMT, OECD TG 406)This test has been used for almost 40 years to detect the sensitising potential of chemicalsthrough a test system maximizing the sensitivity by both intradermal and epidermal inductionand use of an adjuvant (Freund’s Complete Adjuvant). The intradermal induction is made byinjection. Consequently the test is not suited for products which cannot be made up into aliquid formulation.The GPMT was originally designed to maximise the ability to identify a sensitisation hazard,rather than to determine skin sensitisation potency. Yet, when only a GPMT test result isavailable, potency categorisation is possible on the basis of the concentration of test materialused for intradermal induction and the percentage of guinea pigs sensitised. However, itshould be recognised that there is often a degree of uncertainty associated with the derivationof allergenic potencies from the GPMT. If the test has been conducted in full compliance withOECD Test Guideline 406 and with the technical details ensuring proper data interpretation asspecified by Schlede and Eppler (1995), the following scheme may be used.The following scheme could be used for determination of potency categories for sensitisers.However, classification into potency categories is currently not a requirement in theclassification of sensitisers.Table 3.4.2.3.4.2: Potency on basis of the Guinea Pig Maximisation TestConcentration for intradermalinduction (% w/v)Incidence sensitised guinea pigs (%)Potency≤ 0.1 ≥ 60 Extreme≤ 0.1 >30 - 0.1 - 0.1 - 30 - 1.0 ≥ 30 Moderate273

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Potency may be considered when setting a specific concentration limit for a substance inmixtures (see Section 3.4.2.5).3.4.2.3.4.3 Buehler occluded patch test (OECD TG 406)This test has been in use for the last 40 years as a more realistic, although still a sensitive, testto detect skin sensitisers using epidermal occluded exposure. The skin barrier of the testspecies (guinea pig) is kept intact in this assay, thus providing for a more relevant exposurescenario for the human situation. Potency can be categorised using the results of the Buehlertest on the basis of the number of animals sensitised and the concentration of the test materialused for the epidermal induction. However, it should be recognised that there is often a degreeof uncertainty associated with the derivation of allergenic potencies from the Buehler test.The following scheme could be used for determination of potency categories, if the test hasbeen conducted in full compliance with OECD TG 406 and with the technical details ensuringproper data interpretation as specified by Robinson et al (1990).Classification into potency categories is currently not a requirement in the classification ofsensitisers.Table 3.4.2.3.4.3: Potency on basis of the Buehler Occluded Patch TestConcentration for intradermalinduction (% w/v)Incidence sensitised guinea pigs (%)Potency≤ 0.2 ≥ 60 Extreme≤ 0.2 >15 - 0.2 - 0.2 - 15 - 20 ≥ 15 ModeratePotency may be considered when setting a specific concentration limit for a substance inmixtures (see Section 3.4.2.5).3.4.2.3.4.4 Non-compliant skin sensitisation testsIn vivo test methods which do not comply with recognised guidelines are stronglydiscouraged for the identification of skin sensitisers or assessment of skin sensitising potency.The results of such tests have to be evaluated carefully, but may provide supportive evidence.If doubts exist about the validity and the interpretation of the results, the evaluation needs tobe taken by using a weight-of-evidence approach.3.4.2.3.4.5 Animal test methods conducted for purposes other than sensitisationOccasionally signs of skin sensitisation occur in repeated dose tests. These tests are oftendermal toxicity tests on rats. Clearly, if signs of erythema/oedema occur in animals afterrepeated application, the possibility of skin sensitisation should be considered, and ideallyassessed in an appropriate study.3.4.2.3.5 Weight of evidencePositive effects seen in either humans or animals for skin sensitisation will normally justifyclassification. Evidence from animal studies on skin sensitisation is usually more reliable thanevidence from human exposure, although reliable human data is, of course, most relevant toman. In cases where evidence is available from both sources, and there is conflict between theresults, the quality and reliability of the evidence from both sources must be assessed in order274

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012to decide on the classification on a case-by-case basis. Negative human data should notnormally negate positive findings in animal studies.Since the data used in hazard or risk assessment should be relevant, reliable and sufficient forthe regulatory purpose, it is necessary to base the assessment on the totality of availableinformation, i.e. to apply Weight of Evidence (WoE) considerations.The WoE assessment can be based on the total of experimental data, as well as post-marketsurveys and/or occupational experience data. In the case of mixtures, extrapolation fromsimilar mixtures or from data available on the components may often provide reliable meansof assessment. Estimated data might be used to supplement and increase confidence in theavailable experimental data, whereas in some others, such data might be used instead ofexperimental data.WoE assessment can be divided into two stages:a) Assessment of each single test result and, if needed, of other data. It may be helpful toapply criteria for reliability as defined by Klimisch et al (1997). These criteria includedetails on the recognition of the test method, reporting detail, method relevance, testparameters, etc.b) Comparison of the weighed single test results.Good quality data on the substance itself have more weight than such data extrapolated fromsimilar substances.3.4.2.4 Decision on classificationAccording to CLP Annex I, Section 3.4.2.1 substances fulfilling the criteria for respiratorysensitisation will be classified as such in Category 1, and according to CLP Annex I, Section3.4.2.2 substances fulfilling the criteria for skin sensitisation will be classified as such inCategory 1. In addition substances classified in Category 1 for skin sensitisation can beallocated specific concentration limits as described in Section 3.4.2.5.3.4.2.5 Setting of specific concentration limitsRespiratory sensitisers cannot be identified reliably on the basis of animal tests as yet, sinceno recognised validated test exists to determine sensitising potential and potency byinhalation. Therefore specific concentration limits (SCLs) cannot be set on the basis of animaldata. Moreover, there is no concept available to set SCLs on the basis of human data forrespiratory sensitisers.SCLs for skin sensitisation can be set based on the assumption that a substance can becategorised according to their skin sensitisation potency based on the results from animaltesting as reported under Section 3.4.2.3.4.1, 3.4.2.3.4.2 and 3.4.2.3.4.3 above. SCL are set onthe basis of testing of the substance and never on the basis of testing of a mixture containingthe sensitising substance (see 3.4.3.1.1 of Annex I).The generic concentration limit (GCL) for the classification of sensitisers in mixtures is 1%(CLP Annex I, Table 3.4.3). However, for certain sensitisers 1% is not sufficiently protective.Therefore a specific concentration limit (SCL) shall be set (CLP, Article 10) which will betterreflect the hazard of mixtures containing that skin sensitiser.SCLs shall be set when there is adequate and reliable scientific information available showingthat the specific hazard is evident below the GCL, 1%, for classification. As such the SCLshould normally be as suggested in Table 3.4.2.5. However, supported by reliable data theSCL could have some other value below 1%. Reliable data could be human data from e.g.work place studies where the exposure is defined.275

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012It is more difficult to prove the absence of sensitising properties at certain concentrationlevels. Therefore an SCL above the GCL, 1%, may only be set in exceptional circumstances,if scientific information is adequate, reliable and conclusive for that particular skin sensitiser.However there is currently no guidance on how to set SCL above the GCL.The concentration limits for skin sensitisers categorised according to their sensitisationpotency in the Table 3.4.2.5 are recommendations from an EU expert group on skinsensitisation (Basketter et al., 2005a).Table 3.4.2.5: Skin sensitising potency for substances and recommendations on specificconcentration limitsPotencyExtremeStrongModerateConcentration Limit (% w/v)0.001 (SCL)0.1 (SCL)1 (GCL)276

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.4.2.6 Decision logic for classification of substancesIt is strongly recommended that the person responsible for classification study the criteria forclassification before and during use of the decision logic.Decision logic for respiratory sensitisationDoes the substance have respiratory sensitisation data?NOClassificationnot possibleYESa) Is there evidence in humans that the substance can lead tospecific respiratory hypersensitivity, and/or(b) are there positive results from an appropriate animal test?YESCategory 1NODangerNot classified277

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Decision logic for skin sensitisationDoes the substance have skin sensitisation data?NOClassificationnot possibleYES(a) Is there evidence in humans that the substance can lead tosensitisation by skin contact in a substantial number ofpersons, or(b) are there positive results from an appropriate animal test?YESCategory 1NOWarningNot classified278

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.4.3 Classification of mixtures for respiratory or skin sensitisation3.4.3.1 General considerations for classificationThe same principles apply as for substances (see Section 3.4.2).3.4.3.2 Identification of hazard information for skin sensitisationFor identification of the sensitisation potential of a mixture the following information may beavailable:(a) test results on one or more, preferably all of its potentially sensitising components; or(b) test results on the mixture itself; or(c) test results of a similar mixture.Test methods are outlined in the Section 3.4.2.3.4. However, these animal tests have beendeveloped to identify sensitising substances and not mixtures. Therefore the results obtainedon mixtures need to be evaluated with care. For a mixture the cut-off in the mouse LLNAshould be seen as a threshold for identification of a sensitiser rather than as a threshold forsensitisation. A conclusion on the absence of sensitising potential of a mixture based on thenegative outcome in a test must be taken with great caution.On the other hand test data on a mixture take into account effects of possible interactions ofits components. For instance, it is known that presence of a vehicle may significantlyinfluence the skin sensitising potency, by influencing the penetration of the sensitisingcomponent(s) through the skin, (Basketter et al. 2001, Dearman et al. 1996, Heylings et al.1996) or through other mechanisms involved in the acquisition of sensitisation (Cumberbatchet al. 1993; Dearman et al. 1996).Repeated exposure to mixtures, that are non-sensitising under standard LLNA exposureconditions, might induce skin sensitisation, if the sensitising component in this mixture hassufficient accumulation potential in the skin to reach the minimum concentration for apositive effect (De Jong et al. 2007). Uncertainty also exists about the effect of such mixtureafter exposure on a larger skin area. Therefore additional information is important, if theoutcome of sensitisation tests on mixtures contrasts with the classification based on thecontent of sensitising component(s). For example, the validity of a well conducted LLNA on amixture with a negative outcome can scientifically be confirmed by spiking the test mixturewith another sensitiser (positive control) at different concentrations, or by showing a doseresponse relationship. Such LLNA tests could have been designed to provide suchinformation without use of extra animals. Additional animal testing for the purpose ofclassification and labelling shall be undertaken only where no other alternatives, whichprovide adequate reliability and quality of data, are possible (CLP Article 7(1)).3.4.3.3 Classification criteria3.4.3.3.1 When data are available for all components or only for some componentsAnnex I: 3.4.3.3.1. The mixture shall be classified as a respiratory or skin sensitiser when at leastone ingredient has been classified as a respiratory or skin sensitiser and is present at or above theappropriate generic concentration limit as shown in Table 3.4.3 below for solid/liquid and gasrespectively.3.4.3.3.2. Some substances that are classified as sensitisers may elicit a response, when present in amixture in quantities below the concentrations established in Table 3.4.1, in individuals who are279

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012already sensitised to the substance or mixture (see Note 1 to Table 3.4.3).Table 3.4.3Generic concentration limits of ingredients of a mixture classified as either skin sensitisers orrespiratory sensitisers that trigger classification of the mixtureConcentration triggering classification of a mixture as:Ingredient classified as:Skin SensitiserRespiratory SensitiserAll physical states Solid/Liquid GasSkin Sensitiser≥ 0,1 %(Note 1)≥ 1,0 %(Note 2)– –– –Respiratory Sensitiser–≥ 0,1 %(Note 1)≥ 0,1 %(Note 1)–≥ 1,0 %(Note 3)≥ 0,2 %(Note 3)Note 1This concentration limit is generally used for the application of the special labelling requirements ofAnnex II section 2.8 to protect already sensitised individuals. A SDS is required for the mixturecontaining an ingredient above this concentration.Note 2This concentration limit is used to trigger classification of a mixture as a skin sensitiser.Note 3This concentration limit is used to trigger classification of a mixture as a respiratory sensitiser.All sensitising components of a mixture at or above their generic or specific concentrationlimit should be taken into consideration for the purpose of classification. Specificconcentration limits (see Section 3.4.2.5) will always take precedence over the genericconcentration limits.The additivity concept is not applicable for respiratory or skin sensitisation, i.e. if one singleclassified substance is present in the mixture above the generic concentration limit, themixture must be classified for that hazard. If the mixture contains two substances each belowthe generic concentration limits, the mixture will not be classified, as far as no SCL has beenset.3.4.3.3.2 When data are available for the complete mixtureAnnex I: 3.4.3.1.1. When reliable and good quality evidence from human experience or appropriatestudies in experimental animals, as described in the criteria for substances, is available for themixture, then the mixture can be classified by weight-of-evidence evaluation of these data. Careshall be exercised in evaluating data on mixtures, that the dose used does not render the resultsinconclusive.280

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In case classification of a mixture is based on test results for the mixture as a whole, this datamust be shown to be conclusive. Especially it should be taken into account that in case of skinsensitisation current test methods are based on application of maximised dose, which only canbe obtained using a substance by itself and not diluted in a mixture.It is recognised that mixtures not showing sensitisation in a test, may still contain a lowconcentration of sensitising component.For specific guidance on the test methods and evaluation of the results see Section 3.4.2.3.4,Section 3.4.3.1and CLP Annex I, 3.4.3.1.1.3.4.3.3.3 When data are not available for the complete mixture: Bridging PrinciplesAnnex I: 3.4.3.2.1. Where the mixture itself has not been tested to determine its sensitisingproperties, but there are sufficient data on the individual ingredients and similar tested mixtures toadequately characterise the hazards of the mixture, these data shall be used in accordance with thebridging rules out in section 1.1.3.In absence of a test on the mixture, data from tests on a similar mixture, i.e. containing thesame sensitising component in a similar concentration, may be used for skin sensitisingpotential estimation.281

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.4.3.4 Decision logic for classification of mixturesIt is strongly recommended that the person responsible for classification study the criteria forclassification before and during use of the decision logic.Decision logic for respiratory sensitisationDoes the mixture as a whole or its ingredients haverespiratory sensitisation data?NOClassificationnot possibleYESDoes the mixture as a whole have respiratory sensitisation data?YESCategory 1NOa) Is there evidence in humans that the mixture can leadto specific respiratory hypersensitivity, and/or(b) are there positive results from an appropriate animaltest?YESDangerNOCan bridging principles beapplied?YESClassify inappropriatecategoryNOCare shall be be exercised exercised in evaluating evaluating data on mixtures, data onmixtures, that the dose that the used dose does used not does render not the render results theresults inconclusive. inconclusive. Is this the Is case? this the case?NONot classifiedYESDoes the mixture contain one or more ingredients classified asa respiratory sensitiser at ≥ 1.0% w/w (solid/liquid) or ≥ 0.2%v/v (gas) or above a SCL set for the ingredient(s)?YESCategory 1NODangerNot classified282

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Decision logic for skin sensitisationDoes the mixture as a whole or its ingredients have skinsensitisation data?NOClassificationnot possibleYESDoes the mixture as a whole have skin sensitisation data?YESCategory 1NO(a) Is there evidence in humans that the mixture can leadto sensitisation by skin contact in a substantial number ofpersons, orYESWarning(b) are there positive results from an appropriate animal?NOCan bridging principles beapplied?YESClassify inappropriatecategoryNOCare shall be exercised in evaluating data onmixtures, that the dose used does not render theresults inconclusive. Is this the case?NONot classifiedYESDoes the mixture contain one or more ingredients classified as a skinsensitiser at ≥ 1.0 % or above a SCL set for the ingredient(s)?YESCategory 1NOWarningNot classified283

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.4.4 Hazard communication for respiratory or skin sensitisation3.4.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 3.4.4.1. Label elements shall be used for substances or mixtures meeting thecriteria for classification in this hazard class in accordance with Table 3.4.2Table 3.4.4Respiratory or skin sensitisation label elementsClassificationGHS PictogramsRespiratory sensitisationSkin sensitisationCategory 1 Category 1Signal Word Danger WarningHazard StatementPrecautionary StatementPreventionPrecautionary StatementResponsePrecautionary StatementStoragePrecautionary StatementDisposal284H334: May cause allergy orasthma symptoms orbreathing difficulties ifinhaledP261P285P304 + P341P342 + P311P501H317: May cause an allergicskin reactionP261P272P280P302 + P352P333 + P313P321P363P501If the hazard pictogram “GHS08” applies for respiratory sensitisation, the hazard pictogram“GHS07” shall not appear for skin sensitisation or for skin and eye irritation (CLP, Article26).In the SDS for a substance, information on the generic or specific concentration limit shouldbe provided.3.4.4.2 Additional labelling provisionsAnnex II: 2.8. Mixtures not classified as sensitising but containing at least one sensitisingsubstanceThe label on the packaging of mixtures containing at least one substance classified as sensitisingand present in a concentration equal to or greater than 0,1% or in a concentration equal to or

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012greater than that specified under a specific note for the substance in part 3 of Annex VI shall bearthe statement:EUH208 - “Contains (name of sensitising substance). May produce an allergic reaction”3.4.5 Re-classification of substances and mixtures classified for respiratory or skinsensitisation according to DSD and DPD3.4.5.1 Is direct “translation” of classification and labellingpossible?Direct translation from DSD to CLP is possible for sensitising substances.Any existing SCLs may be transferred across to CLP and used for classification of mixtures.Where there is no existing SCL for an already classified substance, the substance shall beclassified in the default Category 1, and a generic concentration limit of 1% applied.3.4.5.2 Re-evaluation of the skin sensitisation dataRe-evaluation of non-tested mixtures has to be done on the basis of any relevant new data thatmight have become available after the time of the latest classification or if an SCL has beenset.3.4.6 Examples of classification for skin sensitisation3.4.6.1.1 Example 13.4.6.1 Example of substance fulfilling the criteria forclassification for skin sensitisationSubstance X gave a positive result in the LLNA with an EC3-value of 10.4%. As this EC3-value is above the cut-off of 2%, the substance is considered to be a moderate skin sensitiser,and should be classified as a Category 1 skin sensitiser. The GCL for classification ofmixtures containing substance X is 1%.3.4.6.1.2 Example 2Substance Y tested positive in the LLNA with an EC3-value of 0.5%. In the GPMT a dermalinduction concentration of 0.375% produced a positive response in 70% of the animals. Onthe basis of both these positive results, the substance is considered to be a strong sensitiserrequiring classification as a Category 1 skin sensitiser. A specific concentration limit of 0.1%is suggested.3.4.6.1.3 Example 3Herby is a herbicide formulation containing 28 g/l substance X, a moderate skin sensitiser(see example 1). There is no sensitisation data for the formulation itself. As Herby containsmore than the GCL (1%) of this sensitising a.i., and in the absence of any additionalinformation, it should be classified as a Category 1 skin sensitiser. The label must also bearthe statement EUH208.3.4.6.1.4 Example 4Methyl/Chloromethyl-isothiazolinone is an example of an extreme sensitiser. This substanceis listed in CLP Annex VI (Index-No. 613-167-00-5) with harmonised classification. Being anextreme sensitiser it has a specific concentration limit with regard to skin sensitisation, and285

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012due to this property any mixture containing the substance in a concentration ≥ 0.0015% mustbe classified with Skin Sens. 1. The label must also bear the statement EUH208.3.4.6.2.1 Example 52863.4.6.2 Example of substances or mixtures not fulfilling thecriteria for classification for skin sensitisationSubstance A was tested in the LLNA and gave a maximum stimulation index of 2.4. On thebasis of a stimulation index below 3, substance A is not considered to be a skin sensitiser, anddoes not require classification.3.4.6.2.2 Example 6Insecto super is an insecticide formulation containing 9 g/l substance X (see Example 1).Substance X is a moderate skin sensitiser (generic concentration limit in mixtures 1%). Basedon the classification of substance X, the insecticide formulation shall not be classified assensitising as the concentration of the a.i. is below the GCL of 1%. The label must bear thestatement EUH208.3.4.7 ReferencesBasketter DA, Andersen KE, Liden C, Van Loveren H, Boman A, Kimber I, Alanko K,Berggren E. (2005a): Evaluation of the skin sensitizing potency of chemicals by using theexisting methods and considerations for the relevance of elicitation. Contact Derm 52:39-43.Basketter DA, Clapp C, Jefferies D, Safford R, Ryan C, Gerberick G, Dearman R, Kimber I.(2005b): Predictive identification of human skin sensitisation thresholds. Contact Derm53:260-267.Basketter DA, Evans P, Fielder RJ, Gerberick GF, Dearman RJ, Kimber I. (2002a): Locallymph node assay – validation and use in practice. Fd Chem Toxic 40:593-598.Basketter DA, Wright ZM, Colson NR, Patlewicz GY, and Pease CK. (2002b): Investigationof the skin sensitizing activity of linalool. Contact Derm. 47(3): 161-164.Basketter DA, Gerberick GF, Kimber I. (2001): Skin sensitisation, vehicle effects and thelocal lymph node assay. Fd Chem Toxic 39: 621-627.Cumberbatch M, Scott RC, Basketter DA, Scholes EW, Hilton J, Dearman RJ, Kimber I(1993): Influence of sodium lauryl sulfate on 2.4-dinitrochlorobenzene-induced lymph nodeactivation. Toxicology 77: 181-191De Jong WH, De Klerk A, Beek MT, Veenman C, Van Loveren H. (2007): Effect of prolongedrepeated exposure to formaldehyde donors with doses below the EC3 value on draining lymph noderesponses. J Immunotoxicol. 4(3):239-46.Dearman RJ, Cumberbatch M, Hilton J, Clowes HM, Heylings JR, Kimber I (1996):Influence of dibutyl phthalate on dermal sensitization to fluorescein isothiocyanate. FundAppl Toxicol 33: 24-30Heylings JR, Clowes HM, Cumberbatch M, Dearman RJ, Fielding I, Hilton J, Kimber I(1996): Sensitization to 2,4-dinitrochlorobenzene: influence of vehicle on absorption andlymph node activation. Toxicology 109: 57-65Kimber I, Basketter DA, Butler M, Gamer A, Garrigue JL, Gerberick GF, Newsome C,Steiling W, Vohr HW,. (2003): Classification of contact allergens according to potency:Proposals. Fd Chem Toxic 41: 1799-1809.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Klimisch HJ, Andreae M, Tillmann U, (1997): A Systematic Approach for Evaluating theQuality of Experimental Toxicological and Ecotoxicological Data. Reg. Toxicol. Pharmacol.25:1-5.Robinson MK, Nusair TL, Fletcher ER, Ritz HL. (1990): A review of the Buehler guinea pigskin sensitization test and its use in a risk assessment process for human skin sensitization.Fundam Appl Toxicol 17: 103-119.Schlede E, Eppler R. (1995): Testing for skin sensitisation according to the notificationprocedure for new chemicals. The Magnusson and Kligman test. Contact Derm. 32: 1-4.Schneider K, Akkan Z. 2004. Quantitative relationship between the local lymph node assay andhuman skin sensitization assays. Regul Toxicol Pharmacol 39: 245-2553.5 GERM CELL MUTAGENICITY3.5.1 Definitions and general considerations for classification for germ cellmutagenicityAnnex I: 3.5.1.1. A mutation means a permanent change in the amount or structure of the geneticmaterial in a cell. The term “mutation” applies both to heritable genetic changes that may bemanifested at the phenotypic level and to the underlying DNA modifications when known(including specific base pair changes and chromosomal translocations). The term “mutagenic” and“mutagen” will be used for agents giving rise to an increased occurrence of mutations inpopulations of cells and/or organisms.3.5.1.2. The more general terms “genotoxic” and “genotoxicity” apply to agents or processes whichalter the structure, information content, or segregation of DNA, including those which cause DNAdamage by interfering with normal replication processes, or which in a non-physiological manner(temporarily) alter its replication. Genotoxicity test results are usually taken as indicators formutagenic effects.Germ cell mutations are those that occur in the egg or sperm cells (germ cells) and thereforecan be passed on to the organism's offspring. Somatic mutations are those that happen in cellsother than the germ cells, and they cannot be transmitted to the next generation. This is animportant distinction to keep in mind in terms of both the causes and the effects of mutation.Annex I: 3.5.2.1 This hazard class is primarily concerned with substances that may causemutations in the germ cells of humans that can be transmitted to the progeny. However, the resultsfrom mutagenicity or genotoxicity tests in vitro and in mammalian somatic and germ cells in vivoare also considered in classifying substances and mixtures within this hazard class.Annex I: 3.6.2.2 Specific considerations for classification of substances as carcinogens3.6.2.2.6. Mutagenicity: It is recognised that genetic events are central in the overall process ofcancer development. Therefore evidence of mutagenic activity in vivo may indicate that a substancehas a potential for carcinogenic effects.Hazard classification for germ cell mutagenicity primarily aims to identify substances causingheritable mutations or being suspected of causing heritable mutations. A secondary aim is thatthe hazard class germ cell mutagenicity offers supporting information with respect to theclassification of carcinogenic substances. This is expressed by the broad meaning of thehazard statements “H340: May cause genetic defects” and “H341: Suspected of causinggenetic defects” which comprises heritable genetic damage as well as somatic cell287

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012mutagenicity. Thus, classification as a germ cell mutagen (Category 1A, 1B, and.2) classifiesfor the hazard heritable genetic damage as well as providing an indication that the substancecould be carcinogenic.It is also warranted that where there is evidence of only somatic cell genotoxicity, substancesare classified as suspected germ cell mutagens. Classification as a suspected germ cellmutagen may also have implications for potential carcinogenicity classification. This holdstrue especially for those genotoxicants which are incapable of causing heritable mutationsbecause they cannot reach the germ cells (e.g. genotoxicants only acting locally, "site ofcontact” genotoxicants). This means that if positive results in vitro are supported by at leastone positive local in vivo, somatic cell test, such an effect should be considered as enoughevidence to lead to classification in Category 2. If there is also negative or equivocal data, aweight of evidence approach using expert judgement has to be applied.3.5.2 Classification of substances for germ cell mutagenicity2883.5.2.1 Identification of hazard information3.5.2.1.1 Identification of human dataOccasionally, studies of genotoxic effects in humans exposed by, for example, accident,occupation or participation in clinical studies (e.g. from case reports or epidemiologicalstudies) may be available. Generally, cells circulating in blood are investigated for theoccurrence of various types of genetic alterations; see alsoIR/CSA, Section R.7.7.3.2.3.5.2.1.2 Identification of non human dataAnimal dataSome test methods have an officially adopted EU/OECD guideline for the testing procedure,although for many test methods this is not the case. Furthermore, modifications to OECDprotocols have been developed for various classes of substances and may serve to enhance theaccuracy of test results. Use of such modified protocols is a matter of expert judgement andwill vary as a function of the chemical and physical properties of the substance to beevaluated. Commonly used non-guideline in vivo tests employ methods by which any tissueof an animal can be examined for effects on the genetic material, giving the possibility toexamine site-of-contact tissues (i.e., skin, epithelium of the respiratory or gastro-intestinaltract) in genotoxicity testing. In addition, test methods developed over the past decades inDrosophila and in various species of plants and fungi are available; see also IR/CSA, SectionR.7.7.3.Other in vivo tests in somatic cells which provide supporting evidence ongenotoxicity/mutagenicity may include, for example, a Comet single cell gel electrophoresisassay for DNA strand breaks, or a test for gene mutations in transgenic rodent models usingreporter genes.With the exception of in vivo studies proving “site of contact” effects, genotoxicity data fromsuch non-standard in vivo studies are not sufficient but may offer supporting information forclassification.In vitro dataTypically, in vitro tests are performed with cultured bacterial cells, human or othermammalian cells. The sensitivity and specificity of tests will vary with different classes ofsubstances; see alsoIR/CSA, Section R.7.7.3.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Use of other dataSee IR/CSA, Section R.7.3.3.1.Existing test methodsSee IR/CSA, Section R.7.3.3.1.3.5.2.2 Classification criteria for substancesAnnex I: 3.5.2.2. For the purpose of classification for germ cell mutagenicity, substances areallocated to one of two categories as shown in Table 3.5.1.Table 3.5.1Hazard categories for germ cell mutagensCategoriesCATEGORY 1:Category 1A:Category 1B:CATEGORY 2:CriteriaSubstances known to induce heritable mutations or to be regarded asif they induce heritable mutations in the germ cells of humans.Substances known to induce heritable mutations in the germ cells ofhumans.The classification in Category 1A is based on positive evidence fromhuman epidemiological studies.Substances to be regarded as if they induce heritable mutations in thegerm cells of humans.The classification in Category 1B is based on:– positive result(s) from in vivo heritable germ cell mutagenicitytests in mammals; or– positive result(s) from in vivo somatic cell mutagenicity tests inmammals, in combination with some evidence that the substancehas potential to cause mutations to germ cells. It is possible toderive this supporting evidence from mutagenicity/genotoxicitytests in germ cells in vivo, or by demonstrating the ability of thesubstance or its metabolite(s) to interact with the genetic materialof germ cells; or– positive results from tests showing mutagenic effects in the germcells of humans, without demonstration of transmission toprogeny; for example, an increase in the frequency of aneuploidyin sperm cells of exposed people.Substances which cause concern for humans owing to the possibilitythat they may induce heritable mutations in the germ cells of humans.The classification in Category 2 is based on:– Positive evidence obtained from experiments in mammals and/orin some cases from in vitro experiments, obtained from:– Somatic cell mutagenicity tests in vivo, in mammals; or– Other in vivo somatic cell genotoxicity tests which aresupported by positive results from in vitro mutagenicity assays.Note: Substances which are positive in in vitro mammalianmutagenicity assays, and which also show chemical structure activityrelationship to known germ cell mutagens, shall be considered forclassification as Category 2 mutagens.289

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.5.2.3 Evaluation of hazard informationAnnex I: 3.5.2.3.3 Classification for heritable effects in human germ cells is made on the basis ofwell conducted, sufficiently validated tests, preferably as described in Regulation (EC) No440/2008 adopted in accordance with Article 13(3) of Regulation (EC) No 1907/2006 (‘TestMethod Regulation’) such as those listed in the following paragraphs. Evaluation of the test resultsshall be done using expert judgement and all the available evidence shall be weighed in arriving ata classification.3.5.2.3.1 Evaluation of human dataHuman data have to be assessed carefully on a case-by-case basis. The interpretation of suchdata requires considerable expertise. Attention should be paid especially to the adequacy ofthe exposure information, confounding factors, co-exposures and to sources of bias in thestudy design or incident. The statistical power of the test may also be considered (SeeIR/CSA, Section R.7.4.4.2).3.5.2.3.2 Evaluation of non human dataEvaluation of genotoxicity test data should be made with care. Regarding positive findings,responses generated only at highly toxic/cytotoxic concentrations should be interpreted withcaution, and the presence or absence of a dose-response relationship should be considered. Incase of negative findings in vivo toxicokinetic and other available information should beconsidered e.g. to verify whether the substance has reached the target organ (for detailedguidance see See IR/CSA, Section R.7.7.4.1).3.5.2.4 Decision on classificationAnnex I: 3.5.2.3.1. To arrive at a classification, test results are considered from experimentsdetermining mutagenic and/or genotoxic effects in germ and/or somatic cells of exposed animals.Mutagenic and/or genotoxic effects determined in in vitro tests shall also be considered.Annex I: 3.5.2.3.9. The classification of individual substances shall be based on the total weight ofevidence available, using expert judgement (See 1.1.1). In those instances where a single wellconductedtest is used for classification, it shall provide clear and unambiguously positive results. Ifnew, well validated, tests arise these may also be used in the total weight of evidence to beconsidered. The relevance of the route of exposure used in the study of the substance compared tothe most likely route of human exposure shall also be taken into account.Classification as a Category 1A mutagenEpidemiological studies have been to date unable to provide evidence to classify a substanceas a Category 1A mutagen. Hereditary diseases in humans for the most part have an unknownorigin and show a varying distribution in different populations. Due to the random distributionof mutations in the genome it is not expected that one particular substance would induce onespecific genetic disorder. Therefore, it is unlikely that such evidence may be obtained byepidemiological studies to enable you to classify a substance as a Category 1A mutagen.Classification as a Category 1B mutagenClassification in Category 1B may be based on positive results of at least one valid in vivomammalian germ cell mutagenicity test. In case there are also negative or equivocal data, aweight of evidence approach using expert judgement has to be applied.290

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012If there are only positive results of at least one valid in vivo mammalian somatic mutagenicitytest but no respective data on mammalian germ cells are available, additional evidence isrequired to be able to classify as mutagen in Category 1B. Such additional data must provethat the substance or its metabolite(s) interacts in vivo with the genetic material of germ cells.It is also possible to obtain supporting evidence in an in vivo genotoxicity test withmammalian germ cells. In addition, genetic damage to germ cells in exposed humans provento be caused by substance exposure may offer respective information. In case of othersupporting evidence or where there are also negative or equivocal data, a weight of evidenceapproach using expert judgement has to be applied.It could be argued that in a case where in vivo mutagenicity/genotoxicity is proven and thesubstance under consideration is systemically available, then that substance should also beconsidered as a Category 1B mutagen. Germ cell mutagens as the spermatogonia aregenerally not protected from substance exposure by the blood-testes barrier formed by theSertoli cells. In such circumstances the relevant criteria are as follows:Annex I: 3.5.2.2. (extract from Table 3.5.1)Category 1B…– positive result(s) from in vivo somatic cell mutagenicity tests in mammals, in combination withsome evidence that the substance has potential to cause mutations to germ cells. It is possible toderive this supporting evidence from mutagenicity/genotoxicity tests in germ cells in vivo, or bydemonstrating the ability of the substance or its metabolite(s) to interact with the geneticmaterial of germ cells;…This wording expresses that supporting evidence in addition to an in vivo somatic cellmutagenicity test in mammals is needed to be able to classify a substance in a Category 1Bmutagen. The second sentence in the green box above gives examples for such evidence, fromthese examples it is clear that such supporting evidence is experimental evidence. There hasto be either data indicating that germ cell mutagenicity/genotoxicity is caused by thesubstance or data showing that the substance or its metabolite(s) interact with the geneticmaterial of germ cells. Thus, in such circumstances, in addition to an in vivo somatic cellmutagenicity test, further experimental evidence is needed to be able to classify a substance asa Category 1B mutagen.Classification as a Category 2 mutagenClassification in Category 2 may be based on positive results of a least one in vivo validmammalian somatic cell mutagenicity test, indicating mutagenic effects in somatic cells. ACategory 2 mutagen classification may also be based on positive results of a least one in vivovalid mammalian genotoxicity test, supported by positive in vitro mutagenicity results.Genetic damage to somatic cells in exposed humans shown to be caused by substanceexposure supported by positive in vitro mutagenicity results may also offer respectiveinformation warranting classification as a Category 2 mutagen. In vitro results can only leadto a Category 2 mutagen classification in a case where there is support by chemical structureactivity relationship to known germ cell mutagens. In the case where there are also negativeor equivocal data, a weight of evidence approach using expert judgement has to be applied.In general, mutations can be differentiated into gene mutations (e.g. point or frame shiftmutation), chromosome mutations (structural chromosome changes) and genome mutations(loss or gain of whole chromosomes). Different mutagenicity tests may detect different typesof mutations and genotoxic effects which have to be taken into account in the weight ofevidence determination. For instance, a substance which only causes chromosome mutations291

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012may be negative in a test for detecting point mutations. A complex data situation with positiveand negative results might still lead to classification. This is because all tests detecting acertain type of mutation (e.g. point mutations) have been positive and all tests detectingchromosome mutations have been negative. Such circumstances clearly warrant classificationalthough several tests have been negative which is plausible in this case.A positive result for somatic or germinal mutagenicity in a test using intraperitonealadministration only shows that the tested substance has an intrinsic mutagenic property, andthe fact that negative results are exhibited by other routes of dosage may be related to factorsinfluencing the distribution/ metabolism of the substance which may be characteristic to thetested animal species. It cannot be ruled out that a positive test result in intraperitoneal studiesin rodents only may be relevant to humans.If there are positive results in at least one valid in vivo mutagenicity test using intraperitonealapplication, or from at least one valid in vivo genotoxicity test using intraperitonealapplication plus supportive in vitro data, classification is warranted. In cases where there areadditional data from further in vivo tests with oral, dermal or inhalative substance application,a weight of evidence approach using expert judgement has to be applied in order to to come toa decision. For instance, it may be difficult to reach a decision on whether or not to classify inthe case where there are positive in vivo data from at least one in vivo test usingintraperitoneal application but (only) negative test data from (an) in vivo test(s) using oral,dermal, or inhalative application. In such a case, it could be argued thatmutagenicity/genotoxicity can only be shown at internal body substance concentrations whichcan not be achieved using application routes other than intraperitoneal. However, it also has tobe taken into account that there is generally no threshold for mutagenicity unless there isspecific proof for the existence of such a threshold as may be the case for aneugens. Thus, ifmutagenicity/genotoxicity can only be demonstrated for the intraperitoneal route exclusively,then this may mean that the effect in the in vivo tests using application routes other thanintraperitoneal may have been present, but it may not have been detected because it wasbelow the detection limit of the oral, dermal, or inhalative test assays.In summary, classification as a Category 2 mutagen would generally apply if onlyintraperitoneal in vivo tests show mutagenicity/genotoxicity and the negative test results fromthe in vivo tests using other routes of application are plausible. Factors influencing plausibilityare e.g. the doses tested and putative kinetic data on the test substance. However, on a caseby-caseanalysis using a weight of evidence approach and expert judgement, nonclassificationmay also result.2923.5.2.5 Setting of specific concentration limitsThere is no detailed and accepted guidance developed for the setting of specific concentrationlimits (SCLs) for mutagenicity, as is the case for carcinogenic substances. Guidance such asthe T 25 concept for carcinogens covering all relevant aspects would need to be developed inorder to derive SCLs for mutagens in a standardized manner. There are several reasons why itis considered impossible to set SCLs for mutagens without a comprehensive guidance, one ofthem being that mutagenicity tests have not been specifically developed for the derivation of aquantitative response. Moreover, different mutagenicity tests have different sensitivities indetecting mutagens. Thus, it is very difficult to describe the minimum data requirementswhich would allow a standardized SCL derivation. Another drawback in practice is that theresults obtained for the most part do not offer sufficient information on dose-response,especially in the case for in vivo tests. In conclusion, the possibility to set SCL for germ cellmutagenicity is therefore not considered possible in the process of self-classification as thereis no standardized methodical approach available which adequately takes into account allrelevant information.

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.5.2.6 Decision logic for substancesThe decision logic which follows is provided as additional guidance. It is stronglyrecommended that the person responsible for classification study the criteria before andduring use of the decision logic.Does the substance have data on mutagenicity?YESNOClassificationnot possibleAccording to the criteria, is the substance:(a) Known to induce heritable mutations in germ cellsof humans, or(b) Should it be regarded as if it induces heritablemutations in the germ cells of humans?Application of the criteria needs expert judgement in aweight of evidence approach.YESCategory 1DangerNOAccording to the criteria, does the substance cause concernfor humans owing to the possibility that it may induceheritable mutations in the germ cells of humans?Application of the criteria needs expert judgement in aweight of evidence approach.YESCategory 2WarningNONot classified293

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.5.3 Classification of mixtures for germ cell mutagenicity2943.5.3.1 Classification criteria for mixturesClassification of mixtures will be based on the available test data for the individualingredients of the mixture, using concentration limits for those ingredients. Under rarecircumstances, the classification may be modified on a case-by-case basis based on theavailable test data for the mixture as a whole or based on bridging principles (see Article6(3)).3.5.3.1.1 When data are available for the complete mixtureAnnex I: 3.5.3.2.1. Classification of mixtures will be based on the available test data for theindividual ingredients of the mixture using concentration limits for the ingredients classified asgerm cell mutagens. On a case-by-case basis, test data on mixtures may be used for classificationwhen demonstrating effects that have not been established from the evaluation based on theindividual ingredients. In such cases, the test results for the mixture as a whole must be shown to beconclusive taking into account dose and other factors such as duration, observations, sensitivity andstatistical analysis of germ cell mutagenicity test systems. Adequate documentation supporting theclassification shall be retained and made available for review upon request.3.5.3.1.2 When data are not available for the complete mixture: bridging principlesAnnex I: 3.5.3.3.1. Where the mixture itself has not been tested to determine its germ cellmutagenicity hazard, but there are sufficient data on the individual ingredients and similar testedmixtures (subject to paragraph 3.5.3.2.1), to adequately characterise the hazards of the mixture,these data shall be used in accordance with the applicable bridging rules set out in section 1.1.3.3.5.3.2 Generic concentration limits for substances triggeringclassification of mixturesAnnex I: 3.5.3.1.1. The mixture shall be classified as a mutagen when at least one ingredient hasbeen classified as a Category 1A, Category 1B or Category 2 mutagen and is present at or above theappropriate generic concentration limit as shown in Table 3.5.2 for Category 1A, Category 1B andCategory 2 respectively.Table 3.5.2Generic concentration limits of ingredients of a mixture classified as germ cell mutagens thattrigger classification of the mixture.Concentration limits triggering classification of a mixture as:Ingredient classified as: Category 1A mutagen Category 1B mutagen Category 2 mutagenCategory 1A mutagen ≥ 0,1 % — —Category 1B mutagen — ≥ 0,1 % —Category 2 mutagen — — ≥ 1,0 %NoteThe concentration limits in the table above apply to solids and liquids (w/w units) as well as gases(v/v units).The option to set SCL for germ cell mutagenicity is not considered possible in the process ofself-classification as there is no standardized methodical approach available which adequatelytakes into account all relevant information (See Section 3.5.2.5).

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.5.3.3 Decision logic for mixturesThe decision logic which follows is provided as additional guidance. It is stronglyrecommended that the person responsible for classification study the criteria before andduring use of the decision logic. This decision logic deviates (slightly) from the original GHSguidance, to meet CLP requirements.295

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Classification based on individual ingredients of the mixtureCategoryDoes the mixture contain one or more ingredientsclassified as a Category 1 mutagen at 0.1%?YESNODangerDoes the mixture contain one or more ingredientsclassified as a Category 2 mutagen at 1.0%?YESCategory 2NONot classifiedWarningModified classification on a case-by-case basisTest data on mixtures may be used for classification when demonstrating effects that have notbeen established from the evaluation based on the individual ingredients (CLP Section3.5.3.2.1, see also CLP Article 6(3)).Are test data availablefor the mixture itselfdemonstrating amutagenic effect notidentified from thedata on individualsubstances?NOYESNOAre the test results on themixture conclusive takinginto account dose andother factors such asduration, observations andanalysis (e.g. statisticalanalysis, test sensitivity) ofgerm cell mutagenicity testsystems?YESClassify inappropriatecategoryDanger orWarningCan bridging principlesbe applied?YESorNo classificationNOSee above: Classification based onindividual ingredients of the mixture.296

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.5.4 Hazard communication in form of labelling for germ cell mutagenicity3.5.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 3.5.4.1. Label elements shall be used in accordance with Table 3.5.3, for substances ormixtures meeting the criteria for classification in this hazard class.Table 3.5.3Label elements of germ cell mutagenicityClassification Category 1A or Category 1B Category 2GHS PictogramsSignal Word Danger WarningHazard StatementPrecautionary StatementPreventionPrecautionary StatementResponsePrecautionary StatementStoragePrecautionary StatementDisposalH340: May cause genetic defects(state route of exposure if it isconclusively proven that no otherroutes of exposure cause thehazard)P201P202P281P308 + P313P405P501H341: Suspected of causinggenetic defects (state route ofexposure if it is conclusivelyproven that no other routes ofexposure cause the hazard)P201P202P281P308 + P313The hazard statement to be applied for the classification germ cell mutagenicity has to beamended to state the route of exposure if it is conclusively proven that no other routes ofexposure will lead to the respective effect. A conclusive proof means that valid in vivo testdata need to be available for all three exposure routes clearly indicating that only oneexposure route leads to positive results. Moreover, such findings should be plausible withrespect to the mode of action. It is estimated that such circumstances rarely, if ever, exist.Therefore, amending the hazard statement with the route of exposure generally does not haveto be considered.3.5.4.2 Additional labelling provisionsThere are no additional labelling provisions for substances and mixtures classified for germcell mutagenicity in CLP, however there are provisions laid out in Annex XVII to REACH.The packaging of substances classified for germ cell mutagenicity Category 1A or Category1B, and mixtures containing such substances, "must be marked visibly, legibly and indeliblyas follows: ‘Restricted to professional users’." (REACH, Annex XVII, point 29).P405P501297

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.5.5 Re-classification of substances classified for germ cell mutagenicity accordingto DSD and DPDDirect translation of classification and labelling is generally possible for substances andmixtures classified as germ cell mutagens.In CLP, there is clear discrimination of in vivo mutagenicity tests and in vivo genotoxicitytests with respect to their relevance for classification. Moreover, in some circumstances whichare assumed to occur very rarely if at all, a different classification may be the consequence ifexpert judgement is not applied.For instance, positive results from studies showing mutagenic effects in germ cells of exposedhumans can lead to classification as a Category 1B mutagen under CLP. However, using thecriteria in DSD it is not clear how to classify in such a case. Moreover, in vivo somatic cellgenotoxicity tests need to be supported by in vitro data in order to classify as a Category 2mutagen under CLP. In such circumstances under DSD, in vivo data do not necessarily needto be supported by in vitro data. However, it has to be taken into account that suchcircumstances will rarely occur as the testing strategy uses in vitro tests as a starting point.3.6 CARCINOGENICITY3.6.1 Definitions and general considerations for classification for carcinogenicityAnnex I: 3.6.1.1. Carcinogen means a substance or a mixture of substances which induce cancer orincrease its incidence. Substances which have induced benign and malignant tumours in wellperformed experimental studies on animals are considered also to be presumed or suspected humancarcinogens unless there is strong evidence that the mechanism of tumour formation is not relevantfor humans.More explicitly, chemicals are defined as carcinogenic if they induce tumours, increasetumour incidence and/or malignancy or shorten the time to tumour occurrence. Benigntumours that are considered to have the potential to progress to malignant tumours aregenerally considered along with malignant tumours. Chemicals can potentially induce cancerby any route of exposure (e.g., when inhaled, ingested, applied to the skin or injected), butcarcinogenic potential and potency may depend on the conditions of exposure (e.g., route,level, pattern and duration of exposure).Carcinogenic chemicals have conventionally been divided according to the presumed mode ofaction; genotoxic or non-genotoxic, see 3.6.2.3.2 (k).Classification of a substance as a carcinogen is based on consideration of the strength of theevidence of available data for classification with considerations of all other relevantinformation (weight of evidence) being taken into account as appropriate. Strength ofevidence involves the enumeration of tumours in human and animal studies and determinationof their level of statistical significance. A number of other factors need to be considered thatinfluence the overall likelihood that a substance poses a carcinogenic hazard in humans(weight of evidence determination). The list of factors for additional consideration is long andrequires the most up-to-date scientific knowledge. It is recognised that, in most cases, expertjudgement is necessary to be able to determine the most appropriate category forclassification for carcinogenicity.298

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.6.2 Classification of substances for carcinogenicity3.6.2.1 Identification of hazard informationCarcinogens may be identified from epidemiological studies, from animal experiments and/orother appropriate means that may include (Quantitative) Structure-Activity Relationships((Q)SAR) analyses and/or extrapolation from structurally similar substances (read-across). Inaddition some information on the carcinogenic potential can be inferred from in vivo and invitro germ cell and somatic cell mutagenicity studies, in vitro cell transformation assays, andgap junction intercellular communication (GJIC) tests.Extensive guidance on data requirements, information sources and strategies for theidentification of potential carcinogens are given in Section R.7.7.9 (Information requirementson carcinogenicity) and Section R.7.7.10 (Information and its sources on carcinogenicity) andfor potential mutagens Section R.7.7.3 (Information and its sources on mutagenicity),IR/CSA.For more about non testing data see Section 3.6.2.3.4.3.6.2.2 Classification criteria for substancesSubstances are classified according to their potential to cause cancer in humans. In somecases there will be direct evidence on the carcinogenicity to humans from epidemiologicalstudies. However, in most cases the available information on carcinogenicity will be primarilyfrom animal studies. In this case the relevance of the findings in animals to humans must beconsidered.299

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: 3.6.2.1. For the purpose of classification for carcinogenicity, substances are allocated toone of two categories based on strength of evidence and additional considerations (weight ofevidence). In certain instances, route-specific classification may be warranted, if it can beconclusively proved that no other route of exposure exhibits the hazard.Table 3.6.1Hazard categories for carcinogensCategoriesCATEGORY 1:Category 1A:Category 1B:CriteriaKnown or presumed human carcinogensA substance is classified in Category 1 for carcinogenicity on the basisof epidemiological and/or animal data. A substance may be furtherdistinguished as:Category 1A, known to have carcinogenic potential for humans,classification is largely based on human evidence, orCategory 1B, presumed to have carcinogenic potential for humans,classification is largely based on animal evidence.The classification in Category 1A and 1B is based on strength ofevidence together with additional considerations (see section 3.6.2.2).Such evidence may be derived from:– human studies that establish a causal relationship between humanexposure to a substance and the development of cancer (knownhuman carcinogen); or– animal experiments for which there is sufficient ( 1 ) evidence todemonstrate animal carcinogenicity (presumed human carcinogen).In addition, on a case-by-case basis, scientific judgement may warrant adecision of presumed human carcinogenicity derived from studiesshowing limited evidence of carcinogenicity in humans together withlimited evidence of carcinogenicity in experimental animals.CATEGORY 2:( 1 ) Note: See 3.6.2.2.4.Suspected human carcinogensThe placing of a substance in Category 2 is done on the basis ofevidence obtained from human and/or animal studies, but which is notsufficiently convincing to place the substance in Category 1A or 1B,based on strength of evidence together with additional considerations(see section 3.6.2.2). Such evidence may be derived either from limited( 1 ) evidence of carcinogenicity in human studies or from limitedevidence of carcinogenicity in animal studies.3.6.2.3 Evaluation of hazard informationAnnex I: 3.6.2.2.1. Classification as a carcinogen is made on the basis of evidence from reliable andacceptable studies and is intended to be used for substances which have an intrinsic property to causecancer. The evaluations shall be based on all existing data, peer-reviewed published studies andadditional acceptable data.3.6.2.2.2. Classification of a substance as a carcinogen is a process that involves two interrelateddeterminations: evaluations of strength of evidence and consideration of all other relevant informationto place substances with human cancer potential into hazard categories.300

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Classification of a substance as a carcinogen requires expert judgement and consideration ofmany different factors (weight and strength of evidence) included in the hazard informationon carcinogenicity. The guidance provides an approach to data analysis rather than hard andfast rules. A stepwise approach to the classification can be taken where all the factors, bothweight and strength of evidence, that may influence the outcome are consideredsystematically. Such approach, including consideration of these factors is outlined, inMcGregor et al, 2009 and Boobis et al, 2006. Also the IPCS “Conceptual Framework forEvaluating a Mode of Action for Chemical carcinogenesis” (2001), ILSI “Framework forHuman Relevance Analysis of Information on Carcinogenic Modes of Action” (Meek et al.,2003; Cohen et al, 2003, 2004) and the International Agency for Research on Cancer (IARC,2006 - Preamble Section B) provide a basis for systematic assessments which may beperformed in a consistent fashion internationally; however they are not intended to providelists of criteria to be checked off.Specific considerations that are necessary are outlined in CLP, Annex I, Section 3.6.2.3 (seeSection 3.6.2.3.1) and other important factors to consider in CLP, annex I, Section 3.6.6.2.6(see Section 3.6.2.3.2). Further guidance on these important factors is given in this document.3.6.2.3.1 Specific considerations for classificationThere is a strong link between CLP and the IARC classification criteria. The definitions forsufficient and limited evidence as defined by IARC are part of the criteria (Annex I, section3.6.2.2.3). IARC, however, understands the criteria of “sufficient" and "limited" as follows:‘It is recognized that the criteria for these evaluations, described below, cannot encompass allof the factors that may be relevant to an evaluation of carcinogenicity. In considering all ofthe relevant scientific data, the Working Group may assign the agent to a higher or lowercategory than a strict interpretation of these criteria would indicate.’ (IARC 2006 preambleSection 6, Evaluation and rationale). This sentence emphasises that in certain circumstancesexpert judgement may overrule the strict interpretation of the IARC criteria for "sufficient"and "limited". These same limitations apply with the current criteria in that expert judgementis necessary and can override the strict interpretation of the definitions.Annex I: 3.6.2.2.3. Strength of evidence involves the enumeration of tumours in human and animalstudies and determination of their level of statistical significance. Sufficient human evidencedemonstrates causality between human exposure and the development of cancer, whereas sufficientevidence in animals shows a causal relationship between the substance and an increased incidenceof tumours. Limited evidence in humans is demonstrated by a positive association betweenexposure and cancer, but a causal relationship cannot be stated. Limited evidence in animals isprovided when data suggest a carcinogenic effect, but are less than sufficient. The terms 'sufficient'and 'limited' have been used here as they have been defined by the International Agency forResearch on Cancer (IARC) and read as follows:(a) Carcinogenicity in humansThe evidence relevant to carcinogenicity from studies in humans is classified into one of thefollowing categories:– sufficient evidence of carcinogenicity: a causal relationship has been established betweenexposure to the agent and human cancer. That is, a positive relationship has been observedbetween the exposure and cancer in studies in which chance, bias and confounding couldbe ruled out with reasonable confidence;– limited evidence of carcinogenicity: a positive association has been observed betweenexposure to the agent and cancer for which a causal interpretation is considered to becredible, but chance, bias or confounding could not be ruled out with reasonable301

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012confidence.(b) Carcinogenicity in experimental animalsCarcinogenicity in experimental animals can be evaluated using conventional bioassays,bioassays that employ genetically modified animals, and other in-vivo bioassays that focus onone or more of the critical stages of carcinogenesis. In the absence of data from conventionallong-term bioassays or from assays with neoplasia as the end-point, consistently positiveresults in several models that address several stages in the multistage process of carcinogenesisshould be considered in evaluating the degree of evidence of carcinogenicity in experimentalanimals. The evidence relevant to carcinogenicity in experimental animals is classified into oneof the following categories:– sufficient evidence of carcinogenicity: a causal relationship has been established betweenthe agent and an increased incidence of malignant neoplasms or of an appropriatecombination of benign and malignant neoplasms in (a) two or more species of animals or(b) two or more independent studies in one species carried out at different times or indifferent laboratories or under different protocols. An increased incidence of tumours inboth sexes of a single species in a well-conducted study, ideally conducted under GoodLaboratory Practices, can also provide sufficient evidence. A single study in one speciesand sex might be considered to provide sufficient evidence of carcinogenicity whenmalignant neoplasms occur to an unusual degree with regard to incidence, site, type oftumour or age at onset, or when there are strong findings of tumours at multiple sites;– limited evidence of carcinogenicity: the data suggest a carcinogenic effect but are limitedfor making a definitive evaluation because, e.g. (a) the evidence of carcinogenicity isrestricted to a single experiment; (b) there are unresolved questions regarding the adequacyof the design, conduct or interpretation of the studies; (c) the agent increases the incidenceonly of benign neoplasms or lesions of uncertain neoplastic potential; or (d) the evidenceof carcinogenicity is restricted to studies that demonstrate only promoting activity in anarrow range of tissues or organs.For human studies, the quality and power of the epidemiology studies require expertconsideration and would normally lead to a Category 1A classification if data of adequatequality shows causality of exposure and cancer development. IR/CSA, Section R.7.7.10.2,further discusses the types of human epidemiology data available and the limitations of thedata. Where there is sufficient doubt in the human data then classification in Category 1B maybe more appropriate. On the other hand epidemiological studies may fail, because ofuncertainties in the exposure assessment and/or limited sensitivity and statistical power, toconfirm the carcinogenic properties of a substance as identified in animal studies (WHOWorking group, 2000).3.6.2.3.2 Additional considerations for classificationAnnex I: 3.6.2.2.4. Additional considerations (as part of the weight of evidence approach (see1.1.1)). Beyond the determination of the strength of evidence for carcinogenicity, a number of otherfactors need to be considered that influence the overall likelihood that a substance poses acarcinogenic hazard in humans. The full list of factors that influence this determination would bevery lengthy, but some of the more important ones are considered here.3.6.2.2.5. The factors can be viewed as either increasing or decreasing the level of concern forhuman carcinogenicity. The relative emphasis accorded to each factor depends upon the amountand coherence of evidence bearing on each. Generally there is a requirement for more completeinformation to decrease than to increase the level of concern. Additional considerations should beused in evaluating the tumour findings and the other factors in a case-by-case manner.3.6.2.2.6. Some important factors which may be taken into consideration, when assessing the302

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012overall level of concern are:(a)(b)(c)(d)(e)(f)(g)(h)(i)(j)(k)tumour type and background incidence;multi-site responses;progression of lesions to malignancy;reduced tumour latency;whether responses are in single or both sexes;whether responses are in a single species or several species;structural similarity to a substance(s) for which there is good evidence of carcinogenicity;routes of exposure;comparison of absorption, distribution, metabolism and excretion between test animals andhumans;the possibility of a confounding effect of excessive toxicity at test doses;mode of action and its relevance for humans, such as cytotoxicity with growth stimulation,mitogenesis, immunosuppression, mutagenicity.As indicated above, the evaluation of animal carcinogenicity data requires consideration of anumber of important additional factors which may increase or decrease the level of concernand the classification category. The list in Annex I, 3.6.2.6 is not exhaustive. Each of thesefactors is discussed individually below.(a) Tumour type and background incidenceKnowledge about the tumour type including its tumour biology is indispensable to decide onthe relevance of observed tumours for humans.By default, carcinogenic effects in experimental animals are considered relevant to humansand are considered for classification as carcinogens. Only when there is sufficient evidenceshowing that a certain type of tumour is not relevant to humans should this tumour type beexcluded for classification.Certain tumour types observed in animal carcinogenicity studies are of questionable or norelevance to humans. In case of multiple tumours anticipated to have no relevance forhumans justification should be given for each tumour type. The justification for dismissingany particular tumour should be presented as a scientifically robust and transparent argument.There are several reasons why a tumour observed in animals may be judged to be not relevantfor humans or may be judged to be of lower concern. In most of these cases the tumour arisesvia a mode of action which does not occur in humans (see this Section part k). In some casesthe tumour may arise in a tissue known to be overly susceptible in the species tested todevelopment of certain tumours and consequently may be judged to be less relevant forhumans. In a few cases a tumour may occur in a tissue with no equivalent in humans.Tumours occurring in tissues with no human equivalentSome of the commonly used animal species have some tissues with no equivalent in humans.Tumours occurring in these tissues include the following– Forestomach tumours in rodents following administration by gavage of irritating orcorrosive, non mutagenic substances. In rodents, the stomach is divided into two parts bythe muco-epidermoid junction separating squamous from glandular epithelium. Theproximal part, or forestomach, is non-glandular, forms a continuum with the oesophagus,and is lined by keratinized, stratified squamous epithelium. While humans do not have a303

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012forestomach, they do have comparable squamous epithelial tissues in the oral cavity andthe upper two-thirds of the oesophagus. See also this Section (k), IARC (2003), and RIVM(2003).– Tumours in the Zymbal’s glands. Zymbal’s glands are located beneath squamousepithelium at the anterior and posterior aspect of the ear canal. The external portion of thegland in rats is 3 to 5 millimetres in diameter.– Tumours in the Harderian glands. Harderian glands are found in all vertebrates that possessa nictitating membrane, or third eyelid. They are located behind the eyeball in the orbitnictitating membrane, encircling the optic nerve. Humans have a rudimentary one.Tumours occurring in such tissues indicate that the substance has the potential to inducecarcinogenic effects in the species tested. It cannot automatically be ruled out that thesubstance could cause similar tumours of comparable cell/tissue origin (e.g. squamous celltumours at other epithelial tissues) in humans. Careful consideration and expert judgement ofthese tumours in the context of the complete tumour response (i.e. if there are also tumours atother sites) and the assumed mode of action is required to decide if these findings wouldsupport a classification. However, tumours observed only in these tissues, with no otherobserved tumours are unlikely to lead to classification. However, such determinations must beevaluated carefully in justifying the carcinogenic potential for humans; any occurrence ofother tumours at distant sites must also be considered.Considering the background incidence and use of historical control dataAny statistically significant increase in tumour incidence, especially where there is a doseresponserelationship, is generally taken as positive evidence of carcinogenic activity.However, in some cases the results involve an increase incidence of tumours in treatedanimals which lies at the borderline of biological and/or statistical significance or there is anincrease in a spontaneous tumour type, then comparison of the tumour incidence withhistorical control tumour data is strongly encouraged.Historical control data provide useful information on the normal pattern and range of tumourtypes and incidences for a particular strain/species, which may not be reflected by the tumourfindings in the concurrent controls in any individual study. This can be particularly relevantfor animal strains which have a propensity to develop a particular type of tumourspontaneously with variable and potentially high incidence. In such a case the tumourincidence in the treated group may be significantly above the concurrent control but could stillbe within the historical incidence range for that tumour type in that species and therefore maynot be providing reliable evidence of treatment related carcinogenicity.Some examples of animal tissues with a high spontaneous tumour incidence are:‒ Adrenal pheochromocytoma in male F344 rats (NTP, 2007a), Sprague-Dawley rats (NTP,2005; RIVM, 2001; Ozaki et al., 2002);‒ Pituitary adenomas in F344 rats (NTP, 2007a), Sprague-Dawley rats (NTP 2005; RIVM2005);‒ Mammary gland tumours (adenomas and carcinomas) in female Sprague-Dawley rats(NTP, 2005);‒ Mononuclear cell leukaemia in F344 rats (NTP, 2007a; RIVM, 2005);‒ Liver tumours in B6C3F1 mice (NTP, 2007b; Haseman et al. 1998; Battershill, J.M. andFielder, R.J., 1998);‒ Leydig cell adenomas in male F344 rats (Cook et al., 1999; Mati et al., 2002; RIVM, 2004;EU Specialised Experts Report, 2004).304

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Historical control data can also be useful to judge the biological significance of marginalincreases in uncommon tumours. If there is a small increase in a particular tumour type whichhistorical data shows to be very uncommon and unlikely to have occurred by chance then thismay support a conclusion of carcinogenicity without the requirement for a statisticallysignificant increase.Use of historical control data should be on a case by case basis with due consideration of theappropriateness and relevance of the historical control data for the study under evaluation. Ina general sense, the historical control data set should be matched as closely as possible to thestudy being evaluated. The historical data must be from the same animal strain/species, andideally, be from the same laboratory to minimise any potential confounding due to variationsin laboratory conditions, study conditions, animal suppliers, husbandry etc. It is also knownthat tumour incidences in control animals can change over time, due to factors such as geneticdrift, changes in diagnostic criteria for pathological changes/tumour types, and husbandryfactors (including the standard diet used), so the historical data should be contemporary to thestudy being evaluated (e.g. within a period of up to around 5 years of the study). Historicaldata older than this should be used with caution and acknowledgement of its lower relevanceand reliability. (RIVM, 2005; Fung et al, 1996; Greim et al, 2003).Even when a particular tumour type may be discounted, expert judgment must be used inassessing the total tumour profile in any animal. However, appearance of only spontaneoustumours, especially if they appear only at high dose levels, may be sufficient to downgrade aclassification from Category 1B to Category 2, or even no classification. Where the onlyavailable tumour data are liver tumours in certain sensitive strains of mice, without any othersupplementary evidence, the substance may not be classified in any of the categories,(Battershill and Fielder, 1998). Expert judgment is required to evaluate the relevance of theresults.(b) Multi-site responsesIn general, chemicals are evaluated for carcinogenic potential in two-year bioassaysconducted in mice and rats. The chemicals produce a spectrum of responses ranging from noeffects in either species to induction of malignant neoplasms in multiple tissues in bothspecies. Between these two extremes, there are variable responses in tissues, sexes andspecies, which demonstrate that there are important differences among the carcinogens, aswell as between the species in which they are tested. The tumour profile observed with asubstance should be taken into account when considering the most appropriate classification.Evidence shows that substances which cause tumours in either multiple sites and/or multiplespecies tend to be more potent carcinogens than those causing tumours at only one site in onespecies (Dybing et al., 1997). This is often true for substances which are mutagenic. Also,where human carcinogens have been tested in two or more species, the majority have causedcancer in several species (Tennant, 1993). Thus, if a substance causes tumours at multiplesites and/or in more than one species then this usually provides strong evidence ofcarcinogenicity. Typically such a tumour profile would lead to a classification in category 1B.(c) Progression of lesions to malignancyIn general, if a substance involves a treatment related increase in tumours then it will meet thecriteria for classification as a carcinogen.If the substance has been shown to cause malignant tumours this will usually constitutesufficient evidence of carcinogenicity supporting Category 1B (Annex I, Section 3.6.2.2.3)The induction of only benign tumours usually provides a lower strength of evidence forcarcinogenicity than the induction of malignant tumours and will usually support Category 2305

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(Annex I, Section 3.6.2.2.3). However, benign tumours may also be of significant concern andthe strength of evidence for carcinogenicity that they provide should be considered usingexpert judgement. For instance, some benign tumours may have the potential to progress tomalignant tumours and therefore any indication that the observed tumours have the potentialto progress to malignancy may increase the level of concern. Also, some benign tumours, forexample brain tumours, may be of concern in themselves.(d) Reduced tumour latencyThe latency of tumour development i.e. how quickly a substance induces tumours, oftenreflects the potency of a carcinogen. This is particularly true for mutagenic substances whichoften induce tumours with relatively short latency and usually more rapidly than nongenotoxicagents. Tumour latency is not generally investigated in detail in standardcarcinogenicity studies, although some information may be provided if the study used serialsacrifices.The latency of tumour formation does not materially affect the classification and hazardcategory. Any substance causing cancer will attract classification regardless of the latency fortumour development. This also includes tumour responses at late treatment/life periods ifsubstance-related. However unusual tumour types or tumours occurring with reduced latencymay add to the weight of evidence for the carcinogenic potential of a substance, even if thetumours are not statistically significant.(e) Whether responses are in single or both sexesIn general, in standard carcinogenicity studies both male and female animals are tested. Theremay be cases where tumours are only observed in one sex.Tumours in one sex only may arise for two broad reasons. The tumours may occur in agender-specific tissue, for instance the uterus or testes (sex-specific tissue), or in a non sexspecifictissue, in one sex only. Tumours may also be induced by a mechanism that is gender(or sex) -specific, for instance a hormonally-mediated mechanism or one involving gender (orsex) -specific differences in toxicokinetics. As with all cases the strength of evidence ofcarcinogenicity should be assessed based on the totality of the information available using aweight of evidence type approach. A default position is that such tumours are still evidence ofcarcinogenicity and should be evaluated in light of the total tumorigenic response to thesubstance observed at other sites (multi-site responses or incidence above background) indetermining the carcinogenic potential and the classification category.If tumours are seen only in one sex of an animal species, the mode of action should becarefully evaluated to see if the response is consistent with the postulated mode of action.Effects seen only in one sex in a test species may be less convincing than effects seen in bothsexes, unless there is a clear patho-physiological difference consistent with the mode of actionto explain the single sex response. However, there is no requirement for a mechanisticunderstanding of tumour induction in order to use these findings to support classification. Ifthere is clear evidence for induction of either a gender (or a sex)-specific tumour thenclassification in Cat 1B may be appropriate. However, it has to be taken into account thataccording to the criteria additional data are required to provide sufficient evidence for animalcarcinogenicity (1B).(f) Whether responses are in a single species or several speciesThe criteria indicate that carcinogenicity in a single animal study (both sexes, ideally in aGLP study) could be sufficient evidence and could therefore lead to a Category 1Bclassification in the absence of any other data. This represents a change compared to the306

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012previous EU-system where such a study would rarely lead to the equivalent of a Category 1Bclassification. For classification as a Category 2 carcinogen under DSD either positive resultsin two animal species should be available or clear positive evidence in one species, togetherwith supporting evidence such as genotoxicity data, metabolic or biochemical studies,induction of benign tumours, structural relationship with other known carcinogens, or datafrom epidemiological studies suggesting an association.However, as defined under 'sufficient' evidence (Annex I, section 3.6.2.2.3 (b)), a single studyin one species and sex might be considered to provide sufficient evidence of carcinogenicitywhen malignant neoplasms occur to an unusual degree with regard to incidence, site, type oftumour or age at onset, or when there are strong findings of tumours at multiple sites.Moreover a single study in one species and sex in combination with positive in-vivomutagenicity data would be considered to provide sufficient evidence of carcinogenicity.Positive responses in several species add to the weight of evidence, that a chemical is acarcinogen.(g) Structural similarity or not to a chemical(s) for which there is good evidence ofcarcinogenicitySee Section 3.6.2.3.4.(h) Routes of exposure;Annex I: 3.6.2.2.8. The classification shall take into consideration whether or not the substance isabsorbed by a given route(s); or whether there are only local tumours at the site of administrationfor the tested route(s), and adequate testing by other major route(s) show lack of carcinogenicity.The classification for carcinogenicity generally does not specify specific routes of exposure. Ifa chemical has been shown to cause tumours by any route of administration then it mayrequire classification, unless there is a robust justification for dismissing the findings from aparticular route. However, under the previous EU system (Annex VI to DSD), classificationspecifically via inhalation was accepted by application of the risk phrase R49; May causecancer by inhalation and a specific hazard statement has been established in CLP, H350i; Maycause cancer by inhalation (CLP, Annex VII, Table 1.1).Most standard carcinogenicity studies use physiological routes of exposure for humans,namely inhalation, oral or dermal exposure. The findings from such routes are usuallyconsidered directly relevant for humans. Studies using these routes will generally takeprecedence over similar studies using other routes of exposure.Sometimes other non-physiological routes are used, such as intra-muscular, sub-cutaneous,intra-peritoneal and intra-tracheal injections or instillations. Findings from studies using theseroutes may provide useful information but should be considered with caution. Usually dosingvia these routes provides a high bolus dose which gives different toxicokinetics to normalroutes and can lead to atypical indication of carcinogenicity. For instance, the high localconcentration can lead to local tumours at the site of injection. These would not normally beconsidered reliable indications of carcinogenicity as they most likely arose from theabnormally high local concentration of the test substance and would lead to a lower categoryclassification or no classification.Where findings are available from studies using standard routes and non-physiological routes,the former will generally take precedence. Usually studies using non-standard routes providesupporting evidence only.The hazard statement allows for identifying the route of exposure “if it is conclusively proventhat no other routes of exposure cause the hazard” (Annex I, section 3.6.4.1). In this case the307

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012hazard statement may be modified accordingly. Genotoxic carcinogens are generallysuspected to be carcinogenic by any route.(i) Comparison of absorption, distribution, metabolism and excretion between test animalsand humans;Annex I: 3.6.2.2.9. It is important that whatever is known of the physico-chemical, toxicokineticand toxicodynamic properties of the substances, as well as any available relevant information onchemical analogues, i.e. structure activity relationship, is taken into consideration when undertakingclassification.Consideration of absorption, distribution, metabolism and excretion (toxicokinetics) of thesubstance in the test animal species and in humans is one important consideration, includingwhere a substance is metabolised to an active carcinogenic metabolite. Toxicokineticbehaviour is normally assumed to be similar in animals and humans, at least from aqualitative perspective. On the other hand, certain tumour types in animals may be associatedwith toxicokinetics or toxicodynamics that are unique to the animal species tested and maynot be predictive of carcinogenicity in humans. Where significant qualitative and quantitativedifferences in toxicokinetics exist between animals and humans this can impact on therelevance of the animal findings for humans and in certain instances may influence thecategory of classification. Where a carcinogenic metabolite identified in animals isdemonstrated not to be produced in humans, no classification may be warranted where it canbe shown that this is the only mechanism of action for carcinogenicity.The use of physiologically-based pharmacokinetic (PB/PK) modelling requires morevalidation and while it may not lead directly to a modification of classification, howeverexpert judgement in conjunction with PB/PK modelling may help to modify the concern forhumans.(j) The possibility of a confounding effect of excessive toxicity at test dosesIn lifetime bioassays compounds are routinely tested using at least three dose levels to enablehazard identification and hazard characterisation as part of risk assessment. Of these doses,the highest dose needs to induce minimal toxicity, such as characterised by an approximately10% reduction in body weight gain (maximal tolerated dose, MTD dose). The MTD is thehighest dose of the test agent during the bioassay that can be predicted not to alter theanimal’s normal longevity from effects other than carcinogenicity. Data obtained from a subchronicor other repeated dose toxicity study are used as the basis for determining the MTD.Excessive toxicity, for instance toxicity at doses exceeding the MTD, can affect thecarcinogenic responses in bioassays. Such toxicity can cause effects such as cell death(necrosis) with associated regenerative hyperplasia, which can lead to tumour development asa secondary consequence unrelated to the intrinsic potential of the substance itself to causetumours at lower less toxic doses.Tumours occurring only at excessive doses associated with severe toxicity generally have amore doubtful potential for carcinogenicity in humans. In addition, tumours occurring only atsites of contact and/or only at excessive doses need to be carefully evaluated for humanrelevance for carcinogenic hazard. For example, as indicated in this Section (a) 'Tumour typeand background incidence', forestomach tumours, following administration by gavage of anirritating or corrosive, non-mutagenic chemical, may be of questionable relevance, both dueto the lack of a corresponding tissue in humans, but importantly, due to the high dose directeffect on the tissue. However, such determinations must be evaluated carefully in justifyingthe carcinogenic potential for humans; any occurrence of other tumours at distant sites mustalso be considered.308

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The proceedings of a WHO/IPCS workshop on the Harmonization of Risk Assessment forCarcinogenicity and Mutagenicity (Germ cells) - A Scoping Meeting (IPCS, 1995; Ashby etal, 1996), points to a number of scientific questions arising for classification of chemicals,e.g. mouse liver tumours, peroxisome proliferation, receptor-mediated reactions, chemicalswhich are carcinogenic only at toxic doses and which do not demonstrate mutagenicity.If a test compound is only found to be carcinogenic at the highest dose(s) used in a lifetimebioassay, and the characteristics associated with doses exceeding the MTD as outlined aboveare present, this could be an indication of a confounding effect of excessive toxicity. This maysupport a classification of the test compound in Category 2 or no classification.(k) Mode of action and its relevance for humans, such as mutagenicity, cytotoxicity withgrowth stimulation, mitogenesis, immunosuppressionCarcinogenic chemicals have conventionally been divided into two categories according tothe presumed mode of action; genotoxic or non-genotoxic. Genotoxic modes of action involvegenetic alterations caused by the chemical interacting directly with DNA to possibly result ina change in the primary sequence of DNA after cell division. A chemical can also causegenetic alterations indirectly following interaction with other cellular processes (e.g.,secondary to the induction of oxidative stress). Non-genotoxic modes of action includeepigenetic changes, i.e., effects that do not involve alterations in DNA but that may influencegene expression, altered cell-cell communication, or other factors involved in the carcinogenicprocess. For example, chronic cytotoxicity with subsequent regenerative cell proliferation isconsidered a mode of action by which tumour development can be enhanced: the induction ofurinary bladder tumours in rats may, in certain cases, be due to persistentirritation/inflammation, tissue erosion and regenerative hyperplasia of the urotheliumfollowing the formation of bladder stones. Other modes of non-genotoxic action can involvespecific receptors (e.g., peroxisome proliferator-activated receptor-alpha (PPARα) which isassociated with liver tumours in rodents; or tumours induced by various hormonalmechanisms). More detail is given in IR/CIS Section R7.7.8.Some modes of action of tumour formation are considered to be not relevant to humans.Where such a mechanism is identified then classification may not be appropriate Only if amode of action of tumour development is conclusively determined not to be operative inhumans may the carcinogenic evidence for that tumour be discounted. However, a weight ofevidence evaluation for a substance calls for any other tumorigenic activity to be evaluated aswell. In addition, the existence of a secondary mechanism of action with the implication of apractical threshold above a certain dose level (e.g., hormonal effects on target organs or onmechanisms of physiological regulation, chronic stimulation of cell proliferation) may lead toa downgrading of a Category 1 to Category 2 classification.The various international documents on carcinogen assessment all note that mode of action inand of itself, or consideration of comparative metabolism, should be evaluated on a case-bycasebasis and are part of an analytic evaluative approach. One must look closely at any modeof action in animal experiments taking into consideration comparativetoxicokinetics/toxicodynamics between the animal test species and humans to determine therelevance of the results to humans. This may lead to the possibility of discounting veryspecific effects of certain types of chemicals. Life stage-dependent effects on cellulardifferentiation may also lead to qualitative differences between animals and humans.To establish a mode of action will usually require specific investigative studies over andabove the standard carcinogenicity study. All available data must be considered carefully tojudge if it can be concluded with confidence that the tumours are being induced through thatspecific mechanism. The IPCS Framework for Analyzing the Relevance of a Cancer Mode of309

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Action for Humans (2007) can be a useful way to construct and present a robust andtransparent assessment of such data.Some mechanisms of tumour formation considered not relevant for humans:‒ Kidney tumours in male rats associated with substances causing α2μ-globulin nephropathy(IARC, 1999)‒ Pheochromocytomas in male rats exposed to particulates through inhalation secondary tohypoxemia (Ozaki et al, 2002)‒ Leydig cell adenomas induced by dopamine antagonists or gonadotropin-releasinghormone (GnRH) (EU Specialised Experts, 2004; RIVM, 2004)‒ Urinary bladder tumours due to crystals in the bladder (IARC, 1999)‒ Forestomach tumours in rodents following administration by gavage of irritating orcorrosive, non-genotoxic substances (RIVM, 2003; IARC 2003)‒ Certain thyroid tumours in rodents mediated by UDP glucuronyltransferase (UGT)induction (IARC, 1999; EU Specialised Experts, 1999)‒ Liver tumours in rodents conclusively linked to peroxisome proliferation (IARC, 1994)3.6.2.3.3 Consideration of mutagenicityAnnex I: 3.6.2.2.6. Mutagenicity: It is recognised that genetic events are central in the overallprocess of cancer development. Therefore evidence of mutagenic activity in vivo may indicate thata substance has a potential for carcinogenic effects.As indicated in Section 3.6.2.1 and above, carcinogenic chemicals have conventionally beendivided according to the presumed mode of action; genotoxic or non-genotoxic. Evidence ofgenotoxic activity is gained from studies on mutagenic activity.It should be noted that in general if a substance is mutagenic then it will be considered to bepotentially carcinogenic in humans however mutagenicity data alone are insufficientinformation to justify a carcinogen classification. In some cases where only in vitro and invivo mutagenicity are present without carcinogenicity data, a Category 2 classification can beconsidered when all factors have been considered such as type and quality of the mutagenicitydata, structure activity relationships etc. A single positive carcinogenicity study in one speciesand sex in combination with positive in-vivo mutagenicity data would be considered toprovide sufficient evidence of carcinogenicity.Lack of genotoxicity is an indicator that other mechanisms are in operation as indicated inSection 3.6.2.3.2(k) above. Thus careful analysis based on all available information isrequired to identify the mechanism and derive a classification category taking into account thefactors leading to the tumours observed, in the animals.3.6.2.3.4 Non testing dataAnnex I: 3.6.2.2.7. A substance that has not been tested for carcinogenicity may in certaininstances be classified in Category 1A, Category 1B or Category 2 based on tumour data from astructural analogue together with substantial support from consideration of other important factorssuch as formation of common significant metabolites, e.g. for benzidine congener dyes.A chemical that has not been tested for carcinogenicity may in certain instances be classifiedas a carcinogen based on tumour data from a structurally similar chemical with which it ispredicted to have similar carcinogenic activity. Such an approach must always be based on arobust and transparent argument to support this supposition. There may also be evidencedemonstrating similarity in terms of other important factors such as toxicokinetics or310

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012mutagenic activity etc. (OECD 2004, 2005, 2007; IR/CSA, Section R.6, QSARs and groupingof chemicals).In the absence of carcinogenicity data, read-across can be used to support a classification forcarcinogenicity when the chemical in question is similar to a known or suspected carcinogen(Category 1A, 1B or 2). The similarity between chemicals is considered in terms of structuralfeatures, physico-chemical properties and overall toxicological profile.In general the chemicals will share a common structural element or functional group (i.e., atoxiphore) that has been shown to be integral to the underlying mechanism of carcinogenicityfor chemicals with this toxiphore in well conducted studies. These toxiphores can beidentified through expert judgement or through automated systems such as (Q)SARs. Theread-across should also consider the physico-chemical properties of the chemical and datafrom other toxicity studies to judge the similarity between the chemicals in terms ofbioavailability by relevant routes of exposure and toxicokinetics. The toxicity profile fromother studies should also be compared (e.g., acute and repeated-dose toxicity andmutagenicity) and should share similarities in nature and severity. Data from shorter termtoxicity studies may be useful, particularly for non-genotoxic carcinogens, to indicate that thechemicals cause the same underlying pathological changes (e.g., hyperplasia), and act via acommon mode of action. Any predictions made on the basis of read-across should take intoaccount the totality of data on the chemicals in question, including the physico-chemicalproperties, toxicological profile, toxicokinetics, structural analogy and the performance of any(Q)SAR models used, in a weight of evidence approach driven by expert judgement. The finaldecision must be clear, scientifically defensible and transparent.The specific category depends on the category of the known carcinogen and the degree ofconfidence in the robustness of the read-across prediction. The category will not be higherthan the chemical used to read-across from, but normally may be the same. However a lowercategory may be applied if the read-across highlights a possible carcinogenic hazard, and thussupports a classification, but there is uncertainty as to the robustness of the read-acrossprediction or there is evidence, for instance from mechanistic or other studies, that thechemical may be of lower concern for carcinogenicity.If a chemical is similar to a substance known to be carcinogenic and shares the toxiphore thatis considered to be causally related to carcinogenicity, then it is unlikely that there will besufficient confidence in a prediction of no hazard (for instance based on arguments relating todifferences in physico-chemical or steric properties), to justify no classification in the absenceof supporting negative experimental data. However, the bioavailability of the toxiphore willneed evaluation (IR/CSA R.6).3.6.2.4 Decision on classificationAs mentioned throughout, classification as a carcinogen is based on consideration of thestrength of evidence with additional considerations (weight of evidence) being taken intoaccount as appropriate. It is recognised that, in most cases, expert judgment is necessary todetermine the classification category.3.6.2.5 Setting of specific concentration limitsExperimental studies have revealed large variations in the doses of various carcinogenicsubstances needed to induce tumours in animals. Thus, the amounts of chemical carcinogensrequired to induce tumours vary with a factor of up to 10 8 -10 9 for different compounds. It isreasonable to assume that there is similar variation in the potency of substances carcinogenicto humans (Sanner and Dybing, 2005).311

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The carcinogenic properties of mixtures are normally not tested. The classification andlabelling of mixtures for carcinogenicity is therefore based on the classification of theingredients and the percentage of each ingredient in the mixture. As indicated in Section3.6.3, the criteria contain default percentages for classification of mixtures with carcinogenicproperties but CLP, Article 10.1 allows the use of specific concentration limits (SCL) basedon the potency of the carcinogen(s). The EU has adopted the T25 concept for carcinogenicity(Dybing et al., 1997) with additional considerations as a measure for intrinsic potency and aguidance document (EC, 1999) to assist in establishing SCLs for carcinogens. By using thisapproach the SCL may occasionally be reduced or raised from the default genericconcentration limits.312

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.6.2.6 Decision logic for classification of substancesThe decision logic which follows is taken from the GHS Guidance. It is stronglyrecommended that the person responsible for classification, study the criteria for classificationbefore and during use of the decision logic.Does the subststance have carcinogenicity data?NOClassificationnot possibleYESAccording to the criteria, is the substance:(a) Known to have carcinogenic potential for humans, or(b) Presumed to have carcinogenic potential for humans?Application of the criteria needs expert judgement in a strengthand weight of evidence approach.YESCategory 1DangerNOCategory 2According to the criteria (see 3.6.2), is the substance asuspected human carcinogen?Application of the criteria needs expert judgement in astrength and weight of evidence approach.YESWarningNONot classified313

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.6.3 Classification of mixtures for carcinogenicity3.6.3.1 Classification criteria for mixturesClassification of mixtures will be based on the available test data for the individualingredients of the mixture, using cut-off values/concentration limits for those ingredients andtaking into account potency consideration. The classification may a case-by-case basis bebased on the available test data for the mixture as a whole (see Section 3.6.3.1.2) or based onbridging principles (see Section 3.6.3.1.3).3.6.3.1.1 When data are available for all ingredients or only for some ingredientsAnnex I: 3.6.3.1.1. The mixture will be classified as a carcinogen when at least one ingredient hasbeen classified as a Category 1A, Category 1B or Category 2 carcinogen and is present at or abovethe appropriate generic concentration limit as shown in Table 3.6.2 below for Category 1A, Category1B and Category 2 respectively.Table 3.6.2Generic concentration limits of ingredients of a mixture classified as carcinogen that triggerclassification of the mixtureIngredient classified as:Generic concentration limits triggering classification of a mixture as:Category 1AcarcinogenCategory 1BcarcinogenCategory 2carcinogenCategory 1A carcinogen 0,1 % — —Category 1B carcinogen — 0,1 % —Category 2 carcinogen — — 1,0 % [Note 1]NoteThe concentration limits in the table above apply to solids and liquids (w/w units) as well as gases(v/v units).Note 1If a Category 2 carcinogen is present in the mixture as an ingredient at a concentration≥ 0,1% a SDS shall be available for the mixture upon request.In case a SCL has been established for one or more ingredients these SCLs have precedenceover the respective GCLs. See Section 3.6.2.5for the setting of SCLs for substances.3.6.3.1.2 When data are available for the complete mixtureAnnex I: 3.6.3.2.1. Classification of mixtures will be based on the available test data for theindividual ingredients of the mixture using concentration limits for the ingredients classified ascarcinogens. On a case-by-case basis, test data on mixtures may be used for classification whendemonstrating effects that have not been established from the evaluation based on the individualingredients. In such cases, the test results for the mixture as a whole must be shown to beconclusive taking into account dose and other factors such as duration, observations, sensitivity andstatistical analysis of carcinogenicity test systems. Adequate documentation supporting theclassification shall be retained and made available for review upon request.314

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.6.3.1.3 When data are not available for the complete mixture: bridging principlesAnnex I: 3.6.3.3.1. Where the mixture itself has not been tested to determine its carcinogenichazard, but there are sufficient data on the individual ingredients and similar tested mixtures(subject to the provisions of paragraph 3.6.3.2.1) to adequately characterise the hazards of themixture, these data shall be used in accordance with the applicable bridging rules set out in section1.1.3.Note that not all bridging principles in Annex I, section 1.1.3. are applicable when classifyingfor carcinogenicity.315

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.6.3.2 Decision logic for classification of mixturesThe decision logic which is based on the GHS Guidance is revised to meet CLP requirements.It is strongly recommended that the person responsible for classification, study the criteria forclassification before and during use of the decision logic.Classification based on individual ingredients of the mixtureCategory 1Does the mixture contain one or more ingredientsclassified as a Category 1 carcinogen at 0.1 %, orabove a SCL set for the ingredient(s)?YESNODangerDoes the mixture contain one or more ingredientsclassified as a Category 2 carcinogen at 1.0 %, orabove a SCL set for the ingredient(s)?YESCategory 2NONot classifiedWarningModified classification on a case-by-case basisTest data on mixtures may be used for classification when demonstrating effects that have not beenestablished from the evaluation based on the individual ingredients (CLP Section 3.6.3.1.1, see alsoCLP Article 6(3)).Are test data availablefor the mixturedemonstrating acarcinogenic effect notidentified from the dataon individualsubstances?YESAre the test results on themixture conclusive taking intoaccount dose and other factorssuch as duration, observationsand analysis (e.g. statisticalanalysis, test sensitivity) ofcarcinogenicity test systems?YESClassify inappropriate categoryNONODanger orWarningCan bridging principles beapplied?YESorNo classificationNOSee above: Classification based onindividual ingredients of the mixture.316

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.6.4 Hazard communication in form of labelling for carcinogenicity3.6.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I; 3.6.4.1 Label elements shall be used in accordance with Table 3.6.3, for substances ormixtures meeting the criteria for classification in this hazard class.Table3.6.3Label elements for carcinogenicityClassification Category 1A or Category 1B Category 2GHS PictogramsSignal Word Danger WarningHazard StatementPrecautionary StatementPreventionPrecautionary StatementResponsePrecautionary StatementStoragePrecautionary StatementDisposalH350: May cause cancer (stateroute of exposure if it isconclusively proven that noother routes of exposure causethe hazard)P201P202P281P308 + P313P405P501H351: Suspected of causingcancer (state route of exposureif it is conclusively proven thatno other routes of exposurecause the hazard)P201P202P281P308 + P313The wording of the Precautionary Statements is found in CLP Annex IV, Part 2.Where there is conclusive proof that cancer is caused only by certain route(s), then this routemay be stated in the hazard statement. In case of Category 1 carcinogens where there isconclusive proof that cancer is caused only by inhalation, the hazard phrase “H350i: Maycause cancer by inhalation” applies (CLP Annex VII Table 1.1).3.6.4.2 Additional labelling provisionsP405P501There are no additional labelling provisions for carcinogenic substances and mixtures in CLP,however there are provisions laid out in Annex XVII to REACH. The packaging ofsubstances with harmonised classification as carcinogenic Category 1A or Category 1B, ormixtures containing such substances, "must be marked visibly, legibly and indelibly asfollows: ‘Restricted to professional users’." (REACH, Annex XVII, point 28).317

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.6.5 Re-classification of substances and mixtures classified for carcinogenicityaccording to DSD and DPD3.6.5.1 Is direct “translation” of classification and labellingpossible?A direct translation as indicated in the translation table in Annex VII to CLP is generallypossible. Translation from classification according to DSD and DPD to the classificationaccording to CLP is as follows:Carc. Cat. 1 is translated into Carc. 1A;Carc. Cat. 2 is translated into Carc. 1B, andCarc. Cat. 3 is translated into Carc. 2, respectively.3.6.5.2 Some additional considerations for re-classificationThere are only few situations where the direct translation may lead to different results,however, these are likely to be very rare.The first difference in applying the CLP criteria is that sufficient evidence (Carc. 1B) forcarcinogenicity in animals can also be derived from two or more independent studies in onespecies carried out at different times or in different laboratories or under different protocols.The second difference applying the CLP criteria is that sufficient evidence (Carc. 1B) forcarcinogenicity in animals can be derived from an increased incidence of tumours in bothsexes of a single species in a well-conducted study, ideally conducted under GLP. The criteriaaccording to DSD allowed classification in Carc. Cat. 2 (analogous to CLP Carc. 1B) wherethere were positive results in two animal species or clear positive evidence in one species,together with supporting evidence such as genotoxicity data, metabolic or biochemicalstudies, induction of benign tumours, structural relationship with other known carcinogens, ordata from epidemiological studies suggesting an association.Another difference can be derived from the IARC classification as ‘possibly carcinogenic tohumans (IARC 2B)’. This category is used for substances for which there is less than sufficientevidence of carcinogenicity in experimental animals. According to IARC, classification as‘possibly carcinogenic to humans’ may be derived from solely strong evidence from mechanisticand other relevant data. This means that no in vivo carcinogenicity nor (Q)SAR data need to beavailable to arrive at classification for limited evidence of carcinogenicity.3.6.6 Examples of classification for carcinogenicityClassification for carcinogenicity involves the consideration of many different factors, asoutlined above, and is a complex task which needs expert judgement. Therefore no examplesof classification for carcinogenicity are included in this guidance document.3.6.7 ReferencesAshby J, Waters M.D., Preston J., Adlerd I.-D., Douglas G.R., Fielder R., Shelby M.D.,Anderson D., Sofuni T., Gopalan H.N.B., Becking G. and Sonich-Mullin C.: IPCSharmonization of methods for the prediction and quantification of humancarcinogenic/mutagenic hazard, and for indicating the probable mechanism of action ofcarcinogens. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis352 (1-2), 153-157, 1996.318

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Battershill J.M. and Fielder R.J.: Mouse-specific carcinogens: an assessment of hazard andsignificance for validation of short-term carcinogenicity bioassays in transgenic mice. Hum.Exp. Toxicol. 17(4), 193-205, 1998.Boobis A.R., Cohen S.M., Dellarco V., McGregor D., Meek M.E., Vickers C., Willcocks D.,Farland W.: IPCS framework for analyzing the relevance of a cancer mode of action forhumans. Crit. Rev. Toxicol. 36(10), 781-92, 2006.Cohen S.M.; Meek M.E., Klaunig J.E.; Patton D.E., Fenner-Crisp P.A.: The HumanRelevance of Information on Carcinogenic Modes of Action: Overview. Critical Reviews inToxicology 33(6), 581-589, 2003.Cohen S.M.: Human carcinogenic risk evaluation: an alternative approach to the two-yearrodent bioassay. Toxicol. Sci. 80, 225-229, 2004.Cook J.C., Klinefelter G.R., Hardisty J.F., Sharpe R.M., Foster P.M.: Rodent Leydig celltumorigenesis: a review of the physiology, pathology, mechanisms, and relevance to humans.Crit. Rev. Toxicol. 29 (2), 169-261, 1999.Dybing E., Sanner T., Roelfzema H., Kroese D, Tennant R.W.: T25: A simplifiedcarcinogenic potency index. Description of the system and study correlation betweencarcinogenic potency and species/site specificity and mutagenicity. Pharmacol. Toxicol. 80,272-279, 1997.EC: Guidelines for setting specific concentration limits for carcinogens in Annex I ofdirective 67/548/EEC. Inclusion of potency considerations. Commission working group onthe classification and labelling of dangerous substances. Office for the Official Publications ofthe European Communities, Luxembourg, ISBN 92-828-7443-5, 1999.EU Commission Group of Specialised Experts in the fields of carcinogenicity, mutagenicityand reprotoxicity: Non genotoxic thyroid carcinogens in the rodent bioassay, ECBI/49/99Add. 1 Rev. 2 excerpt of agenda point 3.1, 1999.http://ecb.jrc.ec.europa.eu/documents/Classification-Labelling/GUIDANCE_DOCUMENTS/4999a1r2_spec0999_excerpt.pdfEU Commission Group of Specialised Experts in the fields of carcinogenicity, mutagenicityand reprotoxicity: Leydig tumours 2004, ECBI/08/04 Rev. 2, 2004http://ecb.jrc.ec.europa.eu/documents/Classification-Labelling/GUIDANCE_DOCUMENTS/0804r2_sr_SE_01_2004_LCT_excerpt.pdfFung K.Y., Krewski D., and Smythe R.T: A comparison of tests for trend with historicalcontrols in carcinogen bioassay. Canadian J. of Statist. 24, 431-454, 1996.Greim H., Gelbke H.-P., Reuter U., Thielmann H.W. and Edler L.: Evaluation of historicalcontrol data in carcinogenicity studies. Human Exp. Toxicol. 22:541-549, 2003.Haseman J.K, Hailey J.R., Morris R.W.: Spontaneous neoplasm incidences in Fischer 344 ratsand B6C3F-1 mice in two-year carcinogenicity studies: A national toxicology programupdate. Toxicol Pathol. 26(3) 428-441, 1998.IARC (1994) Peroxisome. Proliferation and its role in carcinogenesis views and expertopinions of an IARC Working Group, Lyon, 7-11 December 1994, Technical Publication No.24, Lyon, France.IARC (1999) Scientific Publications No. 147 Species Differences in Thyroid, Kidney andUrinary Bladder Carcinogenesis.319

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012IARC (2003). Predictive value of rodent forestomach and gastric neuroendocrine tumours inevaluating carcinogenic risks to humans. Views and expert opinions of an IARC WorkingGroup, Lyon, 29 November – 1 December 1999. IARC Technical Publication No. 39, Lyon,France, p.1-73.IARC (2006) Monographs on the Evaluation of Carcinogenic Risks to Humans, PreambleSection B, Lyon, France.http://monographs.iarc.fr/ENG/Preamble/CurrentPreamble.pdfIPCS (1995). Harmonization of the Assessment of Risk for Carcinogenicity and Mutagenicity(Germ Cells), Carshalton, UK 1995. IPCS/95.x.IPCS (2007) Boobis A.R., Cohen S.M., Dellarco V., McGregor D., Meek M.E., Vickers C.,Willcocks D., Farland W.: IPCS Framework for Analyzing the Relevance of a Cancer Modeof Action for Humans in IPCS Harmonization Project Document No. 4, Part 1, IPCSframework for analysing the relevance of a cancer mode of action for humans and casestudies.http://www.who.int/ipcs/methods/harmonization/areas/cancer_mode.pdfMati W., Lam G., Dahl C., Thorup Andersen J., Balslev E.: Leydig cell tumour – a raretesticular tumour. Int. Urol. Nephrol. 33(1), 103-106, 2002.McGregor D., Binaglia M., Boobis A., Botham P., Hennes E.C., Hoffstadt L., Hubbard S.,Petry T., Riley A., Schwartz D.: 2008 Guidance for the Classification of Carcinogens underthe Globally Harmonised System of Classification and Labelling of Chemicals (GHS).Submitted for publication, 2009.Meek M.E., Bucher J.R., Cohen S.M., Dellarco V., Hill R.N., Lehman-McKeeman L.D.,Longfellow D.G., Pastoor T., Seed J., Patton D.E.: A Framework for Human RelevanceAnalysis of Information on Carcinogenic Modes of Action. Critical Reviews in Toxicology33(6), 591-653, 2003.NTP (2005) Historical Control Report, Gavage, Corn oil, Sprague-Dawley, Females, 2005.http://ntp.niehs.nih.gov/files/HIST_2005SD_TUMOR_RATS1.pdfNTP (2007a) Historical Controls Report All Routes And Vehicles. RATS, October 2007.http://ntp.niehs.nih.gov/files/HistCont-2007-10-17-Rats_All_Routes.pdfNTP (2007b) Historical Controls Report All Routes And Vehicles. MICE, October 2007.http://ntp.niehs.nih.gov/files/HistCont-2007-10-17-Mice_All_Routes.pdfOECD (2004). OECD Principles for the Validation, for Regulatory Purposes, of(Quantitative) Structure-Activity Relationship Models.http://www.oecd.org/document/23/0,2340,en_2649_34379_33957015_1_1_1_1,00.htmlOECD (2005). 3.2 Guidance on the Development and Use of Chemical Categories in HPVChemicals Programmehttp://www.oecd.org/document/7/0,2340,en_2649_34379_1947463_1_1_1_1,00.htmlOECD (2007). Report on the Regulatory Uses and Applications in OECD Member Countriesof (Q)SAR Models in the Assessment of New and Existing Chemicals.ENV/JM/MONO(2006)25http://www.oecd.org/dataoecd/55/22/38131728.pdfOzaki, K., Haseman, J.K., Hailey, J.R., Maronpot, R.R., Nyska, A.: Association of adrenalpheochromocytoma and lung pathology in inhalation studies with particulate compounds inthe male F344 rat – the National Toxicology Program Experience, 2002.320

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012RIVM (2001) report 601516007: Factsheets for the (eco)toxicological risk assessmentstrategy of the National Institute of Public Health and the Environment (RIVM). Editors: R.Luttik and M.T.M. van Raaij.RIVM (2003) report 601516012: Factsheets for the (eco)toxicological risk assessmentstrategy of the National Institute for Public Health and the Environment - Part IV. Editors:C.E. Smit and M.T.M. van Raaij.RIVM (2004) report 601516012: Factsheets for the (eco)toxicological risk assessmentstrategy of the National Institute for Public Health and the Environment - Part IV. Editors:C.E. Smit and M.T.M. van Raaij.RIVM (2005) report 601516013: Factsheets for the (eco)toxicological risk assessmentstrategy of the National Institute for Public Health and the Environment Part V. J.W.A.Scheepmaker, C.E. Smit, M.T.M. van Raaij.Sanner T. and E. Dybing: Comparison of Carcinogen Hazard Characterisation Based onAnimal Studies and Epidemiology Basic & Clinical Pharmacology & Toxicology 96, 66–70,2005.Sonich-Mullin C., Fielder R., Wiltse J., Baetcke K., Dempsey J., Fenner-Crisp P., Grant D.,Hartley M., Knaap A., Kroese D., Mangelsdorf I., Meek E., Rice J.M., Younes M.: IPCSConceptual Framework for Evaluating a Mode of Action for Chemical Carcinogenesis.,Regulatory Toxicology and Pharmacology 34(2), 146-152, 2001.http://www.who.int/ipcs/methods/harmonization/areas/cancer/en/Tennant RW. Stratification of rodent carcinogenicity bioassay results to reflect relative humanhazard. Mutation Research 286, 111-118, 1993.WHO working group: Evaluation and use of epidemiological evidence for environmentalhealth risk assessment: Who Guideline Document. Env. Health Perspect. 108, 997-1002,2000.3.7 REPRODUCTIVE TOXICITY3.7.1 Definitions and general considerations for reproductive toxicityAnnex I: 3.7.1.1. Reproductive toxicity includes adverse effects on sexual function and fertility inadult males and females, as well as developmental toxicity in the offspring. The definitionspresented below are adapted from those agreed as working definitions in IPCS/EHC DocumentN°225, Principles for Evaluating Health Risks to Reproduction Associated with Exposure toChemicals. For classification purposes, the known induction of genetically based heritable effectsin the offspring is addressed in Germ Cell Mutagenicity (section 3.5), since in the presentclassification system it is considered more appropriate to address such effects under the separatehazard class of germ cell mutagenicity.In this classification system, reproductive toxicity is subdivided under two main headings:(a)(b)Adverse effects on sexual function and fertility;Adverse effects on development of the offspring.Some reproductive toxic effects cannot be clearly assigned to either impairment of sexual functionand fertility or to developmental toxicity. Nonetheless, substances with these effects, or mixturescontaining them, shall be classified as reproductive toxicants.321

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.1.2. For the purpose of classification the hazard class Reproductive Toxicity is differentiatedinto:– adverse effects– on sexual function and fertility, or– on development;– effects on or via lactation3.7.1.3. Adverse effects on sexual function and fertilityAny effect of substances that has the potential to interfere with sexual function and fertility. Thisincludes, but is not limited to, alterations to the female and male reproductive system, adverseeffects on onset of puberty, gamete production and transport, reproductive cycle normality, sexualbehaviour, fertility, parturition, pregnancy outcomes, premature reproductive senescence, ormodifications in other functions that are dependent on the integrity of the reproductive systems.3.7.1.4. Adverse effects on development of the offspringDevelopmental toxicity includes, in its widest sense, any effect which interferes with normaldevelopment of the conceptus, either before or after birth, and resulting from exposure of eitherparent prior to conception, or exposure of the developing offspring during prenatal development, orpostnatally, to the time of sexual maturation. However, it is considered that classification under theheading of developmental toxicity is primarily intended to provide a hazard warning for pregnantwomen, and for men and women of reproductive capacity. Therefore, for pragmatic purposes ofclassification, developmental toxicity essentially means adverse effects induced during pregnancy,or as a result of parental exposure. These effects can be manifested at any point in the life span ofthe organism. The major manifestations of developmental toxicity include (1) death of thedeveloping organism, (2) structural abnormality, (3) altered growth, and (4) functional deficiency.3.7.1.1 Special considerations on effects on or via lactationThis classification is intended to indicate when a substance may cause harm due to its effectson or via lactation. This can be due to the substance being absorbed by women and adverselyaffecting milk production or quality, or due to the substance (or its metabolites) being presentin breast milk in amounts sufficient to cause concern for the health of a breastfed child.Annex I: 3.7.1.5. Adverse effects on or via lactation are included under reproductive toxicity, butfor classification purposes such effects are treated separately. This is because it is desirable to beable to classify substances specifically for an adverse effect on lactation so that a specific hazardwarning about this effect can be provided for lactating mothers.Therefore, if the adverse effects that lead to impaired development in the offspring also occurafter in utero exposure then the substance would also be classified for developmental toxicity.In other words, the classification for effects on or via lactation is independent of considerationof the reproductive toxicity of the substance, and a substance can be classified for effects onor via lactation whether or not the substance is also classified for reproductive toxicity.Classification for effects on or via lactation alone is not sufficient for a substance to be subjectto harmonised classification and labelling in accordance with CLP, Article 36.322

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.2 Classification of substances for reproductive toxicity3.7.2.1 Identification of hazard information3.7.2.1.1 Identification of human dataEpidemiological studies as well as clinical data and case reports may be available as stated inCLP, Annex I, 3.7.2.2.3 and further under IR/CSA, Section R.7.6.3.2.3.7.2.1.2 Identification of non human dataIn-vitro, animal data and non-testing information used for classification is outlined in CLPAnnex I, 3.7.2.5. and further specific references to different testing methods are listed inIR/CSA, Section R.7.6.3.1.3.7.2.2 Classification criteriaAnnex I: 3.7.2.1.1. For the purpose of classification for reproductive toxicity, substances areallocated to one of two categories. Within each category, effects on sexual function and fertility,and on development, are considered separately. In addition, effects on lactation are allocated to aseparate hazard category.Table 3.7.1 (a)Hazard categories for reproductive toxicantsCategoriesCATEGORY 1Category 1ACategory 1BCATEGORY 2CriteriaKnown or presumed human reproductive toxicantSubstances are classified in Category 1 for reproductive toxicity when theyare known to have produced an adverse effect on sexual function andfertility, or on development in humans or when there is evidence fromanimal studies, possibly supplemented with other information, to provide astrong presumption that the substance has the capacity to interfere withreproduction in humans. The classification of a substance is furtherdistinguished on the basis of whether the evidence for classification isprimarily from human data (Category 1A) or from animal data (Category1B).Known human reproductive toxicantThe classification of a substance in this Category 1A is largely based onevidence from humans.Presumed human reproductive toxicantThe classification of a substance in this Category 1B is largely based ondata from animal studies. Such data shall provide clear evidence of anadverse effect on sexual function and fertility or on development in theabsence of other toxic effects, or if occurring together with other toxiceffects the adverse effect on reproduction is considered not to be asecondary non-specific consequence of other toxic effects. However, whenthere is mechanistic information that raises doubt about the relevance ofthe effect for humans, classification in Category 2 may be moreappropriate.Suspected human reproductive toxicantSubstances are classified in Category 2 for reproductive toxicity whenthere is some evidence from humans or experimental animals, possibly323

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012324supplemented with other information, of an adverse effect on sexualfunction and fertility, or on development, and where the evidence is notsufficiently convincing to place the substance in Category 1. If deficienciesin the study make the quality of evidence less convincing, Category 2could be the more appropriate classification.Such effects shall have been observed in the absence of other toxic effects,or if occurring together with other toxic effects the adverse effect onreproduction is considered not to be a secondary non-specific consequenceof the other toxic effects.3.7.2.2.1 Classification in the presence of parental toxicity3.7.2.2.1.1 Effects to be considered in the presence of marked systemic effectsIn general all findings on reproductive toxicity should be considered for classificationpurposes irrespective of the level of parental toxicity. A comparison between the severity ofthe effects on fertility/development and the severity of other toxicological findings must beperformed.Fertility effectsAdverse effects on fertility and reproductive performance seen only at dose levels causingmarked systemic toxicity (e.g. lethality, dramatic reduction in absolute body weight, coma)are not relevant for classification purposes.There is no established relationship between fertility effects and less marked systemictoxicity. Therefore it should be assumed that effects on fertility seen at dose levels causingless marked systemic toxicity are not a secondary consequence of this toxicity. However,mating behaviour can be influenced by parental effects not directly related to reproduction(e.g. sedation, paralysis), and such effects on mating behaviour may not warrantclassification.Developmental effects:Annex I: 3.7.2.4. Maternal toxicity3.7.2.4.1. Development of the offspring throughout gestation and during the early postnatal stagescan be influenced by toxic effects in the mother either through non-specific mechanisms related tostress and the disruption of maternal homeostasis, or by specific maternally-mediated mechanisms.In the interpretation of the developmental outcome to decide classification for developmentaleffects it is important to consider the possible influence of maternal toxicity. This is a complexissue because of uncertainties surrounding the relationship between maternal toxicity anddevelopmental outcome. Expert judgement and a weight of evidence approach, using all availablestudies, shall be used to determine the degree of influence that shall be attributed to maternaltoxicity when interpreting the criteria for classification for developmental effects. The adverseeffects in the embryo/foetus shall be first considered, and then maternal toxicity, along with anyother factors which are likely to have influenced these effects, as weight of evidence, to help reacha conclusion about classification.3.7.2.4.2. Based on pragmatic observation, maternal toxicity may, depending on severity, influencedevelopment via non-specific secondary mechanisms, producing effects such as depressed foetalweight, retarded ossification, and possibly resorptions and certain malformations in some strains ofcertain species. However, the limited number of studies which have investigated the relationshipbetween developmental effects and general maternal toxicity have failed to demonstrate aconsistent, reproducible relationship across species. Developmental effects which occur even in thepresence of maternal toxicity are considered to be evidence of developmental toxicity, unless it canbe unequivocally demonstrated on a case-by-case basis that the developmental effects aresecondary to maternal toxicity. Moreover, classification shall be considered where there is a

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012significant toxic effect in the offspring, e.g. irreversible effects such as structural malformations,embryo/foetal lethality, significant post-natal functional deficiencies.3.7.2.4.3. Classification shall not automatically be discounted for substances that producedevelopmental toxicity only in association with maternal toxicity, even if a specific maternallymediatedmechanism has been demonstrated. In such a case, classification in Category 2 may beconsidered more appropriate than Category 1. However, when a substance is so toxic that maternaldeath or severe inanition results, or the dams are prostrate and incapable of nursing the pups, it isreasonable to assume that developmental toxicity is produced solely as a secondary consequence ofmaternal toxicity and discount the developmental effects. Classification is not necessarily theoutcome in the case of minor developmental changes, when there is only a small reduction infoetal/pup body weight or retardation of ossification when seen in association with maternaltoxicity.Adverse effects on postnatal survival and growth seen only at dose levels causing maternaltoxicity may be due to lack of maternal care or other causes such as adverse effects on or vialactation or developmental toxicity. In case post-natal effects are caused by lack of maternalcare classification for developmental effects may not be warranted.3.7.2.2.1.2 Relevance of specific effects in the parentAll types of reproductive toxic effects may be considered as secondary to parental toxicity.With current knowledge it is not possible to identify specific effects indicating toxicity inparental animals which do not have any relevance to reproductive toxicity (e.g. peroxisomeproliferation). However parental toxicity that is less than marked should not influence theclassification for reproductive toxicity independent of the specific parental effects observed.In general it is very difficult to prove a causal relationship between a parentally mediatedmechanism and adverse effects in the offspring. Usually data are insufficient to conclude if aneffect on the offspring is a direct effect or secondary to parental toxicity. In order to determinewhether a reproductive toxic effect is independent or secondary to a parental effect, it wouldbe most appropriate to correlate individual data for offspring and their parents. Nevertheless,associations between parental and offspring effects do not by default prove a causalrelationship.In cases where a causal relationship is established between reproductive and parental toxicityand the effects on the offspring can be proved to be secondary to maternal toxicity, they maystill be relevant for developmental classification, dependent on the severity of the effects.A comparison between the severity of the maternal toxicity and the severity of the findings inthe offspring must be performed. There are several examples showing that the developingorganism can be more susceptible and the long-term consequences can be more severe than inthe adult. The mother might recover while the offspring could be permanently affected.Annex I: 3.7.2.4.4. Some of the end points used to assess maternal effects are provided below. Dataon these end points, if available, need to be evaluated in light of their statistical or biologicalsignificance and dose response relationship.Maternal mortality:an increased incidence of mortality among the treated dams over the controls shall be consideredevidence of maternal toxicity if the increase occurs in a dose-related manner and can be attributedto the systemic toxicity of the test material. Maternal mortality greater than 10 % is consideredexcessive and the data for that dose level shall not normally be considered for further evaluation.Mating index(no. animals with seminal plugs or sperm/no. mated x 100)( 1 )325

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Fertility index(no. animals with implants/no. of matings x 100)Gestation length(if allowed to deliver)Body weight and body weight change:Consideration of the maternal body weight change and/or adjusted (corrected) maternal bodyweight shall be included in the evaluation of maternal toxicity whenever such data are available.The calculation of an adjusted (corrected) mean maternal body weight change, which is thedifference between the initial and terminal body weight minus the gravid uterine weight (oralternatively, the sum of the weights of the foetuses), may indicate whether the effect is maternal orintrauterine. In rabbits, the body weight gain may not be useful indicators of maternal toxicitybecause of normal fluctuations in body weight during pregnancy.Food and water consumption (if relevant):The observation of a significant decrease in the average food or water consumption in treated damscompared to the control group is useful in evaluating maternal toxicity, particularly when the testmaterial is administered in the diet or drinking water. Changes in food or water consumption needto be evaluated in conjunction with maternal body weights when determining if the effects notedare reflective of maternal toxicity or more simply, unpalatability of the test material in feed orwater.Clinical evaluations (including clinical signs, markers, haematology and clinical chemistry studies):The observation of increased incidence of significant clinical signs of toxicity in treated damsrelative to the control group is useful in evaluating maternal toxicity. If this is to be used as thebasis for the assessment of maternal toxicity, the types, incidence, degree and duration of clinicalsigns shall be reported in the study. Clinical signs of maternal intoxication include: coma,prostration, hyperactivity, loss of righting reflex, ataxia, or laboured breathing.Post-mortem data:Increased incidence and/or severity of post-mortem findings may be indicative of maternal toxicity.This can include gross or microscopic pathological findings or organ weight data, includingabsolute organ weight, organ-to-body weight ratio, or organ-to-brain weight ratio. When supportedby findings of adverse histopathological effects in the affected organ(s), the observation of asignificant change in the average weight of suspected target organ(s) of treated dams, compared tothose in the control group, may be considered evidence of maternal toxicity.( 1 ) It is recognised that the Mating index and the Fertility index can also be affected by the male.3.7.2.2.2 Substances causing effects on or via lactation326Annex I: Table 3.7.1 (b)Hazard category for lactation effectsEFFECTS ON OR VIA LACTATIONEffects on or via lactation are allocated to a separate single category. It is recognised that for manysubstances there is no information on the potential to cause adverse effects on the offspring vialactation. However, substances which are absorbed by women and have been shown to interfere withlactation, or which may be present (including metabolites) in breast milk in amounts sufficient tocause concern for the health of a breastfed child, shall be classified and labelled to indicate thisproperty hazardous to breastfed babies. This classification can be assigned on the:(a) human evidence indicating a hazard to babies during the lactation period; and/or

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(b) results of one or two generation studies in animals which provide clear evidence of adverse effectin the offspring due to transfer in the milk or adverse effect on the quality of the milk; and/or(c) absorption, metabolism, distribution and excretion studies that indicate the likelihood that thesubstance is present in potentially toxic levels in breast milk.There are the two general criteria for this classification.(i) …are absorbed by women and have been shown to interfere with lactation.This relates to effects in the mother that impact adversely on the breast milk, either in terms ofthe quantity produced or the quality of the milk produced (i.e. the composition). Any effect onthe quantity or quality of the breast milk is likely to be due to systemic effects in the mother.However, overt maternal toxicity may not be seen (e.g. the substance may just affect thetransfer of a nutrient into the milk with no consequence for the mother). The type andmagnitude of the maternal effects and their potential influence on lactation/milk productionneed to be considered on a case-by-case basis to determine whether classification for effectson or via lactation is necessary.If a substance causes marked overt systemic toxicity in the mother at the same dose level thenit is possible that this may indirectly impair milk production or impair maternal care as a nonspecificsecondary effect. The type and magnitude of the maternal effects and their potentialinfluence on lactation/milk production needs to be considered on a case-by-case basis usingexpert judgment. If there is robust evidence to indicate that the effects on lactation are notcaused directly by the substance then it should not be classified as such.A substance which does not cause overt toxicity in the mother but which interferes with milkproduction or quality will normally be classified for effects on or via lactation because in thiscase the effect on lactation is most likely a direct substance-related effect.(ii) … may be present (including metabolites) in breast milk in amounts sufficient to causeconcern for the health of a breastfed child.This relates to the ability of the substance (including metabolites), to enter the breast milk inamounts sufficient to cause a concern. When the effect on the offspring is caused by thesubstance (or metabolite) after transport through the milk then the maternal toxicity has norelevance for classification. In general, positive data should usually be available to show thata substance leads to an adverse effect in offspring due to effects on lactation to supportclassification. However, in exceptional circumstances, if there are substantiated grounds forconcern that the substance may have an adverse effect via lactation then it may be classifiedas such in the absence of direct evidence. This should be based on a quantitative comparisonof the estimated transfer via the milk and the threshold for toxicity in the pups. This mightapply in cases where the substance has the capacity to bioaccumulate which would lead to apotentially higher burden in the offspring, or where there is evidence that the offspring maybe more sensitive to the substance’s toxicity than adult.The mere presence of the substance in the milk alone, without a strong justification for aconcern to offspring, would normally not support classification for effects on or via lactation.3.7.2.3 Evaluation of hazard informationAppropriate classification will always depend on an integrated assessment of all availabledata and their interrelationship using a weight of evidence approach. Individual datasetsshould be analysed case by case using expert judgment.327

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.2.3.1 Use of data from standard repeat dose testsFertility effects:Toxicological effects, including marked effects, observed in a standard repeat dose studycould be considered valid for the pre-mating phase for adult females and the pre- and postmatingphase for adult males. However in case of contradictions between the standard repeatdose studies and reproductive studies, the result from the latter should be considered morerelevant.For pregnant and lactating females and juveniles data from standard repeat dose studiescannot easily be extrapolated.Developmental effects:A detailed assessment of toxicity in pregnant animals cannot be extrapolated from studieswith non-pregnant animals. However information from general toxicity studies might give anindication of the maternal toxicity that could be anticipated in a subsequent developmentaltoxicity study.3.7.2.3.2 Study designAssessment of the dose-response relationships of parental and reproductive toxicity endpoints and their possible interrelationship require study designs where the dose intervals arenot too far apart. This will improve dose-response assessment and will also reduce the chanceof masking malformations by severe toxicity (e.g. resorptions, lethality) at high dose levels.This may lead to experimental designs in which more than the standard three dose groups anda control are tested. Endpoints from repeat dose toxicity studies may be considered useful forinclusion in subsequent reproductive toxicity studies. These endpoints should be evaluatedboth in parental animals and in offspring.3.7.2.3.3 Evaluation of evidence relating to effects on or via lactation(a) Human evidence indicating a hazard to babies during the lactation period;This criterion acknowledges that human data, e.g. from epidemiological studies or casereports, indicating a hazard to babies during the lactation period can also be used to supportclassification for effects on or via lactation. The use of human data is self-explanatory and anystudy should be assessed on its merits for which expert judgment may be required.Observations in humans that give evidence of adverse effects in breastfed babies of mothersexposed to the chemical in question should be taken to provide clear evidence supportingclassification. Such studies which do not show an adverse effect need to be consideredcarefully. Human studies investigate the risk under the specific conditions of exposure, and anegative finding may just reflect inadequate methods to detect effects or insufficientexposures rather than prove the absence of a hazard.In practice, useful human data are likely to be rare due to the nature of the endpoint. Morelikely are survey type studies which measure the levels of the chemical in breast milk. Suchstudies may provide useful information on the potential for maternal exposure to lead to thepresence of the chemical in the breast milk and so they may be of use in assessing the needfor classification for effects on or via lactation.(b) Results of one or two generation studies in animals which provide clear evidence ofadverse effect in the offspring due to transfer in the milk or adverse effect on the quality ofthe milk;Ideally, studies will be available which inform directly on whether the substance causesadverse effects in the offspring due to an adverse effect on lactation. One generation or multi-328

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012generation reproductive toxicity studies, which involve direct exposure or exposure via themilk of the offspring postnatally, usually provide information on this. The most commonstudy performed today is the two-generation study, but one-generation studies with new studydesigns, like the screening study OECD TG421/422 or the developmental neurotoxicity studyOECD TG426, also exist. The value of these studies is that they directly observe the pupsduring lactation and any adverse effects, such as deaths, decreased viability, clinical signssuch as reduced bodyweight gain etc, can be directly observed and quantified. However,expert judgement is required to decide whether these effects in pups are due to a directadverse effect on lactation, or are due to impaired nursing behaviour which is a non specificsecondary consequence of maternal toxicity. If the impaired nursing behaviour is proven to bea substance related specific effect on behaviour, then classification for effects on or vialactation may be appropriate. It should also be noted that some developmental effectsresulting from exposure in utero would only manifest post-natally and those should not beused for classification for effects on or via lactation. Cross-fostering studies, where available,may help establish whether effects are due to in utero or lactational exposure. If there issufficient data that animal results are not relevant to humans, they should not be taken intoaccount.(c) Absorption, metabolism, distribution and excretion studies that indicate the likelihood thatthe substance is present in potentially toxic levels in breast milk;The criterion indicates that toxicokinetic studies showing that the substance can be present atpotentially toxic levels in breast milk can support classification. The implicit assumptionbehind this clause is that the pups may receive a body burden of the toxic entity throughsuckling that is sufficient to cause toxicity when the level of the toxic entity in the milk isabove a certain threshold level (“a level to cause concern”). There is no robust way toestimate what this threshold is, although the likely body burden expected in the breastfedchild may be compared to the toxicity data in adults (e.g. an appropriate NOAEL or BMD) toindicate whether toxicity is likely. The mere presence of a substance in the milk, without arobust argument that these levels may be potentially toxic to offspring would not normallysupport classification.The toxicokinetics of a substance and the likelihood that it will enter the breast milk may bepredicted on the basis of the physico-chemical properties of the chemical (e.g. using pKa,logP, water solubility, and molecular weight etc) and this information could be used as part ofthe argumentation outlined above. The potential of a substance to bioaccumulate followingrepeated exposure may also be an important factor to consider as this may contribute to thebody burden reaching a potentially toxic level in the offspring. Studies where theoffspring/neonates have extended exposure, such as multi-generation studies, implicitly allowfor bioaccumulation and so findings from these studies can, in themselves, be taken to provideinformation on the potential effects of bioaccumulation. Where these types of studies are notavailable, potential bioaccumulation can be taken into consideration as part of thetoxicokinetic assessment using expert judgement.There may be toxicokinetic and toxicodynamic reasons why neonates may potentially bemore or less vulnerable to a particular adverse effect than adults due to the fact that certainsystems (e.g. the immune and metabolic systems) and tissues/organs are immature and arestill developing. Whether the neonate is more or less vulnerable than adults will depend on thespecific chemical and will be determined by factors such as the hazardous properties of thechemical, its’ physico-chemical properties and how it is metabolised. Therefore, the relativesensitivity of neonates and adults to a substance must be judged on a case by case basis usingexpert judgement. In the absence of any reliable and robust information to inform on this, it329

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012should be assumed that neonates and adults are equivalent in terms of sensitivity to thesubstance.Overall, classification for effects on or via lactation can be assigned on the basis oftoxicokinetic data or a well substantiated estimate of the exposure through the milk aloneprovided that it is supported by an argument clearly justifying that the level present in thebreast milk would be likely to harm developing offspring.3.7.2.4 Decision on classificationAccording to CLP Annex I, Section 3.7.2.1.1, reproductive toxic substances are allocated toeither Category 1A, 1B or 2. Effects on lactation are allocated to a separate hazard categoryand should be ascribed to a substance irrespective if it classified in any other category forreproductive toxicity or not.3.7.2.5 Setting of specific concentration limitsArticle 10(1) Specific concentration limits and generic concentration limits are limits assigned to asubstance indicating a threshold at or above which the presence of that substance in anothersubstance or in a mixture as an identified impurity, additive or individual constituent leads to theclassification of the substance or mixture as hazardous.Specific concentration limits shall be set by the manufacturer, importer or downstream user whereadequate and reliable scientific information shows that the hazard of a substance is evident when thesubstance is present at a level below the concentrations set for any hazard class in Part 2 of Annex Ior below the generic concentration limits set for any hazard class in Parts 3, 4 and 5 of Annex I.In exceptional circumstances specific concentration limits may be set by the manufacturer, importeror downstream user where he has adequate, reliable and conclusive scientific information that ahazard of a substance classified as hazardous is not evident at a level above the concentrations setfor the relevant hazard class in Part 2 of Annex I or above the generic concentration limits set for therelevant hazard class in Parts 3, 4 and 5 of that Annex.3.7.2.5.1 ProcedureThe available data from animal and human studies are evaluated to establish the reproductivetoxicity dose descriptor, ED 10 (effective dose with a 10% effect level above the background),as described below. A preliminary conclusion as to whether the substance shows high,medium or low potency is taken based on the ED 10 data. The preliminary potency evaluationmay be modified after due consideration of a number of modifying factors as described inchapter 3.7.2.5.5. This results in the final potency group. Each final potency group isconnected with a generic concentration limit (GCL) or a specific concentration limit (SCL). Inthis way SCLs are then set taking into account all relevant considerations. See figure3.7.2.5.1. A background document containing the justification of the boundaries of thepotency groups and the SCLs is available in Annex VI to this document.It is noted that there may be alternative approaches to assess potency, such as basing it on theBMD Methodology (Bench Mark Dose). However such alternative methods are notelaborated in this current guidance, although this does not exclude their use. If alternativeapproaches are used, they have to be clearly justified from a scientific and regulatory point ofview (see Article 10, CLP) and they must be able to provide robust scientific proposals andjustifications.330

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 3.7.2.5.1 Procedure for setting SCL for reproductive toxicityDetermine ED 10 usingthe available dataDetermine preliminarypotency groupDetermine final potencygroup considering themodifying factorsDetermine SCL3.7.2.5.2 Cases where potency evaluation is difficult or unfeasibleThe process for evaluating potency assumes the availability of certain types of data. However,these data may not always be available. Also, the classification of substances as reproductivetoxicants may be based on information such as grouping, read-across and the use of QSARs(Guidance IR/CSA, sections R.6 and R.7.2.3.1). In such cases, no direct estimate of thereproductive toxicity potency based on an ED 10 value is possible. While there are often goodreasons for extrapolation of the hazardous properties from one or more substances to another,the expected potency of the individual substances within the group may vary. In these cases apotency evaluation may be difficult or impossible. However, determination of theclassification and the potency using non-testing methods is possible in some cases. Thesecases could include interpolation of an ED 10 within a group of substances with comparablestructures and effects or correction for molecular weight in case of extrapolation betweendifferent salts with comparable availability. If the classification of a substance in Category 2is done on the basis of "limited evidence", the quality of the available data will in such casesdetermine whether a potency assessment is possible. In cases where no further evaluation ispossible, the generic concentration limits of CLP apply. In general, more conclusive evidenceis required when moving a substance to a lower potency group than to a higher potencygroup.3.7.2.5.3 Determination of the ED 10 valueThe ED 10 value (as used for reprotoxicity SCLs) is the lowest dose which inducesreproductive toxic effects which fulfil the criteria for classification for reproductive toxicitywith an incidence or magnitude of 10% after correction for the spontaneous incidence (see in3.7.2.5.3.2).Determining exactly which effect or combination of effects is the one that fulfils theclassification criteria may seem difficult. However, for the majority of substances in thedatabase, the developmental effect(s) observed at the lowest dose level was(/were) an increasein malformations and/or lethalities of the offspring. The ED 10 for effects on sexual functionand fertility is mainly based on effects on fertility and histopathological changes of the331

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012reproductive organs. These effects clearly fulfil the classification requirements. Also,allocation to the final SCLs is based on a limited number of potency groups and not on theexact ED 10 value. Therefore, in practice, it is likely that the ED 10 values for several differenteffects fall into the same potency grouping, resulting in the same SCL.The ED 10 may be obtained either directly or by linear interpolation from experimental data orestimated using Bench Mark Dose (BMD) software. The use of BMD software will result in amore precise estimate of the ED 10 because all data from the dose-response curve are used. Theuse of BMD software is needed when an ED 10 cannot be determined using linear interpolationdue to the absence of a NOAEL when the LOAEL has an effect size above 10%. In general,however, the use of BMD software is not required because of the wide potency groups usedfor setting the SCLs. However, it could be important for substances which are close to theboundary of a potency group. When an ED 10 cannot be calculated by direct or linearinterpolation from experimental data or by the use of BMD software, interpolation betweenthe control group and the LOAEL should be used to determine the ED 10 . In such cases, onlySCLs below the GCL can be determined and not those above the GCL, if no other reliableinformation is available, because it may be difficult in these cases to prove the absence ofeffects at lower dose levels.3.7.2.5.3.1 Determination in practiceIn practice, often several effects on reproduction are observed in various studies, and theclassification is based on the weight of evidence of all results. As a first step, it should bedetermined whether the classification is for effects on development, for effects on sexualfunction and fertility or both. The effects used for classification for developmental toxicityshould be used to determine the potency for developmental toxicity only. The same applies toeffects on sexual function and fertility. This means that for substances fulfilling the criteriafor classification for both developmental effects and effects on sexual function and fertility,two ED 10 values are derived which may differ and lead eventually to different SCLs. Forboth developmental effects and effects on sexual function and fertility, the lowest ED 10 for theeffect(s) that fulfil the criteria for classification in the different studies, is then used as theED 10 that determines the potency of that substance. Where there are doubts as to whether aspecific effect fulfils the classification criteria, ED 10 values for different effects could be takenforward to the next step, when modifing factors are considered, to determine the impact.The calculation of the ED 10 by linear interpolation requires a different approach depending onwhether the effect is measured as an incidence (quantal data, non-parametric data), amagnitude (continuous data, parametric data) or both.3.7.2.5.3.2 Quantal or non-parametric dataFor effects that are measured as changes in incidence, such as an increase in the number ofmalformations or resorptions, the ED 10 is defined as the dose level at which 10% of the testpopulation above the incidence in the concurrent control shows the effect. There may beoccasions where the historical control data have to be taken into account (for example whenthe concurrent control data are atypical and close to the extremes of the historical data). Inthe example in Table 3.7.2.5.1, the ED 10 is 90 mg/kg bw/day because at this dose level 12% -2% (control) = 10% of the test population shows the effect above the incidence in the controlgroup.332

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Table 3.7.2.5.1 Example of the calculation of the ED 10Dose 0mg/kg10mg/kg30mg/kg90mg/kgMalformations 2% 3% 7% 12%For some effects the results of the calculation of the ED 10 based on the incidence in pups maybe different from that based on the incidence in litters. Scientific evidence may indicate whichparameter is more appropriate, but in the absence of such information it is not possible toestimate which ED 10 is more appropriate for a specific effect. In such cases, both theincidence in offspring and the incidence in litters should be calculated, and the lower ED 10value should be used.3.7.2.5.3.3 Continuous or parametric dataFor effects that are measured as changes in magnitude such as mean pup weight or testisweight, the ED 10 is defined as the dose at which a change of 10%, compared to the concurrentcontrol group, is observed. In the example in Table 3.7.2.5.2, the ED 10 is 19.3 mg/kg bw/daybecause at this dose level the mean foetal bodyweight is calculated to be 90% of the controlvalue. A 10% reduction of the control value of 6.2 g gives 5.58 g. Interpolation between 10and 30 mg/kg bw/day to a dose level which would be expected to result in a foetalbodyweight of 5.58 g gives a value of 19.3 mg/kg bw/day.Calculations: (30 – 10)/ (6 - 5.1) = 22.2 ; 6.0 – 5.58 = 0.42 ; 0.42 x 22,2 = 9.3 ; 10 + 9.3 = 19.3 mg/kgbw/day.Table 3.7.2.5.2 Example on the calculation of the ED 10dose 0 mg/kg 10 mg/kg 30 mg/kg 90 mg/kgMean foetalbodyweight (g) 6.2 6.0 5.1 4.5NOAELLOAEL333

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.2.5.3.4 Data combining incidence and magnitudeSome effects such as histopathological changes in the testis are a combination of effects onincidence and magnitude (grading of the effect by a pathologist). However, calculation of anED 10 taking both the incidence and the magnitude into account is not possible or at least morecomplex. The ED 10 should therefore be based on the incidence of the effect below or above acertain magnitude. The magnitude of the effects that will be selected as a starting point has tobe chosen carefully. Normally the particular effect size would be the lowest relevant for therespective classification. The ED 10 is then determined as the dose level at which theincidence, of effects with a magnitude above that of the starting point, is 10% above theincidence in the control group. In practice this means that the grading system is converted intoa simplified system where only percentages of animals in each dose group with an effect witha magnitude above the starting point are regarded as positive. However, it is recognised thatthis approach uses only a part of the actual data and is imprecise, and it may be appropriatethat other effects also be considered in determining the ED 10 .Table 3.7.2.5.3 Example on the calculation of the ED 10 for testicular effects (N=10)334Dose(mg/kg)Testicular degeneration (n)none slight moderate marked severe0 4 5 1 0 010 5 5 0 0 0NOAEL 30 5 4 1 0 0LOAEL 90 0 0 4 2 4For the example in Table 3.7.2.5.3, the effects observed in the 10 mg/kg and 30 mg/kg dosegroups have to be considered as equivalent to the effects of the control group so the NOAELis 30 mg/kg. The magnitude of the testicular effect in the control group and the 10 and 30mg/kg bw/day groups is slight or less. Because of the incidence observed in these threegroups, the level of damage estimated as the starting point magnitude is ‘slight’. The ED 10 isthen defined as a 10% increase of moderate effects or more above the control. In this examplethe incidences for moderate testicular degeneration or more are 10%, 0%, 10% and 100% atrespectively 0, 10, 30 and 90 mg/kg bw/day. The ED 10 is then defined as the dose level with20% (control plus 10%) of moderate testicular effects. The ED 10 would be 36.6 mg/kg bw/daybased on interpolation between 30 and 90 mg/kg bw/day to a dose with 20% animals withmoderate testicular degeneration or higher.3.7.2.5.3.5 Specific data typesNon-oral studiesIn most cases only oral studies will be available and used for determination of the potency.However, if the classification is based on the effects seen in non-oral studies or only non-oralstudies are available, then these data should also be used to determine the potency. Thisrequires route-to-route extrapolation of the external dermal or inhalatory dose to acorresponding oral dose. This should be done as described in the ECHA guidance oninformation requirements and chemical safety assessment in REACH (IR/CSA, section R.8).Extrapolation from dermal exposure to oral exposure should only be done when there aresufficient kinetic data on dermal availability because assuming a high dermal availability isnot a worst case assumption. In cases where such data are not available a direct comparison ofthe dermal dose with the oral potency ranges could be performed in exceptional cases.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012However, such comparison should not result in moving the substance to a lower potencygroup (higher ED 10 ) – only moving the substance to a higher potency group (lower ED 10 )should be considered.Extrapolation from inhalatory exposure to oral exposure can only be done when there aresufficient kinetic data on inhaled availability because assuming a high inhaled availability isnot a worst case assumption. If no inhalatory information on availability is available then itshould be assumed that the inhalation and oral availability are comparable. However, suchcomparison should not result in moving the substance to a lower potency group (higher ED 10 )– only moving the substance to a higher potency group (lower ED 10 ) should be considered.Human dataThe use of human data for ED 10 calculation has several drawbacks including limited data onexposure, limited data on the size of the exposed population and limited information onwhether the exposure included the window of sensitivity. For all these reasons, it is difficultto determine an ED 10 based on human data. Therefore, and because in most instances animaldata are also available for determining an ED 10 , these data are evaluated together on a case bycase basis. Guidance on the use of human data for the derivation of DNELs and DMELs hasbeen developed by ECHA and is available at the ECHA website, seehttp://guidance.echa.europa.eu/guidance4_en.htm3.7.2.5.4 Provisional evaluation of the potency classificationA preliminary potency evaluation applying the ED 10 value is made at this stage.ED 10 values can be used to place substances classified as a reproductive toxicant into selectedranges that define potency groups. In this way, it is possible to identify reproductive toxicantsof high, medium and low potency. For the purpose of determining the preliminary potencygroup, the boundaries in Table 3.7.2.5.4 are used.Table 3.7.2.5.4 Boundaries of the potency groups 48 .Potency groupHigh potency groupMedium potency groupLow potency groupBoundaries3.7.2.5.5 Modifying factorsED 10 value ≤ 4 mg/kg bw/day4 mg/kg bw/day < ED 10 value 400 mg/kg bw/dayED 10 value 400 mg/kg bw/day.Modifying factors are a means to account for case-specific data situations which indicate thatthe potency group for a substance as obtained by the preliminary assessment, should bechanged. While most modifying factors would result in a higher potency group than thepreliminary one, also the opposite could occur: If substance-specific knowledge is available(such as e.g. toxicokinetic information on a higher bioavailability in test animals vs. humans),also a lower potency class might be assigned.While some modifying factors should always be taken into account, other modifying factorscould be more relevant when the potency is close to the boundary between two groups (seeTable 3.7.2.5.4 above).48 see Annex VI of this guidance document for more details335

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Some modifying factors are of a more qualitative nature. When applied, they will simplypoint to a potency group different from the one resulting from the preliminary assessment.Other modifying factors might be quantifiable, at least on a semi-quantitative scale. In suchcases, a potency group higher (or lower) than the preliminary one should be chosen if theestimated size of the modifying factor exceeds the distance of the preliminary ED10 to theborder of the relevant (higher or lower) adjacent potency group.Furthermore, for some substances more than one modifying factor will apply. It will then takeexpert judgement to decide on how to reasonably combine all of these individual factors intoone overall modifying factor. In exceptional cases, such a combination of individual factorsmight even result in a change of two potency classes (e.g. assignment of the high potencyclass, where the preliminary assessment had resulted in the low potency class).In this context, it should be noted that several of the modifying factors may be interrelated.Moreover, some factors may have already been taken into account in deciding on theclassification as a reproductive toxicant. Where such considerations have been made, careshould be taken not to use that information again when determining the potency. For example,when the effects determining the ED 10 were observed at dose levels also causing maternaltoxicity, this should already have been taken into consideration during the classification andshould not be used again to set a higher SCL.3.7.2.5.5.1 Type of effect / severityThe type of effect(s) resulting in the same classification as reproductive toxicant differsbetween substances. Some effects could be considered as more severe than others, however,ranking different effects based on their severity is controversial and difficult to establishcriteria. Further, the effects of a developmental toxicant can differ between dose levels fromvariations via malformations to death of the foetuses. The adverse effects on fertility andsexual function of a substance can differ between dose levels from small changes in testeshistopathology through effects on fertility to an irreversible and complete absence of fertility.As the difference between the dose levels is often smaller than the proposed potency groups(factor 10-100) this will make no difference in most cases. Also classification is in most casesbased on severe effects like malformations or death of the foetuses for developmentaltoxicants and effects on fertility or histopathological changes of the reproductive organs forfertility toxicants. For most classified substances such severe effects were already observed atthe lowest dose with reproductive effects (Muller et al, 2012). Therefore, differentiationbetween types of effect is considered to have limited added value. Exceptions can be dealtwith on a case by case basis.For example, if the ED 10 results in a preliminary conclusion for the medium potency groupbut is close to the border for the high potency group and the ED 10 is based on a severe effectlike malformations or irreversible effects on sexual function and fertility then using the higherpotency group (lower ED 10 ) for that substance should be considered. To determine what is“close to the border” is to compare the distance to the next category border with thesignificance of modifying factors.3.7.2.5.5.2 Data availabilityThere are several aspects to this modifying factor, some of which are: limited data availability where certain test protocols are lacking and therefore certainparameters have not been evaluated, limited data availability where the spectrum of evaluated parameters is sufficient, butonly studies with limited duration are available, and336

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012 limited data availability where only a LOAEL, but no NOAEL could be identified.Where only limited data are available, such as a screening study (OECD 421 and 422), a 28-day repeated dose toxicity study or non-OECD studies which do not exclude the presence ofreproductive effects at lower dose levels, the calculated ED 10 should not be used to set a SCLabove the GCL.Furthermore it should be considered to assign a modifying factor accounting for thelimitations in the database in a similar approach to the one used in deriving DNELs underREACH. Guidance regarding the potential size of such a factor can be obtained from ECHA’sGuidance on IR/CSA R.8 (‘Characterisation of dose [concentration]-response for humanhealth’). Section R.8.4.3.1 ‘Assessment of factors relating to extrapolation’, givesrecommendations on how to set factors for extrapolating to longer study durations as well asfor compensation of the lack of a NOAEL or of the generally poor quality of a database.If there are only limited data which result in an ED 10 in the medium potency group which isclose to the border for the high potency group, then using the higher potency group should beconsidered. For example an ED 10 of 8 mg/kg bw/day might have been estimated based on aLOAEL for malformations in the absence of a NOAEL, This ED 10 is only higher by a factorof 2 (i.e 2 times the border of the high potency group of 4 mg/kg bw/d : see. Table 3.7.2.5.4above), and assigning the high potency group should be considered until additional data atlower dose levels are available. Thus, there is uncertainty, if the ED 10 based on extrapolationfrom and below the LOAEL in the absence of a NOAEL and a correction may be justified.The size of this uncertainty could be determined by the BMDL (Benchmark dose lower 95%-confidence bound). In such cases, the BMDL could be used as a potency estimate instead ofthe ED 10 .3.7.2.5.5.3 Dose-response relationshipThe ED 10 will in most cases probably be in the same range as the NOAEL and LOAEL.However, in cases of a shallow dose effect relationship curve, the LOAEL may sometimes beclearly below the ED 10 . In such situations, if a substance would fall into a lower potencygroup based on the ED 10 but into a higher potency group based on the LOAEL then the higherpotency group should be used for that substance.3.7.2.5.5.4 Mode or mechanism of actionIt is assumed that effects observed in animal studies are relevant to humans. Where it isknown that the mode or mechanism of action is not relevant for humans or is of doubtfulrelevance to humans, this should have been taken into account in the classification and shouldnot be used again as a modifying factor for potency. However, quantitative differences intoxicodynamics can be taken into account when not already taken into account in theclassification. In cases where mechanistic information shows a lower sensitivity in humansthan in experimental animals, this may move substances which are close to the potencyboundaries to a lower potency group. In cases where mechanistic information indicates ahigher sensitivity in humans than in experimental animals, this may move substances near thepotency boundaries to a higher potency group. In general, more conclusive evidence isrequired when moving a substance to a lower potency group than to a higher potency group.3.7.2.5.5.5 ToxicokineticsThe toxicokinetics of a substance can differ between the tested animal species and humans.Where a difference is known this should be taken into account when determining the potencygroup of a substance. This should be based on a comprehensive knowledge of all involvedtoxicokinetic factors and not only on a single parameter. Also differences in kinetics between337

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012pregnant and non-pregnant animals and transport to the foetus should be taken into account.Based on the available data, quantification of this modifying factor has to be performed on acase by case basis. This modifying factor can work in both directions, as e.g. bioavailability inhumans might be known to be lower or higher than in the animal species tested.. In general,more conclusive evidence is required when moving a substance to a lower potency group thanto a higher potency group.3.7.2.5.5.6 Bio-accumulation of substancesThe study design of, for example, developmental studies is aimed at exposure only duringdevelopment. For substances which bio-accumulate, the actual exposure in the time windowof sensitivity for some developmental effects may therefore be much lower than whenexposure at the same external dose level would have started long before the sensitivitywindow. Furthermore, human exposure may occur for a long period before the sensitivewindow. This should be taken into account when determining the potency group. Forsubstances for which no experimental data are available with respect to their potential foraccumulation, section R.7.12 of ECHA’s IR/CSA Guidance R.7c (‘Endpoint specificguidance’) provides some hints on how to make an informed estimate about a respectiveconcern.“Suspected” bio-accumulating substances should be considered as to whether they should bemoved into the next higher potency group (lower ED 10 ). However this should be consideredon a case by case basis and the “suspected” bio-accumulation ability should be justifed. In thecase that the following evidence should be available, the higher potency group would not benecessary: the relevant studies used for the ED 10 were performed in a way that internal dosescould have been expected to have reached a steady state during a sufficiently long partof the study time, and in particular with developmental studies during critical timewindows of development, or the increase in the internal dose caused by the accumulation versus that following asingle administration, is smaller than the distance between the ED 10 and the border tothe next higher potency group.For example, if a substance preliminarily assigned to the medium potency group is known orsuspected to be bio-accumulative and the ED 10 for development has been obtained from a prenataldevelopmental study in rats without any significant pre-treatment of the dams beforemating, assignment to the high potency category should be considered. Conversely, if reliabletoxicokinetic data demonstrate that steady state plasma levels after prolonged repeatedadministration do not exceed those after single exposure by more than a factor of 2, while thepreliminary ED 10 is 20 mg/kg bw/d (i.e. factor 5 from the border to the high potency category)changing the potency class might not appear necessary.338

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.2.5.6 Assigning specific concentration limits (SCLs)Based upon the preliminary potency evaluation using only the ED 10 and applying themodifying factors, a substance can be placed in the final potency group using the table below.The GCL or SCL of that substance can then be found in the same table.Table 3.7.2.5.5 SCLs for substances in each potency group and classification categoryCategory 1 Category 2Dose SCL Dose SCLGroup 1highpotencyGroup 2mediumpotencyGroup 3lowpotencyED 10 below 4mg/kg bw/dayED 10 > 4 mg/kgbw/day, and 4 mg/kgbw/day, and

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.6 Examples3.7.6.1 Examples of the determination of SCLsFour examples are given below:3.7.6.1.1 Example 11.IdentificationSubstance Name:XXXXXXReproH2. EU CLP classification3. ED 10 in animalsBrief summary1B360DOECD 414, Wistar rats, GD 6-19, 0, 20, 60, 180 mg/kg bw. The number of live foetuses per litterwas significantly reduced and the postimplantation loss was 43 % at the high dose compared toonly 8 % in the control being statistically significant.The mean foetal body weight was reduced by 14 %. Further, the incidence of externalmalformations (anasarca and/or cleft palate) was significantly increased. About 10 % of the highdose foetuses were affected (13/132 foetuses; in 7/22 litters) while no such changes were observedin the control.Skeletal malformations were also statistically significantly increased: 7.8 % affected foetuses perlitter (7/73 foetuses in 5/21 litters) were noted in the high dose group compared to 1.1 % in thecontrol. The incidences of shortened scapula (4/73 foetuses), bent radius/ulna (2/73 foetuses),malpositioned and bipartite sternebrae (2/73 foetuses) were statistically significantly increased. Softtissue variations (dilated renal pelvis and ureter) were significantly increased in foetuses from highdose dams compared to controls (27.1 % vs. 6.4 %).At 0, 20, 60, 180 mg/kg 7.9, 14.8, 9.6, 43 % postimplantation loss was found, respectively.Remarks on the study used for the determination of the ED 10Species, strain, sex:Female Wistar ratStudy type: OECD 414Route of administration:Effect descriptor for LOAEL:Mode of action:Genotoxicity classification:Potential to accumulate:Oral gavagePost-implantation loss, anasarca, cleft palateNot knownNoneNo data. not known340

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Determination of the ED 10 valueControl resorption rate (= postimplantation loss) is 7.9%. ED 10 rate would be 17.9%. Interpolationbetween NOAEL (classification) (9.6% at 60 mg/kg) and LOAEL (classification) (43% at 180mg/kg) leads to an ED 10 of 89.8 mg/kg bw/d.Calculation:(180 – 60 ) / (43 – 9.6) = 3.593 mg/kg per % (steepness). Going from 9.6% to 17.9% requiresaddition of 8.3%. This equals 8.3% * 3.593 mg/kg per % = 29.8 plus 60 as the starting point = 89.8mg/kg bw/day.The ED 10 for other relevant effects was above 89.8 mg/kg bw/day.Preliminary potency groupmedium4. Elements that may modify the preliminary potency evaluation4.1. Dose-response relationshipNot relevant as ED 10 not borderline.4.2. Type of effect / severityNot relevant as ED 10 not borderline.4.3. Data availabilityNot relevant. Only one valid study available.4.4. Mode of actionNo data.4.5. ToxicokineticsNo data.4.6. Bio-accumulationLittle information, only environmental. Accumulation in organisms is not to be expeceted due tothe calculated BCF at 3.16. The substance tends not to accumulate in biota due to the lowcalculated BCF (

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.6.1.2 Example 2 (developmental part only)1. IdentificationSubstance Name :XXXXXX2. EU CLP classificationReproH3. ED 10 in animalsBrief summary1B360 FDStudy used for the determination of the ED 10 :Pregnant females received daily gavage doses of 0, 25, 50, 100 or 175 mg/kg during thegestation period (GD 6-19).LOAEL effectSkeletalmalformationsRemarks on the study used for the determination of the ED 10Species, strain, sex:Rabbit, New Zealand White, femaleStudy type: Developmental 6-19Route of administration:GavageEffect descriptor for LOAEL: Skeletal malformations (axial skeleton, ribs)Mode of action:Genotoxicity classification:Potential to accumulate:0mg/kg2/22(9 %)Determination of the ED 10 value25mg/kg2/17(12 %)Substance is metabolised to a substance whichcauses the developmental effectNoneUnknown50mg/kg5/15(33%)100mg/kg10/19(53%)Clear maternal toxicity was evident only at the highest dose level.175mg/kg6/12(50%)ED 10 was determined as 33 mg/kg.Control skeletal malformations is 9%. ED 10 rate would be 19%. Interpolation betweenNOAEL (classification) (12% at 25 mg/kg) and LOAEL (classification) (33% at 50 mg/kg)leads to an ED 10 of 33.3 mg/kg bw/day.Calculation:(50– 25 ) / (33 – 12) = 1.19 mg/kg per % (steepness). Going from 12% to 19% requiresaddition of 7%. This equals 7% * 1.19 mg/kg per % = 8.3 plus 25 as the starting point = 33.3mg/kg bw/day.342

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Preliminary potency groupMedium potency group.4. Elements that may modify the preliminary potency evaluation:4.1 Dose-response relationshipThe effect on which the classification is based is the occurrence of malformations. Asthe lowest ED 10 was the ED 10 for skeletal malformations, this ED 10 was chosen as thebasis for the SCL. The dose effect relationship is clear. The ED 10 (33 mg/kg) is notborderline with the LOAEL. There is no reason to consider the dose-responserelationship to modify the potency of the substance.Skeletal malformations6050% malformations4030201000 50 100 150 200mg/kg4.2 Type of effect / severityThe effect on which the classification is based is the occurrence of malformations, which is asevere effect. Moving the substance to a higher potency group should be considered.4.3 Data availabilityNot relevant. Different studies are available showing a developmental effect on differentspecies (rat, mouse, rabbit).343

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.4 Mode of actionThe toxic metabolite has been extensively investigated and established as a strongembryotoxicant and teratogen. There is no mechanistic information showing a higher or alesser sensitivity in humans than in experimental animals.4.5 ToxicokineticsHuman and rat liver microsomal preparations (mixtures) have been shown to producequalitatively and quantitively similar oxidative metabolic products suggesting that the humanpathways for this substance may be similar to those observed in experimental animals.4.6 Bio-accumulationUnknown5. Allocation of potency group and SCLThe effect on which the classification is based is the occurrence of malformations. This is asevere effect.Due to the fact that the ED 10 (33 mg/kg) is not a borderline case, it is not justified to move thesubstance to the highest potency group although the ED 10 is based on a severe effect likemalformations.Medium potency, GCL.6. ReferencesConfidential344

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.6.1.3 Example 3 (limited to developmental toxicity)1. IdentificationSubstance Name :XXXXXX2. EU CLP classificationRepro1BH360 fD3. ED 10 in animalsBrief summaryRemarksonSeveral studies in rats were available for the evaluation of the developmental effect of thissubstance. These included 2-generation studies, developmental toxicity studies, and studieswith exposure in sensitive periods during gestation. The most relevant study for the evaluationof potency was considered to be a two-generation study performed according to the revisedOECD Test Guideline 416. In this study the substance was administered in the diet.Developmental toxicity was evident as reduced absolute and adjusted AGD in F1 and F2offspring as well as and reduced foetal and testicular weight in offspring. The NOAEL was 50mg/kg bw/day based on reduced AGD from 250 mg/kg bw/day. These effects were reportedin the absence of marked maternal toxicity. Effects on the reproductive organs were alsoreported in male offspring in the developmental toxicity studies at higher doses.the study used for the determination of the ED 10Species, strain, sex:CD(Sprague-Dawley) rats male and female:Study type:2-generation according to OECD 416Route of administration:Effect descriptor for LOAEL:Mode of action:Genotoxicity classification:Potential to accumulate:Oral in feedOverall: reduced anogenital distanceClassification: increase in areolae in malesAntiandrogenic effect, mechanism relevant forhumansNot classified for germ cell mutagenicityNo345

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Determination of the ED 10 valueCalculation of the ED 10 value: 416 mg/kg bw/dayDose (mg/kg bw/day)% male F1 with areola0 2.6350 0.0250 (NOAEL) 0.76750 (LOAEL) 32.3The ED 10 is calculated by interpolation between 250 and 750 mg/kg bw/day to a dose levelwith 10% above control level. Roughly, an increase of 30% above control was found at 750mg/kg bw/day. Interpolation between 250 and 750 mg/kg bw/day results in a dose of 16.67mg/kg bw/day for each % of increase in areola ((750-250)/30). A 10% increase (ED 10 ) isexpected at 250 + 10 * 16.67 = 416 mg/kg bw/day.Preliminary potency groupLow potency4. Elements that may modify the preliminary potency evaluation4.1 Dose-response relationshipA dose-response relationship on decreased AGD was evident for decrease in AGD in the twogenerationstudy. (AGD was decreased in male offspring in a dose-related pattern from 250mg/kg bw/day (1. 89 mm at 250 mg/kg bw/day and 1.70 mm at 750 mg/kg bw/day (control:2.06 mm)).4.2 Type of effect / severityDevelopment: reduced anogenital distance (absolute and adjusted) from 250 mg/kg bw/day inF1 and F2 offspring. Weight changes in the reproductive organs in F1 and F2 male offspring,and macroscopic and microscopic lesions in the reproductive organs in male offspring at 750mg/kg bw/day.Maternal toxicity: organ weight changes, and histopahological lesions in the liver graded asminimal in females at 750 mg/kg bw/day.NOAEL for developmental effects: 50 mg/kg bw/day based on reduced anogenital distancefrom 250 mg/kg bw/day in F1 and F2 offspring.NOAEL for maternal toxicity: 250 mg/kg bw/day.4.3 Data availabilityA two-generation study is considered relevant for the assessment of development toxicity.4.3 Mode of actionThe mechanism (antiandrogen activity) is considered relevant for humans.346

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.5 ToxicokineticsWhen metabolites are measured in urine, they are related to the day before exposure. Themetabolites of the substance in rats differ quantitatively from those in humans. In severalstudies the pattern of malformations induced by some of the metabolites were similar to thatproduced by the substance, suggesting that the metabolic products may be responsible for thedevelopmental toxicity.Although there is a difference in toxicokinetics between rats and humans, this difference isnot expected to result in a difference in potency between rats and humans as the available dataindicate comparable effects and potency of the metabolites.4.6 Bio-accumulationLow to medium bioaccumulation5. Allocation of potency group and SCLThe ED 10 was 416 mg/kg bw/day. The elements that may modify the potency evaluation wereconsidered to not modify the potency. This substance is shown to have a low potency.Therefore an SCL of 3 % should be applied.6. ReferencesConfidential.347

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.6.1.4 Example 41 IdentificationSubstance Name :XXXXXX2 EU CLP classificationRepro 2H361f3 ED 10 in animalsBrief summary:Only two repeated dose studies are available for this substance and no fertility studies. Inthe inhalatory repeated dose study testicular lesions were observed after exposure to 2.87mg/l for 4 exposures of 16 to 20 hours per week during 11 weeks. Other dose levels werenot tested. In the oral 90 day study, effects on the testes were observed after exposure to 660mg/kg bw/day. Other dose levels were not tested.Remarks on the study used for the determination of the ED 10Species, strain, sex:Rats, CD(SD)BR malesStudy type:90 days, 5 days per week, 120 day observationperiodRoute of administration:Effect descriptor for LOAEL:Mode of action:Genotoxicity classification:Potential to accumulate:Determination of the ED 10 valuePreliminary potency groupgavagetesticular atrophy in 50% of the animalsA metabolite is assumed to be causing the testiculareffects. A direct effect of this metabolite on theSertoli cells is postulated.noneunknownThe dose level of 660 mg/kg bw/day is considered as the LOAEL but in the absence of aNOAEL an ED 10 cannot be determined by interpolation or the BMD approach because onlyone dose level was tested. An ED 10 can be estimated based on interpolation between 660mg/kg bw/day (50% of the animals affected) and the control (0 % of the animals affected).This results in an ED 10 of 132 mg/kg bw/day by interpolation.Medium potency group4 Elements that may modify the preliminary potency evaluation4.1 Dose-response relationshipThere is no data available on the dose response relationship.348

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.2 Type of effect / severityThere are clear testicular effects. It is unknown whether these effects will result in functionaleffects on fertility as this has not been tested.4.3 Data availabilityThere is only limited data available at one exposure level.. A LOAEL can be determined but itin the absence of a NOAEL it cannot be excluded that effects on sexual organs occur at levelsbelow the LOAEL. The available data are considered as limited.4.4 Mode of actionA metabolite is assumed to be the cause of the testicular effects. A direct effect of thismetabolite on the Sertoli cells is postulated.4.5 ToxicokineticsUnknown4.6 Bio-accumulationUnknown5 Allocation of potency group and SCLAn ED 10 can only be estimated using interpolation between the only dose tested and thecontrols.. The resulting ED10 indicates medium potency. However, there is only very limiteddata. As there is only an LOAEL and no NOAEL, it cannot be excluded that testicular effectscan be induced at lower levels. However, there is no evidence that this substance can inducetesticular effects at dose levels below 4 mg/kg bw/day. Therefore, a medium potency isconsidered the best estimate based on the available data.6 ReferencesConfidential349

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.2.6 Decision logicThe decision logic which follows is provided here as additional guidance. It is stronglyrecommended that the person responsible for classification study the criteria before andduring use of the decision logic.Classification of substances for fertility or developmental effects:Does the substance have data on reproductive toxicity?NOClassificationnot possibleYESAccording to the criteria, is the substance:(a) Known human reproductive toxicant, or(b) Presumed human reproductive toxicant?Application of the criteria needs expert judgment in aweight of evidence approach.YESCategory 1DangerNOCategory 2According to the criteria, is the substance a suspectedhuman reproductive toxicant?Application of the criteria needs expert judgment in astrength and weight of evidence approach.YESWarningNONot classifiedClassification of substances for effects via lactation:Does the substance according to the criteria cause concernfor the health of breastfed children?YESAdditionalcategory foreffects on orvia lactationNONot classified350

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.3 Classification of mixtures for reproductive toxicity3.7.3.1 Classification criteriaReproductive toxicity classification of mixtures is based on the presence of an ingredientclassified for reproductive toxicity (see CLP Article 6(3) and Annex I, 3.7.3). Only in casethere is data available for the mixture itself which demonstrate effects not retrieved from theingredients, this data might be used for classification. If such data is not available for themixture itself, data on a similar mixture can be used in accordance to the bridging principle(see CLP Annex I, 1.1.3).Annex I: Table 3.7.2Generic concentration limits of ingredients of a mixture classified as reproduction toxicantsor for effects on or via lactation that trigger classification of the mixtureGeneric concentration limits triggering classificationof a mixture as:Ingredient classified as:Category 1AreproductiveCategory 1BreproductiveCategory 2reproductiveAdditionalcategory fortoxicanttoxicanttoxicanteffects on or vialactationCategory 1Areproductive toxicant 0,3 %[Note 1]Category 1Breproductive toxicant 0,3 %[Note 1]Category 2 reproductivetoxicant 3,0 %[Note 1]Additional category foreffects on or vialactation 0,3 %[Note 1]NoteThe concentration limits in the table above apply to solids and liquids (w/w units) as well as gases(v/v units).Note 1If a Category 1 or Category 2 reproductive toxicant or a substance classified for effects on or vialactation is present in the mixture as an ingredient at a concentration above 0,1 %, a SDS shall beavailable for the mixture upon request.3.7.3.1.1 When data are available for the individual ingredientsAnnex I: 3.7.3.1.1. The mixture shall be classified as a reproductive toxicant when at least oneingredient has been classified as a Category 1A, Category 1B or Category 2 reproductive toxicantand is present at or above the appropriate generic concentration limit as shown in Table 3.7.2 belowfor Category 1A, Category 1B and Category 2 respectively.3.7.3.1.2. The mixture shall be classified for effects on or via lactation when at least one ingredienthas been classified for effects on or via lactation and is present at or above the appropriate generic351

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012concentration limit as shown in Table 3.7.2 for the additional category for effects on or vialactation.3.7.3.1.2 When data are available for the complete mixtureAnnex I: 3.7.3.2.1 Classification of mixtures will be based on the available test data for theindividual ingredients of the mixture using concentration limits for the ingredients of the mixture.On a case-by-case basis, test data on mixtures may be used for classification when demonstratingeffects that have not been established from the evaluation based on the individual components. Insuch cases, the test results for the mixture as a whole must be shown to be conclusive taking intoaccount dose and other factors such as duration, observations, sensitivity and statistical analysis ofreproduction test systems. Adequate documentation supporting the classification shall be retainedand made available for review upon request.3.7.3.1.3 When data are not available for the complete mixture: bridging principlesAnnex I: 3.7.3.3.1 Subject to the provisions of paragraph 3.7.3.2.1, where the mixture itself has notbeen tested to determine its reproductive toxicity, but there are sufficient data on the individualingredients and similar tested mixtures to adequately characterise the hazards of the mixture, thesedata shall be used in accordance with the applicable bridging rules set out in section 1.1.3.352

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.3.2 Decision logicThe decision logic which follows is provided here as additional guidance. It is stronglyrecommended that the person responsible for classification study the criteria before andduring use of the decision logic.Classification of mixtures for fertility or developmental effects:Category 1Classification based on individual ingredients of the mixtureDoes the mixture contain one or more ingredients classifiedas a Category 1 reproductive toxicant at 0.3%?NOYESDangerDoes the mixture contain one or more ingredients classifiedas a Category 2 reproductive toxicant at 3 %?NOYESCategory 2Not classifiedWarningModified classification on a case-by-case basisTest data on mixtures may be used for classification when demonstrating effects that have not beenestablished from the evaluation based on the individual ingredients (CLP Annex I, 3.7.3.1.1, see alsoCLP Article 6(3)).Are test data availablefor the mixture itselfdemonstrating areproductive toxic effectnot identified from thedata on individualsubstances?YESAre the test results on themixture conclusive takinginto account dose and otherfactors such as duration,observations and analysis(e.g. statistical analysis, testsensitivity) of reproductivetoxicity test systems?YESClassify inappropriatecategoryNONODanger orWarningorCan bridging principles beapplied?YESNo classificationNOSee above: Classification based onindividual ingredients of the mixture.353

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Classification of mixtures for effects via lactation:Classification based on individual ingredients of the mixtureDoes the mixture contain one or more ingredientsclassified for effects on or via lactation at 0.3 %?NOYESAdditionalcategory foreffects on orvia lactationNot classifiedModified classification on a case-by-case basisTest data on mixtures may be used for classification when demonstrating effects that have not beenestablished from the evaluation based on the individual ingredients (CLP Annex I, 3.7.3.1.1, see alsoCLP Article 6(3)).Are test data availablefor the mixture itselfdemonstrating effects onor via lactation notidentified from the dataon individualsubstances?YESThe test results for the mixtureas a whole must be shown tobe conclusive taking intoaccount dose and other factorssuch as duration, observations,sensitivity and statisticalanalysis of reproductivetoxicity test systems.YESAdditionalcategory foreffects on orvia lactationNONOorCan bridging principles beapplied?YESNoclassificationNOSee above: Classification based onindividual ingredients of the mixture.354

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.7.4 Hazard communication in form of labelling for reproductive toxicity3.7.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 3.7.4.1. Label elements shall be used for substances or mixtures meeting the criteria forclassification in this hazard class in accordance with Table 3.7.3.ClassificationGHS PictogramsTable 3.7.3Label elements for reproductive toxicityCategory 1A orCategory 1BCategory 2Additional categoryfor effects on or vialactationNo pictogramSignal Word Danger Warning No signal wordHazard StatementPrecautionaryStatement PreventionPrecautionaryStatement ResponsePrecautionaryStatement StorageH360: Maydamagefertility or theunborn child (statespecific effect ifknown)(state route ofexposure if it isconclusively proventhat no other routes ofexposure cause thehazard)P201P202P281H361: Suspected ofdamaging fertility orthe unborn child (statespecific effect ifknown) (state route ofexposure if it isconclusively proventhat no other routes ofexposure cause thehazard)P201P202P281H362: May causeharm to breast-fedchildren.P201P260P263P264P270P308 + P313 P308 + P313 P308 + P313P405As shown in CLP Annex I, Table 3.7.3, a substance classified as reproductive toxicant inCategory 1A or 1B shall be assigned the hazard statements H360 and a substance classified inCategory 2 shall be assigned H361. Each of these two hazard statements includes thementioning of the adverse effects on sexual function and fertility or adverse effects ondevelopment of the offspring.Depending on the data available, the hazard statement H360 or H361 shall e.g. be assigned areproductive toxic substance: in the case the criteria for Category 1A/1B or 2 are fulfilled, foreither sexual function or fertility or developmental toxicity and when the other reproductiveeffect cannot be excluded.P405355

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In case reliable and adequate data are available on reproductive toxicity, (so that it is possibleto ascribe one category for the fertility effects and one category for developmental toxiceffects); it is possible to specify the hazard in the hazard statement. The resulting differentvariants of H360 and H361 are shown in the table below, which also provides some exampleswhen they should be assigned a substance.Table 3.7.4.1: Hazard statements for reproductive toxicity: H360 and H361, and theirspecificationsH360H361H360FH360DH361fH361dH360FDH361fdH360Fd“May damage fertility or the unborn child”Examples:1) a substance classified in Repr Cat 1A/B because of adverse effects on fertility andfor which developmental toxic effects cannot be excluded2) a substance classified in Repr Cat 1 A/B but the effects cannot be specified withrespect to fertility or developmental toxicity“Suspected of damaging fertility or the unborn child”Example:1) a substance classified in Repr. Cat 2 on the basis of effects on developmentaltoxicity and for which fertility effects cannot be excluded2) a substance classified in Repr. Cat 2 but the effects cannot be specified withrespect to fertility or developmental toxicity“May damage fertility.”Example: a substance classified in Repr Cat 1A/B on the basis of fertility effects andeffects on developmental toxicity can be excluded according to reliable and adequatedata“May damage the unborn child.”Example: a substance classified in Repr Cat 1A/B on the basis of developmentaltoxicity and effects on fertility can be excluded according to reliable and adequate data“Suspected of damaging fertility”.Example: a substance classified in Repr Cat 2 on the basis of fertility effects and effectson developmental toxicity can be excluded according to reliable and adequate dataSuspected of damaging the unborn child.Example: a substance classified in Repr Cat 2 on the basis of fertility effects and effectson developmental toxicity can be excluded according to reliable and adequate dataMay damage fertility. May damage the unborn child.Example: a substance classified in Repr Cat 1A/B on the basis of fertility effects anddevelopmental toxicity.Suspected of damaging fertility. Suspected of damaging the unborn child.Example: a substance classified in Repr Cat 2 on the basis of fertility effects anddevelopmental toxicity.May damage fertility. Suspected of damaging the unborn child.Example: a substance classified in Repr Cat 1A/B on the basis of fertility effects and inRepr Cat 2 on the basis of developmental toxicity.356

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012H360DfMay damage the unborn child. Suspected of damaging fertility.Example: a substance classified in Repr Cat 1A/B on the basis of developmentaltoxicity and classified in Repr Cat 2 on the basis of fertility effects.According to CLP Annex I, Section 3.7.4.1, the hazard statements shall be amended byspecifying the route of exposure if it is conclusively proven that no other routes of exposurewill lead to an adverse effect on sexual function or fertility or development of the offspring.When conclusively proven, it is meant that valid in vivo test data need to be available for allthree exposure routes clearly indicating that only one exposure route has caused positiveresults i.e. adverse effects on the reproduction. Moreover, such a finding should be consideredplausible with respect to the mechanism or mode of action. It is estimated that such a situationwould rarely occur. Thus, amendment of the hazard statement with the route of exposuregenerally does not have to be considered.3.7.4.2 Additional labelling provisionsThere are no additional labelling provisions for reproductive toxic substances and mixtures inCLP, however there are provisions laid out in Annex XVII to REACH. The packaging ofsubstances harmonised classified for reproductive toxicity Category 1A or Category 1B, andmixtures containing such substances, "must be marked visibly, legibly and indelibly asfollows: ‘Restricted to professional users’." (REACH, Annex XVII, point 30).3.7.5 Re-classification of substances and mixtures classified for reproductivetoxicity according to DSD and DPD3.7.5.1 Is direct “translation” of classification and labellingpossible?Generally yes. In case there is no re-evaluation of the data, the hazard statement specifyingboth 'damage to fertility' and 'damage to the unborn child' should be assigned. It is possible toomit the hazard statement specifying fertility or developmental effects; in case there areclearly negative results (see Section 3.7.4.1).However, in some very rare situations, a reproductive toxicant classified with Repr. Cat. 3;R62 may need classification with Repr. Cat. 1B H360 under CLP. According to Annex VI toDSD, for the classification of a substance into Category 2 for impaired fertility, there shouldnormally be clear evidence in one animal species, with supporting evidence on mechanism ofaction or site of action, or chemical relationship to other known anti-fertility agents or otherinformation from humans which would lead to the conclusion that effects would be likely tobe seen in humans. According to CLP, such supporting evidence is not needed.Classification for effects on or via lactation according to CLP is directly equivalent toassignment of R64 according to DSD as the criteria are essentially the same. Therefore, directtranslation of R64 to H362 is possible.357

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8 SPECIFIC TARGET ORGAN TOXICITY – SINGLE EXPOSURE (STOT-SE)3.8.1 Definitions and general considerations for STOT-SEAnnex 1: 3.8.1.1. Specific target organ toxicity (single exposure) is defined as specific, non lethaltarget organ toxicity arising from a single exposure to a substance or mixture. All significant healtheffects that can impair function, both reversible and irreversible, immediate and/or delayed and notspecifically addressed in Chapters 3.1 to 3.7 and 3.10 are included (see also 3.8.1.6).There are two hazard classes for single exposure toxicity: “Acute toxicity” and “STOT-SE”.These are independent of each other and both may be assigned to a substance or a mixture ifthe respective criteria are met. Acute toxicity refers to lethality and STOT-SE to non lethaleffects. However, care should be taken not to assign both classes for the same toxic effect,essentially giving a “double classification”, even where the criteria for both classes arefulfilled. In such a case the most appropriate class should be assigned.Acute toxicity classification is generally assigned on the basis of evident lethality (e.g.anLD 50 /LC 50 value) or where the potential to cause lethality can be concluded from evidenttoxicity (e.g. from fixed dose procedure). STOT-SE should be considered where there is clearevidence of toxicity to a specific organ especially when it is observed in the absence oflethality.Furthermore, specific toxic effects covered by other hazard classes are not included in STOT-SE. STOT-SE should only be assigned where the observed toxicity is not covered moreappropriately by another hazard class. For example, specific effects caused after a singleexposure like corrosion of skin or effects on the reproductive organs should be used forclassification for skin corrosion or reproductive toxicity, respectively, but not for STOT-SE.Annex 1: 3.8.1.4. Assessment shall take into consideration not only significant changes in a singleorgan or biological system but also generalised changes of a less severe nature involving severalorgans.3.8.1.5. Specific target organ toxicity can occur by any route that is relevant for humans, i.e.principally oral, dermal or inhalation.3.8.1.7. The hazard class Specific Target Organ Toxicity – Single Exposure is differentiated into:Specific target organ toxicity – single exposure, Category 1 and 2;Specific target organ toxicity – single exposure, Category 3.The hazard class STOT-SE has 3 categories, with Categories 1 and 2 being distinct fromCategory 3 in terms of the toxicity they cover and the criteria. Categories 1 and 2 for nonlethal “significant and/or severe toxic effects” are the basis for classification with the categoryreflecting the dose level required to cause the effect. Category 3 covers “transient effects”occurring after single exposure, specifically respiratory tract irritation (RTI) and narcoticeffects (NE). The relationship between Categories 1/2 vs. Category 3 is discussed in Section3.8.2.43.358

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.2 Classification of substances for STOT-SE3.8.2.1 Identification of hazard informationAnnex 1: 3.8.2.1.5. The information required to evaluate specific target organ toxicity comes eitherfrom single exposure in humans, such as: exposure at home, in the workplace or environmentally,or from studies conducted in experimental animals.CLP does not require testing of substances or mixtures for classification purposes. Theassessment is based on the respective criteria together with available adequate and robust testdata/information. Generally, information relevant to STOT-SE can be obtained from humanexperience or acute toxicity studies in animals.3.8.2.1.1 Identification of human dataRelevant information with respect to toxicity after single exposure may be available from casereports, epidemiological studies, medical surveillance and reporting schemes and nationalpoisons centres.Data on sensory irritation of the airways may be available from volunteer studies includingobjective measurements of RTI such as electrophysiological responses, data fromlateralization threshold testing, biomarkers of inflammation in nasal or bronchoalveolarlavage fluids (IR/CSA, Section 7.2.3.2). For more details see IR/CSA, Section 7.4.3.2 andR.7.2.3.8.2.1.2 Identification of non human dataAnnex 1: 3.8.2.1.5 The standard animal studies in rats or mice that provide this information areacute toxicity studies which can include clinical observations and detailed macroscopic andmicroscopic examination to enable the toxic effects on target tissues/organs to be identified. Resultsof acute toxicity studies conducted in other species may also provide relevant information.Annex 1: 3.8.2.1.7.3. Evidence from appropriate studies in experimental animals can furnish muchmore detail, in the form of clinical observations, and macroscopic and microscopic pathologicalexamination, and this can often reveal hazards that may not be life-threatening but could indicatefunctional impairment. Consequently all available evidence, and relevance to human health, mustbe taken into consideration in the classification process, …Non-testing dataPhysicochemical dataPhysicochemical properties, such as pH, physical form, solubility, vapour pressure, particlesize, can be important parameters in evaluating toxicity studies and in determining the mostappropriate classification especially with respect to inhalation where physical form andparticle size can have a significant impact on toxicity.(Q)SAR models, Read across“Non-testing” data (i.e. data not obtained from experimental methods) can be provided by theuse of techniques such as grouping/category formation, Quantitative and qualitative StructureActivity Relationship (Q)SAR models and expert systems, which generally relate physicochemicalproperties and chemical structure to toxicity. The use of these methods is describedin more detail in Section 2.3.2 and IR/CSA, Section R.7.4.4.1.359

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The potential use of (Q)SAR models for predicting effects relevant to STOT-SE Categories1/2 is currently quite limited and may only be applicable in specific cases. However, they maybe somewhat more useful for STOT-SE Category 3 where there are some well establishedrelationships between physicochemical properties or chemical structure and effects such asnarcosis and respiratory tract irritation. For instance substances such as aldehydes,unsaturated carbonic esters and reactive inorganic compounds are generally found to berespiratory tract irritants.In addition, there are systems which can predict the metabolism of substances. These can beuseful in providing information on the potential for the substance to be metabolised tosubstances with known toxicity. An example is certain esters, which after enzymatic cleavageto carbonic acids and alcohols in the nasal region, cause respiratory irritation.For more details see IR/CSA, Section 7.4.3.1.Testing dataAnimal dataThe standard tests on acute toxicity are listed in IR/CSA, Section R.7.4.3.1.For Category 1 and 2, in general terms, most studies involving single exposure via anyrelevant route of exposure, such as acute toxicity studies, can be used for classificationpurposes. Older acute toxicity studies which tended to only measure lethality as anobservational endpoint (e.g. to determine LD 50 /LC 50 ) will generally not provide usefulinformation for STOT-SE. However, newer acute toxicity test protocols, such as the fixeddoseand up-down procedures, have a wider range of observations on signs of toxicity andtherefore may provide information relevant for STOT-SE. Other standard studies, e.g.neurotoxicity tests, or ad-hoc studies designed to investigate acute toxicity, can also providevaluable information for STOT-SE.Care must be taken not to classify for STOT-SE for effects which are not yet lethal at acertain dose, but would lead to lethality within the numeric classification criteria. In otherwords, if lethality would occur at relevant doses then a classification for acute toxicity wouldtake precedence and STOT-SE would not be assigned.Although classification in Category 3 is primarily based on human data, if available, animaldata can be included in the evaluation. These animal data on RTI and NE will generally comefrom standard acute inhalation studies, although it is possible that narcosis could be observedin studies using other routes. Standard acute toxicity tests are often more useful for Category3 than for STOT-SE Categories 1/2 because overt findings of narcosis and RTI are more oftenreported in clinical observations.The Alarie test gives specific information on the potential for sensory irritation. Further,information on this test and its limitations can be found in IR/CSA, Section R.7.2.Furthermore the Inhalation Hazard Test (Annex to OECD TG 403) might give information onthe potential for RTI of volatile substances. Though the focus of STOT-SE is on effectscaused by single exposure, data from studies with repeated exposure might give additionalvaluable information, especially with respect to the underlying mode of action of RTI.In vitro dataSince there are currently no in vitro tests that have been officially adopted by the EU orOECD for assessment of acute toxicity, there are also no useful test systems for STOT-SE(see IR/CSA, Section R.7.4.3.1). Any available studies should be assessed using expertjudgement.360

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.2.2 Classification criteria for Categories 1 and 2Annex I: 3.8.2.1.1. Substances are classified for immediate or delayed effects separately, by the useof expert judgement (see 1.1.1) on the basis of the weight of all evidence available, including theuse of recommended guidance values (see 3.8.2.1.9). Substances are then placed in Category 1 or 2,depending upon the nature and severity of the effect(s) observed (Table 3.8.1).CategoriesCategory 1Category 2Table 3.8.1Categories for specific target organ toxicity-single exposureCriteriaSubstances that have produced significant toxicity in humans or that, on the basisof evidence from studies in experimental animals, can be presumed to have thepotential to produce significant toxicity in humans following single exposureSubstances are classified in Category 1 for specific target organ toxicity (singleexposure) on the basis of:(a) reliable and good quality evidence from human cases or epidemiologicalstudies; or(b) observations from appropriate studies in experimental animals in whichsignificant and/or severe toxic effects of relevance to human health wereproduced at generally low exposure concentrations. Guidancedose/concentration values are provided below (see 3.8.2.1.9) to be used as partof weight-of-evidence evaluation.Substances that, on the basis of evidence from studies in experimental animals canbe presumed to have the potential to be harmful to human health following singleexposureSubstances are classified in Category 2 for specific target organ toxicity (singleexposure) on the basis of observations from appropriate studies in experimentalanimals in which significant toxic effects, of relevance to human health, wereproduced at generally moderate exposure concentrations. Guidancedose/concentration values are provided below (see 3.8.2.1.9) in order to help inclassification.In exceptional cases, human evidence can also be used to place a substance inCategory 2 (see 3.8.2.1.6).Note: Attempts shall be made to determine the primary target organ of toxicity and to classify forthat purpose, such as hepatotoxicants, neurotoxicants. The data shall be carefully evaluated and,where possible, secondary effects should not be included (e.g. a hepatotoxicant can producesecondary effects in the nervous or gastro-intestinal systems).3.8.2.1.2. The relevant route or routes of exposure by which the classified substance producesdamage shall be identified (see 3.8.1.5).STOT-SE Category 1 and 2 is assigned on the basis of findings of “significant” or “severe”toxicity. In this context “significant” means changes which clearly indicate functionaldisturbance or morphological changes which are toxicologically relevant. “Severe” effects aregenerally more profound or serious than “significant” effects and are of a considerablyadverse nature with significant impact on health. Both factors have to be evaluated by weightof evidence and expert judgement.361

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.2.2.1 Guidance valuesAnnex 1: 3.8.2.1.9.1 In order to help reach a decision about whether a substance shall be classifiedor not, and to what degree it shall be classified (Category 1 or Category 2), dose/concentration‘guidance values’ are provided for consideration of the dose/concentration which has been shownto produce significant health effects.Annex 1: 3.8.2.1.9.3. The guidance value (C) ranges for single-dose exposure which has produceda significant non-lethal toxic effect are those applicable to acute toxicity testing, as indicated inTable 3.8.2.Table 3.8.2Guidance value ranges for single-dose exposures aGuidance value ranges for:*Route of exposure Units Category 1 Category 2 Category 3Oral (rat)Dermal (rat or rabbit)mg/kg bodyweightmg/kg bodyweightC ≤ 300 2000 ≥ C > 300 Guidancevalues do notapply bC ≤ 1000 2000 ≥ C > 1000Inhalation (rat) gas ppmV/4h C ≤ 2500 20000 ≥ C > 2500Inhalation (rat) vapour mg/l/4h C ≤ 10 20 ≥ C > 10Inhalation (rat)dust/mist/fumeNotemg/l/4h C ≤ 1.0 5,0 ≥ C >1,0(a) The guidance values and ranges mentioned in Table 3.8.2 above are intended only for guidancepurposes, i.e. to be used as part of the weight of evidence approach, and to assist with decisionabout classification. They are not intended as strict demarcation values.(b) Guidance values are not provided for Category 3 substances since this classification is primarilybased on human data. Animal data, if available, shall be included in the weight of evidenceevaluation.* Note: There is a misprint inAnnex I, Table 3.8.2; the heading 'Guidance value ranges for:' should also belongto the column 'Category 1'.Where significant or severe toxicity has been observed in animal studies, the dose/exposurelevel causing these effects is compared to the guidance values provided to determine ifclassification in Category 1 or 2 is most appropriate.In cases of inhalation studies with exposure times different to 4 hours an extrapolation can beperformed similar to the one described in the Chapter 3.1 for Acute Toxicity.3.8.2.3 Classification criteria for Category 3: Transient target organ effectsCurrently, the criteria for classification in Category 3 only cover the transient effects of“respiratory tract irritation” and “narcotic effects”.362

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: Table 3.8.1 (continued)Categories for specific target organ toxicity-single exposureCategoriesCategory 3CriteriaTransient target organ effectsThis category only includes narcotic effects and respiratory tract irritation.These are target organ effects for which a substance does not meet the criteriato be classified in Categories 1 or 2 indicated above. These are effects whichadversely alter human function for a short duration after exposure and fromwhich humans may recover in a reasonable period without leaving significantalteration of structure or function. Substances are classified specifically forthese effects as laid down in 3.8.2.2Annex 1: 3.8.2.2.1 Criteria for respiratory tract irritationThe criteria for classifying substances as Category 3 for respiratory tract irritation are:(a) respiratory irritant effects (characterized by localized redness, oedema, pruritis and/or pain)that impair function with symptoms such as cough, pain, choking, and breathing difficulties areincluded. This evaluation will be based primarily on human data.(b) subjective human observations could be supported by objective measurements of clearrespiratory tract irritation (RTI) (such as electrophysiological responses, biomarkers ofinflammation in nasal or bronchoalveolar lavage fluids).(c) he symptoms observed in humans shall also be typical of those that would be produced in theexposed population rather than being an isolated idiosyncratic reaction or response triggeredonly in individuals with hypersensitive airways. Ambiguous reports simply of “irritation” shallbe excluded as this term is commonly used to describe a wide range of sensations includingthose such as smell, unpleasant taste, a tickling sensation, and dryness, which are outside thescope of classification for respiratory irritation.(d) there are currently no validated animal tests that deal specifically with RTI, however, usefulinformation may be obtained from the single and repeated inhalation toxicity tests. Forexample, animal studies may provide useful information in terms of clinical signs of toxicity(dyspnoea, rhinitis etc) and histopathology (e.g. hyperemia, edema, minimal inflammation,thickened mucous layer) which are reversible and may be reflective of the characteristicclinical symptoms described above. Such animal studies can be used as part of weight ofevidence evaluation.(e) this special classification would occur only when more severe organ effects including in therespiratory system are not observed.It is clearly indicated in the CLP that there are currently no validated animal tests that dealspecifically with RTI, but that animal studies can be used as a part of weight of evidenceevaluation (3.8.2.2.1.2(d)). However when there are no data in human and animal datasuggesting RTI effects, expert judgement is needed to estimate the severity of the effectsobserved in animals, the conditions of the test, the physical-chemical properties of thesubstance and whether those considerations alone might be sufficient for a classification inCategory 3 for RTI.The generic term RTI covers two different effects: “sensory irritation” and “local cytotoxiceffects”. Classification in STOT-SE Category 3 for respiratory tract irritation is generallylimited to local cytotoxic effects.363

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Sensory irritation refers to the local and central reflex interaction of a substance with theautonomic nerve receptors, which are widely distributed in the mucosal tissues of the eyesand upper respiratory tract. It helps to minimize exposure by decreasing the respiration-timevolumeand inducing the exposed to leave the areas of irritant concentrations, if possible.Sensory irritation-related effects are fully reversible given that its biological function is toserve as a warning against substances that could damage the airways.Local cytotoxic irritant effects induce tissue changes at the site of contact which can bedetected by clinico-pathological or pathological methods. Such effects may induce longlasting functional impairment of the respiratory system.The basic mechanisms underlying morphological changes comprise cytotoxicity andinduction of inflammation. Based on the quality and severity of morphological changes, thefunction of the respiratory system will be impaired, which may lead to the development ofconsequential systemic effects, i.e. there might be consequences on distal organs by adiminution of the oxygen supply. As the functional impairment is seldom evaluated byexperimental inhalation studies in animals, data on functional changes will mainly beavailable from experience in humans.Further see IR/CSA, Section R.7.2.Annex 1: 3.8.2.2.2. Criteria for narcotic effectsThe criteria for classifying substances as Category 3 for narcotic effects are:(a)(b)central nervous system depression including narcotic effects in humans such as drowsiness,narcosis, reduced alertness, loss of reflexes, lack of coordination, and vertigo are included.These effects can also be manifested as severe headache or nausea, and can lead to reducedjudgment, dizziness, irritability, fatigue, impaired memory function, deficits in perceptionand coordination, reaction time, or sleepiness.narcotic effects observed in animal studies may include lethargy, lack of coordination, lossof righting reflex, and ataxia. If these effects are not transient in nature, then they shall beconsidered to support classification for Category 1 or 2 specific target organ toxicity singleexposure.3.8.2.4 Evaluation of hazard information on STOT-SE forsubstances3.8.2.4.1 Evaluation of human dataAnnex I: 3.8.2.1.6. In exceptional cases, based on expert judgement, it is appropriate to placecertain substances with human evidence of target organ toxicity in Category 2:(a) when the weight of human evidence is not sufficiently convincing to warrant Category 1classification, and/or(b) based on the nature and severity of effects.Dose/concentration levels in humans shall not be considered in the classification and any availableevidence from animal studies shall be consistent with the Category 2 classification. In other words,if there are also animal data available on the substance that warrant Category 1 classification, thesubstance shall be classified as Category 1.Annex 1: 3.8.2.1.7.2. Evidence from human experience/incidents is usually restricted to reports ofadverse health consequence, often with uncertainty about exposure conditions, and may not providethe scientific detail that can be obtained from well-conducted studies in experimental animals.364

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex 1: 3.8.2.1.10.2. When well-substantiated human data are available showing a specific targetorgan toxic effect that can be reliably attributed to single exposure to a substance, the substanceshall normally be classified. Positive human data, regardless of probable dose, predominates overanimal data. Thus, if a substance is unclassified because specific target organ toxicity observed wasconsidered not relevant or significant to humans, if subsequent human incident data becomeavailable showing a specific target organ toxic effect, the substance shall be classified.Human data are potentially very valuable for determining an appropriate classification as theyprovide direct evidence on the effects of a substance in humans. However, the evaluation ofhuman data is often made difficult by various limitations frequently found with the types ofstudies and data highlighted in Section 3.8.2.4.1. These include uncertainties relating toexposure assessment (i.e. unreliable information on the amount of a substance the subjectswere exposed to or ingested) and confounding exposures to other substances. As a result itshould be acknowledged that human data often do not provide sufficiently robust evidence ontheir own to support classification but may contribute to a weight of evidence assessment withother available information such as animal studies.Categories 1 and 2In general, where reliable and robust human data are available showing that the substancecauses significant target organ toxicity these take precedence over other data, and directlysupport classification in Category 1. Available animal data may support this conclusion but donot detract from it (e.g. if the same effect is not observed in animals).In exceptional cases, where target organ toxicity is observed in humans but the data reportedare not sufficiently convincing to support Category 1 because of the lack of details in theobservations or in the exposure conditions, and/or with regard to the nature and the severityof the effects observed, then classification in Category 2 could be justified (CLP Annex I,3.8.2.1.6). In this case, any animal data must also be consistent with Category 2 and notsupport Category 1 (see below). In this case, if the animal data support Category 1, they willtake precedence over the human data. This is because the reliability of the human data in thiscase is probably lower than the reliability of data from standard well conducted animal studiesand should accordingly have less weight in the assessment.When using human data, there is no consideration of the human dose/exposure level thatcaused those effects.Category 3Respiratory Tract IrritationHuman evidence for RTI often comes from occupational case reports where exposure isassociated with signs of RTI. Such reports should be interpreted carefully using expertjudgement to ensure that they provide reliable information. For instance, there should be aclear relationship between exposure and the development of signs of RTI, with RTI appearingrelatively soon after the start of exposure. A solid substance which causes RTI due tophysical/mechanical irritation when inhaled as a dust should not be classified. For moredetails on RTI, see R7a.7.2.1, and example n° 3 for sulfur dioxide.Narcotic EffectsNarcotic effects may range from slight dizziness to deep unconsciousness and may be causedby several mechanisms:– pharmaceutical drugs (designed effect; often receptor-mediated; effective dose usuallylow; patient under professional observation; limited importance for industrial chemicalsand their safety assessment.)365

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012– unspecific effects of many organic industrial chemicals on CNS-membranes at high doselevels (often solvent vapours, ≥ 6000 ppm in respired air volume). Such effects can beexpected at high exposure levels due to otherwise low toxicity.– organic chemicals with similarities to and interference with CNS-transmitters; oftenmetabolic transformation necessary; certain solvents, e.g. butandiol, butyrolactone,methoxyethanol; medium levels of effective dose. Children may be considerably moresusceptible than adults.– chemicals with high specific CNS toxicity; narcotic effects usually close to near-lethaldoses (example: H 2 S).Narcotic effects are usually readily reversible on cessation of exposure with no permanentdamage or changes.Human evidence relating to narcosis should be evaluated carefully. Often the reporting ofclinical signs is relatively subjective and reports of effects such as severe headache anddizziness should be interpreted carefully to judge if they provide robust evidence of narcosis.Where relevant human data do not mirror realistic exposure conditions, for instance in casereports from accidental over-exposure situations, supportive information may be needed tocorroborate the observed effects. A single case report from accidental or deliberate exposure(i.e. abuse) is unlikely to provide sufficiently robust evidence to support classification withoutother evidence. For more details on evaluation of available human information see alsoSection 3.1.2.3.1 and IR/CSA, Section R.7.4 (especially R.7.4.4.2). Example n° 4 for tolueneillustrates the procedure.3.8.2.4.2 Evaluation of non human dataAnnex 1: 3.8.2.1.5. The standard animal studies in rats or mice that provide information are acutetoxicity studies which can include clinical observations and detailed macroscopic and microscopicexamination to enable the toxic effects on target tissues/ organs to be identified. Results of acutetoxicity studies conducted in other species may also provide relevant information.Annex 1: 3.8.2.1.10.1. When a substance is characterised only by use of animal data (typical ofnew substances, but also true for many existing substances), the classification process includesreference to dose/concentration guidance values as one of the elements that contribute to the weightof evidence approach.Annex 1: 3.8.2.1.10.3. A substance that has not been tested for specific target organ toxicity may,where appropriate, be classified on the basis of data from a validated structure activity relationshipand expert judgement-based extrapolation from a structural analogue that has previously beenclassified together with substantial support from consideration of other important factors such asformation of common significant metabolites.The type of evidence mentioned in CLP Annex I, 3.8.2.1.7 and 3.8.2.1.8 to support or not tosupport classification (e.g. clinical biochemistry, changes in organ weights with no evidenceof organ dysfunction) is rarely obtained from animal tests designed to measure acutelethality/toxicity (See section 3.8.2.1.2).Categories 1 and 2Generic guidance on data evaluation is presented in IR/CSA, Section R.7.4 and R.7.4.4.2. Allavailable animal data which are of acceptable quality should be used in a weight of evidenceapproach based on a comparison with the classification criteria described above. Theassessment should be done for each route of exposure.366

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012For each study the effects seen in each sex at or around the guidance values (GV) forCategory 1 and Category 2 should be compared with the effects warranting classification inCategory 1 and 2. In general findings in the most sensitive sex would be used to determine theclassification. If the NOAEL from the study is above the GV, the results of that study do notindicate classification for that category (situations 1 and 2 in Figure 1). If the NOAEL isbelow the GV then the effective dose (ED) level, the lowest dose inducing significant/severetarget organ toxicity as defined in Section 3.8.2.2.1 should be determined based on the criteriadescribed above. If the ED is below the GV then this study indicates that classification iswarranted (situations 2 and 4 in Figure 1).In a case where the ED is above a GV but the NOAEL is below the GV (situations 3 and 5)then interpolation between the ED and the NOAEL is required to determine whether theeffects expected at or below the GV would warrant classification .Figure 3.8.2.4.2 Comparison between the NOAEL and the ED versus the guidance valuesSituation 1 Situation 2 Situation 3 Situation 4 Situation 5GVCategory 2- NOAEL 1- ED 2- ED 3- NOAEL 3- NOAEL 2- ED 5GVCategory 1- ED 4- NOAEL 5- NOAEL 4NC Category 2 Interpolation Category 1 InterpolationWhere a number of studies are available these should be assessed using a weight of evidenceapproach to determine the most appropriate classification. Where the findings from individualstudies would lead to a different classification then the studies should be assessed in terms oftheir quality, species and strain used, nature of the tested substance (including the impurityprofile and physical form) etc to choose the most appropriate study to support classification.In general, the study giving the most severe classification will be used unless there are goodreasons that it is not the most appropriate. If the effects observed in animals are notconsidered relevant for humans then these should not be used to support classification.Similarly, if there is robust evidence that humans differ in sensitivity or susceptibility to theeffect observed in the study then this should be taken into account, possibly leading to anincrease or decrease in the classification assigned. The final classification based on nonhuman data will be the most severe classification of the three exposure routes.Category 3367

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012There are no similar guidance values for Category 3. Therefore, if the study shows clearevidence for narcotic effects or respiratory tract irritation at any dose level then this couldsupport classification with Category 3.In evaluating inhalation studies a differentiation of respiratory tract effects and systemiceffects should always be attempted. In addition, the region in the respiratory tract and thequalitative nature of observed effects is pivotal. Often, the lesions observed are representingstages of a reaction pattern leading to severe and irreversible functional and structuralalterations. Therefore reversibility of effects is a significant discriminator. For further detailssee also Section 3.8.2.3.3.8.2.4.3 Evaluation of non-testing and in vitro dataNon-testing and in vitro data can contribute to the weight of evidence supporting aclassification. As described in Annex XI of REACH approaches such as (Q)SAR, groupingand read-across can provide information on the hazardous properties of substances in place oftesting and can be used for classification purposes. Also see R7.4.4.1.3.8.2.4.4 ConversionsThe guidance values are given in mg/kg bodyweight. Where the doses in a study are given indifferent units they will need to be converted as appropriate. For instance the dosages infeeding and drinking water studies are often expressed in ppm, mg test substance/ kg (feed) ormg (test substance)/l (drinking water).The conversion from mg/l to ppm assuming an ambient pressure of 1 at 101.3 kPa and 25°C isppm = 0.0245 mg/l 1/MW.3.8.2.4.5 Weight of evidenceAnnex 1: 3.8.2.1.6. In exceptional cases, based on expert judgement, it is appropriate to placecertain substances with human evidence of target organ toxicity in Category 2:1) when the weight of evidence is not sufficiently convincing to warrant Category 1classification, and/or2) based on the nature and severity of effects.Dose/concentration levels in humans shall not be considered in the classification and any availableevidence from animal studies shall be consistent with the Category 2 classification. In other words,if there are also animal data available on the substance that warrant Category 1 classification, thesubstance shall be classified as Category 1.The available information should be considered using expert judgement and a weight ofevidence assessment, as described in CLP Annex I, 1.1.1 and Module 1.Valid human data generally take precedence over animal and other non-test data. If there arehuman data indicating no classification but there are also non-human data indicatingclassification then the classification is based on the non-human data unless it is shown that thehuman data cover the exposure range of the non-human data or that the non-human data arenot relevant for humans. If the human and non-human data both indicate no classification thenclassification is not required.If there are no human data then the classification is based on the non-human data.368

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.2.5 Decision on classification of substancesDecision on classification for STOT-SE is based on the results of weight of evidenceapproach described in 2.3.STOT-SE and acute toxicity are independent of each other and both may be assigned to asubstance if the respective criteria are met. However, care should be taken not to assign eachclass for the same effect, in other words a double classification for the same effect has to beavoided. STOT-SE will be considered where there is clear evidence for a specific organtoxicity especially in absence of lethality, see examples no 1 and no 3 (methanol andtricresylphosphate).If no classification has been warranted for acute toxicity despite significant toxic effect, thesubstance should be considered for classification as STOT-SE.Normally, the assignment of STOT-SE Category 1 or 2 is independent to the assignment ofCategory 3. Therefore, a substance may be classified in both Category 1/2 and Category 3 ifthe respective criteria are met, for instance, in the case of a neurotoxic substance that alsocauses transient narcotic effects. If Category 1/2 is assigned on the basis of effects in therespiratory tract then Category 3 should not be assigned as this would provide no additionalinformation.Classification as acutely toxic and/or corrosive is considered to cover and communicate thespecific toxicological effect(s) adequately. An additional classification as specific target organtoxicant (single exposure, Category 1 or 2) is not indicated if the severe toxicological effect isthe consequence of the local (i.e. corrosive) mode of action.It is a reasonable assumption that corrosive substances may also cause respiratory tractirritation when inhaled at exposure concentrations below those causing frank respiratory tractcorrosion. If there is evidence from animal studies or from human experience to support thisthen Category 3 may be appropriate. In general, a classification for corrosivity is consideredto implicitly cover the potential to cause RTI and so the additional Category 3 is considered tobe superfluous, although it can be assigned at the discretion of the classifier. The Category 3classification would occur only when more severe effects in the respiratory system are notobserved.Category 3 effects should be confined to changes, whether functional or morphological,occurring in the upper respiratory tract (nasal passages, pharynx and larynx). Localizedirritation with associated adaptive responses (e.g., inflammation, epithelial metaplasia, gobletcell hyperplasia, proliferative effects) may occur and are consistent with Category 3responses. Injury of the olfactory epithelium should be distinguished in terms of irritationrelated(non-specific) and metabolic/ non-irritant (specific).3.8.2.6 Setting of specific concentration limits for STOT-SEArticle 10(1) Specific concentration limits and generic concentration limits are limits assigned to asubstance indicating a threshold at or above which the presence of that substance in anothersubstance or in a mixture as an identified impurity, additive or individual constituent leads to theclassification of the substance or mixture as hazardous.Specific concentration limits shall be set by the manufacturer, importer or downstream user whereadequate and reliable scientific information shows that the hazard of a substance is evident when thesubstance is present at a level below the concentrations set for any hazard class in Part 2 of Annex Ior below the generic concentration limits set for any hazard class in Parts 3, 4 and 5 of Annex I.369

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In exceptional circumstances specific concentration limits may be set by the manufacturer, importeror downstream user where he has adequate, reliable and conclusive scientific information that ahazard of a substance classified as hazardous is not evident at a level above the concentrations setfor the relevant hazard class in Part 2 of Annex I or above the generic concentration limits set for therelevant hazard class in Parts 3, 4 and 5 of that Annex.Specific concentration limits (SCLs) for STOT-SE may be set by the supplier in somesituations according to Article 10 of CLP. For STOT-SE, this may only be done forsubstances inducing STOT-SE Category 1 at a dose level or concentration clearly (more thanone magnitude) below the guidance values according to Table 3.8.2, e.g. below 30 mg/kgbodyweight from the oral single exposure study. This will be mainly based on data inexperimental animals but can also be based on human data if reliable exposure data areavailable. The SCL (SCL Cat. 1) for a Category 1 substance triggering classification of amixture in Category 1 can be determined using the following formula:EDSCLCat . 1 100%Equation 3.8.2.6(a)GV 1SCL Cat 1: 0.7 mg/kgbw/300 mg/kgbw x 100%=0.22% --> 0.2%In this formula the ED is the dose of the Category 1 substance inducing significant specifictarget organ toxicity and GV1 is the guidance value for Category 1 according to Table 3.8.2of Annex I. The resulting SCL is rounded down to the nearest preferred value 49 (1, 2 or 5).Example of determining STOT-SE SCL for a Category 1 substance:0.7mg/ kgbw100%300mg/ kgbw= 0.22% --> 0.2%Though classification of a mixture in Category 1 is not triggered if a Category 1 constituent ispresent in lower concentrations than the established SCL, a classification in Category 2should be considered.The SCL (SCL Cat. 2)for a Category 1 substance triggering classification of a mixture inCategory 2 can be determined using the following formula:EDSCLCat . 2 100%Equation 3.8.2.6(b)GV 2In this formula the ED is the dose of the Category 1 substance inducing specific target organtoxicity and GV2 is the upper guidance value for Category 2 according to Table 3.8.2 ofAnnex I. The resulting SCL is rounded down to the nearest preferred values (1, 2 or 5).However, if the calculated SCL for classification in Category 2 is above 1%, which is theGeneric Concentration Limit, then no SCL should be set.Example for a substance in SCL Category 2:0.7mg/ kgbw100%2000mg/ kgbw= 0.035 --> 0.02% (rounded down)49 This is the “preferred value approach” as used in EU and are values to be established preferentially as thenumerical values 1.2 or 5 or multiples by powers of ten.370

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012For example, a Category 1substance inducing specific target organ toxicity at 0.7 mg/kgbw/day in an acute oral study would generate an SCL for classification of mixtures inCategory 1 at 0.2% and in Category 2 at 0.02% (Cat1: C ≥ 0.2% ; Cat 2: 0.02% ≤ C < 0.2%).It is not appropriate to determine SCLs for substances classified in Category 2 sinceingredients with a higher potency (i.e. lower effect doses than the lower guidance values ofCategory 2) will be classified in Category 1; substances with higher effect doses than theupper guidance value of Cat2 will generally not be classified.Classification in STOT-SE Category 3 for RTI and narcotic effects does not take potency intoaccount and consequently does not have any guidance values. A pragmatic default GCL of20% is suggested, although a lower or higher SCL may be used where it can be justified.Therefore, an SCL can be determined on a case-by-case basis for substances classified asSTOT-SE Category 3 and expert judgement shall be exercised.Specific concentration limits for each of the hazard classes skin and eye irritation, and STOT-SE Category 3 for respiratory tract irritation need to be addressed separately, while unjustifiedread-across of SCLs from one hazard class to another is not acceptable.For narcotic effects, the factors to be taken into consideration in order to set lower or higherSCLs are the effective dose/concentration, and in addition for liquids, the volatility (saturatedvapour concentration) of the substance.371

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.2.7 Decision logicThe decision logic is provided as additional guidance. It is strongly recommended that theperson responsible for classification study the criteria for classification before and during useof the decision logic.This decision logic deviates slightly from the original UNGHS in separating the connectionbetween Category 2 and Category 3, since, different from the procedure in other hazardclasses, they have to be regarded as independent.Classification in Category 1 and Category 2Does the substance have data and/or information to evaluatespecific target organ toxicity following single exposure?NOClassificationnot possibleYESFollowing single exposure,(a) Can the substance produce significant toxicity in humans,or(b) Can it be presumed to have the potential to producesignificant toxicity in humans on the basis of evidencefrom studies in experimental animals?See CLP Annex I, 3.7.3 for criteria and guidance values.Application of the criteria needs expert judgment in a weight ofevidence approach.NOYESCategory 1DangerFollowing single exposure,Can the substance be presumed to have the potential to beharmful to human health on the basis of evidence from studies inexperimental animals?See CLP Annex I, 3.7.3 for criteria and guidance values.Application of the criteria needs expert judgment in a weight ofevidence approach.NOYESCategory 2WarningNot classified372

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Classification in Category 3Does the substance have data and/or information to evaluatespecific target organ toxicity following single exposure withrelevance for RTI or narcotic effects?NOClassificationnot possibleYESFollowing single exposure,Can the substance produce respiratory tract irritation ornarcotic effects?See CLP Annex I, 3.7.3 for criteria and guidance values.Application of the criteria needs expert judgment in a weightof evidence approach.YESCategory 3WarningNONot classified373

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.3 Classification of mixtures for STOT-SE3.8.3.1 Identification of hazard informationWhere toxicological information is available on a mixture this should be used to derive theappropriate classification. Such information may be available from the mixture manufacturer.Where such information on the mixture itself is not available information on similar mixturesand/or the component substances in the mixture must be used, as described below.3.8.3.2 Classification criteria for mixturesAnnex 1: 3.8.3.1. Mixtures are classified using the same criteria as for substances, or alternativelyas described below.3.8.3.2.1 When data are available for the complete mixtureAnnex 1: 3.8.3.2.1. When reliable and good quality evidence from human experience orappropriate studies in experimental animals, as described in the criteria for substances, is availablefor the mixture, then the mixture shall be classified by weight of evidence evaluation of these data(see 1.1.1.3). Care shall be exercised in evaluating data on mixtures, that the dose, duration,observation or analysis, do not render the results inconclusiveIn cases where test data for mixtures are available, the classification process is exactly thesame as for substances.3.8.3.2.2 When data are not available for the complete mixture: bridging principlesAnnex 1: 3.8.3.3.1. Where the mixture itself has not been tested to determine its specific targetorgan toxicity, but there are sufficient data on the individual ingredients and similar tested mixturestoadequately characterise the hazards of the mixture, these data shall be used in accordance with thebridging principles set out in section 1.1.3.When there are no test data on the mixture as a whole, so called “Bridging principles” may beapplied where there are data available on similar tested mixtures and on the individualhazardous ingredient substances within the mixture that are sufficient to adequately assess thehazards of the mixture.3.8.3.2.3 When data are available for all components or only for some components ofthe mixtureAnnex 1: 3.8.3.4.1. Where there is no reliable evidence or test data for the specific mixture itself,and the bridging principles cannot be used to enable classification, then classification of the mixtureis based on the classification of the ingredient substances. In this case, the mixture shall beclassified as a specific target organ toxicant (specific organ specified), following single exposure,when at least one ingredient has been classified as a Category 1 or Category 2 specific target organtoxicant and is present at or above the appropriate generic concentration limit as mentioned inTable 3.8.3 below for Category 1 and 2 respectively.A mixture not classified as corrosive but containing a corrosive ingredient should beconsidered for classification in Category 3 RTI on a case-by-case basis following theapproach explained above (see Section 3.8.2.3). More information on classification ofmixtures into Category 3 is provided below (Section 3.8.3.3)374

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.3.2.4 Components of a mixture that should be taken into account for the purposeof classificationComponents with a concentration equal to or greater than the generic concentration limits(1% for Category 1 components and 10% for Category 2. See Table 3.8.3) or with a SpecificConcentration Limit (see Section 3.8.2.6) will be taken into account for classificationpurposes. For Category 3, the GCL is 20%. Specific concentration limits have preference overthe generic ones.3.8.3.3 Generic concentration limits for substances triggeringclassification of mixtures for STOT-SEThe STOT-SE hazard class does not foresee summation of Category 1 or 2 substances in theclassification process of a mixture. Furthermore, as Category 1 and 2 depict different hazardsthan Category 3 the assessment must be done independently from each other.Annex 1: Table 3.8.3Generic concentration limits of ingredients of a mixture classified as a specific target organ toxicantthat trigger classification of the mixture as Category 1 or 2Ingredient classified as:Category 1Specific Target OrganToxicantCategory 2Specific Target OrganToxicantNote 1:Generic concentration limits triggering classificationof the mixture as :Category 1 Category 2Concentration 10% 1.0% concentration 10%Concentration 10%[(Note 1)]If a Category 2 specific target organ toxicant is present in the mixture as an ingredient at aconcentration ≥ 1.0% a SDS shall be available for the mixture upon request.3.8.3.4.4. Care shall be exercised when toxicants affecting more than one organ system arecombined that the potentiation or synergistic interactions are considered, because certain substancescan cause target organ toxicity at < 1% concentration when other ingredients in the mixture areknown to potentiate its toxic effect.3.8.3.4.5. Care shall be exercised when extrapolating toxicity of a mixture that contains Category 3ingredient(s). A generic concentration limit of 20% is appropriate; however, it shall be recognisedthat this concentration limit may be higher or lower depending on the Category 3 ingredient(s) andthat some effects such as respiratory tract irritation may not occur below a certain concentrationwhile other effects such as narcotic effects may occur below this 20% value. Expert judgementshall be exercised.Categories 1 and 2Each single classified component in a concentration range given in Table 3.8.3 triggers theclassification of the mixture, i.e. additivity of the concentrations of the components is notapplicable.Category 3375

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012When a mixture contains a number of substances classified with Category 3 and present at aconcentration below the GCL (i.e. 20%), an additive approach to determine the classificationof the mixture as a whole may be appropriate. In the additive approach the concentrations ofthe individual substances with the same hazard (i.e. RTI or narcotic effects) are totalledseparately. If each individual total is greater than the GCL then the mixture should beclassified as Category 3 for that hazard. A mixture may be classified either as STOT SE 3(RTI) or STOT SE 3 (narcotic effects) or both.ExampleThe following example shows whether or not additivity should be considered for SpecificTarget Organ Toxicity – Single Exposure (STOT-SE) Category 3 transient effects.Ingredient information:Ingredient Wt% ClassificationIngredient 1 0.5 -Ingredient 2 3.5 Category 3 – Respiratory Tract IrritationIngredient 3 15 Category 3 - Narcotic effectsIngredient 4 15 Category 3 - Narcotic effectsIngredient 5 66 -376Answer:Mixture is Category 3 – Narcotic effects∑%Category 3 – Narcotic effects = 15% + 15% = 30% which is > 20%%, thereforeclassify as Category 3 – Narcotic Effects∑%Category 3 – Respiratory Irritation = 3.5%, which is < 20%, not classified forRespiratory IrritationRationale:(a) Classification via application of substance criteria is not possible since test data wasnot provided for the mixture (paragraph 3.8.3.2);(b) Classification via the application of bridging principles is not possible since data ona similar mixture was not provided (paragraph 3.8.3.3.1);(c) Application of paragraph 3.8.3.4.5 is used for classification. Expert judgement isnecessary when applying this paragraph. Paragraph 3.8.3.4.5 notes that a cut-offvalue/concentration limit of 20% has been suggested, but that the cut-offvalue/concentration limit at which effects occur may be higher or less depending onthe Category 3 ingredient(s). In this case, the classifiers judged that 30% is sufficientto classify.SCLsIn the case where a specific concentration limit has been established for one or moreingredients these SCLs have precedence over the generic concentration limit.3.8.3.4 Decision logic for mixturesA mixture should be classified either in Category 1 or in Category 2, according to the criteriadescribed above. The corresponding hazard statement (H370 for Category 1 or H371 for

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Category 2) should be used without specifying the target organs, except if the classification ofthe mixture is based on data available for the complete mixture, in which case the targetorgans may be given. In the same way, the route of exposure should not be specified, except ifdata are available for the complete mixture and it is conclusively demonstrated that no otherroutes of exposure cause the hazard.If the criteria are fulfilled to classify also the mixture in Category 3 for respiratory irritation ornarcotic effects, only the corresponding hazard statement (H335 and/or H336) will be addedin hazard communication.The decision logic is provided as additional guidance. It is strongly recommended that theperson responsible for classification study the criteria for classification before and during useof the decision logic.This decision logic deviates slightly from the original UNGHS in separating the connectionbetween Category 2 and Category 3, since different from the procedure in other hazard classesthey have to be regarded as independent.377

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Classification in Category 1 or 2Does the mixture as a whole have data/information toevaluate specific target organ toxicity following singleexposure?NOYESSee decisionlogics forsubstancesCan bridging principles be applied?NOYESClassify inappropriatecategoryDoes the mixture contain one or more ingredients classifiedas a Category 1 specific target organ toxicant at aconcentration of 10%?YESCategory 1NODangerDoes the mixture contain one or more ingredientsclassified as a Category 1 specific target organ toxicant ata concentration of 1.0 and < 10%?Or one or more ingredients classified as a Category 2specific target organ toxicant at a concentration of : 10%?YESCategory 2NOWarningNot classified378

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Classification in Category 3Does the mixture as a whole have data and/or informationto evaluate specific target organ toxicity following singleexposure with relevance for RTI or narcotic effects?YESSee decisionlogics forsubstancesNOCan bridging principles be applied?YESClassifyappropriatecategoryinNOCategory 3Does the mixture contain one or more ingredientsclassified as a Category 3 specific target organ toxicantat a concentration 20%?YESNOWarningNot classified379

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.4 Hazard communication in form of labelling for STOT-SE3.8.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 3.8.4.1. Label elements shall be used in accordance with Table 3.8.4., for substances ormixtures meeting the criteria for classification in this hazard class.Table 3.8.4Label elements for specific target organ toxicity after single exposureClassification Category 1 Category 2 Category 3GHS PictogramsSignal word Danger Warning WarningHazard statementPrecautionarystatement PreventionPrecautionaryStatement ResponsePrecautionaryStatement StoragePrecautionaryStatement DisposalH370: Causes damageto organs (or state allorgans affected, ifknown) (state route ofexposure if it isconclusively proventhat no other routes ofexposure cause thehazard)P260P264P270P307 + P311P321H371: May causedamage to organs (orstate all organsaffected, if known)(state route ofexposure if it isconclusively proventhat no other routes ofexposure cause thehazard)P260P264P270P309 + P311H335: May causerespiratory irritation;In the same way, the route of exposure should not be specified, except if data are available forthe complete mixture and if it is conclusively demonstrated that no other routes of exposurecause the hazard. It is recommended to include no more than three primary target organs forpractical reasons and because the classification is for specific target organ toxicity. If more380orH336: May causedrowsiness anddizzinessP261P271P304 + P340P312P405 P405 P403 + P233P405P501 P501 P501The hazard statement should include the primary target organ(s) of toxicity. Organs in whichsecondary effects were observed should not be included. The route of exposure should not bespecified, except if it is conclusively demonstrated that no other routes of exposure cause thehazard. When a mixture is classified for STOT-SE on basis of test data, the hazard statementwill specify the target organs, in the same way as for a substance. If a mixture is classified onbasis of the ingredients, the hazard statement (H370 for Category 1 or H371 for Category 2)may be used without specifying the target organs, as appropriate.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012target organs are effected it is recommended that the overall systemic damage should bereflected by using the phrase “damage to organs”.AnnexI: 3.8.2.1.10.43.8.4.2 Additional labelling provisionsSaturated vapour concentration shall be considered, where appropriate, as an additional element toprovide for specific health and safety protection.According to CLP Annex I, 3.8.2.1.10.4 the saturated vapour concentration shall beconsidered as an additional element for providing specific health and safety protection. Thusif a classified substance is highly volatile a supplementary precautionary advice (e.g.“Special/additional care should be taken due to the high saturated vapour pressure”) might begiven in order to emphasize the hazard in case it is not already covered by the generalPrecuationary statements. (As a rule substances for which the ratio of the effectconcentration at

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012If translation results in a classification in STOT SE 1 for one route and in STOT SE 2 foranother route only classification in Category 1 is required (for both routes).Classification as STOT SE is not route specific as it was for classification with R39/X andR68/X. The route specificity of STOT SE is included in the hazard statement and includesroute-to-route extrapolation by default unless conclusively shown otherwise. Therefore, theroute specific data on STOT SE should be re-evaluated. A re-evaluation is also necessarybecause the primary target organs for STOT SE should be stated in the hazard statement.All substances or mixtures classified with R67 shall be classified as STOT SE Category 3H336.All substances or mixtures classified with R37 shall be classified as STOT SE Category 3H335. Also additional labelling with EUH071 (Corrosive to the respiratory tract) shall beconsidered.3.8.5.2 Re-evaluation of the STOT-SE dataGasses classified with R39/23 or R39/26 should be re-evaluated because of the change fromguidance values in mg/L into ppm.Substances or mixtures not classified for STOT-SE, should be considered for re-evaluationbecause less adverse effects and higher guidance values are required for classificationaccording to CLP compared to DSD. Also, effects in humans are now considered forclassification without restrictions to the exposure level.382

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.6 Examples of classification for STOT-SE3.8.6.1 Examples of substances fulfilling the criteria forclassification3.8.6.1.1 Example 1: MethanolApplicationAvailableinformationRemarksUse of adequate and reliable human data, where animal data are not appropriate.Independent classification for STOT-SE and Acute toxicity due to different effectsTest Data Classification RationaleAnimal data:LD 50 rat > 5,000 (mg/kg)No specific target organtoxicity (impairment ofseeing ability) observed inrats, even in high doses.Human experience:Broad human experiencefrom many case reports aboutblindness following oralintake. Methanol is known tocause lethal intoxications inhumans (mostly viaingestion) in relatively lowdoses: ” …minimal lethaldose in the absence ofmedical treatment is between300 and 1000 mg/kg” (IPCS)Classification notpossibleSTOT-SECategory 1The rat is known to beinsensitive to the toxicity ofmethanol and is thus notconsidered to be a good modelfor human effects (differenteffect/mode of action)The classification criteria forCategory 1 are fulfilled: clearhuman evidence of a specifictarget organ toxicity effectwhich is not covered by Acutetoxicity.The standard animal species for single exposure (acute) tests, the rat, is notsensitive, i.e. no appropriate species for this specific target organ effect.Methanol is classified independently for acute toxicity, since the impairmentof vision is not causal for the lethality, i. e. there are different effects.Labelling:Pictogram GHS 08; Signal word: Danger; Hazard statement: H370 Causesdamage to the eye.383

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.6.1.2 Example 2: Tricresyl phosphateApplicationAvailableinformationRemarksUse of valid human evidence supported by animal dataTest Data Classification RationaleHuman experience:There are well documentedcase reports about severeneurotoxic effectsAnimal experiments:Severe neurotoxic effects(Paralysis) were observedafter single exposure of doses< 200 mg/kgLD 50 rat oral 3000 - 3900mg/kgSTOT-SECategory 1The classification criteria areclearly fulfilled based onhuman experience as well as onresults of animal studiesLabelling:Pictogram GHS 08; Signal word: Danger; Hazard Statement: H370 Causes damageto the central nervous system.3.8.6.1.3 Example 3: Sulfur dioxideApplicationAvailableinformationRemarksUse of valid human evidenceTest Data Classification RationaleHuman experience:Broad, well documentedhuman experience onirritating effect to respiratorysystem.STOT-SECategory 3The classification criteria forCategory 3 (Respiratory TractIrritation) are fulfilled based onwell documented experience inhumansLabelling:Pictogram GHS 07; Signal word: Warning; Hazard statement: H335 May causerespiratory irritation3.8.6.1.4 Example 4: TolueneApplicationAvailableinformationRemarksTest Data Classification RationaleAnimal data:In valid animal experimentsnarcotic effects (transienteffect on nervous system) at>= 8 mg/l were observed.Labelling:STOT-SECategory 3The classification criteria forCategory 3 (Narcotic Effects)are fulfilled based on welldocumented result s in animalexperimentsPictogram GHS 07; Signal word: Warning; Hazard statement: H336 May causedrowsiness and dizziness384

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.8.6.2.1 Example 5: ABCApplicationAvailableinformation3.8.6.2 Examples of substances not fulfilling the criteria forclassificationSE in case same effect leading to Acute toxicity classificationTest Data Classification RationaleAnimal data:In a study in rats after singleexposure at 2,000mg/kgsevere damage in liver(macroscopic examination)and mortality in 6/10animals were observed3.8.6.2.2 Example 6: N,N-DimethylanilineApplicationAvailableinformationRemarksNo classificationin STOT- SEThough a specific organ isdamaged, the substance will beclassified in Acute Toxicity(Category 4), since lethalitywas observed which was due tothe liver impairment. It isassumed that the LD 50 =ATE is≤ 2,000 mg/kg. There should beno double classification for thesame effect/mechanism causinglethality by impairment of aspecific organ, thus noclassification for STOT-SENo classification for STOT-SE in case same effect leading to Acute toxicityclassificationTest Data Classification RationaleAnimal data:Acute oral toxicity: LD 50values > 1,120-1,300 oral ratand 1,690 mg/kg bw dermalrabbit; ca. 50 mg/kg are lethalin cats due to high Met HBformation ; no specific targetorgan toxicity (blood toxicity)observed in rats.Human experience:Broad human experience frommany case reports about lethalintoxications caused bymethemoglobinemiafollowingoral/dermal/inhalationexposure to aromatic aminesNo classificationin STOT-SENo classificationin STOT-SEThe criteria for STOT-SEclassification are not fulfilleddespite a clear specific targetorgan effect in humans and ina relevant animal species. Thesubstance is classified inCategory 3 Acute Toxicitysince the Met HB formation iscausative for the lethality inhumans and in animals (cats)in low doses.The standard animal species for single exposure (acute) tests, the rat, is not sensitive,i.e. no appropriate species for this specific effect.385

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9 SPECIFIC TARGET ORGAN TOXICITY – REPEATED EXPOSURE(STOT-RE)3.9.1 Definitions and general considerations for STOT-REAnnex I: 3.9.1.1. Specific target organ toxicity (repeated exposure) means specific, target organtoxicity arising from a repeated exposure to a substance or mixture. All significant health effectsthat can impair function, both reversible and irreversible, immediate and/or delayed are included.However, other specific toxic effects that are specifically addressed in Chapters 3.1 to 3.8 andChapter 3.10 are not included here.According to CLP Annex I, 3.9.1.1, specific toxic effects covered by other hazard classes arenot included in STOT-RE. STOT-RE should only be assigned where the observed toxicity isnot covered more appropriately by another hazard class. For example specific effects liketumours or effects on the reproductive organs should be used for classification forcarcinogenicity or reproductive toxicity, respectively, but not for STOT-RE.Annex I: 3.9.1.3. These adverse health effects include consistent and identifiable toxic effects inhumans, or, in experimental animals, toxicologically significant changes which have affected thefunction or morphology of a tissue/organ, or have produced serious changes to the biochemistry orhaematology of the organism and these changes are relevant for human health.3.9.1.4. Assessment shall take into consideration not only significant changes in a single organ orbiological system but also generalised changes of a less severe nature involving several organs.3.9.1.5. Specific target organ toxicity can occur by any route that is relevant for humans, i.e.principally oral, dermal or inhalation.Annex I: 3.9.2.2. The relevant route or routes of exposure by which the classified substanceproduces damage shall be identified.The purpose of STOT-RE is to identify the primary target organ(s) of toxicity (CLP Annex I,3.9.1.4) for inclusion in the hazard statement. Where possible secondary effects are observedin other organs, they should be carefully considered for the classification. The STOT-REclassification should identify those routes by which the substance causes the target organtoxicity (CLP Annex I, 3.9.1.5 and 3.9.2.2). This is usually based on the available evidencefor each route. There are no compelling reasons to do route to route extrapolation to attemptto assess the toxicity by other routes of exposure for which there are no data.Annex I: 3.9.1.6. Non-lethal toxic effects observed after a single-event exposure are classified asdescribed in Specific target organ toxicity — Single exposure (section 3.8) and are thereforeexcluded from section 3.9.Where the same target organ toxicity of similar severity is observed after single and repeatedexposure to a similar dose, it may be concluded that the toxicity is essentially an acute (i.e.single exposure) effect with no accumulation or exacerbation of the toxicity with repeatedexposure. In such a case classification with STOT-SE only would be appropriate.386

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9.2 Classification of substances for STOT-RE3.9.2.1 Identification of hazard informationAnnex 1: 3.9.2.5. The information required to evaluate specific target organ toxicity comes eitherfrom repeated exposure in humans, such as exposure at home, in the workplace or environmentally,or from studies conducted in experimental animals.CLP does not require testing of substances and mixtures for classification purposes. Theassessment is based on the respective criteria and consideration of all available adequate andreliable information, primarily such relating to repeated-dose exposures but also taking intoaccount the general physico-chemical nature of the substance. The most useful information isgenerally from human epidemiology, case studies and animal studies, but informationobtained using read-across from similar substances and from appropriate in vitro models canalso be used, where appropriate.3.9.2.1.1 Identification of human dataRelevant information with respect to repeated dose toxicity may be available from casereports, epidemiological studies, medical surveillance and reporting schemes, and nationalpoisons centres.Details are given in IR/CSA, Section 7.5.3.2.3.9.2.1.2 Identification of non human dataAnnex 1: 3.9.2.5. …. The standard animal studies in rats or mice that provide this information are28 day, 90 day or lifetime studies (up to 2 years) that include haematological, clinicochemical anddetailed macroscopic and microscopic examination to enable the toxic effects on targettissues/organs to be identified. Data from repeat dose studies performed in other species shall alsobe used, if available. Other long-term exposure studies, such as on carcinogenicity, neurotoxicity orreproductive toxicity, may also provide evidence of specific target organ toxicity that could be usedin the assessment of classification.Non-testing dataPhysico-chemical dataPhysicochemical properties, such as pH, physical form, solubility, vapour pressure, andparticle size, can be important parameters in evaluating toxicity studies and in determining themost appropriate classification especially with respect to inhalation where physical form andparticle size can have a significant impact on toxicity.(Q)SAR modelsStructurally or mechanistically related substance(s), read-across/grouping/chemical categoryand metabolic pathway approach: A (Q)SAR analysis for a substance may give indications fora specific mechanism of action and identify possible organ or systemic toxicity upon repeatedexposure. Overall, (Q)SAR approaches are currently not well validated for repeated dosetoxicity. (IR/CSA, Section R7.5.4.1). Data on structurally analogous substances may beavailable and add to the toxicity profile of the substance under investigation. The concept ofgrouping, including both read-across and the related chemical category concept have beendeveloped under the OECD HPV program. For certain substances without test data theformation of common significant metabolites or information with those of tested substancesor information from precursors may be valuable information. (For more details see IR/CSA,387

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Sections R.6.1 and R.6.2.5.2 and OECD (2004)). OECD Principles for the Validation, forRegulatory Purposes, of (Quantitative) Structure-Activity Relationship Models)Testing dataAnimal data”The most appropriate data on repeated dose toxicity for use in hazard characterisation andrisk assessment are primarily obtained from studies in experimental animals conforming tointernationally agreed test guidelines. In some circumstances repeated dose toxicity studiesnot conforming to conventional test guidelines may also provide relevant information for thisendpoint” (IR/CSA, Section R.7.5.3.1). Studies not performed according to Standard TestGuidelines and/or GLP have to be evaluated on case by case basis by expert judgement and inthe context of a total weight of evidence assessment if there are more data (for moreinformation see Section 3.9.2.3.4 and IR/CSA, Section R.7.5.4.1.The standard test guidelines are described in IR/CSA, Section R.7.5.4.1. There may also bestudies employing different species and routes of exposure. In addition, special toxicitystudies investigating further the nature, mechanism and/or dose relationship of a critical effectin a target organ or tissue may also have been performed for some substances. Other studiesproviding information on repeated dose toxicity: although not aiming at investigating repeateddose toxicity per se and other available EU/OECD test guideline studies involving repeatedexposure of experimental animals may provide useful information on repeated dose toxicity,e.g reproduction toxicity or carcinogenicity studies. For more details see IR/CSA, Section R.7.5.4.1 (ECHA, 2008).In vitro dataAt present available in vitro data is not useful on its own for regulatory decisions such asclassification and labelling. However, such data may be helpful in the assessment of repeateddose toxicity, for instance to detect local target organ effects and/or to clarify the mechanismsof action. Since, at present, there are no validated and regulatory accepted in vitro methods,the quality of each of these studies and the adequacy of the data provided should be carefullyevaluated” (IR/CSA, Section R.7.5.4.1).3.9.2.2 Classification criteria for substancesAnnex 1: 3.9.2.1. Substances are classified as specific target organ toxicants following repeatedexposure by the use of expert judgement (see 1.1.1), on the basis of the weight of all evidenceavailable, including the use of recommended guidance values which take into account the durationof exposure and the dose/concentration which produced the effect(s), (see 3.9.2.9), and are placedin one of two categories, depending upon the nature and severity of the effect(s) observed (Table3.9.1).Table 3.9.1Categories for specific target organ toxicity-repeated exposureCategoriesCategory 1CriteriaSubstances that have produced significant toxicity in humans or that, on the basis ofevidence from studies in experimental animals, can be presumed to have thepotential to produce significant toxicity in humans following repeated exposure.Substances are classified in Category 1 for target organ toxicity (repeat exposure) onthe basis of:reliable and good quality evidence from human cases or epidemiological studies; orobservations from appropriate studies in experimental animals in which significant388

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012and/or severe toxic effects, of relevance to human health, were produced at generallylow exposure concentrations. Guidance dose/concentration values are providedbelow (see 3.9.2.9), to be used as part of a weight-of- evidence evaluation.Category 2Substances that, on the basis of evidence from studies in experimental animals canbe presumed to have the potential to be harmful to human health following repeatedexposure. Substances are classified in category 2 for target organ toxicity (repeatexposure) on the basis of observations from appropriate studies in experimentalanimals in which significant toxic effects, of relevance to human health, wereproduced at generally moderate exposure concentrations. Guidancedose/concentration values are provided below (see 3.9.2.9) in order to help inclassification.In exceptional cases human evidence can also be used to place a substance inCategory 2 (see 3.9.2.6).NoteAttempts shall be made to determine the primary target organ of toxicity and classify for thatpurpose, such as hepatotoxicants, neurotoxicants. One shall carefully evaluate the data and, wherepossible, not include secondary effects (a hepatotoxicant can produce secondary effects in thenervous or gastro-intestinal systems).In the Note above "classify" would mean to identify the primary target organ.STOT-RE is assigned on the basis of findings of “significant” or “severe” toxicity. In thiscontext “significant” means changes which clearly indicate functional disturbance ormorphological changes which are toxicologically relevant. “Severe” effects are generallymore profound or serious than “significant” effects and are of a considerably adverse naturewhich significantly impact on health. Both factors have to be evaluated by weight of evidenceand expert judgement.Annex 1: 3.9.2.9.4. The decision to classify at all can be influenced by reference to thedose/concentration guidance values at or below which a significant toxic effect has been observed.Annex 1: 3.9.2.9.6. Thus classification in Category 1 is applicable, when significant toxic effectsobserved in a 90-day repeated-dose study conducted in experimental animals are seen to occur at orbelow the guidance values (C) as indicated in Table 3.9.2 below:Table 3.9.2Guidance values to assist in Category 1 classificationRoute of exposure Units Guidance values(dose/concentration)Oral (rat) mg/kg body weight/day C ≤ 10Dermal (rat or rabbit) mg/kg body weight/day C ≤ 20Inhalation (rat) gas ppmV/6h/day C ≤ 50Inhalation (rat) vapour mg/litre/6h/day C ≤ 0,2Inhalation (rat)dust/mist/fumemg/litre/6h/day C ≤ 0,02389

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex 3.9.2.9.7. Classification in Category 2 is applicable, when significant toxic effects observedin a 90-day repeated-dose study conducted in experimental animals are seen to occur within theguidance value ranges as indicated in Table 3.9.3 below:Table 3.9.3Guidance values to assist in Category 2 classificationRoute of Exposure Units Guidance Value Ranges:(dose/concentration)Oral (rat) mg/kg body weight/day 10 < C ≤ 100Dermal (rat or rabbit) mg/kg body weight/day 20 < C ≤ 200Inhalation (rat) gas ppmV/6h/day 50 < C ≤ 250Inhalation (rat) vapour mg/litre/6h/day 0,2 < C ≤ 1,0Inhalation (rat) dust/mist/fume mg/litre/6h/day 0,02 < C ≤ 0,2Annex 1 3.9.2.9.8. The guidance values and ranges mentioned in paragraphs 3.9.2.9.6 and 3.9.2.9.7are intended only for guidance purposes, i.e., to be used as part of the weight of evidence approach,and to assist with decisions about classification. They are not intended as strict demarcation values.Annex 1 3.9.2.9.5.The guidance values refer to effects seen in a standard 90-day toxicity studyconducted in rats. They can be used as a basis to extrapolate equivalent guidance values for toxicitystudies of greater or lesser duration, using dose/exposure time extrapolation similar to Haber’s rulefor inhalation, which states essentially that the effective dose is directly proportional to theexposure concentration and the duration of exposure. The assessment shall be done on a case-bycasebasis; for a 28-day study the guidance values below is increased by a factor of three.Haber’s rule is used to adjust the standard guidance values, which are for studies of 90-dayduration, for studies of longer or shorter durations. It should be used cautiously with dueconsideration of the nature of the substance in question and the resulting value produced.In particular, care should be taken when using Haber’s rule to assess inhalation data onsubstances which are corrosive or local active or have the potential to accumulate withrepeated exposure.One particular problem to note is that when adjusting the guidance value for very short studydurations this can lead to very high guidance values which are not appropriate. For instance,for a 4 day exposure a guidance value of 2250 mg/kg bw/day for classification as STOT REcategory 2 could potentially be produced. This is above the limit for acute toxicity of 2000mg/kg bw and it does not make sense to have a guidance value for repeated dose toxicity thatis above the guidance value for mortality after acute exposure. To address this problem apragmatic approach is proposed. For studies with exposure durations shorter than 9 days (i.e10% of the 90 days to which the default general guidance value applies) the guidance valueused should be no greater than 10 times the default guidance value. For example, the effectsin an oral range-finding study of 9 days or less should be compared with a guidance value of1000 mg/kg bw/day for STOT-RE Category 2.Expert judgement is needed for the establishment of equivalent guidance values because oneneeds to know about the limitations of the applicability of the proportionality. In thefollowing table the equivalents for 28-day and 90-day studies according to Haber's rule aregiven:390

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Table 3.9.2.2 Equivalent guidance values for 28-day and 90-day studiesStudy type Species Unit Category 190-dayCategory 128-dayCategory 290-dayCategory 228-dayOral Rat mg/kg bw/d ≤ 10 ≤ 30 ≤ 100 ≤ 300Dermal Rat mg/kg bw/d ≤ 20 ≤ 60 ≤ 200 ≤ 600Inhalation, gas Rat ppmV/6 h/d ≤ 50 ≤ 150 ≤ 250 ≤ 750Inhalation, vapor Rat mg/l/6 h/d ≤ 0.2 ≤ 0.6 ≤ 1 ≤ 3Inhalation,dust/mist/fumeRat mg/l/6 h/d ≤ 0.02 ≤ 0.06 ≤ 0.2 ≤ 0.6Annex 1: 3.9.2.9.9. Thus it is feasible that a specific profile of toxicity occurs in repeat-dose animalstudies at a dose/concentration below the guidance value, such as < 100 mg/kg bw/day by the oralroute, however the nature of the effect, such as nephrotoxicity seen only in male rats of a particularstrain known to be susceptible to this effect may result in the decision not to classify. Conversely, aspecific profile of toxicity may be seen in animal studies occurring at above a guidance value, suchas ≥ 100 mg/kg bw/day by the oral route, and in addition there is supplementary information fromother sources, such as other long-term administration studies, or human case experience, whichsupports a conclusion that, in view of the weight of evidence, classification is the prudent action totake.3.9.2.3 Evaluation of hazard informationAnnex 1: 3.9.2.4. .……Evaluation shall be based on all existing data, including peer-reviewedpublished studies and additional acceptable data.3.9.2.3.1 Evaluation of human dataAnnex 1: 1.1.1.4. For the purpose of classification for health hazards (Part 3) established hazardouseffects seen in appropriate animal studies or from human experience that are consistent with thecriteria for classification shall normally justify classification. Where evidence is available fromboth humans and animals and there is a conflict between the findings, the quality and reliability ofthe evidence from both sources shall be evaluated in order to resolve the question of classification.Generally, adequate, reliable and representative data on humans (including epidemiological studies,scientifically valid case studies as specified in this Annex or statistically backed experience) shallhave precedence over other data. However, even well-designed and conducted epidemiologicalstudies may lack a sufficient number of subjects to detect relatively rare but still significant effects,to assess potentially confounding factors. Therefore, positive results from well-conducted animalstudies are not necessarily negated by the lack of positive human experience but require anassessment of the robustness, quality and statistical power of both the human and animal data.Annex 1 3.9.2.7.2. Evidence from human experience/incidents is usually restricted to reports ofadverse health consequence, often with uncertainty about exposure conditions, and may not providethe scientific detail that can be obtained from well-conducted studies in experimental animals.Where relevant human data do not mirror realistic exposure conditions, supportiveinformation may be needed to corroborate the observed effects. A single case report from391

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012deliberate exposure (i.e. abuse) is unlikely to provide sufficiently robust evidence to supportclassification without other evidence.IR/CSA, Section R.7.5.4.2 gives a detailed description on the use of human hazardinformation3.9.2.3.2 Evaluation of non human dataAnnex 1 3.9.2.7.3. Evidence from appropriate studies in experimental animals can furnish muchmore detail, in the form of clinical observations, haematology, clinical chemistry, and macroscopicand microscopic pathological examination, and this can often reveal hazards that may not be lifethreateningbut could indicate functional impairment.All available animal data which are of acceptable quality should be used in a weight ofevidence approach based on a comparison with the classification criteria described above.This should be done separately for each route for which data are available.For each study the effects seen in each sex at or around the guidance values for Category 1and Category 2 should be compared with the effects warranting classification in Category 1and 2. In general findings in the most sensitive sex would be used to determine theclassification. If the NOAEL from the study is above the guidance value (GV), the results ofthat study do not indicate classification for that category (situations 1 and 2 in Figure3.9.2.3.2). If the NOAEL is below the GV then the effective dose level (ED), i.e. the lowestdose inducing significant/severe target organ toxicity as defined in Section 3.9.2.2, should bedetermined based on the criteria described above. If the ED is below the GV then this studyindicates that classification is warranted (situations 2 and 4 in Figure 1).In a case where the ED is above a GV but the NOAEL is below the GV (situations 3 and 5)then interpolation between the ED and the NOAEL is required to determine whether theeffects expected at or below the GV would warrant classification .Figure 3.9.2.3.2Comparison between the NOAEL and the ED versus the guidance valuesSituation 1 Situation 2 Situation 3 Situation 4 Situation 5GVCategory 2- NOAEL 1- ED 2- ED 3- NOAEL 3- NOAEL 2- ED 5GVCategory 1- ED 4- NOAEL 5- NOAEL 4NC Category 2 Interpolation Category 1 Interpolation392

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Where a number of studies are available these should be assessed using a weight of evidenceapproach to determine the most appropriate classification. Where the findings from individualstudies would lead to a different classification then the studies should be assessed in terms oftheir quality, species and strain used, nature of the tested substance (including the impurityprofile and physical form) etc to choose the most appropriate study to support classification.In general, the study giving the most severe classification will be used unless there are goodreasons that it is not the most appropriate. If the effects observed in animals are notconsidered relevant for humans then these should not be used to support classification.Similarly, if there is robust evidence that humans differ in sensitivity or susceptibility to theeffect observed in the study then this should be taken into account, possibly leading to anincrease or decrease in the classification assigned.If there are differences in effects at the GV between studies with different duration then moreweight is usually given to studies of a longer duration (28 days or more). This is becauseanimals may not have fully adapted to the exposure in studies of shorter durations and alsobecause longer duration studies tend to include more thorough and extensive investigations(e.g. in terms of detailed pathology and haematological effects etc) which can generally givemore substantial information compared to shorter duration studies. If a 90-day as well as a 28-day study are available expert judgement has to be used and not just Haber's rule.If there are differences in effects between good quality data in the same sex, species and strainthen other variables such as particle size, vehicle, substance purity and impurities andconcentration should be considered. If the results are considered to be depending on a specificimpurity then different classifications depending on the concentration of the impurity could beconsidered.Any information pertaining to the relevance of findings in animals to humans must be takeninto account and may be used to modify the classification from how it would be if based onthe available animal data. For instance, it may be shown that the findings in animals are notrelevant for humans, for example if the toxicity in animals is mediated by a mode of actionthat does not occur in humans. This would potentially provide a supporting case for noclassification. Similarly, evidence may suggest that the potency of the substance may behigher or lower in humans than in animals, for example because of differences intoxicokinetics/toxicodynamics between the species. Such evidence could be used to increaseor decrease the severity of the classification as appropriate. It should be noted that sucharguments for modifying the classification must be robust and transparent (see 3.9.2.3.4).The final classification based on non human data will be the most severe classification of thethree routes. If it is shown that classification for this endpoint is not required for a specificroute then this can be included in the hazard statement. Evaluation of non human data canresult in no classification, STOT RE 1 or STOT RE 2. The results of the evaluation in nonhuman data should be used in combination with the results of the evaluation of human data.If it is shown that classification for this endpoint is not required for a specific route then thiscan be included in the hazard statement according to the table below.Table 3.9.2.4.1 Inclusion of route of exposure in Hazard statement393

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Route 1 Route 2 Route 3 H-statementCategory 1 Category 2 unknown Causes damage to organs through prolonged or repeatedexposureCategory 1 Category 2 NC Causes damage to organs via route 1 and 2Category 1 NC unknown Causes damage to organs through prolonged or repeatedexposureCategory 1 unknown unknown Causes damage to organs through prolonged or repeatedexposureCategory 1 NC NC Causes damage to organs via route 13.9.2.3.3 ConversionsThe guidance values are giving in mg/kg bodyweight. Where the doses in a study are given indifferent units they will need to be converted as appropriate. For instance the dosages infeeding and drinking water studies are often expressed in ppm, mg test substance/ kg (feed) ormg (test substance)/l (drinking water).Where insufficient information is reported in the study to perform the conversion, Table3.9.2.3.3.1 and Table 3.9.2.3.3.2 can be used as “Approximate relations”. These tables arederived from the following documents: IR/CSA, Chapter 8, Table 17; and OECDENV/JM/MONO (2002)19, 04-Sep-2002, Table 1; L.R. Arrington (Introductory LaboratoryAnimal Science, 1978).Table 3.9.2.4.2(a) Food conversionAnimal Weight (kg) Food consumed per day (g) Factor 1mg/kgbw/dequivalent to ppm in dietRat, young 0.10 10 10Rat ,older 0.40 20 20Mouse 0.02 3 7Dog 10 250 40Table 3.9.2.4.2(b) Conversion drinking waterAnimal Weight (kg) Drinkingwaterconsumed per day(g)Factor 1mg/kgbw/d equivalent to ppmin drinking waterRat, young 0.25 28 (25-30) 9Rat ,older 0.40 28 (25-30) 14Mouse 0.025 5 (4-7) 8Dog 13 350 37The conversion is performed according to the following simple equation:mg/kgbw = ppm/factorExample: In a 4 week study rats received the 1000 ppm test substance in feedDosage (mg/kg bw): 1000:10= 100 mg/kgbw.394

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In any case a calculation of the average substance intake based on measured bodyweight andconsumption data is preferable and should be performed where possible.Gases: mg/l into ppm:Effect doses from gases given in the unit mg/l have to be converted into the unit ppm as usedby the CLP via the following simplified formula assuming values for ambient pressure of 1atm = 101.3 kPa and 25 ° c:mg/l = ppm x MW/0.024453.9.2.3.4 Weight of evidenceAnnex 1: 3.9.2.3. Classification is determined by expert judgment (see section 1.1.1), on the basisof the weight of all evidence available including the guidance presented below.3.9.2.4. Weight of evidence of all data (see section 1.1.1), including human incidents,epidemiology, and studies conducted in experimental animals, is used to substantiate specific targetorgan toxic effects that merit classification. This taps the considerable body of industrial toxicologydata collected over the years. Evaluation shall be based on all existing data, including peerreviewedpublished studies and additional acceptable data.Annex 1: 3.9.2.10.2. When well-substantiated human data are available showing a specific targetorgan toxic effect that can be reliably attributed to repeated or prolonged exposure to a substance,the substance shall normally be classified. Positive human data, regardless of probable dose,predominates over animal data. Thus, if a substance is unclassified because no specific target organtoxicity was seen at or below the dose/concentration guidance value for animal testing, ifsubsequent human incident data become available showing a specific target organ toxic effect, thesubstance shall be classified.3.9.2.10.3. A substance that has not been tested for specific target organ toxicity may, whereappropriate, be classified on the basis of data from a validated structure activity relationship andexpert judgment-based extrapolation from a structural analogue that has previously been classifiedtogether with substantial support from consideration of other important factors such as formation ofcommon significant metabolites.In cases where there is sufficient human evidence that meets the criteria given in CLP AnnexI, Table 3.9.1 to support classification then this will normally lead to classification inCategory 1, irrespective of other information available.Where human evidence does not meet this criterion, for example when the weight of evidenceis not sufficiently convincing (limited number of cases or doubt on causal relationship) orbecause of the nature and severity of the effects (CLP Annex I, 3.9.2.7.3 and 3.9.2.8.1), thenclassification is based primarily on the non-human dataIf there are no human data then the classification is based on the non-human data. If there ishuman data indicating no classification but there is also non-human data indicatingclassification then the classification is based on the non-human data unless it is shown that thehuman data cover the exposure range of the non-human data and that the non-human data arenot relevant for humans. If the human and non-human data both indicate no classification thenclassification is not required.3.9.2.4 Decision on classificationAnnex 1: 3.9.2.7.1. Reliable evidence associating repeated exposure to the substance with a consistentand identifiable toxic effect demonstrates support for the classification.395

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex 1: 3.9.2.7.3. Evidence from appropriate studies in experimental animals can furnish muchmore detail, in the form of clinical observations, haematology, clinical chemistry, and macroscopicand microscopic pathological examination, and this can often reveal hazards that may not be lifethreateningbut could indicate functional impairment. Consequently all available evidence, andrelevance to human health, shall be taken into consideration in the classification process, includingbut not limited to the following toxic effects in humans and/or animals:(a)(b)(c)(d)(e)(f)(g)morbidity or death resulting from repeated or long-term exposure. Morbidity or death mayresult from repeated exposure, even to relatively low doses/concentrations, due tobioaccumulation of the substance or its metabolites, and/or due to the overwhelming of thede-toxification process by repeated exposure to the substance or its metabolites.significant functional changes in the central or peripheral nervous systems or other organsystems, including signs of central nervous system depression and effects on special senses(e.g., sight, hearing and sense of smell).any consistent and significant adverse change in clinical biochemistry, haematology, orurinalysis parameters.significant organ damage noted at necropsy and/or subsequently seen or confirmed atmicroscopic examination.multi-focal or diffuse necrosis, fibrosis or granuloma formation in vital organs withregenerative capacity.morphological changes that are potentially reversible but provide clear evidence of markedorgan dysfunction (e.g., severe fatty change in the liver).evidence of appreciable cell death (including cell degeneration and reduced cell number) invital organs incapable of regeneration.Annex 1: 3.9.2.8. Effects considered not to support classification for specific target organ toxicityfollowing repeated exposure3.9.2.8.1. It is recognised that effects may be seen in humans and/or animals that do not justifyclassification. Such effects include, but are not limited to:(a) Clinical observations or small changes in bodyweight gain, food consumption or water intakethat have toxicological importance but that do not, by themselves, indicate “significant" toxicity.(b) Small changes in clinical biochemistry, haematology or urinalysis parameters and/or transienteffects, when such changes or effects are of doubtful or minimal toxicological importance(c) Changes in organ weights with no evidence of organ dysfunction.(d) Adaptive responses that are not considered toxicologically relevant.(e) Substance-induced species-specific mechanisms of toxicity, i.e. demonstrated with reasonablecertainty to be not relevant for human health, shall not justify classification.If the evaluation of available data on a substance shows that the criteria for classification in acategory are fulfilled than the substance shall be classified in that category for STOT-RE.If the data show that classification is warranted in Category 1 for one route and in Category 2for another route then the substance shall only be classified in Category 1. The correspondinghazard statements are provided in Section 3.9.4.1.396

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012If only data is available for one route showing that classification is warranted than no routeshould be stated in the hazard statement. If the data conclusively show that no classificationfor STOT-RE is warranted for a specific route then the remaining routes should be stated. Ifthe data show that classification is warranted in Category 1 for one route and in Category 2for another route then the hazard statement for Category 1 should include both routes becausesubstances are placed in one of two categories.3.9.2.5 Additional considerationsIn the following sections some special aspects in the decision process on classification aredescribed in more detail.3.9.2.5.1 Irritating/corrosive substancesSubstances (or mixtures) classified as corrosive may cause severe toxicological effectsfollowing repeated exposure, especially in the lungs following inhalation exposure. In suchcases, it has to be evaluated whether the severe effect is a reflection of true repeated exposuretoxicity or whether it is in fact just acute toxicity (i.e. corrosivity). One way to distinguishbetween these possibilities is to consider the dose level which causes the toxicity. If the doseis more than half an order of magnitude lower than that mediating the evident acute toxicity(corrosivity) then it could be considered to be a repeated-dose effect distinct from the acutetoxicity. In this case, classification as specific target organ toxicant (repeated exposure) wouldbe warranted even if the substance (or mixture) is also classified as acutely toxic and/orcorrosive.In assessing non systemic effects caused by irritating/corrosive substances it should be kept inmind, that the guidance values /criteria for R48 in the DSD and later on those for STOT-REof the CLP were derived from acute toxicity criteria (lethality based) assuming that systemiceffects show a time dependent increase of severity due to accumulation of toxicity and takingalso adaptive and detoxification processes into account. The effect considered in this contextwas lethality. This indicates that classification was intended for the presence of severe healthdamage, only. (see ECBI/67/00)3.9.2.5.2 HematotoxicityMethaemoglobin generating agentsMethaemoglobinemia has often been regarded as an acute clinical symptom resulting from theaction of methemoglobin-generating agents. If lethality is observed in humans or in animals 50or can be predicted (QSAR), methemoglobin generating substances should be classified in theAcute Toxicity Hazard Class. Since this effect is difficult to detect in rodents, expertjudgement should be used (cf. Guidance on Acute toxicity, Example2). Ifmethemoglobinemia does not result in lethality but exposure to methaemoglobin generatingagents results in signs of damage to the erythrocytes and haemolysis, anaemia or hypoxemia,the formation of methaemoglobin shall be classified accordingly either in STOT-SE orSTOT-RE. (Muller A. et al., 2006).Haemolytic anaemiaThe guidance developed for classification of substances inducing haemolytic anaemiaaccording to 67/548/EEC (Muller A. et al., 2006) cannot directly be used under CLP because50 Observation of lethality following methemoglobin formation is not usual, as several animals are more tolerantto it. Extrapolation to the human situation must be the critical decision key.397

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012of the changes in criteria (see CLP Annex I, 3.9.2.7.3 c and 3.9.2.8.b, d ). The major criterionfor haemolytic anaemia changed:o From “Any consistent changes in haematology which indicate severe organdysfunction.”o To “Any consistent and significant adverse changes in haematology.”This indicates that less adverse effects are considered for classification according to CLP.This is consistent with the changes in the other criteria for classification for repeatedexposure.Adaptation towards the criteria according to CLP results in the following guidance:It is evident that anaemia describes a continuum of effects, from sub-clinical to potentiallylethal in severity. Overall, the interpretation of study findings requires an assessment of thetotality of findings, to judge whether they constitute an adaptive response or an adversetoxicologically significant effect. If a haemolytic substance induces one or more of the serioushealth effects listed as examples below within the critical range of doses, classification iswarranted. It is sufficient for classification that only one of these criteria is fulfilled.Annex I: 2.9.2.7.3.(a) morbidity or death resulting from repeated or long-term exposure. Morbidity or death mayresult from repeated exposure, even to relatively low doses/concentrations, due to bioaccumulationof the substance or its metabolites, and/or due to the overwhelming of the de-toxification process byrepeated exposure to the substance or its metabolites;Example: Premature deaths in anaemic animals that are not limited to the first three days oftreatment in the repeated dose study. (Mortality during days 0–3 may be relevant foracute toxicity.) Clinical signs of hypoxia, e.g. cyanosis, dyspnoea, pallor, in anaemic animals thatare not limited to the first three days of treatment in the repeated dose study.(b) significant functional changes in the central or peripheral nervous systems or other organsystems, including signs of central nervous system depression and effects on special senses (e.g.sight, hearing and sense of smell);(c) any consistent and significant adverse effect in clinical biochemistry, haematology or urinalysisparameters;398Examples: Reduction in Hb at ≥20%. Reduction in functional Hb at ≥20% due to a combination of Hb reduction andMetHb increase. Haemoglobinuria that is not limited to the first three days of treatment in therepeated dose study in combination with other changes indicating significanthaemolytic anaemia (e.g. a reduction in Hb at ≥10%). Haemosiderinuria supported by relevant histopathological findings in the kidney incombination with other changes indicating significant haemolytic anaemia (e.g. areduction in Hb at ≥10%).(d) significant organ damage noted at necropsy and/or subsequently seen or confirmed atmicroscopic examination;(e) multifocal or diffuse necrosis, fibrosis or granuloma formation in vital organs with regenerative

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012capacity;Example: Multifocal or diffuse fibrosis in the spleen, liver or kidney.(f) morphological changes that are potentially reversible but are clear evidence of markedorgan dysfunction (e.g. severe fatty change in the liver)Example: Tubular nephrosis(g) evidence of appreciable cell death (including cell degeneration and reduced cell number) in vitalorgans incapable of regeneration.In the case where multiple less severe effects with regenerative capacity were observed, theclassification should apply as“Assessment shall take into consideration not only significant changes in a single organ orbiological system but also generalised changes of a less severe nature involving severalorgans.”Example: Marked increase of haemosiderosis in the spleen, liver or kidney in combination withother changes indicating significant haemolytic anaemia (e.g. a reduction in Hb at≥10%) in a 28 day study. Significant increase in haemosiderosis in the spleen, liver or kidney in combinationwith microscopic effects like necrosis, fibrosis or cirrhosis.Annex I: 3.9.2.8.1. It is recognised that effects may be seen in humans and/or animals that do notjustify classification. Such effects include, but are not limited to:(a) clinical observations or small changes in bodyweight gain, food consumption or water intakethat have toxicological importance but that do not, by themselves, indicate ‘significant’ toxicity;(b) small changes in clinical biochemistry, haematology or urinalysis parameters and/or transienteffects, when such changes or effects are of doubtful or minimal toxicological importance;Example: Significant decrease in Hb without any other significant indicators of haemolyticanaemia. Minimal to slight increase in MetHb formation without any other indications ofsignificant haemolytic anaemia.(c) changes in organ weights with no evidence of organ dysfunction;(d) adaptive responses that are not considered toxicologically relevant.Example: Only adaptive or compensating effects without significant signs of haemolyticanaemia.(e) substance-induced species-specific mechanisms of toxicity, i.e. demonstrated with reasonablecertainty to be not relevant for human health, shall not justify classification.3.9.2.5.3 Mechanisms not relevant to humans (CLP Annex I, 3.9.2.8.1. (e))399

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In general, valid data from animal experiments are considered relevant for humans and areused for hazard assessment/classification. However, it is acknowledged that there are caseswhere animal data are not relevant for humans and should not be used for that purpose. This isthe case when there is clear evidence that a substance – induced effect is due to a speciesspecificmechanism which is not relevant for humans. Examples for such species differencesare described in this section.-2-μ globulin nephropathy in male ratsThe protein α-2-μ globulin, which is primarily synthesized in male rats, has the capability tobind to certain chemicals. The resultant adducts accumulate as droplets in the kidneys andcauses progressive renal toxicity within a few weeks which can ultimately lead to kidneytumours. This specific mechanism is unique to male rats and has no relevance for humans.Examples of chemicals causing -2-μ globulin nephropathy are: unleaded gasoline,chlorinated paraffins, isophorone, d-limonene.Specific thyroid toxicity via liver enzyme inductionCertain chemicals cause induction of liver enzymes and are interfering with the regulation ofthyroid hormones. An increase in the activity of hepatic UDPG-transferase results inincreased glucuronidation of thyroid hormones and increased excretion. It is known thatrodents are highly sensitive to a reduction in thyroid hormone levels (T4), resulting in thyroidtoxicity (e.g. hypertrophy, hyperplasia) after repeated stimulation / exposure of this organ.This in turn is related to an increase in the activity of hepatic UDPG-transferase. Humans,unlike rodents, possess a T 4 binding protein that greatly reduces susceptibility to plasma T 4depletion and thyroid stimulation. Thus, such a mechanism/effect cannot be directlyextrapolated to humans, i.e. these thyroid effects observed in rodents caused by an increase inhepatic UDPG-transferase are therefore considered of insufficient concern for classification(see ECBI/22/98-Add1).Peroxisome induction/proliferationPeroxisomes are cell-organelles which can be induced to a specifically high level in rats andmice under certain conditions, e.g. by repeated exposure to long chain and branched fattyacids. Peroxisome proliferation which is especially occurring in the liver causes liver toxicity(e.g. hyperplasia, oxidative stress) and can ultimately after longterm exposure also may leadto tumours.There is no evidence of e.g. hepatomegaly from clinical studies in humans treatedwith peroxisome proliferators ( I.H.F.Purchase, Human & Experimental Toxicology (1994),13, Suppl.2 S47-S48). Examples are Clofibrat and Diethylhexylphthalate (DEHP).Lung OverloadThe relevance of lung overload in animals to humans is currently not clear and is subject tocontinued scientific debate.3.9.2.5.4 Adaptive responses (CLP Annex I, 3.9.2.8.1. (d))Adaptive (compensatory) changes generally constitute a normal biochemical or physiologicalresponse to a substance or to the effect of the substance (e.g. in response to methaemoglobinformation), usually manifested as an increase in background processes such as metabolism orerythropoiesis etc, which are generally reversible with no adverse consequences on cessationof exposure. In some cases the adaptive response may also be associated with pathologicalchanges which reflect the normal response of the target tissue to substances. For example,liver hypertrophy in response to enzyme induction, the increase in alveolar macrophagesfollowing inhalation of insoluble particles that must be cleared from the lungs, and the400

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012development of epithelial hyperplasia and metaplasia in the rat larynx in response toinhalation of irritants.Determination of whether adaptive changes support a classification requires a holisticassessment of the nature and severity of the observations and their dose-response relationshipusing expert judgement. Exposure to a substance can lead to a spectrum of effects which varyin incidence and severity with dose. At lower doses there may be adaptive changes which arenot considered to be toxicologically significant or adverse, whereas at higher doses thesechanges may become more severe and/or other effects may occur which together constitutefrank toxicity. Also, sometimes the adaptive effect is observed but the primary effect is notbecause the relevant parameter is not determined or not determined at the right time. Forexample, irritation of the larynx after inhalation of irritants is not observed at the end of arepeated dose study because of the quick response. The adaptive effect can then be used as anindication of the primary effect. It is often difficult to clearly distinguish between changeswhich are adaptive in nature and those which represent clear overt toxicity and thisassessment requires expert judgement. Where the response to a substance is considered to bepurely adaptive at dose levels relevant for classification then no classification would beappropriate.3.9.2.5.5 Post-observation periods in 28 day and 90 day studiesFor subacute/subchronic testing protocols, the usual guideline procedure is to sacrifice theexposed animals immediately after the end of the exposure period (d 29 or 91).Japanese agencies often require a 14 days postobservation period for 28 day studies (OECD407). This means that 10 more animals in the top dose and 10 more animals as an additionalcontrol group are then necessary.The reversibility of organotoxic effects can in most cases be estimated by the pathologist fromhistologic findings without a post-observation period. Certain effects are entirely reversible such as simple irritation or many forms of liver,testicular and hematotoxicity. Other effects may be reversible in morphological terms but the reserve capacity of theorganism may be irreversibly compromised (such as in the case of kidney toxicity with apersistent loss in kidney nephrons). Some forms of tissue toxicity may be fundamentally irreversible, such as CNS- and neurotoxicitywith specific histological findings, cardiac toxicity and lung toxicity. Often, sucheffects do not return to normal morphology and may deteriorate even after the end ofexposure.3.9.2.6 Setting of specific concentration limitsSpecific concentration limits (SCLs) for STOT-RE may be set by the supplier in somesituations according to Article 10 of CLP. For STOT-RE, this may only be done forsubstances inducing target organ toxicity at a dose level or concentration clearly (more thanone magnitude) below the guidance values according to CLP Annex I, Table 3.9.2, thatcorresponds to ED below 1 mg/kg bodyweight from the 90-day oral study. Where theexposure duration is not 90 days the ED has to be adjusted to an equivalent for 90 days usingHaber’s law and expert judgement (as described above). This will be mainly based on data inexperimental animals but can also be used for human data if reliable exposure data areavailable. Setting of SCLs above the GCL is not applicable for STOT RE becauseclassification for STOT RE is based on potency. Substances with a low potency do not requireclassification for this hazard class and substances with a medium or high potency areclassified in a category defined by the GV.401

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The SCL for a Category 1 substance (SCLCat.1) can be determined using the followingformula:EDSCLCat . 1 100%Equation 3.9.2.6(a)GV 1SCL Cat 1: 0.12 mg/kgbw/10 mg/kgbw x 100%= 1.2% --> 1%ED (effective dose) is the dose inducing specific target organ toxicity and GV1 is theguidance value for Category 1 according to CLP Annex I, Table 3.9.2 of Annex I correctedfor the exposure duration. The resulting SCL is rounded down to the nearest preferred value(1, 2 or 5).Though classification of a mixture in Category 1 is not triggered if a Category 1 constituent ispresent in lower concentrations than the established SCL, a classification in Category 2should be considered. The SCL for classification of a mixture in Category 2 (SCLCat. 2)based on substances classified in Category 1can be determined using the following formula:EDSCLCat . 2 100%Equation 3.9.2.6(b)GV 2SCL Cat 2: 0.12 mg/kgbw/100 mg/kgbw x 100%=0.12% --> 0.1%In this formula the ED (effective dose) is the dose inducing specific target organ toxicity andGV2 is the upper guidance value for Category 2 according to CLP Annex I, Table 3.9.3corrected for the exposure duration. The resulting SCL is rounded down to the nearestpreferred values (1, 2 or 5).It is not appropriate to determine SCLs for substances classified in Category 2 sinceingredients with a higher potency (i.e. lower effect doses than the guidance values ofCategory 2) will be classified in Category 1 and substances with respective higher effectdoses will generally not be classified. For example, a substance inducing significant specifictarget organ toxicity at 0.12 mg/kg bw/day in a 90-day oral study would require a SCL forCategory 1 of 1% and for Category 2 of 0.1%.402

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9.2.7 Decision logic for classification of substancesThe decision logic which follows is provided as additional guidance to the criteria. It isstrongly recommended that the person responsible for classification, study the criteria forclassification before and during use of the decision logic.Does the substance have data and/or information to evaluate specific targetorgan toxicity following repeated exposure?NoClassificationnot possibleYesFollowing repeated exposure,Can the substance produce significant toxicity in humans, orCan it be presumed to have the potential to produce significant toxicity inhumans on the basis of evidence from studies in experimental animals?See 3.9.2 for criteria and guidance values. Application of the criterianeeds expert judgment in a weight of evidence approach.YesCategory 1DangerNoFollowing repeated exposure,Can the substance be presumed to have the potential to be harmful tohuman health on the basis of evidence from studies in experimentalanimals?See 3.9.2 for criteria and guidance values. Application of the criterianeeds expert judgment in a weight of evidence approach.YesCategory 2WarningNoNot classified403

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9.3 Classification of mixtures for STOT-RE3.9.3.1 Identification of hazard informationWhere toxicological information is available on a mixture this should be used to derive theappropriate classification. Such information may be available from the mixture manufacturer.Where such information on the mixture itself is not available information on similar mixturesand/or the component substances in the mixture must be used, as described below.Further, the hazard information on all individual components in the mixture could beidentified as described in Section 3.9.3.3.2.3.9.3.2 Classification criteria for mixturesAnnex 1: 3.9.3.1. Mixtures are classified using the same criteria as for substances, or alternativelyas described below. As with substances, mixtures shall be classified for specific target organtoxicity following repeated exposure.3.9.3.3 When data are available for the complete mixtureAnnex 1: 3.9.3.2.1. When reliable and good quality evidence from human experience orappropriate studies in experimental animals, as described in the criteria for substances, is availablefor the mixture (see 1.1.1.3), then the mixture shall be classified by weight of evidence evaluationof these data. Care shall be exercised in evaluating data on mixtures, that the dose, duration,observation or analysis, do not render the results inconclusive.In cases where test data for mixtures are available, the classification process is exactly thesame as for substances.3.9.3.3.1 When data are not available for the complete mixture: bridging principlesAnnex 1: 3.9.3.3.1. Where the mixture itself has not been tested to determine its specific targetorgan toxicity, but there are sufficient data on the individual ingredients and similar tested mixturesto adequately characterise the hazards of the mixture, these data shall be used in accordance withthe bridging principles set out in section 1.1.3.When there are no test data on the mixture as a whole, so called “Bridging principles” may beapplied where there are data available on similar tested mixtures and on the individualhazardous ingredient substances within the mixture that are sufficient to adequately assess thehazards of the mixture.3.9.3.3.2 When data are available for all components or only for some components ofthe mixtureAnnex 1: 3.9.3.4.1. Where there is no reliable evidence or test data for the specific mixture itself,and the bridging principles cannot be used to enable classification, then classification of the mixtureis based on the classification of the ingredient substances. In this case, the mixture shall beclassified as a specific target organ toxicant (specific organ specified), when at least one ingredienthas been classified as a Category 1 or Category 2 specific target organ toxicant and is present at orabove the appropriate generic concentration limit as laid out in Table 3.9.4 below for Category 1and 2 respectively.404

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9.3.3.3 Components of a mixture that should be taken into account for the purposeof classificationComponents with a concentration equal to or greater than the generic concentration limits(1% for Category 1 components and 10 % for Category 2; see CLP Annex I, Table 3.9.4) orwith a specific concentration limit (see also 3.9.3.5) will be taken into account forclassification purposes. Specific concentration limits have preference over the genericconcentration limits.3.9.3.4 Generic concentration limits for substances triggeringclassification of mixturesAnnex 1: Table 3.9.4Generic concentration limits of ingredients of a mixture classified as a specific target organtoxicant that trigger classification of the mixture.Ingredient classified as:Generic concentration limits triggering classification of themixture as:Category 1 Category 2Category 1Concentration 10% 1.0% concentration 10%Specific Target Organ ToxicantCategory 2Specific Target Organ ToxicantConcentration 10%(Note 1)Note 1If a Category 2 specific target organ toxicant is present in the mixture as an ingredient at aconcentration ≥ 1,0 % a SDS shall be available for the mixture upon request.Annex 1: 3.9.3.4.4. Care shall be exercised when toxicants affecting more than one organ systemare combined that the potentiation or synergistic interactions are considered, because certainsubstances can cause target organ toxicity at < 1% concentration when other ingredients in themixture are known to potentiate its toxic effect.In the case a specific concentration limit has been established for one or more ingredientsthese SCLs have precedence over the respective generic concentration limit.When classifying a mixture for STOT-RE the additive approach, where the concentrations ofindividual components with the same hazards are summed, is not used. If any individualcomponent is present at a concentration higher than the relevant generic or specificconcentration limit then the mixture will be classified.405

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9.3.5 Decision logic for mixturesA mixture should be classified either in Category 1 or in Category 2, according to the criteriadescribed above. When a mixture is classified for STOT-SE on basis of test data, the hazardstatement will specify the target organs, in the same way as for a substance. If a mixture isclassified on basis of the ingredients, the hazard statement (H372 for Category 1 or H373 forCategory 2) may be used without specifying the target organs, as appropriate. In the sameway, the route of exposure should not be specified, except if data are available for thecomplete mixture and if it is conclusively demonstrated that no other routes of exposure causethe hazard.The decision logic which follows is provided as additional guidance to the criteria. It isstrongly recommended that the person responsible for classification study the criteria forclassification before and during use of the decision logic.Does the mixture have data and/or information to evaluate?YesSeeSubstancesNoCan bridging principles be applied?YesClassify inappropriatecategoryNoDoes mixture contain one or more ingredients classified as a Category 1specific target organ toxicant at a concentration of: ≥ 10% ?YesCategory 1No Does mixture contain one or more ingredients classified as aCategory 2 specific target organ toxicant at a concentration of:≥ 10%(A SDS is required if a cat 2 substance is present at or above 1%)YesDangerCategory 2WarningNoNot classified406

Guidance on the Application of the CLP Criteria Version 3.0 - November 20123.9.4 Hazard communication in form of labelling for STOT RE3.9.4.1 Pictograms, signal words, hazard statements andprecautionary statementsAnnex I: 3.9.4.1. Label elements shall be used in accordance with Table 3.9.5 for substances ormixtures meeting the criteria for classification in this hazard class.GHS PictogramsTable 3.9.5Label elements for specific target organ toxicity after repeated exposureClassification Category 1 Category 2Signal word Danger WarningHazard statementPrecautionary statementpreventionH372: Causes damage toorgans (state all organsaffected, if known) throughprolonged or repeatedexposure (state route ofexposure if it is conclusivelyproven that no other routes ofexposure cause the hazard)P260P264P270H373: May cause damage toorgans (state all organsaffected, if known) throughprolonged or repeatedexposure (state route ofexposure if it is conclusivelyproven that no other routes ofexposure cause the hazard)P260Precautionary statement response P314 P314Precautionary statement storagePrecautionary statement disposal P501 P501The hazard statement should include the primary target organ(s) of toxicity. Organs in whichsecondary effects were observed should not be included. The route of exposure should not bespecified, except if it is conclusively demonstrated that no other routes of exposure cause thehazard.When a mixture is classified for STOT-RE on basis of test data, the hazard statement willspecify the target organs, in the same way as for a substance. If a mixture is classified onbasis of the ingredients, the hazard statement (H372 for Category 1 or H373 for Category 2)may be used without specifying the target organs, as appropriate.In the same way, the route of exposure should not be specified, except if data are available forthe complete mixture and if it is conclusively demonstrated that no other routes of exposurecause the hazard.It is recommended to include no more then three primary target organs for practical reasonsand because the classification is for specific target organ toxicity. If more target organs areaffected it is recommended that the overall systemic damage should be reflected by using themore general term “damage of organs”.407

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124083.9.4.2 Additional labelling provisionsAnnex 1: 3.9.2.10.4 Saturated vapour concentration shall be considered, where appropriate, as anadditional element to provide for specific health and safety protectionAccording to CLP Annex I, 3.9.2.10.4 the saturated vapour concentration shall be consideredas an additional element for providing specific health and safety protection. Thus if aclassified substance is highly volatile a supplementary precautionary advice (e.g.“Special/additional care should be taken due to the high saturated vapour pressure”) might begiven in order to emphasize the hazard in case it is not already covered by the general Pstatements. (As a rule substances for which the ratio of the effect concentration at ≤ 4h to theSVC at 20° C is ≤ 1/10).Although not according to the criteria of STOT-RE, the following EU-special hazardstatement “Repeated exposure” may be used when appropriate:EUH066- “Repeated exposure may cause skin dryness or cracking” (see Chapter 3.2Skin/Corrosion/Irritation).3.9.5 Re-classification of substances and mixtures classified for STOT-REaccording to DSD and DPDClassification with STOT–RE according to CLP is comparable to the classification withR48/X according to DSD. Also substances and mixtures currently classified with R33 shouldbe considered because there is no corresponding classification in CLP. However, differencesare present regarding the approach to route-to-route extrapolation.3.9.5.1 Is direct “translation” of classification and labellingpossible for STOT-RE substances?Direct translation of substances or mixtures classified with R48/X is possible becauseclassification criteria are based on the dose and the severity of a toxic effect and arecomparable in both, CLP and DSD. However, in some cases a change in the Category mayresult by reviewing the data.Substances or mixtures classified with R48/23, R48/20 (for vapour), R48/24 and/or R48/25shall be classified as STOT-RE Category 1 because less adverse effects and higher guidancevalues are required for classification according to CLP compared to DSD. Notable, there isone exception: dust/mist/fume with an ED > 0.02 and ≤ 0.025 mg/l/6h which are classifiedaccording to DSD with R48/23 might not be classified in Category 1 according to CLP.Setting of SCL may be considered for substances showing STOT-RE at levels clearly belowthe guidance values (see 3.9.2.6).All substances or mixtures classified with R48/20 (for dust/mist/fume), R48/21 and/or R48/22shall be classified generally at least as STOT-RE Cat 2. Again, dust/mist/fume with an ED >0.2 and ≤ 0.25 mg/l/6h which are classified according to DSD with R48/20 might not beclassified according to CLP. However, due to the general increase in guidance values, therequirement for less severe effects classification in Category 2 should also be considered but.If translation results in a classification in STOT-RE Category 1 for one route and in STOT-RE Category 2 for another route only classification in Category 1 is required (for both routes).In contrast to DSD where the route of exposure is included in the classification and correlateswith the routes tested (or extrapolated), according to CLP the exposure route should bespecified only when it is conclusively proven that no other routes of exposure cause thehazard. Therefore, the route specific data on STOT-RE should be re-evaluated. A re-

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012evaluation is also necessary because the primary target organs for STOT-RE should be statedin the hazard statement.3.9.5.2 Re-evaluation of the STOT-RE dataGasses classified with R48/20 or R48/23 should be re-evaluated because the guidance valueschanged from general guidance values in mg/L for aerosols, vapours and gasses to a specificguidance value for gasses in ppm.Substances or mixtures not classified for, STOT-RE including substances or mixturesclassified with R33, should be re-evaluated because less adverse effects and higher guidancevalues are required for classification according to CLP compared to DSD. Also, effects inhumans are now considered for classification generally without restrictions to the exposurelevel.3.9.6 Examples of classification for STOT-RERemarks:The classification proposals for the examples refer only to STOT-RE.Labelling is done only with respect to hazard statements (statement with respect of organsaffected = target organs).3.9.6.1 Examples of substances fulfilling the criteria forclassification3.9.6.1.1 Example 1: Hydroxylamine / Hydroxylamonium salts (CAS no. 7803-49-8)Application of criteria for evaluation/classification and decision on classification: Use ofstudies with different duration; Haber’s rule; Expert judgementAvailable information:1) Human experience: No information available2) Animal data:Background:Hydroxylamine and its salts are direct MetHb producers in contrast to aromatic amines, whichrequire metabolic activation (XI/484/92).Several studies are available for the assessment of the toxicity after repeated administration:– 4-week drinking water study (BASF, 1989)– 3-month drinking water study (BASF, 1989)– Combined chronic/carcinogenicity study in drinking water in rats (BASF, 2001)Though not explicitly stated in the criteria the "... study with the longest duration shouldnormally be used".– In the 3-month-study at the dose level of 21 mg/kg bw only “slight to moderatehematotoxic effects” were observed. Thus this dose would not be a sufficient EDcausing ”significant/severe” effects, but it can be concluded that via interpolation anED would result within the Guidance Value Range for Cat 2 (10-100 mg/kg bw).– A classification in Category 2 would be warranted based on the 3-month-study.409

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In the combined chronic/carcinogenicity study (BASF, 2001), the effects observed after 12and 24 months are to be considered separately:12 month study:– 0 ppm (control): hemosiderin storage of low degree in males and females (spleen)– 5 ppm (males 0.3 mg and females 0.4 mg/kg bw/day): No substance-induced effects;hemosiderin storage of low degree in males and females, comparable to controls.– 20 ppm (males 1.1 mg and females 1.6 mg/kg b.w./day): Here, hemosiderin depositswith the gradation of moderate was observed in the spleens of the males; hemosiderinstorage of low degree in females comparable to controls. This effect is not to beregarded as serious since hematology did not reveal any findings whatsoever withregard to anemia. This is supported by the fact that no substantial (1/10 moderate, but1/10 severe in the male control group) extramedullary hematopoiesis was observed inthis group. In the histopathological examination, the spleen was not found to beimpaired morphologically. Thus, this dose is to be regarded as the NOAEL for maleswhereas it is the NOEL for females.– 80 ppm (males 4.5 mg and females 6.2 mg/kg b.w./day): The clinicochemical findingsare assessed as mild anemia in the males (e.g. decrease of RBC, HB and HT (< 10%);MCV increased at the beginning and compensatory normalization later) and, also asmild anemia in the females (decrease in RBC < 12%, HB < 10% and HT < 10%). Theincrease of MCV, PLT and RET and of Howell-Jolly bodies is regarded as acompensatory effect, and the bone marrow still reacts, i.e. it does not demonstrate "...decreased bone marrow production of red blood cells" within the meaning of thecriteria. The only slight increase of the Heinz bodies is considered to be a sign of aweak hematotoxic effect. From the point of view of histopathology, the effects(hemosiderin storage, extramedullary hematopoesis) can be regarded as signs ofanemia, but not within the meaning of "serious" (the effect was more pronounced in thefemales than in the males). The extramedullary hematopoiesis observed is thus againcompensatory in the sense of a functional counterreaction.Assessment:For a 12-month study, cut-off values of 25 and 2.5 mg/kg bw/day (100 mg/kg/day : 4) haveto be regarded for STOT-RE Category 1 vs. Category 2 respectively. At the dose level of 1.1(m) or 1.6 mg/kg bw/day (f), no hematotoxic effects whatsoever or extramedullaryhematopoiesis were observed, nor substantial hemosiderin deposits. The effects at 4.5 (f) and6.2 (m) mg/kg bw/day are regarded as mild anemia; however, more distinct effects may beexpected to occur up to the cut-off value (25 mg/kg/day). Therefore, a classification inCategory 2 seems justified.24-month study:In contrast to the 12-month study, no complete hematological examination was carried out,i.e. only morphological parameters were evaluated, yet full histopathology. The followingfindings relevant to classification – with the exception of the neoplasias – were obtained:– 5 ppm (males 0.2 mg and females 0.4 mg/kg b.w./day): No nonneoplastic effects– 20 ppm (males 1 mg and females 1.6 mg/kg b.w./day): Increased proportion ofhemosiderin deposits in the spleens of the females, but no extramedullaryhematopoiesis, which demonstrates that there was no clear anemia before.410Remark:The fact that, at this dose level, hemosiderin was detected only in the males in the 12-

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012month study and an increased proportion of it only in the females in the 24-month studyshows that this effect was only borderline.– 80 ppm (males 3.7 mg and females 6.2 mg/kg b.w./day): Again hemosiderin storage andextramedullary hematopoesis were observed, yet no serious effects in hematology norhistopathology.Furthermore, the results of the study do not indicate that any animal diedprematurely on account of the anemia.Remark:No effects at all were observed in kidneys nor in liver in the 12-month study. In the 3month study only in the highest dose the relative liver weights were increased in themales; in the 3 month as well as in the 24-month study only marginal effects (diffusehemosiderin storage in the liver) in both sexes was observed in the highest dose.Assessment:The results of the 24 month study show that effects as seen after 12 month exposure are notsubstantially increased.Classification:Based on the evaluation of the 3-month-study and the more relevant 12-month-study byexpert judgement a classification in Category 2 is warranted.Labelling:Hazard statement: H373 May cause damage to blood system through prolonged or repeatedexposure(See also ECBI/ 14/3/ Add 3 (2003) and ECBI/56/04 Rev 1)3.9.6.1.2 Example 2: But-2-yn-1,4-diol (EC No 203-788-6; CAS No 110-65-6)Current classification according to DSD: Xn; R48/22Application of criteria for evaluation/classification and allocation of hazard statements withrespect to specific target organs and route of exposureAvailable information:1) Human experience: no information available2) Animal data: 28 d oral study 28 d inhalation study Acute oral toxicity: LD50 rat 132 (males) and 176 (females) mg/kg -> Category 3 Acute dermal toxicity: LD50 424 (males) and 983 (female) mg/kg -> Category3 Acute inhalation toxicity: LC50 rat 0.69 mg/l -> Category 2 Corrosivity in animal experiments (Category 1)STOT-RE oral:28 d rat oral (gavage): doses 0; 1; 10; 50 mg/kgbw/d1 mg/kg: NOEL10-mg/kg: LOELIncreased liver weight (not statistically significant)Hepatic and spleenic changes.(no clear desription of severity given)Diminished RBC counts in females, yet no other changes in blood chemisry411

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012 Histopathology: in 2/10 males and 3/10 females swelling of parenchymal cells andincreased polymorphism of the hepatocyte nuclei and the nuclear cells. Theseeffects are regarded as not “significant/severe toxic effects ” 50 mg/kg: mortality (3/8 males; 3/8 females); hepato-and nephrotoxicityresponsible for mortality; no distinct hepato-and nephrotoxicity described forsurvivors Hematology: Decrease in RBC count ca. 20% and 21% in HB both in females;decrease in Hematocrite 11%. These effects are regarded as “ moderatehematotoxicity”.Conclusion for the highest dose group: severe effects.Assessment:The substance has a high acute toxicity (s. a.). Since the factor between the acute LD50and the subacute lethal dose (20 applications) is only 2-3, it can be assumed that thesubstance has a low cumulative potential. On the other hand there is a steep dose responsein the 4 week study, thus it can be concluded by interpolation that at 30 mg/kg moderatebut no “significant/severe“ toxicity could be expected; 30 mg/kg is the guidance value forCategory 1 in a 4 week study according to Habers rule: 10 mg/kg x 3 )STOT-RE inhalationIn a valid 4 week inhalation study (vapour) rats were exposed to 0.5; 5; and 25mg/m3/6h/d. 0.5 mg/m3: NOAEC for local effects in therespiratory tract 5 mg/ m3: minimal –slight focal squamous metaplasia and inflammation in thelarynx 25 mg m3: minimal –slight focal squamous metaplasia and inflammation in thelarynx 25 mg/ m3: NOAEC for systemic effects including hematology, clinicalchemistry, histopathology and neuropathology examinationsAssessment:Up to the highest concentration tested there were no systemic effects. Since the substanceis classified as corrosive an irritation of the respiratory tract by the vapour could beexpected and has been observed in minimal-slight degree at 5-25 mg/m3. It is assumedthat the irritation would increase with higher concentrations.The corrosive/irritationpotential is covered by the classification as “corrosive” Category1, thus no classificationas STOT-RE with respect to the inhalation route would result.Classification:Category 2 for the oral route is proposed since within the guidance values of 30-300 mg/kgin a 4 week study serious effect occurred. According to a total weight of evidence approach itis concluded that these significant effects would not be observed below 30 mg/kg, theconcentration limit for Category 1.Classification via the inhalation route is not warranted, since at the highest concentrationtested only local effects, but no systemic effects were observed. The local effects(corrosivity/irritancy) are covered by the respective classification.Labelling:HAZARD STATEMENT: H373 MAY CAUSE DAMAGE TO LIVER AND KIDNEYTHROUGH PROLONGED OR REPEATED EXPOSURE412

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Remark. Since the substance is classified as STOT-RE via the oral route and specific toxicityhas not been conclusively excluded for the dermal route ( rather it can be expected due to highdermal absorbtion in acute toxicity, Category 3) the Hazard statement for STOT-RE in totalwithout specifying a route has to be applied based on the classification via the oral route.(See also Risk assessment report BUT-2YNE-1,4-DIOL; EC 2005)3.9.6.1.3 Example 3: XYZApplication of criteria for evaluation/classification and allocation of hazard statements withrespect to specific target organs and route of exposureAvailable information:1) Human experience: No information available2) Animal data:Key chronic toxicity data (underlined for EU classification)Type of study - EffectsNOAELppm (mg/kg/d)LOAELppm (mg/kg/d)CLP RepeatedExposure (STOT)classificationmouse, oral 28 days0, 300, 600, 1200 ppm(M: 0, 51-58, 101-115, 177-226mg/kg/d, F: 0, 59-66, 111-127, 221-281 mg/kg/d)M: no NOAELF: 300 (59-66)M: 300 (51-58)F: 600 (111-127)Category 2 basedon the effects onbloodhematological changes in M ( RBC count, Hb, Ht)rat, oral 13 weeks0, 50, 500, 1000 ppm(M: 0, 3.5, 38, 67 mg/kg/d, F: 0, 4,38, 80 mg/kg/d)hematological changes in F ( RBC count, Hb, Ht)male rat, oral 30, 60, 90 days0, 5, 10, 25 mg/kg/d (by gavage)(open literature)mortality at 5 (5/25), 10 (7/25) & 25(8/25) mg/kg50 (M: 3.5, F: 4) 500 (M: 38, F:38)Category 2 basedon the effects onbloodNo classification isproposed on basisof this studybecause the deathobserved in the 3groups are incontradiction withthe other relevantexperiments in thisspecies. (death nodose related, someanimals (2/6) arealready dead after30 days at 5 mg/kg)rat, oral 2-years0, 30, 150, 300 ppm(M : 0, 1.46, 7.31, 14.66 mg/kg/d, F: 0, 1.8, 8.86, 18.57 mg/kg/d)30 (M: 1.46, F:1.8)150 (M: 7.31, F:8.86)Category.2based onthe effects on blood(haemolyticanaemiaaccompanied by413

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012eyelid masses: 1 F/50 at 150 ppm, 5M/50 & 3 F/49 at 300 ppmchanges in erythroid parameters (RBC count, MC Hb, MCV in Fat 300 ppm)extramedullary hemopoiesis in liver(M : 150 & 300 ppm, F : 300 ppm),spleens myeloid hyperplasia in BM infemur & sternum ofF at 300 ppm i. hemorrhages w/i mesentericlymph nodes at 150 & 300 ppmrat, oral 80 weeksM: 0, 5, 20, 52 mg/kg/dF: 0, 6, 26, 67 mg/kg/d(open literature)ataxic syndrom in F at 67 mg/kg/d(unusual gait). The condition ofthese rats worsened, leading toparalysis posterior to the lumbarregion atrophy of the hing legs. Nospecific hystopathological lesion ofCNS or PNS.rat, oral, 104 weeks0, 3, 30, 300 ppm(M: 0, 0.1, 1.2, 11.6 mg/kg/d, F: 0,0.1, 1.4, 13.8 mg/kg/d)(open literature)anemia in 300 ppm (F) (not in 30ppm)regressive changes of sciatic nerve(degeneration) + atrophy of calfmuscle in F at 300 ppm, but noneurologcal signsprogression of myocardial lesions at300 ppmmouse, oral, 97/98 weeksM : 0, 15, 150, 300 ppm ( 0, 3, 24,50 mg/kg/d)F : 0, 15, 300, 600 ppm (0, 3, 57,112 mg/kg/d)retinal atrophy at 150 ppm ( orabsence of outer nuclear cell layerof retina) turnover of erythrocytes41415(M: 5.2, F: 3.1)compensatorymechanisms)No classification(effects above thecut-off values)Category 2 basedon the effects onblood and nervoussystemCategory 2 basedon the effects onblood.Category 2 basedon the effects on theretina

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Classification for XYZ : STOT-RE Category 2Labelling :Symbol: GHS08Signal word: WarningHazard statement: H373 May cause damage to the blood and nervous systems throughprolonged or repeated exposure.Justification :The effects on blood are reported in the 2 species (mouse, rat), at doses low enough to justifyCategory 2. The effects on NS are reported in the rat at doses low enough to justify Category2.3.9.6.2 Examples of substances not fulfilling the criteria forclassification3.9.6.2.1 Example 4: MCCPs (Medium Chain Chlorinated Paraffins) = Alkanes, C 14-17, Chloro- (EC No 287-477-0; CAS No 85535-85-9)Application of criteria for evaluation/classification with regard to mechanisms not relevant tohumans (see Section 3.9.2.5.3)Available information:1) Human experience: No information available2) Animal data: see summaryKey chronic toxicity data: Summary of data for repeated exposureThe only available data relate to a number of oral dosing studies (up to 90 days duration) that haveinvestigated the repeated dose toxicity of MCCPs (C 14-17 , 40% or 52% chlorinated paraffins) inrodents. However, only two studies emerge as providing helpful dose-response information inrespect of classification and labelling (IRDC 1984, Poon et al 1995). The others, all presented inmore detail in the ESR RAR, were generally mechanistic studies on the interplay between liver andthyroid and the relevance of effects on these organs to human health, conducted at relatively highexposure levels.In rats, the liver, thyroid and kidney are the target organs for repeated dose toxicity of MCCPs.For the liver, increases in weight and changes in enzyme activity are seen in rats at exposure levelsof 36 mg/kg/day or more (Poon et al 1995). These effects are considered part of an adaptiveresponse to an increase in metabolic demand. There is also the possibility that peroxisomeproliferation plays a role. These findings were not considered to justify classification. At higherexposure levels (around 360 mg/kg/day), single cell necrosis was observed in rats (Poon et al1995), but this is above the cut-off level for classification.Increased thyroid weight was observed in a 90-day study only at the highest exposure level tested,625 mg/kg/day (IRDC 1984). Histopathologically, lesions such as hyperplasia have been observeddown to the lowest exposure levels tested (eg. 0.4 mg/kg/day by Poon et al 1995) with an exposurerelatedincrease in severity. However, the severity only ranged from “mild” to “moderate” evenwith an increase in exposure of 3 orders of magnitude. The thyroid changes (increased weight andfollicular hypertrophy and hyperplasia) are considered to occur as a result of repeated stimulation ofthis organ caused by the well-characterised negative feedback control effect arising from plasma T 4depletion. This in turn is related to an increase in the activity of hepatic UDPG-transferase.Humans, unlike rodents, possess a T 4 binding protein that greatly reduces susceptibility to plasmaT 4 depletion and thyroid stimulation. The thyroid effects observed in rats are therefore considered415

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012of insufficient concern for classification.No adverse renal effects were seen in males and female rats at 0.4 mg/kg/day in a 90-day study(Poon et al, 1995). Inner medullary tubular dilatation was seen at 4 mg/kg / day in the kidneys offemales only. These lesions were slight, with changes increasing only marginally in severity andincidence at higher levels (up to 420 mg/kg/day for females). An exposure-related increase in theincidence and severity of a mixed population of interstitial inflammatory cells, tubular regenerationand minimal degenerative changes in the tubular epithelium was seen in treated males and femalesat 10 mg/kg/day or more. At 10 mg/kg/day the severity of these changes was graded as ‘trace’, andeven at the highest exposure level, 625 mg/kg/day it was only ‘mild’. As the effects observed in thehighest dose group do not seem to be severe, no classification is proposed for repeated-exposureeffects.Mechanistic studies conducted using short-chain chlorinated paraffins (SCCPs, C 10-13 ) indicatedeposition of β2μ globulin in proximal convoluted tubules and this may be the primary mechanismfor renal toxicity in male rats.Classification for MCCP’s: No classification for STOT-REJustification:Effects on the liver: the effects justifying the classification (necrosis) are above the cut-offlimit values.Effects on the thyroid: the effects observed are specific for the rat and do not justifyclassification.Effects on the kidneys: the data are not detailed enough to have an idea what are effectivelythe effects around the cut-off values (10-100 mg/kg) instead of 50 mg/kg (DSD cut-off value)but probably we could come to the same conclusion, i.e. the effect is not enough to justify theclassification in any category.3.9.6.3.1 Example 5:4163.9.6.3 Examples of mixtures fulfilling the criteria forclassificationApplication of criteria for mixture classification: 'When data are available for the completemixture' (see Section 3.9.3.3).Available information:A mixture with a suspect ingredient (8%) has been tested in a valid 90-day oral studyaccording to TG OECD 408 and GLP. At the dose of 90 mg/kg bw/day severe liver damage(necrosis) has been observed, at 30 mg/kg bw/day slight-moderate liver impairment. TheNOAEL was 9 mg/kg bw/day.Classification: STOT-RE Category 2Justification:The classification is based on data of a valid, appropriate animal study for the completemixture. Therefore the criteria for substances (CLP Annex I, Table 3.9.3) are applied.3.9.6.3.2 Example 6Application of criteria for mixture classification: 'When data are available for all components'(Section 3.9.3.3). Components of a mixture that should be taken into account are listed belowtogether with their concentrations. Generic concentration limits should be used, non-additivityis applied.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Available information:Ingredient % w/w Classification1 39 NC2 5.5 STOT-RE Category13 54 NC4 1.5 STOT-RE Category 2Classification of the mixture: STOT-RE Category 2Justification:No test data with respect to STOT-RE are available for the complete mixture. Bridgingprinciples can not be applied since no respective test data on a similar mixture are available.The classification of the mixture will be based on the classified ingredients (CLP Annex I,Table 3.9.4).There is one Category 1 ingredient in a concentration of

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012A Cat 1 0.3 SCL 0.2%C Cat 2 9Classification of the mixture: Category 1 since the concentration of A, even if being lowerthan the generic concentration limit, is higher than the SCL; C does not contribute to theclassification.3.9.6.4.1 Example 93.9.6.4 Example of mixtures not fulfilling the criteria forclassificationApplication of criteria for mixture classification: 'When data are available for all components'(Section 3.9.3.3); components of a mixture that should be taken into account are listed belowtogether with their concentrations. Generic concentration limits should be used, non-additivityis applied:Available information:IngredientConcentration(% w/w)1 39 NCClassification2 9 STOT-RE Category23 49.5 NC4 2.5 STOT-RE Category 2Classification of the mixture: NC (no classification).Justification:No test data with respect to STOT-RE are available for the mixture as a whole. Bridgingprinciples can not be applied, since no respective test data on a similar mixture are available(CLP Annex I, Table 3.9.4).The classification of the mixture is based on the classified ingredients. No ingredient isclassified in Category 1. Therefore the mixture cannot be classified in Category 1. Though thesum of the Category 2 ingredients (11.5 %) is above the generic concentration limit of 10%,the mixture is not classified. This is because for STOT-RE the no additivity approach appliesand no individual ingredient ≥ 10% is present in the mixture.3.9.7 ReferencesMuller, A. et al (2006) Regulatory Toxicology and Pharmacology 45, 229-241418

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124 PART 4: ENVIRONMENTAL HAZARDS4.1 HAZARDOUS TO THE AQUATIC ENVIRONMENT4.1.1 IntroductionGuidance for the application of the criteria covering effects on the aquatic compartment wasdeveloped by OECD and incorporated as Annexes 9 and 10 in the “Globally HarmonisedSystem of classification and labelling of chemicals (UN GHS)” (United Nations GHS (Rev.3) 2009)).The text in this chapter, and even more so in some of the Annexes to this chapter, is largelybased on the text in UN GHS (Rev. 3, 2009). The guidance given in Annexes 9 and 10 of UNGHS relates to substances, but not mixtures. Some parts have therefore been slightly revisedto take into account recent developments and additional guidance documents provided byECHA. Furthermore guidance on the classification of mixtures has been brought into thischapter as well as classification examples for both substances and mixtures.4.1.2 ScopeAnnex I: 4.1.1.3.1 Classification of substances and mixtures for environmental hazards requires theidentification of the hazards they present to the aquatic environment. The aquatic environment isconsidered in terms of the aquatic organisms that live in the water, and the aquatic ecosystem ofwhich they are part. The basis, therefore, of the identification of acute (short-term) and long-termhazards is the aquatic toxicity of the substance or mixture, although this shall be modified by takingaccount of further information on the degradation and bioaccumulation behaviour, if appropriate.The classification scheme has been developed with the objective of identifying thosechemicals that present, through their intrinsic properties, a hazard to the aquatic environmentcovering the aquatic freshwater and marine ecosystems. For most substances, the majority ofdata available addresses this environmental compartment. The classification scheme is limitedin scope in that it does not, as yet, include aquatic sediments, nor higher organisms at the topend of the aquatic food-chain, although these may to some extent be covered by the criteriaselected.Although limited in scope, it is widely accepted that this compartment is vulnerable, in that itis the receiving environment for many harmful substances, and the organisms that live therecan be very sensitive. It is also complex since any system that seeks to identify hazards to theenvironment must seek to define those effects in terms of wider effects on ecosystems ratherthan on individuals within a species or population. However, for practical reasons a limitedset of specific properties has been selected through which the acute (short-term) and longtermhazards, can be best described: acute aquatic toxicity; chronic aquatic toxicity; lack ofrapid degradability; and potential or actual bioaccumulation. Relevant definitions for aquatichazard classification of substances i.e. acute and/or chronic aquatic toxicity, availability andbioavailability to the aquatic environment are outlined in the CLP Regulation, Annex I,Section 4.1.1.1. Some further guidance can be viewed in the IR/CSA 51 , Chapter B.6.3. Therationale for the selection of these properties as the means to define the aquatic hazard will bedescribed in more detail in the following sections of this guidance.51 IR/CSA … Guidance on Information Requirements and Chemical Safety Assessment (ECHA, 2008).419

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3 Classification of substances hazardous to the aquatic environment4.1.3.1 Information applicable for classification of substances hazardous to theaquatic environment4.1.3.1.1 Substance properties used for classificationGenerally speaking, in deciding whether a substance should be classified, a search ofappropriate databases and other sources of data should be made for at least the followingsubstance properties: water solubility, octanol/water partition coefficient (log K ow ), acuteaquatic toxicity (L(E)C 50 ), chronic aquatic toxicity (NOEC or equivalent EC 52 x ), degradation(evidence of rapid degradability, hydrolysis) and bioaccumulation (preferablybioconcentration factor in fish (BCF)). Other information might be considered on a case-bycasebasis.Although not used directly in the criteria, the water solubility and stability data are importantsince they are a valuable help in the data interpretation of the other properties. However,water solubility may be difficult to determine and is frequently recorded as simply being low,insoluble or less than the detection limit. This may create problems in interpreting aquatictoxicity and bioaccumulation studies (see also Annex III). Hydrolysis data (Test MethodsRegulation (EC) N o 440/2008; OECD Test guideline 111) and information on the hydrolysisproducts as well as their behaviour in water might be helpful as well. As an example, forsubstances where the degradation half-life (DT 50 ) is less than 12 hours, environmental effectsare likely to be attributed to the hydrolysis products rather than to the parent substance itself(IR/CSA, Chapter R7.8).4.1.3.1.2 Information and data availabilityAnnex I: 4.1.1.2.2 Preferably data shall be derived using the standardised test methods referred toin Article 8(3). In practice data from other standardised test methods such as national methods shallalso be used where they are considered as equivalent. Where valid data are available from nonstandardtesting and from non-testing methods, these shall be considered in classification providedthey fulfil the requirements specified in section 1 of Annex XI to Regulation (EC) No 1907/2006.In general, both freshwater and marine species toxicity data are considered suitable for use inclassification provided the test methods used are equivalent. Where such data are not availableclassification shall be based on the best available data. See also part 1 of Annex I to Regulation(EC) No 1272/2008.The data used to classify a substance can be drawn from data required for other regulatorypurposes as well as the relevant literature. A number of internationally recognised databasesexist which can act as a good starting point. Such databases vary widely in quality andcomprehensiveness and it is unlikely that any one database will hold all the informationnecessary for classification to be made. Some databases specialise in aquatic toxicity andothers in environmental fate. Information can also be gathered from data submitted underplant protection products and/or biocidal products legislation.Non-testing informationInformation derived from (Q)SAR and read-across, grouping and categorisation can also beused, see also IR/CSA, Chapter R.6.52 if available, preference is given to EC 10 , see OECD 2006420

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Information sourcesIR/CSA Chapter R.3.4.1 specifies a selection of freely available databases and databankswhich might be consulted for classification purposes. All ECHA guidance documents areavailable on the Agency’s website (http://echa.europa.eu/web/guest/support/guidance-onreach-and-clp-implementation).Data can also be found through the eChemPortal, which is a global portal to information onchemical substances. The eChemPortal provides access to a number of databases, includingthe OECD HPV (Existing Chemicals Database) and the SIDS UNEP (Screening InformationDataset for High Volume Chemicals). The eChemPortal is currently hosted by the OECD:(http://www.echemportal.org/)Further guidance is given in Annex V to this document.4.1.3.2 Evaluation of available information4.1.3.2.1 General considerationsThe term substance covers a wide range of chemicals (INS 53 , Chapter 3) many of which posechallenges to a classification system based on rigid criteria. This section will thus providesome guidance on how these challenges can be dealt with based both on experience in use andclear scientific rationale.The range of interpretational problems can be extensive and as a result such interpretation willalways rely on the ability and expertise of the individuals responsible for classification.However, it is possible to identify some commonly occurring difficulties and provideguidance. Such difficulties can fall into a number of overlapping issues:(a) The difficulty in applying the current test procedures to some types of substances;(b)The difficulty in interpreting the data derived both from these “difficult to test”substances and from other substances;(c) The difficulty in interpretation of diverse datasets derived from a wide variety ofsources (e.g. Weight of Evidence).(d) The difficulty of interpreting ‘other’ informationRegarding the use of test data, in general, only reliable information (i.e. with a Klimischreliability score of 1 (reliable without restrictions) or 2 (reliable with restrictions)) should be usedfor classification purposes. However, good quality data may not always be available for alltrophic levels. It will be necessary to consider data of lower quality for those trophic levels forwhich good quality data are not available. Consideration of such data, however, will also needto consider the difficulties that may have affected the likelihood of achieving a valid result.For larger data sets, preference should be given to information with Klimisch score 1, whileinformation with Klimisch score 2 can be used as supporting information. For more informationon the Klimisch reliability scoring system, see IR/CSA, Chapter R.4.2.4.1.3.2.2 Substances difficult to testFor many organic substances, the testing and interpretation of data present no problems whenapplying both the relevant Test Methods Regulation (EC) N o 440/2008 and/or OECD TestGuidelines and the classification criteria. There are a number of typical interpretationalproblems, however, that can be characterised by the properties of the substance being studied.These are commonly called “difficult substances”:53 INS means Guidance on Identification and Naming of substances in REACH (ECHA, 2007)421

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(a)(b)(c)(d)(e)(f)poorly soluble substances: these substances are difficult to test because they presentproblems in the preparation of a test solution, maintenance of test concentrationsand verification of exposure during aquatic toxicity testing. In addition, manyavailable data for such substances have been produced using “solutions” in excessof the water solubility resulting in major interpretational problems in defining thetrue L(E)C 50 or NOEC/EC x for the purposes of classification. Interpretation of thepartitioning behaviour can also be problematic where the poor solubility in waterand octanol may be compounded by insufficient sensitivity in the analyticalmethod. Water solubility may be difficult to determine and is frequently recordedas simply being less than the detection limit, creating problems in interpreting bothaquatic toxicity and bioaccumulation studies. In biodegradation studies, poorsolubility may result in low bioavailability and thus lower than expectedbiodegradation rates. The specific test method or the choice of procedures used canthus be of key importance;unstable substances: such substances that degrade (or react) rapidly in the testsystem present both testing and interpretational problems. It will be necessary todetermine whether the correct methodology in line with the guidance provided insection 4.1.3.3 has been used, whether it is the substance or thedegradation/reaction product that has been tested, and whether the data produced isrelevant to the classification of the parent substance;volatile substances: such substances that can clearly present testing problems whenused in open systems should be evaluated to ensure adequate maintenance ofexposure concentrations. Loss of test material during biodegradation testing isinevitable in certain methods and will lead to misinterpretation of the results;complex or multi-constituent 54 substances: such substances, for example, complexhydrocarbons , or other UVCB 55 substances, frequently cannot be dissolved into ahomogeneous solution, and the multiple components make monitoring impossible.For organics, consideration therefore needs to be given to using the data derivedfrom the testing of water-accommodated fractions (WAFs) for aquatic toxicity, andthe use of such data in the classification scheme 56 . Biodegradation,bioaccumulation, partitioning behaviour and water solubility all present problemsof interpretation, where each component of these complex or multi-constituentsubstances may behave differently;polymers: such substances frequently comprise a wide range of molecular masses,which individually might have different water solubilities. Special methods areavailable to determine the water soluble fraction and these data will need to be usedin interpreting the test data against the classification criteria;inorganic compounds and metals: such substances, which can interact with themedia, can produce a range of aquatic toxicities dependent on factors such as pH,water hardness etc. Difficult interpretational problems also arise from the testing ofessential elements that are beneficial at certain levels. For metals and inorganicmetal compounds, the concept of degradability as applied to organic compounds54 Further definitions are provided in the Guidance on Identification and Naming of Substances (INS) in REACH(ECHA, 2007).55 UVCB means Substances of Unknown or Variable composition, Complex reaction products or Biologicalmaterials, see Chapter 4.3 in INS.56 Note that the toxicity is sometimes expressed as LL 50 , related to the lethal loading level. This loading levelfrom the WSF or WAF may be used directly in the classification criteria (see also Annex I.4.5 of this guidancedocument).422

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012has limited or no meaning. Equally the use of bioaccumulation data should betreated with care (see also Annex IV);(g) surface active substances: such substances can form emulsions in which thebioavailability is difficult to ascertain, even with careful preparation of solutions.Micelle formation can result in an overestimation of the bioavailable fraction evenwhen “solutions” are apparently formed. This presents significant problems ofinterpretation in each of the water solubility, partition coefficient, bioaccumulationand aquatic toxicity studies;(h) ionisable substances: such substances can change the extent of ionisation accordingto the level of counter ions in the media. Acids and bases, for example, will showradically different partitioning behaviour depending on the pH;(i) coloured substances: such substances can cause problems in the algal/aquatic planttesting because of the blocking of incident light;(j) impurities: some substances can contain impurities that can change in percentageand in chemical nature between production batches. Interpretational problems canarise where either or both the toxicity and water solubility of the impurities aregreater than the parent substance, thus potentially influencing the toxicity data in asignificant way. In general, the substance as manufactured including impuritiesshould be tested and the classification should be based on these test results. Toassess the sameness of two substances containing the same impurity in differentamount see INS, Chapter 5;(k) essential substances: some substances are essential to life, even though, like anysubstance, excessive concentrations can be harmful. This can lead to complexconcentration/dose-response curves;(l) substances which can chelate or sequester essential elements, leading to the sameproblems of interpretation as in (k).For further details see the OECD Guidance Document on aquatic toxicity testing of difficultsubstances and mixtures (OECD 2000) and also the IR/CSA Guidance, Chapter R.7b,Appendix 7.8.1 and Annex I to this guidance.4.1.3.2.3 Interpretation of data for aquatic toxicity, degradation andbioaccumulation4.1.3.2.3.1 Aquatic toxicityAnnex I: 4.1.2.7.1 Acute aquatic toxicity is normally determined using a fish 96 hour LC 50 , acrustacea species 48 hour EC 50 and/or an algal species 72 or 96 hour EC 50 . These species cover arange of trophic levels and taxa and are considered as surrogate for all aquatic organisms. Data onother species (e.g. Lemna spp.) shall also be considered if the test methodology is suitable. Theaquatic plant growth inhibition tests are normally considered as chronic tests but the EC 50 s aretreated as acute values for classification purposes (see note 2).4.1.2.7.2 For determining chronic aquatic toxicity for classification purposes data generatedaccording to the standardised test methods referred to in Article 8(3) shall be accepted, as well asresults obtained from other validated and internationally accepted test methods. The NOECs orother equivalent EC x (e.g. EC 10 ) shall be used.Fish, crustacea and algae or other aquatic plants are tested as surrogate species representing arange of trophic levels and taxa, and the test methods are highly standardised (see Annex I forfurther details). Valid data for short- and long-term tests on other species at the same trophic423

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012level shall also be considered, provided they are equivalent in terms of species relevance,testing conditions and test endpoints.The purpose of classification is to characterise both the acute and long-term hazards in theaquatic environment. The acute and long-term hazards represent distinct types of hazard andshould be applied independently.The lowest available toxicity value(s) between and within the different trophic levels (fish,crustacea, algae/aquatic plants) will normally be used to define the appropriate hazardcategory(ies), although there may be circumstances where a weight of evidence approach isrequired (see section 4.1.3.2.4).Care should be taken when classifying substances like ionisable organic chemicals or organometallicsubstances as the observed results may express different toxicities in freshwater andmarine environments and/or poorly soluble substances (water solubility < 1 mg/l), wherethere is evidence that the acute test does not provide a true measure of the intrinsic toxicity.Relevant descriptions of the type of acute and/or chronic aquatic toxicity tests have beenoutlined in detail in Annex I to this guidance and in IR/CSA, Sections R.7.8.3-R.7.8.4. Forclassification and labelling purposes, tests using organisms outside the specified size(generally smaller) and/or tests with a differing test duration could be used if no otheracceptable data are available.Currently in vitro studies are only validated for some human health endpoints and accordingto IR/CSA, Chapters R.7.8.3-R.7.8.4, there are currently no validated fish cell systemsavailable for use as alternative data to determine acute and long-term hazards within the scopeof classification and labelling.4.1.3.2.3.2 DegradationAnnex I: 4.1.2.9.1 Substances that rapidly degrade can be quickly removed from the environment.While effects of such substances can occur, particularly in the event of a spillage or accident, theyare localised and of short duration. In the absence of rapid degradation in the environment asubstance in the water has the potential to exert toxicity over a wide temporal and spatial scale.4.1.2.9.2 One way of demonstrating rapid degradation utilises the biodegradation screening testsdesigned to determine whether an organic substance is "readily biodegradable". Where such dataare not available, a BOD(5 days)/COD ratio ≥ 0,5 is considered as indicative of rapid degradation.Thus, a substance which passes this screening test is considered likely to biodegrade "rapidly" inthe aquatic environment, and is thus unlikely to be persistent. However, a fail in the screening testdoes not necessarily mean that the substance will not degrade rapidly in the environment. Otherevidence of rapid degradation in the environment may therefore also be considered and are ofparticular importance where the substances are inhibitory to microbial activity at the concentrationlevels used in standard testing. Thus, a further classification criterion is included which allows theuse of data to show that the substance did actually degrade biotically or abiotically in the aquaticenvironment by > 70 % in 28 days. Thus, if degradation is demonstrated under environmentallyrealistic conditions, then the criterion of "rapid degradability" is met.The definition of degradation covers both biotic (biodegradation) and abiotic degradationprocesses. Data on degradation properties of a substance may be available from standardisedtests, from other types of investigations, or they may be estimated from the structure of themolecules (see section 1.4). In section II.2 of Annex II to this guidance a general overview ofrelevant definitions on how to use different (bio)degradability tests and guidance for theinterpretation of test data in the context of classification and labelling is given. Additionalinformation on (bio)degradation testing methods can be found in IR/CSA, Chapter R.7.9. TheOECD test methods 301A-F (C.4-A to F of the Test Methods Regulation 440/2008),424

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012OECD310, or equivalent tests, are commonly used to determine ‘ready biodegradability’.Some guidance on the use of QSAR methods for degradability is presented in IR/CSA,Chapter R.7.9.3.1.The paragraphs below will focus on the guidance for using degradability data forclassification & labelling under CLP. It should be noted that the guidance on degradabilitypertains primarily to individual substances. In the case of complex or multi-constituentsubstances, the proposed test approaches do not normally allow an unequivocal interpretationof the degradability of the individual components of the substances. Thus, results ofbiodegradability tests on complex or multi-constituent substances should be carefullyevaluated before use for classification purposes is considered.Annex I: 4.1.2.9.3 Many degradation data are available in the form of degradation half-lives andthese can be used in defining rapid degradation provided that ultimate biodegradation of thesubstance, i.e. full mineralisation, is achieved. Primary biodegradation does not normally suffice inthe assessment of rapid degradability unless it can be demonstrated that the degradation productsdo not fulfil the criteria for classification as hazardous to the aquatic environment.4.1.2.9.4 The criteria used reflect the fact that environmental degradation may be biotic or abiotic.Hydrolysis can be considered if the hydrolysis products do not fulfil the criteria for classification ashazardous to the aquatic environment.4.1.2.9.5 Substances are considered rapidly degradable in the environment if one of the followingcriteria holds true:(a)(b)(c)if, in 28-day ready biodegradation studies, at least the following levels of degradation areachieved:(i) tests based on dissolved organic carbon: 70 %;(ii)tests based on oxygen depletion or carbon dioxide generation: 60 % of theoreticalmaximum.These levels of biodegradation must be achieved within 10 days of the start of degradationwhich point is taken as the time when 10 % of the substance has been degraded; unless thesubstance is identified as an UVCB or as a complex, multi-constituent substance withstructurally similar constituents. In this case, and where there is sufficient justification, the10-day window condition may be waived and the pass level applied at 28 days, orif, in those cases where only BOD and COD data are available, when the ratio ofBOD 5 /COD is 0,5; orif other convincing scientific evidence is available to demonstrate that the substance can bedegraded (biotically and/or abiotically) in the aquatic environment to a level > 70 % withina 28-day period.The following decision scheme may be used as a general guidance to facilitate decisions inrelation to rapid degradability in the aquatic environment and classification of chemicalshazardous to the aquatic environment.A substance is considered to be not rapidly degradable unless at least one of the following isfulfilled:(a)The substance is demonstrated to be readily biodegradable in a 28-day test for readybiodegradability. The pass level of the test (70 % DOC removal or 60 % theoreticaloxygen demand) must be achieved within 10 days from the onset of biodegradation, ifit is possible to evaluate this according to the available test data (the ten-day windowcondition may be waived for complex multi-component substances and the pass levelapplied at 28 days, as discussed in point II.2.3 of Annex II to this document). If this is425

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012not possible, then the pass level should be evaluated within a 14 days time window ifpossible, or after the end of the test; or(b) The substance is demonstrated to be ultimately degraded in a surface water simulationtest with a half-life of < 16 days (corresponding to a degradation of >70 % within 28days); or(c) The substance is demonstrated to be primarily degraded biotically or abiotically e.g.via hydroysis, in the aquatic environment with a half-life 70 % within 28 days), and it can be demonstrated that the degradationproducts do not fulfill the criteria for classification as hazardous to the aquaticenvironment.When these preferred data types are not available rapid degradation may be demonstrated ifone of the following criteria is justified:(d) The substance is demonstrated to be ultimately degraded in an aquatic sediment or soilsimulation test with a half-life of < 16 days (corresponding to a degradation of > 70 %within 28 days); or(e) In those cases where only BOD 5 and COD data are available, the ratio of BOD 5 /CODis greater than or equal to 0.5. The same criterion applies to ready biodegradabilitytests of a shorter duration than 28 days, if the half-life furthermore is < 7 days; or(f) A weight of evidence approach based on read-across provides convincing evidencethat a given substance is rapidly degradable.If none of the above types of data are available then the substance is considered as not rapidlydegradable. This decision may be supported by fulfilment of at least one of the followingcriteria:(i) the substance is not inherently degradable in an inherent biodegradabilitytest; or(ii) the substance is predicted to be slowly biodegradable by scientificallyvalid QSARs, e.g. for the Biodegradation Probability Program, the scorefor rapid degradation (linear or non-linear model) < 0.5; or(iii) the substance is considered to be not rapidly degradable based on indirectevidence, such as knowledge from structurally similar substances; or(iv) no other data regarding degradability are available.The percentage degradation reached after 28 days in ready biodegradability tests may be useddirectly for the assessment of ‘rapid degradability’ if no specific information on the timewindow is available or if the data were derived with the MITI 1 test (OECD 301C, 2006 orC.4-E of the Test Methods Regulation 440/2008). In the Closed Bottle test (OECD 301D, orC.4-F of the Test Methods Regulation 440/2008) a 14-day window may be used whenmeasurements have not been made after 10 days. For some industrial chemicals that in termsof composition can be seen as multi-component substances testing for ‘readybiodegradability’ can lead to interpretational problems (see Annex II to this guidance).Selection of test systemsAs regards paragraph 4.1.2.9.5 point c in Annex I to CLP, the evaluation of the fulfilment ofthis criterion should be conducted on a case-by-case basis by expert judgement. Test systemsthat can be used to demonstrate the occurrence of rapid degradability are listed in Annex II.This includes e.g. simulation tests under realistic conditions, mesocosms and field monitoring.426

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Inherent- (OECD 302A and B, or C.9 and C.12 of the Test Methods Regulation 440/2008)and sewage treatment simulation (OECD 303, or C.10 of the Test Methods Regulation440/2008) tests are not normally used in this context, due to the high levels of adaptedbiomass. Anaerobic degradation tests (OECD 311/ISO 11734 and analogous tests) do notqualify because of the specificity of the anaerobic compartments. Also the newly definedcategory of ‘Enhanced Ready Biodegradation (Screening) Tests’ in IR/CSA, Chapter R.7.9 donot qualify for use in classification and labelling, as they are presently not reviewed andinternationally standardised.Use of SARs and QSARsThe estimation of degradation via SARs and/or QSARs for hydrolysis and biodegradation is arapidly developing field. The predictions from QSAR models may be considered ascontributing to a decision on ready or rapid degradation for classification purposes. QSARmodels should be used with great care, taking into account the applicability domain andvalidation of the models. Current practice is to use the outcome of these biodegradationmodels to predict that a substance is not readily degradable, rather than vice versa. This isbecause models such as BIOWIN tend to predict non-biodegradability more accurately thanbiodegradability. However, QSAR information can be used as a part of expert judgement andWeight of Evidence practices, for example where very consistent measured and predicted dataare available for a structurally analogous compound.General interpretation problems and substances difficult to testBoth the UN GHS Annex 9 and the INS discuss substances that are inherently difficult to testfor biodegradability, and possible adjustments to overcome testing problems. Testing orinterpretational problems may occur with e.g. complex multi-constituent substances, surfaceactive agents, highly volatile or insoluble substances, substances that are toxic to microorganismsat normal test concentrations, and unstable molecules.4.1.3.2.3.3 BioaccumulationAnnex I: 4.1.2.8.1 Bioaccumulation of substances within aquatic organisms can give rise to toxiceffects over longer time scales even when actual water concentrations are low. For organicsubstances the potential for bioaccumulation shall normally be determined by using theoctanol/water partition coefficient, usually reported as a log K ow . The relationship between the logK ow of an organic substance and its bioconcentration as measured by the bioconcentration factor(BCF) in fish has considerable scientific literature support. Using a cut-off value of log K ow 4 isintended to identify only those substances with a real potential to bioconcentrate. While thisrepresents a potential to bioaccumulate, an experimentally determined BCF provides a bettermeasure and shall be used in preference if available. A BCF in fish of ≥ 500 is indicative of thepotential to bioconcentrate for classification purposes. Some relationships can be observed betweenchronic toxicity and bioaccumulation potential, as toxicity is related to the body burden.The potential for bioaccumulation is an important criterion to determine whether a chemicalsubstance is a potential hazard to the environment. Bioaccumulation of a substance into anorganism is not a hazard in itself, but should be considered in relation to potential long-termeffects. Chemical concentration and accumulation may result in internal concentrations of asubstance in an organism (body burden), which may or may not lead to toxic effects overlong-term exposures. Further guidance on bioaccumulation is given in Annex III to thisguidance. Bioaccumulation of metals is discussed in Annex IV.Information on actual bioaccumulation of a substance may be available from standardisedtests (e.g. Test Methods Regulation (EC) N o 440/2008, OECD 305: Bioconcentration – Flowthrough fish test) or information on the bioaccumulation potential, for organic substances,may be estimated from the structure of the molecule.427

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In general, the potential of an organic substance to bioconcentrate is primarily related to thelipophilicity of the substance. A surrogate measure of lipophilicity is the n-octanol/waterpartition coefficient (K ow ) which, for lipophilic non-ionised organic substances, undergoingminimal metabolism or biotransformation within the organism, is correlated with thebioconcentration factor. Therefore, K ow is often used for estimating the bioconcentration ofnon-ionised organic substances, based on the empirical relationship between log BCF and logK ow . For those organic substances, estimation methods are available for calculating the K ow .Data on the bioconcentration properties of non-ionised organic substances may thus be1. experimentally determined2. estimated from experimentally determined K ow , or3. estimated from K ow values derived by use of Quantitative Structure ActivityRelationships (QSARs)Experimentally derived BCF values of high quality are ultimately preferred for classificationpurposes. BCF results from poor or questionable quality studies should not be used forclassification purposes if high quality data on log K ow are available. If no BCF is available forfish species, high quality data on the BCF for some invertebrates (e.g. blue mussel, oysterand/or scallop) may be used as a worst case surrogate.For non-ionised organic substances, experimentally derived high quality K ow values arepreferred. If no experimental data of high quality are available validated QuantitativeStructure Activity Relationships (QSARs) for log K ow may be used in the classificationprocess. If data are available but not validated, expert judgement should be used. For ionisedorganic substances problems may occur with e.g. changes in pH which may significantlyaffect the water solubility and partition coefficient of the substance. Further guidance on howto deal with such difficulties is provided in the OECD Guidance Document on aquatic toxicitytesting of difficult substances and mixtures (OECD 2000).428

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.2.4 Using weight of evidence in evaluations in the context of C&L4.1.3.2.4.1 General aspects of weight of evidenceThe weight of evidence approach is described in IR/CSA, Chapter B.4.4 as follows: “Theweight of evidence (WoE) approach is not a scientifically well-defined term or an agreedformalised concept. It involves assessing the relevance, reliability and adequacy of each pieceof available information, holding the various pieces of information up against each other andreaching a conclusion on the hazard. This process always involves expert judgement. It isimportant to document and communicate how the evidence-based approach was used in areliable, robust and transparent manner.”Where there is only one experimental data entry per endpoint, classification and labellingdecisions are relatively straightforward. However this is often not the case when dealing withdata deficient substances or substances for which more than one valid piece of data isavailable for a given data element. In both situations, available information needs to beevaluated carefully. Data deficiency may occur for substances for which there are no, orlimited experimental data with relevance for classification and labelling. This might be thecase for substances exempted from REACH such as polymers or substances manufactured inquantities < 1 tonne/annum.The taxa chosen, fish, crustacea and aquatic plants that represent the “base-set” in mosthazard profiles, represent a minimum dataset for a fully valid description of hazard. Thelowest of the available toxicity values will normally be used to define the hazard category.Given the wide range of species in the environment, the three taxa tested can only be a poorsurrogate and the lowest value is therefore taken for precautionary reasons to define thehazard category. In doing so, it is recognised that the distribution of species sensitivity can beseveral orders of magnitude wide, and that there will thus be both more and less sensitivespecies in the environment. Therefore, when data are limited, the use of the most sensitivespecies tested gives a cautious but acceptable definition of the hazard. There are somecircumstances where it may not be appropriate to use the lowest toxicity value as the basis forclassification. This will usually only arise where it is possible to define the sensitivitydistribution with more accuracy than would normally be possible, such as when large datasetsare available. Such large datasets should be evaluated with due caution.Conversely, as CLP allows the use of expert judgment in employing non-testing informationsuch as QSARs, the classification of data deficient substances could potentially be conductedin the absence of any experimental data.In applying the WoE approach, the reliability of the experimental information underevaluation needs to be taken into due account. Typically, this information originates fromstudies which have been ranked according to the Klimisch criteria. The scores assigned to thestudies may serve as an indication of the ‘weight’ that the corresponding information couldhave in ‘weighing the evidence’.4.1.3.2.4.2 Guidance on WoE for data deficient substancesEither for those substances for which the standard data set of acute aquatic testing in fish,crustacea and algae/aquatic plants is not available or where there are data gaps, REACHintroduces the concept of an “Integrated Testing Strategy” (for further guidance see IR/CSA,Chapter R.7B, Figure R.7.8-2). This outlines a stepwise approach on the use of test data andnon-testing information, such as reliable QSARs and in vitro testing. It outlines how therelevant information is collected and evaluated and in the final step, expert judgement is usedto reach an overall assessment of the aquatic toxicity of the substance under evaluation, takinginto consideration also metabolites, reaction products, analogues.429

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012For classification purposes, representative species should be chosen which cover a range oftrophic levels and taxonomic groups, namely fish, crustacea and primary producers. Annex Ito this document also provides guidance on the following where no experimental data areavailable:QSARs can be relied upon to provide predictions of acute toxicity to fish, crustacea(Daphnia and Mysid) and algae for non-electrolytes, non-electrophilic, and otherwisenon-reactive substances. Care should be taken when evaluating the toxicity of poorlywater soluble substances, where the quoted toxicity may be greater than the watersolubility.4.1.3.2.4.3 Guidance on WoE for substances for which more than one valid piece of datais available for a given data elementThe best quality data should be used as the fundamental basis for classification. Classificationshould preferably be based on primary data sources. It is essential that test conditions beclearly and completely articulated.Where multiple studies for a taxonomic group are available, all studies that are assessed tohave sufficient quality should be taken into consideration. The study showing the highesttoxicity (e.g. the one with the lowest L(E)C 50 or NOEC or EC x ) should normally be chosen askey study for aquatic hazard classification for that taxonomic group. However, in a WoEapproach, a different weight may be given to studies irrespective the test results. For example:a judgement has to be made on a case-by-case basis whether Klimish 1 studies in a dataset aregiven more weight than Klimish 2 studies or valid QSAR data available for the sametaxonomic group.Lower quality information showing no or low toxicity should specifically be treated with care,especially where the quality assessment has revealed points of concern regardingmethodology and reporting (e.g. maintenance of test concentrations). In addition it should benoted that substances which are difficult to test may yield apparent results that are notindicating the true toxicity. Expert judgement would also be needed for classification in thesecases.Assessment of data quality includes assessment of adequacy of the information forclassification purposes and an assessment of both relevance and reliability. Details on theassessment of quality can be found in IR/CSA, Chapter R.4.Where more than one acceptable test is available for the same taxonomic group, the mostsensitive (the one with the lowest L(E)C 50 or NOEC/EC 10 ) is generally used for classification.However, this must be dealt with on a case-by-case basis. When larger data sets (four or morevalues) are available for the same species, the geometric mean of toxicity values may be usedas the representative toxicity value for that species. In estimating a mean value, it is notadvisable to combine tests of different species within a taxonomic group or in different lifestages or tested under different conditions or duration. This implies that for substances, wherefour or more ecotoxicity data on the same species and endpoint are available, the data shouldbe grouped, and the geometric mean used as a representative toxicity value for that species.In case of very large data sets meeting the criteria for applying the Species SensitivityDistribution (SSD) approach (see IR/CSA, Chapter R.10), statistical techniques (e.g. HC 5derivation) can be considered to estimate the aquatic toxicity reference value for classification(equivalent to using the lowest EC 50 or NOEC), in a weight of evidence approach.4.1.3.2.4.4 OutliersThe WoE approach would also address potential outliers, since as a starting point, all datapoints for a specific trophic level/taxonomic group would be considered to come from the430

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012same sensitivity distribution. Only if a sufficiently large number of data were available,appropriate statistical tests would be performed to confirm or disprove a particular value as anoutlier.The issue of possible ‘outliers’, which may exist, particularly in large data sets can be tackledaccording to a proposal in IR/CSA, Chapter R.7.8.4.1.4.1.3.2.4.5 Weight of evidence in degradationWhere multiple or conflicting datasets exist for a single chemical, the most reliable datashould be selected first, and subsequently a “weight of evidence” approach followed based onthese data. This implies that if both positive (i.e. above the pass level) and negative results(below pass level) have been obtained for a substance in rapid degradability tests, then thedata of the highest quality and the best documentation should be used for determining therapid degradability of the substance. Thus, given the conservative nature of readybiodegradability tests positive results could be used irrespective of negative results when thescientific quality is good and the test conditions are well documented, i.e. the guidelinecriteria are fulfilled. See Annex II for further guidance.4.1.3.2.4.6 Weight of evidence in bioaccumulationWhen conflicting bioaccumulation data is available, see Annex III for guidance.4.1.3.3 Classification categories and criteria4.1.3.3.1 Outline of the core classification systemAnnex I: 4.1.2.2. The core classification system for substances consists of one acute hazardclassification category and three long-term hazard classification categories. The acute and the longtermhazard classification categories are applied independently.Annex I: 4.1.2.3. The criteria for classification of a substance in category Acute 1 are defined on thebasis of acute aquatic toxicity data only (EC 50 or LC 50 ). The criteria for classification of a substanceinto the categories Chronic 1 to 3 follow a tiered approach where the first step is to see if availableinformation on chronic toxicity merits long-term hazard classification. In absence of adequatechronic toxicity data, the subsequent step is to combine two types of information, i.e. acute aquatictoxicity data and environmental fate data (degradability and bioaccumulation data) (see Figure4.1.1).431

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure 4.1.1Categories for substances long-term hazardousto the aquatic environmentAre thereadequate chronictoxicity data availablefor all threetrophic levels?YesClassify according to the criteria given in Table 4.1.0(b)(i)or 4.1.0(b)(ii) depending on information on rapiddegradationNoAre thereadequate chronictoxicity data availablefor one or twotrophic levels?NoYesAssess both:(a) according to the criteria given in Table 4.1.0(b)(i) or4.1.0(b)(ii) (depending on information on rapiddegradation), and(b) (if for the other trophic level(s) adequate acute toxicitydata are available) according to the criteria given inTable 4.1.0(b)(iii),and classify according to the most stringent outcomeAre thereadequate acutetoxicity dataavailable?YesClassify according to the criteria given in Table 4.1.0(b)(iii)4.1.2.1. The system for classification recognises that the intrinsic hazard to aquatic organisms isrepresented by both the acute and long-term hazard of a substance. For the long-term hazardseparate hazard categories are defined representing a gradation in the level of hazard identified. Thelowest of the available toxicity values between and within the different trophic levels (fish,crustacean, algae/aquatic plants) shall normally be used to define the appropriate hazardcategory(ies). There are circumstances, however, when a weight of evidence approach isappropriate.Where adequate chronic toxicity data exist for the three trophic levels and the lowest chronictoxicity value (that normally would define the appropriate hazard category) is below or equalto 1 mg/l, a long-term hazard classification is warranted. The actual category is alsodepending on the information on rapid degradation.While recognising that for packaged goods the long-term hazard represents the principalconcern, it must also be recognised that chronic toxicity data are expensive to generate andgenerally not readily available for most substances. On the other hand, acute toxicity data aremore often readily available than chronic toxicity data, or can be generated according tohighly standardised test protocols. It is this acute toxicity which has therefore been used as thecore property in defining both the acute and the long-term hazard if no adequate chronic testdata are available. Nevertheless, it has been recognised that chronic toxicity data, if available,should be preferred in defining the long-term hazard category.432

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Chronic toxicity data (ECx or NOEC) would normally override acute data for long-termhazard classification. However, when assessing the adequacy there may be some cases (suchas data poor substances) where the chronic data do not represent the species that is consideredthe most sensitive in available short-term tests. In such cases the classification should bebased on the data (acute or chronic) that gives the most strict classification and M-factor.The combination of chronic toxicity and degradation properties reflects the potential hazard ofa substance. Substances that do not rapidly degrade have a higher potential for longer termexposures and therefore should be classified in a more severe category than substances whichare rapidly degradable.A review of the existing adequate appropriate acute toxicity data and environmental fate data(degradability and bioaccumulation) is required for those trophic levels where adequatechronic toxicity data may be absent; to decide if a long-term hazard classification may bewarranted.While recognising that acute toxicity itself is not a sufficiently accurate predictor of chronictoxicity to be used solely and directly for establishing hazard, it is considered that, incombination with either a potential to bioaccumulate (i.e. experimentally determined BCF 500 or, if absent, the log K ow 4) or potential longer term exposure (i.e. lack of rapiddegradation) it can be used as a suitable surrogate for classification purposes. Substancesrapidly degrading that show acute toxicity with a significant degree of bioaccumulation willnormally show chronic toxicity at a significantly lower concentration. Equally, substancesthat do not rapidly degrade have a higher potential for giving rise to longer term exposureswhich again may result in long-term toxicity being realised.The hazard categories for acute and chronic aquatic toxicity and their related criteria are setout in CLP, Annex I, Section 4.1, Table 4.1.0.Annex I: Table 4.1.0Classification categories for hazardous to the aquatic environment(a) Acute (short-term) aquatic hazardCategory Acute 1: (Note 1)96 hr LC 50 (for fish) 1 mg/l and/or48 hr EC 50 (for crustacea) 1 mg/l and/or72 or 96 hr ErC 50 (for algae or other aquatic plants) 1 mg/l. (Note 2)433

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(b) Long-term aquatic hazard(i) Non-rapidly degradable substances (Note 3) for which there are adequate chronic toxicitydata availableCategory Chronic 1: (Note 1)Chronic NOEC or EC x (for fish)Chronic NOEC or EC x (for crustacea)Chronic NOEC or EC x (for algae or other aquatic plants)0,1 mg/l and/or0,1 mg/l and/or0,1 mg/l.Category Chronic 2:Chronic NOEC or EC x (for fish)Chronic NOEC or EC x (for crustacea)Chronic NOEC or ECx (for algae or other aquatic plants)1 mg/l and/or1 mg/l and/or1 mg/l.(ii) Rapidly degradable substances (Note 3) for which there are adequate chronic toxicitydata availableCategory Chronic 1: (Note 1)Chronic NOEC or EC x (for fish)Chronic NOEC or EC x (for crustacea)Chronic NOEC or EC x (for algae or other aquatic plants)0,01 mg/l and/or0,01 mg/l and/or0,01 mg/lCategory Chronic 2:Chronic NOEC or EC x (for fish)Chronic NOEC or EC x (for crustacea)Chronic NOEC or EC x (for algae or other aquatic plants)0,1 mg/l and/or0,1 mg/l and/or0,1 mg/lCategory Chronic 3:Chronic NOEC or EC x (for fish)Chronic NOEC or EC x (for crustacea)Chronic NOEC or ECx (for algae or other aquatic plants)1 mg/l and/or1 mg/l and/or1 mg/l.434

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(iii) Substances for which adequate chronic toxicity data are not availableCategory Chronic 1: (Note 1)96 hr LC 50 (for fish) 1 mg/l and/or48 hr EC 50 (for crustacea) 1 mg/l and/or72 or 96 hr ErC 50 (for algae or other aquatic plants) 1 mg/l. (Note 2)and the substance is not rapidly degradable and/or the experimentally determined BCF ≥ 500 (or, ifabsent, the log K ow 4). (Note 3).Category Chronic 2:96 hr LC 50 (for fish) >1 to 10 mg/l and/or48 hr EC 50 (for crustacea) >1 to 10 mg/l and/or72 or 96 hr ErC 50 (for algae or other aquatic plants) >1 to 10 mg/l. (Note 2)and the substance is not rapidly degradable and/or the experimentally determined BCF ≥ 500 (or, ifabsent, the log K ow 4). (Note 3).Category Chronic 3:96 hr LC 50 (for fish) > 10 to 100 mg/l and/or48 hr EC 50 (for crustacea) > 10 to 100 mg/l and/or72 or 96 hr ErC 50 (for algae or other aquatic plants) > 10 to 100 mg/l. (Note 2)and the substance is not rapidly degradable and/or the experimentally determined BCF ≥ 500 (or, ifabsent, the log K ow 4). (Note 3).NOTE 1: When classifying substances as Acute Category 1 and/or Chronic Category 1 it isnecessary at the same time to indicate then appropriate M-factor(s) (see table 4.1.3).NOTE 2: Classification shall be based on the ErC 50 [= EC 50 (growth rate)]. Incircumstances where the basis of the EC 50 is not specified or no ErC 50 is recorded, classificationshall be based on the lowest EC 50 available.NOTE 3: When no useful data on degradability are available, either experimentallydetermined or estimated data, the substance should be regarded as not rapidly degradable.Classifications may also be made in cases where data are not available on all three trophiclevels. In these cases, the classification may be subject to further information becomingavailable. In general, all the data available will need to be considered prior to assigning aclassification. Where good quality data are not available, lower quality data will need to beconsidered. In these circumstances, a judgement will need to be made regarding the true levelof hazard. For example, where good quality data are available for a particular species or taxa,this should be used in preference to any lower quality data which might also be available forthat species or taxa. However, good quality data may not always be available for all trophiclevels. It will be necessary to consider data of lower quality for those trophic levels for which435

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012good quality data are not available. Consideration of such data, however, will also need toconsider the difficulties that may have affected the likelihood of achieving a valid result. Forexample, the test details and experimental design may be critical to the assessment of theusability of some data, such as that from hydrolytically unstable chemicals, while less so forother chemicals. Such difficulties are described further in Annex I to this guidance.Normally, the identification of hazard, and hence the classification will be based oninformation directly obtained from testing of the substance being considered. There areoccasions, however, where this can create difficulties or the outcomes do not conform tocommon sense. For example, some chemicals, although stable in the bottle, will react rapidly(or slowly) in water giving rise to degradation products that may have different properties.Where such degradation is rapid, the available test data will frequently define the hazard ofthe degradation products since it will be these that have been tested. These data may be usedto classify the parent substance in the normal way. However, where degradation is slower, itmay be possible to test the parent substance and thus generate hazard data in the normalmanner. The subsequent degradation may then be considered in determining whether an acuteor long-term hazard category should apply. There may be occasions, however, when asubstance so tested may degrade to give rise to a more hazardous product. In thesecircumstances, the classification of the parent compound should take due account of thehazard of the degradation product, and the rate at which it can be formed undernormalenvironmental conditions (for detailed information please check also the Annexes tothis guidance).4.1.3.3.2 The “safety net”4.1.2.4 The system also introduces a "safety net" classification (referred to as category Chronic 4)for use when the data available do not allow classification under the formal criteria for acute 1 orchronic 1 to 3 but there are nevertheless some grounds for concern (see example in Table 4.1.0).“Safety net” classificationAnnex I: 4.1.2.6. Table 4.1.0. continuedChronic Category 4Cases when data do not allow classification under the above criteria but there are neverthelesssome grounds for concern. This includes, for example, poorly soluble substances for which noacute toxicity is recorded at levels up to the water solubility (Note 4), and which are not rapidlydegradable in accordance with Section 4.1.2.9.5 and have an experimentally determined BCF≥ 500 (or, if absent, a log K ow 4), indicating a potential to bioaccumulate, which will beclassified in this category unless other scientific evidence exists showing classification to beunnecessary. Such evidence includes chronic toxicity NOECs > water solubility or > 1 mg/l, orother evidence of rapid degradation in the environment than the ones provided by any of themethods listed in Section 4.1.2.9.5.NOTE 4: "No acute toxicity" is taken to mean that the L(E)C 50 (s) is/are above the watersolubility. Also for poorly soluble substances, (water solubility < 1 mg/l), where there is evidencethat the acute test does not provide a true measure of the intrinsic toxicity.Category Chronic 4 is for example triggered in the following cases. For some poorly solublesubstances, which are normally considered as those having a water solubility < 1 mg/l, noacute toxicity is expressed in toxicity tests performed at the solubility limit. If for such asubstance, however, the BCF 500, or if absent, the log K ow 4 (indicating a bioaccumulatingpotential) and the substance is also not rapidly degradable, a safety netclassification, category Chronic 4 is assigned. For these types of substances the exposure436

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012duration in short-term tests may well be too short for a steady-state concentration of thesubstance to be reached in the test organisms. Thus, even though no acute toxicity has beenmeasured in a short-term (acute) test, it remains a real possibility that such non-rapidlydegradable and bioaccumulative substances may exert chronic effects, particularly since suchlow degradability may lead to an extended exposure period in the aquatic environment.The precise definitions of the core elements of this system are described in detail in AnnexesI-III to this guidance document.4.1.3.3.3 Setting an M-factor for highly toxic substances4.1.2.5 Substances with acute toxicities below 1 mg/l or chronic toxicities below 0,1 mg/l (if nonrapidlydegradable) and 0,01 mg/l (if rapidly degradable) contribute as components of a mixture tothe toxicity of the mixture even at a low concentration and shall normally be given increasedweight in applying the summation of classification approach (see Note 1 of Table 4.1.0 and4.1.3.5.5).When a substance is classified as category Acute 1 and/or category Chronic 1, (a) multiplyingfactor(s) (M-factor) has/have to be assigned (as described Article 10 of CLP). Whereappropriate, M-factors shall be set for acute and long-term hazards separately. This means thatthere can be two different M-factors (one for acute and one for long-term hazard) for onesubstance. It is important to also include the M-factor(s) in the SDS as other users in thesupply chain might need it, e.g. for classification of mixtures containing that substance.The M-factor itself can be taken from the table below and is dependent on the toxicity band ofthe substances. For a substance with an acute toxicity of 0.005 mg/l for example an M-factorof 100 needs to be assigned. Whereas e.g. with a chronic toxicity of 0.005 mg/l an M-factor of10 needs to be assigned for non-rapidly degrable substance and an M-factor of 1 to rapidlydegradable substances.Annex I: Table 4.1.3Multiplying factors for highly toxic components of mixturesAcute toxicity M factor Chronic toxicity M factorL(E)C 50 value NOEC value NRD acomponentsRD bcomponents0,1 < L(E)C 50 ≤ 1 1 0,01 < NOEC ≤ 0,1 1 -0,01 < L(E)C 50 0,1 10 0,001 < NOEC ≤ 0,01 10 10,001 < L(E)C 50 0,01 100 0,0001 < NOEC ≤ 0,001 100 100,0001 < L(E)C 50 0,001 1000 0,00001 < NOEC ≤ 0,0001 1000 1000,00001 < L(E)C 50 0,0001 10000 0,000001 < NOEC ≤ 0,00001 10000 1000(continue in factor 10 intervals)(continue in factor 10 intervals)abNon-rapidly degradableRapidly degradableThe NOEC value in Table 4.1.3 (Annex I to CLP) refers to both NOEC and EC x (toxicityvalues are in mg/l). The first two columns in Table 4.1.3 refer to the classification system inTable 4.1.0 (a)(b, point iii), the last three columns refer to the respective classification systemin Table 4.1.0 (b, points i & ii). In cases where chronic data are not available and Table 4.1.0(a)(b, point iii) is used for defining long-term aquatic hazard, the resulting M-factor derivedfor acute aquatic hazard classification is also applied to the long-term aquatic hazardclassification.437

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.4 Decision on classification: examples for substancesIf the evaluation shows that the criteria are fulfilled, one category for acute aquatic hazardand/or one for long-term aquatic hazard should be assigned, as well as (an) M-factor(s) whereapplicable. For the labelling elements, such as hazard pictograms, signal words, hazardstatements and precautionary statements, see section 4.1.6 of this guidance.Further classification examples specific to metals and metal compounds are given in AnnexIV to this guidance document.The examples in this section are focussed on self-classification based on relevant dataavailable. Mandatory use of harmonised classification for substances included in Table 3.1 ofAnnex VI, the use of information from the classification and labelling inventory and the useof the translation Table in Annex VII are not taken into account in these examples.After data collection self-classification starts with evaluation of the adequateness of the datacollected and assessment of the results and concluding on endpoints relevant forenvironmental hazard classification. Where the assessment shows that criteria forenvironmental classification are fulfilled, one category for acute aquatic hazard and/or onecategory for long-term aquatic hazards should be assigned and M-factor(s) should bededucted where applicable.List of the examples on substance classification included in this section: Example A: Hydrophilic substance, straightforward classification based on acute andchronic toxicity data; Example B: Hydrophilic substance, straightforward classification based on acute data, nochronic toxicity data available; Example C: Moderately water soluble substance, straightforward classification based onacute data, chronic toxicity data available for two trophic levels; combined set of QSARdata and experimental data; Example D: Substance with several toxicity data for one trophic level; Example E: “Safety net” classification category Chronic 4; Example F: Substance difficult to test, toxicity above level of water solubility.Further classification examples specific to metals and metal compounds are given in AnnexIV to this guidance.The examples are presented using a logical format starting with a table listing for all relevantdata elements the information available, followed by an aquatic hazard assessment for eachdata element, a section showing the aquatic hazard classification, a section with the reasoningbehind the conclusions and finally a table presenting the applicable labelling elements.438

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Explanation of data elements used in the examples:Physico-chemical properties important for evaluation of aquatic hazards for the purpose ofclassification: Generally this consists of water solubility (mg/l) and log octanol/waterpartition coefficient (log K ow );Acute aquatic toxicity: Generally expressed in terms of LC 50 or EC 50 (mg/l);Long-term aquatic toxicity: Generally expressed in terms of NOEC or EC x (mg/l);Degradation (evidence of rapid degradation): Generally expressed in terms of biotic orabiotic degradation of organic substances (or transformation of inorganic substances). Incase of rapid primary degradation, information shall be given whether the degradationproducts can be classified as hazardous to the aquatic environment or not;Bioaccumulation: Generally expressed in terms of bioconcentration factor in fishInformation on reliability is not taken into account in the exemplification. For the purpose ofthe examples the reliability score is assumed to be high (e.g. for experimental tests, Klimischscore 1 or 2) unless otherwise stated. Note that assigning a reliability score to studies isimportant - if a study is assessed as poorly reliable it is normally not usable for classificationpurposes.Besides the conclusion from studies on relevant endpoints for classification the followinginformation is presented for each example in a separate column: Referral to applicable test method according to the EU Test Methods Regulation (EC) No440/2008 or OECD test guideline or QSAR model used; Some basic information on the test design (pH of the test media, renewal regime of testmedia (static, semi-static, flow-through); Use of measured or nominal test concentrations; Compliance of the experiment and reporting with OECD Good Laboratory Practice (GLP)rules; Specific information related to the relevant endpoints, as appropriate.This information plays a crucial role when the adequacy of the data and the assessment of thestudy results are being evaluated for their applicability in the classification and labellingscheme. However, in these examples this information is included mainly to make the datamore realistic.439

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.4.1 Example A: Hydrophilic substance, straightforward classification based onacute and chronic toxicity dataDATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksPhysico-chemical propertiesWater solubility: 1200 mg/l A.6. / pH:7.0, non-GLPLog octanol/water partition coefficient (log K ow ): 2.75 A.8. / pH:7.5, GLPAcute aquatic toxicityFishOncorhynchus mykiss:12 mg/l (96 h LC 50 )C.1. / static, non-GLPLepomis macrochirus:2.7 mg/l (96 h LC 50 )C.1. / static, GLPCrustacea Daphnia magna: 18 mg/l (48 h EC 50 ) C.2. / static, non-GLPAlgae/aquatic plantsScenedesmus subspicatus:0.056 mg/l (96 h ErC 50 )C.3. / static, GLPLemna gibba:Chronic aquatic toxicity0.031 mg/l (7 d ErC 50 )C.26. / semi-static, GLPFish: Danio rerio: 1.2 mg/l (21 d NOEC) OECD 210 / Early Life Stagetoxicity test, flow-through,GLPCrustacea: Daphnia magna: 1.1 mgl (21 d NOEC) C.20. / semi-static, GLPAlgae/aquatic plants: Scenedesmus subspicatus: 0.01 mg/l (96 h NOEC) C.3. / static, GLPDegradation (evidence of rapid degradation)Biotic degradation:86 % in 28 days (10day-window fulfilled)C.4-C / pH:7.5, GLPAbiotic degradation, hydrolysis: (half-life (d)):BioaccumulationBioconcentration factor in fish (BCF)No dataNo data440

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Aquatic hazard assessment, conclusions and comments:Physico-chemical properties: The substance is readily soluble. Log K ow < 4, indicating low potential forbioaccumulation, which can be used in absence of BCF data.Acute aquatic toxicity: The acute aquatic toxicity based on the lowest of the available toxicity values is between0.01 and 0.1 mg/l.Long-term aquatic toxicity:The long-term aquatic toxicity based on the lowest of the available toxicity values isbetween 0.001 and 0.01 mg/l.Degradation (evidence of rapid degradation): > 70 % degradation in 28 days based on dissolved organic carbon (DOC) fulfils thecriteria for rapid degradation.Aquatic hazard classification and, where applicable, established M-factor(s):Acute (short-term) aquatic hazard: category Acute 1, M-factor: 10.Long-term aquatic hazard: category Chronic 1, M-factor: 1.Reasoning:Acute aquatic hazard: acute toxicity L(E)C 50 ≤ 1 mg/l. M-factor based on L(E)C 50 between0.01 and 0.1 mg/l.Long-term aquatic hazard:The criteria for classification of a substance into the categories Chronic 1 to 3 follow a tieredapproach where the first step is to see if adequate information on long-term toxicity isavailable allowing long-term hazard classification. In absence of adequate long-term toxicitydata for some or all trophic levels, the subsequent step is to combine two types of information,i.e. acute aquatic toxicity data and environmental fate data (degradability and bioaccumulationdata). For details see section 4.1.3.3 and Table 4.1.0. Adequate long-term toxicity data for all three trophic levels, long-term toxicity NOEC≤ 0.01 mg/l, rapidly degradable. M-factor based on NOEC between 0.001 and 0.01mg/l (rapidly degradable).Labelling elements based on the classification:441

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012ElementGHS PictogramSignal WordCodeGHS09WARNINGHazard Statement H410 57Precautionary statement(s)P273, P391, P50157 Note that in accordance with CLP Article 27 the hazard statement H400 may be considered redundant andtherefore not included on the label because hazard statement H410 also applies, see section 4.1.6 of thisdocument.442

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.4.2 Example B: Hydrophilic substance, straightforward classification based onacute data, no chronic data availableDATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksPhysico-chemical propertiesWater solubility: 1200 mg/l A.6. / pH:7.0, non-GLPLog octanol/water partition coefficient (log K ow ): 2.75 A.8. / pH:7.5, GLPAcute aquatic toxicityFishOncorhynchus mykiss:12 mg/l (96 h LC 50 )C.1. / static, non-GLPLepomis macrochirus:2.7 mg/l (96 h LC 50 )C.1. / static, GLPCrustacea Daphnia magna: 18 mg/l (48 h EC 50 ) C.2. / static, non-GLPAlgae/aquatic plantsScenedesmus subspicatus:0.056 mg/l (96 h ErC 50 )C.3. / static, GLPLemna gibba:Chronic aquatic toxicityFish:Crustacea:Algae/aquatic plants:Degradation (evidence of rapid degradation)Biotic degradation:Abiotic degradation, hydrolysis: (half-life (d)):Bioaccumulation0.031 mg/l (7 d ErC 50 )No dataNo dataNOEC not reported86 % in 28days (10 daywindowfulfilled)No dataC.26. / semi-static, GLPC.4-C / pH:7.5, GLPBioconcentration factor in fish (BCF) 560 l/kg C.13. / pH: 7.8, GLP, BCF(related to total radioactiveresidues because data forparent compound notavailable)443

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Aquatic hazard assessment, conclusions and comments:Physico-chemical properties: The substance is readily soluble. Log K ow < 4, indicating low potential forbioaccumulation, which can be used in absence of BCF data (see bioaccumulationassessment).Acute aquatic toxicity: The acute aquatic toxicity based on the lowest of the available toxicity values is between0.01 and 0.1 mg/l.Long-term aquatic toxicity: No adequate chronic toxicity data available for all three trophic levels.Degradation (evidence of rapid degradation): > 70 % degradation based on dissolved organic carbon (DOC) fulfils the criteria for rapiddegradation.Bioaccumulation: BCF > 500, hence high potential for bioaccumulation. BCF value overrules the use oflogK ow value which in this case is lower than the cut-off value of 4.Aquatic hazard classification and, where applicable, established M-factor(s):Acute aquatic hazard: category Acute 1, M-factor: 10.Long–term aquatic hazard: category Chronic 1, M-factor: 10.Reasoning:Acute (short-term) aquatic hazard: acute toxicity L(E)C 50 ≤ 1 mg/l. M-factor based on L(E)C 50between 0.01 and 0.1 mg/l.Long-term aquatic hazard:The criteria for classification of a substance into the categories Chronic 1 to 3 follow a tieredapproach where the first step is to see if adequate information on long-term toxicity isavailable allowing long-term hazard classification. In absence of adequate long-term toxicitydata for some or all trophic levels, the subsequent step is to combine two types of information,i.e. acute aquatic toxicity data and environmental fate data (degradability and bioaccumulationdata). For details see section 4.1.3.3 and Table 4.1.0. No adequate long-term toxicity data available (for all three trophic levels); Lowest acute toxicity L(E)C 50 ≤ 1 mg/l; Substance is rapidly degradable but the experimentally determined BCF > 500; Since the conclusion is based on Table 4.1.0 (b) (iii), therefore the M-factor is based onthe acute toxicity between 0.01 and 0.1 mg/l. In this case, the same factor M applies forboth acute and long-term hazard.Labelling elements based on the classification:444

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012ElementGHS PictogramSignal WordCodeGHS09WARNINGHazard Statement H410 58Precautionary statement(s)P273, P391, P50158 Note that in accordance with CLP Article 27 the hazard statement H400 may be considered redundant andtherefore not included on the label because hazard statement H410 also applies, see section 4.1.6 of thisdocument.445

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.4.3 Example C: Moderately water soluble substance, straightforwardclassification based on acute data, chronic data available for two trophiclevels only; combined set of QSAR data and experimental dataDATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksPhysico-chemical propertiesWater solubility: 25 mg/l A.6. / pH: 7.0, non-GLPLog octanol/water partition coefficient (log K ow ): 5.75A.8. / pH: 7.5, GLP3.9QSAR KOWINN, valid, non-GLPAcute aquatic toxicityFishOncorhynchus mykiss:12.3 mg/l (96 h LC 50 )C.1. / static, non-GLPLepomis macrochirus:22.5 mg/l (96 h LC 50 )C.1. / static, GLPCrustaceaDaphnia magna:0.79 mg/l (48 h EC 50 )C.2. / static, non-GLPDaphnia magna:1.06 mg/l (48 h EC 50 )QSAR, ECOSAR, valid, non-GLPAlgae/aquatic plants Scenedesmus subspicatus: 1.53 mg/l (96 h ErC 50 ) C.3. / static, GLPChronic aquatic toxicityFish: Oncorhynchus mykiss: 0.56 mg/l (21 d NOEC) OECD 210 / Early Life Stagetoxicity test, flow-through,GLPCrustacea:No dataAlgae/aquatic plants: Scenedesmus subspicatus: 0.23 mg/l (96 h NOEC) C.3. / static, GLPDegradation (evidence of rapid degradation)Biotic degradation:45 % in 28 daysC.4-C / pH: 7.5, GLPAbiotic degradation, hydrolysis: (half-life (d)):BioaccumulationBioconcentration factor in fish (BCF):No dataNo data446

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Aquatic hazard assessment, conclusions and comments:Physico-chemical properties: The substance is moderately soluble. Log K ow 5.75. Based on weight of evidence, validK ow estimated with QSAR is overruled by valid GLP experimental data.Note that use of experimental data and QSAR data for estimation log K ow should becarefully considered on a case by case basis. The validity of data may be dependant on thestructure of the chemical. See Annex III, section 2.2 for more details on the use of log K owdata and Annex III, section 3.4.1 for details on chemical classes that need special attentionin this respect.Acute aquatic toxicity: The acute aquatic toxicity based on the lowest of the available toxicity values is between0.1 and 1 mg/l; For Daphnia magna two valid values are presented. A weight of evidence approach isapplied in which the QSAR data are outweighed by the valid experimental data. Hence,the lowest acute toxicity value of 0.79 mg/l is used for crustaceans.Long-term aquatic toxicity:Adequate chronic toxicity data available only for fish and algae/aquatic plants, not forcrustaceans;The chronic aquatic toxicity based on the lowest of the available toxicity values for fishand algae/aquatic plants is between 0.1 and 1 mg/l.Since there is adequate chronic toxicity data available for two trophic levels, assess both:(a) according to the criteria given in Table 4.1.0(b)(i) or 4.1.0(b)(ii) (depending oninformation on rapid degradation), and(b) (if for the other trophic level(s) adequate acute toxicity data are available) accordingto the criteria given in Table 4.1.0(b)(iii),and classify according to the most stringent outcome.Degradation (evidence of rapid degradation):< 70 % degradation in 28 days based on dissolved organic carbon (DOC), does not fulfilthe criteria for rapid degradation.Bioaccumulation: Log K ow 5.75, indicating high potential for bioaccumulation, which can be used inabsence of BCF data.Aquatic hazard classification and, where applicable, established M-factor(s):Acute aquatic hazard: category Acute 1, M factor: 1.447

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Long-term aquatic hazard: category Chronic 1, M factor: 1.Reasoning:Acute (short-term) aquatic hazard: lowest acute aquatic toxicity L(E)C 50 ≤ 1 mg/l. M-factorbased on L(E)C 50 between 0.1 and 1 mg/l.Long-term aquatic hazard:The criteria for classification of a substance into the categories Chronic 1 to 3 follow a tieredapproach where the first step is to see if adequate information on long-term toxicity isavailable allowing long-term hazard classification. In absence of adequate long-term toxicitydata for some or all trophic levels, the subsequent step is to combine two types of information,i.e. acute aquatic toxicity data and environmental fate data (degradability and bioaccumulationdata). In this example the absence of long-term study for the species/trophic level (i.e.Daphnia/Crustacea) with the lowest acute toxicity value supports using the surrogate system.For details see section 4.1.3.3 and Table 4.1.0. NOEC-based system (Table 4.1.0 (b)(i): lowest long-term aquatic toxicity NOEC ≤ 1mg/l, not rapidly degradable, hence category Chronic 2; Surrogate system (Table 4.1.0 (b)(iii): lowest acute aquatic toxicity L(E)C 50 < 1 mg/l, notrapidly degradable (and Log K ow >4), hence category Chronic 1; Conclusion: category Chronic 1 applies following the most stringent outcome; Since the conclusion is based on the surrogate system (Table 4.1.0 (b) (iii)) the M-factor isbased on the acute aquatic toxicity between 0.1 and 1 mg/l.Labelling elements based on the classification:ElementGHS PictogramSignal WordCodeGHS09WARNINGHazard Statement H410 59Precautionary statement(s)P273, P391, P50159 Note that in accordance with CLP Article 27 the hazard statement H400 may be considered redundant andtherefore not included on the label because hazard statement H410 also applies, see section 4.1.6 of thisdocument.448

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.4.4 Example D: Substance with several toxicity data for a trophic levelDATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksPhysico-chemical propertiesWater solubility: 120 mg/l A.6. / pH:7.0, non-GLPLog octanol/water partition coefficient (log K ow ): 4.9 A.8. / pH:7.5, GLPAcute aquatic toxicityFish: Lepomis macrochirus: 108 mg/l (96 h LC 50 ) C.1. / static, GLPCrustacea 60 :Daphnia magna:40 mg/l (48 h EC 50 )C.2. / static, GLPProcambarus clarkii:0.12 mg/l (48 h EC 50 )Method na. / static, GLPAsellus aquaticus:0.4 mg/l (48 h EC 50 )Method na. / static, non-GLPMysidopsis bahia:0.5 mg/l (48 h EC 50 )Method na. / static, GLPChironomus tentans:0.8 mg/l (48 h EC 50 )Method na. / static, GLPAlgae/aquatic plantsPseudokirchneriellasubcapitata: 22 mg/l (96 h ErC 50 ) C.3. / static, GLPChronic aquatic toxicityFish: Pimephales promelas 1.1 mg/l (21 d NOEC) OECD 210 / Early Life Stagetoxicity test, flow-through,GLP, endpoint: growthCrustacea: Daphnia magna 1.2 mg/l (21 d NOEC) C.20. / semi-static, GLP,endpoint: reproductionAlgae/aquatic plants:PseudokirchneriellaDegradation (evidence of rapid degradation)subcapitata: 8.5 mg/l (96 h NOEC) C.3. / static, GLPBiotic degradation:No dataAbiotic degradation, hydrolysis (half-life (d)):No dataBioaccumulationBioconcentration factor in fish (BCF):No data60 Some species in this trophic level may be representatives of other taxonomic groups than crustecea e.g. thenon-biting midge Chironomus tentans is a representative of the subphylum Hexapoda (class Insecta).449

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Aquatic hazard assessment, conclusions and comments:Physico-chemical properties: The substance is water soluble. Log K ow 4.9.Acute aquatic toxicity: The acute aquatic toxicity (based on the lowest of the available toxicity values) is between0.1 and 1 mg/l. The classification in this example should be based on the most sensitivespecies which is the crustacean Procambarus clarkii; Note that in general for substances for which multiple toxicity data is available for ataxonomic group (in this case crustaceans) on a case-by-case basis the toxicity data maybe evaluated by weighting the evidence. If for example four or more acute LC 50 valueswere available for the same fish species, then a geometric mean may be calculated (seesection 4.1.3.2.4.3). In this specific example, acute toxicity data on five separatecrustacean species is available and all – except one – are from GLP studies that areweighed equally in a weight of evidence approach. Accordingly, the lowest value is usedfor classification purposes.Chronic aquatic toxicity:Adequate long-term toxicity data available only for fish and algae/aquatic plants. Thechronic aquatic toxicity (based on the lowest of the two available toxicity values) is above1 mg/l;For crustaceans chronic data is available for Daphnia magna which based upon therelatively large acute dataset is clearly the least sensitive of the species for which data isavailable. Hence, the chronic aquatic toxicity data on Daphnia magna in this case shouldbe considered not in conformity with the definition of ‘adequate chronic data’.Degradation (evidence of rapid degradation):No data available for this substance. In this case the substance is considered as not rapidlydegradable (see Table 4.1.0, Note 3).Bioaccumulation: Log K ow 4.9, indicating high potential for bioaccumulation, which can be used in absenceof BCF data.Aquatic hazard classification and, where applicable, established M-factor(s):Acute aquatic hazard: category Acute 1, M factor: 1.Long-term aquatic hazard: category Chronic 1, M factor 1.Reasoning:Acute aquatic hazard: Acute aquatic toxicity L(E)C 50 > 0.001 and < 0.01 mg/l;Long-term aquatic hazard:450

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The criteria for classification of a substance into the categories Chronic 1 to 3 follow a tieredapproach where the first step is to see if adequate information on long-term toxicity isavailable allowing long-term hazard classification. In absence of adequate long-term toxicitydata for some or all trophic levels, the subsequent step is to combine two types of information,i.e. acute aquatic toxicity data and environmental fate data (degradability and bioaccumulationdata). For details see section 4.1.3.3 and Table 4.1.0. Adequate Chronic toxicity data available for two out of three trophic levels (fish andalgae/aquatic plants), lowest NOEC above 1 mg/l. Conclusion for these two trophic levels:NOEC-based system (Table 4.1.0 (b)(i): lowest long-term aquatic toxicity NOEC > 1mg/l, hence not classified; Surrogate system (Table 4.1.0 (b)(iii): lowest acute aquatic toxicity L(E)C 50 < 1 mg/l(0.12 mg/l Procambarus clarkii), not rapidly degradable (and log K ow > 4), hence categoryChronic 1; Conclusion: category Chronic 1 applies following the most stringent outcome; Since the conclusion is based on the surrogate system (Table 4.1.0 (b) (iii)) the M-factor isbased on the acute aquatic toxicity between 0.1 and 1 mg/l.Labelling elements based on the classification:ElementGHS PictogramSignal WordCodeGHS09WARNINGHazard Statement H410 61Precautionary statement(s)P273, P391, P50161 Note that in accordance with CLP Article 27 the hazard statement H400 may be considered redundant andtherefore not included on the label because hazard statement H410 also applies, see section 4.1.6 of thisdocument.451

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.4.5 Example E: “Safety net” classification category Chronic 4DATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksPhysico-chemical propertiesWater solubility: 0.009 mg/l A.6. / pH:7.0, non-GLPLog octanol/water partition coefficient (log K ow ): 5.4 A.8. / pH:7.5, GLPAcute aquatic toxicityFish:No dataCrustacea Daphnia magna: > 1 mg/l (48 h EC 50 ) C.2. / static, nominalconcentration, non-GLPAlgae/aquatic plants:Chronic aquatic toxicityFish:Crustacea:Algae/aquatic plants:No dataNo dataNo dataNo dataDegradation (evidence of rapid degradation)Biotic degradation:No dataAbiotic degradation, hydrolysis (half-life (d)):No dataBioaccumulationBioconcentration factor in fish (BCF):No dataAquatic hazard assessment, conclusions and comments:Physico-chemical properties: The substance is poorly soluble. Log K ow > 4, indicating high potential forbioaccumulation, which can be used in absence of BCF data.Acute aquatic toxicity: Data poor substance. No acute toxicity recorded at levels up to the limit of watersolubility.452

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Long-term aquatic toxicity:No adequate chronic toxicity data available for all three trophic levels.Degradation (evidence of rapid degradation):The substance is considered not rapidly degradable by default in absence of measureddata.Bioaccumulation: Log K ow 5.4, indicating high potential for bioaccumulation, which can be used in absenceof BCF data.Aquatic hazard classification and, where applicable, established M-factor(s):Acute hazard: Not classified.Long-term hazard: ‘Safety net’ classification category Chronic 4.Reasoning:Acute hazard: No acute aquatic toxicity recorded at levels up to the limit of water solubility;Long-term hazard: No adequate chronic toxicity data available for all three trophic levels.Substance nevertheless of concern based on the following findings: Poorly soluble substance; No acute aquatic toxicity recorded at levels up to the limit of water solubility; Not rapidly degradable (by default in absence of measured data); High potential for bioaccumulation (in absence of BCF data, log K ow > 4).No evidence on NOEC being > water solubility for all three trophic levels.No other evidence of rapid degradation in the environmentLabelling elements based on the classification:ElementCodeGHS Pictogram -Signal Word -Hazard StatementPrecautionary statement(s)H413P273, P501453

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.4.6 Example F: Substance difficult to test, toxicity above level of watersolubilityDATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksPhysico-chemical propertiesWater solubility: < 0.2 mg/l A.6. / pH: 7.0, non-GLPLog octanol/water partition coefficient (log K ow ): No data Not determined due toinstability of the substance inwaterAcute aquatic toxicityFish: Oncorhynchus mykiss: 12 mg/l (96 h LC 50 ) C.1. / static, nominalconcentration, non-GLPCrustacea Daphnia magna: 18 mg/l (48 h EC 50 ) C.2. / static, nominalconcentration, non-GLPAlgae/aquatic plants PseudokirchneriellasubcapitataChronic aquatic toxicityFish:Crustacea:Algae/aquatic plants:3.56 mg/l (96 h ErC 50 ) C.3. / static, nominalconcentration, non-GLPNo dataNo dataNo dataDegradation (evidence of rapid degradation)Biotic degradation:No dataAbiotic degradation, hydrolysis: (half-life (d)):BioaccumulationBioconcentration factor in fish (BCF):< 0.5 days (longest halflifewithin pH 4-9)No dataC.7. / pH: 7.0, non-GLP454

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Aquatic hazard assessment, conclusions and comments:Physico-chemical properties: The water solubility test is not considered to be valid (Klimisch 3) as the substance isknown to rapidly hydrolyse and this was not considered in this study. Log K ow notdetermined.Acute aquatic toxicity: This data is based on initial measured concentrations in the suspension and the reportedEC 50 values are far above the water solubility (Klimisch score 3). Tests undertaken in astatic regime which is inappropriate for a substance which rapidly hydrolyses (see alsoIR/CSA R.7b for guidance on how to test difficult substances) It is not clear whether the reported effects in the acute toxicity studies are due to physicaleffects of the undissolved substance particles in the test media on the test species orinherent toxicity of the substance.Long-term aquatic toxicity:No adequate long-term toxicity data available for all three trophic levels.Degradation (evidence of rapid degradation):In the assessment of rapid degradability hydrolysis can be considered if the hydrolysisproducts do not fulfil the criteria for classification as hazardous to the aquaticenvironment. In this example hydrolysis is sufficient to show a rapid degradability of theparent substance in the environment but no information is available about the breakdownproduct(s). More data on degradation of this/these compound(s) would be necessary;In absence of data to show a rapid degradation of the breakdown product(s) the parentsubstance is considered not rapidly degradable.Bioaccumulation: Log K ow could not be determined experimentally. The parent substance has a lowpotential for bioaccumulation due to hydrolytical instability.Aquatic hazard classification and, where applicable, established M-factor(s):Acute aquatic hazard: Not classified in absence of adequate data (data of poor quality).Long-term aquatic hazard: category Chronic 4.Reasoning:Acute hazard (Table 4.1.0 (a)): No acute aquatic toxicity as no adequate acute data available;Long-term hazard: No adequate long-term toxicity data available for all three trophic levels.Substance nevertheless of concern based on the following findings: Poorly soluble substance (< 0.2 mg/l); No acute aquatic toxicity recorded at levels up to the limit of water solubility;455

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Not rapidly degradable (see section 4.1.3.2.3.2 of this guidance (CLP legal text: point4.1.2.9.3);No evidence of NOEC being > water solubility for all three trophic levels.No information available on the hydrolysis products and hence dataset not decisivewhether these fulfil the criteria for classification as hazardous to the aquatic environmentbased upon:o Toxicityo Rapid degradabilityo BioaccumulationIn this case the safety net classification should be applied because of the large uncertaintyon the fate and effects of the hydrolysis products.Labelling elements based on the classification:ElementCodeGHS Pictogram -Signal Word -Hazard StatementPrecautionary statement(s)H413P273, P501456

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4 Classification of mixtures hazardous to the aquatic environment4.1.4.1 General considerations for classification of mixtures hazardous to the aquaticenvironmentNote that general principles for classification of mixtures under CLP are given in section1.1.6.2 and section 1.6 of part 1 of this guidance document.The basic principle of mixture classification under CLP is shown in the green box below andin Figure 4.1.2, which is also explained in the text below the box.Annex I: 4.1.3.2 The approach for classification of aquatic environmental hazards is tiered, and isdependent upon the type of information available for the mixture itself and for its components.Figure 4.1.2 outlines the process to be followed.Elements of the tiered approach include:– classification based on tested mixtures;– classification based on bridging principles;– the use of "summation of classified components" and/or an "additivity formula".Figure 4. 1.2Tiered approach to classification of mixturesfor acute and long-term aquatic environmental hazardsAquatic toxicity test data available on the mixture as a wholeNoYesCLASSIFYfor acute/long-term aquatic hazard(see 4.1.3.3)Sufficient data available onsimilar mixtures to estimatehazardsYesApply bridging principles(see 4.1.3.4.)CLASSIFYfor acute/long-term aquatic hazardNoEither aquatic toxicity orclassification data available for allrelevant componentsYesApply summation Method (see4.1.3.5.5) using:• Percentage of all componentsclassified as "Chronic"CLASSIFYfor acute/long-term aquatic hazard• Percentage of componentsclassified as "Acute"• Percentage of components withacute or chronic toxicity data:Noapply addititivity formulas (see4.1.3.5.2) and convert thederived L(E)C50 or EqNOECm tothe appropriate "Acute" or"Chronic" CategoryUse available hazard dataof known components.Apply Summation method and/orAdditivity Formula (see 4.1.3.5) andapply 4.1.3.6CLASSIFYfor acute/long-term aquatic hazard457

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Explanation of Figure 4.1.2: Horizontal arrow in first row: In some cases, particularly where specific and valid testdata are already available on the mixture, there is a general obligation to use these dataon the mixture itself for classification purposes. Valid data must normally then beavailable on each of fish, crustacea and algae or other aquatic plants, unless a decisionto classify in the most stringent category(ies) (Acute 1 and/or Chronic 1) can be madewithout a full dataset (see section 4.1.4.3 of this document). Horizontal arrows in second row: In other cases, sufficient data may be available onsimilar tested mixtures to estimate hazards using the bridging principles (see section4.1.4.4 of this document).Horizontal arrows in third row: In general, however, where either aquatic toxicity orclassification data are available for all relevant components of a mixture the aquatichazard classification shall be made through the identification of the hazards of therespective components in a first step, and then in a second step through the summationof the quantities of these hazardous components, applying the summation method (seesection 4.1.4.5. of this document).When doing so:o The percentage of all components classified as Acute 1 and/or Chronic 1, 2, 3& 4 is fed straight into the summation method (for relevant components seepoint 4.1.3.1 of Annex I to CLP);o For the percentage of the other components with acute or long-term toxicitydata, the addititivity formulas (see point 4.1.3.5.2 of Annex I to CLP) may beapplied. The derived L(E)C 50 or EqNOECm is converted to the appropriate"Acute" or "Chronic" Category and then, in a second step, fed into thesummation method. 62Horizontal arrows in fourth (last) row: Use available hazard data of knowncomponents.o This applies to mixtures containing unknown components and/or knowncomponents, for which neither toxicity data nor classifications are known. Inthese cases, apply the additional statement on the label and in the safety datasheet: "Contains x % of components with unknown hazards to the aquaticenvironment" (see the green box below). For classification based on the knownpart of the mixture, use the Summation Method and/or the Additivity Formula(see section 4.1.4.5 of this document).Annex I: 4.1.3.6.1 In the event that no useable information on acute and/or long-term aquatichazard is available for one or more relevant components, it is concluded that the mixture cannot beattributed to one or more definitive hazard category(ies). In this situation the mixture shall beclassified based on the known components only, with the additional statement on the label and inthe SDS that: "Contains x % of components with unknown hazards to the aquatic environment".62 As manufacturers and importers are obliged to classify all substances placed on the market within the EU, thesummation method can usually be directly applied and the addititivity formula will be of limited application.458

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4.2 Information requirementsBefore a classification can be made, available information on toxicity of the mixture as awhole as well as all the available information on the composition of the mixture and thehazard category of relevant components (substances) should be gathered. Note thatmanufacturers, importers or downstream users are not requested by the CLP Regulation togenerate new data for determining the aquatic hazard classification of the mixture. Rather thesupplier should be contacted if it is considered that the information on the substance ormixture supplied is not sufficient for classification purposes.Generally, therefore, the constituent substance classifications should be used as the basis forto derivation of the correct hazard classification for the final mixture (see also section 1.6.4 ofthis guidance document).Article 11 of the CLP-Regulation refers to cut-off values. These values are the minimumconcentrations for a substance to be taken into account for classification purposes. Thesubstances meeting these criteria are relevant ingredients or relevant components. When aclassified substance is present in a concentration above the generic cut-off value it contributesto the mixture classification even if it may not trigger classification of the mixture directly.Annex I: 4.1.3.1. The classification system for mixtures covers all classification categories whichare used for substances, i.e. categories Acute 1 and Chronic 1 to 4. In order to make use of allavailable data for purposes of classifying the aquatic environmental hazards of the mixture, thefollowing is applied where appropriate:The "relevant components" of a mixture are those which are classified "Acute 1"or "Chronic 1" andpresent in a concentration of 0.1 % (w/w) or greater, and those which are classified "Chronic 2","Chronic 3" or "Chronic 4" and present in a concentration of 1 % (w/w) or greater, unless there is apresumption (such as in the case of highly toxic components (see 4.1.3.5.5.5)) that a componentpresent in a lower concentration can still be relevant for classifying the mixture for aquaticenvironmental hazards. Generally, for substances classified as "Acute 1" or "Chronic 1" theconcentration to be taken into account is (0.1/M) %. (For explanation M-factor see 4.1.3.5.5.5).For aquatic hazards the cut-off values are further addressed under point 1.1.2.2.2 (b) of AnnexI to CLP. The calculation referred to in point (b)(i) of that point, is found in point 4.1.3.1 ofAnnex I to CLP (see the green box above).This signals that highly toxic components will need to be considered at lower levels than thegeneric cut-off values, and this applies to any substance to which an M-factor greater than 1has been assigned (see section 4.1.4.5 of this document)..Note that generic concentration limits (GCLs) should be given in weight percentages exceptfor certain gaseous mixtures where they may be best described in volume percentage, e.g. asingle hazardous component in an inert diluent, e.g. nitrogen or helium.When the information on the mixture has been gathered and validated, the following guidanceshould be followed depending on the type and level of information available.459

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4.3 Classification criteria for mixtures hazardous to the aquatic environmentbased on test data on the mixture as a wholeThe testing of a mixture for aquatic toxicity is highly complex, both in terms of the conduct ofthe test, and in the interpretation of data from such testing. The different physico-chemicalproperties, such as water solubility, vapour pressure, and adsorption, make it almostimpossible to prepare an exposure concentration that is characteristic of the mixture, while themulti-component analysis needed to verify such an exposure concentration is both complexand expensive.Therefore, before any such new testing is conducted, alternative approaches such as thesummation method, should be considered, particularly where testing would involve the use ofvertebrate animals such as fish (see also section 1.1.6.2 of this document). Nevertheless, thereare circumstances where test data may already be available, and should then be examined toassess its relevance for the purposes of classification. Data which has been prepared forRegulatory use in compliance with standard guidelines, such as test data on plant protectionor biocidal products, may be considered as acceptable for classification. Where such valid testdata, both acute and chronic, are available, they may be used in accordance with the generalguidance below.Annex I: 4.1.3.3.1 When the mixture as a whole has been tested to determine its aquatic toxicity,this information can be used for classifying the mixture according to the criteria that have beenagreed for substances. The classification is normally based on the data for fish, crustacea andalgae/plants (see sections 4.1.2.7.1 and 4.1.2.7.2). When adequate acute or chronic toxicity data forthe mixture as a whole are lacking, “bridging principles” or “summation method” should be applied(see sections 4.1.3.4 and 4.1.3.5).4.1.3.3.2 The long-term hazard classification of mixtures requires additional information ondegradability and in certain cases bioaccumulation. Degradability and bioaccumulation tests formixtures are not used as they are usually difficult to interpret, and such tests may be meaningfulonly for single substances.4.1.3.3.3 Classification for category Acute 1(a) When there are adequate acute toxicity test data (LC 50 or EC 50 ) available for themixture as a whole showing L(E)C 50 1 mg/l:Classify mixture as Acute 1 in accordance with point (a) of Table 4.1.0.(b) When there are acute toxicity test data (LC 50 (s) or EC 50 (s)) available for themixture as a whole showing L(E)C 50 (s) 1 mg/l for normally all trophic levels:No need to classify for acute hazard.4.1.3.3.4 Classification for categories Chronic 1, 2 and 3(a) When there are adequate chronic toxicity data (EC x or NOEC) available for themixture as a whole showing EC x or NOEC of the tested mixture ≤ 1mg/l:(i)(ii)Classify the mixture as Chronic 1, 2 or 3 in accordance with point(b)(ii) of Table 4.1.0. as rapidly degradable if the available informationallows the conclusion that all relevant component of the mixture arerapidly degradable;Classify the mixture as Chronic 1 or 2 in all other cases in accordancewith point (b)(i) of Table 4.1.0.as non-rapidly degradable;460

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(b) When there are adequate chronic toxicity data (EC x or NOEC) available for themixture as a whole showing EC x (s) or NOEC(s) of the tested mixture > 1 mg/l fornormally all trophic levels:No need to classify for long-term hazard in categories Chronic 1, 2 or 3.4.1.3.3.5 Classification for category Chronic 4If there are nevertheless reasons for concern:Classify the mixture as Chronic 4 (safety net classification) in accordance with Table4.1.0.Where a classification is made based on test data, valid data should normally be available oneach of fish, crustacea and algae or other aquatic plants, unless a decision to classify in themost stringent category(ies) (Acute 1 and/or Chronic 1) can be made without a full dataset. Tobe valid, it would normally be necessary to show that the tested organism has been exposed tothe toxic components of the mixture in proportion to the composition of the mixture, and thatthis exposure has been maintained for the duration of the test. If this cannot be accomplishedthe classification should be based on information on the individual components. It isinsufficient to simply prepare a water-accommodated fraction (WAF) for testing.When there is adequate toxicity test data available for the mixture as a whole, this may besimplified to two basic rules for each of acute and long-term hazard classification:Classification for acute (short-term) aquatic hazard:i) If the lowest valid acute/short-term L(E)C 50 is ≤ 1 mg/l, classify as Acute 1.ii)If valid acute/short-term test data are available on fish, crustacea and algae/aquaticplants (i.e. all three trophic levels), and all showing L(E)C 50 > 1 mg/l, there is noneed to classify for acute aquatic hazard.Classification for long-term aquatic hazardi) If the lowest valid chronic toxicity test data (NOEC or EC x ) is ≤ 1 mg/l, classify asChronic 1, 2 or 3, depending on the information on components degradability, e.g.if all components are known to be rapidly degradable.ii)If valid chronic toxicity test data are available on fish, crustacea and algae/aquaticplants (i.e. all three trophic levels), and all showing NOEC or EC x >1 mg/l, there isno need to classify for long-term aquatic hazard in Chronic 1, 2 or 3.4.1.4.4 When experimental aquatic toxicity data are not available for the completemixture: bridging principlesAnnex I: 4.1.3.4.1 Where the mixture itself has not been tested to determine its aquaticenvironmental hazard, but there are sufficient data on the individual components and similar testedmixtures to adequately characterise the hazards of the mixture, this data shall be used in accordancewith the bridging rules set out in Section 1.1.3. However, in relation to application of the bridgingrule for dilution, sections 4.1.3.4.2 and 4.1.3.4.3 shall be used.4.1.3.4.2 Dilution: if a mixture is formed by diluting another tested mixture or a substanceclassified for its aquatic environmental hazard with a diluent which has an equivalent or loweraquatic hazard classification than the least toxic original component and which is not expected toaffect the aquatic hazards of other components, then the resulting mixture may be classified as461

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012equivalent to the original tested mixture or substance. Alternatively, the method explained insection 4.1.3.5 may be applied.4.1.3.4.3 If a mixture is formed by diluting another classified mixture or substance with water orother totally non-toxic material, the toxicity of the mixture can be calculated from the originalmixture or substance.For circumstances where no or inadequate test data are available on the mixture itself, theclassification of a mixture may be determined based on sufficient data for individualcomponents of the mixture and on another similar tested mixture by an appropriateapplication of any of the specified "bridging principles". The identified relevant informationneeds to be evaluated for the purpose of classification, by comparing it with the criteria inpoint 1.1.3 of Annex I to CLP. Those rules allow characterisation of the hazards of themixture without performing tests on it, but rather by building on the available information onsimilar tested mixtures (see also Part 1, section 1.6.3.2 of this guidance document).4.1.4.5 When hazard data (information on toxicity or classification) are available forall the components of the mixtureAnnex I: 4.1.3.5.1 The classification of a mixture is based on summation of the classification of itscomponents. The percentage of components classified as "Acute" or "Chronic" is fed straight in tothe summation method. Details of the summation method are described in 4.1.3.5.5.Where no or inadequate test data on the mixture itself is available and the bridging principlesare not applicable, the classification of the mixture is based on information on thecomponents. The information that will most usually be available to aid classification of amixture will be the classification applied to the individual components (substances). Thesedata and any associated M-factor(s) are included in the Safety Data Sheets (SDS) and also inthe Classification and Labelling Inventory (C&L Inventory) established and maintained bythe Agency in the form of a database [link to be added once the public Inventory is available].In cases the aquatic hazard classification of a mixture will be made based on data on thecomponents, it is therefore generally the summation of the quantities of the hazardouscomponents that should be used to determine a specific hazard classification of the mixture.Provided the classification data, in part or in total, and the % of these components in themixture are known, a classification of the mixture can be made according to the summationmethod. The following text from CLP describes the application of this method.Annex I: 4.1.3.5.5 Summation method4.1.3.5.5.1 Rationale4.1.3.5.5.1.1 In case of the substance classification categories Chronic 1 to Chronic 3, theunderlying toxicity criteria differ by a factor of 10 in moving from one category to another.Substances with a classification in a high toxicity band therefore contribute to the classification of amixture in a lower band. The calculation of these classification categories therefore needs toconsider the contribution of any substance classified as Chronic 1, 2 or 3.462

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.3.5.5.2. Classification procedure4.1.3.5.5.2.1 In general a more severe classification for mixtures overrides a less severeclassification, e.g. a classification with Chronic 1 overrides a classification with Chronic 2. As aconsequence, in this example, the classification procedure is already completed if the result of theclassification is Chronic 1. A more severe classification than Chronic 1 is not possible. Therefore itis not necessary to undergo the further classification procedure.4.1.3.5.5.3 Classification for category Acute 14.1.3.5.5.3.1 First all components classified as Acute 1 are considered. If the sum of theconcentrations (in %) of these components multiplied by their corresponding M-factors is greaterthan 25 % the whole mixture is classified as Acute 1.4.1.3.5.5.3.2 The classification of mixtures for acute hazards based on this summation of classifiedcomponents is summarised in Table 4.1.1.Table 4.1.1Classification of a mixture for acute hazards,based on summation of classified componentsSum of components classified as:Mixture is classified as:Acute 1 M a ≥ 25 %Acute 1a For explanation of the M-factor see 4.1.3.5.5.54.1.3.5.5.4 Classification for the categories Chronic 1, 2, 3 and 44.1.3.5.5.4.1 First all components classified as Chronic 1 are considered. If the sum of theconcentrations (in %) of these components multiplied by their corresponding M-factors is equal toor greater than 25 %, the mixture is classified as Chronic 1. If the result of the calculation is aclassification of the mixture as Chronic 1, the classification procedure is completed.4.1.3.5.5.4.2 In cases where the mixture is not classified as Chronic 1, classification of the mixtureas Chronic 2 is considered. A mixture is classified as Chronic 2 if 10 times the sum of theconcentrations (in %) of all components classified as Chronic 1 multiplied by their correspondingM-factors plus the sum of the concentrations (in %) of all components classified as Chronic 2 isequal to or greater than 25 %. If the result of the calculation is classification of the mixture asChronic 2, the classification process is completed.4.1.3.5.5.4.3 In cases where the mixture is not classified either as Chronic 1 or Chronic 2,classification of the mixture as Chronic 3 is considered. A mixture is classified as Chronic 3 if 100times the sum of the concentrations (in %) of all components classified as Chronic 1 multiplied bytheir corresponding M-factors plus 10 times the sum of the concentrations (in %) of all componentsclassified with Chronic 2 plus the sum of the concentrations (in %) of all components classified asChronic 3 is ≥ 25 %.4.1.3.5.5.4.4 If the mixture is still not classified in Chronic 1, 2 or 3, classification of the mixture asChronic 4 shall be considered. A mixture is classified as Chronic 4 if the sum of the concentrations(in %) of components classified as Chronic 1, 2, 3 and 4 is equal to or greater than 25 %.4.1.3.5.5.4.5 The classification of mixtures for long-term hazards, based on this summation of theconcentrations of classified components, is summarised in Table 4.1.2.463

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Table 4.1.2Classification of a mixture for long-term hazards,based on summation of the concentrations of classified componentsSum of components classified as:Mixture is classified as:Chronic 1 M a ≥ 25 % Chronic 1(M 10 Chronic 1) + Chronic 2 ≥ 25 % Chronic 2(M 100 Chronic 1) + (10 Chronic 2)+ Chronic 3 ≥ 25 %Chronic 1 + Chronic 2 + Chronic 3+ Chronic 4 ≥ 25 %Chronic 3Chronic 4a For explanation of the M-factor, see 4.1.3.5.5.54.1.3.5.5.1.2 When a mixture contains components classified as Acute 1 or Chronic 1, attentionmust be paid to the fact that such components, when their acute toxicity is below 1 mg/l and/orchronic toxicity is below 0,1 mg/l (if non-rapidly degradable) and 0.01 mg/l (if rapidly degradable)contribute to the toxicity of the mixture even at a low concentration. Active ingredients inpesticides often possess such high aquatic toxicity but also some other substances likeorganometallic compounds. Under these circumstances the application of the normal genericconcentration limits leads to an "under-classification" of the mixture. Therefore, multiplyingfactors shall be applied to account for highly toxic components, as described in section 4.1.3.5.5.5.For those components for which only toxicity data are available the additivity formulas offer away for estimating what the toxicity of a mixture would be if the individual substancetoxicities could be ‘added’ to each other in a straightforward way. Thus it assumes a similar‘mode of action’ for each component.To make full use of this approach access to the whole aquatic toxicity dataset and thenecessary knowledge to select the best and most appropriate data is required. Clearly, the bestuse would be to add up separately each of the fish toxicity data, the crustacean toxicity dataand the algae/aquatic plants toxicity data to derive a specific toxicity value for each trophiclevel. The lowest of the toxicity values would normally be used to define the appropriatehazard category for the mixture. Indeed, if it is only possible to characterise part of themixture in this way, that part can be assigned a hazard category (and an M-factor forcategories Acute 1 and/or Chronic 1) and then, in a second step, be used in the summationmethod.The use of the additivity formulae is limited to those circumstances where the substancehazard category is not known. The following text from CLP describes the application of theadditivity formulae.464

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Annex I: 4.1.3.5.2 Mixtures can be made of a combination of both components that are classified(as Acute 1 and/or Chronic 1, 2, 3, 4) and others for which adequate toxicity test data is available.When adequate toxicity data are available for more than one component in the mixture, thecombined toxicity of those components is calculated using the following additivity formulas (a) or(b), depending on the nature of the toxicity data:(a)where:Based on acute aquatic toxicity:CiL(E)CCiL(E)50m C50iC i = concentration of component i (weight percentage);L(E)C 50 i = (mg/l) LC 50 or EC 50 for component i; = number of components, and i is running from 1 to n;L(E)C 50 m = L(E) C 50 of the part of the mixture with test data;The calculated toxicity may be used to assign to that portion of the mixture an acute hazardcategory which is then subsequently used in applying the summation method;(b)Based on chronic aquatic toxicity:where:C i =Cj =Ci CjEqNOEC mnCiNOECinCj0,1 NOECjconcentration of component i (weight percentage) covering the rapidlydegradable components;concentration of component j (weight percentage) covering the nonrapidlydegradable components;NOEC i = NOEC (or other recognized measures for chronic toxicity) for componenti covering the rapidly degradable components, in mg/l;NOECj = NOEC (or other recognized measures for chronic toxicity) for componentj covering the non-rapidly degradable components, in mg/l;n = number of components, and i and j are running from 1 to n;EqNOEC m = Equivalent NOEC of the part of the mixture with test data;The equivalent toxicity thus reflects the fact that non-rapidly degrading substances are classifiedone hazard category level more “severe” than rapidly degrading substances.The calculated equivalent toxicity may be used to assign that portion of the mixture a long-termhazard category, in accordance with the criteria for rapidly degradable substances (point (b)(ii) ofTable 4.1.0.), which is then subsequently used in applying the summation method.4.1.3.5.3. When applying the additivity formula for part of the mixture, it is preferable to calculatethe toxicity of this part of the mixture using for each substance toxicity values that relate to thesame taxonomic group (i.e. fish, crustacean, algae or equivalent) and then to use the highesttoxicity (lowest value) obtained (i.e. use the most sensitive of the three taxonomic groups).However, when toxicity data for each component are not available in the same taxonomic group,465

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012the toxicity value of each component is selected in the same manner that toxicity values areselected for the classification of substances, i.e. the higher toxicity (from the most sensitive testorganism) is used. The calculated acute and chronic toxicity is then used to assess whether this partof the mixture shall be classified as Acute 1 and/or Chronic 1, 2 or 3 using the same criteriadescribed for substances.Note that generic concentration limits (GCLs) should be given in weight percentages exceptfor certain gaseous mixtures where they may be best described in volume percentage, e.g. asingle hazardous component in an inert diluent, e.g. nitrogen or helium.NOTICE: With the aquatic toxicity data at hand the ingredient substance classification andM-factor(s) could easily be gained by a direct comparison with the substance criteria, whichthen could be fed straight into the summation method. It will therefore usually not benecessary to use the additivity formulae.4.1.4.6 When hazard data (information on toxicity or classification) are available foronly some components of the mixtureThis section is related to Figure 4.1.1. where one can decide to apply the summationmethod and/or the additivity formulae (see point 4.1.3.5 of Annex I to CLP) and applypoint 4.1.3.6 of Annex I to CLP.Use available hazard data of known components.o This applies to mixtures containing unknown components and/or knowncomponents, for which neither toxicity data nor classifications are known. Inthese cases, for labelling purposes consider the provisions of point 4.1.3.6 inAnnex I to CLP. For classification based on the known part of the mixture, usethe summation method and/or the additivity formula (see point 4.1.3.5 ofAnnex I to CLP).oNOTE: If a mixture is classified in more than one way, the method yielding themost stringent result should be used.4.1.4.7 Decision on classification: examples for mixturesIf the evaluation shows that the criteria are fulfilled, one category for acute aquatic hazardand/or one category for long-term aquatic hazards should be assigned. For the labellingelements, such as: hazard pictograms, signal words, hazard statements and precautionarystatements, see Section 4.1.6.466

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012List of the examples on mixtures classification included in this section:The classification system for mixtures is complex as different methods are available. Whichmethod to use is dependent upon the type of information available.● Example A: When classification data are available for some or all components of a mixture:straightforward application of the summation method● Example B1: When toxicity test data on the mixture as a whole are available for all threetrophic levels: classification based on test data on the mixture● Example B2: When information on the classification of the components and test data on themixture as a whole are available for some, but not all three trophic levels: classification basedon the summation method● Example C: When no data are available on the mixture as a whole and its components, buttest data are available on a similar tested mixture: use of the bridging principles – dilutionwith water● Example D: When only test data are available for some, but not all components of themixture: use of the additivity formulae and of the summation method467

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4.7.1 Example A: When classification data are available for some or all componentsof a mixture: straightforward application of the summation methodInformation on ingredients classification and concentrationAcuteaquatichazardMLong-termaquatic hazardM C (%)AstralamidAcute 110Chronic 1101BastralamidAcute 11Chronic 2-3CastralamidDastralamidEstralamidNotclassifiedNotclassified---Chronic 2Chronic 3Not classified---101010FestralamidNotclassified-Not classified-66NotclassifiedM = M-factor; C = ConcentrationAquatic hazard classification:Acute aquatic hazard :Not classified.Long-term aquatic hazard:Category Chronic 2Reasoning:Valid test data on the mixture as a whole (for all three trophic levels) are not available.Valid test data on similar tested mixtures are not available, either, meaning that anybridging principle cannot be used.Therefore, classification should be considered based on individual components using thesummation method.Acute aquatic hazard: Information on classification including associated M-factors and the %of the components in the mixture are available.Classify for acute hazard if: ∑ (Acute 1 M) ≥ 25%Using the classification of the components of the mixture: (1 10) + (3 1) = 13 (which is

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Long-term aquatic hazard:Step 1: Classify as Chronic 1 if: ∑ (Chronic 1 M) ≥ 25% (if not, then go to Step 2).Step 2: Classify as Chronic 2 if: ∑ (10 Chronic 1 M) + ∑ (Chronic 2) ≥ 25% (if not, thengo to Step 3).Step 3: Classify as Chronic 3 if: ∑ (100 Chronic 1 M) + ∑ (10 Chronic 2) + ∑ (Chronic3) ≥ 25% (if not, then go to Step 4).Step 4: Classify as Chronic 4 if: ∑ (Chronic 1) + ∑ (Chronic 2) + ∑ (Chronic 3) + ∑ (Chronic4) ≥ 25%Using the classification of the components of the mixture:Step 1: (1 10) = 10 (which is < 25% → Step 2).Step 2: (10 1 10) + 3+10 = 113 (which is > 25%). Hence, classify as Category Chronic 2.Labelling elements based on the classification:ElementGHS PictogramAquatic hazardinformation thatcould appear on thelabelGHS09Signal Word -Hazard StatementPrecautionary statement(s)H411P273, P391, P501469

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4.7.2 Example B1: When toxicity data on the mixture as a whole is available for allthree trophic levels: classification based on test data for the mixtureInformation on components classification and concentrationAcuteaquatichazardMLong-termaquatic hazardM C (%)FrusthrinAcute 11Chronic 1140GladobrinM = M-factor; C = ConcentrationAcute (short-term) aquatic toxicityFish:Mixture (Cyprinus carpio)Crustacea:Mixture (Daphnia magna)Algae/aquatic plants:Mixture (Scenedesmus subspicatus)Chronic (long-term) aquatic toxicityFish:Mixture (Cyprinus carpio)Acute 11Value19 mg/l(96 hr LC 50 )3.5 mg/l(48 hr EC 50 )15 mg/l(72 or 96 hr ErC 50 )0.09 mg/l(12 d NOEC)Chronic 3-60Test method ((EC) No.440/2008) or OECDguideline / remarksC.1 / static, GLPC.2 / static, GLPC.3 / static, GLPOECD 210 / Early LifeStage, flow through, GLPCrustacea:Mixture (Daphnia magna)Algae/aquatic plants:Mixture (Scenedesmus subspicatus)0.05 mg/l(21 d NOEC)1.5 mg/l(96 h NOEC)C.20 / semi-static, GLPC.3 / static, GLPAquatic hazard classification:Acute aquatic hazard: Not classified.Long-term aquatic hazard: Chronic 1.Reasoning:470Acute aquatic hazard:Valid test data for all the three trophic levels are available for the mixture as a whole,therefore no need to consider bridging principles or classification of individual

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012components for acute hazard classification of the mixture. Test data showed thatL(E)C 50 > 1 mg/L. Consequently - no classification for acute aquatic hazard.Long-term aquatic hazard:Valid test data for all the three trophic levels are available for the mixture as a whole,therefore no need to consider classification of individual components for long-termhazard classification of the mixture. Test data showed that NOEC < 0.1 mg/l. Noinformation on rapid degradation. Hence, the mixture is considered being not rapidlydegradable. The mixture is classified as category Chronic 1.Labelling elements based on the classification:ElementGHS PictogramSignal WordHazard StatementPrecautionary statement(s)Aquatic hazardinformation thatcould appear on thelabelGHS09WARNINGH410P273, P391, P501471

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4.7.3 Example B2: When information on the classification of the components isavailable and toxicity data on the mixture as a whole is available for some, but not allthree trophic levels: use of the summation methodInformation on components classification and concentrationAcuteaquatichazardMLong-termaquatic hazardM C (%)FrusthrinAcute 11Chronic 1140GladobrinAcute 11Chronic 3-60M = M-factor; C = ConcentrationAcute (short-term) aquatic toxicityValueTest method ((EC) No.440/2008) or OECDguideline / remarksAlgae/aquatic plants:Mixture (Scenedesmus subspicatus)Chronic (long-term) aquatic toxicityAlgae/aquatic plants:Mixture (Scenedesmus subspicatus)15 mg/l(72 or 96 hr ErC 50 )1.5 mg/l(96 h NOEC)C.3 / static, GLPC.3 / static, GLPAquatic hazard classification:Acute aquatic hazard: Acute 1.Long-term aquatic hazard: Chronic 1.Reasoning: Valid test data on the mixture as a whole are available for one, but not for all the threetrophic levels and we don’t know if algae is clearly the most sensitve trophic level for themixture. Neither is valid test data on similar tested mixtures available, meaning that the bridgingprinciples could not be used.Therefore, classification should for both acute hazard and long-term hazard be consideredbased on individual components using the summation method. Testing should not beconducted for the mixture for the remaining trophic levels.Acute aquatic hazard:Information on classification including associated M-factors and the % of the components inthe mixture are available.Classify for acute hazard if: ∑ (Acute 1 M) ≥ 25%472

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Using the classification of the components of the mixture: (40 1) + (60 1) = 100 (which is≥ 25%). Hence - category Acute 1.Long-term aquatic hazard:Information on classification including associated M-factors and the % of the components inthe mixture are available.Step 1: Classify as Chronic 1 if: ∑ (Chronic 1 M) ≥ 25% (if not, then go to Step 2).Using the classification of the components of the mixture:Step 1: (40 1) = 40 (which is ≥ 25%). Hence - Category Chronic 1.Labelling elements based on the classification:ElementGHS PictogramSignal WordAquatic hazardinformation thatcould appear on thelabelGHS09WARNINGHazard Statement H410 63Precautionary statement(s)P273, P391, P50163 Note that in accordance with article 27 hazard statement H400 may be considered redundant and therefore notincluded on the label because hazard statement H410 also applies, see section 4.1.6.473

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4.7.4 Example C: When no data are available on the mixture as a whole and itscomponents, but test data are available on a similar tested mixture: use of the bridgingprinciples – dilution with waterTest SpeciesFishCrustaceaAlgaeInformation / DataNo data availableNo data availableNo data availableA reference mixture has shown a LC 50 of 0.5 mg/l and adequate NOECs in the range 0.07 to

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012ElementGHS PictogramAquatic hazardinformation thatcould appear on thelabelGHS09Signal Word -Hazard StatementPrecautionary statement(s)H411P273, P391, P501475

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.4.7.5 Example D: When test data are available for some, but not all components ofthe mixture: use of the additivity formula and of the summation methodInformation on components classification and concentrationAcuteaquatichazardMLong-termaquatic hazardM C (%)Component 1----50Component 2----10Component 3----10Component 4Notclassified-Chronic 1-30Component toxicity data:Data elements Component 1(50% of the mixture)Component 2(10% of the mixture)Test method((EC) No.440/2008) orOECD guideline /remarksPhysico-chemicalpropertiesWater solubility (S w ): 200 mg/l 1000 mg/l A.6 / pH: 7.0, non-GLPLog octanol/waterpartition coefficient(log K ow ):No dataNo dataAcute (short-term)aquatic toxicityFish:Oncorhynchus mykissNo data0.3 mg/l(96 hr LC 50 )C.1 / static, GLPCrustacea:Daphnia magna0.55 mg/l(48 hr EC 50 )No dataC.2 / static, non-GLPAlgae/aquatic plants:Scenedesmus4760.37 mg/l 1.37 mg/l

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012subspicatus (72 hr E r C 50 ) (72 hr E r C 50 ) C.3 / static, GLPLong-term aquatictoxicityFish:Oncorhynchus mykiss0.07 mg/l(28 d NOEC)1.3 mg/l(28 d NOEC)OECD 210 / semistaticCrustacea:Daphnia magna0.09 mg/l(21 d NOEC)1.4 mg/l(21 d NOEC)C.20 / semi-static,GLPAlgae/aquatic plants:ScenedesmussubspicatusDegradation(evidence of rapiddegradation)Biotic degradation (%degradation in 28 days(or, if absent, half-lifein water (d)):Abiotic degradation(Hydrolysis) (half-life(d)):BioaccumulationBioconcentrationfactor in fish (BCF):0.13 mg/L(72 hr NOEC)No dataNo dataNo data0.53 mg/L(72 hr NOEC) C.3 / static, GLPNo dataNo dataNo dataChronic classification is known for 30% of the mixture.Test data is available for 60% of the mixture.For 10% of the mixture no information is available.Aquatic hazard classification:Acute aquatic hazard: Category Acute 1.Long-term aquatic hazard: Category Chronic 1.Reasoning: Valid test data on the mixture as a whole (for all three trophic levels) are not available.477

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Valid test data on similar tested mixtures are not available, either, meaning that anybridging principle cannot be used.Therefore, classification should be considered based on individual components using thesummation method.NOTICE! In the case of the downstream user or importer not having theclassification of all the components, further dialogue with the supplier maybe necessary to obtain additional information. The suppliers in a supplychain shall cooperate to meet the requirements for classification, labellingand packaging (see CLP Article 4(9)). This particular example, however,shows what could be done if the classification of some components in anycase is not available (which, for example, could be the case whenimporting certain mixtures).Acute aquatic hazard:For component 1 the most sensitive species showed a L(E)C 50 0.37mg/l. Thus,component 1, comprising 50% of the mixture, is classified as Acute 1; M factor 1.Subsequently used in the summation method, more than 25% of the mixture isclassified as category Acute 1. Hence, the mixture is classified as Acute 1.Alternatively: You can calculate the combined toxicity for components 1 and 2applying the Additivity Formula 64 :L(E)C 50m = 60 / (50/0.37 mg/L + 10/0.3mg/L) = 0.36 mg/LAssign category Acute 1. This means that 60% of this mixture is classified as categoryAcute 1 and hence, subsequently used in the summation method, the whole mixture isclassified as Acute 1.Long-term aquatic hazard:Assign hazard categories for each component for which there are adequate chronictoxicity data available:Relevant information Category C (%)Component 10.07 mg/L Assign Chronic 1, 50 %(28 d NOEC Fish);No information ondegradation. Hence, thesubstance is considerednot rapidly degradable.M factor 1Component 20.53 mg/L Assign Chronic 2 10%(72 hr NOEC Algae);No information ondegradationComponent 3 No data - 10%Component 4 Not classified Chronic 1 30 %64 In many cases it is possible to use the summation method straight away by assigning hazard categories tosingle components of a mixture when data is available.478

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012More than 25% of the mixture is classified as category Chronic 1 and thus, the mixtureis classified as category Chronic 1.Alternatively: You can apply the Additivity Formula 65 to calculate the combinedtoxicity for components 1 and 2 (60% of the mixture)EqNOEC m = 60 / (50/(0.1 x 0.07) + 10/(0.1 x 1.3)) = 0.008 mg/l for fishEqNOEC m = 60 / (50/(0.1 x 0.09)) + 10/(0.1 x 1.4)) = 0.011 mg/l for crustaceansEqNOEC m = 60 / (50/(0.1 x 0.13) + 10/(0.1 x 0.53)) = 0.015 mg/l for algaeThe lowest calculated EqNOEC m is 0.008 mg/l.Apply table 4.1.0 b (i). Assign category Chronic 1, M factor 10 to that part of themixture.In addition component 4 of the mixture is classified as category Chronic 1 andcomprises 10% of the mixture.The long-term hazard category assigned to that part of the mixture the mixture is thensubsequently used in applying the summation method:Classify as Chronic 1 if: ∑ (Chronic 1 M) ≥ 25%∑ (60 10) + 10 = 610Thus, the mixture is classified as category Chronic 1.Labelling elements based on the classification:ElementGHS PictogramSignal WordAquatic hazardinformation thatcould appear onthe labelGHS09WARNINGHazard Statement H410 66Precautionary statement(s)P273, P391, P501In the SDS and on the label it has to be stated: “Contains 10% of components with unknownhazards to the aquatic environment”.65 See also section 4.1.4.6 of this guidance.66 Note that in accordance with CLP Article 27, the hazard statement H400 may be considered redundant andtherefore not included on the label because hazard statement H410 also applies, see section 4.1.6 of tisdocument.479

Guidance on the Application of the CLP Criteria Version 3.0 - November 20124.1.5 Metal and metal compounds4.1.2.10. Inorganic compounds and metals4.1.2.10.1. For inorganic compounds and metals, the concept of degradability as applied to organiccompounds has limited or no meaning. Rather, such substances may be transformed by normalenvironmental processes to either increase or decrease the bioavailability of the toxic species.Equally the use of bioaccumulation data shall be treated with care*.4.1.2.10.1. Poorly soluble inorganic compounds and metals may be acutely or chronically toxic inthe aquatic environment depending on the intrinsic toxicity of the bioavailable inorganic speciesand the rate and amount of this species which enter solution. All evidence must be weighed in aclassification decision. This would be especially true for metals showing borderline results in theTransformation/Dissolution Protocol._____________(*) Specific guidance has been issued by the European Chemicals Agency on how these data forsuch substances may be used in meeting the requirements of the classification criteria.”Annex IV provides the detailed guidance on the classification of metals and metalcompounds.The guidance on classification of alloys and complex metal containing materials is limited sofar. More guidance is needed (see also Annex IV 5.5.1).4.1.6 Hazard communication for hazards to the aquatic environmentA substance or mixture classified as hazardous and contained in packaging shall bear a labelin accordance with the rules in Title III of CLP. The elements to be included in labels shouldbe specified in accordance with the hazard pictograms, signal words, hazard statements andprecautionary statements which form the core information of the CLP system. For generalguidance on labelling see the Introductory Guidance on the CLP Regulation (ECHA, 2009)and also the Guidance on Labelling and Packaging in accordance with Regulation (EC) No1272/2008 (ECHA, 2011).Label elements shall be used for substances or mixtures meeting the criteria for classificationin the hazard class Hazardous to the Aquatic Environment in accordance with Table 4.1.4 ofAnnex I to CLP.PictogramThe hazard pictogram shall satisfy the provisions of Annex V and Annex I, part 1.2 to theRegulation.Symbol: Environment; Pictogram Code: GHS09The pictogram GHS09 is required only for substances or mixtures classified as:- Acute hazard category 1 and/or- Long-term hazard categories 1 or 2Signal word480

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The label shall include the relevant signal word in accordance with the classification of thehazardous substance or mixture. The signal word relevant for the hazard class Hazardous tothe Aquatic Environment is:WARNINGSignal Word Code: WngThe signal word ‘Warning’ is required only for substances or mixtures classified as:- Acute 1 and/or- Chronic 1Where the signal word ‘Danger’ is used on the label due to classification into another hazardclass(es), the signal word ‘Warning’ shall not appear on the label.Hazard statementsThe label shall include the relevant hazard statements in accordance with the classification ofthe hazardous substance or mixture and shall be worded in accordance with Annex III to CLP.The hazard statements (and the Hazard statement Codes) relevant for the hazard classHazardous to the Aquatic Environment are:- Very toxic to aquatic life (H400)- Very toxic to aquatic life with long lasting effects (H410)- Toxic to aquatic life with long lasting effects (H411)- Harmful to aquatic life with long lasting effects (H412)- May cause long lasting harmful effects to aquatic life (H413)The hazard statement H400 is required only for substances or mixtures classified as:- Acute 1The hazard statements H410 to H413 are respectively required for substances or mixturesclassified as:- Chronic 1, 2, 3 or 4Article 27 of CLP states that if a substance or mixture is classified within several hazardclasses or differentiations of a hazard class, all hazard statements resulting from theclassification shall appear on the label, unless there is evident duplication or redundancy.This means that where the hazard statement H410 is used on the label due to classificationinto long-term hazard category Chronic 1, the hazard statement H400 shall not appear on thelabel. Furthermore, where a substance or a mixture is classified both in acute and long-termhazard categories, the hazard statement required to appear on the label shall for this hazardclassification be H410 (see Table 4.1.6).481

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Table 4.1.6Aquatic hazardclassificationAssociated hazardstatementAssociated hazard statement that couldappear on the labelAcute 1 H400 H400Acute 1 and Chronic 1 H400; H410 H410Acute 1 and Chronic 2 H400; H411 H410Acute 1 and Chronic 3 H400; H412 H410Acute 1 and Chronic 4 67 H400; H413 H410Chronic 1 H410 H410Chronic 2 H411 H411Chronic 3 H412 H412Chronic 4 H413 H413Precautionary statementsIn accordance with CLP Articles 17 and 22 the label shall include the relevant precautionarystatements. The precautionary statements that can in principle be used for the hazard classHazardous to the Aquatic Environment are:- Avoid release to the environment (P273)- Collect spillage (P391)- Dispose of contents/container to … (P501)4.1.7 Re-classification of substances and mixtures classified as hazardous to theaquatic environment according to DSD/DPDFor the re-classification of substances and mixtures with regard to their hazards to the aquaticenvironment, a supplier has to apply the classification criteria specified in Annex I, part 4 ofCLP. For this reason, all available information shall be re-evaluated in order to apply thecriteria, as stated in CLP, accordingly. It is not suggested that new testing should beperformed, but instead, available information should be evaluated for its relevance andreliability.Besides the fact that M-factors need to be established for Acute 1 and Chronic 1classifications, a direct translation of classification from the DSD/DPD to CLP can only bedone in absence of chronic toxicity data. But also then, the translation for substances is notstraightforward in all cases, for example: Differences between the CLP classification and the DSD classification of substances towhich R53 - alone or in combination with R50, R51 or R52 - is applied. This is based on67 Please note that this combined classification only applies for mixtures.482

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012the slightly different criteria for classification, in particular higher cut-off values forlogKow (i.e. 4 in CLP compared to 3 in DSD) and BCF (i.e. 500 in CLP compared to 100in DSD). That means that only for those substances for which adequate chronic toxicitydata is not available, for which the long-term aquatic hazard classification is based on acombination of acute toxicity data and bioaccumulation data (without data on rapidbiodegradability affecting classification) and to which the currently applied R53 is basedexclusively on a BCF between 100 and 500 or a logK ow between 3 and 4 the classificationwould be subject to re-consideration.4.1.8 ReferencesEuropean Communities, 2003: Technical guidance Document on Risk Assessment. Part II.European Commission, Joint Research CentreOECD 2000: Series on Testing and Assessment Number 23, Guidance Document on aquatictoxicity Testing of difficult substances and mixtures. ENV/JM/MONO(2000)6OECD 2006: Series on Testing and Assessment Number 54, Current approaches in thestatistical analysis of ecotoxicity data: a guidance to application. ENV/JM/MONO(2006)18483

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012PART 5: ADDITIONAL HAZARDS5.1 HAZARDOUS TO THE OZONE LAYERThe criteria chapter for classification and labelling of substances and mixtures hazardous tothe ozone layer are short and the need for guidance is limited to the actual ODP-value thatwould trigger classification for a substance.Annex I:5.1.2 Classification criteria for substances5.1.2.1. A substance shall be classified as Hazardous to the Ozone Layer (Category 1) if the availableevidence concerning its properties and its predicted or observed environmental fate and behaviourindicate that it may present a danger to the structure and/or the functioning of the stratosphericozone layer.5.1.3 Classification criteria for mixtures5.1.3.1. Mixtures shall be classified as Hazardous to the Ozone Layer (Category 1) on the basis of theindividual concentration of the substance(s) contained therein that are also classified as Hazardousto the Ozone Layer (Category 1), in accordance with Table 5.1.Any substances having an Ozone Depleting Potential (ODP) greater or equal to the lowestODP (i.e. 0.005) of the substances currently listed in Annex I to Regulation (EC) No1005/2009 68 should be classified as hazardous to the ozone layer (category 1).68 Regulation (EC) No 1005/2009 of the European Parliament and of the Council of 16 September 2009 on substances thatdeplete the ozone layer484

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012ANNEXESIANNEX I: AQUATIC TOXICITYI.1 IntroductionThe basis for the identification of a hazard to the aquatic environment for a substance is theaquatic toxicity of that substance. Classification is predicated on having toxicity data for fish,crustacea, and algae/aquatic plant available. These taxa are generally accepted asrepresentative of aquatic fauna and flora for hazard identification. Data on these particulartaxa are more likely to be found because of this general acceptance by regulatory authoritiesand the chemical industry. Other information on the degradation and bioaccumulationbehaviour is used to better delineate the aquatic hazard. This section describes the appropriatetests for ecotoxicity, provides some basic concepts in evaluating the data and usingcombinations of testing results for classification. Further detailed guidance is given in theIntegrated Testing Strategy (ITS) for aquatic toxicity for the substance (IR/CSA (R.7A)Chapters 7.8.3 – 7.8.5).I.2 Description of testsFor classifying substances in the harmonised system, freshwater and marine species toxicitydata can be considered as equivalent data. It should be noted that some types of substances,e.g. ionisable organic chemicals or organometallic substances may express different toxicitiesin freshwater and marine environments. Since the purpose of classification is to characterisehazard in the aquatic environment, the result showing the highest toxicity should normally bechosen. However, there are circumstances where a weight of evidence approach isappropriate.The criteria for determining aquatic hazards should be test method neutral, allowing differentapproaches as long as they are scientifically sound and validated according to internationalprocedures and criteria already referred to in existing systems for the hazard of concern andproduce mutually acceptable data. Where valid data are available from non-standard testingand from non-testing methods, these shall be considered in classification provided they fulfilthe requirements specified in Section 1 of Annex XI to the REACH Regulation (EC) No1907/2006.According to the proposed system (OECD 1998):“Acute toxicity would normally be determined using a fish 96 hour LC 50 (OECD TestGuideline 203 or equivalent), a crustacea species 48 hour EC 50 (OECD Test Guideline 202or equivalent) and/or an algal species 72 or 96 hour EC 50 (OECD Test Guideline 201 orequivalent). These species are considered as surrogate for all aquatic organisms and data onother species such as the duckweed Lemna may also be considered if the test methodology issuitable.”Chronic testing involves an exposure that covers a significant period of time when compared tothe organism´s life cycle. The term can signify periods from days to a year, or more dependingon the reproductive cycle of the aquatic organism. Chronic tests can be done to assess certaininformation relating to growth, survival, reproduction and development.485

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012“Chronic toxicity data are less available than acute data and the range of testing proceduresless standardised. Data generated according to the OECD Test Guidelines 210 (Fish EarlyLife Stage), 202 Part 2 or 211 (Daphnia Reproduction) and 201 (Algal Growth Inhibition) orequivalent can be accepted. Other validated and internationally accepted tests could also beused. The NOECs or other equivalent EC x should be used.”It should be noted that several of the test guidelines cited as examples for classification arebeing revised or are being planned for updating. Such revisions may lead to minormodifications of test conditions. Therefore, the expert group that developed the harmonisedcriteria for classification intended some flexibility in test duration and/or species and numberof animals used.Guidelines for conducting acceptable tests with fish, crustacea, and algae can be found inmany sources (Test Methods Regulation 440/2008; OECD e.g. the OECD monograph No.11,Detailed Review Paper on Aquatic Toxicity Testing for Industrial Chemicals and Pesticides,1999; EPA, 1996; ASTM, 1999; ISO EU).I.2.1Fish testsI.2.1.1Acute testingAcute tests are generally performed with young juveniles 0.1 – 5 g in size for a period of 96hours. The observational endpoint in these tests is mortality. Fish larger than this range and/ordurations shorter than 96 hours are generally less sensitive. However, for classification, theycould be used if no acceptable data with the smaller fish for 96 hours are available or the resultsof these tests with different size fish or test durations would influence classification in a morehazardous category. Tests consistent with OECD Test Guideline 203 (Fish 96 hour LC 50 ) orequivalent should be used for classification.I.2.1.2Chronic testingChronic or long-term tests with fish can be initiated with fertilized eggs, embryos, juveniles,or reproductively active adults. Tests consistent with OECD Test Guideline 210 (Fish EarlyLife Stage), the fish life-cycle test (US EPA 850.1500), or equivalent can be used in theclassification scheme. Durations can vary widely depending on the test purpose (anywherefrom 7 days to over 200 days). Observational endpoints can include hatching success, growth(length and weight changes), spawning success, and survival. Technically, the OECD 210Guideline (Fish Early Life Stage) is not a “chronic” test, but a sub-chronic test on sensitivelife stages. It is widely accepted as a predictor of chronic toxicity and is used as such forpurposes of classification in the harmonised system. Fish early life stage toxicity data aremuch more available than fish life cycle or reproduction studies.I.2.2Tests with CrustaceaeI.2.2.1Acute testingAcute tests with crustacea generally begin with first instar juveniles. For daphnids, testduration of 48 hours is used. For other crustacea, such as mysids or others, duration of 96hours is typical. The observational endpoint is mortality or immobilisation as a surrogate tomortality. Immobilisation is defined as unresponsive to gentle prodding. Tests consistent withOECD Test Guideline 202 Part 1 (Daphnia acute) or USA-EPA OPPTS 850.1035 (Mysidacute toxicity) or their equivalents should be used for classification.I.2.2.2Chronic testing486

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Chronic tests with crustacea also generally begin with first instar juveniles and continuethrough maturation and reproduction. For daphnids, in particular Daphnia magna, 21 days issufficient for maturation and the production of 3 broods. For mysids, 28 days is necessary.Observational endpoints include time to first brood, number of offspring produced perfemale, growth, and survival. It is recommended that tests consistent with OECD testguidelines 211 and/or 202 Part 2 (Daphnia reproduction) or US-EPA 850.1350 (Mysidchronic) or their equivalents be used in the classification scheme.I.2.3Algae / other aquatic plant testsI.2.3.1Tests with algaeAlgae are cultured and exposed to the test substance in a nutrient-enriched medium. Testsconsistent with OECD Test Guideline 201 (Algal growth inhibition) should be used. Standardtest methods employ a cell density in the inoculum in order to ensure exponential growththrough the test, usually 3 to 4 days duration.The algal growth inhibition test is a short-term test that provides both acute and chronicendpoints. However, the EC 50 is treated as an acute value for classification purposes.Classification shall be based on both, the algal growth rate reduction endpoint, ErC 50 [= EC 50(growth rate)] and NOErC [= NOEC (growth rate)] provided that the control growth isexponential (greater than a factor of 16). This endpoint is preferred because it is notdependent on the test design, whereas the endpoint, biomass (growth) inhibition (EbC 50 )depends on both, growth rate of the test species as well as test duration and other elements oftest design. Thus in circumstances where the basis of the EC 50 is not specified and no ErC 50 isrecorded, classification shall be based on the lowest EC 50 available. Where the algal toxicityErC 50 [ = EC 50 (growth rate)] falls more than 100 times below the next most sensitive speciesand results in a classification based solely on this effect, consideration should be given towhether this toxicity is representative of the toxicity to aquatic plants. Where it can be shownthat this is not the case, professional judgment should be used in deciding if classificationshould be applied.I.2.3.2Tests with aquatic macrophytesThe most commonly used vascular plants for aquatic toxicity tests are duckweeds (Lemnagibba and L. minor). The tests last for up to 14 days and are performed in nutrient enrichedmedia similar to that used for algae, but may be increased in strength. The observationalendpoint is based on change in the number of fronds produced. Tests consistent with OECDTest Guideline on Lemna (2006) and US-EPA 850.4400 (aquatic plant toxicity, Lemna)should be used.Under the REACH Regulation growth inhibition study on aquatic plants, algae are thepreferred species.I.3 Aquatic toxicity conceptsThis section addresses the use of acute and chronic toxicity data in classification, and specialconsiderations for exposure regimes, algal toxicity testing, and use of QSARs.I.3.1Acute toxicityAcute toxicity for purposes of classification refers to the intrinsic property of a substance tobe injurious to an organism in a short-term exposure to that substance. Acute toxicity isgenerally expressed in terms of a concentration which is lethal to 50 % of the test organisms487

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(lethal concentration, LC 50 ), causes a measurable adverse effect to 50 % of the test organisms(e.g. immobilisation of daphnids, EC 50 ), or leads to a 50 % reduction in test (treated)organism responses from control (untreated) organism responses (e.g. growth rate in algae,ErC 50 ).Acute aquatic toxicity is normally determined using a fish 96 hour LC 50 , a crustacea species48 hour EC 50 , an algal species 72 or 96 hour EC 50 and/or aquatic plants 7 days EC50. Thesespecies cover a range of trophic levels and taxa and are considered as surrogate for all aquaticorganisms. Data on other species shall also be considered if the test methodology is suitable.Since the purpose of classification is to characterise hazard in the aquatic environment, theresult showing the highest toxicity should be chosen. However, there are circumstances,when a weight of evidence approach is appropriate.Substances with an acute toxicity determined to be less than one part per million (1 mg/l) aregenerally recognised as being very toxic. The handling, use, or discharge into theenvironment of these substances poses a high degree of hazard and they are classified incategory Acute 1. When classifying substances as Acute 1, it is necessary at the same time toindicate an appropriate Multiplying factor, M-factor. The multiplying factors are definedusing a toxicity value (see Section 4.1.3.3.2).I.3.2Chronic toxicityChronic toxicity, for purposes of classification, refers to the intrinsic property of a substanceto cause adverse effects to aquatic organisms during exposures which are determined inrelation to the life-cycle of the organism. Such chronic effects usually include a range ofsublethal endpoints and are generally expressed in terms of a No Observed EffectConcentration (NOEC), or an equivalent EC x . Observable endpoints typically includesurvival, growth and/or reproduction. Chronic toxicity exposure durations can vary widelydepending on the test endpoint measured and test species used.For the classification based on chronic toxicity a differentiation is made between rapidlydegradable and non-rapidly degradable substances. Substances that do rapidly degrade areclassified in category Chronic 1 when the chronic toxicity NOEC or EC x is determined to be≤ 0.01 mg/l. Decimal bands are accepted for categorising chronic toxicity above thiscategory. Substances with a chronic toxicity NOEC or EC x between 0.01 and 0.1 mg/l areclassified in category Chronic 2 for chronic toxicity. Substances with a chronic toxicityNOEC or EC x between 0.1 and 1.0 mg/l are classified in category Chronic 3 for chronictoxicity. Finally, those substances with chronic toxicity NOECs or EC x s over 1.0 mg/l are notclassifiable for long-term hazard in any of the categories Chronic 1, 2 or 3. For substancesthat do not rapidly degrade or for which such has to be assumed by worst case (i.e. thisapplies in case where no information on rapid degradation is available) two chroniccategories are used: category Chronic 1 if the chronic toxicity NOEC or EC x is determined tobe ≤ 0.1 mg/l and category Chronic 2 if the chronic toxicity NOEC or EC x is determined tobe between 0.1 and 1.0 mg/l.When classifying substances as Chronic 1, it is necessary at the same time to indicate anappropriate M-factor. The multiplying factors are defined using a toxicity value (see Section4.1.3.3.2).Since chronic toxicity data are less common in certain sectors than acute data, forclassification schemes, the potential for long-term hazard is in absence of chronic toxicitydata, is identified by appropriate combinations of acute toxicity, lack of degradability, and/or488

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012the potential or actual bioaccumulation. However, where adequate chronic toxicity data exist,this shall be used in preference over the classification based on the combination of acutetoxicity with degradability and/or bioaccumulation. In this context, the following generalapproach should be used.(a) If adequate chronic toxicity data are available for all three trophic levelsthis can be used directly to determine an appropriate long-term hazardcategory;(b)(c)If adequate chronic toxicity data are available for one or two trophiclevels, it should be examined if acute toxicity data are available for theother trophic level(s). A potential classification is made for the trophiclevel(s) with chronic data and compared with that made using the acutetoxicity data for the other trophic level(s). The final classification shallbe made according to the most stringent outcome.In order to remove or lower a long-term aquatic classification, usingchronic toxicity data, it must be demonstrated that the NOEC(s) (orequivalent ECx) used would be suitable to remove or lower the concernfor all taxa which resulted in classification based on acute data incombination with degradability, and/or bioaccumulation. This can oftenbe achieved by using a long-term NOEC for the most sensitive speciesidentified by the acute toxicity. Thus, if a classification has been basedon a fish acute LC50, it would generally not be possible to remove orlower this classification using a long-term NOEC from an invertebratetoxicity test. In this case, the NOEC would normally need to be derivedfrom a long-term fish test of the same species or one of equivalent orgreater sensitivity. Equally, if classification has resulted from the acutetoxicity of more than one taxonomic group, it is likely that NOECs fromeach taxonomic group will be needed. In case of classification of asubstance as Chronic 4, sufficient evidence should be provided that theNOEC or equivalent ECx for each taxonomic group is greater than 1mg/l or greater than the water solubility of the substances underconsideration.I.3.3Exposure regimesFour types of exposure conditions are employed in both acute and chronic tests and in bothfreshwater and saltwater media: static, static-renewal (semi-static), recirculation, and flowthrough.The choice for which test type to use usually depends on test substancecharacteristics, test duration, test species, and regulatory requirements.I.3.4Test media for algae and LemnaAlgal and Lemna tests are performed in nutrient-enriched media and use of one commonconstituent, EDTA, or other chelators, should be considered carefully. When testing thetoxicity of organic chemicals, trace amounts of a chelator like EDTA are needed to complexmicronutrients in the culture medium; if omitted, growth can be significantly reduced andcompromise test utility. However, chelators can reduce the observed toxicity of metal testsubstances. Therefore, for metal compounds, it is desirable that data from tests with highconcentration of chelators and/or tests with stoichiometrical excess of chelator relative to ironbe critically evaluated. Free chelator may mask heavy metal toxicity considerably, in489

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012particular with strong chelators like EDTA (see Annex IV to this guidance on Metals andinorganic metal compounds). However, in the absence of available iron in the medium thegrowth of algae and Lemna can become iron limited, and consequently data from tests withno or with reduced iron and EDTA should be treated with caution.I.3.5Use of substance categorisation (read across and grouping) and (Q)SARs forclassification and labellingSee Section 1.4 of this guidance.I.4 Substances which are difficult to testFor classification of organic compounds, it is desirable to have stabilised and analyticallymeasured test concentrations. Although measured concentrations are preferred, classificationmay, under certain circumstances, be based on studies where nominal concentrations are theonly valid data available. If the material is likely to substantially degrade or otherwise be lostfrom the water column, care must be taken in data interpretation and classification should bedone taking into account the loss of the toxicant during the test, if relevant and possible.Additionally, metals present their own set of difficulties and are discussed separately (seeAnnex IV on metals).In most cases where test conditions are hard to define, the actual test concentration is likely tobe less than the nominal or expected test concentration. Where acute toxicities (L(E)C 50 ) areestimated to be less than 1 mg/l for a difficult to test substance, one can be fairly confidentthe classification as Acute 1 (and Chronic 1, if appropriate) is warranted. However, if theestimated toxicity is greater than 1 mg/l, the estimated toxicity is likely to under-represent thetoxicity. In these circumstances, expert judgement is needed to determine the acceptability ofa test with a difficult substance for use in classification. In addition, caution is also neededwhen deriving appropriate M-factors, in particular when the nominal effect concentrations areclose to the thresholds for diverging M-factors. Where the nature of the testing difficulty isbelieved to have a significant influence on the actual test concentration when toxicity isestimated to be greater than 1 mg/l and the test concentration is not measured, then the testshould be used with due caution in classification.The following paragraphs provide some detailed guidance on some of these problems ofinterpretation. In doing so it should be remembered that this is guidance and hard and fastrules cannot be applied. The nature of many of the difficulties mean that expert judgementmust always be applied both in determining whether there is sufficient information in a testfor a judgement to be made on its validity, and also whether a toxicity level can bedetermined suitable for use in applying the classification criteria.I.4.1Unstable substancesWhile testing procedures should ideally have been adopted which minimise the impacts ofinstability in the test media, in practice, in certain tests, it can be almost impossible tomaintain a concentration throughout the test. Common causes of lack of constant exposureconcentration during the test are oxidation, hydrolysis, photodegradation and biodegradation.While the latter forms of degradation can more readily be controlled, such controls arefrequently absent in much existing testing. Nevertheless, for some testing, particularly acuteand chronic fish toxicity testing, a choice of exposure regimes is available to help minimiselosses due to instability, and this should be taken into account in deciding on the test datavalidity.490

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Where instability is a factor in determining the level of exposure during the test, an essentialprerequisite for data interpretation is the existence of measured exposure concentrations atsuitable time points throughout the test. In the absence of analytically measuredconcentrations at least at the start and end of test, no valid interpretation can be made and thetest should be considered as invalid for classification purposes. Where measured data areavailable, a number of practical rules can be considered by way of guidance in interpretation:(a) where measured data are available for the start and end of test (as is normal forthe acute Daphnia and algal tests), the L(E)C 50, for classification purposes, may becalculated based on the geometric mean concentration of the start and end of test.Where concentrations at the end of test are below the analytical detection limit, suchconcentrations shall be considered to be half that detection limit;(b) where measured data are available at the start and end of media renewalperiods (as may be available for the semi-static tests), the geometric mean for eachrenewal period should be calculated, and the mean exposure over the whole exposureperiod calculated from these data;(c) where the toxicity can be attributed to a degradation breakdown product, andthe concentrations of this are known, the L(E)C 50 for classification purposes may becalculated based on the geometric mean of the degradation product concentration,back calculated to the parent substance;(d) similar principles may be applied to measured data in chronic toxicity testing.I.4.2 Poorly soluble substancesThese substances, usually taken to be those with a solubility in water of < 1 mg/l, arefrequently difficult to dissolve in the test media, and the dissolved concentrations will oftenprove difficult to measure at the low concentrations anticipated. For many substances, thetrue solubility in the test media will be unknown, and will often be recorded as < detectionlimit in purified water. Nevertheless such substances can show toxicity, and where no toxicityis found, judgement must be applied to whether the result can be considered valid forclassification. Judgement should err on the side of caution and should not underestimate thehazard.Ideally, tests using appropriate dissolution techniques and with accurately measuredconcentrations within the range of water solubility should be used. Where such test data areavailable, they should be used in preference to other data. It is normal, however, particularlywhen considering older data, to find such substances with toxicity levels recorded in excess ofthe water solubility, or where the dissolved levels are below the detection limit of the analyticalmethod. Thus, in both circumstances, it is not possible to verify the actual exposureconcentrations using measured data. Where these are the only data available on which toclassify, some practical rules can be considered by way of general guidance:(a) where the acute toxicity is recorded at levels in excess of the water solubility,the L(E)C 50 for classification purposes may be considered to be equal to or below themeasured water solubility. In such circumstances it is likely that category Chronic 1and/or category Acute 1 should be applied. In making this decision, due attentionshould be paid to the possibility that the excess undissolved substance may have givenrise to physical effects on the test organisms. Where this is considered the likely causeof the effects observed, the test should be considered as invalid for classificationpurposes;491

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(b) where no acute toxicity is recorded at levels in excess of the water solubility,the L(E)C 50 for classification purposes may be considered to be greater than themeasured water solubility. In such circumstances, consideration should be given towhether the category Chronic 4 should apply. In making a decision that the substanceshows no acute toxicity, due account should be taken of the techniques used toachieve the maximum dissolved concentrations. Where these are not considered asadequate, the test should be considered as invalid for classification purposes;(c) where the water solubility is below the detection limit of the analytical methodfor a substance, and acute toxicity is recorded, the L(E)C 50 for classification purposesmay be considered to be less than the analytical detection limit. Where no toxicity isobserved, the L(E)C 50 for classification purposes, may be considered to be greaterthan the water solubility. Due consideration should also be given to the quality criteriamentioned above;(d) where chronic toxicity data are available, the same general rules should apply.In principle, only data showing no observed effect concentrations at levels above thewater solubility limit, or greater than 1 mg/l need be considered. Again, where thesedata cannot be validated by measuring the concentrations, the techniques used toachieve the maximum dissolved concentrations must be considered as appropriate.I.4.3 Other factors contributing to concentration lossA number of other factors can also contribute to losses of test material from solution and,while some can be avoided by correct study design, interpretation of data where these factorshave contributed may, from time to time, be necessary.(a) sedimentation: this can occur during a test for a number of reasons. A commonexplanation is that the substance has not truly dissolved despite the apparent absence ofparticulates, and agglomeration occurs during the test leading to precipitation. In thesecircumstances, the L(E)C 50 for classification purposes, may be considered to be based onthe end of test concentrations. Equally, precipitation can occur through reaction with themedia. This is considered under instability above;(b) adsorption: this can occur for substances of high adsorption characteristics suchas high log Kow substances. Where this occurs, the loss of concentration is usually rapidand exposure may best be characterised by the end of test concentrations;(c) bioaccumulation: losses may occur through the bioaccumulation of asubstance into the test organisms. This may be particularly important where the watersolubility is low and log K ow correspondingly high. The L(E)C 50 for classificationpurposes, may be calculated based on the geometric mean of the start and end of testconcentrations.I.4.4Perturbation of the test mediaStrong acids and bases may exert their toxicity through extreme pH. Generally howeverchanges of the pH in aquatic systems are normally prevented by buffer systems in the testmedium. If no data are available on a salt, the salt should generally be classified in the sameway as the anion or cation, i.e. as the ion that receives the most stringent classification. If theeffect concentration is related to only one of the ions, the classification of the salt should takethe molecular weight difference into consideration by correcting the effect concentration bymultiplying with the ratio: MW salt /MW ion.492

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Polymers are typically not available in aquatic systems. Dispersible polymers and other highmolecular mass materials can perturb the test system and interfere with uptake of oxygen, andgive rise to mechanical or secondary effects. These factors need to be taken into accountwhen considering data from these substances. Many polymers behave like complexsubstances, however, having a significant low molecular mass fraction which can leach fromthe bulk polymer. This is considered further below.I.4.5Complex substancesComplex substances are characterised by a range of chemical structures, frequently in ahomologous series, but covering a wide range of water solubilities and other physicochemicalcharacteristics. On addition to water, equilibrium will be reached between thedissolved and undissolved fractions which will be characteristic of the loading of thesubstance. For this reason, such complex substances are usually tested as a WSF or WAF,and the L(E)C 50 recorded based on the loading or nominal concentrations. Analytical supportdata are not normally available since the dissolved fraction will itself be a complex mixtureof components. The toxicity parameter is sometimes referred to as LL 50 , related to the lethalloading level. This loading level from the WSF or WAF may be used directly in theclassification criteria.Polymers represent a special kind of complex substance, requiring consideration of thepolymer type and their dissolution/dispersal behaviour. Polymers may dissolve as suchwithout change, (true solubility related to particle size), be dispersible, or portions consistingof low molecular weight fractions may go into solution. In the latter case, in effect, the testingof a polymer is a test of the ability of low molecular mass material to leach from the bulkpolymer, and whether this leachate is toxic. It can thus be considered in the same way as acomplex mixture in that a loading of polymer can best characterise the resultant leachate, andhence the toxicity can be related to this loading.I.5 ReferencesUS EPA 1996. Ecological Effects Test Guidelines - OPPTS Harmonized Test GuidelinesSeries 850.1000 -- Public Drafts, EPA 712-C-96-113. http://www.epa.gov/opptsfrs/publications/OPPTS_Harmonized/850_Ecological_Effects_Test_Guidelines/Drafts/ASTM 1999. Annual book of ASTM standards, Vol. 11.04. American Society for Testingand Materials, Philadelphia, PAISO guidelines: http://www.iso.org/iso/iso_catalogue.htmTest Methods Regulation (EC) No 440/2008.http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2008:142:0001:0739:en:PDF493

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012IIANNEX II: RAPID DEGRADATIONII.1IntroductionDegradability is one of the important properties of substances that have impact on thepotential for substances to exert an aquatic hazard. Non-degradable substances will persist inthe environment and may consequently have a potential for causing long-term adverse effectson biota. In contrast, degradable substances may be removed in the sewers, in sewagetreatment plants or in the environment. It should be noted that data from degradability testson mixtures are difficult or impossible to interpret, and are therefore not used in classificationand labelling.Classification of substances is primarily based on their intrinsic properties. However, thedegree of degradation depends not only on the intrinsic degradability or recalcitrance of themolecule, but also on the actual conditions in the receiving environmental compartment suchas redox potential, pH, temperature, presence of suitable micro-organisms, concentration ofthe substance and occurrence and concentration of other substrates. The interpretation of thedegradation properties in an aquatic hazard classification context therefore requires detailedcriteria that balance the intrinsic properties of the substance and the prevailing environmentalconditions into a concluding statement on the potential for long-term adverse effects.The term degradation is defined in Section 4.1 of Annex I to CLP as “the decomposition oforganic molecules to smaller molecules and eventually to carbon dioxide, water and salts”.For inorganic compounds and metals, the concept of degradability has no meaning. Ratherthe substance may be transformed by normal environmental processes to either increase ordecrease the bioavailability of the toxic species. Therefore, the present section applies only toorganic and organo-metal compounds. A separate section on the classification & labelling(C&L) of metals is provided in Part 4, section 4.1.5 and Annex IV to the CLP guidance.Data on degradation properties of a substance may be available from standardised tests, orfrom other types of investigations, or they may be estimated from the structure of themolecules i.e. via SAR or QSAR approaches. The interpretation of such degradation data forclassification purposes often requires detailed evaluation of the (test) data. The use ofbiodegradation data for classification purposes is only applicable to substances.Biodegradation data on mixtures cannot be used as it does not provide a reliable indication ofenvironmental fate (CLP Annex I, point 4.1.3.3.1).II.2Interpretation of degradability dataOften a diverse range of test data is available that does not necessarily fit directly with theclassification criteria. Consequently, guidance is needed on interpretation of existing test datain the context of the aquatic hazard classification. Based on the harmonised criteria, guidancefor interpretation of degradation data is prepared below for several types of data comprisedby the expression “rapid degradation” in the aquatic environment.494

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012II.2.1Ready biodegradabilityReady biodegradability is defined in the OECD Test Guidelines No. 301 methods A-F(OECD 1992), OECD 306 (marine water) and OECD 310 (OECD 2006). All organicsubstances that degrade to a level higher than the pass level in a standard OECD readybiodegradability test or in a similar test should be considered readily biodegradable, andconsequently also rapidly degradable. Many test data found in the open literature, however,do not specify all of the conditions that should be evaluated to demonstrate whether or not thetest fulfils the requirements of a ready biodegradability test. Expert judgement is thereforeneeded as regards the validity of the data before use for classification purposes. Beforeconcluding on the ready biodegradability of a test substance, however, at least the followingparameters should be considered.II.2.1.1Concentration of test substanceRelatively high concentrations of test substance are used in the OECD ready biodegradabilitytests (2-100 mg/l). Many substances may however be toxic to the inocula at such highconcentrations, resulting in a low degradation of the substances in these tests, although thesubstances might be rapidly degradable at lower non-toxic concentrations. A toxicity testwith micro-organisms, or inhibition of the inoculum observed with a positive controlsubstance may demonstrate the toxicity of the test substance. Guidance on the selection ofsuitable microbial inhibition test methods is provided in IR/CSA Parts R7.8.14. When it islikely that inhibition is the reason for a substance being not readily degradable, results from atest employing lower non-toxic concentrations of the test substance should be used whenavailable.II.2.1.2Time windowThe harmonised criteria include a general requirement for all of the ready biodegradabilitytests on achievement of the pass level within ten days. This is not in line with the OECD TestGuideline 301 in which the ten-day time window applies to the OECD ready biodegradabilitytests except to the MITI I test (OECD Test Guideline 301C). In the Closed Bottle test (OECDTest Guideline 301D), a 14-days window may be used instead when measurements have notbeen made after ten days. Moreover, often only limited information is available in referencesof biodegradation tests. Thus, as a pragmatic approach the percentage of degradation reachedafter 28 days may be used directly for assessment of ready biodegradability when noinformation on the ten days time window is available. This should, however, only beaccepted for existing test data and data from tests where the ten-day window does not apply.Where there is sufficient justification, the ten-day window condition may be waived forcertain complex substances and the pass level is applied at 28 days. This applies to multiconstituentand certain UVCB substances (such as oils and surfactants) consisting ofstructural similar constituents with different chain-lengths, degree and/or site of branching orstereo-isomers, even in their most purified commercial forms. Testing of each individualconstituent may be costly and impractical. If a test on such a complex substance is performedand it is anticipated that a sequential biodegradation of the individual constituents is takingplace, then the ten-day window should not be applied to interpret the results of the test. Acase by case evaluation should however take place on whether a biodegradability test on sucha substance would give valuable information regarding its biodegradability as such i.e.regarding the degradability of all the constituents, or whether instead an investigation of thedegradability of carefully selected individual constituents of the complex substance isrequired (OECD 2006).495

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012II.2.2BOD 5 /CODInformation on the 5-day biochemical oxygen demand (BOD 5 ) will be used for classificationpurposes only when no other measured degradability data are available. Thus, priority isgiven to data from ready biodegradability tests and from simulation studies regardingdegradability in the aquatic environment. Therefore, this test should not be performedanymore for assessment of the ready biodegradability of substances. Older test data mayhowever be used when no other degradability data are available. For substances where thechemical structure is known, the theoretical oxygen demand (ThOD) can be calculated andthis value should be used instead of the chemical oxygen demand (COD).II.2.3Other convincing scientific evidenceRapid degradation in the aquatic environment may be demonstrated by other data than aready biodegradability test, or a BOD 5 /COD ratio. These may be data on biotic and/or abioticdegradation. Data on primary degradation can only be used where it is demonstrated that thedegradation products shall not be classified as hazardous to the aquatic environment, i.e. thatthey do not fulfil the classification criteria.The fulfilment of criterion (c) of paragraph 4.1.2.9.5 of CLP requires that the substance isdegraded in the aquatic environment to a level of > 70 % within a 28-day period. If first-orderkinetics are assumed, which is reasonable at the low substance concentrations prevailing inmost aquatic environments, the degradation rate will be relatively constant for the 28-dayperiod. Thus, the degradation requirement will be fulfilled with an average degradation rateconstant, k > -(ln 0.3 - ln 1)/28 = 0.043 day -1 . This corresponds to a degradation half-life, t ½

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012II.2.3.2(d) realistic temperature e.g. 5 °C to 25 °C; and(e) ultimate degradation is determined i.e. determination of themineralisation rate or the individual degradation rates of the totalbiodegradation pathway.Field investigationsParallel to laboratory simulation tests are field investigations or mesocosm experiments. Insuch studies, fate and/or effects of chemicals in the environment or in environmentalenclosures may be investigated. Fate data from such experiments can in principle be used forassessing the potential for a rapid degradation. This may, however, often be difficult, as itrequires that ultimate degradation can be demonstrated. This may be documented bypreparing mass balances showing that no non-degradable intermediates are formed, andwhich take the fractions into account that are removed from the aqueous system due to otherprocesses such as sorption to sediment or volatilisation from the aquatic environment.II.2.3.3Monitoring dataMonitoring data may demonstrate the removal of contaminants from the aquatic environment.Such data are, however, very difficult to use for classification purposes. The followingaspects should be considered before use:(a) Is the removal a result of degradation, or is it a result of other processessuch as dilution or distribution between compartments (sorption,volatilisation)?(b) Is formation of non-degradable intermediates excluded?Only when it can be demonstrated that removal as a result of ultimate degradation fulfils thecriteria for rapid degradability, can such data be considered for use for classificationpurposes. In general, monitoring data should only be used as supporting evidence fordemonstration of either persistence in the aquatic environment, or of rapid degradation.II.2.3.4Inherent and Enhanced Ready Biodegradability testsSubstances that are degraded more than 70% in tests for inherent biodegradability (OECDTest Guidelines 302) have the potential for ultimate biodegradation. However, because of theoptimised conditions in these tests, the rapid biodegradability of inherently biodegradablesubstances in the environment cannot be assumed. The optimised conditions in inherentbiodegradability tests stimulate adaptation of the micro-organisms thus increasing thebiodegradation potential, compared to natural environments. Therefore, positive results ingeneral should not be interpreted as evidence for rapid degradation in the environment.IR/CSA Chapters R.7B and R.11 refer in the context of persistence testing to a new categoryof tests, i.e. the ‘enhanced ready (screening) biodegradability tests’. These are in essenceready biodegradability tests to which more flexibility is given to demonstrate the occurrenceof degradation e.g. via prolonged testing times, larger test volumes, adaptation, etc. Thesemethods are not yet validated and/or standardised for C&L.II.2.3.5Sewage treatment plant simulation testsResults from tests simulating the conditions in a sewage treatment plant (STP) e.g. the OECDTest Guideline 303 cannot be used for assessing the degradation in the aquatic environment.The main reasons for this are that the microbial biomass in a STP is significantly different497

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012from the biomass in the environment, that there is a considerably different composition ofsubstrates, and that the presence of rapidly mineralised organic matter in waste water mayfacilitate degradation of the test substance by co-metabolism.II.2.3.6Soil and sediment degradation dataIt has been argued that for many non-sorptive substances more or less the same degradationrates are found in soil and in surface water. For sorptive substances, a lower degradation ratemay generally be expected in soil than in water due to a lower bioavailability caused bysorption. Thus, when a substance has been shown to be degraded rapidly in a soil simulationstudy, it is most likely also rapidly degradable in the aquatic environment. It is thereforeproposed that an experimentally determined rapid degradation in soil is sufficientdocumentation for a rapid degradation in surface waters when:(a) no pre-exposure (pre-adaptation) of the soil micro-organisms has takenplace, and(b)an environmentally realistic concentration of substance is tested, and(c) the substance is ultimately degraded within 28 days with a half-life < 16days corresponding to a degradation rate > 0.043 day -1 .The same argumentation is considered valid for data on degradation in sediment underaerobic conditions.II.2.3.7Anaerobic degradation dataData regarding anaerobic degradation cannot be used in relation to deciding whether asubstance should be regarded as rapidly degradable, because the aquatic environment isgenerally regarded as the aerobic compartment where the aquatic organisms, such as thoseemployed for aquatic hazard classification, live.II.2.3.8HydrolysisData on hydrolysis e.g. OECD Test Guideline 111 might be considered for classificationpurposes only when the longest half-life t ½ determined within the pH range 4-9 is shorter than16 days. However, hydrolysis is not an ultimate degradation and various intermediatedegradation products may be formed, some of which may be only slowly degradable. Onlywhen it can be satisfactorily demonstrated that the hydrolysis products formed do not fulfilthe criteria for classification as hazardous for the aquatic environment, data from hydrolysisstudies could be considered.When a substance is quickly hydrolysed e.g. with t ½ < a few days, this process is a part of thedegradation determined in biodegradation tests. Hydrolysis may be the initial transformationprocess in biodegradation.II.2.3.9Photochemical degradationInformation on photochemical degradation e.g. OECD 1997 is difficult to use forclassification purposes. The actual degree of photochemical degradation in the aquaticenvironment depends on local conditions e.g. water depth, suspended solids, turbidity as wellas seasonal influences, and the hazard of the degradation products is usually not known.Probably only seldom will enough information be available for a thorough evaluation basedon photochemical degradation.498

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012II.2.3.10Estimation of degradationHydrolysis: Certain QSARs have been developed for prediction of an approximate hydrolysishalf-life, which should only be considered when no experimental data are available, or in aWeight of Evidence approach. However, a hydrolysis half-life can only be used with greatcare in relation to classification, because hydrolysis does not concern ultimate degradability(see “Hydrolysis” of this Section). Furthermore the QSARs developed until now have a ratherlimited applicability and are only able to predict the potential for hydrolysis on a limitednumber of chemical classes (see also IR/CSA Chapter R.7.9.3.1).Biodegradation: In general, no quantitative estimation method (QSAR) for estimating thedegree of biodegradability of organic substances is yet sufficiently accurate to unequivocallypredict rapid degradation. However, results from such methods may be used to predict that asubstance is not rapidly degradable, or be used in a Weight of Evidence approach. Forexample, when in the Biodegradation Probability Program e.g. BIOWIN version 3.67,Syracuse Research Corporation the probability is < 0.5 estimated by the linear or non-linearmethods, the substances should be regarded as not rapidly degradable (OECD, 1994;Pedersen et al., 1995 & Langenberg et al., 1996). Also other (Q)SAR methods may be usedas well as expert judgement, for example, when degradation data for structurally analoguecompounds are available, but such judgement should be conducted with great care. See alsoIR/CSA Chapter R.7.9.3.1.In general, a QSAR prediction that the substance is not rapidly degradable is considered abetter documentation for classification than application of a default classification, when nouseful degradation data are available.Degradation data from structurally related substances may provide evidence that a givensubstance displays very similar degradation properties. Such information may be employedin a read-across or weight of evidence approach for C&L.II.2.3.11VolatilisationChemicals may be removed from some aquatic environments by volatilisation. The intrinsicpotential for volatilisation is determined by the Henry's Law constant (H) of the substance.Volatilisation from the aquatic environment is highly dependent on the environmentalconditions of the specific water body in question, such as the water depth, the gas exchangecoefficients (depending on wind speed and water flow) and stratification of the water body.Because volatilisation only represents removal of a chemical from the water phase, and notdegradation, the Henry's Law constant cannot be used for assessment of degradation inrelation to aquatic hazard classification of substances (see also Pedersen et al., 1995).II.2.4No degradation data availableWhen no useful data on degradability are available - either experimentally determined orestimated data - the substance should be regarded by default as not rapidly degradable.II.3General interpretation problemsII.3.1Complex substancesThe harmonised criteria for classification of chemicals as hazardous for the aquaticenvironment focus on single substances. Some intrinsically complex substances are multiconstituentsubstances. They are typically of natural origin and need occasionally to beconsidered. This may be the case for chemicals that are produced or extracted from mineral499

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012oil or plant material. Such complex chemicals are normally considered as single substances ina regulatory context. In most cases they are defined as a homologous series of substanceswithin a certain range of carbon chain length and/or degree of substitution. When this is thecase, no major difference in degradability is foreseen and the degree of degradability can beestablished from tests of the complex chemical. One exception would be when a borderlinedegradation is found because in this case some of the individual substances may be rapidlydegradable and others may not be rapidly degradable. This requires a more detailedassessment of the degradability of the individual constituents in the complex substance.When the constituents that are not-rapidly-degradable constitute a significant part of thecomplex substance e.g. more than 20 %, or for a hazardous constituent, an even lowercontent, the substance should be regarded as not rapidly degradable.II.3.2Availability of the substanceThe present standard methods for investigating degradability of substances are developed forreadily soluble test compounds. However, many organic substances are only slightly solublein water. As the standard tests require 2-100 mg/l of the test substance, sufficient availabilitymay not be reached for substances with low water solubility. In general, the DOC Die-Awaytest (OECD Test Guideline 301A) and the Modified OECD Screening test (OECD TestGuideline 301E) are less suitable for testing the biodegradability of poorly soluble substancessince adsorption may be confused with degradation. In such cases, test adaptations may beconsidered with e.g. continuous mixing and/or an increased exposure time. Also tests with aspecial design, where concentrations of the test substance lower than the water solubility havebeen employed e.g. with radiolabelled test chemicals, could be relevant.II.3.3Test duration less than 28 daysSometimes degradation is reported for tests terminated before the 28 day period specified inthe standards e.g. the MITI, 1992. These data are of course directly applicable when adegradation greater than or equal to the pass level is obtained. When a lower degradationlevel is reached, the results need to be interpreted with caution. One possibility is that theduration of the test was too short and that the chemical structure would probably have beendegraded in a 28-day biodegradability test. If substantial degradation occurs within a shorttime period, the situation may be compared with the criterion BOD 5 /COD 0.5 or with therequirements on degradation within the 10-days time window. In these cases, a substancemay be considered readily degradable (and hence rapidly degradable), if:(a) the ultimate biodegradability exceeds 50 % within 5 days; or(b) the ultimate degradation rate constant in this period is greater than 0.1day -1 corresponding to a half-life of 7 days.These criteria are proposed in order to ensure that rapid mineralisation did occur, althoughthe test was ended before 28 days and before the pass level was attained. Interpretation of testdata that do not comply with the prescribed pass levels must be made with great caution. It ismandatory to consider whether a biodegradability result below the pass level was due to apartial degradation of the substance and not a complete mineralisation. If partial degradationis the probable explanation for the observed biodegradability, the substance should beconsidered not readily biodegradable.II.3.4Primary biodegradation500

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012In some tests, only the disappearance of the parent compound i.e. primary degradation isdetermined for example by following the degradation by specific or group specific chemicalanalyses of the test substance. Data on primary biodegradability may be used fordemonstrating rapid degradability only when it can be satisfactorily demonstrated that thedegradation products formed do not fulfil the criteria for classification as hazardous to theaquatic environment.II.3.5Conflicting results from screening testsThe situation where more degradation data are available for the same substance introducesthe possibility of conflicting results. In general, conflicting results for a substance which hasbeen tested several times with an appropriate biodegradability test could be interpreted by a“weight of evidence approach”. This implies that if both positive i.e. higher degradation thanthe pass level and negative results have been obtained for a substance in readybiodegradability tests, then the data of the highest quality and the best documentation shouldbe used for determining the ready biodegradability of the substance. However, positiveresults in ready biodegradability tests could be considered valid, irrespective of negativeresults, when the scientific quality is good and the test conditions are well documented, i.e.guideline criteria are fulfilled, including the use of non-pre-exposed (non-adapted) inoculum.The suitability of the inoculum for degrading the test substance depends on the presence andamount of competent degraders. When the inoculum is obtained from an environment thathas previously been exposed to the test substance, the inoculum may be adapted asdemonstrated by a degradation capacity greater than that of an inoculum from a non-exposedenvironment. As far as possible the inoculum must be sampled from an unexposedenvironment, but for substances that are used ubiquitously in high volumes and releasedwidespread or more or less continuously, this may be difficult or impossible. Whenconflicting results are obtained, the origin and density of the inoculum should be checked inorder to clarify whether or not differences in the adaptation of the microbial community maybe the reason.As mentioned above, many substances may be toxic or inhibitory to the inoculum at therelatively high concentrations tested in ready biodegradability tests. Especially in theModified MITI (I) test (OECD Test Guideline 301C) and the Manometric Respirometry test(OECD Test Guideline 301F) high concentrations (100 mg/l) are prescribed. The lowest testsubstance concentrations are prescribed in the Closed Bottle test (OECD Test Guideline301D) where 2-10 mg/l is used. The possibility of toxic effects may be evaluated byincluding a toxicity control in the ready biodegradability test or by comparing the testconcentration with toxicity test data on micro-organisms (for test methods see IR/CSAChapter R.7.8.14).Volatile substances should only be tested in closed systems as the Closed Bottle test (OECDTest Guideline 301D), the MITI I test (OECD Test Guideline 301C) the ManometricRespirometry test (OECD Test Guideline 301F), or OECD 310 (CO2 in sealed vessels –Headspace Test). Results from other tests should be evaluated carefully and only consideredif it can be demonstrated, e.g. by mass balance estimates, that the removal of the testsubstance is not a result of volatilisation.II.3.6Variation in simulation test dataA number of simulation test data may be available for certain high priority chemicals. Oftensuch data provide a range of half-lives in environmental media such as soil, sediment and/orsurface water. The observed differences in half-lives from simulation tests performed on the501

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012same substance may reflect differences in test conditions, all of which may beenvironmentally relevant. A suitable half-life in the higher end i.e. a realistic worst case ofthe observed range of half-lives from such investigations should be selected for classificationby employing a weight of evidence approach and taking the realism and relevance of theemployed tests into account in relation to environmental conditions. In general, simulationtest data of surface water are preferred relative to aquatic sediment or soil simulation test datain relation to the evaluation of rapid degradability in the aquatic environment.II.4Decision schemeThe following decision scheme may be used as a general guidance to facilitate decisions inrelation to rapid degradability in the aquatic environment and classification of chemicalshazardous to the aquatic environment.A substance is considered to be not rapidly degradable unless at least one of the following isfulfilled:a) The substance is demonstrated to be readily biodegradable in a 28-day test for readybiodegradability. The pass level of the test (70 % DOC removal or 60 % theoreticaloxygen demand) must be achieved within 10 days from the onset of biodegradation, ifit is possible to evaluate this according to the available test data (the ten-day windowcondition may be waived for complex multi-component substances and the pass levelapplied at 28 days, as discussed in II.2.3). If this is not possible, then the pass levelshould be evaluated within a 14 days time window if possible, or after the end of thetest; orb) The substance is demonstrated to be ultimately degraded in a surface water simulationtest 3 with a half-life of < 16 days (corresponding to a degradation of >70 % within 28days); orc) The substance is demonstrated to be primarily degraded biotically or abiotically e.g.via hydroysis, in the aquatic environment with a half-life 70 % within 28 days), and it can be demonstrated that thedegradation products do not fulfill the criteria for classification as hazardous to theaquatic environment.When these preferred data types are not available rapid degradation may be demonstratedif one of the following criteria is justified:a) The substance is demonstrated to be ultimately degraded in an aquatic sediment orsoil simulation test 3 with a half-life of < 16 days (corresponding to a degradationof > 70 % within 28 days); orb) In those cases where only BOD 5 and COD data are available, the ratio ofBOD 5 /COD is greater than or equal to 0.5. The same criterion applies to readybiodegradability tests of a shorter duration than 28 days, if the half-lifefurthermore is < 7 days; orc) A weight of evidence approach based on read-across provides convincingevidence that a given substance is rapidly degradable.If none of the above types of data are available then the substance is considered as notrapidly degradable. This decision may be supported by fulfilment of at least one of thefollowing criteria:502

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(i)(ii)(iii)(iv)the substance is not inherently degradable in an inherent biodegradabilitytest; orthe substance is predicted to be slowly biodegradable by scientificallyvalid QSARs, e.g. for the Biodegradation Probability Program, the scorefor rapid degradation (linear or non-linear model) < 0.5; orthe substance is considered to be not rapidly degradable based on indirectevidence, such as knowledge from structurally similar substances; orno other data regarding degradability are available.II.5ReferencesOECD (2006). Revised introduction to the OECD guidelines for testing of chemicals, section3. OECD, 23 March 2006.OECD (1992), OECD Test Guidelines No. 301 methods A-FPedersen, F., H. Tyle, J. R. Niemeldi, B. Guttmann, L. Lander, and A. Wedebrand 1995.Environmental Hazard Classification – data collection and interpretation guide. TemaNord1995:581Langenberg J.H., W.J.G.M. Peijnenburg & E. Rorije (1996). On the usefulness and reliabilityof existing QSBRs for risk assessment and priority setting. SAR and QSAR in EnvironmentalResearch 5, 1-16503

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012IIIANNEX III: BIOACCUMULATIONIII.1IntroductionBioaccumulation of a substance by an organism is not in itself a hazard. However, thebioaccumulation of a substance should be considered in relation to the potential for thatsubstance to exert long-term effects. Chemical concentration and accumulation may result ininternal concentrations of a substance in an organism (body burden), which may or may notlead to toxic effects over long-term exposures. For most organic chemicals uptake from water(bioconcentration) is believed to be the predominant route of uptake. Only for veryhydrophobic substances does uptake from food become important. The classification criteriause the bioconcentration factor (BCF) or in the absence of it the octanol/water partitioncoefficient (log K ow ) as the measure of the potential for bioaccumulation. For these reasons,the present guidance document mainly considers bioconcentration and does not discuss indetail uptake via food or other routes. However, the possibility to use information on thebiomagnification factor (BMF) as supportive evidence for bioaccumulation of highlylipophilic substances may be taken into account on a case by case basis.Classification of a substance is primarily based on its intrinsic properties. However, thedegree of bioconcentration also depends on factors such as the degree of bioavailability, thephysiology of test organism, maintenance of constant exposure concentration, exposureduration, metabolism inside the body of the target organism and excretion from the body. Theinterpretation of the bioconcentration potential in a chemical classification context thereforerequires an evaluation of the intrinsic properties of the substance, as well as of theexperimental conditions under which bioconcentration factor (BCF) has been determined.IR/CSA (R.7C) Chapter 7.10.5.1 discusses the suitability of bioconcentration data, log K owdata and other information (e.g. evidence for limited bioaccumulation potential) forclassification purposes. Use of measured biomagnification data is discussed in relation to thescreening approach in IR/CSA (R.7C) Chapter 7.10.4.5. Bioaccumulation of metals isdiscussed in Annex IV.Information on the bioaccumulation potential of a substance may be available fromstandardised tests or may be estimated from the structure of the molecule. The interpretationof such bioconcentration data for classification purposes often requires detailed evaluation oftest data. Guidance has been developed in IR/CSA in order to facilitate this evaluation.Chapter 7.1.8 (R.7A) gives guidance on n-octanol/water partition coefficient and Chapter7.10.4 (R.7C) gives guidance on how to evaluate laboratory data on aquatic bioaccumulation.The use of bioaccumulation data for classification purposes is only applicable to substances.Bioaccumulation data on mixtures cannot be used as it does not provide a reliable indicationof environmental fate (CLP Annex I, point 4.1.3.3.1).III.2Interpretation of bioconcentration dataAquatic hazard classification of a substance is normally based on existing data on itsenvironmental properties. Test data will only seldom be produced with the main purpose offacilitating a classification. Often a diverse range of test data is available which does notnecessarily match the classification criteria. Further guidance on how to use this data is givenin Chapter 7.10.5 of IR/CSA (R.7C).Bioconcentration of an organic substance can be experimentally determined inbioconcentration experiments, during which BCF is measured as the concentration in theorganism relative to the concentration in water under steady-state conditions and/or estimated

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012from the uptake rate constant and the elimination rate constant. In general, the potential of anorganic substance to bioconcentrate is primarily related to the lipophilicity of the substance. Ameasure of lipophilicity is the n-octanol/water partition coefficient (K ow ) which, for lipophilicnon-ionised organic substances, undergoing minimal metabolism or biotransformation withinthe organism, is correlated with the bioconcentration factor. Therefore, K ow is often used forestimating the bioconcentration of non-ionised organic substances, based on the empiricalrelationship between log BCF and log K ow . For those organic substances, estimation methodsare available for calculating the K ow . Data on the bioconcentration properties of non-ionisedorganic substances may thus be (i) experimentally determined, (ii) estimated fromexperimentally determined K ow , or (iii) estimated from K ow values derived by use ofQuantitative Structure Activity Relationships (QSARs). Guidance for interpretation of suchdata is given in Chapters 7.10.4 and 7.10.5 of IR/CSA (R.7C). Guidance is also given onionised chemicals and other classes that need special attention (see section III.3.1).III.2.1Bioconcentration factor (BCF)The bioconcentration factor is defined as the ratio on a weight basis between theconcentration of the chemical in biota and the concentration in the surrounding medium, herewater, at steady state. BCF can thus be experimentally derived under steady-state conditions,on the basis of measured concentrations. In addition BCF can also be calculated as the ratiobetween the first-order uptake and elimination rate constants; a method which does notrequire steady state (equilibrium conditions).Different test guidelines for the experimental determination of bioconcentration in fish havebeen documented and adopted, the most generally applied being the OECD test guideline 30569 (OECD, 1996; C.13 in Test Methods Regulation 440/2008 is a corresponding test).Experimentally derived BCF values of high quality studies are ultimately preferred forclassification purposes as such data override surrogate data, e.g. K ow .High quality data are defined as data where the validity criteria for the test method applied arefulfilled and described. Further guidance is provided in Chapter 7.10.4 of IR/CSA (R.7C).BCF results from poor or questionable quality may give an erroneous BCF value. Therefore,such data should be carefully evaluated before use and consideration should be given to usingK ow instead.If there is no BCF value for fish species, high-quality data on the BCF value for invertebratespecies may be used. An invertebrate (mussel, oyster or scallop) BCF can be used as a worstcase (conservative) value for fish. BCF for algae should not be used.Experimental BCF data on highly lipophilic substances (e.g. with log K ow above 6) will havea higher level of uncertainty than BCF values determined for less lipophilic substances. Forhighly lipophilic substances, e.g. with log K ow above 6, experimentally derived BCF valuestend to decrease with increasing log K ow . Conceptual explanations of this non-linearity mainlyrefer to either reduced membrane permeation kinetics or reduced biotic lipid solubility forlarge molecules. A low bioavailability and uptake of these substances in the organism willthus occur. Other factors comprise experimental artifacts, such as equilibrium not beingreached, reduced bioavailability due to sorption to organic matter in the aqueous phase, andanalytical errors. Special care should thus be taken when evaluating experimental data onBCF for highly lipophilic substances as these data will have a much higher level ofuncertainty than BCF values determined for less lipophilic substances.69 Note that OECD 305 is currently under revision. All adopted OECD guidelines can be freely accessed via theOECD iLibrary.505

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012III.2.1.1BCF in different test speciesBCF values used for classification are based on whole body measurements. As statedpreviously, the optimal data for classification are BCF values derived using the OECD testguideline 305 or corresponding EU test guideline C.13 or internationally equivalent methods,which uses small fish. Due to the higher gill surface-to-weight ratio in smaller organisms thanin larger ones, steady-state conditions will be reached sooner in smaller organisms than inlarger ones. The size of the organisms (fish) used in bioconcentration studies is thus ofconsiderable importance in relation to the time used in the uptake phase, when the reportedBCF value is based solely on measured concentrations in fish and water at steady-state. Thus,if large fish, e.g. adult salmon, have been used in bioconcentration studies, it should beevaluated whether the uptake period was sufficiently long for steady state to be reached or toallow for a kinetic uptake rate constant to be determined precisely. Also possible growthdilution should be taken into account when calculating the BCF values for smaller fish thatgrow during the bioconcentration studies.Furthermore, when using existing data for classification, it is possible that the BCF valuescould be derived from several different fish or other aquatic species (e.g. clams) and fordifferent organs in the fish. Thus, to compare diverse measured BCF data from differentspecies to each other and to the criteria, normalisation to common basis lipid content will berequired to reduce variability. Detailed guidance can be found in IR/CSA (R.7C) Chapter7.10.4.1 for 'correction factors'.Generally, the highest valid BCF value expressed on this common lipid basis is used todetermine the wet weight based BCF-value in relation to the cut off value for BCF of 500 ofthe classification criteria.III.2.1.2Use of radio-labelled substancesThe use of radio-labelled test substances can facilitate the analytical measurents in water andfish samples. However, unless combined with a specific analytical method, the totalradioactivity measurements potentially reflect the presence of the parent substance as well aspossible metabolite(s) and possible metabolised carbon, which have been incorporated in thefish tissue in organic molecules. BCF values determined by use of radio-labelled testsubstances are therefore normally overestimated.When using radio-labelled substances, the labelling is most often placed in the stable part ofthe molecule, for which reason the measured BCF value includes the BCF of the metabolitesas well as the BCF from the parent substance. For some substances it is the metabolite whichis the most toxic or which has the highest bioconcentration potential. Selective measurementsof the parent substance as well as the metabolites may thus be important for the interpretationof the aquatic hazard (including the bioconcentration potential) of such substances.In experiments where radio-labelled substances have been used, high radio-labelconcentrations are often found in the gall bladder of fish. This is interpreted to be caused bybiotransformation in the liver and subsequently by excretion of metabolites in the gall bladder(Comotto et al., 1979; Wakabayashi et al., 1987; Goodrich et al., 1991; Toshima et al., 1992).The BCF from radio-labelled studies should, preferentially, be based on the parent compound.If these are unavailable, for classification purposes, the BCF based on total radio-labelledresidues can be used. If the BCF, in terms of radio-labelled residues, is ≥ 1000, theidentification and quantification of degradation products documented to be ≥ 10 % of totalresidues in fish tissues at steady state, are strongly recommended.When fish do not eat, the content of the gall bladder is not emptied into the gut, and highconcentrations of metabolites may build up in the gall bladder. The feeding regime may thus

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012have a pronounced effect on the measured BCF. In the literature many studies are foundwhere radio-labelled compounds are used, and where the fish are not fed. In these studies thebioconcentration may in most cases have been overestimated.III.2.2 Octanol-water-partitioning coefficient (K ow )For organic substances experimentally derived high-quality K ow values are preferred overother determinations of K ow . When no experimental data of high quality are available,validated Quantitative Structure Activity Relationships (QSARs) for log K ow may be used inthe classification process. Such validated QSARs may be used without modification to theagreed criteria if they are restricted to chemicals for which their applicability domain is wellcharacterised. For substances like strong acids and bases, substances which react with theeluent, or surface-active substances, a QSAR estimated value of K ow or an estimate based onindividual n-octanol and water solubilities should be provided instead of an analyticaldetermination of K ow . Measurements should be taken on ionisable substances in their nonionisedform (free acid or free base) only by using an appropriate buffer with pH below pKfor free acid or above the pK for free base. If multiple log K ow data are available for the samesubstance, the reasons for any differences should be assessed before selecting a value.Generally, the highest valid value should take precedence. Further details are provided inIR/CSA (R.7A) Chapter 7.1. Guidance on pH correction for ionisable substances is given inchapter 7.1.20.III.2.2.1Experimental determination of K owFor experimental determination of K ow values, several different methods are described instandard guidelines. Chapter 7.1.8.3 in IR/CSA (R.7A) gives guidance on directmeasurement methods (Shake Flask Method, Generator Column Method, and Slow StirringMethod), and on one indirect measurement method (Reverse Phase HPLC Method).III.2.2.2 Use of QSARs for determination of log K owWhen an estimated K ow value is found, the estimation method has to be taken into account.Numerous QSARs have been and continue to be developed for the estimation of K ow . Theperformances of top six programs, as evaluated in 2007, are given in Table III.2.2.2 below. Itis recommended that at least one of the below software programs be used for the prediction oflog K ow . If possible, the average of several predictions should be taken. More guidance isprovided is Chapter 7.1.8.3 in IR/CSA (R.7A).507

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Table II.2.2.2 Examples of software programs for the estimation of log K ow (from IR/CSA(R.7A), Chapter 7.1.8.3)Software Website Availability BatchOperation%Predictedwithin0.5 LogunitStandardErrorADMET www.simulationsplus.com Purchase Yes 94.2 0.27ACDLabs www.acdlabs.com Purchase Yes 93.5 0.27ChemSilico www.logp.com Free on line No 93.5 0.30KOWWINwww.epa.gov/oppt/exposure/pubs/episuitedl.htmFree todownloadYes 89.1 0.34SPARC ibmlc2.chem.uga.edu/sparc Free on line No 88.5 0.33ClogP www.daylight.com Purchase Yes 88.4 0.29III.3Chemical classes that need special attention with respect to BCF and K owvaluesThere are certain physico-chemical properties of substances, which can make thedetermination of BCF or its measurement difficult. These may be substances, which do notbioconcentrate in a manner consistent with their other physico-chemical properties, e.g. sterichindrance or substances which make the use of descriptors inappropriate, e.g. surface activity,which makes both the measurement and use of log K ow inappropriate.III.3.1Substances difficult to testThe methods presented above are generally designed for non-ionised organic substances.They are therefore of limited usefulness for a large number of other substances, collectivelytermed difficult substances, which include complex mixtures and chemicals that are chargedat environmental pH (such as inorganic compounds). Substances difficult to test may bepoorly soluble substances, complex mixtures, high molecular weight substances, surfaceactive substances, inorganic substances, ionisable substances, or organic substances that donot partition to lipid. Some guidance is given in this Chapter. More detailed guidance isprovided in IR/CSA (R.7C), mainly in Chapter 7.10.7.In order to bioconcentrate in aquatic organisms, an organic substance needs to be present inthe water, available for transfer across the fish gills and soluble in lipids. Factors that mayalter this availability will thus change the actual bioconcentration of a substance, whencompared with the prediction. For example, readily biodegradable substances may only bepresent in the aquatic compartment for short periods of time. Similarly, volatility, andhydrolysis will reduce the concentration and the time during which a substance is availablefor bioconcentration. A further important parameter, which may reduce the actual exposureconcentration of a substance, is adsorption, either to particulate matter or to surfaces ingeneral. There are a number of substances, which have shown to be rapidly transformed in theorganism, thus leading to a lower BCF value than expected. Substances that form micelles oraggregates may bioconcentrate to a lower extent than would be predicted from simple

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012physico-chemical properties. This is also the case for hydrophobic substances that arecontained in micelles formed as a consequence of the use of dispersants. Therefore, the use ofdispersants in bioaccumulation tests is discouraged. Further guidance is given in IR/CSA(R.7C) Chapter 7.10.3.4 on how to consider the factors that affect the bioaccumulationpotential of many substances and that are important especially in the absence of a fully validBCF test result.In general, for substances difficult to test, measured BCF and K ow values – based on theparent substance – are a prerequisite for the determination of the bioconcentration potential.Furthermore, proper documentation of the test concentration is a prerequisite for thevalidation of the given BCF value.III.3.2Poorly soluble and complex substancesSpecial attention should be paid to poorly soluble substances. Frequently the solubility ofthese substances is recorded as less than the detection limit, which creates problems ininterpreting the bioconcentration potential. Where the test data indicate that the concentrationsin the study are below the limit of detection, then the test is invalid and cannot be used. Forsuch substances the bioconcentration potential should be based on experimental determinationof log K ow or QSAR estimations of log K ow (see Section III. 2.2). Complex substances containa range of individual substances which can have a great variation in their physico-chemicaland toxicological properties. It is generally not recommended to estimate an average orweighted BCF value. It is preferable to identify one or more representative constituents forfurther consideration. Further guidance is given in Chapter 7.10.7.2 in IR/CSA (R.7C) 2008.III.3.3High molecular weight substancesA number of regulatory systems use molecular weight as an indicator for reduced or minimalbioconcentration. It is, however, concluded in IR/CSA (R.7C) 2008, Chapter 7.10.3.4 thatmolecular mass and size should not be used in isolation as confirmatory evidence of lack ofbioaccumulation (ECETOC 2005). However, supported by other data and by employingexpert judgement, it may be concluded by a weight of evidence argument that such substancesare unlikely to have a high bioconcentration factor (regardless of the log K ow value). Moredetails can be found in PBT assessment guidance (IR/CSA (R.11) 2008).III.3.4Surface-active substances (surfactants)Surfactants consist of an apolar, lipophilic part (most often an alkyl chain) (the hydrophobictail) and a polar part (the hydrophilic headgroup). According to the charge of the headgroup,surfactants are subdivided into classes of anionic, cationic, non-ionic, or amphotericsurfactants. Due to the variety of different headgroups, surfactants are a structurally diverseclass of compounds, which is defined by surface activity rather than by chemical structure.The bioaccumulation potential of surfactants should thus be considered in relation to thedifferent subclasses (anionic, cationic, non-ionic, or amphoteric) instead of to the group as awhole. Surface-active substances may form emulsions, in which the bioavailability is difficultto ascertain. Micelle formation can result in a change of the bioavailable fraction even whenthe solutions are apparently formed, thus giving problems in interpretation of thebioaccumulation potential. See Chapter 7.10.7.4 in IR/CSA (R.7C) 2008 for further guidance.Measured (experimentally derived) BCF values on surfactants show that BCF tends toincrease with increasing alkyl chain length and be dependent of the site of attachment of thehead group, other structural features and whether the alkyl part is subject tobiotransformation.509

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012III.3.4.1 Octanol-water-partition coefficient (K ow )The octanol-water partition coefficient for surfactants cannot be determined using theshakeflask or slow stirring method because of the formation of emulsions. In addition, thesurfactant molecules will exist in the water phase almost exclusively as ions, whereas theywill have to pair with a counter-ion in order to be dissolved in octanol. Therefore,experimental determination of K ow does not characterise the partition of ionic surfactants(Tolls, 1998). On the other hand, it has been shown that the bioconcentration of anionic andnon-ionic surfactants increases with increasing lipophilicity (Tolls, 1998). Tolls (1998)showed that for some surfactants, an estimated log K ow value using LOGKOW couldrepresent the bioaccumulation potential; however, for other surfactants some ‘correction’ tothe estimated log K ow value using the method of Roberts (1989) was required. These resultsillustrate that the quality of the relationship between log K ow estimates and bioconcentrationdepends on the class and specific type of surfactants involved. Therefore, the classification ofthe bioconcentration potential based on log K ow values should be used with caution. Furtherguidance is provided in Chapter 7.10.7.4 in IR/CSA (R.7C) 2008.III.4Conflicting data and lack of dataIII.4.1Conflicting BCF dataWhen multiple BCF data are available for the same substance, the possibility of conflictingresults may arise. In general, conflicting results for a substance, which has been tested severaltimes with an appropriate bioconcentration test, should be interpreted by a “weight ofevidence approach”. This implies that if experimentally determined BCF data, both ≥ and

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012III.5Decision schemeBased on the above discussions and conclusions, a decision scheme has been elaboratedwhich may facilitate decisions as to whether or not a substance has the potential forbioconcentration in aquatic species.Experimentally derived BCF values of high quality are ultimately preferred for classificationpurposes. BCF results from poor or questionable quality studies should not be used forclassification purposes. If no BCF is available for fish species, high quality data on the BCFfor some invertebrates (e.g. blue mussel, oyster and/or scallop) may be used as a worst casesurrogate.For non-ionised organic substances, experimentally derived high quality K ow values, or valueswhich are evaluated in reviews and assigned as the “recommended values”, are preferred. Ifno experimentally data of high quality are available validated Quantitative Structure ActivityRelationships (QSARs) for log K ow may be used in the classification process. Such validatedQSARs may be used without modification in relation to the classification criteria, if restrictedto chemicals for which their applicability is well characterised. For difficult substances likestrong acids and bases, metal complexes, and surface-active substances a QSAR estimatedvalue of K ow or an estimate based on individual n-octanol and water solubilities should beprovided instead of an analytical determination of K ow .If data are available but not validated, expert judgement should be used.Whether or not a substance has a potential for bioconcentration in aquatic organisms couldthus be decided in accordance with the following scheme:Valid/high quality experimentally determined BCF value → YES:→BCF ≥ 500: The substance meets the criterion→BCF < 500: The substance does not meet the criterionValid/high quality experimentally determined BCF value → NO:→Valid/high quality experimentally determined log K ow value → YES:→log K ow ≥ 4: The substance meets the criterion→log K ow < 4: The substance does not meet the criterionValid/high quality experimentally determined BCF value → NO:Valid/high quality experimentally determined log K ow value → NO:Use of validated QSAR for estimating a log K ow value → YES:→log K ow ≥ 4: The substance meets the criterion→log K ow < 4: The substance does not meet the criterionIII.6ReferencesComotto, R,M., Kimerle, R.A., Swisher, R.D. 1979. Bioconcentration and metabolism oflinear alkylbenzenesulfonate by Daphnids and Fathead minnows. L.L.Marking, R.A. Kimerle,Eds. Aquatic Toxicology (ASTM, 1979), vol. ASTM STP 667.EC 2008. C.13 Bioconcentration: flow-through fish test in COUNCIL REGULATION (EC)No 440/2008 of 30 May 2008 laying down test methods pursuant to Regulation (EC) No1907/2006 of the European Parliament and of the Council on the Registration, Evaluation,511

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Authorisation and Restriction of Chemicals (REACH). Official Journal of European UnionL142: 571-589.ECETOC 2005. Technical Report No. 97. Alternative testing approaches in environmentalsafety assessment. ISSN-0773-8072-97.Goodrich, M.S., Melancon, M.J., Davis, R.A.1991. The toxicity, bioaccumulation,metabolism and elimination of dioctyl sodium sulfosuccinate DSS in rainbow trout(Oncorhynchus mykiss). Water Res. 25: 119-124.OECD 1996. Bioaccumulation: Flow-through Fish Test. Organisation for EconomicCooperation and Development (OECD), OECD Guideline for the Testing of Chemicals No.305, Paris, France.Roberts, D.W. 1989. Aquatic toxicity of linear alkyl benzene sulphonates (LAS) – a QSARanalysis. Communicaciones Presentadas a Las Jornadas Del Comite Esponol de laDetergencia, 20 (1989) 35-43. Also in J.E. Turner, M.W. England, T.W. Schultz and N.J.Kwaak (eds.) QSAR 88. Proc. Third International Workshop on Qualitative Structure-Activity Relationships in Environmental Toxicology, 22-26 May 1988, Knoxville, Tennessee,pp. 91-98. Available from National Technical Information Service, US Dept. of Commerce,Springfield, VA.Tolls, J. 1998. Bioconcentration of surfactants. Ph.D. Thesis, Utrecht University, Utrecht, TheNetherlands.Toshima, S., Moriya, T., Yoshimura, K. 1992. Effects of polyoxyethylene (20) sorbitanmonooleate on the acute toxicity of linear alkylbenzenesulfonate (C12-LAS) to fish,Ecotoxicol. Environ. Safety 24: 26-36.Wakabayashi, M., Kikuchi, M., Sato, A., Yoshida, T. 1987. Bioconcentration of alcoholethoxalates in carp (Cyprinus carpio), Ecotoxicol. Environ. Safety 13, 148-163.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012IV ANNEX IV: METALS AND INORGANIC METAL COMPOUNDSIV.1 IntroductionThe harmonised system for classifying chemical substances is a hazard-based system, and thebasis of the identification of hazard is the aquatic toxicity of the substances, and informationon the degradation and bioaccumulation behaviour (OECD 2001). Since this document dealsonly with the hazards associated with a given substance when the substance is dissolved in thewater column, exposure from this source is limited by the solubility of the substance in waterand bioavailability of the substance to organisms in the aquatic environment. Thus, the hazardclassification schemes for metals and metal compounds are limited to the acute and long-termhazards posed by metals and metal compounds when they are available (i.e. exist as dissolvedmetal ions, for example, as M+ when present as M-NO 3 ), and do not take into accountexposures to metals and metal compounds that are not dissolved in the water column but maystill be bioavailable, such as metals in foods. This section does not take into account the nonmetallicion (e.g. CN - ) of metal compounds which may be toxic. For such metal compoundsthe hazards of the non-metallic ions must also be considered.Also organometal compounds may be of concern given they may pose bioaccumulation orpersistence hazards. Organometals do not dissociate or dissolve in water as the metal ion, asmetals and inorganic metal compounds do. Organometals (e.g. methyl mercury or tributyltin)that do not release metal ions are thereby excluded from the guidance of this section andshould be classified according to the general guidance provided in section 4. Metalcompounds that contain an organic component but that dissociate easily in water or dissolveas the metal ion should be treated in the same way as metal compounds and classifiedaccording to this annex (e.g. zinc acetate).The level of the metal ion which may be present in solution following the addition of themetal and/or its compounds, will largely be determined by two processes: the extent to whichit can be dissolved, i.e. its water solubility, and the extent to which it can react with the mediato transform to water soluble forms. The rate and extent at which this latter process, known as“transformation” for the purposes of this guidance, takes place can vary extensively betweendifferent compounds and the metal itself, and is an important factor in determining theappropriate hazard class. Where data on transformation are available, they should be takeninto account in determining the classification. The Protocol for determining this rate isavailable as Annex 10 to the UN GHS.Generally speaking, the rate at which a substance dissolves is not considered relevant to thedetermination of its intrinsic toxicity. However, for metals and many poorly soluble inorganicmetal compounds, the difficulties in achieving dissolution through normal solubilisationtechniques are so severe that the two processes of solubilisation and transformation becomeindistinguishable. Thus, where the compound is sufficiently poorly soluble that the levelsdissolved following normal attempts at solubilisation do not exceed the available L(E)C 50 , itis the rate and extent of transformation, which must be considered. The transformation will beaffected by a number of factors, not least of which will be the properties of the media withrespect to pH, water hardness, alkalinity, temperature etc. In addition to these properties, otherfactors such as the size and, in particular, the specific surface area of the particles which havebeen tested, the length of time over which exposure to the media takes place and, of coursethe mass or surface area loading of the substance in the media will all play a part indetermining the level of dissolved metal ions in the water. Transformation data can generally,therefore, only be considered as reliable for the purposes of classification if conductedaccording to the standard protocol in Annex 10 to UN GHS. This protocol aims at513

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012standardising the principal variables such that the level of dissolved ion can be directly relatedto the loading of the substance added. It is this loading level which yields the level of metalion equivalent to the available L(E)C 50 or NOEC/EC 10 that can then be used to determine theacute or long-term hazard category appropriate for classification. The testing methodology isdetailed in Annex 10 to the UN GHS. The strategy to be adopted in using the data from thetesting protocol, and the data requirements needed to make that strategy work, are describedin Annex IV.2, IV.3 and in more detail in Annex IV.5 of this document.In considering the classification of metals and metal compounds, both readily and poorlysoluble, recognition has to be paid to a number of factors. As defined in Annex II, section II.1,the term “degradation” refers to the decomposition of organic molecules. For inorganiccompounds and metals, clearly the concept of degradability, as it has been considered andused for organic substances, has limited or no meaning. Rather, the substance may betransformed by normal environmental processes to either increase or decrease thebioavailability of the toxic species. Equally, the log K ow cannot be considered as a measure ofthe potential to accumulate. Nevertheless, the concept that a substance, or a toxicmetabolite/reaction product may not be rapidly lost from the environment and/or maybioaccumulate, are as applicable to metals and metal compounds as they are to organicsubstances.Speciation of the soluble form can be affected by pH, water hardness and other variables, andmay yield particular forms of the metal ion which are more or less toxic. In addition, metalions could be made non-available from the water column by a number of processes (e.g.mineralisation and partitioning). Sometimes these processes can be sufficiently rapid to beanalogous to degradation in assessing chronic (long-term) aquatic hazard. However,partitioning of the metal ion from the water column to other environmental media does notnecessarily mean that it is no longer bioavailable, nor does it necessarily mean that the metalhas been made permanently unavailable.Information pertaining to the extent of the partitioning of a metal ion from the water column,or the extent to which a metal has been or can be converted to a form that is less toxic or nontoxicis frequently not available over a sufficiently wide range of environmentally relevantconditions, and thus, a number of assumptions will need to be made as an aid in classification.These assumptions may be modified if available data show otherwise. In the first instance itshould be assumed that the metal ions, once in the water, are “not rapidly partitioned” fromthe water column. Underlying this is the assumption that, although speciation can occur, thespecies will remain available under environmentally relevant conditions. This may not alwaysbe the case, as described above, and any evidence available that would suggest changes to thebioavailability over the course of 28 days, should be carefully examined.The bioaccumulation of metals and inorganic metal compounds is a complex process andbioaccumulation data should be used with care. The application of bioaccumulation criteriawill need to be considered on a case-by-case basis taking due account of all the available data.A further assumption that can be made, which represents a cautious approach, is that, in theabsence of any solubility data for a particular metal compound, either measured or calculated,the metal compound will be assumed to be sufficiently soluble to cause toxicity at the level ofthe ecotoxicity reference value (ERV), being the acute ERV (expressed as L(E)C 50 ), and/orthe chronic ERV (expressed as the NOEC/ECx or an HC5 for extensive data sets) and thusmay be classified in the same way as other soluble salts of the metal. Again, this is clearly notalways the case, and it may be wise to generate appropriate solubility data. Absence ofsolubility data on the metallic form for a metal for which the soluble salts are classified forthe environment, will therefore lead to a default classification due to potential hazardconcerns.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012This Annex IV deals with metals and inorganic metal compounds. Within the context of thisguidance document, metals and metal compounds are characterised as follows:(a) metals (M 0 ) in their elemental state are not soluble in water but may transform to yieldthe available form (e.g. Fe 0 will not dissolve as such but the Fe 0 molecules present atthe surface of a massive/powder will be first transformed into Fe 2+ or Fe 3+ compoundsprior to their solubilisation). This means that a metal in the elemental state may reactwith water or a dilute aqueous electrolyte to form soluble cationic or anionic products,and in the process the metal will oxidise, or transform, from the neutral or zerooxidation state to a higher one;(b) in a simple metal compound, such as an oxide or sulphide, the metal already exists inthe oxidised state, so that further metal oxidation is unlikely to occur when thecompound is introduced into an aqueous medium.Organo-metals are outside the scope of this section.While oxidisation may not change, interaction with the media may yield more soluble forms.A sparingly soluble metal compound can be considered as one for which a solubility productcan be calculated, and which will yield a small amount of the available form by dissolution.However, it should be recognised that the final solution concentration may be influenced by anumber of factors, including the solubility product of some metal compounds precipitatedduring the transformation/dissolution test, e.g. aluminium hydroxide.IV.2 Application of aquatic toxicity data and solubility data for classificationIV.2.1 Interpretation of aquatic toxicity dataEcotoxicity data of soluble inorganic compounds are used and combined to define the toxicityof the metal ion under consideration. The ecotoxicity of soluble inorganic metal compounds isdependent on the physico-chemistry of the medium, irrespective of the original metal speciesreleased in the environment. Reading across metal compounds can therefore be conducted bycomparing the soluble metal ion concentration (µg Me/L) causing the ecotoxicity effect andtranslating this towards the compound under investigation. A molecular weight correction ofthe ecotoxicity reference value may be required to classify soluble metal compounds (MWsoluble substance/MW metal ion 70 ). Poorly soluble metal compounds and metals do notrequire Molecular weight correction given the amount used for Transformation Dissolutionalready recognises this into the loading calculation. The comparison is therefore directly doneby comparing the soluble fraction measured after Transformation Dissolution with theecotoxicity reference values of the soluble metal ion (based on the UN GHS, 2009).When evaluating ecotoxicity data, the general guidance on the weight of evidence (see section4.1.3.6 of this document) is also applicable to metals.The term adequacy covers here both the reliability (inherent quality of a test relating to testmethodology and the way that the performance and results of a test are described) and therelevance (extent to which a test is appropriate to be used for the derivation of an ecotoxicityreference value) of the available ecotoxicity data:Under the reliability criteria, metal specific considerations include the description of someabiotic parameters in the test conditions for enabling the consideration of the bioavailablemetal concentration and free metal ion concentration:70 Note that this calculation needs to be adjusted to reflect the stoichiometry of the compound, for example forZn 3 (PO 4 ) 2 the MW metal would be multiplied by three.515

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012‐ Description of the physical test conditions: further to the general parameters (O 2 , T°,pH, …) abiotic parameters such as dissolved organic carbon (DOC), hardness,alkalinity of the water that govern the speciation and hence the metal bioavailability isrequired. A proper description of culture conditions related to the level of essentialmetals is required to avoid artefacts due to acclimatisation/adaptation (see also below)‐ Description of test materials and methods: to calculate the free metal ionconcentration with speciation models the concentrations of dissolved major ions andcations like Al, Fe, Mg, Ca… are required‐ Concentration-effect relationship; hormesis: sometimes an increased performance ingrowth or reproduction is seen at low metal doses that exceed the control values,referred to as hormesis. Such effects can be important especially for major tracenutrients such as Fe, Zn and Cu but can also occur with a wide variety of non-essentialsubstances. In such cases, positive effects should not be considered in the derivation ofacute ERV’s and especially chronic ERV’s, likely other models than the conventionallog-logistic dose-response model should be used to fit the dose-response curve andconsideration should be given to the adequacy of the control diet/exposure. Due to theessential nutritional needs, caution is needed with regards to extrapolation of the doseresponsecurve (e.g. to derive an acute ERV) below the lowest tested concentration.Under the relevancy criteria, certain considerations need to be made, related to the relevancyof the test substance and to acclimatisation/adaptation:‐ Relevance of the test substance: soluble metal salts should be used for the purpose ofclassification of inorganic metals/metal compounds. The ecotoxicity adapted fromorganic metal compounds exposure should not be used.‐ Acclimatisation/adaptation: For essential metals, the culture medium should contain aminimal concentration not causing deficiency for the test species used. This isespecially relevant for organisms used for long-term toxicity tests where the marginbetween essentiality and toxicity may become small. As an example, for algae,depletion of the strong complexing agent EDTA from the medium may result in irondeficiency.Aquatic toxicity studies carried out according to a recognised protocol should normally beacceptable as valid for the purposes of classification. Annex I should also be consulted forgeneric issues that are common to assessing any aquatic toxicity data point for the purposes ofclassification.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012IV.2.1.1 Metal complexation and speciationThe toxicity of a particular metal in solution, appears to depend primarily on (but is notstrictly limited to) the level of dissolved free metal ions and the physico-chemistry of theenvironment. Abiotic factors including alkalinity, ionic strength and pH can influence thetoxicity of metals in two ways: (i) by influencing the chemical speciation of the metal in water(and hence affecting the availability) and (ii) by influencing the uptake and binding ofavailable metal by biological tissues. For the classification of metals,Transformation/Dissolution is carried out over a pH range. Ideally both T/D and ecotoxicitydata are compared at a similar pH since both parameters will vary with pH. However, themajority of ecotoxicity tests are performed at the higher pH range (i.e. > pH 7.5) andecotoxicity data obtained at lower pH are often scarce. Bioavailability and speciation models(e.g. respectively Biotic Ligand Models and WHAM (Tipping, 1994), as discussed below)may allow to normalise ecotoxicity data obtained at a given pH to other pH values, relevant tothe T/D data. The applicability of the bioavailability models to the biological species forwhich data are available must be evaluated. Guidance on the Bioavailability correction formetals can be found in IR/CSA Annex R.7.13.2).Where chemical speciation is important, it may be possible to model the concentrations of thedifferent chemical forms of the metal, including those that are likely to cause toxicity.Analysis methods for quantifying exposure concentrations, which are capable ofdistinguishing between the complexed and uncomplexed fractions of a test substance, may notalways be available or economic.Complexation of metals to organic and inorganic ligands in test media and naturalenvironments can be estimated from metal speciation models. Speciation models for metals,including pH, hardness, DOC, and inorganic substances such as MINTEQ (Brown andAllison, 1987), WHAM (Tipping, 1994) and CHESS (Santore and Driscoll, 1995) can be usedto calculate the uncomplexed and complexed fractions of the metal ions.Alternatively, and when available for the metal, the Biotic Ligand Model (BLM), allows, forthe calculation of the acute and/or chronic ERV’s of the metal ion, for different pH values,through integration of metal speciation and its interaction with the organism. The BLM modelhas at present been validated for a number of metals, organisms, and end-points (Santore andDi Toro, 1999). The models and formula used for the characterisation of metal complexationin the media should always be clearly reported, allowing for their translation back to naturalenvironments (OECD, 2000). In case a metal-specific BLM is available covering anappropriate pH range, a normalised comparison of aquatic toxicity data can be made using theentire effects database for different reference pH values.IV.2.2 Interpretation of solubility dataWhen considering the available data on solubility, their validity and applicability to theidentification of the hazard of metal compounds should be assessed. In particular, the pH andthe medium in which the data were generated should be known.IV.2.2.1 Assessment of existing dataExisting data will be in one of the three forms: for soluble, insoluble metal compounds andthe metallic form. For some well-studied metals, there will be solubility products and/orsolubility data for the various inorganic metal compounds. It is also possible that the pHrelationship of the solubility will be known. However, for many metals or metal compounds,517

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012it is probable that the available information will be descriptive only, e.g. poorly soluble orresulting from the water solubility test form the OECD 105 physico-chemical waterdissolution test. Unfortunately there appears to be very little (consistent) guidance about thesolubility ranges for such descriptive terms. Where these are the only information available itis most probable that solubility data will need to be generated using theTransformation/Dissolution Protocol (Annex 10 to the UN GHS).IV.2.2.2 Screening T/D test for assessing solubility of metal compoundsIn the absence of solubility data, a simple “Screening Test” for assessing solubility, based onthe high rate of loading (100 mg/l) for 24 h and rigid stirring conditions, should be used formetal compounds as described in the Transformation/Dissolution Protocol (Annex 10 to theUN GHS). The function of the screening test is to identify those metal compounds whichundergo either dissolution or rapid transformation such that they are indistinguishable fromsoluble forms and hence may be classified based on the dissolved ion concentration and thosewho dissolves slowly and can be assessed in the same way as the metallic form. Where dataare available from the screening test detailed in the Transformation/Dissolution Protocol, themaximum solubility obtained over the tested pH range should be used. Where data are notavailable over the full pH range, a check should be made that this maximum solubility hasbeen achieved by reference to suitable thermodynamic speciation models or other suitablemethods (see section IV.2.1.1 of this document). It should be noted that this test is onlyintended to be used for inorganic metal compounds. Metals should immediately be assessed atthe level of the full T/D test.IV.2.2.3 Full T/D test for assessing solubility of metals and metal compoundsThe Full Transformation Dissolution test should be carried out at the pH 71 that maximises theconcentration of dissolved metal ions in solution and that expresses the highest toxicity.Based on the data from the Full Test, it is possible to generate a concentration of the metalions in solution after 7 days (short-term test) for each of the three loadings (i.e. 1 mg/l as“low”, 10 mg/l as “medium” and 100 mg/l as “high loading”) used in the test. If the purposeof the test is to assess the long-term hazard of the substance, then the loadings 72 should be0.01 mg/l, 0.1 mg/l or 1 mg/l depending on the transformation rate and the duration of the testbeing extended to 28 days (long-term test).IV.2.3 Comparison of aquatic toxicity data and solubility data71 The UN GHS transformation/dissolution protocol specifies a pH range of 6-8.5 for the 7days test and 5.5 to 8 .5 for the 28days test. Considering the difficulty in carrying out transformation/dissolution tests at pH 5.5, the OECD only validated thetest in the pH range of 6-to 8.5.72 The standard protocol in Annex 10 to UN GHS presently only foresees a long-term loading rate of 1 mg/l and lowerloading rates may not even be practically feasible for each case. While TDp testing at lower loading rates is in principle thebest way forward it is technically often not feasible for the lower chronic loading rates. Extensive experience with the T/Dprotocol demonstrated that reliable predictions can be made for other loading rates. In order to make maximal use of existingTransformation Dissolution data, the 28 days results for the lower chronic loading rates (0,1 and 0,01 mg/l) can therefore bederived by extrapolation from TDp evidence from other loading rates. Such read across should be justified on a case by casebasis and supported by reliable information on the T/D at different loading rates, e.g. over 7 and/or 28 days. It should benoted that the relationship between loading rate and dissolved metal concentration may well not be linear. Thereforeextrapolation of T/D data to lower loadings should preferably be made by using the equations of section A10.6.1 of the UN-Annex 10 transformation dissolution protocol or alternatively by extrapolating in a precautionary way.The UN announced to change/update Annex 10 in the near future to bring it better in line with the chronic classificationstrategy an aim that is already anticipated in this guidance note for the CLP.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012A decision on whether or not the substance is classified will be made by comparing aquatictoxicity data and solubility data. Depending on the available data two approaches can befollowed.1) When only a limited dataset is available existing data should be taken togetherirrespective of whether the toxicity and dissolution data are at the same pH and thelowest data point should give the basis for classification (this should be used as thedefault approach). This default approach may lead to the lowest toxicity data pointcompared with the highest Transformation Dissolution result each derived at differentpH levels used for the purpose of classification.2) When a more extensive toxicity/dissolution dataset is available, a split of the acute andchronic ecotoxicity reference values can be performed according to their pH usedduring T/D test. The worst case classification entry across pHs should be used basedon comparing TDp data with relevant ecotox data across the pH range. Meaning thattoxicity data and transformation data are in this case always compared at the same pH.This split of the effects data into pH classes would apply in an equal way to the acute and thelong-term effects data sets.IV.3 Assessment of environmental transformationEnvironmental transformation of one species of a metal to another species of the same metaldoes not constitute “degradation” as applied to organic compounds and may increase ordecrease the availability and bioavailability of the toxic species. In addition naturallyoccurring geochemical processes can partition metal ions from the water column while alsoother processes may remove metal ions from the water column (e.g. by precipitation andspeciation). Data on water column residence time, the processes involved at the water –sediment interface (i.e. deposition and re-mobilisation) are fairly extensive for some metals.Using the principles and assumptions discussed above in section IV.1 of this document, itmay therefore be possible to incorporate this approach into the classification.Such assessments are difficult to give guidance for and will normally be addressed on a caseby-caseapproach. However, the following may be taken into account:(a) Changes in speciation if they are to non-available forms, however, the potential forthe reverse change to occur must also be considered;(b) Changes to a metal compound which is considerably less soluble than that of themetal compound being considered.Some caution is recommended; see section IV.1 of this document, the 5 th and 6 th paragraph.Comment by ECHA: Please note that in the light of a lack of scientific consensus andcontinuing discussions on the interpretation of rapid removal from the water column in thecontext of classification, it has been decided to remove certain parts from the Annex IV forthe time being until agreement on the validity of use of the concept of rapid removal forclassification purposes has been reached.IV.4 BioaccumulationWhile log Kow is a good predictor of BCF for certain types of organic compounds e.g.nonpolar organic substances, it is irrelevant for inorganic substances such as inorganic metalcompounds because metals, in contrast to organic substances, are not lipophilic and are notpassively transported through cellular membranes. Uptake of metal ions occurs through activeprocesses.519

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The mechanisms for uptake and depuration rates of metals are very complex and variable andthere is at present no general model to describe this. Instead the bioaccumulation of metalsaccording to the classification criteria should be evaluated on a case-by-case basis usingexpert judgement.While BCFs are indicative of the potential for bioaccumulation there may be a number ofcomplications in interpreting measured BCF values for metals and inorganic metalcompounds. For most metals and inorganic metal compounds the relationship between waterconcentration and BCF in aquatic organisms is inverse, and bioconcentration data shouldtherefore be used with care. This is particularly relevant for metals that are biologicallyessential. Metals that are biologically essential are actively regulated in organisms in whichthe metal is essential (homeostasis). Removal and sequestration processes that minimisetoxicity are complemented by an ability to up-regulate concentrations for essentiality. Sincenutritional requirement of the organisms can be higher than the environmental concentration,this active regulation can result in high BCFs and an inverse relationship between BCFs andthe concentration of the metal in water. When environmental concentrations are low, highBCFs may be expected as a natural consequence of metal uptake to meet nutritionalrequirements and can in these instances be viewed as a normal phenomenon. Also, while ametal may be essential in a particular organism, it may not be essential in other organisms.Therefore, where the metal is not essential or when the bioconcentration of an essential metalis above nutritional levels, special consideration should be given to the potential forbioconcentration and environmental concern.Non- essential metals are also actively regulated to some extent and therefore also for nonessentialmetals, an inverse relationship between the metal concentration and the externalconcentration may be observed (McGeer et al., 2003).Consequently for both essential and non-essential elements, measured BCFs decline asexternal concentration increases. When external concentrations are so high that they exceed athreshold level, or overwhelm the regulatory mechanism, this can cause harm to the organismBCF and BAF may be used to estimate metal accumulation by:a) Considering information on essentiality and homeostasis of metals/ metal compounds. As aresult, of such regulation, the “bioaccumulative” criterion is not applicable to these metals.b). Assessing bioconcentration factors for non-essential metals, should preferably be donefrom BCF studies using environmentally relevant concentrations in the test media.IV.5 Classification strategies for metals and metal compoundsIV.5.1 IntroductionNotice! Acute and long-term hazards are assessed individually.For determination of long-term hazards preference should be given in applying the approachbased on chronic toxicity data. Such evidence is often frequently available for the bioavailableforms of metals.The schemes for the determination of acute and long-term aquatic hazards of metals and metalcompounds are described below and summarised diagrammatically in the figures:IV.5.2.1 (acute hazard classification of metals),IV.5.2.2 (a and b) (long-term hazard of metals);IV.5.3.1 (acute hazard classification of metal compounds);

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012IV.5.3.2 (a and b) (long-term hazard of metal compounds).There are several stages in these schemes where data are used for decision purposes. It is notthe intention of the classification schemes to generate new ecotoxicity data. In the absence ofvalid data, it will be necessary to use all available data and expert judgement.In the following sections, the reference to the acute and chronic ERV’s refer to the datapoint(s) that will be used to select the hazard category(ies) for the metal or metal compound.When considering acute and chronic ERV’s data for metal compounds, it is important toensure that the data point to be used as the justification for the classification is expressed inthe weight of the molecule of the metal compound to be classified. This is known ascorrecting for molecular weight. Thus while most metal data is expressed in, for example,mg/l of the metal (ion), this value will need to be adjusted to the corresponding weight of themetal compound. Thus:Acute ERV compound = acute ERV of the metal compound = acute ERV of metal ion x(Molecular weight of metal compound /atomic weight of the metal).Chronic ERV compound = chronic ERV of the metal compound = chronic ERV of metal ion x(Molecular weight of metal compound /atomic weight of the metal).IV.5.2 Classification strategies for metalsNotice!Acute and long-term hazards are assessed individually.IV.5.2.1 Classification strategy for determining acute aquatic hazard for metalsThe scheme for the determination of acute aquatic hazard for metals are described in thissection and summarised diagrammatically in Figure IV.5.2.1.Where the acute ERV for the metal ions of concern is greater than 1 mg/l the metals need notbe considered further in the classification scheme for acute hazard.Where the acute ERV for the metal ions of concern is less than or equal to 1 mg/lconsideration must be given to the data available on the rate and extent to which these ionscan be generated from the metal. Such rate and extend data, to be valid and useable shouldhave been generated using the Transformation/Dissolution Protocol (Annex 10 to UN GHS)for a 7d period.Where 7d data from the Transformation/Dissolution protocol are available, then the resultsshould be used to classify, according to the following rule:Classify the metal as Category Acute 1 if the dissolved metal ion concentration after aperiod of 7 days (or earlier for a significant time period) at a loading rate of 1 mg/lexceeds that of the acute ERV, an M-factor must also be established as part of thisclassification (see IV 5.4).521

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure IV.5.2.1 Classification strategy for determining acute aquatic hazard for metals.7 days T/D full test data availableYESNONot possible to classify foracute aquatic hazard due toinsufficient data.Concentration at 1 mg/l loading rate ≥acute ERV of dissolved formNOYESClassify Acute 1 and addM-factor (see IV.5.4)Do not classify for acutehazardIV.5.2.2 Classification strategy for determining long-term aquatic hazard for metalsThe scheme for the determination of long-term aquatic hazard for metals are described in thissection and summarised diagrammatically in Figures IV.5.2.2 (a and b).Metals can be classified for long-term aquatic hazards:1) using chronic reference data when available; or2) using the surrogate approach in absence of appropriate chronic toxicity reference data.In case relevant chronic ecotoxicity data (chronic ERV) are available the approach comparingchronic ERV with 28 days transformation/dissolution reference should be applied asdescribed under IV.5.2.2.1 while otherwise the surrogate approach (see IV.5.2.2.2) should befollowed.IV.5.2.2.1 Approach based on available chronic toxicity reference dataWhere the chronic ERV for the metal ions of concern is greater than 1 mg/l, the metals neednot be considered further in the classification scheme.Where the chronic ERV for the metal ions of concern is less than or equal to 1 mg/lconsideration must be given to the data available on the rate and extent to which these ionscan be generated from the metal. Such rate and extend data, to be valid and useable shouldhave been generated using the Transformation/Dissolution Protocol (Annex 10 to UN GHS)for a 28 d period.Where such T/Dp data are unavailable the surrogate approach should be applied (see section5.2.2.2). Where 28d data from the Transformation/Dissolution protocol are available, then,the results should be used to aid classification according to the following rules:a) Classify the metal as Category Chronic 1 if the dissolved metal ion concentrationobtained at a loading rate of 0.1 mg/l is greater than or equal to the chronic ERV, anM-factor must also be established as part of this classification (see IV.5.4); orb) Classify the metal as Category Chronic 2 if the dissolved metal ion concentrationobtained at a loading rate of 1 mg/l is greater than or equal to the chronic ERV.If there is evidence of rapid environmental transformation:

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012c) Classify the metal as Category Chronic 1 if the dissolved metal ion concentrationobtained at a loading rate of 0.01 mg/l is greater than or equal to the chronic ERV, anM-factor must also be established as part of this classification (see IV 5.4); ord) Classify the metal as Category Chronic 2 if the dissolved metal ion concentrationobtained at a loading rate of 0.1 mg/l is greater than or equal to the chronic ERV; ore) Classify the metal as Category Chronic 3 if the dissolved metal ion concentrationobtained at a loading rate of 1 mg/l is greater than or equal to the chronic ERV.Do not classify for long-term hazard if the dissolved metal ion concentration obtained fromthe 28 day Transformation/Dissolution test at a loading rate of 1 mg/l is less than the chronicERV of the metal ion.IV.5.2.2.2 The surrogate approachWhere the acute ERV for the metal ions of concern is less than or equal to 100 mg/lconsideration must be given to the data available on the rate and extent to which these ionscan be generated from the metal. Such rate and extend data, to be valid and useable shouldhave been generated using the Transformation/Dissolution Protocol (Annex 10 to UN GHS)for a 7d period.Where such T/Dp data are unavailable, i.e. there is no clear data of sufficient validity to showthat the transformation to metal ions will not occur; the safety net classification (CategoryChronic 4) should be applied since the known classifiable toxicity of these soluble forms isconsidered to give rise to sufficient concern.Where T/Dp data are available classification should be according to the following rules:(a)(b)(c)(d)Classify the metal as Category Chronic 1 if the dissolved metal ion concentrationobtained from the 7 day transformation test at the low loading rate (1 mg/l) isgreater than or equal to the acute ERV, an M-factor must also be established aspart of this classification (see IV.5.4);Classify the metal as Category Chronic 2 if the dissolved metal ion concentrationobtained from the 7 day transformation test at the medium loading rate (10 mg/l) isgreater than or equal to the acute ERV;Classify the metal as Category Chronic 3 if the dissolved metal ion concentrationobtained from the 7 day transformation test at the high loading rate (100 mg/l) isgreater than or equal to the acute ERV.Classify the metal as Category Chronic 4 if the dissolved metal ion concentrationobtained from the 7 day transformation test at the high loading rate (100 mg/l) islower than the acute ERV.523

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure IV.5.2.2a Classification strategy for determining long-term aquatic hazard for metals.NOIs chronic ERV available?YESChronic ERV < 1 mg/lNODo not classify for longterm aquatic hazardYESGo to FigureIV.5.2.2bNO28 days T/D full test data availableYESIs there evidence of rapid environmentaltransformation?YESNOClassifyChronic 1and addM-factor(see IV5.4)YESConcentration at 0.01 mg/lloading rate ≥ chronic ERV ofdissolved formNOConcentration at 0.1 mg/lloading rate ≥ chronic ERVof dissolved formNOYESClassifyChronic 1and addM-factor(see IV5.4)ClassifyChronic 2YESConcentration at 0.1 mg/lloading rate ≥ chronic ERV ofdissolved formNOConcentration at 1 mg/lloading rate ≥ chronic ERVof dissolved formNOYESClassifyChronic 2ClassifyChronic 3YESConcentration at 1 mg/l loadingrate ≥ chronic ERV of dissolvedformNODo not classify for longterm aquatic hazard

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure IV.5.2.2b Classification strategy for determining long-term aquatic hazard for metalsin absence of appropriate chronic toxicity reference and/or T/Dp data.Is acute ERV available and ≤ 100 mg/l?YES7 days T/D full test data availableNONot possible to classify for long-termaquatic hazard due to insufficient dataClassify Chronic 4unless Acute 1 appliesYESNOYESNOIs there evidence of both rapidenvironmental transformation and nobioaccumulation?NOYESIs chronic ERV available and ≤ 1 mg/l?Concentration at 1 mg/lloading rate ≥ acute ERVof dissolved formNOYESClassifyChronic 1and addM-factor(IV5.4))Concentration at 10 mg/lloading rate ≥ acute ERVof dissolved formYESClassifyChronic 2NOConcentration at 100 mg/lloading rate ≥ acute ERVof dissolved formNOYESClassifyChronic 3Classify Chronic 4IV.5.3 Classification strategies for metal compoundsNotice! Acute and long-term hazards are assessed individually.A metal compound will be considered as readily soluble if:‐ the water solubility (measured through a 24-hour Dissolution Screening test orestimated e.g. from the solubility product) is greater or equal to the acute ERV of thedissolved metal ion concentration; or‐ If such data are unavailable, i.e. there are no clear data of sufficient validity to showthat the transformation to metal ions will not occur;Care should be exercised for metal compounds whose solubility is close to the acute toxicityreference value as the conditions under which solubility is measured could differ significantlyfrom those of the acute toxicity test. In these cases the results of the Dissolution ScreeningTest are preferred.525

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Metal compounds that have lower water solubility than the acute ERV through a 24-hourDissolution Screening test or estimated from the solubility product, are considered as poorlysoluble metal compound.IV.5.3.1 Classification strategies for determining acute aquatic hazard for metalcompoundsThe scheme for the determination of acute aquatic hazard for metal compounds are describedin this section and summarised diagrammatically in Figure IV.5.3.1.Where the acute ERV for the metal ions of concern corrected for the molecular weight of thecompound (further called as acute ERV compound ) is greater than 1 mg/l, the metal compoundsneed not to be considered further in the classification scheme for acute hazard.Where the acute ERV compound is less than or equal to 1 mg/l, consideration must be given tothe data available on the rate and extent to which these ions can be generated from the metalcompound. Such data, to be valid and useable should have been generated using the T/D(Annex 10 to UN GHS).Readily soluble metal compoundsClassify the metal compound as Category Acute 1 if the acute ERV compound ≤ 1 mg/l,M-factor must also be established as part of this classification (see IV.5.4).Poorly soluble metal compoundsWhere 7d data from the Transformation/Dissolution protocol are available, then the resultsshould be used to classify sparingly soluble metal compounds, according to the followingrule:Classify the metal compound as Category Acute 1 if the dissolved metal ionconcentration after a period of 7 days (or earlier for a significant time period) at aloading rate of 1 mg/l exceeds that of the acute ERV, an M-factor must also beestablished as part of this classification(see IV.5.4).an

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure IV.5.3.1 Classification strategy for determining acute aquatic hazard for metalcompounds.Is it readily soluble(solubility ≥ acute ERV)?YESAcute ERV compound ≤ 1 mg/lYESClassify Acute 1 andadd M-factor (see IV5.4)NONODo not classify for acuteaquatic hazard7 days T/D full test data availableYESNONot possible to classify foracute aquatic hazard due toinsufficient dataConcentration at 1 mg/l loading rate ≥acute ERV of dissolved formNOYESClassify Acute 1 andadd M-factor (see IV5.4)Do not classify for acuteaquatic hazardIV.5.3.2 Classification strategy for determining long-term aquatic hazard for metalcompoundsThe scheme for the determination of long-term aquatic hazard for metal compounds aredescribed in this section and summarised diagrammatically in Figures IV.5.3.2 (a and b).Metal compounds can be classified for long-term aquatic hazards:1) using chronic reference data when available; or2) using the surrogate approach in absence of appropriate chronic toxicity reference data.In case relevant chronic ecotoxicity data (chronic ERV) are available the approach comparingchronic ERV of the dissolved metal ion with release data of 28 daystransformation/dissolution , should be applied as described under IV.5.3.2.1 while otherwisethe surrogate approach (see IV.5.3.2.2) should be followed.IV.5.3.2.1 Approach based on available chronic toxicity reference dataWhere the chronic ERV for the metal ions of concern corrected for the molecular weight ofthe compound (further called as chronic ERV compound ) is greater than 1 mg/l, the metalcompounds need not to be considered further in the classification scheme for long-termhazard.Readily soluble metal compoundsReadily soluble metal compounds are classified on the basis of chronic ERV of the dissolvedmetal ion, corrected for the molecular weight of the compound (further called as chronicERV compound ) .527

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012If there is no evidence of rapid environmental transformation:a) Classify the metal compound as Category Chronic 1 if the chronic ERV compound ≤ 0.1mg/l, an M-factor must also be established as part of this classification (see IV.5.4); orb) Classify the metal compound as Category Chronic 2 if the chronic ERV compound >0.1mg/l and ≤ 1 mg/l.If there is evidence of rapid environmental transformation:c) Classify the metal compound as Category Chronic 1 if the chronic ERVcompound ≤0.01 mg/l,an M-factor must also be established as part of this classification (seeIV.5.4); ord) Classify the metal compound as Category Chronic 2 if the chronic ERV compound >0.01mg/l and ≤ 0.1 mg/l; ore) Classify the metal compound as Category Chronic 3 if the chronic ERV compound >0.1mg/l and ≤ 1 mg/l.Poorly soluble metal compoundsWhere the chronic ERV for the metal ions of concern is greater than 1 mg/l, the metals neednot be considered further in the classification scheme.Where the chronic ERV compound is less than or equal to 1 mg/l consideration must be given tothe data available on the rate and extent to which these ions can be generated from the metalcompound. Such rate and extend data, to be valid and useable should have been generatedusing the Transformation/Dissolution Protocol (Annex 10 to UN GHS) for a 28d period.Where 28d T/Dp data are unavailable, the surrogate approach should be applied (see section5.3.2.2).Where 28d data from the Transformation/Dissolution protocol are available, then classifyaccording to the following rules:a) Classify the metal compound as Category Chronic 1 if the dissolved metal ionconcentration obtained from the 28 day transformation test at a loading rate of 0.1mg/l is greater than or equal to the chronic ERV, an M-factor must also be establishedas part of this classification (see IV.5.4); orb) Classify the metal compound as Category Chronic 2 if the dissolved metal ionconcentration obtained from the 28 day transformation test at a loading rate of 1 mg/lis greater than or equal to the chronic ERV.If there is evidence of rapid environmental transformation:c) Classify the metal compound as Category Chronic 1 if the dissolved metal ionconcentration obtained from the 28 day transformation test at a loading rate of 0.01mg/l is greater than or equal to the chronic ERV, an M-factor must also be establishedas part of this classification (see IV.5.4); ord) Classify the metal compound as Category Chronic 2 if the dissolved metal ionconcentration obtained from the 28 day transformation test at a loading rate of 0.1mg/l is greater than or equal to the chronic ERV; ore) Classify the metal compound as Category Chronic 3 if the dissolved metal ionconcentration obtained from the 28 day transformation test at a loading rate of 1 mg/lis greater than or equal to the chronic ERV.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Do not classify for long-term hazard if the dissolved metal ion concentration obtained fromthe 28 day Transformation/Dissolution test at a loading rate of 1 mg/l is less than the chronicERV of the dissolved metal ion.IV.5.3.2.2 The surrogate approachReadily soluble metal compoundsIn absence of relevant chronic toxicity data, and unless there is evidence of both rapidenvironmental transformation and evidence of no bioaccumulation (see sections IV.3 andIV.4), Readily soluble metal compounds are classified as:a) Category Chronic 1 if the acute ERVcompound ≤ 1 mg/l, an M-factor must also beestablished as part of this classification (see IV.5.4); orb) Category Chronic 2 if the acute ERVcompound > 1mg/l and ≤ 10 mg/l; orc) Category Chronic 3 if the acute ERVcompound > 10mg/l and ≤ 100 mg/l.Poorly soluble metal compoundsWhere the acute ERV compound is less than or equal to 100 mg/l consideration must be given tothe data available on the rate and extent to which these ions can be generated from the metal.Such rate and extend data, to be valid and useable should have been generated using theTransformation/Dissolution Protocol (Annex 10 to UN GHS) for a 7d period.Where such 7d T/Dp data are unavailable, i.e. there is no clear data of sufficient validity toshow that the transformation to metal ions will not occur; the safety net classification(Category Chronic 4) has to be applied.Where T/Dp data are available but relevant chronic ERVs are absent, the results should beused to aid classification according to the following rules:a) Classify the metal compound as Category Chronic 1 if the dissolved metal ionconcentration obtained from the 7 day transformation test at the low loading rate (1mg/l) is greater than or equal to the acute ERV and there is no evidence of rapidenvironmental transformation and no bioaccumulation, an M-factor must also beestablished as part of this classification(see IV.5.4);b) Classify the metal compound as Category Chronic 2 if the dissolved metal ionconcentration obtained from the 7 day transformation test at the medium loading rate(10 mg/l) is greater than or equal to the acute ERV and there is no evidence of rapidenvironmental transformation and no bioaccumulation;c) Classify the metal compound as Category Chronic 3 if the dissolved metal ionconcentration obtained from the 7 day transformation test at the high loading rate (100mg/l) is greater than or equal to the acute ERV and there is no evidence of rapidenvironmental transformation and no bioaccumulation;d) Classify the metal compound as Category Chronic 4 if the dissolved metal ionconcentration obtained from the 7 day transformation test at the high loading rate (100mg/l) is lower than the acute ERV and there is no evidence of rapid environmentaltransformation and no bioaccumulation.529

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure IV.5.3.2a Classification strategy for determining long-term aquatic hazard for metalcompounds.Is chronic ERV available?YESNOGo to FigureIV.5.3.2bIs it readily soluble(solubility ≥ acuteERV)?YESIs there evidence of rapid environmentaltransformation?NOClassifyChronic 1and addM-factor(see IV5.4)YESYESChronic ERV compound ≤ 0.01 mg/lNONOChronic ERV compound ≤ 0.1mg/lNOYESClassifyChronic 1and addM-factor(see IV5.4)ClassifyChronic 2YESChronic ERV compound > 0.01mg/land ≤ 0.1 mg/lNOChronic ERV compound >0.1mg/l and ≤ 1 mg/lNOYESClassifyChronic 2ClassifyChronic 3YESChronic ERV compound > 0.1mg/land ≤ 1 mg/lNODo not classify for longtermaquatic hazardChronic ERV compound < 1 mg/lYESNODo not classify forlong-term hazard28 days T/D full test data availableYESNOGo to FigureIV.5.3.2b anduse the surrogateapproachIs there evidence of rapid environmentaltransformation?ClassifyChronic 1and addM-factor(see IV5.4)YESYESConcentration at 0.01 mg/lloading rate ≥ chronic ERVNONOConcentration at 0.1 mg/lloading rate ≥ chronic ERVNOYESClassifyChronic 1and addM-factor(see IV5.4)ClassifyChronic 2YESConcentration at 0.1 mg/l loadingrate ≥ chronic ERVNOConcentration at 1 mg/lloading rate ≥ chronic ERVNOYESClassifyChronic 2ClassifyChronic 3YESConcentration at 1 mg/l loadingrate ≥ chronic ERVNODo not classify for longtermaquatic hazard

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Figure IV.5.3.2b Classification strategy for determining long-term aquatic hazard for metalcompounds in absence of appropriate chronic toxicity reference and/or T/Dp data.Is acute ERV available and≤ 100 mg/l?YESNONot possible to classify for long-termaquatic hazard due to insufficient dataIs it readily soluble(solubility ≥ acuteERV)?NOYESIs there evidence of both rapidenvironmental transformation and nobioaccumulation?NOYESacute ERV compound ≤ 1 mg/lClassifyChronic 1and addM-factor(see IV5.4)NOYESacute ERV compound > 1mg/land ≤ 10 mg/lYESClassifyChronic 27 days T/D full test dataavailable?NOClassify Chronic4, unless Acute 1appliesNOacuteERV compound > 10 mg/land ≤ 100 mg/lYESClassifyChronic 3YESYESIs there evidence of both rapidenvironmental transformation andno bioaccumulation?YESIs chronic ERV availableand ≤ 1 mg/l?NONot possible to classify forlong-term aquatic hazarddue to insufficient dataNOConcentration at 1 mg/lloading rate ≥ acute ERVNOYESClassifyChronic 1and addM-factor(see IV5.4)Concentration at 10 mg/lloading rate ≥ acute ERVYESClassifyChronic 2NOConcentration at 100 mg/lloading rate ≥ acute ERVYESClassifyChronic 3NOClassify Chronic4, unless Acute 1applies531

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012IV.5.4 Setting M-factors for metals and inorganic metal compoundsFor the hazard class “Hazardous to the Aquatic Environment”, SCLs are not applicable.Instead the M-factors concept is used.The M-factors are used in application of summation method for classification of mixturescontaining substances that are classified as very toxic. The concept of M-factors has beenestablished to give an increased weight to very toxic substances when classifying mixtures.M-factors are only applicable to the concentration of a substance classified as hazardous tothe aquatic environment (categories Acute 1 and Chronic 1) and are used to derive by thesummation method the classification of a mixture in which the substance is present. They are,however, substance-specific and it is important that they are being established already whenclassifying substances.M-factors should have been established in accordance with Article 10 of CLP and beavailable in the C&L Inventory.For the harmonised classifications in Annex VI to CLP, M-factors shall be set by themanufacturer, importer or downstream user in case there is no M-factor provided, inaccordance with CLP Article 10(4).For soluble metal compounds M-factors are applied as for organic substances (see tableIV.5.4.1).For poorly soluble metal compounds and metals M-factors can be estimated from the ratio ofthe soluble metal ions concentrations obtained from Transformation Dissolution (atrespectively 7 d or 28 d’s for a loading of 1 mg/l) and the ERV of the dissolved metal iontaking the considerations mentioned in I.V.2.3 into account. If this ratio is:- below 10 then an M-factor of 1 should be applied- > 10 and < 100 then the M-factor would be 10,- > 100 and < 1000 then the M-factor would be 100,Continue in factor 10 intervalsTable IV.5.4.1: M-factors for inorganic substances.Acute ERV (mg/L)Multiplying factors (M)0,1 < Acute ERV < 1 10,01 < Acute ERV < 0,1 100,001 < Acute ERV < 0,01 1000,0001 < Acute ERV < 0,001 1000Continue in factor 10 intervals 10000Chronic ERV (mg/L)Multiplying factors (M)No rapidenvironmentaltransformationRapid environmentaltransformation0,01 < Chronic ERV < 0,1 1 10,001 < Chronic ERV < 0,01 10 1

Guidance on the Application of the CLP Criteria Version 3.0 - November 20120,0001 < Chronic ERV < 0,001 100 100,00001 < Chronic ERV < 0,0001 1000 100Continue in factor 10 intervalsIV.5.5 Particle size and surface areaSurface area is a crucial parameter in that any variation in surface area tested may cause asignificant change in the levels of metals ions released in a given time-window. Thus, particlesize or surface area is fixed for the purposes of the transformation test, allowing thecomparative classifications to be based solely on the loading level. Normally, theclassification data generated would have used the smallest particle size marketed to determinethe extent of transformation. There may be cases where data generated for a particular metalpowder are not considered as suitable for classification of the massive forms. For example,where it can be shown that the tested powder is structurally a different material (e.g. differentcrystallographic structure) and/or it has been produced by a special process and is notgenerally generated from the massive metal, classification of the massive can be based ontesting of a more representative particle size or surface area, if such data are available. Thepowder may be classified separately based on the data generated on the powder. However, innormal circumstances it is not anticipated that more than two classification proposals wouldbe made for the same metal.Metals with a particle size smaller than the default diameter value of 1 mm can be tested on acase-by-case basis. One example of this is where metal powders are produced by a differentproduction technique or where the powders give rise to a higher dissolution (or reaction) ratethan the massive form leading to a more stringent classification.The particle sizes tested and/or used for classification and labelling depend on the substancebeing assessed and are shown in the table below:Type Particle size CommentsMetal compoundsMetals – powdersSmallest representativesize soldSmallest representativesize soldNever larger than 1 mmMay need to consider different sources ifyielding different crystallographic/morphologic propertiesMetals – massive 1 mm Default value may be altered if sufficientjustificationMassives will usually be tested as 1 mm particles. Alternatively, the T/D testing of materialswith different surface area’s may result in highly reliable dissolution kinetic equations thatallows to define the "Critical Particle Diameter" (CPD) for appropriate loadings for the acuteand long-term hazard assessment .For most metals and some metal compounds, it is possible, using the Transformation/Dissolution Protocol (Annex 10 to UN GHS), to obtain a correlation between theconcentration of the metal ion after a specified time interval as a function of the surface arealoadings of the forms tested. Such correlations should be established for the relevant pHranges as specified in the protocol. In such cases, it could then be possible to estimate thelevel of dissolved metal ion concentration at a given pH of the metal with different particles,using the critical surface area approach [Skeaff et. al. (2000)]. From this correlation and alinkage to the appropriate toxicity data at corresponding pH level, it is possible to determine a533

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012"Critical Surface Area" (CSA) of the substance that delivers the L(E)C 50 to the dissolutionmedium and then to convert the CSA to a Critical Particle Diameter (CPD) (see example).This CPD at appropriate mass loadings for acute and long-term hazard assessment can then beused to:- determine the classification category of powders based on the finest representativepowder on the market and- determine an accurate classification of the massive metal by applying a 1 mm (default)diameterWithin the CSA Approach an equation is developed to predict metal ion release (based onpreviously measured metal ion release from different loadings of the metal), which iscorrelated to measured surface area, and a corresponding calculated equivalent particlediameter. The basis of the CSA Approach is that the release of metal ions is dependent onthe surface area of the substance, with this release being predictable once the relationshiphas been established. The CSA is the surface area loading (mm 2 /l) to a medium that delivers aselected ecotoxicity reference value to that medium. The term SA is the measured specificsurface area (m 2 /g) of the metal sample. The measured specific critical surface area (SA crit )(m 2 /g) is the measured specific surface areas for the corresponding low, medium and highloadings which are associated with the respective acute and long-term aquatic toxicityclassification categoriess in the classification scheme for metals and metal compounds. Atypical equation for this relationship for a given substance, aquatic medium, pH and retentiontime is:log (C Me(aq) , mg/l) = a + b log(A meas )C Me(aq) = total dissolved concentration of metal ion (mg/l) at a particular length of test time(i.e. 168 hours for acute toxicity transformation testing) under certain conditions (i.e. pH,specified medium, etc.), as determined by transformation/dissolution testing of differentsurface area loadingsa, b = regression coefficientsA meas = initial surface area loading (mm 2 /l) [equals (measured specific surface area, SA, inm 2 /g) X (substance mass loading in g/l) X 10 6 ], where SA was measured with the BETnitrogen adsorption-desorption technique.IV.5.6 Classification of mixtures of metals and metal compoundsSimple composed metal or metal compound mixtures should be handled as mixtures andclassified according to the mixtures rules described in Section 4.1.4 given they normallyexpress toxicity as a function of their composing ingredients. Ores and concentrates andUVCB inorganics are considered as substances in respect to CLP, but follow in general themixture ruling. to determine their classification unless specific ecotoxicity data are availablefor the mineral(s) under consideration.Ores and concentrates and inorganic UVCBs are considered substances under CLP. In theabsence of substance specific ecotoxicity data, their classification can be assessed by applyingthe mixtures rule. The metals industry has developed classification tools that allow for thehazard ID and environmental classification of these complex materials, by integrating allaspects of this guidance with a knowledge of their mineralogical and other typical metalproperties.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Metal alloys are defined by the CLP as “special preparations” because their (eco)toxicityprofile differs from that of their constituents. Further information on how to assess theenvironmental hazard classification of alloys and other complex metal containing materials isprovided hereunder.IV.5.6.1 Classification of alloys and complex metal containing materialsMetal alloys, or alloy manufacturing products are not simple mixtures of metals or metalcompounds, since the alloy has clearly distinctive properties compared to a classical mixtureof its metal components. Justified by their intrinsic properties, the solubility properties candiffer substantially from what is observed for each individual constituent in that alloy (eg therate and extend of metals release from pure metals are different from the ones from alloys).The rate and extend to which the ingredient of the alloy react with the media to transform towater soluble forms can be measured in the same way as with metals (by using the OECDTransformation/Dissolution test (Annex 10 to UN GHS)). However, alloys often react slowlyand to a very limited extent, making the application of the T/D protocol more complex.Special care should be taken in this respect to the detection limit and the accuratedetermination of the measured surface. Initial testing of alloys, using the T/D protocol, showsthat this can be useful but further additional guidance on this aspect is recommended.More complex metals or metal compounds containing inorganic substances like e.g. ores andconcentrates are not simple mixtures of metals or metal compounds. Justified by their intrinsicproperties, the solubility properties can differ substantially from what is observed for eachindividual constituent of that complex substance (e.g. the rate and extent of metals releasefrom e.g. ores/concentrates are different from the ones from simple metals). All thesematerials are typically not readily soluble in any aqueous medium. In addition, these materialsare often heterogeneous in size and composition on a microscopic/macroscopic scale.Therefore, adequate amounts of the material could be used to evaluate the extent to which thesubstances can be dissolved, i.e. its water solubility and/or the extent to which the metals canreact with the media to transform to water soluble forms e.g. throughTransformation/Dissolution tests. Additional guidance on this aspect is needed for complexmetal mixtures.An ecotoxicity validation step may be important for alloys and complex metal containingmaterials (e.g. ores, concentrates, slags), where binding of the metal to abiotic and biologicalbinding sites will in many cases be competitive. Therefore the “additivity mode” is notnecessarily valid and additional information may be relevant.Therefore, information from ecotoxicity validation steps could be useful in cases where asignificant uncertainty is associated with the existing toxicity data. This ecotoxicity validationshould have been derived from tests using most sensitive species at dissolved ionconcentrations equivalent to those measured in the T/D medium. However, information fromecotoxicity testing directly in the T/D medium is not recommended because the compositionof this medium is unlikely to meet the requirements for standard test media to ensure propersurvival and/or reproduction. Therefore, ecotoxicity tests should have been conducted instandard media dosed at metal concentration equivalent to the concentration level actuallymeasured in the T/D medium.IV.6 References535

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Brown, D.S. and Allison, J.D. (1987). MINTEQA1 Equilibrium Metal Speciation Model: Auser’s manual. Athens, Georgia, USEPA Environmental Research Laboratory, Office ofResearch and DevelopmentFarley KJ, Carbonaro RF and Di Toro DM (2007), Unit World Model Tier 1 Hazard RankingModel for metals in lakes. Report prepared for the International Council of Metals andMining (ICMM)McGeer JC, Brix KV, Skeaff JM, DeForest DK, Brigham SI, Adams WJ and A Green,(2003). Inverse relationship between bioconcentration factor and exposure concentration formetals: Implications for hazard assessment of metals in the aquatic environment.Environmental Toxicology and Chemistry 22(5), 1017-1037.DiToro, M.D.; C. D. Kavvadas; R.Mathew, P.R. Paquin and R.P. Winfield . The persistenceand availability of metals in aquatic environments. ICMM, 2001.OECD 2001. OECD SERIES ON TESTING AND ASSESSMENT, Number 33: HARMONISEDINTEGRATED CLASSIFICATION SYSTEM FOR HUMAN HEALTH AND ENVIRONMENTALHAZARDS OF CHEMICAL SUBSTANCES AND MIXTURES.OECD, Paris. http://www.oecd.org/dataoecd/48/51/37182285.pdfOECD, 2000. Guidance Document on Aquatic Toxicity Testing of Difficult Substances andMixtures, OECD, ParisOECD, 2001. Guidance Document on Transformation/Dissolution of Metals and MetalsCompounds in Aqueous MediaSantore, R.C. and Driscoll, C.T. (1995). The CHESS Model for Calculating ChemicalEquilibria in Soils and Solutions, Chemical Equilibrium and Reaction Models. The SoilSociety of America, American Society of AgronomySantore, R.C. and Di Toro, D.M. et al (1999). A biotic ligand model of the acute toxicity ofmetals. II. Application to fish and daphnia exposure to copper. Environ. Tox. Chem.SubmittedSkeaff, J., Delbeke, K., Van Assche, F. and Conard, B. (2000) A critical surface are conceptfor acute hazard classification of relatively insoluble metal-containing powders in aquaticenvironments. Environ. Tox. Chem. 19:1681-1691Tipping, E. (1994). WHAM – A computer equilibrium model and computer code for waters,sediments, and soils incorporating discrete site/electrostatic model of ion-binding by humicsubstances. Computers and Geoscience 20 (6): 073-1023IV.7 Decision on classification: examples for metals and metal compoundsList of examples: Example A: Soluble metal compound with acute and chronic toxicity data and noevidence of rapid environmental transformation (Me 2 (SO4) 2 ). Example B: Poorly soluble metal compound with acute and chronic toxicity data,Transformation/Dissolution data at 7 days (low loading rate) and at 28 days (only lowand medium loading rates) and no evidence of rapid environmental transformation. Example C: Metal in powder and massive form with acute and chronic toxicity dataand Transformation/Dissolution data at 7 days (low, medium and high loading rates)

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012and at 28 days (only the high loading rate) and no evidence of rapid environmentaltransformation.o Explanatory note to Example D - Critical Surface Area (CSA) Approach. Example D: Hazard classification of a soluble metal salt: the case of rapidenvironmental transformation through speciation in the water column.537

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Example A: Soluble metal compound with acute and chronic toxicity data and noevidence of rapid environmental transformation (Me 2 (SO4) 2 ).DATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksTransformation dissolution protocol evidenceScreening test (24 h) at 100 mg/l loadingpH 6 : 6240 µg/lpH 8 : 840 µg/lMetals TDp, non-GLP7 d TDp test Not applicable28 d TDp test Not applicableMWT of the metal ion versus compound 60 / 312Acute aquatic toxicity of metal ion 73Fish: Oncorhynchus mykiss: 120 µg/l (96 h LC 50 ) at pH 7,8106 µg/l (96 h LC 50 ) at pH 7,8104 µg/l (96 h LC 50 ) at pH 7,878 µg/l (96 h LC 50 ) at pH7,8C.1. / static, GLPC.1. / static, non-GLPC.1. / static, GLPC.1. / static, non-GLP(species mean: 102 µg/l at pH 7,8 )Crustacea Daphnia magna: 180 µg/l (48 h EC 50 ) at pH 8 C.2. / static, non-GLPAlgae/aquatic plantsScenedesmus subspicatus:154 µg/l (72 h ErC 50 ) at pH 8C.3. / static, GLPLemna gibba:670 µg/l (7 d ErC 50 ) at pH 8C.26. / semi-static, GLPChronic aquatic toxicity 74Fish:Danio rerio:24 µg/l (28 d NOEC) at pH 6OECD 210 / 28 d flowthrough,non-GLPMarine Fish87 µg/l (28 d NOEC) at pH 81414 µg/l (28 d EC10)OECD 210 /28 d flowthrough, GLP)OECD 210 /28 d flowthrough, GLP)Crustacea:Daphnia magna:37 µg/l (21 d EC 10 ) at pH 7.8C.20. / semi-static, GLP8.6 µg/l (21 d NOEC) at pH 6.4C.20./semi-static non-GLPMarine decapoda1612 µg/l (21 d NOEC)Non standard testAlgae/aquatic plants: Scenedesmus subspicatus: 21.6 µg/l (72 h NOEC) at pH 8 C.3. / static, GLP73 Tests performed with readily soluble salts such as metal sulphates and metal chlorides.74 Tests performed with readily soluble salts such as metal sulphates and metal chlorides.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Degradation (evidence of rapid degradation)8.7 µg/l (72 h NOEC) at pH 6.2 C.3. / static, non-GLPRapid environmental transformationNo evidence.BioaccumulationBioconcentration factor in fish+/- 200 at NOEC levelAquatic hazard assessment, conclusions and comments:Transformation Dissolution : The substance passes the 24 h screening TDp test at pH 6 given the dissolution at aloading of 100 mg/l is 6240 µg/l > acute ERV of the soluble ion being 102 µg/l at pH 7.8.Acute aquatic toxicity: The acute ecotoxicity reference value is driven by the Fish data. No data are available forthe low pH end. The acute ERV for the metal compound is 102 * (312/(2*60)) = 265 µg/l.Evidence of rapid environmental transformation:No information available, so substance considered as not rapidly transformed by normalenvironmental processes.Chronic aquatic toxicity: The chronic aquatic ecotoxicity reference toxicity value based on the lowest of theavailable toxicity values is slightly below 10 µg/l for Daphnia magna at pH 6,4 for themetal ion. The chronic ERV for the metal compound is 8.6 * (312/(2*60)) = 22.4 µg/l.Aquatic hazard classification and, where applicable, established M-factor(s): Acute (short-term) aquatic hazard: category Acute 1, M-factor: 1 Long-term aquatic hazard: category Chronic 1, M-factor: 1Reasoning:Acute aquatic hazard The acute ecotoxicity reference value is driven by the Fish data. A species mean of 102µg/l for the metal ion, is calculated for Oncorhynchus mykiss given 4 or more toxicity datafor the same species under comparable conditions are available. Acute aquatic hazard expressed as the ERV for the metal compound after molecularweight correction ≤ 1 mg/l. M-factor is 1 given the acute ERV is between 1 and 0.1 mg/l.539

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The molecular weight correction recognises that 2 metal ions are included.The substance passes the 24 h screening dissolution test by comparing acute toxicity dataat pH 7.8 with TDp data at pH6 given an acute toxicity data set at pH 6 is lacking and thechronic data indicate more toxic behaviour of the metal at the lower pH end.Long-term aquatic hazard: Adequate information on chronic toxicity (all 3 trophic levels) is available allowing longtermhazard classification (no use of the surrogate approach). 75 Marine toxicity data are not included in the chronic ERV assessment given far lesssensitive as fresh water toxicity references and data for 3 trophic levels for the freshwaterare available. The Daphnia magna reference at pH6 is the lowest and determines the chronic ERV. A molecular weight correction is applied to the substance recognising that 2 metal ions areincluded. Rapid environmental transformation cannot be demonstrated given the lack of sufficientinformation. The M-factor of 1 is based on the chronic ERV of 22 µg/l (so between 0.01 and 0.1 mg/l.)without rapid environmental transformation.Labelling elements based on the classification:ElementGHS PictogramSignal WordCodeGHS09WARNINGHazard Statement H400, H410 H410 76Precautionary statement(s)P273, P391, P50175 In absence of adequate chronic toxicity data for all trophic levels, the subsequent step is to combine two typesof information, i.e. chronic info for the trophic level with such data and acute aquatic toxicity data andenvironmental fate information for lacking info on trophic levels. For details see section 4.1.3.3 and Table 4.1.0.76 In accordance with CLP Article 27, the hazard statement H400 may be considered redundant on the label andtherefore not included on the label because hazard statement H410 also applies, see section 4.1.6 of thisdocument.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Example B: Poorly soluble metal compound with acute and chronic toxicity data,transformation/dissolution data at 7 days (low loading rate) and at 28 days (only lowand medium loading rates) and no evidence of rapid environmental transformationDATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksTransformation dissolution protocol evidenceScreening test (24 h) at 100 mg/l loadingpH 6: 74 µg/lpH 8: 34 µg/lMetals TDp, non-GLP7 d TDp test at 1 mg/l loading pH 6: 50 µg/lpH 8: 16 µg/lMetals TDp, non-GLPMetals TDp, non-GLP28 d TDp test at 0.1 mg/l loadingat 0.01 mg/l loadingpH 6: no data availablepH 8: no data availablepH 6: 9 µg/lMetals TDp, non-GLPMetals TDp, non-GLPMetals TDp, non-GLPpH 8:

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Chronic aquatic toxicity 78Fish: Danio rerio: 24 µg/l (28 d NOEC) at pH 687 µg/l (28 d NOEC) at pH 8Crustacea: Daphnia magna 37 µg/l (21 d EC 10 ) at pH 7.88.6 µgl (21 d NOEC) at pH 6.4Algae/aquatic plants: Scenedesmus subspicatus: 21.6 µg/l (96 h NOEC) at pH 88.7 µg/l (72 h EC 10 ) at pH 6.2OECD 210 / 28 d flowthrough,non-GLPOECD 210 /28 d flowthrough, GLP)C.20. / semi-static, GLPC.20. / semi-static, non-GLPC.3. / static, GLPC.3. / static, non-GLPDegradation (evidence of rapid degradation)Rapid environmental transformationNo data available thereforeconsidered as not rapidlytransformed.BioaccumulationBioconcentration factor in fish+/- 200 at NOEC levelAquatic hazard assessment, conclusions and comments:Transformation Dissolution screening outcome: The substance fail the 24 h screening Transformation Dissolution test given thedissolution at a loading of 100 mg/l :‣ at pH 6 is 74 µg/l < acute ERV of the soluble ion being 78 µg/l (borderline case)‣ at pH 8 is 34 µg/l < acute ERV of the soluble ion being 102 µg/lAcute aquatic toxicity: Adequate data on pH 6 and 8 are available allowing to derive an acute ERV for the(soluble) metal ion :‣ at the lower pH end (around pH 6) : 78 µg/l‣ at the higher pH end (around pH 8) : 102 µg/l7 days Transformation/Dissolution outcome : The acute release after 7 d is the highest at pH 6 (50 µg/l) being lower than the acutetoxicity level (78 µg/l) at this corresponding pH The acute release is lower at or around pH 8 (16 µg/l), which is significantly lower thanthe acute toxicity level (102 µg/l) at this corresponding pH78 Tests performed with readily soluble salts such as metal sulphates and metal chlorides.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Evidence of rapid environmental transformation: No information available and therefore substance considered as not rapidly transformedby normal environmental processes.Chronic aquatic toxicity for a substance not rapidly transformed: The chronic ERV for the (soluble) metal ion is 8.6 µg/l around pH 6 and 21.6 µg/l aroundpH 8.28 days Transformation dissolution outcome for a substance not rapidly transformed: The release after 28 d at pH 6 at a loading of 0.1 mg/l is not available and needs to beextrapolated from the 0.01 loading rate assuming a 10 times higher dissolution level(10x9=90 µg/l), which is significantly larger than the chronic ERV at pH 6 (8.6 µg/l). The release for the 0.1 mg/l loading is also extrapolated in the same way and is muchlower at pH 8. The calculated release rate of < 10 µg/l is still lower than the chronictoxicity level 21.6 µg/l at this pH level. The calculated release rates at 1 mg/l loadingwould be < 100 µg/l which is significantly larger than the chronic ERV at pH 8.Aquatic hazard classification and, where applicable, established M-factor(s):Acute (short-term) aquatic hazard: no acute classificationLong-term aquatic hazard: category Chronic 1, M-factor 10Reasoning:The metal compound is considered as poorly soluble since it fails the OECD transformationdissolution screening test at a 100 mg/l loading. The test confirmed pH 6 as the pH of thehighest release rate.Acute aquatic hazards: The acute ecotoxicity reference value is driven by the Fish data for the high pH and byalgae data for the low pH level. For the high pH end (around pH 8) a species mean of 102µg/l for the metal ion is calculated for Oncorhynchus mykiss and a single reference of 78µg/l for Scenedesmus subspicatus at around pH 6. A poorly soluble substance is evaluated for classification by comparing the dissolvedmetal ion level resulting from the TDp at 7d, at a loading rate of 1 mg/l with the acuteERV as determined for the (soluble) metal ion. A molecular weight correction for thepoorly soluble metal compound is consequently not required given this factor has alreadybeen included for the loading rate of the TDp test. The dissolution level of the poorly soluble metal compound from the 7d TDp at 1 mgloading is lower than the acute ERVs of the soluble metal ion for both pH levels, therebynot resulting in an acute classification.Long-term aquatic hazard:543

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Adequate information on chronic toxicity (all 3 trophic levels) for the higher and lowerpH levels are available allowing direct long-term hazard classification (no use of thesurrogate approach).No valid info is available on rapid transformation by normal environmental processes sothe poorly soluble metal compound is considered to be not rapidly transformed.No Molecular Weight Correction is applied for the poorly soluble metal compound giventhe classification scheme is based on the comparison of the dissolved fraction of thepoorly metal compound with the chronic ERV of the soluble metal ion at both pH 6 andpH 8.No TDp data are available for the 0.1 mg/l and 1 mg/l loading. The calculated dissolutionlevel from the 28d TDp at pH 6 at 0.1mg/l loading (+/- 90 µg/l) for the poorly solublemetal compound is much higher than the chronic ERV’s of the soluble metal ion for pH 6(8.6 µg/l) warranting a chronic 1 classification. The classification is much less sensitive atpH 8 given a less toxic and a lower dissolution rate.The M-factor associated with the long-term hazard classification is derived by using thesolubility level derived from the 28d TDp test at the 0,1 mg/l loading (90 µg/l at pH 6)divided by the ERV of the dissolved metal ion (8.6 µg/l at pH 6): 90/8.6=10.45.Accordingly to section IV.5.5.2 the substance will get an M-factor 10, given this factorwas between 10 and 100.Labelling elements based on the classification:ElementGHS PictogramSignal WordHazard StatementPrecautionary statement(s)CodeGHS09WARNINGH410P273, P391, P501

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Example C: Metal in powder and massive form with acute and chronic toxicity dataand Transformation/Dissolution data at 7 days (low, medium and high loading rates)and at 28 days (only the high loading rate) and no evidence of rapid environmentaltransformationDATA ELEMENTS Value Test method ((EC) No.440/2008) or OECDguideline / remarksTransformation dissolution protocol evidenceFor metal in POWDER formScreening test (24 h) at 100 mg/l loading Not applicable for metals Metals TDp, non-GLP7 d TDp test at 1 mg/l loading pH 6 : 1.7 µg/l (.)Metals TDp, non-GLPat 10 mg/l loadingat 100 mg/l loadingpH 8 : 3 µg/lpH 6 : 24 µg/lpH 8 : 29 µg/lpH 6 : 340 µg/lpH 8 : 280 µg/l28 d TDp test at 1 mg/l loading pH 6: 2.3 µg/lMetals TDp, non-GLPat 0.1 mg/l loadingat 0.01 mg/l loadingpH 8: 3.5 µg/lno measured data availableno measured data availableMWT of the metal 59Acute aquatic toxicity of metal ion 79Fish: Large data sets available for the 2pH ends but less sensitive thancrustacean at high pH end andAlgae at low pH endCrustaceaAlgae/aquatic plantsCeriodaphnia dubiaPseudokirchneriella subcapitataMost sensitive species at high phend (pH 8.3-8.7) : Geometric meanfor 6 values under comparable testconditions (EC 50 48h ): 68 µg metalion/lData sets available for the 2 pHends but less sensitive thancrustacean at high pH end and mostsensitive endpoint at low end.Most sensitive value (96 h EC 10 ) atthe low pH range: 120 µg metalion/lC.1. / static, non-GLPC.1. / static, GLPC.2. / static, non-GLPC.3. / static, GLPAnd non-GLPC.26. / static, non GLP79 Tests performed with readily soluble salts such as metal sulphates and metal chlorides.545

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Chronic aquatic toxicity 80Fish: Large data sets available fordifferent pHs but less sensitive thancrustacean at high and low pHCrustacea:Ceriodaphnia dubiaMost sensitive species at high andlow pH end:C.20. / semi-static, non-GLP- At low pH (NOEC 21d): 20 µg/l- At high pH: (EC10 21d): 2.4 µg /lAlgae/aquatic plants:Large data sets available fordifferent pH’s but less sensitivethan crustacean at high and low pHC.3. / static, GLPC.3. / static, non-GLPDegradation (evidence of rapid degradation)Rapid environmental transformation No information. .BioaccumulationBioconcentration factor in fish toxicity referencevalue?1 low 0.0017 0.12 No10 low 0.024 0.12 No100 low 0.35 0.12 Yes1 high 0.003 0.068 No10 high 0.029 0.068 No80 Tests performed with readily soluble salts such as metal sulphates and metal chlorides.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012100 high 0.28 0.068 Yes* pH value at which dissolution testing was conducted and similar to the pH for the acute toxicity reference valueThe release from the metal powder 81 at a loading of 100 mg/l is for both pH ranges higherthan the acute ERV.7 days Transformation/Dissolution outcome for the massive form :The CSA Approach can be used to calculate a Critical Particle Diameter (CPD) for thedissolution rates from the metal powder. The metal in massive form will be classified ashazardous to the aquatic environment if the CPD is above or equal to 1 mm. The measuredcritical surface area (SA crit ) that releases sufficient ions to reach the acute ERV for the mostcritical pH (6) is SA crit 0.101 m 2 /g corresponding to an equivalent critical spherical particlediameter (CD spec ) of 6.67 m at a 100 mg/l loading rate. This is far less than 1 mm.Evidence of rapid environmental transformation: No information available and therefore substance considered as not rapidly transformedby normal environmental processes.Chronic aquatic toxicity: The chronic ERV for the (soluble) metal ion is 2.4 µg/l at around pH 8 and 20 µg/l aroundpH 6 which is an inverse relationship with pH as for the acute level.28 days Transformation/Dissolution outcome for a substance not rapidly transformed: The release after 28 d at a loading of 1 mg/l is slightly higher at pH 8 (3.5 µg/l) than at pH6 (2.3 µg/l). TDp data for lower loadings are not available and were calculated given that the rate ofmetal ion release from the metal in the OECD 203 medium at high pH at the 28 days canbe predicted by the equation: log (C Me(aq) ) = -5.144 + 1.0229log(A meas ), wherebyC me(aq) = total dissolved concentration of metal (mg/l)A meas = initial surface area loading (mm 2 /l) [equals (measured specific surface area,SA, in m 2 /g) (substance mass loading in g/l) X 10], where SA was measuredwith the BET nitrogen adsorption-desorption technique.An equal approach can be followed for the lower pH level.81 The finest representative metal powder should be used for TDp testing.547

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Measured and estimated transformation dissolution data for the metal powder are listed inthe table belowLoading (mgmetal ion/l)MeasuredcalculatedorpH*Highestdissolutionmetal/l)(mgReference toxicityvalue (mg metal/l)Dissolution > toxicityreference value?1 Measured low 0.0023 0.020 No1 Measured high 0.0035 0.0024 Yes0.1 Estimated Low 0.00023 0.020 No0.1 Estimated High 0.00035 0.0024 No* pH value at which dissolution testing was conducted and similar to the pH for the acute toxicity reference valueThe release after 28 days at the 1 mg/l loading for the higher pH level slightly exceeds the chronicERV, while no such effect is noted at pH 6 mainly due to the lower sensitivity of the species.Aquatic hazard classification and, where applicable, established M-factor(s):Acute (short-term) aquatic hazard:- for the powder form: no acute hazard classification- for the massive form: no acute hazard classificationLong-term aquatic hazard:- for the powder form: category Chronic 2- for the massive form: no long-term hazard classificationReasoning :The single environmental classification for all metal powders (spherical diameter ≤ 1 mm) of theconsidered metal can be derived by comparing the transformation/dissolution data for the smallestcommercially representative metal powder with the acute and chronic toxicity reference values (for thesoluble metal compounds).Acute hazard classification: The dissolution rate for the finest powder on the market does not reach the concentrationcorresponding with the ERV, within 7 days at a loading of 1 mg/l. This is only reached at aloading of 100 mg/l. Therefore, no acute hazard classification is required. The dissolution rate for the massive forms (spherical diameter > 1 mm) is lower than those forpowders given the lower available surface area. The Critical surface area approach confirms thatabove a diameter of 6.7 µm the acute ERV cannot be reached within 7 days at a loading of 1 mg/l.(Not even at a 100 mg/l loading.) Thereby confirming no need for an acute hazard classification.More explanation on the CSA assessment of the powder form for this metal is included in theexplanatory note to example D (see below).Long-term hazard classification:The metal does not fulfil the criterion for rapid environmental transformation.T/D data are only available for 1 mg/l loading rate. The medium loading rate of 0,1 mg/l requiredfor the long-term hazard assessment could be safely extrapolated from existing evidence givenclear relationships between concentration and dissolution were established for both pH levels.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012The comparison of chronic ERV’s with the 28 days TDp results concludes that the chronic ERVfor the metal ion is only reached at a loading rate of 1 mg/l at pH 8. Therefore, chronic 2 hazardclassification for the metal in the powder form is warranted..Given the surface of the particle reference for massive metal is > 100 larger than for thesmallest commercially representative form this corresponds to a Critical Particle Diameter >1 mm at the high loading rate. Therefore there is no need to classify the massive form forlong-term hazard.Labelling elements based on the classification for the powder form:ElementGHS PictogramSignal WordHazard StatementPrecautionary statement(s)CodenonenoneH411P273, P391, P501Labelling elements based on the classification for the massive form: noneElementGHS PictogramSignal WordHazard StatementPrecautionary statement(s)Codenonenonenonenone549

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Explanatory note to Example C - Critical Surface Area (CSA) approachAcute hazard:For the metal powder in this example, the data showed that the concentration of metal released in the OECD 203medium at pH 8 at the 168 hr can be predicted by the equation:log (C Me(aq) ) = -5.122 + 0.9875 log (A meas )C Mel(aq) = total dissolved concentration of Metal ion (mg/l) at 168 hr and pH 8;A meas = initial surface area loading (mm 2 /l) [equals (measured specific surface area, SA, in m 2 /g) (substance mass loading in g/l) 10 6 ], where SA was measured with the BET nitrogenadsorption-desorption technique.The CSA approach can subsequently determine what surface areas and particle diameters would result indifferent levels of aquatic toxicity classification using the regression coefficients from the above equation, a (-5.122) and b (0.9875), and the proposed acute toxicity reference value (0.068 mg Me/l) as the C Me(aq) . Thecritical surface area (CSA) would be the A meas at which the metal ion is released at the concentration of the acutetoxicity reference value. The following equations can be used to derive these values for this case:log L(E)C 50 = -5.122 + 0.9875 log CSAL(E)C 50 = acute ecotoxicity reference value for classification (mg/l)CSA = critical surface area (mm 2 /l) that releases metal ion in the concentration of the acute ecotoxicityreference value to the aquatic mediumThe CSA can be derived as follows: log L(E)C logCSA50 0.98755.122 For an acute toxicity reference value of 0.068 mg Me/l, the CSA is thus 10,100 mm 2 /l. This is the surface arealoading of metal that will deliver the reference value amount of metal ion to the OECD 203 medium at pH 8 andat a time of 168 hr.The critical specific surface areas, SA crit s for a loading of 1 mg/l will deliver the acute toxicity reference value tothe OECD 203 medium at pH 8 and a time of 168 hr can be calculated by:SA crit = critical specific surface area (m 2 /g) corresponding to the acute ecotoxicity reference valueCP = classification cut-off loading of 1 mg/l that yield a classification as acute 1)Thus, for the metal powder under consideration a CSA of 10.100 mm 2 /l and the CP of 1 mg/l, the SA crit is 10,1m 2 /g.The equivalent critical spherical particle diameter (CD spec ) associated with the acute ecotoxicity reference valueis determined by:CDspec SAcrit6 Me Me = density of the metal (g/cm 3 )CD spec = critical diameter of the sphere (m) corresponding to the acute ecotoxicity reference valueFor the above SA crit of 10,1 m 2 /g, corresponding to the 1 mg/l loading, the critical diameter would be 0,067 m.The EU-CLP system defines that the finest representative metal powder should be used for TDp testing andclassification of the metal powder form.An acute toxicity classification can therefore be assigned to all metal powders (diameter ≤ 1 mm) by measuringthe real surface area using the BET nitrogen adsorption-desorption technique and comparing it to SA crit . If thesurface area of the reference material is greater than the SA crit for the associated acute toxicity classification thenthe representative metal sample would classify for that acute hazard category and classify all powder types ofthat metal in the same way. If the measured surface area is less than the SA crit s of all of the classificationcategories then all powders of this metal would not classify for aquatic toxicity.The CSA Approach can consequently be used to assign an acute hazard classification to the metal powders basedon measured surface area using the measured surface area of0.43 m 2 /g for the smallest representative size

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012powder on the EU market. Since this surface area is greater than 0.1 m 2 /g but less than 1 m 2 /g, there is accordingto this approach no need for an acute hazard classification of the metal powders in this example.The CSA Approach can also be used to calculate a Critical Particle Diameter (CPD) to be used to determine anaccurate classification of the metal massive (diameter > 1 mm), where the measured surface area of the testedgranules is 0.086 m 2 /g. This surface area is far less than all of the SA crit so there is no need for an acuteclassification for the metal massive.Long-term hazard: For this example it has been shown that rate of metal ion release from the metal in the OECD203 medium at high pH at the 672 hr can be predicted by the equation:log (C Me(aq) ) = -5.144 + 1.0229log(A meas )C me(aq) = total dissolved concentration of metal (mg/l)A meas = initial surface area loading (mm 2 /l) [equals (measured specific surface area, SA, in m 2 /g) (substance mass loading in g/l) X 10 6 ], where SA was measured with the BET nitrogenadsorption-desorption technique.The CSA Approach can determine what surface areas and particle diameter would result in chronic (long-term)hazard classification by using the regression coefficients from the above equation, a (-5.144) and b (1.0229), andthe proposed chronic toxicity reference value (0.0024 mg Me/l) as the C Me(aq) . The critical surface area (CSA)would be the A meas at which metal ion is released at the concentration of the chronic toxicity reference value. Thefollowing equations can be used to derive these values.log chronic toxicity = -5.144 + 1.0229log CSAchronic toxicity = chronic ecotoxicity reference value for classification (mg/l), using calculated EC 10 s ormeasured NOECs (if the EC 10 is less than the NOEC)CSA = critical surface area (mm 2 /l) that releases metal in the concentration of the chronic toxicityreference value to the aquatic mediumThe CSA can be derived as follows: logchronictoxicity 5.144 logCSA 1.0229 For the chronic hazard classification derivation exactly the same approach as for the acute hazard assessment canbe followed to define SA crit and CD spec . For this metal powder example this results in a CSA of 3,420 mm 2 /l andthe CP of 1 mg/l, the SA crit is 0.342 m2/g.For a SA crit of 0.342 m 2 /g, corresponding to the 1 mg/l loading, the critical diameter would be 2 m.Equivalent as for the assessment of the acute hazard the CSA Approach can be used to assign a long-term hazardclassification to all powders based on measured surface area of the reference powder, using the measured surfacearea at 100 mg/l loading (0.43 m 2 /g) for the smallest representative size powder on the EU market. Since thissurface area is greater than 0.342 m 2 /g, all metal powders would be classified as Chronic 3.The CSA Approach can also be used to classify the massive metal (diameter > 1 mm), where the measuredsurface area of the massive at 100 mg/l loading) is 0.086 m 2 /g. This surface area is less than the chronic SA crit sothe massive metal form would not be classified for long-term environmental hazard.551

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Example D: Hazard classification of a soluble metal salt: the case of rapidenvironmental transformation through speciation in the water columnGeneral approachThis example was selected to:(i) illustrate the use of information on the metal oxidation and resulting transformationof metal ions in the water column for classification decisions;(ii) provide further information related to testing of sparingly soluble metal salts.The metal ion selected for this example, Me(II), is unstable when its solutions are exposed toair, and it oxidises to the Me(III), which then forms the familiar insoluble, hydrated,amorphous, gelatinous precipitate, Me(OH) 3 (metal hydroxide). The question then arises as towhether the metal hydroxide precipitate forms rapidly enough to decrease the concentration ofMe(II) and Me(III) ions to levels below which there is no cause for concern over the aquaticenvironment. Consideration of the rates at which Me(II) oxidises to Me(III) is relevant to thisquestion to proof rapid environmental transformation.Additionally, the classification of substances of concern for the aquatic environment requiresevaluation of aquatic toxicity. Results for this case were evaluated against standardacceptability criteria for use in this classification assessment.ResultsAssessment of the rapid environmental transformation:A review of the scientific literature on the oxidation of metal sulphate reveals the following:Metal sulphate reacts with oxygen in water to form metal hydroxide (MeOH 2 ), moderatelyinsoluble, Ksp = 1.6 10 -14 ) this in turn undergoes further oxidation to form metal hydroxide(MeOH 3 ) which is highly insoluble (Ksp = 1 10 -36 ). Formation of metal hydroxide at pHlevels above 5.0 limits the presence of metal ions in aqueous systems. In sediments the metalhydroxide is expected to result in enriched concentrations of insoluble metal sulphide.The rates at which dissolved metal sulphate (Me ++ ) oxidises to (Me +++ ) and forms the metalhydroxide [Me(OH) 3 ] precipitate: Is highly dependent on pH (100 fold from pH 6 to 8); decreases with increase in ionic strength of the aqueous medium (pristine waters containless metal ions); dependent to some extent on the anions present in solution such as sulphate and chloride; increases 10-fold for a 15 C increase in temperature; exhibits a linear dependence on the partial pressure of oxygen; and dependent on the initial concentration of metal sulphate and exhibits linear reactionkinetics at Me(II) loadings less than ~50 micromolar (~3 mg/l). At concentrations greaterthan 50 micromolar, rates of reaction increase with increasing concentration of metalsulfate (about 4 for each order of magnitude).Based on literature data and empirical reaction kinetics, it can be calculated that, at low pH(reasonable worst case scenario) in the OECD 203 medium (diluted by 10 as per theTransformation/Dissolution Protocol), the half-times for the oxidation of Me(II) are 11, 9 and3.6 hr, for 1, 10 and 100 mg/l loadings of MeSO 4 , respectively. At high pH, the reaction isestimated to be as short as 8 seconds. The rapid precipitation of metal ions from aqueoussystems accounts for low “metal” concentrations found in most natural aquatic systems (all

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012except natural waters at very low pH values (i.e. < pH 5.5)). Under the reasonable worst casescenario of low pH and a low initial concentration of 1 mg/l MeSO 4 , the 70 % removal fromsolution is calculated to be achieved in 19hr and 90 % removal would be achieved by 36hr.Since the removal of the metal sulphate are due to reaction with oxygen in water to formhighly insoluble and non classifiable metal hydroxide and the half life for the removal of thesoluble species are less than 16 days this can be considered as rapidly transformed in thewater column and the substance considered for classification purposes as rapidly degradable.To support this, evidence of rapid loss of “Metal ions” (and other metals) from the watercolumn has been reported in mesocosm lake experiments (Perch Lake). The data arepresented as half lives as a function of time, partition coefficient and first stability constant.Half lives for metal ions in the mesocosms are calculated to be approximately 11 days underthe given conditions. The data support that half lives are short and loss from the watercolumn can be related to both formation of the metal hydroxide but also to sorption tosuspended particles that are settling.Aquatic Toxicity:Acute ERV values lie in the range of 1-37 mg/l (see Table). Two values for Daphnia magnawere less than 10 mg/l. Four Daphnia magna studies were performed and the geometric meanvalue for this species is 5.77 mg/l. The values for fish were all greater than 10 mg/l. No algalstudies were deemed reliable. All these values are expressed as mg/l Me. If the classificationrelates specifically to metal sulphate of which the most common form is the heptahydrateMeSO 4 .7H 2 O. The numerical ERV values detailed should be adjusted according to the tablebelow and the species under consideration to calculate the toxicity on a metal sulfate basis.Chemical Species Molecular Weight RatioMeSO 4 7H 2 O 278.0 4.978MeSO 4 H 2 O 169.91 3.043MeSO 4 151.90 2.720Me 55.84 1.0The data cover all the reliable results available for aquatic toxicity of binary “metal” and any observedtoxicity effects could relate to the Me ion which could be in Me(II) or metal Me(III) oxidation states.Conversion of the acute ERV values for the metal ion to those appropriate for MeSO 4 .7H 2 O implies anacute toxicity range of 6.4 to 199 mg/l.Table IV.7.1 Acute toxicity data deemed reliable for “Metal” are presented as mg/l Me.Test substance Test organism Duration Endpoints L(E)C 50 (mg Me L -1 )MeCl 3 .6H 2 O Pimephales promelas 96h Survival 21.8Lepomis macrochirus 96h Survival 20.3MeSO 4 .7H 2 O Oncorhynchus mykiss 96h Survival 16.6Me 2 (SO 4 ) 3 Oncorhynchus mykiss 96h Survival >27.9MeSO 4 Daphnia pulex 24h Immobility 36.9MeSO 4 Daphnia magna 24h Immobility 17MeCl 3 .6H 2 O Daphnia pulex 48h Immobility 12.9Me 2 (SO 4 ) 3 Daphnia longispina 48h Immobility 11.5553

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Test substance Test organism Duration Endpoints L(E)C 50 (mg Me L -1 )MeCl 3 .6H 2 O Daphnia magna 48 h Immobility 9.6MeSO 4 Daphnia magna 24h Immobility 5.25MeSO 4 .7H 2 O Daphnia magna 48h Immobility 1.29Table IV.7.2Chronic toxicity data deemed reliable for “Metal” are presented as mg/l Me.Test substance Test organism Duration Endpoints NOEC/LOECFe(OH) 3 Salvelinus fontinalis 30 days HatchingGrowthSurvivalFe(OH) 3 Oncorhynchus kisuth 30 days HatchingGrowthSurvivalFeCl 3 .6H 2 O Pimephales promelas 33 days SurvivalLengthWeightFeCl 3 .6H 2 O Daphnia pulex 21 days ImmobilityTotal offspringBrood sizeFeCl 3 .6H 2 O Daphnia magna 21 days ImmobilityReproduction(mg Me L -1 )>10.3>10.32.81/>10.3>10.31.0/1.61.61/2.812.51/5.010.63/1.261.26/2.515.9 EC504.4 EC16Aquatic hazard classification:Acute hazard: Not classified.Long-term hazard: Not classified.Reasoning:Acute aquatic toxicity > 1 mg/l.Since all chronic aquatic toxicity values are higher than 1 mg/l and rapid transformation to a metalhydroxide takes place by normal environmental processes, no classification is warranted.Labelling elements based on the classification:ElementGHS PictogramSignal WordHazard StatementPrecautionary statement(s)Codenonenonenonenone

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012V ANNEX V: COLLECTION OF INTERNET LINKS FOR THE USERS OF THEGUIDANCEReference/Site name Host URLECHA website ECHA http://echa.europa.eu/web/guestUN GHS UN http://www.unece.org/trans/danger/publi/ghs/ghs_welcome_e.htmleChemPortal OECD http://www.echemportal.org/OECD Series onTesting and AssessmentEU Test MethodRegulation 440/2008OECDEChttp://www.oecd.org/document/30/0,3746,en_2649_34377_1916638_1_1_1_1,00.htmlREACH guidance ECHA http://echa.europa.eu/web/guest/support/guidance-on-reach-and-clpimplementationhttp://eurlex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32008R0440:EN:NOTOECD test guidelines OECD http://www.oecd.org/findDocument/0,3354,en_2649_34377_1_1_1_1_1,00.htmlPublic C&L Inventory ECHA http://www.echa.europa.eu/web/guest/information-on-chemicals/clinventory-database555

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012VIANNEX VI: BACKGROUND DOCUMENT TO THE GUIDANCE FORSETTING SPECIFIC CONCENTRATION LIMITS FOR SUBSTANCESCLASSIFIED FOR REPRODUCTIVE TOXICITY ACCORDING TOREGULATION (EC) NO 1272/20081 Executive summaryRegulation (EC) No 1272/2008 on the classification, labelling and packaging of substancesand mixtures (CLP Regulation or CLP) contains rules including criteria for the classificationof substances and mixtures. While the classification of substances for human health hazards isbased on specific criteria for each hazard class, the classification of mixtures is mainly basedon the concentration and the classification of the substances contained in the mixture. CLPincludes generic concentration limits (GCLs) which are specific for a hazard class andcategory and which indicate a threshold above which the presence of a substance in a mixtureleads to classification of the mixture. However, under certain conditions specificconcentration limits (SCLs) must or may be used . As the Regulation itself does not provideany further guidance on when and how to set SCLs, guidance has been developed for certainhazard classes (see the respective chapters on setting SCLs in Part 3 of the Guidance on theApplication of the CLP Criteria).This Annex provides a background to the method for the determination of SCLs forsubstances classified as reproductive toxicants as outlined in the guidance in Part 3.The potency, expressed as the dose for the induction of reproductive effects was identified asthe best determinant for setting SCLs. The ED 10 for effects warranting classification wasselected as the most appropriate parameter for estimating the potency. The ED 10 is the doselevel which induces reproductive effects in 10% of the animals above the control group or achange of 10% in the effect compared to the control group. Based on the ED 10 the substanceis placed in a potency group. However, modifying factors can alter the potency group,especially when the potency estimate is close to the boundary between two groups.The distribution of the potency of a large number of substances classified in Annex VI to CLPas developmental toxicants and/or substances affecting sexual function and fertility wasdetermined by means of establishing two databases. In line with other methods for settingSCLs for other hazard classes, it is proposed to define three potency groups. The boundariesfor the potency groups were determined in line with the provisions outlined in Article 10(1) ofCLP, the results of the database analyses and policy considerations. Most substances areforeseen to fall into the medium potency group which is linked to the GCL. For substances inthe high and low potency group, the SCLs included in the table below are proposed.HighpotencygroupCategory 1 Category 2Dose SCL Dose SCLED 10 below 4mg/kg bw/day0.03%(factors of 10lower forextremely potentsubstances B )ED 10 below 4mg/kg bw/day0.3%(factors of 10lower forextremely potentsubstances)

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012MediumpotencygroupLowpotencygroupED 10 > 4 mg/kgbw/day, and 4 mg/kgbw/day, and < 400mg/kg bw/day3% ED 10 above 400mg/kg bw/day3% (GCL)3-10% AA The limit of 10% may be considered in certain cases, such as for substances with a ED 10 value above 1000mg/kg bw/day and a NOAEL below 1000 mg/kg bw/dayB For substances with an ED 10 more than 10 fold below 4 mg/kg bw/day, meaning an ED 10 below 0.4 mg/kgbw/day, a 10-fold lower SCL should be used. For even more potent substance the SCL should be lowered with afactor of 10 for every factor of 10 the ED 10 is below 4 mg/kg bw/day.2 Introduction2.1 General description of the classification system for reprotoxic substances andmixturesRegulation (EC) No 1272/2008 (CLP) contains rules for the classification of substances andmixtures. In chapter 3.7 of Annex I to this Regulation, criteria are given for the classificationof substances as reprotoxicants in one of the following categories:Category 1: Known or presumed human reproductive toxicantSubstances are classified in Category 1 for reproductive toxicity when they areknown to have produced an adverse effect on sexual function and fertility, oron development in humans or when there is evidence from animal studies,possibly supplemented with other information, to provide a strong presumptionthat the substance has the capacity to interfere with reproduction in humans.The classification of a substance is further distinguished on the basis ofwhether the evidence for classification is primarily from human data (Category1A) or from animal data (Category 1B).Category 1A: Known human reproductive toxicantThe classification of a substance in Category 1A is largely based on evidencefrom humans.Category 1B: Presumed human reproductive toxicantThe classification of a substance in Category 1B is largely based on data fromanimal studies. Such data must provide clear evidence of an adverse effect onsexual function and fertility or on development in the absence of other toxiceffects, or if occurring together with other toxic effects the adverse effect onreproduction is considered not to be a secondary non-specific consequence ofother toxic effects. However, when there is mechanistic information that raisesdoubt about the relevance of the effect for humans, classification in Category 2may be more appropriate.Category 2: Suspected human reproductive toxicantSubstances are classified in Category 2 for reproductive toxicity when there issome evidence from humans or experimental animals, possibly supplementedwith other information, of an adverse effect on sexual function and fertility, oron development, and where the evidence is not sufficiently convincing to placethe substance in Category 1. If deficiencies in the study make the quality of557

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012evidence less convincing, Category 2 could be the more appropriateclassification. Such effects shall have been observed in the absence of othertoxic effects, or if occurring together with other toxic effects the adverse effecton reproduction is considered not to be a secondary non-specific consequenceof the other toxic effects.Effects on or via lactation are also part of the hazard class reproductive toxicity. Classificationfor these effects is independent of the classification in the classes 1A, 1B or 2 as describedabove. Development of a method for the determination of SCLs for substances with effects onor via lactation is outside the scope of this document. Therefore, these effects and thisclassification are not further considered in this document.The classification of mixtures containing substances classified for reproductive toxicity and ofsubstances containing impurities, additives or constituents classified for reproductive toxicityis based on the concentration of the reproductive toxic component(s). Table 3.7.2 of Annex Ito CLP contains GCLs above which classification for reproductive toxicity is required. TheGCL is 0.3% for reprotoxicants Category 1A and 1B and 3.0% for Category 2. However, aGCL for all substances may not be protective for high potency substances and may beoverprotective for substances with a low potency. Therefore, SCLs may be needed for suchsubstances.According to CLP Article 10, SCLs shall be set where adequate and reliable scientificinformation shows that the hazard of a substance is evident at a level below the GCL. Thisresults in SCLs below the GCLs. SCLs above the GCLs may be set in exceptionalcircumstances where adequate, reliable and conclusive scientific information shows that ahazard of a substance is not evident at a concentration above the GCL. Normally, substancesthat fulfil the criteria for reproductive toxicity are subject to a harmonised classification andlabelling and included in Annex VI to CLP. In such cases, SCLs are set via the procedure forharmonisation of classification and labelling of substances in line with CLP Article 37. Whenthere is no such harmonised entry in Annex VI to CLP, a manufacturer, importer ordownstream user must self-classify reproductive toxic substances and must set lower or mayset higher SCLs than the GCLs if justified according to CLP Article 10(1). He may alsoprovide a proposal for a harmonised classification (CLP Article 37(2)), including an SCLwhere appropriate.2.2 Description of the process for the development of a method to set SCLs forreproductive toxic substancesThere are no hazard specific criteria for the setting of SCLs in CLP . According to CLPArticle 10 (7), the European Chemicals Agency (ECHA) is required to provide furtherguidance on the setting of SCLs. A working group was established to develop such guidancefor the hazard class reproductive toxicity, with the exception of the effects on or via lactation.The work on the proposal for guidance on the determination of SCLs for reproductivetoxicants was initiated by an EU working group of the TC C&L (Technical Committee onClassification and Labelling of Dangerous Substances), continued under the REACHImplementation Project (RIP) 3.6 and subsequently under the auspices of ECHA.To get an impression of the possible parameters for potency and their distribution, twodatabases were compiled, containing several parameters for a large number of substancesclassified for developmental toxicity and impaired fertility. Based on the compiled datachoices were made for the most appropriate parameter, the boundaries of the potency groupsand the associated SCLs.In the course of the guidance development, three documents have been produced. The firstdocument is the actual guidance chapter included in the Guidance on the Application of the

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012CLP Criteria. The second document is this annexed background document, describing theprocess and considerations and providing the rationale for the proposed guidance. The thirddocument is a publication of the databases of parameters for developmental toxicants andsubstances with an effect on sexual function or fertility and the analyses of the databases[(Muller et al., 2012)]Chapter 2 of this document describes potency parameters and contains a number of theoreticalconsiderations on the determination of the most appropriate parameter and the SCLs. Adescription of the databases and the analyses is also provided in this chapter. Chapter 4 isdedicated to the non-modifying factors. Chapter 5 describes and justifies the potencyboundaries and corresponding SCLs.2.3 Considering potency in setting specific concentration limits for various healthhazardsThe criteria for classification for reproductive toxicity are based on the strength of scientificevidence that the substance can cause reproductive toxicity. In general, no specificconsiderations are given to the potency of the substance to induce reproductive toxicity.On the other hand, classification for several other health hazard classes is based on potency.Substances with different potency are classified in different categories within the hazard class.The classification of mixtures for that hazard class is then based on the concentration of thesubstance in the mixture and the hazard category or the potency (for acute toxicity) of thesubstance.For acute toxicity, the potency is based on the acute toxicity estimate (ATE). The ATE is thedose level which induces 50% mortality in a acute toxicity study (LD 50 or LC 50 ) or theestimated LD 50 or LC 50 using fixed dose procedure or the acute toxic class method. This valueis used to classify a substance into one of several categories. For mixtures, the ATE value isused to estimate the potency of a mixture by calculation. The estimated potency is then usedto classify the mixture into a hazard category.For specific target organ toxicity (STOT) after single and repeated exposure, the potency isdefined as the dose at which the substance shows significant toxic effects in a study. Based onthe potency, a substance is either classified for STOT into one of two hazard categories or notclassified. The classification of a mixture containing a substance classified for STOT dependson the percentage of the substance in the mixture and the hazard category of the substance. Aminimal percentage is included in the criteria. SCLs have to be determined for substanceswith a very high potency.Classification for carcinogenicity is, as for reproductive toxicity, based on the strength ofscientific evidence and again no specific consideration is given to the potency. Theclassification of mixtures containing a carcinogenic substance is based on the GCL unless aSCL has been allocated for that substance as provided in Annex VI to CLP. SCLs forcarcinogenic substances are determined based on the potency for carcinogenic effects basedon the T25. The T25 is defined as the daily dose (in mg/kg bw) inducing a tumour incidenceof 25% upon lifetime exposure after correction for the spontaneous incidence. This is mainlybased on animal studies. Substances are divided into 3 groups based on the T25. High potencysubstances have a T25 < 1mg/kg bw/ day, medium potency substances have a T25 between 1-100 mg/kg bw/day, and T25> 100 mg/kg bw/day for low potency substances. Besides theT25, other elements were included that modify the potency evaluation (Commission WorkingGroup, date unknown). This method has been included in the Guidance on the Application ofthe CLP Criteria.The use of potency for the classification into different categories for several other hazardclasses and the use of the potency to set SCLs for carcinogenic substances, justifies the use of559

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012potency as a first approach also for setting SCLs for reproductive toxic substances. As nodefinition of potency for reproductive toxicants was available, the following definition is usedas a working definition:Reproductive toxicity potency is defined as the dose which induces reproductive toxic effectswith a specific type, incidence and magnitude, considering the study design in terms ofspecies and strain, exposure route, exposure duration, exposure window in the life cycle, andpossible concomitant parental toxicity.According to this definition ‘Potency’ is primarily based on applied dose and can be modifiedby consideration of ‘severity’. Within this definition the dose is defined as the amount ofsubstance to which the animals or humans that showed the effect (meaning type, incidenceand magnitude) were exposed on an mg/kg bw/day basis. The incidence is the proportion ofanimals or humans that showed the effect. The type of effect describes which property of anorgan or system of the animal or human is affected and the magnitude describes the level ofchange compared to the control. Together, the incidence, type and magnitude describe the‘severity’ of the effect, meaning how adverse the effect or combination of effects is. Withspecific incidence, type and magnitude (together specific severity) a comparable level ofseverity is indicated for different effects.The working definition above allows potency to be defined at different levels of specificseverity, for example at the ED 10 and the LOAEL (Lowest Observed Adverse Effect Level),and for different type of effects. Therefore, several possible estimates for potency wereinvestigated.2.4 Parameters for potency for reproductive toxicityA consistent database to derive potency estimates for reproductive toxicity was lacking.Therefore, data on substances classified for effects on reproduction were collected andanalysed. This was done separately for substances with an effect on development andsubstances with an effect on sexual function and fertility because the types of effects clearlydiffer between these two main types of reproductive effects. Therefore, this chapter falls intotwo parts, namely one for parameters for potency of substances with developmental effects(chapter 2.3.1) and one for parameters for potency of substances with effects on sexualfunction and fertility (chapter 2.3.2). As potency is primarily based on the dose in mg/kgbw/day at which different adverse effects are observed, a number of parameters/dosedescriptors (e.g. NOAEL 82 , LOAEL 83 , ED 10 etc.) exist for each type of adverse effect. Thecollected data included the NOAEL, LOAEL and ED 10 (effective dose with a 10% incidenceor effect level above the background) as parameters for the effect on reproduction of eachsubstance. They were further divided into effects fulfilling the criteria for classification(named “LOAEL (classification)” for example) and any effects on reproduction (named“NOAEL (overall)” for example). Together, this sub-division results in 6 different potencyparameters, see Table 1. Other data, e.g. a mutagenicity classification of a substance, the typeof effect at the LOAEL and species used in the test, were also collected. These parameterswere analysed and the results tabulated and plotted graphically. The results are published byMuller et al., 2012. As the data for these two main types of reproductive toxicity wereanalysed separately, the results are provided separately.2.4.1 Potency parameters for developmental toxicants (Muller et al, 2012)Data for one or more of the parameters for development were available for 99 substancesclassified for developmental toxicity when the work on this guidance development started.82 NOAEL means No Observed Adverse Effect Level83 LOAEL means Lowest Observed Adverse Effect Level

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012For almost all substances a LOAEL is available but a NOAEL and ED 10 were sometimesmissing. The absence of a NOAEL is mostly caused by the absence of a dose level without aneffect in the study or database of a substance. The absence of an ED 10 value is mainly causedby the absence of a NOAEL and in most of those cases an ED 10 could only be derived by abenchmark dose (BMD) approach to avoid interpolation between the LOAEL and the vehiclecontrol. Another cause for the absence of ED 10 values is the limited reporting of effect levelsin the consulted study summaries or study reports.The difference in the average value between the highest and lowest of the 6 parameters forpotency is a factor of 4 or less. This is very small compared to the difference in potencybetween substances for each parameter of up to 1,000,000 fold (Table 1). The potencydifference is more pronounced for a NOAEL or LOAEL compared to an ED 10 mainly becausefor most potent substances only a NOAEL and/or a LOAEL was available but not an ED 10 .The available data indicate that there is a close relation between the NOAEL, LOAEL andED 10 for most substances. The average LOAEL is between a factor of 2 and 3 above theaverage NOAEL. The fact that it is not closer to the factor of 3 to 4 that is normally usedbetween dose levels is probably due to the absence of a NOAEL for a number of substances.The average ED 10 (classification), is slightly higher than the average LOAEL (classification).The difference is more pronounced for the “overall” values, namely approximately a factor of2. These findings are caused by both the dose spacing in the studies and the limiteddiscriminative power of the NOAEL approach.Table 1. Average values (assuming log/normal distribution) (in mg/kg bw/day) and potencydifferences for parameters for all developmental toxicants of the database (Muller et al, 2012)Parameter N Average StandarddeviationLowestvalueHighestvaluePotencydifferenceNOAEL (overall) 68 12 10 0.002 684 342000LOAEL (overall) 98 25 13 0.002 2281 1140500ED 10 (overall) 59 43 6 0.3 785 2617NOAEL(classification)LOAEL(classification)ED 10(classification)76 18 11 0.002 1100 55000097 40 13 0.002 2281 114050063 48 6 0.3 933 3110A part of the differences in average values and potency between the different parameters inTable 1 is probably caused by the difference in the number of substances for which aparticular variable is present. When only substances are used for which all 6 parameters werepresent, this reduces the database to 44 substances (Table 2). A part of the difference betweenthe parameters in potency difference can be explained by the unusual dose levels (NOAEL0.026 mg/kg bw/day and LOAEL 0.26 mg/kg bw/day) used in the study for the substance thathad the lowest values for all parameters (cadmium oxide).Table 2. Average values (assuming log/normal distribution) (in mg/kg bw/day) and potencydifferences for parameters for developmental toxicants (N=44) with all 6 parameters (Mulleret al, 2012)Parameter Average Standard Lowest Highest Potency561

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012deviation value value differenceNOAEL (overall) 19 7 0.026 684 26308LOAEL (overall) 58 7 0.260 2281 8773ED 10 (overall) 44 5 0.300 570 1900NOAEL (classification) 25 7 0.026 684 26308LOAEL (classification) 71 6 0.260 2281 8773ED 10 (classification) 49 6 0.300 933 3110Comparing Tables 1 and 2 indicates no major changes in average, standard deviation andhighest value for each parameter. However, the lowest value changes for several parameters.The resulting potency difference becomes much more comparable between the parameters.This indicates that the difference between the parameters in potency difference in Table 1 ismainly due to the absence of an ED 10 for some very potent substances.2.4.2 Potency parameters for substances with an adverse effect on sexual function andfertility (Muller et al, 2012)Data for one or more of the potency parameters were available for 93 substances classified foradverse effects on sexual function and fertility (hereafter called fertility toxicants) when thework with the guidance development started. For all substances, an LOAEL was available buta NOAEL and an ED 10 were sometimes missing. The absence of a NOAEL is mostly causedby the absence of a dose level without an effect in the study or database of a substance. Theabsence of an ED 10 value is mainly caused by the absence of a NOAEL and in most of thosecases an ED 10 could only be derived by a Benchmark Dose (BMD) approach to avoidinterpolation between the LOAEL and the vehicle control. Another cause for the absence ofan ED 10 values is the limited reporting of effect levels in the consulted study summaries orstudy reports.The difference in the average values between the highest and lowest of the 6 parameters forpotency is less than a factor of 4. This is small compared to the difference in potency betweensubstances for each parameter of up to 30,000 (Table 3). The difference in potency within theparameters is more pronounced for the NOAEL values than for the values of LOAEL andED 10 , which is mainly due to one substance with a NOAEL of 0.032 mg/kg bw/day but anLOAEL of 10 mg/kg bw/day. The available data indicate that there is a close relation betweenthe NOAEL, LOAEL and ED 10 for most substances. The average LOAEL is between a factor2 and 3 above the average NOAEL. The fact that it is not closer to the factor of 3 to 4 that isnormally used between dose levels is probably due to the absence of an NOAEL for a numberof substances. The average ED 10 is between the average NOAEL and LOAEL.Table 3. Average values (assuming log/normal distribution) (in mg/kg bw/day) and potencydifferences for parameters for all fertility toxicants of the databaseParameter N Average StandarddeviationLowestvalueHighestvaluePotencydifferenceNOAEL (overall) 68 20 7 0.032 635 19844LOAEL (overall) 93 54 7 0.25 2060 8240ED 10 (overall) 37 31 5 0.6 1065 1775NOAEL 70 24 7 0.032 940 29375

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012(classification)LOAEL(classification)ED 10(classification)93 62 7 0.33 2060 624237 33 6 0.6 1065 1775A part of the differences in the average values and in potency between the differentparameters in Table 3 is probably caused by the difference in the number of substances forwhich a particular parameter is present. When only substances are used for which all 6parameters were present, this reduces the database to 34 substances (Table 4).Table 4. Average values (assuming log/normal distribution) (in mg/kg bw/day) and potencydifferences for parameters for fertility toxicants (N=34) with all 6 parametersParameter Average StandarddeviationLowestvalueHighestvalueNOAEL (overall) 19 6 0.3 250 833LOAEL (overall) 72 6 0.7 1000 1429ED 10 (overall) 35 5 1.3 1065 819NOAEL(classification) 24 6 0.3 940 3133LOAEL(classification) 89 6 0.7 1580 2257ED 10 (classification) 39 5 1.3 1065 819PotencydifferenceComparing Tables 3 and 4 indicates no major changes in average, standard deviation andhighest value for each parameter. However, the lowest value changes for some parameters.The resulting potency difference becomes much more comparable between the parameters.This indicates that part of the differences between the parameters in potency difference inTable 3 is due to the absence of an ED 10 for some very potent substances.2.4.3 Conclusions on the most appropriate parameter for potencyAs LOAELs are available for almost all substances, this could be considered the most usefulinformed parameter on which to base potency. However, in the absence of a NOAEL, aLOAEL is not a suitable parameter for potency because there is no indication to what extentthe real LOAEL could be lower than the LOAEL observed. The lower number of substancesfor which an ED 10 is available is probably due to the limitations of the available studysummaries for several substances. Use of the ED 10 requires access to a detailed summary ofthe study or the study report itself which was not available for several substances in thedatabase.However, this guidance will be applied by both industry and Member State CompetentAuthorities when preparing proposals for harmonised classification and labelling, and byindustry in case of self-classification of a reproductive toxic substance for which there is noentry in Annex VI to CLP.Companies have access to their own studies. It is expected that by the completion of theREACH registration deadlines, more detailed information including ED 10 will be available formore substances than in this database used to develop this guidance.563

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Member States will have access to the study summaries in the registrations. The full studiescould be requested by ECHA or by a Member State Competent Authority, according to CLPArticle 49(3).It should be noted that in the absence of a NOAEL, an ED 10 cannot be determined byinterpolation, in case the size of the effect at the LOAEL is more than 10%. However, anED 10 can be estimated using bench mark dose (BMD) software when sufficient data areavailable. A NOAEL and LOAEL cannot be estimated using the BMD approach. In addition,a fixed level of effect of e.g. 10% (ED 10 ) is considered to be more representative for thepotency and facilitates comparisons of relative potency between substances to a greaterextent, than a LOAEL which is a chosen dose level.For most other hazard classes, the SCLs are based on effect levels. For carcinogenicity theT25 is used, and for skin sensitisation the EC 3 value or the dose level with a certain level ofresponders is used. Therefore, the LOAEL or ED 10 is considered a more appropriateparameter for determination of an SCL than the NOAEL.For substances where there is a difference in the LOAEL overall (lowest dose with any effecton reproduction) versus the LOAEL classification (lowest dose with an effect on reproductionfulfilling the classification criteria), this is in most cases due to non-significant increases inlethalities or malformations or decreases in foetal body weight at the LOAEL overall versussignificant increases in lethalities or malformations at the LOAEL classification. Thedifference between significant and non-significant effects will disappear if the ED 10 is used asparameter for potency.The difference in parameters between “overall” and “classification” was sometimes due tolimited effects that normally do not warrant classification such as a small increase invariations at the LOAEL and to more severe effects warranting classification at a higher doselevel. To have a more consistent parameter for potency, it was preferred to use the parametersfor effects warranting classification.Overall, the use of the ED 10 for effects warranting classification is proposed as the mostappropriate estimate for the potency. The advantage of this parameter is that it is a dose levelwith a specified level of effects of at least a certain severity. This is in line with mostclassification criteria and with other methods for the determination of SCLs.Furthermore, not all aspects included in the working definition of reproductive potency arefully taken into account in the ED 10 . Therefore, certain additional parameters should beconsidered which can change the potency group as determined by using the ED 10 , resulting inthe setting of lower or higher concentration limits. See chapter 4 for such modifying factors.3 Modifying factorsSeveral possible elements of reproductive toxicity were considered as elements which shouldalso be taken into account when determining the potency group for reproductive toxicity of asubstance (modifying factors). Modifying factors may change the potency group for asubstance. While some modifying factors should always be taken into account, othermodifying factors could be more relevant when the potency is close to the boundary betweentwo groups (see Table 8 above). It should be noted that several of the elements may beinterrelated.Some factors may have already been taken into account in deciding on the classification as areproductive toxicant. Where such considerations have been made, care should be taken not touse that information again when determining the potency. For example, when the effectsdetermining the ED 10 were observed at dose levels also causing maternal toxicity, this should

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012already have been taken into consideration during the classification and should not be usedagain to set a higher SCL. Factors considered not to be used as modifying factors are includedin section 4.7 of this Annex. The following factors are used as modifying factors: Type of effect / severity Data availability Dose-response relationship Mode or mechanism of action Toxicokinetics Bio-accumulation of substancesThe justification of the use of these modifying factors is provided in the guidance (see section3.7.2.5.5).4 Non-modifying factorsA wide range of parameters were considered as possible modifying factors for thedetermination of reproductive potency. Parameters selected as modifying factors are includedabove. Parameters or factors considered but not included as modifying factors are listedbelow:4. 1 Species and strainsThe species used to determine the ED 10 could be considered as a modifying factor if it isshown that a certain species is generally more sensitive to reproductive toxicants, meaningshowing effects at a lower exposure level, and this can be considered relevant to humans.However, comparison of the different parameters between the two most used species fordevelopmental effects, rats and rabbits, did not indicate a difference in average NOAEL,LOAEL or ED 10 in this analysis. Furthermore, almost all studies that were determinative forthe classification for fertility were studies in rats. Therefore, species is not regarded as amodifying factor. The most sensitive species for each substance has to be used to determinethe potency parameter unless there is clear evidence that the observed effects are not relevantto humans or when there is good evidence for a difference in sensitivity between humans andthe test species. This also applies to different strains.4. 2 Systemic or maternal toxicityAdverse effects on fertility and sexual function may be caused as a secondary effect ofsystemic toxicity to other organs. Developmental effects may be caused as a secondary effectof maternal toxicity. However, this should have already been taken into account forclassifying a substance in a specific category. Therefore, this should not also be used formodifying the concentration limit.4. 3 MutagenicityAnalyses of the databases [(Muller et al., 2012)] indicate that substances classified both forreproductive toxicity and mutagenicity have a higher potency (lower ED 10 ) than substancesclassified for reproductive toxicity only. However, as this higher potency is already includedin the lower ED 10 , there is no need to use mutagenicity as a modifying factor.4.4 VolatilityVolatility is a physical property related to exposure rather than to the intrinsic hazardouspotency of a substance. However, the exposure level to a substance in a mixture is not onlyinfluenced by the concentration but also by the volatility of the substance. The higher thevolatility of a substance the higher the inhalation exposure may be when handling such asubstance in a mixture. Inhalation exposure to vapours are not covered by the experimental565

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012oral testing limit of 1000 mg/kg bw/day as the exposure at workplaces can be more than oneorder of magnitude above the extrapolated exposure level covered by the limit dose(Schneider et al., 2007). This is probably the reason why no limit dose for classification isincluded in the classification criteria (see appendix I, 3.7.2.5.4). Therefore, volatility could beconsidered as a modifying factor.However this argument is not specific for reproductive toxicity and should then apply to allrelevant hazard classes. In methods for setting SCLs for other hazard classes such ascarcinogenicity, the volatility is not used as a modifying factor, although it is suggested to bea factor to take into consideration when setting SCLs for narcotic effects (STOT-SE 3).Further, volatility is not specifically mentioned in the criteria for classification for any otherhazard class other than STOT-SE and -RE (3.8.2.1.10.4 and 3.9.2.10.4) for which theguidance recommends a specific precautionary statement on the label for highly volatilesubstances.However for some hazard classes, volatility is taken into account in the classification ofsubstances and mixtures by using different numeric criteria (acute toxicity, table 3.1.1) orguidance values (STOT-SE table 3.8.2 and STOT-RE, table 3.9.2 and 3.9.3) for vapours thanfor dusts and mists. For STOT-SE and STOT-RE, the method for setting SCLs is directlydepending on these guidance values.It was decided not to include volatility as a modifying factor because it is a physical propertythat depends also on other factors (e.g. temperature and composition of the mixture) and istherefore more related to exposure rather that to the intrinsic hazardous potency of thesubstance.5 Potency groups and specific concentration limits5.1 Justification of the proposed potency boundaries and specific concentrationlimitsIn the following some general considerations on potency groups are first provided, followedby justifications for the approach taken and for the suggested boundaries of the potencygroups and the corresponding concentration limits.5.1.1 General considerations on potency groups5.1.1.1 Legal requirementsAccording to the second subparagraph of CLP Article 10(1)“Specific concentration limits shall be set by the manufacturer, importer or downstream user whereadequate and reliable scientific information shows that the hazard of a substance is evident when thesubstance is present at a level below the concentrations set for any hazard class in Part 2 of Annex I orbelow the generic concentration limits set for any hazard class in Parts 3, 4 and 5 of Annex I.”According to the third subparagraph of CLP Article 10(1)“In exceptional circumstances specific concentration limits may be set by the manufacturer, importer ordownstream user where he has adequate, reliable and conclusive scientific information that a hazard of asubstance classified as hazardous is not evident at a level above the concentrations set for the relevanthazard class in Part 2 of Annex I or above the generic concentration limits set for the relevant hazard classin Parts 3, 4 and 5 of that Annex.”.5.1.1.2 Scientific results of the database analysisThe databases with ED 10 values for substances (Category 1 and 2) with an effect ondevelopment and with an effect on sexual function and fertility were compared to determine

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012whether there is a difference in potency between Category 1 and Category 2 substances[(Muller et al, 2012)]. The results should be carefully interpreted because of the limitations ofthe database: the database is based on a limited number of substances and the available dataper substance is reduced to a single number (ED 10 ) and some modifying factors. Reducing thedata in the database would have included removal of differences in effects and doubtsbetween Category 1 and Category 2. In any case, the comparisons indicate that the averagepotency of substances with an effect on development and with an effect on sexual functionand fertility are comparable and that also the average potencies of Category 1 and 2substances are comparable and certainly do not differ by a factor of 10.5.1.1.3 Policy related considerations and proposed methodData derived from an insensitive test method could in some cases not be regarded as adequate,reliable and conclusive evidence, as mentioned in Article 10 (1) (3 rd para). For example, ascreening assay which only uses a limited number of animals and studied endpoints, cannotbe used to set higher SCLs (but can be used to set lower SCLs). Also a study resulting in anLOAEL without an NOAEL cannot be used to set higher SCLs.Determination of the boundaries of the potency groups (see Table 8) and the SCL or GCL foreach group is a policy related issue. CLP Article 10, the criteria in Annex I to CLP and theavailable data do not give a clear direction. Therefore, a simple system was developed.Furthermore, the approach taken is similar to the one developed for other hazard classes suchas skin sensitization and carcinogenicity, which should be an appropriate justification for thecurrent method.Determination of the potency for reproductive toxicity will in most cases be based on limiteddata from one or a few studies. It was recognised that an exact SCL for each substance thatalso differs for each substance would indicate a precision that is not realistic or scientificallyjustified. Also, Janer (2007) has shown that the variation in the NOAELs of 2-generationstudies for one substance is considerable. Therefore, it is proposed to divide the substancesinto large potency groups with associated SCLs as it is done for other hazard classes. Threepotency groups are proposed. As shown in Table 10 below, substances with the lowestpotency (highest ED 10 ) fall in a group with an SCL above the GCL. Most substances shouldfall in the group with the GCL. Only substances with a very high potency (low ED 10 ) shouldfall in the group with a SCL below the GCL. It is proposed to include approximately 70 –80% in the GCL potency group and 5 to 15% in the low and high potency groups. Further, asthe average potency of developmental toxicants and substances affecting sexual function andfertility are comparable, it is proposed to use the same boundaries for both types of effect.Also, the database shows there is no difference in potency between substances in Category 1and Category 2. Therefore it is proposed to use the same boundaries for Category 1 and 2substances.5.1.1.4. Other methods consideredSeveral other options for a method for determining SCLs were discussed including a methodthat was used by the TC C&L in a limited number of cases in the past. This method is basedon the limit dose of 1000 mg/kg bw/day, as described in the test guideline OECD 414 and416.The concentration limit expressed as a % in mixtures is derived by dividing the NOAEL bythe limit dose followed by multiplication by 100 (see ECBI/47/02 Add.7). This method wouldresult in an individual SCL for each substance. This would indicate a precision that cannot beexpected from standard reproduction studies. Also this would result in an SCL for mostsubstances and in a GCL for only some substances. Therefore, this method was not567

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012considered. Potency groups are used in the proposed method because this does not give theimpression of a high precision and allow the placing of many substances in the mediumpotency group with the connected GCL.5.1.2 Justification of the boundaries between the three potency groups.The estimated percentages of already classified substances in each group for both Category 1and 2 substances with an effect on development or an adverse effect on fertility and sexualfunction are provided in the tables below. They are based on the distribution of potencies ofknown developmental toxicants and of known fertility toxicants (Muller et al., 2012).Several possible values of the boundaries between the three groups are tested. Theestimations are based on counting the number of substances above or below a number ofpossible boundaries and applying some of the modifying factors such as the presence of aNOAEL and considering also the saturated vapour concentration for substances in the lowpotency group. However, the saturated vapour concentration, reflecting volatility, is notproposed as a modifying factor in the guidance.Taking into account all modifying factors for all substances would imply a full assessment ofthe potency for all substances. This was not possible within the available resources. As mostmodifying factors result in a shift from the low potency group into the medium potency groupand from the medium potency group into the high potency group, it is likely that thepercentages in the low potency group may decrease and the percentages in the high potencygroup may increase. (Thus, the effect of volatility on the frequencies in Table 9 should bemarginal.)Based on the ED 10 distribution a rough estimate was made by the Working group of theoptimal boundaries using a range of a factor of 100 for the medium potency group. Then thenumber of substances falling into several combinations of boundaries was estimated.Table 9. Percentages of substances in the three potency groups using the ED 10 and some of themodifying factors for different boundaries of the potency groups and considering the saturatedvapour concentration of low potency substances.Type ofeffect Classification Potency groupBoundaries of the high and low potency groups200mg/kg300mg/kg400mg/kg500mg/kg600mg/kg700mg/kgDevelopment Cat 1A/1B High potency 12,1 13,8 17,2 20,7 20,7 20,7H360D Medium potency 75,9 77,6 79,3 77,6 79,3 79,3Low potency 12,1 8,6 3,4 1,7 0,0 0,0% with SCL 24,1 22,4 20,7 22,4 20,7 20,7Cat 2 High potency 10,3 13,8 13,8 17,2 17,2 20,7H361d Medium potency 72,4 72,4 79,3 75,9 82,8 79,3Low potency 17,2 13,8 6,9 6,9 0,0 0,0% with SCL 27,6 27,6 20,7 24,1 17,2 20,7Fertility Cat 1A/1B High potency 3,4 3,4 3,4 6,9 10,3 13,8H360F Medium potency 89,7 93,1 96,6 93,1 89,7 86,2Low potency 6,9 3,4 0,0 0,0 0,0 0,0% with SCL 10,3 6,9 3,4 6,9 10,3 13,8

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Type ofeffect Classification Potency groupAllBoundaries of the high and low potency groups200mg/kg300mg/kg400mg/kg500mg/kg600mg/kg700mg/kgCat 2 High potency 6,3 9,4 10,9 15,6 15,6 17,2H361f Medium potency 71,9 76,6 81,3 78,1 79,7 79,7Low potency 21,9 14,1 7,8 6,3 4,7 3,1% with SCL 28,1 23,4 18,8 21,9 20,3 20,3avg high potency 8.0 10.1 11.3 15.1 16.0 18.1avg mediumpotency 77.5 79.9 84.1 81.2 82.9 81.1avg low potency 14.5 10.0 4.5 3.7 1.2 0.8avg % with SCL 22,5 20,1 15,9 18,8 17,1 18,9As shown in Table 9 boundaries of 4 to 400 mg/kg bw/day would result in the maximumnumber of substances being included in the medium potency range for most types of effectsand classifications and for both type of effects and classifications combined. Fordevelopmental effects Category 1 and 2 the percentage of substances in the medium potencygroup is within the target of ca. 70-80%. For effects on sexual function and fertility Category2 this is almost the case. Only for Category 1 is this not the case. The percentage ofsubstances in the medium potency group could be reduced by reducing the factor of 100between the boundaries. However, because of the large difference in potency of thesubstances classified for reproductive toxicity of up to a million, this was not considerednecessary. The percentage of substances in the high potency group is higher than thepercentage in the lower potency group for the boundaries of 4 to 400 mg/kg bw/day.However, the percentage of substances in the high potency group was above 15% forsubstances classified for an effect on development in Category 1.Following the PEG consultation, it was agred that volatility was not considered a modifyingfactor and thus, the ED 10 distribution changes as shown in table 10. Borders of 4 to 400 mg/kgbw/day would result in the maximum number of substances being included in the mediumpotency range for most type of effects and classifications and for both type of effects andclassifications combined. However, the same value also applies to some of the other borders.For developmental effects Category 1 and 2 the percentage of substances in the mediumpotency group is within the target of ca. 70-80%. For effects on sexual function and fertilityCategory 2 this is not the case. The percentage of substances in the medium potency groupcould be reduced by reducing the factor of 100 between the borders. However, because of thelarge difference in potency of the substances classified for reproductive toxicity of up to amillion, this was not considered necessary. The percentage of substances in the high potencygroup is approximately the same as the percentage in the lower potency group for the bordersof 4 to 400 mg/kg bw/day.569

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012Table 10. Percentages of substances in the three potency groups using the ED 10 and some ofthe modifying factors but not volatility for different borders of the potency groups.Type ofeffect Classification Potency groupBorders of the high and low potency groups≤2mg/kg≥200mg/kg≤3mg/kg≥300mg/kg≤4mg/kg≥400mg/kg≤5mg/kg≥500mg/kg≤6mg/kg≥600mg/kg≤7mg/kg≥700mg/kgDevelopment Cat 1A/1B High potency 12.1 13.8 17.2 20.7 20.7 20.7H360D Medium potency 67.2 74.1 77.6 75.9 79.3 79.3Low potency 20.7 12.1 5.2 3.4 0 0% with SCL 32.8 25.9 22.4 24.1 20.7 20.7Cat 2 High potency 7.3 9.8 9.8 12.2 12.2 14.6H361d Medium potency 68.2 65.8 70.7 70.7 75.6 78.1Low potency 24.4 24.4 19.5 17.1 12.2 7.3% with SCL 31.7 34.2 29.3 29.3 24.4 21.9Fertility Cat 1A/1B High potency 3.4 3.4 3.4 6.9 10.3 13.8AllH360F Medium potency 86.3 89.7 93.2 89.7 86.3 86.2Low potency 10.3 6.9 3.4 3.4 3.4 0% with SCL 13.7 10.3 6.8 10.3 13.7 13.8Cat 2 High potency 6.3 9.4 10.9 15.6 15.6 17.2H361f Medium potency 68.7 73.4 78.2 75.0 76.6 76.5Low potency 25.0 17.2 10.9 9.4 7.8 6.3% with SCL 31.3 26.6 21.8 25.0 23.4 23.5avg high potencyavg mediumpotencyavg low potencyavg % with SCLOn average, combining both effect types and both classification categories, the goal of 70-80% of the substances in the medium potency group and 5 -15% of the substances in the lowand high potency group was fulfilled with boundaries of 4 and 400 mg/kg bw/day. However,other combinations of boundaries such as 3 and 300 and 5 to 500 mg/kg bw/day also fulfillthese requirements. Using these boundaries would result in a change of potency group for 10to 14 substances (5 – 7%). Further it could be considered to lower the factor of 100 betweenthe borders to increase the number of substances. For example, using boundaries of 5 to 300mg/kg bw/day would result in 13.9% high potency substances, 15.2% low potency substancesand 71% substances in the medium potency group. Also, the percentages provided in thetables 9 and 10 are calculated not using every modifying factor. Therefore, it can be statedthat the choice of the boundaries is arbitrary. However, based on the available information,the boundaries of 4 to 400 mg/kg bw/day seem to be reasonable.7.372.620.127.49.175.715.224.310.379.99.820.113.977.88.322.214.779.45.920.616.680.03.420.0

Guidance on the Application of the CLP Criteria Version 3.0 - November 20125.1.3 Concentration limits for Category 1 and Category 2 substancesThe generic concentration limit (GCL) from the respective categories will be used formedium potency substances (group 2). As mentioned earlier the GCL is 0.3% forreproductive toxicants Category 1A and 1B and 3.0% for Category 2.Category 1A and 1BDifferent concentration limits have to be used for the different potency groups. Substancesclassified in Category 1 in the low potency group (group 3) can have a SCL above the GCL of0.3%. We propose to use an SCL of 3% which is tenfold of the GCL. A factor of 10 is usedoften in CLP as difference in GCL between hazard categories. This factor is also used in theguidance for setting SCLs for carcinogens. For substances in group 1 (high potency), it isproposed to use a SCL of 0.03%. For extremely potent reproductive toxicants with an ED 10(classification) of more than 10 fold below the boundary limit of 4 mg/kg bw/day it isproposed to use even lower SCLs. For every factor of 10 below the upper limit the SCL isreduced with a factor of 10.Category 2Substances classified in Category 2 in the low potency group (group 3) can have a SCL abovethe GCL of 3%. We propose to use an SCL of 3-10% which is one to 3-fold of the GCL. AnSCL above 10% was considered too high. The upper SCL of 10% can only be used inexceptional cases (NOAEL below 1000 mg/kg bw/day but ED 10 above 1000 mg/kg bw/day).This would account for none of the substances in the database. For high potency substances(group 1), it is proposed to use an SCL of 0.3%. For extremely potent reproductive toxicantswith an ED 10 (classification) of more than 10-fold below the boundary limit of 4 mg/kgbw/day it is proposed to use even lower SCLs. For every factor of 10 below the upper limit,the SCL is reduced by a factor of 10.The resulting SCLs for each potency group are presented in Table 11.Table 11. SCLs for substances in each potency group and classification categoryCategory 1 Category 2Dose SCL Dose SCLGroup 1highpotencyGroup 2mediumpotencyGroup 3lowpotencyED 10(classification)below 4 mg/kgbw/dayED 10 > 4 mg/kgbw/day, and 4 mg/kgbw/day, and < 400mg/kg bw/day3% ED 10(classification)above 400 mg/kgbw/day0.3%(factors of 10lower forextremely potentsubstances)3% (GCL)3-10% AA The limit of 10% may be considered in certain cases, such as for substances with an ED 10value above 1000 mg/kg bw/day and a NOAEL below 1000 mg/kg bw/day.571

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012B For substances with an ED 10 more than 10 fold below 4 mg/kg bw/day, meaning an ED 10 below 0.4 mg/kgbw/day, a 10-fold lower SCL should be used. For even more potent substance the SCL should be lowered with afactor of 10 for every factor of 10 the ED 10 is below 4 mg/kg bw/day.Assigning two SCLs to a substanceA reproductive toxic substance is classified in one category for both effects on developmentand on sexual function and fertility. Within each category effects on development and onsexual function & fertility are considered separately. The potency and resulting concentrationlimits have to be determined separately for the two main types of reproductive toxic effects.In case the potency and resulting specific concentration limits are different for sexualfunction/fertility and development for a substance, the substance needs to be assigned oneSCL for developmental toxicity and another SCL for effects on sexual function and fertility.These concentration limits will in all cases trigger different specifications of the hazardstatements for the two main types of effects, to be applied to mixtures containing thesubstance (see also 3.7.4.1, Annex I, CLP).5.2 Assigning SCLsThe SCL or GCL for each substance can be determined using the final potency group of thesubstance using Table 9.6 ReferencesGemma Janer, Betty C. Hakkert, Aldert H. Piersma, Theo Vermeire and Wout Slob (2007) Aretrospective analysis of the added value of the rat two-generation reproductive toxicity studyversus the rat subchronic toxicity study. Reproductive Toxicology 24,103-123.Schneider K, Oltmans J, van Gelder R and Gebel T (2007) Suitability of the limit dose inevaluating reproductive toxicity of substances and preparations. International Journal ofToxicology, 26: 183-195.Regulation (EC) No 1272/2008 on the classification, labelling and packaging of substancesand mixtures. Official Journal of the European Union L353, 31-12-2008.Regulation (EC) NO 1907/2006 concerning the Registration, Evaluation, Authorisation andRestriction of Chemicals (REACH). Official Journal of the European Union L136, 29-05-2007.R.8 Guidance on information requirements and chemical safety assessment. Chapter R.8:Characterisation of dose [concentration]-response for human health. ECHA, 2008.Commission Working Group on the Classification and Labelling of Dangerous Substances(date unknown) Guidelines for setting specific concentration limits for carcinogens in Annex Ito Directive 67/548/EEC - Inclusion of potency considerations.http://ecb.jrc.ec.europa.eu/documents/Classification-Labelling/GUIDANCE_DOCUMENTS/Guidelines_for_carcinogens.pdfMuller A, Blaude M-N, Ihlemann C, Bjorge C, Ohlsson A and Gebel T (2012) A regulatoryapproach to assess the potency of substances to the reproduction. Regulatory Toxicology andPharmacology 63, 97-105.

Guidance on the Application of the CLP Criteria Version 3.0 - November 2012EUROPEAN CHEMICALS AGENCYANNANKATU 18, P.O. BOX 400,FI-00121 HELSINKI, FINLANDECHA.EUROPA.EU573