Occupational Exposure to Carbon Nanotubes and Nanofibers
Occupational Exposure to Carbon Nanotubes and Nanofibers
Occupational Exposure to Carbon Nanotubes and Nanofibers
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
filter. In the method evaluation, different samplertypes (open-face 25-mm <strong>and</strong> 37-mm cassettes,298 personal cascade impac<strong>to</strong>rs, <strong>and</strong> four pro<strong>to</strong>typeimpac<strong>to</strong>rs) were used <strong>to</strong> collect diesel exhaustaerosol at an express mail facility. The relative st<strong>and</strong>arddeviation (RSD) for the mean EC concentrationwas 5.6% [Birch <strong>and</strong> Cary 1996]. Based on the95% confidence limit (19%; 13 degrees of freedom,n = 14) on the accuracy, the NIOSH accuracy criterion[Kennedy et al. 1995] was fulfilled. Variabilityfor the OC results was higher (RSD = 12.3%),which is <strong>to</strong> be expected when different samplers areused <strong>to</strong> collect aerosols that contain semi-volatile(<strong>and</strong> volatile) components, because these may havea filter face velocity dependence. The method precision(RSD) for triplicate analyses (1.5 cm 2 filterportions) of a 37-mm quartz-fiber filter sample ofDPM was normally better than 5%, <strong>and</strong> often 2%or less [NIOSH 1994a].In the method evaluation, the limit of detection(LOD) was estimated two ways: (1) through analysisof low-level calibration st<strong>and</strong>ards [Birch <strong>and</strong>Cary 1996; NIOSH 1994], <strong>and</strong> (2) through analysisof pre-cleaned media blanks. In the first approach,OC st<strong>and</strong>ard solutions (sucrose <strong>and</strong> ethylene diaminetetraaceticacid [EDTA]) covering a range from0.23 <strong>to</strong> 2.82 µg C (or from 0.15 <strong>to</strong> 1.83 µg C per cm 2of filter) were analyzed. An aliquot (usually 10 µL)of the st<strong>and</strong>ard was applied <strong>to</strong> one end of a 1.5-cm 2rectangular filter portion that was pre-cleaned in thesample oven just before application of the aliquot.The filter portion was pre-cleaned <strong>to</strong> remove anyOC contamination, which can greatly increase theEC LOD when TC results are used for its estimation.After cleaning the filter portion, metal tweezers areused <strong>to</strong> remove from the sample oven the quartzspatula that holds the portion. External <strong>to</strong> the oven,the spatula is held in place by a metal bracket suchthat the st<strong>and</strong>ard can be applied without removingthe filter portion from the spatula. This avoids potentialcontamination from h<strong>and</strong>ling.Results of linear regression of the low-level calibrationdata were used <strong>to</strong> calculate the LOD as 3 σy/m,where σy is the st<strong>and</strong>ard error of the regression <strong>and</strong>m is the slope of the regression line. TC results wereused rather than OC because the pyrolysis correctionmay not account for all of the char formedduring analysis of the st<strong>and</strong>ard (because of lowsample loading <strong>and</strong>/or the position of the aliquotin the laser). If not, a small amount of the OC willbe reported as EC, introducing variability in theOC results <strong>and</strong> increasing the LOD. The LOD estimatedthrough the linear regression results was0.24 µg C per filter portion, or 0.15 µg/cm 2 .A simpler approach for LOD determination isthrough analysis of media blanks. In the methodevaluation, TC results for pre-cleaned, 1.5-cm 2portions of the filter media were used <strong>to</strong> calculatethe LOD estimate. The mean (n = 40) TC blankwas 0.03 ±0.1 µg TC. Thus, the LOD estimated asthree times the st<strong>and</strong>ard deviation for pre-cleanedmedia blanks (3 σ blank) was about 0.3 µg C. Thisresult agrees well with the value (0.24 µg C) estimatedthrough analysis of the st<strong>and</strong>ard solutions.Considering a 960-L air sample collected on a 37-mm filter <strong>and</strong> a 1.5-cm 2 sample portion, this LODtranslates <strong>to</strong> an air concentration of about 2 µg/m 3([0.3 µg TC/1.5 cm 2 ] [8.5 cm 2 ]/0.960 m 3 = 1.78 µg/m 3 ), corresponding <strong>to</strong> the reported upper LOQ ofabout 7 µg/m 3 (LOQ = 3.3 × LOD).As with all analytical methods, the LOD is a varyingnumber. However, the EC LOD (about 2 µg/m 3 ,or an LOQ of 7 µg/m 3 ) reported for NIOSH Method5040 is a high estimate. As discussed in Section6 of the CIB, it was based on analysis of precleanedmedia blanks from different filter lots, overa 6-month period, <strong>and</strong> by different analysts at twodifferent labora<strong>to</strong>ries. Further, variability for theTC results, rather than the EC results, was used <strong>to</strong>estimate the LOD. These combined fac<strong>to</strong>rs gave aconservative (high) estimate of the EC LOD. Moretypical values, under different sampling conditions,are discussed in Section 6.1 of the CIB.C.3 Inter-Labora<strong>to</strong>ryComparisonsWhen results of the initial method evaluationwere published [Birch <strong>and</strong> Cary 1996], an interlabora<strong>to</strong>rycomparison was not possible becauseNIOSH CIB 65 • <strong>Carbon</strong> <strong>Nanotubes</strong> <strong>and</strong> <strong>Nanofibers</strong>151