International Congress on the Deterioration and Conservation of Stone

12 th <strong>International</strong> <strong>Congress</strong> on the Deterioration and
Conservation of Stone
Wednesday 24 October 2012
Poster Presentations—Methods and Materials of
Cleaning, Conservation, Repair and Maintenance
Session XI: 3:00 – 5:00

THE EVALUATION OF NANOSILICA PERFORMANCE FOR
CONSOLIDATION TREATMENT OF AN HIGHLY POROUS CALCARENITE
1 Angela Calia, 1 Maurizio Masieri, 2 Giovanni Baldi, 2 Caterina Mazzotta
1 CNR-IBAM, National Research Council - Institute of Archaeological Heritage.
Campus Universitario Prov.le Lecce - Monteroni73100 Lecce Italy
2 Ce.Ri.Col - Colorobbia Research Centre
Via Pietramarina, 53 50053 SOVIGLIANA, Vinci (FI) - Italy
Abstract
The innovative properties of the nano - materials can have advantageous application
in the restoration and conservation of the cultural heritage with relation to the tailoring
of new products for protection and consolidation of stone. Their potential use in this
field needs to be assessed taking into account specific requirements such as
effectiveness, harmfulness and durability. This paper reports on the experimental
activity concerning the application of a nanosilica based product for the consolidation of
stones having high porosity. The study deals with the assessment of the suitable methods
and amounts of product to be applied and the determination of several basic
requirements in order to evaluate the properties of the treated stone, such as the surface
distribution and the penetration depth within the porosimetric network, the stone surface
colorimetric parameters, the superficial strengthening effects, as well as the behaviour
with respect to the capillary water penetration and vapour permeability.
Keywords: stone consolidation, silica nanoparticles, porous stone
1. Introduction
New products are, increasingly on, available for stone conservation. Thanks to the
advanced material technologies many new materials with innovative properties and new
potential applications, such undoubtedly are the nanostructured materials, have been
realized for applications in the field of the cultural heritage, i.e. protection and
consolidation. The properties of these new advanced materials have been explored as
nanoparticles based single systems (Salvadori and Dei, 2001; Zendri et al., 2007;
Ciliberto et al., 2008; Licciulli et al., 2011) or in addiction to other components. This is
the case of the chemical engineering of PMC (Miliani et al., 2007, Kim et al , 2008) and
polymers-nanoparticles composites, where various kinds of inorganic oxides
nanoparticles are combined with pre-existing products, thus introducing a large
spectrum of new advanced performances (de Ferri et al., 2011; Manoudis et al., 2007).
Other than from their chemical properties, the issues of the products for stone
conservation depends on the support characteristics, thus meaning that treatments might
to be tuned on the different substrates taking into account specific requirements such as
effectiveness, harmfulness and durability with respect to the stones to whom they are
applied.
This paper reports on the experimental activity concerning the application of
nanosilica for stone consolidation. The study moves from a collaboration activity with
the Cericol Research Center, Colorobbia Italia, that is the supplier of the product, within

a research project aimed to study conservation treatments for highly porous stone
materials. A commercial water based product, already experienced for capitals marbles
of the Pisa Tower (Baldi, 2008; Baldi, 2012), was tested on a soft and porous calcarenite
used in the Apulian region (Souther Italy), in order to assess its performances on porous
stones. Generally speaking, soft and porous stones were widely used in the past as
building materials, due to a relative facility of extraction and cutting in spite of their low
durability. The poor resistance to the chemical–physical decay processes make them
particularly affected by decohesion problems, thus demanding consolidation treatments
for restoring physical–mechanical characteristics on their surface and adhesion to the
non-deteriorated support.
The suitable methodology of the treatment and the optimal amounts of the products
to apply to the stone were evaluated by the assessment of the superficial distribution and
penetration depth of nanosilica, as well as by the evaluation of the colour surface
properties of the treated stones. The best treatment that was identified by this
preliminary screening was then evaluated with reference to the strengthening effect on
the stone surface, as well as with respect to the water vapor permeability and capillary
water absorption, by comparing mechanical and hydric parameters before and after the
product’s application.
2. Materials and methods
2.1 The nanosilica product
A nanosilica based product in aqueous medium was used, PARNASOS® ZG00009.
It was formulated by Ce.Ri.Col and is produced by Colorobbia Italia.
The synthesis procedure is based on the hydrolysis in water of tetraethyl orthosilicate
Si(OC2H5)4, following the reaction:
Si(OR)4 + H2O � HO-Si(OR)3 + R-OH
The amount of water and catalyst present are chosen to obtain a complete replacement
of OR groups by OH groups. In addition, a condensation reaction takes part to form a
siloxane [Si-O-Si] bond from partially hydrolyzed molecules.
The sol is then de-stabilized arranging the pH value < 4.0.
The residue is washed, dried at 80°C and re-dispersed in water to a pH value > 9.0.
The stable nanodispersion of rounded silica particles has a mean size of about 30 nm
(PdI 0.40) with a Z-potential of 35 mV.
PARNASOS® ZG00009, has a concentration in nanosilica of 30 % w/w. Density of this
product is 1.20 g/mL and viscosity at 20°C is 12 (mPas/sec).
2.2. The stone
A calcarenite, locally named “pietra gentile” (GS stone), was used for the
application of the nanosilica treatments. This stone is representative of soft and porous
stones, widely used within historical-architectural heritage, as well as in many sites and
archaeological artifacts in the Apulia region (Southern Italy) (Tucci et al, 2008; Sileo,
2012).
It is a very fine calcarenite, white coloured and with a massive structure, made of
fine fossil remains having the average size of 200 microns ca., and lythoclasts within a
micritic groundmass finely mixed with poor microsparitic cement. From a petrographic
point of view it ranges from medium-fine wackestone to packstone (Dunham R.J.,

1962), (Figure 1), due to the irregular and very variable structural characteristics. The
open integral porosity of the stone choose for this work is about 30%, with pore radius
distribution mainly between 4 and 0.5 µm.
Figure 1. GS medium-fine calcarenite [wackestone, thin section photomicrograph, polarised light,
crossed nicols (left); macrophotograph (right)]
2.3 The stone treatments
As it is well known, the treatments have to be tuned with relation to the
characteristics of the stones to be treated. Previous experiences of application of this
product were performed on the weathered marbles of the Pisa tower (Italy), where the
capitals were treated by capillarity. This method is suitable for “ mobile” objects of
limited extension, whilst it would not be proposed as a work site method for large scale
surface interventions. Moreover, in the case of the highly porous stones, the great
absorption of the product involved in the treatment is an aspect that needs to be
evaluated with reference to the sustainability in both economic terms and compatibility
with the original characteristics of the preexisting stones. By consequence, the first
problem of the treatment of the stone under study with the nanosilica was the choice of
the product amount and the application method. Different treatments (A,B,C,D) were
realised, following the application methodology by capillarity and by brush.
After the cutting and cleaning with a soft brush, the samples were washed with
deionized water in order to remove the stone dust and then they were dried in oven at
60°C. The dry weight was assumed when the difference between two consecutive
weight measurements was less than 0.1% of the initial weight of the sample. Before the
treatments, the stones were stabilized at the laboratory controlled conditions (22 ± 2°C,
45 ± 5% R.H.) for 24 hours. After the treatments, the samples were dried until the
constant weight at T = 22 °C, R.U. = 40%.
The following treatments were applied to the stone.
Treatment A. Application by capillarity; moving from the results of the capillary
absorption test, the absorption time of one hour was adopted in order to ensure almost
the material saturation by the solution. The GS stone has high water uptake (400
mg/cm2 at 8 days) and quick over time (85% of the total absorbed water at 1 hour).
Treatment B. Application by brush; several consecutive applications were realised until
the product was refused by the stone specimens.
Treatment C. Application by brush; the product was applied in four steps, with a time
lapse of 10 minutes from each other, until the stone refuse.

Treatment D. Application by brush; half a maximum solution’s amount experienced in
B was applied (100 mg/cm 2 ) by several consecutive applications.
The application details are summarised in Table 1. With reference to the amounts
of the applied solution, it was found that the application of the treatment step by step,
with a time lapse from each other, inhibited the absorption of the solution. In this case it
was noticeably lower than the maximum allowed by consecutive applications by brush.
Table 1. Application details of the treatments
Treatment
Application method
Amount of applied
solution
(mg/cm 2 )
Amount of
nanosilica
A Capillarity
Brush, continuous
280 84
B applications up to the stone
refuse
Brush, four time lapsed
200 60
C steps of appl. up to the stone
refuse
Brush, consecutive
36 11
D applications without stone
refuse
100 30
2.4 Measurements, analyses and tests
Before and after the treatments the following measurements, analyses and tests
were carried out on the stone specimens.
ESEM observations and EDS analyses
Morphological observations by Environmental Scanning Electron Microscopy (ESEM,
Mod. XL30, FEI Company) were performed in order to study the distribution of the
treatments on the stone surface.
Qualitative/quantitative elemental analyses by Energy-Dispersive X-ray spectroscopy (EDS),
allowing traceability of suitable atoms, were carried out on cross sections of the treated
specimens to detect the penetration depth of the nanosilica within the stones. The following
test conditions were adopted: low vacuum mode, pressure of 0.7 Torr, beam accelerating
voltage of 25kV, 100 Lsec acquisition time,100 x 100µm area of each analysis.
The EDS spectra were normalized to the Ca peak., and Si was chosen as the indicator of the
presence of the treatments. Indeed, it is an effective marker for the evaluation of the
penetration depth of Si-based preservative products within pure calcareous stones.
In particular, ESEM observations and EDS analyses were applied to a preliminary
screening for the evaluation of the optimal amounts and methods of the product
application. Moving from these preliminary assessment, the suitable treatment was
characterised by further investigations, as in the following items.
PH test
Phenolphthalein was used as Ph indicator, in order to detect the presence and
distribution of the nanosilica product within the stone. A solution of phenolphthalein at

1% concentration in ethanol was applied to the cross sections of the treated samples. Ph
conditions lower than 8,2% make the solution uncoloured, while Ph conditions higher
than 9.8, as it is in the case of the nanosilica product PARNASOS® ZG00009, lead to a
red colour change.
Color measurements
They were performed in the CIELab space, using a reflectance colorimeter
(Minolta Chroma Meter CR 300, illuminant C). The colour parameters L*a*b* (CIE
1976) were measured on each stone sample (NORMAL Rec. 43/93); ten measurements
were taken on each sample area measuring 5x5 cm.
Porosimetric analyses
They were performed by Mercury Intrusion Porosimetry (Thermo Quest- Pascal
140 and 240)
Abrasion test
The abrasion resistance (UNI EN 14157-2004) was determined on the treated and
untreated face of the same stone samples.
Static contact angle measurements
The determination of the static contact angle on laboratory specimens (UNI
11207:2007) was carried out by means of a Lorenzen and Wettre apparatus (Costech
instrument); 15 measurements were performed on each sample area of 5x5 cm.
Permeability test
The permeability test was carried out on 5 specimens measuring 5x5x1 cm
(NORMAL Rec. 21/85); the results were expressed as the mean value.
Capillary test
The capillary test was performed on 5 specimens measuring 5x5x2 (UNI
10589:2000); the results were expressed as the mean value.
3. Results and discussion
3.1 The suitable application’s method and product amount
Stone treatments might to ensure the homogenous distribution of the products on
the surface of the stones, avoiding accumulation on the surface that hide the original
colour characteristics. Further requirement of the consolidation treatments is a good
penetration within the stone, in order to realize the adhesion of the inconsistent
superficial layers with the unweathered stone beneath.
The morphological observation of the samples involved in the different treatments
revealed that the best result in terms of surface distribution of the nanosilica product was
obtained for the D treatment. By comparing the sample surface of the untreated and
treated samples, it was evident that the treatment D didn’t alter the original grain shaped
morphology of the stone surface (Figures 2, 5). On the contrary, the treatments A, B, C
led to wide product accumulation areas, hiding the original morphology of the stone, as
it can be observed in Figure 3, 4, where micro-cracks are also evident in the nanosilica
coating on the stone surface.
Macroscopically, the accumulation of the product led to a translucent appearance of
the stone surface and/or to the presence of nanosilica white powder. Figure 6 shows the
surface samples of the untreated and treated stone, with reference to each treatment,
observed by the stereomicroscope, using oblique illumination.

Figure 2. Untreated stone surface
Figure 4. Treatment B. Morphology of the
sample surface
Figure 3. Treatment A. Morphology of the
sample surface
Figure 5. Treatment D. Morphology of the
sample surface
Untreated stone Treatment D Treatment A Treatment C Treatment B
Figure 6. Surfaces of the stone after the treatments
Nevertheless, the observed alteration of the chromatic surface properties was not
recorded by the colorimetry. In Table 2 the L * , a * , b * color parameters before and after
the treatments are listed, with the colour variation (∆E). The lowest ∆E resulting from
the application of the treatments A and B contrast with the aspect of the surface that was
clearly observed by naked eye. In this case, the colorimetry seems not to be a suitable
method to record the color properties of the treated surfaces. The explanation would be
found in the presence of the nanosilica on the surface, that realize a glossy layer. The
reflection phenomena induced by the highly reflective glossy surfaces may cause
measurement errors in colorimetric measurements in direct light, since most of the
incident light is reflected out of the field of detection of the optical-fiber cable for
measuring specimen.

Table 2. Color parameters
before treatment after treatment
L * a * b * L * a * a * ∆E *
A 92,67 0,39 6,41 91,00 0,48 8,16 2,42
B 92,75 0,31 5,82 91,71 0,36 6,13 1,09
C 91,49 0,39 8,16 90,41 0,49 8,24 1,09
D 93,07 0,41 5,70 90,07 0,47 7,13 3,33
A large representation of the nanosilica distribution within the stone was given by
the phenolphthalein Ph test (Figure 7). The inhomogeneous penetration of the product
under the surface was also evidenced, following the heterogeneity of the stone structure.
The penetration depth was found to range from 5 to 10 mm under the surface, most
frequently it was up to 7-8 mm.
Figure 7. Depth of penetration
by the phenolphthalein color
test (Treatment A)
Figure 8. Distribution profile of Si within the stone by EDS
analyses with reference to the each treatment.
The EDX profile of the Si content (Figure 8), starting just under the surface (50
microns) up to the constant line content of the untreated stone, is shown in Figure 9 for
the A, B, D treatments. The recorded penetration depth was up to 8 mm. In spite of the
different amounts of the solution that were applied by each treatment, the Si content
within the stone is quite similar for A and D treatments, while lower quantity was
detected in the stone that underwent to the treatment B. These results show that,
irrespectively from the quantity of the solution applied by the different methods, the
depth and distribution of the nanosilica particles was almost the same, thus meaning that
the higher quantity of the solution applied by A and B treatments exceed the penetration
capability of the nanosilica allowed by the stone. According to the ESEM observations,
the treatment D was confirmed as the suitable one, giving a penetration effectiveness
that is comparable with the other methods, in spite of the lower amount solution applied;
it also avoid the accumulation of the nanosilica on the surface that come from the excess
of the solution applied.

Moving from the results obtained by this preliminary screening, further
investigations were carried out on the treatment D.
3.2 Basic properties of the treated stone
Other than good penetration within the stone material and preservation of the
chromatic properties of the stone to be treated, consolidation treatment might ensure the
mechanical strengthening of the consolidated material. At the same time the introduction
of the consolidant has not to alter the water migration, in liquid and vapor phase,
specifically the water that easily can be entrapped within the stone. Therefore the
harmfulness with respect of the hydric behavior of the stone material needs to be
assessed, as well as the cohesion effectiveness.
The hardening effect induced by the application of nanosilica was evidenced by the
increase of 7% of the surface resistance recorded by the abrasion test. In fact, the lenght
of the groove measured on the untreated and treated stone was 42 and 45 mm,
respectively. The cohesion effectiveness is related to the filling of the pores due to the
nanoparticles deposition within the pore network of the stone. It reflects on the variation
of the open integral porosity as well as on the porosimetric distribution. The
porosimetric analyses of the stone specimens in the level from 0 up to 0.5 cm under the
surface recorded the variation of the porosity from 31% to 28% after the treatment.
Figure 10 evidences how the porosimetric changes due to the application of the
nanosilica mainly involved the smallest pores measuring less than 1 micron. The
decrease of this pore fraction account for the permeability reduction of the treated stone,
that was evaluated of 26%. Indeed, the decrease of the permeability after the treatment
was from 200 to 147 g/m²·24h.
Figure 10. Cumulative volume curve at the depth’s interval 0-0.5 cm
(b.t.= before treatment; a.t.=after treatment)
Table 3. Evaluation parameters of the capillary absorption
CA (mg/cm 2 s -1/2 ) Qtf (mg/cm 2 ) ICrel
b.t. 8.67 414
a.t. 7.99 414
(b.t.= before treatment; a.t.=after treatment)
With respect to the stone behavior towards the liquid water, the application of the
treatment doesn’t change the water drop absorption on the stone surface, as it was
0.95

expected; the contact angle measurements failed in the same way on the untreated and
treated samples, due to the rapid water absorption.
The capillary test recorded a lower kinetics of the water uptake in the initial time
(the first hour), while at the end of the test (8 days) the total absorbed water remained
unchanged for the treated and untreated stone. The values of the capillarity absorption
coefficient (A.C.), the total amount of the water absorbed (Qft) and the capillary relative
index (C.I. rel) are summarized in Table 3.
4. Conclusions
The work carried out points out some interesting aspects entailed in the treatments
of porous stones with nanosilica products. Firstly, the method of the application and the
amount of the product to be applied are discriminating parameters with respect to the
superficial distribution of the treatment. The application by capillarity, as well as by
brushing up to the saturation of the stone lead to the deposition of nanosilica on the
surface that alter the original appearance of the stone; moreover this application methods
make the treatments not cost effective, due to the great product consuming related to the
treatment of the porous stones. It was also evidenced that the application by brush by not
consecutive steps quickly lead to stone refuse, strongly limiting the absorption of the
solution by the stone. On the contrary, the different application methods do not seem
influence the penetration capability of the nanosilica particles, in terms of depth and
concentration within the stone, rather suggesting a role played by the concentration of
the solution and/or the stone structure. The penetration depth is quite satisfactory –from
5 to 10 mm, most frequently up to 7-8 mm – depending on the heterogeneous stone
structure. The consolidation effect of the nanosilica treatment lead to the increase of the
mechanical resistance of the stone surface, as it was showed by the abrasion test. The
cohesion power arise from the filling of the smallest pores; with relation to this
mechanism of action it is suitable a previous assessment of the amount of nanosilica that
is introduced within the materials, in order to avoid negative effects on the stone
permeability. Finally, the study carried out evidences, once a time again, the importance
of tuning the treatments on the basis of the specific stone characteristics in order to
ensure the sustainability of the interventions. Sustainability is firstly in terms of
compatibility with the constituent stone of the artifacts; for porous stones, that typically
involve high product consuming, due to their high absorption capability, sustainability is
also in terms of cost effectiveness.
5. References
Baldi G., “Nanostructured oxides: from laboratory research to industrial
applications – Multifunctional behavior of nanostructured surfaces and development of
new industrial products”, COST MP0904 Training School, Genoa 2012.
Baldi G., “Nanotechnology for cultural heritage – Nanostructured materials for
consolidation and pictorial integration in cultural heritage”, Nanotec2008.
Ciliberto E., Condorelli G. G., La Delfa S., and Viscuso E., Nanoparticles of
Sr(OH)2: synthesis in homogeneous phase at low temperature and application for
cultural heritage artifacts, Appl. Phys. A, 92, 2008, 137-141.

de Ferri L., Lottici P. P., Lorenzi A., Montenero A., Salvioli-Mariani E., Study of
silica nanoparticles – polysiloxane hydrophobic treatments for stone-based monument
protection, Jour. Cult. Her., 12, 2011, 356-363.
Kim, Eun Kyung, Jongok Won, Jeong-Jin Kim, Yong Soo Kang, and Sa Dug Kim.
2008.TEOS/GPTMS/silica nanoparticle solutions for conservation of Korean heritage
stones. In Proceedings of the 11th <strong>International</strong> <strong>Congress</strong> on Deterioration and
Conservation of Stone, 15–20 September 2008, Torun´, Poland, ed. J. W. Lukaszewicz
and P. Niemcewicz, 915–23. Torun´, Poland: Nicolaus Copernicus University.
Licciulli A., Calia A., Lettieri M., Diso D., Masieri M., Franza S., Amadelli R.,
Casarano G.: Photocatalytic TiO2 coatings on limestone, Journal of Sol-Gel Science and
Technology DOI: 10.1007/s10971-011-2574-9, 2011.
Manoudis P, Papadopoulou S, Karapanagiotis I, Tsakalof A, Zuburtikudis I.
and Panayiotou C: Polymer-Silica nanoparticles composite films as protective coatings
for stone-based monuments, Journal of Physics: Conference Series 61 (2007) 1361–
1365
Miliani C., Velo-Simpson M. L., Scherer G W.: Particle-modified consolidants: A
study on the effect of particles on solegel properties and consolidation effectiveness,
Jour. Cult. Her., 8, 2007, 1-6.
Salvadori B., and Dei L., Langmuir, 17, 2001, 2371-2374.
Sileo M. 2012 - Individuazione e caratterizzazione geologica, chimicomineralogica
e petrofisica di calcareniti tenere della Puglia e della Basilicata in relazione
alle problematiche di provenienza e conservazione dei Beni Culturali. PhD thesis
Tucci P., Armiento G., Menichini P., Esposito R., I materiali lapidei della facciata
de “Le Monacelle” di Ostuni e le loro antiche cave di appartenenza, in III <strong>International</strong>
Symposium on the Conservation of Monuments in the Mediterranean Basin, V.Fassina,
Ott H., Zezza F. ( Eds), Venezia, 22-25 Giugno 1994.
UNI 11207:2007 Standard, “Cultural heritage - Natural and artificial stones -
Determination of static contact angle on laboratory specimens”.
UNI EN 14157-2004, Natural stones. Determination of abrasion resistance
UNI 10589:2000 “ Cultural Heritage – Natural and artificial stones –
Determination of water absorption by capillarity “
Zendri E., Biscontin G., Nardini I., Riato S.: Characterization and reactivity of
silicatic consolidants, Construction and Building Materials 21 (2007) 1098–1106.

THE PRESERVATION OF SANDSTONE RELIEFS AT THE
ARCHAEOLOGICAL SITE OF TAJÍN, MEXICO, USING COLLOIDAL
SILICA
Dulce Ma. Grimaldi 1 , Nora A. Pérez 1,2 , Jennifer H. Porter 1
1 Coordinación Nacional de Conservación del Patrimonio Cultural, INAH, Ex convento
de Churubusco, Xicoténcatl y General Anaya s/n, Col. San Diego Churubusco, Del.
Coyoacán, CP 04120, México D.F.
2 Instituto de Investigaciones en Materiales, Universidad Nacional Autónoma de México,
Circuito exterior s/n, Ciudad Universitaria, Delegación Coyoacán, CP 04510, México
D.F.
Abstract
The Archaeological Site of Tajín was the biggest and most important
prehispanic city of the north coast of the Mexican Gulf; it reached its peak from the
early 9th to the early 13th century C.E. Currently, it is surrounded by jungle and stands
in the vicinity of areas of petroleum extraction. At this World Site Heritage zone several
constructions have detailed religious and symbolic scenes carved on carbonate-cemented
sandstone which suffers from diverse deterioration, mainly weathering, flaking, and
often extensive loss.
The first plasters applied in the conservation of these reliefs were made of synthetic
polymers or lime without good results. These materials differed in color, texture and
porosity from the original material; they were difficult to apply and often stained the
stone during application and after weathering. After time, they also suffered from
cracking and detachment.
Fills prepared with colloidal silica showed good initial results for the treatment of the
sandstone reliefs submitted to tropical environmental conditions and air pollution.
Colloidal silica-based plasters have been tested for one year on site and have shown
better endurance, stability, and appearance than the previously applied plasters. The
plasters were also tested as protective surface coatings for the sandstone against
environmental deterioration and acidic pollution at the site.
Petrographic studies and nitrogen adsorption-desorption techniques were used to
characterize the plasters in order to evaluate their long term durability and compatibility
with the original sandstone, as well as future alteration.
Keywords: Colloidal silica, sandstone, Tajín reliefs
1. Introduction
Sandstone was extensively used at the Archaeological Site of Tajín, the biggest
and most important prehispanic city of the north coast of the Mexican Gulf (Figure 1).
Here, most of the buildings have a core constructed of stone and earth rubble covered

with close-fitting sandstone flagstones assembled with minimal use of mortar. At this
World Heritage Site, detailed bas-reliefs on tablets, columns, friezes, panels and altars
depict religious and mythological scenes that illustrate the beliefs of its inhabitants.
Tajín reached its peak between the early 9th and early 13th centuries C.E. While
originally related to the important site of Teotihuacan, Tajín was able to survive the fall
of its powerful neighbor, probably due to its own strategic location along the trade
routes of Mesoamerica. Although the collapse of Tajín started around the 11th century
(Soto, 2009), it was not completely abandoned; even today small communities have
survived around the site, and it is still part of the inhabitants’ daily walk to school and
work.
Tajín is undergoing formal preservation of its wall painting and stone elements through
a project carried out by Mexico’s Instituto Nacional de Antropología e Historia (INAH)
Department of Conservation (CNCPC). This project, which started in 2008 and is slated
to last until 2020, includes not only preservation and restoration procedures but also
documentation, maintenance, dissemination, training of local employees and applied
research for conservation issues.
This paper describes the preliminary results of ongoing research for the use of colloidal
silica for the conservation of the sandstone reliefs at Tajín.
2. The sandstone problem at Tajín
The sandstone at Tajín is affected by various types of deterioration, presumably
caused mainly by environmental conditions and intensive weathering. Deterioration is
manifested mainly by exfoliation, detachment and erosion, which have resulted in the
extensive loss of material from both flagstones and relief. The decay also includes
cracking, deformation, efflorescence, discoloration and diverse biological colonization
(Figure 2).

(a
While further research is required to fully understand the causes of deterioration at Tajín,
a number of factors undoubtedly play a role. Variations in the composition of
construction materials may account for some of the observed diversity in deterioration.
Tajín is surrounded by rainforest, with a Senegal-type climate characterized by high heat
and humidity throughout most of the year. Hurricane season runs from June to October,
while the “nortes”, cold fronts with winds and sometimes rain that come from the north,
can occur during the rest of the year. The geologic stratigraphy comprises three levels:
the middle layer is semi-permeable, while the third is impermeable due to the presence
of clay, which means that the large quantities of water which fall during the rainy season
are trapped in the uppermost level. During the 1981-1983 season, a drainage system was
installed to reduce the problem (Bruggermann, 1992), but lack of maintenance of the
system has caused water levels to rise again, favoring mobilization and efflorescence of
salts from sources such as cement from previous interventions, while flagstones suffer
stress from the contraction-expansion of their autochthonous clays during wetting and
drying cycles (Bruggermann, 1992).
Concern towards air pollution and acid rain at this area has favored a daily monitoring of
annual basis. Large-scale and local meteorological conditions in the region of Tajín
favor transport from easterly and northerly and are a source of potential acid precursor
emissions. However, the relationship between transport pathway and precipitation
acidity is still not clear and ongoing research is done (Kahl et al., 2007).
3. Selection of conservation materials
Around 1970´s several panels decorated with reliefs were first treated
(Bruggermann, 1992). At that time, cement plasters were used for structural
consolidation, while plasters for fills and edging were made with lime combined with
synthetic polymers. These materials differed in color, texture and porosity from the
original material; they have often stained the stone during application. After time, they
also suffered from weathering, cracking and detachment.
During 2008, some of the most decayed reliefs were again treated to prevent further
deterioration. The treatment included the edging of flakes to prevent detachment, the
filling of areas of loss and the application of a protective surface coating in extensively
weathered areas. Lime, crushed sandstone, earth and sand were employed as the
conservation material for treatment in 2008-2010. The plasters were evaluated
immediately after application as well as every three months, and showed unsatisfactory
(b

esults for the treatment: edging and fill plasters suffered weathering, cracking and
detachment, while the protective surface coating was aesthetically incompatible with
sandstone, even although pigments and sand were used to match colors and texture as
much as possible.
In 2011 an effort was made to find an alternative to the unsatisfactory materials used in
previous treatments. Colloidal silica was selected as a test binder for plasters, due to its
chemical and aesthetic compatibility with the silica stone system, its low toxicity and
ease of use. It was selected over ethyl silicate-type binders because it does not react
chemically with the stone substrate, resulting in a softer and more reversible plaster
material.
4. Colloidal silica in conservation
Colloidal silica is a suspension with a determined volume fraction of SiO2 spheres
in various solvents. The particle size of the silica colloids ranges from 30 to 100 nm.
Depending on the solvent, temperature and humidity, the drying and sedimentation of
the SiO2 particles differs (Okubo et al. 2008, Okubo et al. 2006). Colloidal silica is used
in many industrial applications such as polishing slurries, catalysts, composite coatings
and adsorbents (Kobayashi et al. 2005).
Colloidal silica has most often been applied in conservation as consolidant for stone,
plasters and paint layers (Biscontin 2002, Chandra & Liping 1999, Malaga et al. 2000,
Mangio & Lind 1997, Moropoulou et al. 2000, Mosquera et al. 2005, Schindler 2005,
Zendri et al. 2006, Lithgow & Stewart 2001), although it also has been used as a
component in grouts (Christaras et al. 2002). Colloidal silica has also been applied to
limestones and lime plasters, sandstone and volcanic tuffs (Elías 2007).
A number of studies describe the use of colloidal silica as a binder for fills and
sacrificial layers, and provided useful information for the tests described here. In
examples which included colloidal silica in sacrificial layers, it was found that the
penetration of the silica and therefore adhesion of the protective layer can be controlled
through silica dilution, and by exploiting differences between the silica solution and
substrate pHs (Kozlowski et al. 1992; Stepien et al. 1993). There is also an interesting
example of colloidal silica used as a binder for sandstone repair mortars and sacrificial
layers, in which the compositions of the mortar and sacrificial layer were based on the
tested properties of the sandstone to be repaired. The sacrificial layer was composed of
ground sandstone and silica. This study noted that treatment with high pH silica
solutions may cause severe color changes in stones containing iron compounds sensitive
to changes in pH (Kuhlenthal et al. 2000, Simon et al. 2006, Snethlage 2000).
Although many are case studies, most published applications of colloidal silica include
laboratory testing of proposed treatment systems, including studies of properties such as
penetration, water vapor permeability, salt resistance and optical effects, in relation to
the properties of the materials they have been used to conserve. Generally results
suggest that colloidal silica has a minimal impact on water vapor permeability, is stable
over time, and improves the weathering resistance of stone. No particular negative
impacts have been stated, although this may be due to lack of reported observations over

time in the field. A few studies offer an evaluation of behavior over time, up to two
years after the application of colloidal silica-based tests or treatments in situ. These
authors generally found the tests to be stable and relatively unaffected by deterioration
(Kuhlenthal et al. 2000, Stepien 1993).
5. Materials preparation and testing methodology
Ludox HS40®, a suspension of 30% wt. colloidal silica in water, was selected as a
binder for test plasters in this research, since it has already been applied in conservation
It has a density of 1.2 g/cm 3 and pH 9.
Several Ludox-based plasters were made using varying proportions of sand, crushed
sandstone and sometimes a local earth. The results for plasters made with a
binder:aggregate ratio of 2:3 and 1:3 by volume are presented here. The proportions of
the aggregates were changed according to the color and texture of the sandstone
substrate to which they were applied. Test plasters were first applied to sandstone
fragments collected around the archaeological site and characterized in the laboratory
during and after application. Plaster mixes with desirable characteristics were then
applied in situ to sandstone blocks for exposure to natural weathering and evaluation
over time.
A year after application of the test plasters in the laboratory and in situ, petrographic
analysis was performed in order to observe and compare the crystallization of the plaster
with the original sandstone. Optical microscopy was used to observe the texture and
cohesion degree of the plasters. To ensure that the plaster does not affect the
physisorption properties of the stone, the nitrogen adsorption-desorption technique was
used.
A desired property in conservation materials is reversibility, and therefore removal tests
of the plasters were made in laboratory samples and on site using wooden sticks, after
previous moistening of the plaster. The cleaned surface was observed in the optic
microscope to determine changes in the surface’s texture. The colloidal silica plasters
were compared with the former lime based plasters.
6. Results
Based on petrographic analysis the Tajín’s sandstone can be classified as a
calcareous litarenite with fragments of sedimentary rocks, quartz and feldspars on a clay
matrix, and partially cemented by calcite. The proportion of quartz and calcite varies
according to the selected stone. Therefore, the inclusion of colloidal silica in the
plaster´s composition is important to assess chemical compatibility with the original.
In the petrographic analysis we could observe that the plasters have good integration
with the stone substrate. Colloidal silica plaster can be applied to form thin layers like a
lime plaster would do, or thick layers, as can be seen in Figure 3. The petrographic
description of the first plaster shows the presence of quartz, feldspar and glass that are
the components of the aggregates, and these materials are in a matrix formed from
quartz microcrystals which is a 39% of the plaster’s composition. The second plaster has
also lithic fragments and pyroxenes that come from the mixed earth, the matrix is

formed from the quartz microcrystals of the colloidal silica, but they are a 15% of the
plaster, accordingly to the plaster’s preparation.
(a
(b
In the optical microscope, the difference between the colloidal silica and lime plasters
was evident (Table 1). The silica-based plasters had a texture similar to the original
stone, mostly due to the particle size of the aggregates. The cohesion of the plaster was
also determined by the size and form of the aggregate particles, since all the water in the
colloidal silica evaporates and only the SiO2 crystals remain, causing cohesion
properties to be determined largely by the arrangement between aggregate particles. The
texture of the lime plasters varied significantly from that of the sandstone, but these
plasters showed very good cohesion, since the lime acts directly as a binder between the
particles.
The nitrogen adsorption-desorption isotherms for the colloidal silica and lime plaster
showed that the materials are mesoporous, the desorption rate was lower for the lime
plaster than for the colloidal silica plaster. The stone presented a type II isotherm with a
type H3 hysteresis loop (Rouquérol 1999), which is also for mesoporous materials.
These results indicates that the physiorption dynamic is similar for the three materials,
but since the desorption rate in the lime plaster is the lowest, it could be expected a
change in the material while it is wet.
The reversibility of the treatment was determined by the cohesion of the plaster and its
adhesion with the stone, since the interaction of the plaster and the stone is mechanical.
The lime effectively bound not only the aggregates, but also the plaster to the stone
making it difficult to remove and leaving remains of lime in the smaller pores, changing
the appearance of the stone. This was not a problem for stones with a higher proportion
of calcite, but for the ones with a higher content of feldspars and quartz, the color
changed dramatically.
The colloidal silica plasters were well attached to the stone, but it was very easy to
remove them, since the interaction was only mechanical and was directly related to the
surface texture and the arrangement of the particles. If the aggregate particles were not
sieved, the plaster was easier to remove since the arrangement was less strong than with
a finer particle size, which had a good arrangement. These plasters did not leave
remains on the stone, at least visible at a microscopic level, and after the removal the
texture was the same as the untreated stone. The difference in the particle size also had
an effect in the plasticity of the material, and therefore, in the handling of the plaster in
the application on the reliefs. The comparison of the evaluated properties of the plasters
is presented in Table 2.
(c

7. Discussion
Aggregate particle size significantly affected the adequate composition and good
handling characteristics of the plasters, as well as their behavior once dried. The
interface between the stone and plasters made with fine aggregates was difficult to
distinguish, even at the microscopic level. Fine grained plasters were easy to apply, even
as thin layers that conformed well to the details of the relief. They also formed stable
plasters, since they had good cohesion which was noticeably influenced the durability of
the material: after one year of weathering in the extreme conditions of Tajín finer
plasters remained unchanged.
Another important factor was the proportion of different aggregates. Plasters made with
higher proportions of sand lost cohesion and showed diminished durability. Higher
proportions of earth caused cracking during the setting of the plaster. Finally, crushed
sandstone provided both plasticity and good color and texture matching.
It was easy to match the color and texture of the silica-based plasters with the sandstone.
In Figure 4(b) several colors were achieved by varying the proportion of aggregates, in
order to obtain the plaster with best aesthetic results. The restoration of the relief at
Figure 4(a) included filling the lost areas and providing preventive care for the rest of
the damaged edges. Other areas of this relief were successfully treated with the same
plaster as a protective surface coating. The colloidal silica plaster proved to be a
compatible material (Figure 4c) that helps to recover and preserve the relief´s value.
(b
Colloidal silica provided a good microcrystalline matrix for aggregate cohesion and
therefore it resembles the original sandstone composition, as it can be seen in
petrography analyses. Reversibility is one of the main desired characteristics that were
achieved using this plaster.
8. Conclusions
Colloidal silica-based plasters and protective surface coatings were evaluated after
one year of exposure to tropical conditions and air pollution in Tajín. The plasters
proved to have good endurance, stability, and appearance characteristics, and laboratory
(a)
(c)

testing suggests that they are more compatible with the sandstone substrate than limebased
materials. They are therefore considered much more successful than the
previously applied lime plasters, and are now considered acceptable for use in future
conservation treatments at Tajín.
The tests conducted in this study and future treatments will continue to be monitored
and evaluated over the remainder of the conservation project at the site, which will end
in 2020. It is hoped that the continued study of these treatments will provide a more
thorough evaluation of the long-term success and viability of colloidal silica-based
plasters in conservation.
9. Acknowledgements
Thanks to the authorities of INAH (CNCPC, Tajín Archaeological Site, and Centro
INAH Veracruz) who encouraged this research on behalf of Tajín´s cultural patrimony.
10. References
Brüggermann, J. 1992. “El juego de pelota”. In Tajín, Delgado, A. L. (ed.) 113-132.
México: Gobierno del Estado de Veracruz.
Chandra, S., Wu, L. 1999. ”Colloidal silica in stone conservation”. Bauinstandsetzen
und Baudenkmalpflege: eine internationale Zeitschrift, 5 (1): 15-28.
Christaras, B., Chatziangelou, M., Dimitriou, A., Lemoni, El., Mariolakos,
I., Foundoulis, J. 2002 “Impermeation of Macedonian tombs in N. Greece using
grouting techniques”. In Protection and conservation of the cultural heritage of the
Mediterranean cities: proceedings of the 5th international symposium on the
conservation of monuments in the Mediterranean Basin, Sevilla, Spain, 5-8 April 2000,
Galán, Emilio and Zezza, Fulvio (ed.) 417-423.
Elías, C. O. 2007. ”Estudio comparative de morteros de hidróxido de calcio y óxido de
sílice: materiales aplicados al tratamiento de lagunas faltantes en objectos elaborados
con toba volcánica de Guadalajara”. Unpublished undergraduate thesis for the Escuela
de Conservación y Restauración de Occidente, Guadalajara, Mexico.
Kahl, J.D.W., Bravo, H., Soto, R., et al. 2007. “Characterization of atmospheric
transport to the El Tajín Archaeological Zone in Veracruz, México”. Atmósfera, 20(004):
359-371.
Kobayashi, M., Juillerat, F., Galletto, P., Bowwnd, P., Borkove, M. 2005. ‘Aggregation
and charging of colloidal silica particles : Effect of particle size’. Langmuir,
21(13):5761-5769
Kozłowski, R., Tokarz, M., Persson, M. 1992. “"GYPSTOP'': a novel protective
treatment”. In Proceedings of the 7th <strong>International</strong> <strong>Congress</strong> on Deterioration and
Conservation of Stone: held in Lisbon, Portugal, 15-18 June 1992,Delgado Rodrigues,
J., Henriques, F., Jeremias, F. (ed.) 1187-1196.

Kühlenthal, M., Kaiser, E., Fischer, H. 2000. “Testergebnisse und verfahreneisen bei der
Einstellung des Reparaturmörtels”. ICOMOS Journals of the German National
Committee, 34: 201-208.
Mangio, R., Lind, A. 1997. “Silica sols (GypStop) for simultaneous desalination and
consolidation of Egyptian limestone at the Karnak Temple in Luxor, Egypt”. In
Proceedings: 4th <strong>International</strong> Symposium on the Conservation of Monuments in the
Mediterranean: new concepts, technologies and materials for the conservation of
historic cities, sites and complexes: Rhodes, 6-11 May 1997. Moropoulou, A.,Zezza,
F., Kollias, E., and Papachristodoulou, I. (ed.) 299-312. Technical Chamber of Greece.
Moropoulou, A., Haralampoulos, G., Tsiourva, T., Theoulakis, P., Koui, M. 2000.
“Long term performance evaluation of consolidation treatments in situ". In La prova del
tempo: verifiche degli interventi per la conservatione del costruito. Atti del convegno di
studi, Bressanone, 27-30 giugno 2000, Scienze e beni culturali,16: 239-255.
Mosquera, M. J., Santos, D.M., Montes, A. 2004. “Producing new stone consolidants for
the conservation of monumental stones”. Materials Issues in Art and Archaeology VII:
Materials Research Society Symposium, held November 30-December 3, 2004, 81-87.
Okubo, T., Kimura K., Tsuchida, A. 2008. ‘Drying dissipative patterns of colloidal
crystals of silica spheres on a cover glass at the regulated temperature and humidity’.
Colloid and Polymer Science, 286 (6-7):621-629.
Okubo, T., Nakagawa, N., Tsuchida A. 2006. ‘Drying dissipative patterns of colloidal
crystals of silica spheres in organic solvents’. Colloid and Polymer Science, 285
(11):1247-1255.
Rouquérol, F., Rouquérol, J., Swing, K. 1999. Adsorption by Powders and Porous
Solids:
Principles, Methodology and Applications. Cornwall: Academic Press.
Schindler, C. 2005. “Investigation of consolidation materials for volcanic tuff”. Biuletyn
informacyjny konserwatorów dzieł sztuki 16(1): 41-47.
Simon, S., Shaer, M., Kaiser, E. 2006. “The salt-laden rock-carved tomb facades
in Petra, Jordan: Scientific investigations”. In Proceedings of the ARCCHIP
Workshop ARIADNE 13: Problems of Salts in Masonry; “SALTeXPERT”
(with the Getty Conservation Institute), November 27–December 1, 2002,
Simon, S., Drdácký, M.(ed.) 5: 341–49. European Research on Cultural
Heritage: State-of-the-Art Studies. Prague: Institute for Theoretical and Applied
Mechanics, Academy of Sciences of the Czech Republic.
Soto, A. P. 2009. El Tajín, Arte y Poder. México: INAH-CONACULTA-UNAM.
Stepien, P., Kozlowski, R., Tokarz, M. 1993. "Gypstop - colloidal silica for protective
coating of porous building materials: practical experience at the Wawel castle, Cracow,

Poland". In Structural repair and maintenance of historical buildings III, 303-310.
Southampton: Computational Mechanics Publications.
Zendri, E., Biscontin, G., Nardini, I., Riato, S. 2006. “Characterization and reactivity of
silicatic consolidants”. Construction and Building Materials, 21 (5): 1098–1106.
List of tables.
Table 1. Properties of the plasters observed in the optical microscope. Cohesion, particle size and
distribution are before removal tests.
Composition Cohesion
2 colloidal silica:
3 aggregates good
1 colloidal silica:
3 aggregates regular
1 lime:
2 aggregates
very good
SiO2
size
Particle
SiO 2 Particle
distribution
Cleaned surface
appearance
homogeneous
and medium
homogeneous
homogeneous original surface texture
and small homogeneous original surface texture
abrasion on surface
and surface holes
filled with lime
Table 2. Comparison of lime plasters in respect to colloidal silica plasters.
Characteristic Lime plasters Colloidal silica plasters
Color and Always whitish compared to Easy to match when mixed with sand,
texture sandstone. Pigments were used to
match color, something that can
crushed sandstone and local earth.
cause
changes
undesirable long term
Reversibility Not so easy to remove by Easy to remove by mechanical means
mechanical means. Leaves white after application of water on it. Leaves
color residues difficult to remove as no color residues. The only change in
lime penetrates in the sandstone color of the original surface is due to
pores.
the removal of surface patina.
Ease for Easy to manipulate and store. Requires familiarity with the material
application
as small amount of plaster can be
prepared at a time. Dries fast.
Endurance It cannot endure the constant Good endurance. It can fail if too much
movement of sandstone at the sand is added or if large aggregate is
tropical environment so it finally
detaches.
used.
Cracking after Requires setting under controlled Good. High earth content causes
application conditions as can easily crack under
tropical environment.
cracking.
List of captions.
Figure 1. Localization of Tajín Archaeological Site in Veracruz, Mexico.
Figure 2. (a) Pyramid of the Niches, best known building at Tajín, (b) Exfoliation, flaking, and
weathering at one niche. Courtesy of INAH.

Figure 3. Petrographic images of the plasters on the laboratory probes. 2.5X, 10X objective. The
plasters are marked with a rectangle. a) Colloidal silica plaster with proportion 2:3, b) Colloidal
silica plaster with proportion 1:3, c) Lime plaster with proportion 1:2. Courtesy of INAH.
Figure 4. On site application of colloidal silica at the central south panel from the South Ballgame.
(a) Upper frame with the depiction of a dual body related to Quetzalcoatl. On the left edge the
treated area. (b) Color and texture testing for filler material with colloidal silica. (c) Same area
after treatment. Courtesy of INAH.

THE PERFORMANCE OF AN INDUCED CALCIUM OXALATE
SURFACE ON GLOBIGERINA LIMESTONE
Tabitha Dreyfuss (tabitha.dreyfuss@gov.mt) & JoAnn Cassar (joann.cassar@um.edu.mt)
Department of the Built Heritage, Faculty for the Built Environment, University of
Malta, Malta
Abstract
This paper contains the second set of results obtained from a research programme
that consisted of an induced surface conversion of weathered Globigerina Limestone
from calcium carbonate to calcium oxalate and the prospects of it being used for the
conservation of this stone. The first part of the research confirmed the formation of
calcium oxalate even in salt (sodium chloride) contaminated samples. The second phase
of the research showed that the production of calcium oxalate resulted in the formation
of a more compact, more cohesive, and harder surface when compared to untreated
samples, which showed no colour change, retained the original hydrophilic and wetting
properties of the stone and also increased the stone’s resistance to acid attack (acetic and
hydrochloric acid), and to salt weathering (using sodium sulphate).
Keywords: oxalates, consolidation, protection, Globigerina Limestone
1.0 Introduction/ background
It is a long-standing and ever-present problem that salt-contaminated porous
limestones are extremely difficult to treat successfully. In Malta, this is an especially
serious problem since the locally available stone is extremely porous and is usually
heavily contaminated with marine salt (sodium chloride), and desalination is often not
feasible. This stone, used for the great majority of the Island’s historic buildings,
essentially consists of fine-grained limestone with few to abundant fossils including
planktonic and benthonic foraminifera especially Globigerinae which is from where it
gets its name. Globigerina Limestone is a very porous stone, with a porosity of up to 41%
(Cassar 2002). The pores are mainly channel type. This allows suction of water by
capillary action through the pore structure. Globigerina Limestone used for building in
the Maltese Islands is described by quarry workers to occur as two types: the more
durable “franka” limestone which usually weathers well and the “soll” limestone which
deteriorates badly, even in the same environment. It is believed that the majority of
historical monuments were built with the “franka” type. A research programme, which
commenced in 2004, consisted of an induced surface conversion of weathered
Globigerina Limestone from calcium carbonate to calcium oxalate, with the aim of
using this conversion for the conservation of this stone. The first part of the research
concerned the verification of the surface conversion, and the determination of its extent,
using X-Ray Diffraction. The results obtained confirmed the formation of calcium
oxalate in all cases, and its extent, relative to the stone’s surface texture and its salt
(sodium chloride) content (Mifsud et al. 2006). Whewellite, as opposed to weddellite,
was formed in all cases. The treated non-desalinated samples were found to form more

whewellite than desalinated samples. This was attributed, at least in part, to the
increased surface area that the non-desalinated samples possessed which allowed for
more contact with the ammonium oxalate poultice and thus resulted in more calcium
oxalate being formed. This paper covers the second phase of the research which was
aimed at evaluating the performance of the induced calcium oxalate surface.
2.0 Samples and treatment
The range of samples included fresh quarry samples, artificially weathered quarry
samples (using sodium sulphate for the samples that were to be desalinated and sodium
chloride for those that were not to be desalinated) and naturally weathered samples. All
samples were of the “franka” Globigerina Limestone type as indicated by quarry
workers and confirmed through the insoluble residue test (Cassar 2002). These consisted
of cubes measuring 5cm x 5cm x 5cm. Three samples were prepared for each of the tests.
From each group of samples, both desalinated and non-desalinated samples were studied.
The samples were subjected to a 5% ammonium oxalate poultice for 5 hours, after
which testing took place on both treated and untreated samples.
3.0 Testing methodology
Scanning Electron Microscopy (SEM) was carried out using a LEO 1430, Oxford
Link microscope, to look at the surface morphology and to make direct comparisons
between treated and untreated samples. Any consolidation that may have developed as a
result of ammonium oxalate treatment was evaluated through the Mohs hardness test in
accordance with EN 101:1991 and the Tape test. The tape test which concerns the
investigation of the superficial surface cohesion of the stone was carried out as
explained by Croveri (Croveri 2004). The imparted protection from acid attack of the
treated samples was examined through acid resistance tests using acetic acid at
concentrations of 5% and 10%, and hydrochloric acid at concentrations of 0.5% and 2%
(Matteini et al. 1994). The depth of calcium oxalate formation, and its distribution in the
treated samples, was studied through acid etching following immersion in a 5% acetic
acid solution and then drying. The resistance of the samples to salt crystallisation was
examined by means of artificial weathering tests using sodium sulphate in accordance
with EN 12370:2000, while the hydrophilic and wetting properties were investigated by
means of water absorption tests through capillarity, as per NORMAL 11/85. The
potential change in the appearance of the limestone following ammonium oxalate
treatment was examined by visual inspection and direct comparison of treated and
untreated samples within each group, and their comparison to a Munsell chart.
4.0 Results
4.1 Visual analysis for colour change
Following ammonium oxalate treatment and drying, the colour of the treated
samples was not visibly different, to the naked eye, from the untreated samples. For
each sample type, treated and untreated stones were directly compared to a Munsell
colour chart and in all instances, treated and untreated samples corresponded to one
specific colour only. The quarry non-desalinated samples corresponded to 5Y8/1, and all
the other samples corresponded to 5Y8/2. Therefore, although different sample groups
differed in colour from each other, within each group, treated and untreated samples
were the same colour.

4.2 Tape test
The results obtained from the Tape test are illustrated in Fig 1. This shows that, in
all cases though to varying degrees, the untreated samples left more material on the back
of the tape than the samples that had been treated.
TAPE TEST
16.00
Fig 1. Results of the Tape test
14.00
12.00
10.00
8.00
6.00
% MASS INCREASE
4.00
2.00
0.00
ARTIFICIALLY
ARTIFICIALLY
ARTIFICIALLY
WEATHERED
ARTIFICIALLY
WEATHERED
NATURALLY
NATURALLY
NATURALLY
WEATHERED
NATURALLY
WEATHERED
QUARRY NON-
DESALINATED
TREATED
QUARRY NON-
DESALINATED
UNTREATED
QUARRY
QUARRY
WEATHERED NON-
DESALINATED
WEATHERED NON-
DESALINATED
WEATHERED NON-
DESALINATED
WEATHERED NON-
DESALINATED
DESALINATED
TREATED
DESALINATED
UNTREATED
TREATED
UNTREATED
DESALINATED
TREATED
DESALINATED
UNTREATED
TREATED
UNTREATED
DESALINATED
TREATED
DESALINATED
UNTREATED

4.3 Mohs hardness test
The results from the Mohs hardness test showed that for all of the samples, the
treated stones exhibited an increased scratch resistance when compared to the untreated
ones. The untreated samples had a scratch resistance that varied from 2 (gypsum) to 3
(calcite) whereas in all cases, the treated samples had a scratch resistance of 4
(fluorspar), which is the hardness of calcium oxalate.
4.4 Scanning Electron Microscopy (SEM)
The treated and untreated samples of the same type were examined and
photographed at identical magnifications using the Scanning Electron Microscope. This
facilitated direct comparisons. Magnifications ranged from x500 up to x25K. In general,
the untreated samples showed a surface with loose, individual rounded grains. These
rounded crystals (A), probably calcite, may be observed in Fig 2. In the treated samples,
these rounded grains are present as part of a larger mass/ cluster (B in Fig 3). In addition,
in the treated samples, elongated/ extended grains were observed; these were not
observed in any of the untreated samples. These elongated grains, which could be
whewellite, are sometimes seen to be linking grains together (Fig 3). Fig 3 also shows
arm-like formations (C) branching between different parts of the treated surface. These
again are not seen in the untreated samples.
Fig 2. Naturally weathered non-desalinated sample untreated (left). Fig 3. Naturally weathered
non-desalinated sample treated (right)
4.5 Water absorption through capillarity
The results for the water absorption test through capillarity are expressed
graphically in Fig 4. All the graphs are similar and are seen to contain two portions, an
initial rapid amount of water absorbed per unit area in a given amount of time (within
the first 1 hour 15 minutes), and the later, slower absorption of water following that
(between the first 1 hour 15 minutes and one week).

Fig 4. Water absorption through capillarity for all the samples

4.6 Sodium sulphate crystallisation test
The results for the sodium sulphate crystallisation test showed that in all cases, the
treated samples had a greater resistance to damage than the untreated samples. This was
evaluated through the weighing of material lost. Following the salt weathering tests, the
surfaces of the samples were visually observed. While the untreated samples had a
powdery rough texture, the treated samples all had a smoother surface with less
deterioration visible.
4.7 Acid resistance tests
The resistance to acid attack was evaluated through the determination of the
percentage of mass lost in the samples which were placed in contact with acetic and
hydrochloric acids at different concentrations. The percentage mass lost in the samples
after 5 minutes was calculated. In all instances, the samples that were treated with
ammonium oxalate showed more resistance to acid attack. This was seen as a smaller
percentage of mass of sample lost in the treated samples than in the untreated samples of
the same type.
4.8 Depth of oxalate formation
The unexpected whitening effect produced when the treated samples were placed in
contact with a 5% acetic acid solution was used to determine the depth of calcium
oxalate formed, as this appeared to outline the areas which were treated and
distinguished them from untreated areas. The change in colour from yellowish to white
in the top faces and the top part of the sides was seen in all of the treated samples. None
of the untreated samples became whiter on contact with the acid. The average depth of
the white part and the maximum depth observed on the sides of the treated samples are
shown in Fig 5. These depths may not necessarily be representative of the depths
achieved throughout the sample internally, but are a representation of the depths
achieved on the perimeter of the sample. The samples were treated on the top face only,
but migration into the stone could be easier on the external perimeter than internally.

DEPTH OF OXALATE FORMATION
average depth (mm) maximum depth observed (mm)
Fig 5. Results of the depth of oxalate formation
12.0
10.0
8.0
5.0 Discussion
The lack of a visible colour change in the samples of Globigerina Limestone
following ammonium oxalate treatment is an encouraging sign. It also suggests that little
or no free iron was present, at least in these particular samples that could be mobilised
by the treatment. Further testing on other samples is needed to confirm that this is
generally the case.
6.0
mm
4.0
2.0
0.0
ARTIFICIALLY WEATHERED NON-
DESALINATED TREATED
ARTIFICIALLY WEATHERED DESALINATED
TREATED
NATURALLY WEATHERED NON-
DESALINATED TREATED
QUARRY DESALINATED TREATED QUARRY NON-DESALINATED TREATED NATURALLY WEATHERED DESALINATED
TREATED

A consolidating action, at least on the surface of the samples, as a result of
ammonium oxalate treatment has been confirmed by the fact that less material came
away during the Tape test in all cases of the treated samples. This conclusion is
supported by the consistent increase in hardness to grade 4 (fluorspar) in all of the
treated samples. Also, the SEM image comparisons showed that the surfaces of the
untreated and treated samples differed in their surface morphology. In general, the
treated samples showed a surface that was more compact, with a reduction in loose
surface grains visible. This observation, together with the formation of clusters of grains
and the bridging between clusters observed in the treated samples, suggests that a
consolidation action has in fact occurred. In addition, on visual observation and on
handling of the samples, all the treated samples had a smoother texture with a smaller
amount of loose grains present on the surface. Even though a change in surface porosity
and pore size probably takes place with ammonium oxalate treatment, the hydrophilic
properties of the stone, and thus the water absorption patterns of the stone, are seen to be
preserved in all cases as seen in the water absorption tests. The reduced loss in material
following sodium sulphate weathering of the treated samples, when compared with the
untreated samples, also signifies that protection from deterioration by salt crystallisation
takes place following treatment. The greater damage occurring in the untreated samples
confirms this. Resistance to acid attack in the treated samples was seen as a smaller
percentage of mass of sample lost than in the treated samples. The amount of protection
from acid attack achieved is directly comparable to the amount of calcium oxalate
formed, which was obtained through XRD analyses (Mifsud and Cassar 2006).
The whitening observed in the treated samples only during testing with the 5%
acetic acid solution may be explained by the fact that this acid is corrosive to calcium
carbonate, while calcium oxalate remains undamaged. The calcium carbonate, together
with its colour impurities, thus erodes, leaving the calcium oxalate which is white in
colour, behind. This outcome, which was used to determine the depth of calcium oxalate
formation in the treated samples, showed that in all instances the samples that were seen
to have the greatest averagedepth of formation of calcium oxalate were again the nondesalinated
(sodium chloride) samples. The whiter surface observed in the treated nondesalinated
samples suggests that the formation of calcium oxalate in these samples is
more uniform and better distributed than in the treated desalinated samples. This is in
keeping with the results obtained from XRD analysis, where the treated non-desalinated
samples were seen to have developed more calcium oxalate than the treated desalinated
samples of the same type. This has already been explained as being possibly due to the
increased surface area that the non-desalinated samples possessed, and which the
desalinated samples lost during the desalination process (Mifsud and Cassar 2006). The
deeper calcium oxalate formation in the non-desalinated samples also suggests that
ammonium oxalate treatment in non-desalinated “franka” results in a deeper penetration
of the treatment. This observation merits further investigations.
6.0 Conclusions
These promising results present ammonium oxalate treatment of “franka”
Globigerina Limestone in a positive light. This treatment results in some surface
consolidation which is manifest as a smooth surface texture with a reduced amount of
loose surface material as well as an increased hardness and scratch resistance of the
surface of “franka” limestone, an increased resistance to salt crystallisation and to acid

attack. Additionally, it has been seen that the hydrophilic and wetting properties of the
treated stone are not disturbed, and no colour changes are observed. The results have
also shown that the non-desalinated samples developed a deeper average formation of
calcium oxalate than the desalinated samples. Further research here is being planned and
is also to include the evaluation of the porosity and pore size distribution of the surfaces
of treated and untreated samples. This will allow for the better understanding of the
differences in behaviour of treated samples when compared with untreated ones. The
study of the use of different concentrations of ammonium oxalate, as well as different
duration times and different modes of application, together with an evaluation of the
depth of calcium oxalate formed inside the stone, may lead to an optimum method for
treating exposed naturally weathered “franka” Globigerina Limestone. The practical in
situ application of the treatment to vertical surfaces, such as those found in buildings is
also to be researched, as well as the ongoing monitoring of treated monuments and
buildings. In practical terms, it can in fact be concluded that this treatment can
potentially be used in the field of conservation as a protective treatment, possibly with
some consolidating properties, on historic buildings and monuments built with this stone
in the Maltese Islands.
7.0 References
1. Cassar, J., 2002. Deterioration of the Globigerina Limestone of the Maltese
Islands. In: Siegesmund, S., Weiss, T., and Vollbrecht, A. (eds). Natural Stone,
Weathering Phenomena, Conservation Strategies and Case Studies. 205: 33-49.
Geological Society London, London.
2. Croveri, P., 2004. Metodologie di consolidamento di materiali lapidei nell’area
Mediterraneo: La Globigerina Limestone maltese – degrado e consolidamento.
Dottorato di Ricerca in Scienza per la Conservazione dei Beni Culturali XVII
cicolo, Universita degli Studi di Firenze. Unpublished.
3. EN 101:1991: Ceramic floor and wall tiles. Method for determination of
scratch hardness of surface according to Mohs.
4. EN 12370:2000: Natural stone test methods – Determination of resistance to
salt crystallisation.
5. Matteini, M., Moles, A., Giovannoni, S., 1994. Calcium oxalate as a protective
mineral system for wall paintings: methodology and analyses. In: 3rd
<strong>International</strong> Symposium on the Conservation of Monuments in the
Mediterranean Basin, Venice, Italy.
6. NORMAL 11/85, 1986: Assorbimento d’acqua per capillarita. Coefficiente di
assorbimento capillare. CNR – ICR.
7. Mifsud, T. & Cassar, J. 2006. The treatment of weathered Globigerina
Limestone: the surface conversion of calcium carbonate to calcium oxalate. In:
Heritage, Weathering & Conservation – Fort, Alavrez de Buergo, Gomez-
Heras & Vazquez-Calvo (eds). Taylor & Francis Group, London.

DURABILITY OF CONSOLIDATED POROUS LIMESTONES, A
LABORATORY TESTING APPROACH
Z. Pápay 1 , Á. Török 1
1 Budapest University of Technology and Economics, Department of Construction
Materials and Engineering Geology, Budapest, Hungary, torokakos@mail.bme.hu
Abstract
Three types of porous limestone were tested under laboratory conditions to assess the
durability of various consolidation trials. Fine-, medium and coarse-grained types from
Sóskút quarry were used. Cylindrical samples were consolidated by saturation. In the
tests three consolidants were applied: one type of silica-acid-ester, an aliphatic uretan
resin and a polymethyl methacrylate. The durability was tested on consolidated
cylindrical specimens by accelerated weathering tests in the form of freeze-thaw cycles.
Physical parameters such as density, total porosity, pore-size distribution, ultrasonic
pulse velocity, Duroskop rebound values and indirect tensile strength tests (Brazilian
test) were measured on natural, consolidated, and freeze-thaw affected consolidated
samples. The durability was calculated by comparing test results of consolidated and
freeze-thaw subjected consolidated samples. The indirect tensile strength tests were
proved to model the changes in strength. Our analyses have shown that the tested
consolidants have different penetration depth and different effect on strength of porous
limestone. The durability against freeze-thaw of consolidated limestone types depends
on micro-fabric (pore-size distribution) but cannot be directly assessed from the indirect
tensile strength of the test specimens.
Keywords: porous limestone, stone consolidants, durability, tensile strength
1. Introduction
Porous limestones were commonly used as dimension stone in the 19 th century in
Central Europe especially in Austria and Hungary. Famous monuments such as St.
Stephan’s Dom in Vienna or the Parliament building and the Citadel in Budapest were
constructed from these types of stones (Török et al. 2004). This type of limestone is
very sensitive for weathering processes, which affect not only the aesthetic appearance
of stones but also cause structural damage (Török 2002). Stone consolidation is aimed
to slow down the rapid deterioration and/or strengthen the already weathered stones.
Previous consolidation trials include testing of various carbonate stones (Wheeler et al.
2000, Alvarez de Buergo & Fort 2002, Lukaszewicz 2004, Ahmed et al. 2006, Ferrera
Pinto & Delgado Rodriguez 2012). In this study three different consolidants (silicic acid
ester, aliphatic uretan resin, Paraloid B72) were tested on three porous limestone types
under laboratory conditions. The tests aimed to clarify how the different consolidating
agents change the properties of porous limestone and how modify the durability of
stones. Pápay& Török 2008 showed the first part of the experiments, which examined
the effect of consolidation on air dry samples. The density, ultrasonic pulse velocity and
1

tensile strength were measured on the three limestone types prior and after the
consolidation in order to understand the physical changes caused by the consolidants.
Pápay & Török 2008 found that the treatment by silicic acid ester had the best
efficiency.
2. Materials and Methods
The Miocene limestones from Sóskút were used in the tests, which are also known
as oolitic limestone. From fine-, medium- and coarse-grained limestone blocks
cylindrical specimens were drilled with diameter and height of 5 cm. The experiments
were carried out according to the recommendations of the previous Hungarian standards
and new EU norms at Rock analytical laboratory of Budapest University of Technology
and Economics, Department of Construction Materials and Engineering Geology.
The cylindrical specimens were divided into analytical groups on the basis of nondestructive
testing methods such as density (MSZ EN 1936:2000), ultrasonic pulse
velocity (MSZ EN 14579:2005) according to MSZ 18282-4. In each testing group an
equalized set of samples were grouped, namely to minimize the standard deviation of
fabric related variations, such as density and ultrasonic pulse velocity. The tested
consolidants and their properties are listed in Table 1. The specimens were treated under
laboratory conditions at normal atmospheric pressure. The apparent density was
determined by water saturation under atmospheric pressure. The tensile strength was
determined by indirect tensile test (Brazilian test). In the course of Brazilian tensile
strength test (MSZ 18285-2:1979), the specimens are loaded by parallel plates. Freezethaw
cycles were carried out according to previous Hungarian Standard 18289/2-78
therefore specimens were saturated fully in water and frozen at -20 ˚C for 6 hours. The
freeze-thaw cycles were continued until the first cracks appeared on specimens. The
tested consolidants and their properties are detailed in Table 1.
Table 1: Properties of the applied consolidants
Consolidant Diluting agent
2
Effective
substance
Density
[g/m 3 ]
Silicic acid ester ready to use ca. 20 m% 0,79
Aliphatic uretan
resin
white spirit 50 m% 0,93
Paraloid B72 nitro-thinner 4 m% 0,85
Three different porous limestone types were used for the tests: fine-, medium- and
coarse-grained ones. The fine grained variety has a microfabric of pelloidal wackestone.
It contains micritic peloids and some forams. Intergranular pores dominate however
small amount of intragranular pores also occur. The matrix is characterized by micritic
calcite with patches of microsparitic calcite. Majority of pores are within the range of

0,1 µm and 1 µm, but pores of 10 µm in size were also detected (Fig 1a). The water
absorption rate is 16 m%; while the measured total porosity is 24 m%. The mediumgrained
limestone is characterised by well to moderately rounded calcitic ooids, microoncoids
and visible but evenly scattered small pores (Fig 1b). Its microfabric is oolitic
grainstone with few sand sized quartz grains in the nuclei of ooids (Török 2002). The
coarse-grained limestone has a bioclastic ooidal grainstone micro-fabric (Török et al.
2007). The most common bioclasts are gastropods (Cerithuim sp.). Large mouldic pores
(up to 1 cm-size) are irregularly scattered within the stone (Fig. 1c.).
Figure 1: Fine-grained (a), medium-grained (b) and coarse-grained (c) limestone
3. Changes in physical properties
The measured parameters of treated and untreated test groups are listed in Table 3
for fine-grained, in Table 4 for medium-grained and in Table 5 for coarse-grained
limestone. Tables show the average values of 3 or 4 specimens and the standard
deviations are also listed in brackets.
Ultrasonic pulse velocity data show that the ultrasonic sound pulse velocity
increased due to treatment and decreased after freeze-thaw cycles with no exception; for
all the three limestones and for the three consolidating agents. For fine-grained
limestone there was no signifanct change in ultrasonic pulse velocity. The largest
decrease in ultrasonic pulse velocity was measured on silicic acid ester treated coarse
grained specimens (32,09%). Nearly the same change was observed for silicic acid ester
and Paraloid B72 treated medium-grained test groups (-21,25% and -25,57%). Tensile
strength values show the same trend as ultrasonic pulse velocity values for mediumgrained
and coarse-grained limestone. The change in tensile strength of treated air dry
vs. treated frozen samples was the greatest for Paraloid B72 treated medium-grained
limestone and silicic acid ester treated coarse-grained limestone (-57,9% and -58,1%).
For silicic acid ester treated medium-grained and for Paralod B72 treated coarse-grained
limestone 34,1% and 39,5% decrease were observed. Aliphatic uretan resin caused less
decrease 9,4 % for medium-grained and 6,25 % for coarse-grained limestone. Untreated
samples show no damage in its structure, the tensile strength values are nearly the same.
For fine-grained limestone increase in tensile strength was observed for frozen Paraloid
B72 treated specimens (26,9%). Smaller change was measured for silicic acid ester
(+7,6%) and aliphatic uretan resin (-10,2%) treated test groups. The changes in tensile
strength do not correspond to surface strength changes. There was no significant change
3

in surface strength values except medium grained Paraloid B72 treated, coarse-grained
untreated and silicic acid ester treated samples (-20% and -22,2%).
Before performance of freeze-thaw cycle the water absorption rate was measured for
the three consolidating agents and for the reference test group. Fig. 2, Fig. 3 and Fig. 4
show the result of water absorption in 20 days period. The water absorption rates and
their standard deviations are listed in Table 6.
Table 3: Physical properties of fine-grained limestone (ρ: density before/after treatment; v:
ultrasonic pulse velocity before/after treatment, standard deviations of strength tests are given in
brackets)
Test
group
untreated
silicic
acid ester
Paraloid
B72
aliphatic
uretan
resin
ρ
[kg/m 3 ]
1848/
2060
2004/
2060
1888/
1991
1924/
1923
v
[km/s]
2.27/
2.24
2.70/
2.59
2.58/
2.39
2.55/
2.32
Tensile
strength
[MPa]
2.09
(0.37)
3.43
(1.25)
2.68
(0.52)
2.92
(0.29)
4
Tensile
strength
after
frost
[MPa]
1.92
(0.57)
3,69
(1.03)
3,4
(1.20)
2,64
(0.12)
Surface
strength
[-]
Surface
strength
after
frost [-]
24 22
24 24
25 26
29 29
Table 4: Physical properties of medium-grained limestone (ρ: density before/after treatment; v:
ultrasonic pulse velocity before/after treatment)
Test
group
untreated
silicic
acid ester
Paraloid
B72
aliphatic
uretan
resin
ρ
[kg/m 3 ]
1684/
1705
1720/
1717
1649/
1655
1725/
1722
v
[km/s]
2.32/
2.10
2.40/
1.89
2.19/
1.63
2.54/
2.28
Tensile
strength
[MPa]
0.91
(0.13)
1.32
(0.13)
0.95
(0.18)
1.7
(0.15)
Tensile
strength
after
frost
[MPa]
0.90
(0.17)
0.87
(0.26)
0.40
(0.10)
1.54
(0.14)
Surface
strength
[-]
Surface
strength
after
frost [-]
5 5
6 6
10 8
12 11

Table 5: Physical properties of coarse-grained limestone (ρ: density before/after treatment; v:
ultrasonic pulse velocity before/after treatment)
Test
group
untreated
silicic
acid ester
Paraloid
B72
aliphatic
uretan
resin
ρ
[kg/m 3 ]
1568/
1594
1596/
1602
1609/
1600
1590/
1607
v
[km/s]
1,95/
1,84
2,15/
1,46
2,12/
1,73
2,24/
2,22
Tensile
strength
[MPa]
0.53
(0.15)
0.74
(0.12)
0.81
(0.16)
1.60
(0.13)
Fig. 2 Water absorption of fine-grained limestone
5
Tensile
strength
after
frost
[MPa]
0.66
(0.11)
0.31
(0.02)
0.49
(0.22)
1.50
(0.21)
Surface
strength
[-]
Surface
strength
after
frost [-]
9 7
9 7
9 10
13 13

Fig. 3 Water absorption of medium-grained limestone
Fig. 4 Water absorption of coarse-grained limestone
6

Table 6 Water absorption rate of untreated and (treated) samples
Limestone air dry silicic acid Paralloid B72 aliphatic
type
ester
uretan resin
fine-grained 9.79 (3.79) 7.79 (2.88) 10.24 (4.39) 6.7 (1.38)
mediumgrained
15.46 (1.51) 15.31 (0.96) 19.19 (0.21) 5.9 (0.48)
coarse grained 16.93 (1.35) 18.94 (0.34) 18.93 (0.89) 6.54 (2.24)
4. Conclusions
The freeze-thaw durability tests of untreated and consolidated limestones have shown
that frost resistance of consolidated medium- and coarse-grained limestone samples
became smaller. It is in good agreement with the measured increase in water absorption.
It is difficult to find a clear and similar trend for the fine-grained limestone. The
strength and porosity of treated vs. non-consolidated specimens show high scatter in
data, although most consolidated samples have higher strength and less porosity, than
that of the natural fine-grained limestone (Table 3 and Table 6). From the nondestructive
tests the Duroscope rebound values are less representative, while the
ultrasonic pulse velocities indicate better the effect of consolidation. The treated
samples have higher velocities compared to the non treated ones.
5. References
Ahmed, H., Török Á., Lőcsei J. 2006. Performance of some commercial stone
consolidating agents on porous limestones from Egypt. In: Fort, R, Alvarez de
Buego M., Gomez-Heras M. & Vazquez-Calvo C. (eds): Heritage Weathering
and Conservation, Taylor & Francis/Balkema, London. Vol. II, 735-740.
Alvarez de Buergo, M., Fort, R. 2002. Characterizing the construction materials of a
historic building and evaluating possible preservation treatments for restoration
purposes. In: Siegesmund, S., Weiss, T., S., Vollbrecht, A (Eds.), Natural Stones,
Weathering Phenomena, Conservation Strategies and Case Studies. Geological
Society, London, Special Publications 205: 241-254.
Ferreira Pinto A. P, Delgado Rodrigues, J. 2012. Consolidation of carbonate stones:
Influence of treatment procedures on the strengthening action of consolidants.
Journal of Cultural Heritage, 13, 2, 154–166.
Lukaszewicz, J.W. 2004. The efficiency of the application of tetraethoxysilane in the
conservation of stone monuments. In: Kwiatkowski, D. & Löfvendal, R. (Eds.):
Proceedings of the 10th <strong>International</strong> <strong>Congress</strong> on Deterioration and
Conservation of Stone. ICOMOS Sweden, Stockholm, Vol. I, 479-486.
Pápay Z., Török Á. 2008. Three consolidants and three porous limestones: testing the
effectiveness of consolidants on Hungarian porous limestones from Sóskút
quarry, in terms of physicomechanical properties. In: Lukaszewicz, J. &
7

Niemcewicz, P. (Eds.): Proceedings of the 11th <strong>International</strong> <strong>Congress</strong> on
Deterioration and Conservation of Stone. Nicolaus Copernicus University Press,
Torun, Vol. I, 717-724.
Török Á 2002. Oolitic limestone in polluted atmospheric environment in Budapest:
weathering phenomena and alterations in physical properties. In: Siegesmund, S.,
Weiss, T., S., Vollbrecht, A (Eds.), Natural Stones, Weathering Phenomena,
Conservation Strategies and Case Studies. Geological Society, London, Special
Publications, 205, 363-379.
Török, Á., Rozgonyi N, Prikryl, R., Prikrylová, J. 2004: Leithakalk: the ornamental and
building stone of Central Europe, an overview. In: Prikryl, R. (ed), Dimension
stone. Balkema, Rotterdam, 89-93.
Wheeler, G., Mendez-Vivar, J., Goins, E. S., Fleming, S. A., Brinker, C. J. 2000.
Evaluation of alkoxysilane coupling agents in the consolidation of limestone.
<strong>International</strong> In: Fassina, V. (ed): Proceedings of the 9th <strong>Congress</strong> on
Deterioration and Conservation of Stone, Elsevier, Amsterdam. 541-546.
8

CONSOLIDATION EFFICIENCY OF IN-SITU APPLICATION CONSIDERING
WEATHERING DEGREE FOR KOREAN SANDSTONE CULTURAL
HERITAGE
Myeong Seong Lee 1 , Jae Man Lee 1 , Sung Mi Park 1 , Jiyoung Kim 2
1 Conservation Science Division, National Research Institute of Cultural Heritage,
Daejeon, Korea
2 Department of Cultural Heritage Conservation Sciences, Kongju National University,
Gongju, Korea
Abstract
The aim of this study is to examine the efficiency of ethylsilicate consolidants on
sandstone according to the weathering state for an appropriate application to stone
cultural heritage in Yeongyang area, Korea. There are many sandstone and
conglomeratic sandstone cultural heritages in Yeongyang area, and they require
conservation intervention due to granular disintegration and scaling of the stone surface.
Hyeonri Three-storied Stone Pagoda made of typical sandstone of this area was
investigated for the analyses of the material and deterioration. And both in-situ and
laboratory applications of consolidants were conducted to the outcrop and quarried
stones which showed same characteristics in rock types and weathering degree with the
stone pagoda. As a result, Wacker OH 100 and Remmers 300 showed the highest
consolidating effect, and Remmers 300 especially performed more strengthening effect
for the loosen and granular-disintegrated surface of the Yeongyang sandstone.
Keywords: Sandstone, Weathering Degree, Consolidation, In-situ Application,
Ultrasonic Velocity
1. Introduction
Yeongyang area is known to be the representing district of sandstone distribution of
Korean Peninsula and there are numerous national treasures and stone cultural heritages
in the area. Many of these stone cultural heritages have gone through many restoration
and preservation phases. However, most of the stone cultural heritages are exposed to
the natural environments, as well as artificial damages and it leads to severe weathering
of the exposed surfaces. Furthermore, the cement mortars and epoxy resin that were
used for the restoration and conservation materials in the past are losing its functions,
showing discoloration and cracking which is a problem in the conservation point of
view.
In order to prevent any side effect as mentioned above, there is a necessity to
analyze the characteristics of the resin used for foreign and domestic stone cultural
heritage restoration, as well as the changes in the physical properties of the stones,
before and after the treatment (Shin et al., 2004; Kim et al., 2009; Song et al., 2009a;
2009b; Lee et al., 2010). There is also a necessity for examining the mid-to-long term

stability through regular monitoring and field application tests against natural
weathering. However, scientific research on the characteristics of the conservation
material has not been carried out considering the Korean climate and environment prior
to its use. Now we are observing that the restoration and preservation materials are
turning out to be playing a role in the persisting damage (Lee et al., 2010).
In this research, with having the material and weathering characteristics of the
Korean sandstone cultural heritages in consideration, we have identified the damage
status and the material characteristics of the stone pagodas and Buddha monuments
dispersed in the area in order to choose the proper surface reinforcing material, as well
as an effective treatment method. We chose the Hyeon-ri Three-storied stone pagoda
(Treasure no. 610) as our test subject since it was severely damaged and showed
common weathering phenomena of the sandstone works. Detailed material
characterization and damage assessment were performed. We also searched and
obtained an outcrop specimen that has the same material and similar weathering degree
as the stone pagoda to perform indoor tests of the consolidants and in-situ application
tests. With the test results, we have determined the suitable consolidant and method for
the sandstones in the Yeongyang area.
2. Current Status and Research Methods
At the Yeongyang study area, there are a total of 9 stone cultural heritages including
7 pieces of stone pagoda, 1 piece of stone Buddha monument, and 1 Buddhist flagpole
supports. Most of these stone pagoda and Buddha monuments are composed of
sandstone, and there is a development of cracks, joints and decomposition of the
surfaces along the natural bedding planes. Also, the weathering degree is quite high, due
to the exfoliation, peeling, and falling out of the sandstone. Especially, mechanical
strength of the material of Three-storied stone pagoda of Hyeon-ri is dramatically
decreasing due to natural weathering. Conservation intervention was performed in the
past but further treatment is required due to deterioration and discoloration of the
conservation materials.
We have selected the Three-storied stone pagoda of Hyeon-ri, which bears a wide
range of weathering characteristics, as our research subject in order to establish a
conservation remedy for the sandstone cultural heritages dispersed in the Yeongyang
area. The comprehensive and quantitative examinations regarding the material
characteristics and damage rate of the stone structures were performed. And we have
collected sandstone samples in the Yeongyang area that are showing the same stone type
and weathering degree as the stone pagoda. The samples were tested by laboratory
experiments comparing the characteristics and performance before and after the
consolidation. We also applied the consolidant to the outdoor outcrops and examined its
stability.
3. Material Characteristics and Deterioration Assessment
The material of the Three-storied stone pagoda of Hyeon-ri is composed of red to
purple brown sandstone and conglomerate (Figure 1A). These stone types can be easily
observed in the surrounding areas of the stone pagoda. These rocks are the main
basement rocks of the Dogyedong Formation that is widely distributed in the
Yeongyang area. The sandstone of the pagoda consists of anhedral quartz and feldspar,
and clay matrix with irregular grain size and roundness distribution under polarized

microscope (Figure 1B). The rock-forming minerals are loosely bonded together and the
matrix is generally altered to clay minerals.
The conservation treatment for the stone pagoda was performed in 2005. However,
since the stone pagoda was exposed to nature and climate changes, the weathering
degree has rapidly changed higher due to the contraction-expansion of the minerals,
chemical reactions and contamination. The most significant physical damage on the
pagoda is granular disintegration and exfoliation. In the basement of the pagoda, a
granular disintegration formed over the whole surface of the stone (Figure 1C), and
exfoliation formed predominantly on the southern and eastern faces. In addition, the
treatment materials applied in the past are deteriorated and discolored, and damaging
aesthetic values of the pagoda (Figure 1D).
Figure 1. Lithological characteristics and damage aspects of Three-storied stone pagoda in
Hyeonri. (A) Reddish to purple brown sandstone and conglomeratic sandstone. (B) Polarizing
microscopic images showing heterogeneous roundness and grading of quartz. (C) Granular
disintegration of the basement stone. (D) Discoloration of the conservation material. (F) Average
ultrasonic velocity according to the elevated position of the stone elements.
The ultrasonic velocity measurement has been conducted to the pagoda in order to
determine physical strength of the stones. The obtained velocity was calculated by
directions and individual stone elements (Figure 1F). The results are as follows:
southern part average 1994 ㎧, eastern part average 2012 ㎧, northern part average
2097 ㎧, and western part average 2107 ㎧.
Using the results, the weathering degree was determined. The results showed that
the overall degree is class 4 (Highly weathered) with the exception of the southern face,
which is class 5. The basement and the 3 rd body stones is class 5 (Completely
weathered), whereas the 1 st body and roof stones, 2 nd body and roof stones, and 3 rd roof
stones are categorized as class 4 (Figure 1F). Especially, the foundation stones showed
much lower ultrasonic velocity. This referred to more granular disintegration and
surface deterioration of the stones compared to other parts.

4. Effectiveness Evaluation for Stone Consolidation
The treatment materials of Hyeon-ri Three-storied stone pagoda were severely
deteriorated so that it is required further conservation intervention. Prior to the
retreatment, indoor and outdoor application tests to determine the stability and the
effectiveness of the consolidant were performed. Firstly, we analyzed the changes of the
mechanical properties and strength of the stones before and after the consolidation.
The four consolidants were selected, such as Wacker OH 100, Remmers OH,
Remmers 100 and Remmers 300, according to its ethyl silicate content. We have chosen
to use the vacuum impregnation method for consolidation. Although this method is
limited to indoor application, it performs the maximum effectiveness. The treatment
environment was as follows: temperature 20℃, relative humidity 40 ~ 50%. After the
consolidation, we put the samples until there was no change in weight to settle the
consolidation reaction. When the samples completed its reaction, the physical properties
were measured and compared to the samples before the consolidation.
Of the 4 types of consolidants, Wacker OH 100, which had the highest silicate
content, showed the highest penetration rate at 17.0 g. Remmers 100, with the lowest
silicate content, showed the lowest penetration rate at 12.5 g. Therefore, we could
estimate that higher silicate contents would produce higher consolidating effect. Also, in
every case of the testing, the weight continually decreased as time went on, but
stabilized after 40 days.
After the treatment, porosity and absorption rate decreased by 3.33% and 1.41%,
respectively, and the ultrasonic velocity increased by approximately 677m/s. The
compressive and tensile strength increased by 45.97MPa and 2.58MPa, respectively, and
Young’s Modulus rate has also increased (Figure 2). We can conclude that the physical
properties have improved due to the penetration of the consolidandts by filling the pores
and cavities in the stones.
Focused on the individual consolidants, the Wacker OH 100 showed the highest
increasing rate for the porosity and absorption rate, whereas Remmers 300 showed the
highest increasing rate for the ultrasonic velocity. Remmers 300 showed comparably
higher increasing rate for compressive strength and Poisson’s Coefficient, while all 4
types showed similar characteristic changes for tensile strength (Figure 2). Therefore we
concluded that Wacker OH 100 and Remmers 300 produced the best reinforcement
effect. They both contain 99~100% ethyl silicate contents, and we now understand that
the ethyl silicate concentration plays a large role in producing favorable consolidating
effect (Song et al., 2009a; Han et al., 2008).
5. In-situ Application for Stone Consolidation
Up until recently, it was difficult to gain satisfactory efficiency from the
preservation and restoration methods and treatments used for the stone cultural heritages.
This was because there was not enough information and understanding about the stone
materials and weathering characteristics, and there were not enough scientific
examinations and tests to develop a suitable consolidating agent and application
methods(Song et al., 2009a). In-situ application simulation tests were performed on
outdoor bedrocks with similar weathering degree and stone properties as the Hyeon-ri
Three-storied stone pagoda, using Wacker OH 100 and Remmers 300, which showed
the best performance in the laboratory experiments.

Figure 2. Physical properties of the untreated and treated samples by consolidants.
The selected site is a bedrock surrounding Yeondaeam, near the Samji-dong stone
pagoda. In-situ application areas were divided into ‘highly weathered areas’ (area 1, 2,
3) and ‘completely weathered areas’ (area 4, 5, 6). Two different methods were used to
see the difference in results that are spray and brush applications. The consolidation
treatment progressed over three trials (Song et al., 2009a), and the changes in material
characteristics were monitored using the ultrasonic velocity (PUNDIT PLUS).
After the application, the ultrasonic velocity increased in all areas. However, after
the 3 rd application, the rate began to slowly decrease (Figure 3A, 3B). This means that
the pores and cavities of the stone were filled with the consolidant during the application,
but it slowly regained the cavity space after a period of time.
The ultrasonic velocity increasing rate was calculated by dividing the velocity
difference between the pre-treatment and post-treatment, with the pre-treatment value.
The results are as follows: After the 3 rd application, there was an increase of 26.6% in
area 1, 22.2% in area 2, 23.7% in area 3, 40.1% in area 4, 34.9% in area 5, 37.6% in area
6 (Figure 3C). Overall, the ‘completely weathered areas’ (area 4, 5, 6) showed a higher
increasing rate compared to the ‘highly weathered areas’ (area 1, 2, 3).
By treatment methods, the brush application method produced a 5% higher
increasing rate. Wacker OH 100 brush method produced the best results for areas 1, 2, 3.
Remmers 300 brush method produced the best results for areas 4, 5, 6(Figure 3C).
During the monitoring after the application, Wacker OH 100 spray method produced the
lowest ultrasonic velocity decreasing rate. However, when comparing the rate change,
we can see that the Remmers KSE 300 brush method performed the best consolidating
effect (Figure 3D).

Figure 3. Ultrasonic velocity during and after in-situ consolidation of sandstone. (A) Ultrasonic
velocity by application areas during the consolidation. (B) Ultrasonic velocity by application areas
during the consolidation. (C) Increased ultrasonic velocity by consolidants and application
methods. (D) Ultrasonic velocity change during and after the consolidation.
6. Discussion and Conclusion
By observing the changes of the porosity, absorption rate and ultrasonic velocity,
before and after laboratory experiment with the sandstone samples from the Yeongyang
area, Wacker OH 100 and Remmers 300 produced the best consolidating results. Based
on these results, by applying the consolidants at the field sites, the ultrasonic velocity
increased by 14.5~28.2% by using both Wacker OH 100 and Remmers 300.
Furthermore, as the number of application trials of the consolidants increased, it showed
a consistent increasing rate. Therefore we can verify that a minimum of three trials
process would be effective for the surface consolidation of the stone cultural heritages.
Regardless of which consolidant or method is used, the sandstones with higher
weathering degree showed immediate results after the application process. This is
because as the sandstones suffer more deterioration, there are more pores and cavities
formed within. And by applying the consolidant, the pore space is filled and
strengthened. Of the application methods, the brush method proved to be better than the
spray method. The spray method has a tendency to disperse into the air during the
application process with the volatile components of the consolidant, and the surface
contact ratio is significantly poor than that of the brush method, producing a less
penetration.
The Hyeonri Three-storied stone pagoda, as well as other cultural heritages of
similar material and deterioration, requires consolidation for highly deteriorated areas.
And a minimum of three trials of applications is also deemed necessary to maximize the

inding effect of loosened minerals, and it would effectively increase the strength of the
stone.
7. References
Illiev I. G. 1967. An attempt to measure the degree of weathering of intrusive rocks
from their physico-mechanical properties. Proceedings of the First <strong>International</strong>
<strong>Congress</strong>, <strong>International</strong> Society of Rock Mechanics, Lisbon, 109-114.
Kim S.H., Won J., Kang Y.S., Jang Y.D., Kim S.D., Kim J.J. 2009. Studies on Physical
Characterization of Gyeongju Namsan Granite after Treated with Consolidants. Journal
of Conservation Science 25, 245-254.
Lee J.W., Lee M.S., Choi Y.S., Oh J.H., Kim J.H., Kim S.D. 2010. Scientific
Conservation Treatment and Restoration of the Monument for Jo Heon and the Soldiers
in Chilbaeguichong (Chilbaeguichong jungbongjoheonseonsaengilgunsunuibi). Journal
of Conservation Science 26, 191-201.
Lee T.J., Kim S.D., Gal S.Y. 2010. Selection and Conservation for the Filler of Threestoried
Stone Pagoda at the West of Gameunsaji Site in Gyeongju. Journal of
Conservation Science 26, 361-370.
Shin G.H., Park H.D. 2004. Quantitative Evaluation of the Effectiveness of the
Consolidation with ‘Wacker OH 100’, in Great Sphinx. Geosystem Engineering 41, 7-
16.
Song C.Y., Jun B.K., Han M.S., Lee J.J., Kim S.D. 2009a. Field Experiments of
Consolidant and Filler for Stone Cultural Heritage: Primary Verification Using
Ultrasonic Velocity. Journal of Conservation Science 25, 87-100.
Song C.Y., Han M.S., Lee J.J., Jun B.K., Do M.H., 2009b. Characteristic Analysis on
Mixed Filler of Conservation Materials for Stone Cultural Heritage. Journal of
Conservation Science 25, 439-450.
Han M.S., Lee J.J., Jun B.K., Song C.Y., Kim S.D. 2008. Quantitative Evaluation for
Effectiveness of Consolidation Treatment by using the Ethylsilicate for the Namsan
Granite in Gyeongju. Journal of Mineralogical Society of Korea 21, 183-192.

TiO2-SiO2-PDMS NANO-COMPOSITE HYDROPHOBIC COATING WITH
SELF-CLEANING PROPERTIES FOR MARBLE PROTECTION
C. Kapridaki, N. Maravelaki-Kalaitzaki *
Technical University of Crete, Department of Sciences, Laboratory of Analytical and
Environmental Chemistry, Polytechnioupolis, Akrotiri, 73100 Chania, Crete, Greece.
E-mail address: pmaravelaki@isc.tuc.gr
Abstract
This research work presents the design of a transparent-hydrophobic hybrid
nanocrystalline SiO2-TiO2 coating and its application in the field of monument
conservation. The SiO2-TiO2 coating derived from a mixture of tetraethoxysilane (TEOS)
and titanium-tetra-isopropoxide (TTIP) incorporating an organosilane, the hydroxylterminated
polydimethylsiloxane (PDMS). The complete hydrolysis of TEOS was
achieved by non-toxic catalysts.
The physico-chemical properties of the transparent coating were evaluated by XRD,
FT-IR and SEM analyses. A transparent, non-fractured film derived from the sol-gel
hybrid synthesized material, which is practically amorphous with crystallite size of 5 nm.
The evolution of the hydrolysis of TEOS, as well as the copolymerization of TEOS,
TiO2 and PDMS as a function of time were assessed by FT-IR analysis.
The effectiveness of TiO2-SiO2-PDMS nano-composite as a hydrophobic coating
was investigated by capillary water absorption and contact angle measurements. A
reduction of 85 per cent (±6) in the value of the water capillary coefficient between
treated and untreated marble was observed. The results of the contact angle
measurements revealed an increase from 73 o to 115 o (0 min) between untreated and
treated marble. Furthermore, the treated samples show reduction of 34 per cent (±5) in
the water vapor permeability. The total color differences expressed as ΔΕ*=3.13 (±0.35)
range within acceptable limits. Finally, the self-cleaning property of the coating was
pointed out by the removal of biological micro-organisms of deteriorated marble after
the treatment.
The innovation of this synthesis pertains to the application of a hydrophobic
transparent nano-composite based on SiO2-TiO2 with self-cleaning properties, without
modifying the color of the marble surface and the water vapor permeability.
Keywords: hydrophobic, coating, self-cleaning, marble, nano-TiO2-SiO2-PDMS
1. Introduction
Titanium dioxide (TiO2) due to its great photocatalytic and antimicrobial activity is
classified as the most widely used semiconductor material in various fields, such as
photocatalysis for environmental purification, sensors, solar energy conversion and selfcleaning.
The photocatalysis of TiO2 nano-particles is activated when the TiO2 surface is
irradiated with a photon containing the energy equal to or larger than the band gap;
* Corresponding author. Tel.: +30 (28210) 37661; fax: +30 (28210) 37841.
E-mail address: pmaravelaki@isc.tuc.gr (Noni Maravelaki)
1

therefore, electron-hole pairs are generated and characterized as powerful oxidizing and
reducing agents. Moreover, the latter agents and the absorbed oxygen and water coming
from the surrounding air can react, producing highly active hydroxyl radical (OH • ) and
superoxide ions (O 2– ). It has been confirmed that the holes, the superoxide ions and
hydroxyl radicals are effective in the chain reactions for the breaking down of organic
compounds and microorganisms [Fujishima, Rao and Tryk 2000].
Indeed, TiO2 is an appealing additive in various materials such as aerosols, aerogels,
nanotubes, nanocrystals and mesoporous enhancing their photoreactivity (Fujishima,
Rao and Tryk 2000). Recent researches demonstrate that TiO2 nano-particles have
promising applications on building materials as bactericide agent (Poulios Spathis
Grigoriadou et al. 1999, Pincho and Mosquera 2011). The most appropriate way to
avoid deterioration and control biodeterioration of stone monuments is to prevent water
penetration into the stone bulk, since accumulation of water is mainly responsible for
stone decay. Due to this fact the development of new designed hydrophobic protective
coatings is deemed necessary (Price 1996, Warscheid Braams 2000). Polymeric resins,
acrylic polymers and copolymers, vinyl polymers, organosilicone compounds and
fluorinated materials as well as polymeric nano-composite systems and biodegradable
polymers are used as protective hydrophobic coatings on architectural surfaces (Price
1996, Manoudis Tsakalof Karapanoagiotis et al. 2009). Despite the effectiveness of the
hydrophobic products in confining the deterioration action of the external physicochemical
agents, nevertheless, these products proved to be less efficient towards
reducing the development of biological decay.
Into this framework the capability of TiO2 nanoparticles to degrade organic
pollutants and microorganisms is exploited by designing hybrid nano-composite
material through the sol-gel technique. A new hybrid coating, combining hydrophobic
and bactericide properties, was derived via the sol–gel procedure simultaneously
incorporating the TiO2 nano-particles. The development of this new transparent and
hydrophobic coating, as well as the assessment of its effectiveness as a stone protective
is presented in detail. The synthesis of the TiO2-SiO2-PDMS coating derived from the
following precursors: tetraethoxysilane (TEOS), titanium-tetra-isopropoxide (TTIP) and
the organosilane, hydroxyl-terminated polydimethylsiloxane (PDMS). The incorporation
both of TiO2 nano-particles and PDMS into the acid hydrolyzed TEOS was successfully
achieved. In this study, oxalic acid was used in order to catalyze the hydrolysis of TEOS,
as well as both to control the grain growth of TiO2 and the solvent drying, given its well
established property as a drying control chemical additive (DCCA) (Hench and Orcel
1989). Moreover, the choice of oxalic acid was based on its capability of reaction with
the calcium carbonate producing calcium oxalate which is more stable than calcite
(Manganelli del Fa', Camaiti, Borselli et al. 1989). The presence of PDMS provides the
hydrophobic properties due to its methyl groups, the enhancement of toughness and
flexibility of the silica network, thus preventing the gel from cracking during drying
(Mosquera, De los Santos and Rivas 2010, Zarrega, Cervantes, Salazar-Hernandez et al.
2010). Furthermore, PDMS can also be used as binder between the nano-particles due to
its terminal hydroxyl groups; therefore, prevention from precipitation can be achieved.
The new hybrid synthesized material was evaluated as a hydrophobic protective coating
on marble surfaces through the assessment of the water capillary absorption and contact
angle measurements. Finally, the water vapor permeability, the color changes of treated
marbles and the photocatalytic activity of the coating were assessed.
2

The innovation of this work pertains to successful application of a hydrophobic
transparent nano-composite based on SiO2-TiO2 with self-cleaning properties in the field
of the monument conservation. Neither toxic catalysts nor required calcination treatment
have been employed in this sol-gel process.
2. Experimental
2.1 Synthesis and application of coating on marble
Tetraethylorthosilicate (TEOS, Sigma Adrich), titanium tetraisopropoxide (TTIP,
Sigma Aldrich) and hydroxyl-terminated polythimethylsiloxane (PDMS, Sigma Adrich)
were used as raw materials for the synthesis of the nanocomposite coating, while ethanol
(EtOH, Sigma Adrich) and deionized water were used as solvents. Oxalic acid dihydrate
(Ox, Panreac) assists the catalysis of TEOS and creates a suitable pH environment
which prevents the particle TiO2 agglomeration (Mahshid, Askari, Ghamsari 2007).
In particular, incorporation of TTIP into TEOS was achieved and led to the
synthesis of the hybrid material. TEOS was firstly prehydrolyzed in a solution
containing EtOH, H2O and Ox at room temperature under magnetic stirring for 140 min.
In this solution, PDMS was added in drops and the stirring was continued for another 15
min. Finally, in the previous solution, TTIP was dropped, while the solution remained
under high stirring for 24 h at room temperature, thus providing a transparent sol. A
molar ratio TEOS/EtOH/H2O/PDMS/TTIP/H2C2O4 of 1/4/4/0.04/0.017/0.0001 was used.
The hybrid material was firstly characterized and consequently applied to marble
surfaces. The evaluation of the coating was performed by the following techniques. The
crystallinity of the nano-TiO2 as well as the nano-composite were studied by X-Ray
diffraction (XRD) using Burker D8 Advance diffractometer, operated at 35 kV and
35 mA with Cu Kα radiation with a nickel filter at a scan rate of 2° min −1 and a Bruker
Lynx Eye strip silicon detector. The sample qualitative analysis was carried out by
Fourier Transform Infrared Spectroscopy (FT-IR) using a Perkin-Elmer 1000
spectrometer in the spectral range of 400-4000 cm -1 . After the synthesis the sol was cast
into cylindrical and transparent molds 5.5 cm in diameter and 2 cm in length, while the
mold was sealed on top by a moldable film (Parafilm). The mold was kept under
laboratory conditions to dry and assessed for its cracking behavior. The surface
morphology of both xerogel and treated marble surface was analyzed by Scanning
electron microscopy (SEM) using a FEI - Quanta Inspect D8334 instrument operating at
25 kV on the specimen upon which a thin layer of gold was evaporated.
Dionysos and Thasos marbles were the stone materials tested, which have been
widely used in many historic and modern building constructions. Marble specimens,
both rounded, 5 cm in diameter and 2 cm in height, and rectangular, with dimensions of
10x5x3 cm were shaped and polished. Before the application of the nano-composite
coating on the surfaces, it was necessary for all the specimens to be washed with
deionized water, dried for 24 h at 60 o C, weighted and stored in a dessicator. The
treatment of the specimens was carried out by brushing the designed sol on the
previously impregnated with EtOH surface. The treated specimens were kept under
laboratory conditions until a constant weight (±0.001g) was achieved; then, the same
procedure was repeated once more.
3

2.2 Evaluation of the coating effectiveness
In order to assess the treatment efficiency, on both treated and untreated surfaces
relative analytical techniques were employed and compared accordingly. The
effectiveness of TiO2-SiO2-PDMS as hydrophobic coating was primarily investigated by
comparing the water capillary absorption coefficients (WCA) of treated and untreated
surfaces. The water capillary absorption was performed by the gravimetric sorption
technique according to the NORMAL specifications 11/85. The results were plotted as
the mass of absorbed water per area of sample versus the square root of time. The
hydrophobic behavior was confirmed by water contact angle measurements according to
the sessile drop method (Yildirim Erbil 2012). The measurements were performed using
distilled water and an optical tensionmeter (Thetalite TL 101, KSV) under laboratory
conditions. Three droplets of water (~10 μL volume) were applied to different points of
each specimen by a needle from a distance sufficiently close to the substrate, so the
kinetic energy of the droplets would be negligible. The image of the droplet was
captured immediately for measuring the static contact angle (θs).
Moreover, water vapor permeability (WVP) was carried out on specially shaped
treated and untreated specimens, according to NORMAL 21/85. The results were plotted
as the mass of water vapor permeability per surface versus time and the curve defines
the water vapor permeability, which was calculated as the slope of the linear curve. The
water vapor permeability was estimated by the comparison between the calculated
slopes of treated and untreated specimens.
The effect of changing marble color by the application of the nano-composite was
determined using a colorimeter (Micromatch Plus, Sheen Instruments Ltd). The hybrid
coating material was applied on the half part of rectangular (10x5x3 cm) specimens; the
measurements were taken from both parts of the marbles (untreated and treated). The
results were evaluated as L* (brightness), α* (redness color) and b* (yellowness color)
coordinates. The total color difference (ΔΕ * ) represents the color change between the
treated and the untreated surface (Maravelaki-Kalaitzaki, Kalithrakas-Kontos, Korakaki
et al. 2006).
The self-cleaning evaluation of the hybrid nano-coating was achieved by the
application of the synthesized material on biological decay marble’s surface and
successively exposed to UV irradiation. The biological decay of the marbles was
previously assessed by the identification of the microorganisms by API systems (API
20E and API 20NE) (Cappitelli Principi Pedrazzani et al. 2007).
3. Results and discussion
3.1 Characterization of hybrid nano-coating
The X-Ray diffraction pattern of the sol-gel synthesized TiO2 is shown in Fig. 1a,
where the anatase phase of TiO2 can be observed. The existence of broad XRD peaks
confirms the small particles size of TiO2 nano-particles. Furthermore, the nanocomposite
is practically amorphous with 5 nm crystallite size, calculated from the XRD
data following the Scherrer formula (Patterson 1939). Fig. 1b illustrates the sol-gel
synthesized material, characterized as colorless and highly transparent. In addition, the
gelation time of the nano-composite material was approximately 5 days.
4

Figure 1. X-Ray diffractogram of (a) the sol-gel synthesized TiO 2 and (b) view of the
transparent sol-gel synthesized nano-composite.
FT-IR spectroscopy was carried out in order to evaluate the structural properties of
the TiO2-SiO2-PDMS coating. The FT-IR spectra of the precursor materials and the new
designed material at two times intervals are depicted in Fig. 2. In particular, Figs 2a and
2b show the FT-IR spectra of TEOS and PDMS, respectively, while Figs 2c and 2d
illustrate the spectra of the sol-gel synthesized material obtained on the 1 st and 35 th day
of curing time, respectively. The completion of hydrolysis of TEOS as well as the
copolymerization of the hybrid material was assessed by observing changes in several
absorptions. The absence of the characteristic peak at 1167 cm -1 , corresponding to the
rocking of the C-H bond in -CH3 of TEOS, on the 1 st day indicates the TEOS fully
hydrolysis (Fig. 2c) (Tellez, Rubio Morales et al. 2004, Rubio, Rubio and Oteo 1998,
Orcel, Phallipou and Hench 1986). Additionally, during the procedure of TEOS
hydrolysis new bonds of Si-O-Si are created and confirmed from either the presence of
new peak or the shifting of already existing peaks. The peak at 1077 cm -1 can be
justified by the antisymmetric stretching of the newly created Si-O-Si bonds (Fig. 2d),
moreover the shifting of the absorption at 792 cm -1 (Fig. 2a) to higher frequency at 805
cm -1 (Fig. 2c and 2d) can be explained by the neo-formed Si-O-Si bonds from the
hydrolyzed TEOS (Tellez, Rubio Morales et al. 2004, Orcel, Phallipou and Hench 1986).
In Fig. 2b, the existing peak at 1259 cm -1 corresponds to -CH3 groups in Si-(CH3) of
PDMS molecules (Tellez, Rubio Morales et al. 2004). The latter, on the 35 th day, shifts
to higher intensity (1267 cm -1 ) (Fig. 2d); this can be correlated with the appearance of a
new peak at 847 cm -1 most probably due to copolymerization reactions between Si-OH
groups and Ti-OH with PDMS molecules (Tellez, Rubio Morales et al. 2004). In Fig. 2a
the peak at 961 cm -1 is firstly related with the -CH3 rocking of the TEOS molecule
(Tellez, Rubio Morales et al. 2004, Orcel, Phallipou and Hench 1986). However, given
that at the same position the Si-OH stretching appeared, the shifting of this peak from
961 to 943 cm -1 observed on the 35th day, was most probably due to the incorporation of
TiO2 to the silica network and the creation of Si-O-Ti bond, which is located in the
spectral range of 920-950 cm -1 (Tellez, Rubio Morales et al. 2004). The absence of the
characteristic peaks of C-H located at 2975, 2927, 2887 and 877 cm -1 on the 35 th day,
was due to the formation of a homogeneous organic-inorganic hybrid xerogel (Tellez,
Rubio Morales et al. 2004, Rubio, Rubio and Oteo 1998). Consequently the FT-IR
analysis indicated the copolymerization of PDMS and TiO2 into the silica network
leading to the formation of a homogeneous organic-inorganic hybrid xerogel.
5

Figure 2. FTIR spectra of the raw materials TEOS, PDMS, as well as the sol-gel synthesized
coating at two time intervals: 1 st day (sol) and 35 th day (xerogel).
The SEM micrograph in Fig. 3a demonstrated the transparent and non-fractured
film derived from the sol-gel synthesized material. After treatment, the dense nonfractured
coating observed on the marble surface in Fig. 3b, further supported the
formation of the crack free nano-composite material.
Figure 3. SEM micrographs of: (a) the designed transparent nano-coating and, (b) the
marble surface morphology after the application of the coating.
6

3.2 Efficiency of the applied coatings
The effectiveness of the nano-composite material in providing hydrophobic
protection was evaluated by water capillary absorption (WCA) and contact angle
measurements. Fig. 4a illustrates the curves of the WCA, before and after treatment of a
representative specimen, whereas the corresponding water capillary coefficients are
reported in Table 1. The results of the WCA indicate a significant decrease after
treatment of both the capillary absorbed water and the water capillary coefficient.
Figure 4. (a) Water absorption curves and (b) Water vapor permeability curves of untreated and
treated marble.
Table 1: Variations of the Water Capillary Absorptions (WCA) and Water Vapor
Permeability (WVP) coefficients
Sample Product quantity
(mg/cm 2 WCA
) (mg/cm 2 s -1/2 WVP
)
(mg cm 2 h -1 )
Untreated - 0.1067 (±0.0088) N=5 0.140 (±0.044) N=4
Treated 0.658(± 0.138) N=5 0.0132(±0.0067) N=5 0.093(±0.017) N=4
% Reduction - 88(± 6) N=5 34(± 5) N=4
N= number of samples
The contact angle measurements depicted in Fig. 5 and Table 2 corroborated the
hydrophobic behavior of the applied coating. In particular, Fig. 5 shows the water static
contact angles for untreated and treated specimens in two successive times (0 and 20
sec), while in Table 2 the corresponding contact angle values are reported.
Figure 5. Water static contact angles of marble before and after treatment with the synthesized
coating, at two different times (0 and 20 sec).
7

Table 2: Contact angle measurements and chromatic parameters
Sample
Untreated
Treated
Contact angles Chromatic parameters
0 min 20 min L* α* b* ΔΕ*
72.62 o
(±2.93)
114.84
(±0.40)
unt= untreated, t=treated
59.93 o
(±1.80)
111.34 o
(±1.59)
Μ1unt 83.81 0.03 3.28
Μ2unt 84.14 1.34 5.91
Μ3unt 82.36 1.52 8.50
Μ4unt 85.32 1.29 8.76
Μ1t 80.98 0.37 3.42
Μ2t 81.60 1.58 7.86
Μ3t 78.26 1.86 11.5
Μ4t 87.55 0.99 5.76
3.14
(±0.36)
Based on the results of water capillary absorption and contact angle measurements it
can be stated that a sufficient surface hydrophobicity was achieved by the application of
the new hybrid material on marble surfaces.
The curves of the water vapor permeability of a representative marble sample before
and after treatment (Fig. 4b), as well as the calculated water vapor coefficient presented
in Table 1, indicate a decrease in the water vapor permeability. The 34% decrease of the
water vapor permeability complies with the acceptable ranges of the hydrophobic
products applied to monument surfaces.
The color coordinates (L*, a* and b*) before and after treatment are presented in
Table 2. Given that a color alteration with ΔΕ*=2-3 cannot be detected by the naked eye,
the chromatic variation after treatment expressed by ΔΕ*=3 can be considered
insignificant.
Figure 6. Self-cleaning of biological decayed marble
The self-cleaning efficiency of the designed nano-composite was investigated on
marble specimens with biological decay. Before the treatment of the decayed marbles an
identification of the existent microorganisms was performed by API systems. The
microorganisms identified on the marble specimens were the following: Klebsiella
pneumonia, Pasteurella pneumotripica, Sphingomonas, Burkholderia cepacia
(Pseudomonas). These specimens, after treatment and exposure to UV irradiation,
8

exhibited a color variation of ΔΕ*=21 with parameters ΔL*=20, Δα*=-0.9 and Δb*=0.51
(Fig. 6). These results are in accordance with the macroscopic evidence and can be
considered indicative for both the microorganism removal and the self-cleaning
properties of the sol-gel synthesized nano-material.
Overall, the TiO2-SiO2-PDMS coating seems to behave efficiently as a protective
product in the field of monument conservation, by enhancing the marble hydrophobicity
and influencing insignificantly the water vapor permeability and the color parameters.
Additionally, this new nano-coating possesses self-cleaning properties against
microorganisms and pollutants.
4. Conclusions
In the present study a new hybrid material for the protection of stone has been
successfully synthesized and investigated. The new TiO2-SiO2-PDMS nano-composite is
obtained through a simple sol-gel route in which the acid hydrolyzed TEOS is mixed
with PDMS and TTIP. The acid hydrolysis of TEOS was achieved by oxalic acid that
may affect beneficially the stone due to its reaction with calcium carbonate producing
the less soluble calcium oxalate than calcite. The oxalic acid and PDMS contribute to
the formation of a crack-free coating, whose antibacterial activity against the
microorganism growth and pollutant absorption was achieved by the anatase form of
TiO2.
The protective action on marble of the transparent and non-fractured film was
proven by the decrease of the water capillary measurements and the increase of the
contact angle values. The hydrophobic effectiveness of the protective coating was
attributed to the methyl groups of PDMS. Finally, after treatment the variation of the
water vapor permeability and color parameters ranged within acceptable limits.
Summarizing, a transparent, hydrophobic, crack-free TiO2-SiO2-PDMS nanocomposite
was designed for stone protection exhibiting self-cleaning and hydrophobic
properties, as well as insignificant modification of the water vapor permeability and
color parameters.
Acknowledgment
This research has been co-financed by the European Union (European Social Fund -
ESF) and Greek national funds through the Operational Program "Education and
Lifelong Learning" of the National Strategic Reference Framework (NSRF) - Research
Funding Program: Heracleitus II, Investing in knowledge society through the European
Social Fund. Part of this research has been co-financed by the European Union
(European Social Fund - ESF) and Greek national funds through the Operational
Program "Innovation & Enterprise Unit of the Technical University of Crete" of
Operational Program "Education and Lifelong Learning"-Action: "Nursery of Ideas".
5. References
Cappitelli, F., Principi, P. Pedrazzani, R. Toniolo, L. Sorlini, C. 2007. ‘Bacterial and
fungal deterioration of the Milan Cathedral marble treated with protective synthetic
resins’, Science of the Total Environment 385(1-3): 172-181.
Fujishima, A., Rao, T. Tryk, A. 2000. ‘Titanium dioxide photocatalysis’. Journal of
Photochemistry and Photobiology C: Photochemistry Reviews, 1:1-21
9

Hench, L. Orcel, G. 1989. ‘Drying control chemical additives for rapid production of
large sol-gel derived silicon, boron and sodium containing monoliths’, United States
Patent, Patent number 4.851.150.
Mahshid, S. Askari, M. Ghamsari, M. 2007. ‘Synthesis of TiO2 nanoparticles by
hydrolysis and peptization of titanium isopropoxide solution’, Journal of Materials
Processing Technology, 189(1-3): 296–300.
Manganelli del Fa', C. Camaiti, M. Borselli, G. Maravelaki, P. 1989. ‘The oxalate Films:
Origin and Significance in the Conservation of Works of Art’ Paper presented at the
Proc. <strong>International</strong> Symposium in Milan, Centro CNR Gino Bozza.
Manoudis, P. Tsakalof, A. Karapanoagiotis, I. Zuburtikudis, I. Panayiotou, C. 2009.
‘Fabrication of super-hydrophobic surfaces for enhanced stone protection’, Surface
and Coatings Technology, 203(10-11): 1322-1328.
Maravelaki-Kalaitzaki, P. Kalithrakas-Kontos, N, Korakaki, D. Agioutantis, Z.
Maurigiannakis, S. 2006. ‘Evaluation of silicon-based strengthening agents on porous
limestones’, Progress in Organic Coatings, 57(2):140-148.
Mosquera, M. De los Santos, D. Rivas, T. 2010. ‘Surfactant-Synthesized Ormosils with
Application to Stone Restoration’. Langmuir, 26(9): 6737-6745.
Orcel. G, Phallipou, J, Hench, L. 1986. ‘Structural changes of silica xerogels using low
temperature dehydration’, Journal of Non-Crystalline Solids, 88:114-130.
Patterson, A. 1939. ‘The Scherrer Formula for X-Ray Particle Size Determination’,
Physical Review, 56(10): 978-982.
Rubio, J. Rubio, F. Oteo, J.L. 1998. ‘A FT-IR Study of the Hydrolysis of
Tetraethylorthosilicate (TEOS)’, Spectroscopy Letters, 31(1):199-219.
Tellez. L. J, Rubio, F. Morales, E. Oteo, J.L. 2004. ‘FT-IR Study of the Hydrolysis and
Polymerization of Tetraethyl Orthosilicate and Polydimethyl Siloxane in the Presence of
Tetrabutyl Orthotitanate’, Spectroscopy Letters, 37(1): 11-31.
Poulios, I. Spathis, P. Grigoriadou, A. Delidou, K. Tsoumparis, P. 1999. ‘Protection of
marbles against corrosion and microbial corrosion with TiO2 coatings’, Journal of
Environmental Science and Health, Part A: Toxic/Hazardous Substances and
Environmental Engineering, 34(7): 1455-1471.
Price, C.A. 1996. Stone Conservation, United States of America: The Getty
Conservation Institute.
Yildirim Erbil, H. 2012. ‘Evaporation of pure liquid sessile and spherical suspended
drops: A review’, Advances in Colloid and Interface Science, 170:67-86.
Zarrega, R. Cervantes, J. Salazar-Hernandez, C. Wheeler, G. 2010. ‘Effect of the
addition of hydroxyl-terminated polydimethylsiloxane to TEOS-based stone
consolidants’, Journal of Cultural Heritage, 11(2):138-144.
10

COMPACT LIMESTONES AS HISTORICAL BUILDING MATERIAL:
PROPERTIES OF THE TRANI STONE (APULIA, SOUTHERN ITALY) AND
PRELIMINARY STUDY FOR SELF CLEANING TREATMENTS
Angela Calia 1 , Loredana Matera 2 , Mariateresa Lettieri 3
CNR - Istituto per i Beni Archeologici e Monumentali, Provinciale Lecce-Monteroni,
73100 Lecce (Italy)
1 a.calia@ibam.cnr.it ; 2 l.matera@ibam.cnr.it ; 3 mt.lettieri@ibam.cnr.it
Abstract
In this paper we deal with an experimental activity aimed to the characterization of
a compact local limestone, named Trani stone. Compact limestones are very common
materials used in the monumental and historic built heritage. In particular, with
reference to the Apulian region (Southern Italy), they are the constituent stones of the
numerous Romanesque Cathedrals, as well as of many other important monuments such
as the UNESCO site of Castel del Monte. They have also been employed for the
building of the fortified towns and norman-swabian castles facing the sea. The study
reports mineralogical-petrographical and physical features, with particular reference to
the behaviour with respect to the water. Ultrasonic tests have also been performed as
indirect tool for the qualification of the stone in dry and wet conditions. A superficial
treatment with photocatalytic titania (in water and alcoholic solution) has also been
applied to the stone, in order to study the potential use of self-cleaning and antipollution
nanotitania coatings for stone surface protection; a preliminary assessment of
the morphology and distribution of the titania films on the stone surface and related
colour changes has been carried out.
Keywords: compact limestone, physical characterisation, stone protection, nanotitania
coatings
1. Introduction
Compact limestones are very common materials used in the monumental and historic
built heritage, thanks to their large availability, workability and attractiveness. Many
varieties of limestones can be found and they are selectively used as building or
ornamental stones in minor buildings and monuments (Siegesmund and Török, 2011).
Compact limestones mainly undergo to superficial decay, that includes dissolution of the
soluble carbonatic components, sulphatation processes, deposition of substances coming
from the surrounding environment (Bell, 1993). Dirt on the surface is a threat for the
conservation of the stone surface, other than a drawback from an aesthetical point of
view. This problem has been increasing in urban contexts, where the impact of the dirt
makes the cleaning a relevant work within the conservation activities; moreover it also
makes the maintenance for the preservation of the integrity of the surface very expensive
due to the frequency that would be required. With reference to this problem, advantages
could arise from the of application of new advanced products, such as photocatalytic
TiO2.

Compact limestones are widely spread in the Apulian region (Southern Italy), due to the
geological context of this area. The cultivation of these stone materials dates backs to
the ancient times; until nowadays it has been a relevant economic resource for this area
and the stone exploitation and trading is one of the most important activity in the region.
Other than the common building materials in minor built heritage that give a typical
fingerprint to many local historic towns, compact limestones are used in the numerous
Romanesque Cathedrals of high artistic and architectural value, as well as in many other
relevant monuments such as the UNESCO site of Castel del Monte. They have also been
employed within the fortified towns and norman-swabian castles facing the sea.
Among the Apulian limestone, the "Trani Stone" is the object of the present study,
aimed at characterizing the material for conservation purpose.
The mineralogical-petrographical and physical features have been investigated. With
reference to its conservation, the attention has been devoted to explore the possibilities
of application of new advanced products, such as photocatalytic TiO2. Two experimental
nanotitania-based products (in water and alcoholic solution) have been applied to the
stone, in order to study the potential use of self-cleaning and anti-pollution nanotitania
coatings for stone surface protection. The basic assessment of the treatments has been
carried out, by studying the morphology and distribution of the Titania films on the
stone surface and the related colour changes.
2. Stone characterization
2.1 Mineralogical-petrographical and porosimetric features
Mineralogical-petrographical features have been investigated by optical microscopy
with transmitted light, using a Zeiss microscope. Porosimetric analyses have been
carried out by mercury intrusion porosimetry (Carlo Erba Porosimeter 4000); integral
open porosity and pore size distribution have been determined, according to the
NorMaL Recommendation 4/80 (Raccomandazione NORMAL 4/80, 1980).
Trani stone is a very well cemented grainstone (Dunham, 1962) (Figure 1). It has a
grain supported texture, with the allochemical components mainly made of fossil’s
fragments (pelagic foraminifera, nummulites, ostracods and miliolidi) having fine to
medium size (80-300µm); macrofragments have also been sometimes observed. The
spatic calcite cement is present between the grains and in the inner fossil pocket,
forming a crystalline mosaic that close almost the voids.
Figure 1. Trani stone photomicrograph
(thin section, crossed nicols)
Figure 2. Pore size distribution

Trani stone is a very compact limestone. The integral open porosity measured by
MIP is very low (4%); the pore size has unimodal distribution and is mainly made of
smallest pores (between 0.05 and 0.001 microns) (Figure 2).
2.2 Physical properties
The following tests and analyses have been carried out:
- measurements of the ultrasonic wave velocity (Vp and Vs) in dry and wet conditions,
by direct transmission method and 1-MHz probes (4Plus Epoch, Olympus equipment),
on cubic specimens of 10 cm. For each sample three measurements were taken along X
Y Z directions and the mean values of Vp and Vs have been considered. The anisotropy
index has been calculated as ARS= (Vpmax-Vpmin)/(Vpmax) ∙ 100 (Ruedrich et al.,
2007). Vp, Vs and ARS have been expressed as the mean values on a total of 10
samples;
- colour measurements (Raccomandazione NORMAL 43/93. 1993), using a Chroma
Meter Minolta CR 300 colorimeter; colour parameters have been referred to the CIELab
space and ten measurements have been carried out on each sample area of 5x5 cm; the
results are expressed as the mean values on a total of 5 samples ;
- static contact angle measurements (Raccomandazione NORMAL 33/89, 1989), by a
Lorenzen and Wettre apparatus (Costech instrument); thirty measurements for each
sample area of 5x5cm were performed; the results are expressed as the mean values on a
total of 5 samples;
- capillarity rise test (UNI 10859, 2000) on10 samples measuring 5x5x2 cm; duration of
the test 8 days;
- contact-sponge test (Vandevoorde et al., 2009); for this test a disc-shaped sponge,
having an area of 22.06 cm 2 , has been soaked with 5 ml of water and pressed against the
stone surface for 1 minute; water absorption has been calculated as Wa= (mi - mf)/ (A ∙
t),where mi is the initial weight of the sponge with water, mf is the weight of the sponge
after the application to the stone surface, A is the sponge area, t is the contact time; six
measurements were performed and the reported data were expressed as the mean values;
- total immersion test (Raccomandazione NORMAL 7/81 1981) on 10 samples
measuring 5x5x5 cm; duration of the test 12 days;
- evaporation test (Raccomandazione NorMaL 29/88, 1988) on 10 samples measuring
5x5x5 cm; duration of the test 9 days;
- water vapour permeability test (Raccomandazione NORMAL 21/85, 1985) on 10
samples measuring 5x5x1 cm.
The compactness and the high quality of the stone correspond to high ultrasonic waves
propagation along X, Y and Z directions, in wet and dry conditions (Table 1). Similar
values of the ultrasonic velocities have been found in both silicatic and carbonatic
compact stones (Fort et al., 2011). The anisotropy index ARS give a very low value
(2%), meaning a negligible anisotropy.

Table 1. P and S wave velocity in dry and wet conditions
Sample condition
Wave velocity (m/s)
Wave
X Y Z
ARS (%)
Dry
VP
VS
6238 ± 199
3421 ± 26
6274 ± 188
3440 ± 269
6141 ± 122
3369 ± 88
2.12
Wet
VP
VS
6459 ± 216
3333 ± 124
6620 ± 243
3375 ± 114
6391 ± 82
3265 ± 55
3.46
The values of the L *, a * and b *colour parameters of the stone are reported in
Table 2. The low standard deviation values denote the homogeneous colour of the stone.
The very low stone porosity lead to a water-stone contact angle of 45°.
Table 2. Summary of the physical properties of Trani stone.
Integral open porosity [%] 3.7 ± 0.3
Prevailing range of pore radius [µm] 0.025-0.001
Colour parameters L* = 87.07 ± 0.28
a* = 1.36 ± 0.05
b* = 8.84 ± 0.21
Water-stone contact angle [°] 46 ± 5*
Water absorbed by the “contact sponge test” (Wa) 0.76 ± 0.04
[mg/cm 2 ∙min]
Total water absorbed by capillarity (Qft) [mg/cm 2 ] 69.46 ± 2.69
Capillary absorption coefficient (CA) [mg/cm 2 ∙ s 1/2 ] 0.32 ± 0.02
Capillary absorption index (CI) [mg/cm 2 ∙ s 1/2 ] 0.83 ± 0.01
Imbibition capacity (IC) [%] 0.38 ± 0.09
Dying index (DI) [%] 0.13 ± 0.02
Water vapour permeability (g/m 2 ∙ 24h at 20°C) 28 ± 4
The hydric behaviour of the stone is depicted in Figures 3, 4 and 5. The stone
absorbs low amount of water by capillarity (70 mg/cm 2 , corresponding to a weight
increase of 1.35 ± 0.06%). The most absorption (90% of the total water amount) takes
place throughout the first 24 hours. In the same way the water uptake by total immersion
is very limited; the recorded imbibition capacity (IC) is 0.4% ca.
With reference to the drying, the most water loss by evaporation takes place during
the first 24h (72%) and the residual water within the samples after 3 days is 10%;
negligible water content has been recorded after 10 days.
The water vapour permeability of the stone is 28 g/m 2 per 24 hours .
All the parameters coming from the analyses and tests are summarized in table 2.

Figure 3. Capillarity water absorption curve.
Figure 4. Total immersion water absorption curve.
Figure 5. Water evaporation curve.

3. Application of the Titania coatings
Titanium dioxide in the anatase form has proved to be an attractive compound,
thanks to its chemical stability, non-toxicity and moderate price, to working as a
preventive protection system activated by solar and UV light exposure. The efficiency
of commercial and experimental titania coating as self-cleaning treatments on building
stone materials has been recently investigated (Licciulli et al., 2011; Potenza el al., 2007;
Luvidi et al., 2010; Quagliarini, 2012; Quagliarini, in press; La Russa, 2012). Two
titania based experimental products has been selected for the application on the Trani
stone.
Titania sols have been synthesized through the sol-gel technique starting from
tetrapropylorthotitanate (TPOT) as titania precursor (Licciulli et al, 2011). Distilled
water or a commercial solution of propanol, buthanol and ethanol have been used as
solvents. The TiO2 concentration within the solution is 1% by weight. The titania films
have been applied on the stone specimens by spray coating, using a H.V.L.P. (High
Volume – Low Pressure, AKOKA Model H2000A) spray gun (1 mm nozzle diameter,
operating pressure: 2 bar).
The amount of the applied solution is 18.6 and 37.8 g/m 2 with reference to the
alcoholic (TAL) and aqueous (TAQ) products, respectively. It has been determined by
weight measurements, before and after the treatment of the stone specimens. After the
application of the products, the samples were kept in laboratory environmental
conditions (23 ± 2°C and 50 ± 5% relative humidity). Then the treated stone surfaces
have been subjected to morphological observations by scanning electron microscope
(SEM) in order to study the distribution and characteristics of the coatings on the stone
surface. Investigations have been carried out on samples without metallisation, in “low
vacuum” mode (0.6 torr, 25kV), using an Environmental-SEM (mod. XL30, FEI
Company); both secondary (GSE) and backscattered (BSE) electron detectors have been
used for the morphological observations.
Colour measurements have been also performed in order to assess if the films alter
the original chromatic features of the stone surface.
The first problem related to the application of the titania coatings consists in the
realization of a continuous and homogeneous film on the surface of the stone. The
coating issued from the alcoholic sol (TAL treatment) appears to be characterized by
extended micro-cracks (Figure 6a). On the contrary, good results have been obtained by
the TAQ treatment. A continuous coating, uniformly distributed on the stone surface has
been obtained by the application of the aqueous solution (Figure 6b), and any fractures
have been observed in this case. The quick evaporation of the alcoholic sol and the rapid
precipitation of the titania with the subsequent formation of cracks, account for these
results. Micro-cracks on the coatings have been observed in a previous study (Licciulli
et al, 2011) as depending on the solution’s concentration.

a b
Figure 6. ESEM images of the stone surfaces treated with a) TAL and b) TAQ.
The application of the titania films has been found do not alter in a significant way
the original colour of the stone surface. Colour variations of 1.24 and 0.68, in terms of
∆E values, were measured on the stone treated with the aqueous and alcoholic solution,
respectively.
4. Conclusions
The knowledge of the characteristics of the building materials used within the
hystorical-architectural heritage is the basic condition for conservation purpose.
Properties of the stones are the reference parameters for the choice of the conservation
treatments and the evaluation of their performance. With this regard, other than the wellknown
technical characteristics related to the stone qualification for the trading at
nowadays (Addabbo et al., 1985; Maggiore, 1983; Radina, 1956; Zezza, 1974; AA.VV.
1982), the work here carried out gives additional information on the stone properties of
local historical building materials that is of interest in the conservation field. Among the
building stones (Siegesmund and Dürrast, 2011), Trani stone is a material of high
quality. Its high compactness, low porosity and behaviour to the water absorption
account for the performance observed within the numerous buildings and monuments
that have been realised using this stone. Differently than for the porous stones widely
used in the same context of the Apulian region, whose relevant problem is the deep
penetration of the agents of the decay and the consequent heavy loss of materials from
the surface, Trani stone mainly undergo to the superficial decay. Dissolution processes
and dirt deposition on the surface have been the main decay problems along the time,
that have relatively compromised its conservation. Further threat in the last decades is
coming from the pollution in urban environments, that is also responsible of the increase
of the deposition rate of the dirt on the stone surfaces. With reference to this problem, it
could be advisable to verify the possibility of the application of new advanced products,
such as photocatalitic TiO2 and the advantages that could arise from their use for stone
treatments. Promising results have been obtained in order to the application of the water
based experimental solution, in terms of morphological characteristics of the film and
colour integrity of the stone. In addition the use of the water instead of the other
chemical solvents has undoubtedly lower impact with respect to the environment and
workers during the on-site application. These first basic aspects come from a
preliminary activity within a work in progress, devoted to the evaluation of the

possibilities and limits of new advanced products that, increasingly on, are proposed for
stone conservation .
5. References
A.A.V.V. 1982. ‘Marmi di Puglia’.In Regione Puglia, Novara: Istituto Geografico
De Agostini.
Addabbo, A., Maggiore, M., Mancini, I.L., Pieri P. and Walsh N. 1985. ‘Proposta
di classificazione dei marmi pugliesi sulla base dei caratteri genetici e dei requisiti
tecnici’. In III Convegno Nazionale su Attività estrattiva dei minerali di 2 categoria. Atti
delle giornate di studio 31-40. Parma: PEI.
Bell F.G. 1993. ‘Durability of carbonate rock as building stone with comments on
its preservation’. Environmental Geology 21: 187-200.
Dunham R.J., 1962. Classification of carbonate rocks according to depositional
texture. Classification of carbonate rocks. Mem. American Association Petrology and
Geology, 1, W.E. Ham Ed. 108-121.
Fort, R., Varas, M.J., Alvarez de Buergo, M. and Martin-Freire D. 2011.
‘Determination of anisotropy to enhance the durability of natural stone’. Journal of
Geophysics Engineering 8 S132.
La Russa, M.F., Ruffolo, S.A., Rovella, N., Belfiore, C.M., Palermo, A.M., Guzzi
M.T., Crisci G.M. 2012. ‘Multifunctional TiO2 coatings for Cultural Heritage’. Progress
in Organic Coatings 74: 186-191.
Licciulli, A., Calia, A., Lettieri, M., Diso, D., Masieri, M., Franza, S., Amadelli,
R. and Casarano, G. 2011. ‘Photocatalytic TiO2 coatings on limestone’. Journal of Sol-
Gel Science and Technology 60: 437–444.
Luvidi, L., Laguzzi, G., Gallese, F., Mecchi, A.M. and Sidoti G. 2010.
‘Application of TiO2 based coatings on stone surfaces of interest in the field of Cultural
Heritage’. In Proceedings of 4th <strong>International</strong> congress on science and technology for
the safeguard of Cultural Heritage in the Mediterranean Basin. Ferrari A. (Ed.) 495–
500 (Vol. II). Napoli: Grafica Elettronica.
Maggiore, M. 1983. ‘Relazione tra caratteri genetici e proprietà tecniche dei
"calcari delle Murge" impiegati come pietre ornamentali’. In Studi geologici e geofisici
sulle regioni Pugliese e Lucana, Vol. 14, 3-42. Bari: Adriatica.
Potenza, G., Licciulli, A., Diso, D., Franza, S., Calia, A., Lettieri, M., and
Ciccarella, G. 2007. ‘Surface engineering on natural stone through TiO2 photocatalytic
coatings’. In Proceedings of international RILEM symposium on photocatalysis,
environment and construction materials. Baglioni, P. and Casssar, L. (Eds.) 315–322.
Bagneux (France): RILEM Publications S.a.r.l..
Quagliarini, E., Bondioli, F., Goffredo, G.B., Licciulli, A., Munafò, P. 2012.
‘Smart surfaces for architectural heritage: Preliminary results about the application of
TiO2-based coatings on travertine’. Journal of Cultural Heritage 13: 204-209.
Quagliarini, E., Bondioli, F., Goffredo, G.B., Licciulli, A., Munafò, P. in press.
‘Self-cleaning materials on Architectural Heritage: Compatibility of photo-induced
hydrophilicity of TiO2 coatings on stone surfaces’. Journal of Cultural Heritage DOI:
10.1016/j.culher.2012.02.006.
Raccomandazione NORMAL 21/85 1985. ‘Permeabilità al vapor d’acqua’. Rome:
C.N.R – I.C.R..

Raccomandazione NorMaL 29/88 1988. ‘Misura dell’indice di asciugamento
(DRYING INDEX)’. Rome: C.N.R – I.C.R..
Raccomandazione NORMAL 33/89 1989. ‘Misura dell’angolo di contatto’. Rome:
C.N.R – I.C.R..
Raccomandazione NORMAL 4/80 1980. ‘Distribuzione del volume dei pori in
funzione del loro diametro’. Rome: C.N.R – I.C.R..
Raccomandazione NORMAL 43/93 1993.’Misure colorimetriche di superfici
opache’. Rome: C.N.R – I.C.R..
Raccomandazione NORMAL 7/81 1981. ‘Assorbimento d’acqua per immersione
totale – capacità di imbibizione’. Rome: C.N.R – I.C.R.., 1981.
Radina, B. 1956. ‘La Pietra di Trani’. Geotecnica 5: 206-216.
Ruedrich, J., Seidel, M., Rothert, E. and Siegesmund, S. 2007. ‘Length changes of
sandstones caused by salt crystallization’. In Building stone decay; from diagnosis to
conservation Prikryl R. and Smith B.J. (eds) 199-209. London: Geological Society of
London.
Siegesmund, S. and Dürrast, H. 2011. ‘Physical and Mechanical Properties of
Rocks’. In Stone in Architecture. Properties, Durability, 4th edn, Siegesmund, S. and
Snethlage R. (eds) 97-225. Berlin: Springer-Verlag.
Siegesmund, S. and Török, A. 2011. ‘Building Stones’. In Stone in Architecture.
Properties, Durability, 4th edn, Siegesmund, S. and Snethlage R. (eds) 11-95. Berlin:
Springer-Verlag.
UNI 10859 2000. ‘Determinazione dell’assorbimento d’acqua per capillarità’
Milan: UNI.
Vandevoorde, D., Pamplona, M., Schalm, O., Vanhellemont, Y., Cnudde, V. and
Verhaeven E. 2009. ‘Contact sponge method: Performance of a promising tool for
measuring the initial water absorption’. Journal of Cultural Heritage 10: 41-47.
Zezza, F. 1974. ‘Le pietre da costruzione e ornamentali della Puglia.
Caratteristiche sedimentologico-petrografiche, proprietà fisico-meccaniche e problemi
geologico-tecnici relativi all'attività estrattiva’. Rassegna Tecnica Pugliese - Continuità,
anno VIII, n.3-4.

PERFORMANCE OF NANOCOMPOSITES FOR CONSERVATION OF
ARTISTIC STONES
Franca Persia 1 , Luisa Caneve 2 , Francesco Colao 2 , Rosaria D’Amato 2 ,
Cristina Giancristofaro 1 , Giulia Ricci 1 , Luciano Pilloni 1 , Antonio Rinaldi 1
1 ENEA-UTTMAT, S. Maria di Galeria (Rome), Italy
2 ENEA-UTAPRAD Frascati (Rome), Italy
Abstract
Properties of consolidant and protective materials modified with nanoparticles
were analyzed in this study following their application onto marble and travertine
samples. To this purpose different solutions of an acrylic resin and a silicon-based
polymer with dispersed silica and titania nanoparticles were prepared. Artificial aging
processes, both in climatic chamber and in solar box, were carried out to simulate real
degradation processes in terms of photo-thermal effects and mechanical damage. The
relative durability of the two different consolidants as modified by nanoparticles was
evaluated comparatively by means of diverse diagnostic techniques, namely: scanning
electron microscopy (SEM), laser induced fluorescence (LIF), ultrasonic sound,
colorimetry, total immersion water absorption and contact angle. The results
demonstrate that nanoparticles enhance the effectiveness of consolidant and protective
materials because they induce substantial changes of surface morphology of the coating
layer and counter the physical damage observed during artificial weathering, especially
in alkylsiloxane products.
Keywords: stone conservation, marble, travertine, nanocomposites, hydrophobic
coatings, artificial weathering,
1. Introduction
In the last few years, nanocomposites have been frequently applied to restoration
and conservation of artworks (Mosquera 2008;Manoudis 2009;Kima 2009). In fact,
inorganic oxide nanoparticles, such as silica and titania, improve the performance of
materials used in conservation field.
Consolidant, protective and hydrophobic polymeric materials have been used in
conservation science for several decades (Torraca 1986;Tabasso 2006;Accardo 1981).
Their protective properties were studied with regard to different aspects, such as
different types of stones, natural and artificial environmental factors (e.g. water, solar
light, chemical and biological pollutants, etc.), and degradation mechanisms (Favaro
2006;Kaczmarek 1996;Melo1999).
Water, in any physical state, is a major degradation factor, due to its capability to
transport chemical and biological substances in and out of the material, as well as to
cause cracking in freeze-thaw and wet-dry cycles. The water effect depends on porous
structure, which is responsible of these phenomena in first place (Amoroso 1983;
Lazzarini 1986;Rodrìguez-Gordillo 2006).

For these reasons, consolidant properties and water repellence are often the most
important requirement of a conservative product. Recently, the effectiveness of
polymeric conservation products was improved by means of nanocomposites (Manoudis
2007, 2008;Shi 2008;Miliani 2007), which can be synthesized by different methods
depending on the specific aim (Manoudis 2008).
In the present work SiO2 and TiO2 nanoparticles, with size of about 15 nm and
produced by CO2 laser pyrolysis, were added as nanometric filler to acrylic and siloxane
polymeric dispersions. Nanosilica and nanotitania were chosen for their physical
properties, such as the improved water repellence (Shang 2005) following the increased
roughness. As confirmed by Scanning Electronic Microscopy (SEM), the roughness
achieved from adding nanocomposites resembled the well know surface properties of
the lotus leaf (“LotusEffect”), which confers self-cleaning and superhydrophobic
properties (Manoudis 2009).
The preservation properties of our nanocomposites were tested by the application
on two different litotypes, very common in outdoor cultural heritage: white marble
(statuary and veined Carrara) and travertine. Samples of treated stone were submitted to
artificial aging processes, both in climatic chamber and in solar box, to assess and
compare the performance vs. nanocomposites.
Among the many techniques capable to test the consolidant and the hydrophobic
performance of the individuated nanocomposites, non invasive diagnostic techniques
were chosen and deployed before and after the artificial aging processes, i.e. (i) SEM to
check surface morphology, (ii) ultrasound speed to evaluate the effectiveness of the
applied products in improving substrate cohesion, (iii) laser induced fluorescence (LIF)
to recognize and discriminate different types of nanocomposites, (iv) colorimetric
measurements to evaluate colour alteration, (v) contact angle to measure surface
wettability, and (vi) water absorption through total immersion to calculate the relative
imbibing capacity.
2. Experimental
2.1 Materials
Laboratory experiments were carried out on quarried travertine samples from
Tivoli (Rome, IT, 5x5x4cm in size) and on square blocks of three different types of
marble:(i) Carrara with grey veins, (ii) white statuary Carrara (5x5x2cm), and (iii) 8
pieces of historical marble of irregular size coming out from the deposit of the National
Roman Museum in Rome. Historical marbles were identified as Carrara, Proconnesio
and Pentelic (Bosco 2009).
Before applying the conservation products, the white Carrara marble samples were
aged naturally through open air exposure (inclined at 45° on the horizon, in south
direction) for 14 months. Carrara grey samples were aged through artificial weathering
in climatic chamber by 10 freeze-thaw cycles (between 10h at -18°C to 10 h at 30°C).
Prior to treatment all the samples were washed with deionized water.
The nanocomposites were obtained by using two different commercial products as
binders and two different nanoparticles as fillers with concentration NP/polymer of 0.2,
1 and 2 (%w/vol). The two binders are largely used for conservation of stone monument.
The first one is Paraloid B72, an acrylic resin (methylmethacrilate/ethylmethacrylate
copolymer MA/EMA 30/70 w/w%) sold by SINOPIA. The second one is Rhodorsil

RC80, a polyethysiloxane produced by Rhodia Silicones, Italy. The SiO2 and TiO2
nanoparticles were synthesised by CO2 laser pyrolysis of two liquid precursors, Si(OEt)4
and Ti(i-OPr)4 respectively, with size around 15 nm and low polydispersity. The
nanoparticles were used alone and together, with various mixing ratios. For the
nanocomposite based on Paraloid B72, the calculated amount of nanoparticles were
firstly dispersed in nitro solvent (GOLD 18, ITALCHIMICA LAZIO, IT) in different
concentration by using ultrasonic tip (Branson Sonifier 450) in order to obtain
homogeneous dispersions. Then, 300 mg of Paraloid B72 were added to the different
solution. For the nanocomposite based on polyethylsiloxane the nanoparticles were
directly dispersed in the commercial solution Rhodorsil RC80 and treated with
ultrasonic tip for 20 minutes.
The obtained solutions were applied on the stones by brushing until manifest
refusal to simulate the treatment in real situation, performed by a restorer. The treatment
was repeated after 4 hours (Ferreira Pinto 2008).
2.2 Aging
Artificial aging by sunlight and freeze-thaw cycles were performed to evaluate the
durability of the treated samples. A climatic chamber (SOLARBOX 1550 E) equipped
with a xenon arc light source, 1500 W with spectral range from 280 till 800 nm was
used. Following the standard NORMA UNI 10925-2001 (Cultural heritage- natural and
artificial stone materials- Methodology for sunlight irradiation), samples underwent
constant irradiation of 1000W/m 2 for 556 hours. The irradiation time was sufficient to
insure 2000 MJ/m 2 , the specific irradiation conditions required by the standard. Twenty
samples (from white and grey Carrara and 8 historical specimens) were treated.
The freeze-thaw resistance was carried out in an Angelantoni climatic chamber as
per UNI EN 539 – 2. Accordingly, samples were first immersed in water for 7 days till
complete imbibitions of water. Then, 50 freeze-thaw cycles were carried out by slowly
decreasing the temperature (over 15 minutes) down to -16±3°C, followed by a gradual
increase (over 15 minutes) up till 17 °C, with hold times of 15 minutes between the
ramps.
2.3 Evaluation tests
SEM micrographs were collected to investigate the nanoparticle composite film
morphologies and to evaluate the influence of nanoparticles on the integrity of surface
coatings after the aging processes. The microstructural investigations were performed
with a FEG-SEM LEO 1530 (Zeiss, Obercoken-Germany) equipped with In-lens
secondary electron detector, conventional secondary electron detector and scintillation
detector for backscattered electrons (Centaurus). The observations were carried out at
low voltage in order to reduce charging effects due to the insulating nature of the
applied polymeric films. The low voltage technique allowed observing the true surface
morphology of insulating materials eliminating the artifact created by the usual
metallisation.
The wettability properties of the nanocomposite coatings were assessed by static
water contact angle measurements to evaluate the local water repellence of the surface.
The measurements were carried out through a custom apparatus made in-house in
compliance with standard UNI EN 15802 – 2010 (Conservation of cultural property -
Test methods - Determination of static contact angle). The specimens were placed on a

sample holder (free to translate in XYZ axes) and a 5 μl water drop was deposited onto
the stone surface through a graduated micro-pipette. A stereo microscope Leica Wild
M3Z connected to a digital camera (Moticam 2000) was used to acquire the drop image.
The contact angle analysis was performed by the Low-Bond Axisymmetric Drop Shape
Analysis (LBADSA) Plugin for ImageJ software, which is based on the fitting of the
Young-Laplace equation to the image data (Stalder 2010).
Measurements of water absorption by total immersion were carried out to verify
the water absorption capability. The used method is the gravimetric one with which the
absorption curve and the relative Imbibing Capacity (I.C.) was measured for each
sample. The results are presented according to UNI Document 10921 (Natural and
artificial stones – Water repellents – Application on samples and determination of their
properties in laboratory) in terms of protection ratio percentages, P.R.%, which is
defined as the percentage variation between the imbibing capacity of untreated and
coated stone.
Ultrasonic velocity measurements (US) were carried out with a portable instrument,
a Krautkramer USM 23 with low frequency using a probe of 50 KHz through
transmission method and no sample preparation was necessary. Ultrasonic wave
velocity was measured in all specimens in the three directions X, Y, and Z. For each
direction, four values were recorded and the average was calculated. A thin layer of
water was used as acoustic coupling medium between the stone and the transducer.
LIF measurements were performed with an apparatus realized at the ENEA
laboratory. The radiation source is a Thomsom DIVA diode pulsed Nd:YAG laser able
to produce laser pulses both at 266 and 355 nm at a repetition rate of 20 Hz with a time
duration of 8 ns. A laser fluence of 0.9 mJ/cm 2 was selected in this case. The detected
spectral range was 200-800 nm and the spectrometer entrance was protected from the
intense backscattered radiation by means of an appropriate filter. Since no optical
elements were used to collimate the laser beam, the overall spatial resolution is a
function of the operation distance and in the present case it was possible to infer a
resolution of approximately 1 to 2 mm from the spot size on the target. The digitized
spectrum was transferred to a portable computer where a LabView program allowed the
user to set experimental parameters, to control data acquisition, and to perform a
preliminary data analysis.
Colorimetric measurements were executed to calculate colour changes induced by
photochemical and photothermal degradation of nanocomposite. Colour values, reported
in the CIEL*a*b* space, were obtained using a Minolta CM-525i Spectophotometer,
using a D65 illuminant. These results allowed calculating the total color difference ΔE,
relative to the same area of the sample before and after coatings application.
US, LIF, SEM, contact angle, water absorption and colorimetric measurements
were performed before and after each treatment and after the accelerated weathering
tests.
3. Results and discussion
3.1 Results and discussion before aging
SEM images reveal that the surface morphology of the treated stone primarily
depends on the nanoparticle concentration. Figure 1 shows the surface changes in
Rhodorsil RC80 deposited films as a function of the silica nanoparticle concentration.

The nanoparticles are not homogeneously dispersed in the polymer film, but they lead to
the formation of protruded aggregates with irregular shape.
Figure 1. SEM images of statuary Carrara marble treated with
Rhodorsil RC80 and 0.2% (a), 1% (b), 2% (c) w/v silica nanoparticles.
Few distinct aggregates separated by wide smooth areas of continuous polymer
film can be observed on samples covered with polymeric low particle concentration
films (0.2% w/v, Figure 1a). The increase in nanoparticles concentration (Figure 1b,c)
enhanced the surface roughness of the stone because the protruded aggregates formed by
the treatment are significantly higher. TiO2/SiO2 nanoparticles composite films show
different surface morphology due to the different interaction between silica/titania
nanoparticles and siloxane matrix. Nano-mixtures containing a reduced concentration of
titania nanoparticles are able to form homogeneous coatings, with aggregates embedded
inside a continuous protective film, as shown in Figure 2a,c. On the contrary, the RC80 -
TiO2 (1% w/v) - SiO2 (0.1% w/v) application increases the inhomogeneity of the film,
producing big aggregates followed by nanoparticles-free smooth regions. In fact, Figure
2b suggests that the randomly distribution of titania aggregates can create areas of
continuous visibly cracked polymer film.
Figure 2. SEM images of statuary Carrara marble treated with
(a) RC80 - SiO 2 (1%) - TiO 2 (0.1%), (b) RC80 - SiO 2 (0.1%) –
TiO2 (1%), (c,d) RC80 - SiO 2 (0.2%) - TiO 2 (0.2%).

SEM images of Paraloid B72+nanoparticles films were similar to the Rhodorsil
ones. The size and the surface density of the aggregates were constantly proportional to
nanoparticle concentration.
Higher magnification SEM image (Figure 2d) reveals that the nanoparticle
aggregate morphology induces a micro- and nano-scale roughness at the surface of the
films. This effect is more evident for the acrylic composite. In fact the acrylic film
creates a smooth coating that suppresses the stone roughness. On the contrary, the pure
siloxane forms a continuous film, which follows the underlying substrate with more
fidelity, conserving the stone original roughness.
Static contact angle (°)
165
155
145
135
125
115
105
Paraloid B72 Rhodorsil RC80
112° 2
157°
95
85
Pure RC80 RC80 + 2%
75
SiO2 w/v
0.0 0.5 1.0 1.5 2.0
Concentration of SiO2 nanoparticles (% w/v)
112° 11
Pure RC80 RC80 + 1% w/v SiO2
+ 0,1 w/v TiO2
0-0 0,2 - 0,2 0,1 - 1 1 - 0,1
Figure 3. Static contact angle vs. concentration of silica nanoparticles (left)
and SiO 2-TiO 2 nanomixtures (right).
Figure 3(left) shows the increase of static water contact angle (θs) as a function of
SiO2 particle concentration for Rhodorsil and Paraloid B72 composite films. In the case
of Rhodorsil treatments we obtain a significant contact angle enhancement from about
110° (no particles) to almost 160° (2% w/v particles). Paraloid B72-SiO2 values increase
from around 80° (no particles) to 135° (2% w/v particles). The addition of titania
nanoparticles gives comparable results for both kinds of investigated polymer coatings.
The graph in Figure 3(right) shows the water contact angles regarding silica and
titania nanoparticle mixtures where an increasing up to 151° for siloxane films and up to
133° for acrylic nanocomposites is visible.
These data suggest that nanoparticles induce a significant enhancement in
hydrophobicity and impart highly water repellent properties to the protective films. In
particular Paraloid B72 protective layers change their character from hydrophilic
(θs90°) surface (Shang 2005).
SEM and contact angle analyses provided a useful insight in terms of good
hydrophobicity and acceptable homogeneus surface morphology so that it was possible
to select the optimal nanoparticle concentration for each kind of nanocomposite (silica-
RC80, titania-RC80, silica and titania mixture-RC80) for subsequent analysis. Figure 4
reports the Protection Ratios obtained for marble and travertine samples treated with
siloxane and Paraloid B72 dispersions. The application of pure Rhodorsil RC80
decreases substantially the amount of water absorbed by the stone block. In fact the
151°

elative PR% are above 80% and 88% for marble and travertine substrates, respectively.
The application of nanoparticle-modified dispersions increases the protective efficiency
of the coatings. In fact the best performance is recorded for silica nanocomposites where
P.R.% reaches up to 95% (on travertine specimen). The stones treated with titania
composites rendered values similar to the pure siloxane. Silica-titania mixture shows an
intermediate behavior.
P.R. (%)
100
95
90
85
80
75
Marble Travertine
RC80 +SiO 2 (1%) +TiO 2 (0,2%) +SiO 2 (0,2%)
US velocity variation (%)
45
40
35
30
25
20
15
10
5
0
65
55
45
35
25
15
5
Marble Travertine
Paraloid +SiO 2 (1%) +TiO 2 (0,2%) +SiO 2 (0,2%)
Figure 4. Protection Ratio (%) for marble and travertine substrates treated with
Rhodorsil RC80 (left) and Paraloid B72 nanocomposites (right).
Rhodorsil RC80 Paraloid B72
+TiO 2 (0,2%)
The water absorption data obtained for Paraloid B72 dispersions once again
demonstrate the enhanced efficiency of nanoparticles for marble and travertine
protection. Indeed specimens treated with pure Paraloid B72 achieve P.R.% of order of
10% for marble and 40% for travertine, while by the application of nanocomposites
containing silica or silica-titania mixture nanoparticles the P.R.% values raise to around
20% and 60% for marble and travertine respectively.
Measurements of ultrasonic velocity provided information concerning the effect of
treatments on the mechanical properties of the stone. After the consolidation treatments
all the samples show an increase of ultrasonic velocity with respect to the untreated
specimens due to the improvement of structural cohesion and mechanical resistance of
marble and travertine substrates.
The US results show for both marbles and travertines a significant variation with
siloxane treatment, while
Paraloid B72 increase
reached only 5% (Figure 5).
This effect can be attributed
to the low penetration
capacity of the acrylic resin,
partially ascribed to its high
macromolecular dimensions,
that reduced the
consolidation effect of the
treatment by remaining in
the more external zone of the
stone samples. Higher
enhancement of cohesion from
+TiO 2 (0,2%)
Pure polymer +SiO 2 (1%) +TiO 2 (0,2%) +SiO 2 (0,2%)
+TiO 2 (0,2%)
Figure 5. Ultrasonic velocity variation (%) in marble
samples after the consolidation treatment.

the presence of silica nanoparticles results for Rhodorsil RC80 dispersions: the best
performance were recorded with the addition of the silica and titania mixture at the same
nanoparticle concentration of 0,2% w/v (about an ultrasonic velocity variation over 25%
respect to the pure siloxane application).
The fluorescence measurements have been carried out on the polymers and on the
dispersions separately, at first. LIF spectra of both of RC80 and Paraloid B72 at the
excitation wavelength of 266 nm present typical emission bands at 330 nm and 510 nm.
The presence of nanoparticles determines a variation of the bands intensity (Figure 6). In
particular, the band intensity at 510 nm increases in the presence of SiO2, while the band
disappears in the presence of TiO2. Also in the mixed nanocomposites, the intensity
increases when SiO2 concentration increases, whereas it decreases if the TiO2 quantity
increases, confirming the photoluminescence properties of the nanosilica composite, as
reported in the literature (Borsella 1997).
Figure 6. LIF spectra of silica and titania nanocomposites (1% w/v nanoparticle
concentration) compared to pure RC80 and Paraloid B72 spectra.
By the comparison of the lithotypes fluorescence spectra with those obtained from
the treated stones it can be observed that the emission bands of the substrates at 355 nm
overlap the fluorescence bands of RC80 and Paraloid, not permitting to discriminate the
nanocomposites from the substrates themselves. The excitation wavelength of 266 nm,
on the contrary, allows to observe some differences.
For Rhodorsil RC80 and Paraloid B72 treatments, the addition of SiO2
nanoparticles causes an increase of 345 nm - 366 nm and 505 nm bands respect to the
pure polymer matrix. The presence of TiO2 decreases the signal intensity, in different
bands ratios related to the acrylic and siloxane matrix, differently.
In the case of travertine the missing fluorescence of the substrate on the
wavelength range of the polymers emissions permits to exactly identify a characteristic
band of polymers at 340 nm. Also in this case, SiO2 nanocomposites increase the
fluorescence efficiency, while TiO2 decreases the signal.
Colorimetric measurements show in RC80 films a maximum total color difference
for marble surface treated with pure (no particles) polymer (ΔE=3.2; Figure 7). The
addition of silica and titania nanoparticles prevents the brightness variation and reduces
the global color difference ΔE. The travertine treated substrates undergoes a less
important optical alteration due to the lower decrease of L* (ΔE=1.5). The nanoparticle
filler is always able to reduce the total color difference ΔE.

Total color difference ΔE
5
4
3
2
1
0
Marble Travertine
RC80 +SiO 2 (1%) +TiO 2 (0,2%) +SiO 2 (0,2%)
+TiO 2 (0,2%)
2.5
2
1.5
1
0.5
0
Marble Travertine
Paraloid +SiO 2 (1%) +TiO 2 (0,2%) +SiO 2 (0,2%)
Figure 7. Total color difference ΔE for marble and travertine substrates treated with
Rhodorsil RC80 (left) and Paraloid B72 nanocomposites (right).
The Paraloid B72 nanoparticle coatings determine on marble and travertine
substrates similar slight color difference mainly caused by small changes in L* and b*
(yellow–blue chromatic component). Pure Paraloid B72 application induces on
travertine samples the most pronounced L* and b* variations. The polymer-particle
composite films reduce this effect restoring the brightness of the stone.
The ΔE is

Figure 9. LIF spectra of a marble sample
treated with a Paraloid B72 nanocomposite
before and after Solarbox (S.) aging.
Water absorption properties and
ultrasonic variation after aging processes
are agree with SEM and LIF results,
confirming a correspondence between the
surface properties and the consolidant
effect. The siloxane dispersion shows the
highest water absorption resistance (P.R. %
equal to 83%) for nanosilica composites.
The same result has been registered for
acrylic treatment, showing a P.R. equal to
17%. The major total loss of protection
effectiveness, well visible in RC80
compounds, is linked to the minor
protection capability of the degradated acrylic film. The hydrophobicity of the stone
specimens, measured through contact angle, undergoes a visible reduction in all
nanocomposite surfaces (Table 1). At the end of the degradation process nanosilica
composites nevertheless retain the greatest contact angle.
Treatment CA (°) D.S. Aging CAi (°) D.S. ΔCA(°)
ΔCA
(%)
RC80 112,4 1,2
S.
F-T
110,5
112,1
4,8
4,2
-1,9
-0,3
-1,7
-0,3
RC80 - SiO2 (1%) 146,2 2,6
S.
F-T
120,6
129,5
3,1
3,0
-25,6
-16,7
-17,5
-11,4
RC80 - TiO2 (0,2%) 131,5 3,0
S.
F-T
114,9
126,1
2,2
4,0
-16,6
-5,4
-12,6
-4,1
RC80 - SiO2 (0,2%) - TiO2
(0,2%)
148,3 8,1
S.
F-T
114,7
125,4
1,7
4,1
-33,6
-22,9
-22,7
-15,5
Paraloid B72 81,8 1,9
S.
F-T
71,2
76,1
2,2
3,2
-10,6
-5,7
-13,0
-7,0
Paraloid B72 - SiO2 (1%) 130,4 6,4
S.
F-T
115,1
114,4
3,1
2,6
-15,3
-16,0
-11,7
-12,2
Paraloid B72 - TiO2 (0,2%) 115,2 2,9
S.
F-T
105,2
110,2
1,4
1,7
-10,0
-5,0
-8,7
-4,3
Paraloid B72 - SiO2 (0,2%) -
TiO2 (0,2%)
131,8 2,6
S.
F-T
98,2
110,1
13,7
2,5
-33,6
-21,7
-25,5
-16,5
Table 1. Static contact angle before (CA) and after (CAi) Solarbox (S) and freeze-thaw (F-T)
aging and relative standard deviation (D.S.). Static contact angle variation is also reported.
The physical-mechanical processes induced by freeze-thaw cycles allow to record a
loss of conservative efficiency mainly detected by ultrasonic and water absorption data.
The ultrasonic wave velocity shows a reduction of 50% for RC80 and 35% for Paraloid
B72, without important variation related to the addition of nanoparticle filler. The loss
of mechanical consistency responsible for these results brings a decrease of the
protective capacity maintaining in any case the best performance of long-term durability
and hydrophobicity (Table 1) for the nanosilica composites. LIF measurements after the
aging processes on samples submitted to freeze-thaw cycles don’t show differences on
the spectra. No chemical variation of RC80 and Paraloid B72 is induced by this kind of
treatment.

Chromatic alteration measured in terms of ΔE is acceptable for all the specimens
submitted to both kind of weathering aging, remaining

Kaczmarek H., 1996.‘Changes to polymer morphology caused by U.V. irradiation:
2.Surface destruction of polymer blends’. Polymer, vol.37, No. 4, 547-553.
Kima E.K., Won J., Dob J., et al. 2009. ‘Effects of silica nanoparticle and GPTMS
addition on TEOS-based stone consolidants’.Journal of Cultural Heritage, 10, 214–221.
Lazzarini L., Tabasso M.L.1986. Il restauro della pietra. Cedameditore.
Manoudis P., Papadopoulou S., Karapanagiotis I. et al. 2007. ‘Polymer-Silica
nanoparticles composite films as protective coatings for stone-based monuments’.
Journal of Physics, Conference Series 61, 1361–1365.
Manoudis P.N., Karapanagiotis I., Tsakalof A., et al. 2008a.‘Super-hydrophobic
polymer/nanoparticle composites for the protection of marble monuments’. 9th
<strong>International</strong> Conference on NDT of Art, Jerusalem Israel, 25-30 May 2008.
Manoudis P.N., Karapanagiotis I., Tsakalof A., et al. 2008b. ‘Superhydrophobic
Composite Films Produced on VariousSubstrates’. Langmuir, 24, 11225-11232.
Manoudis P.N., Karapanagiotis I., Tsakalof A., et al. 2009a. ‘Fabrication of superhydrophobic
surfaces for enhanced stone protection’. Surface & CoatingsTechnology,
203, 1322–1328.
Manoudis P.N., Karapanagiotis I., Tsakalof A., et al. 2009b. ‘Superhydrophobic films
for the protection of outdoor cultural heritage assets’. Surface & Coatings Technology,
203, 1322–1328.
Melo M.J., Bracci S., Camaiti M., et al. 1999. ‘Photodegradation of acrylic resins used
in the conservation of stone’. Polymer Degradation and Stability, 66, 23-30.
Miliani C., Velo-Simpson M.L., Scherer G.W. 2007. ‘Particle-modified consolidants: A
study on the effect of particles on sol-gel properties and consolidation effectiveness’.
Journal of Cultural Heritage, 8, 1-6.
Mosquera M.J., De los Santos D.M., Montes A., et al. 2008. ‘New Nanomaterials for
Consolidating Stone’. Langmuir, 24, 2772-277.
Rodrìguez-Gordillo J., Sàez-Pérez M.P. 2006.‘Effects of thermal changes on Macael
marble: Experimental study’. Construction and Building Materials, 20, 355–365.
Shang H.M., Wang Y., Limmer S.J. et al. 2005. ‘Optically transparent superhydrophobic
silica-based films’. Thin Solid Films, 472: 37-43.
Shi H., Liu F., Yang L., et al. 2008. ‘Characterization of protective performance of
epoxy reinforced with nanometer-sized TiO2 and SiO2’. Progress in Organic Coatings,
62, 359–368.
StalderA.R., MelchiorT., Muller M., et al. 2010.‘Low-Bond Axisymmetric Drop Shape
Analysis for Surface Tensionand Contact Angle Measurements of Sessile
Drops’.Colloids Surf A,364, 72–81.
Tabasso M.L., Simon S. 2006.‘Testing methods and criteria for the selection/evaluation
of products for the conservation of porous building materials’. Reviews in conservation,
7, 67-82.
Torraca G. 1986. ‘Momenti nella storia della conservazione del marmo: metodi e
attitudini in varie epoche’. OPD restauro, numero speciale, p. 32-46.

METHODS OF POLARISING MICROSCOPY AND SEM TO ASSESS THE
PERFORMANCE OF NANO-LIME CONSOLIDANTS IN POROUS SOLIDS
Elisabeth GHAFFARI 1 , Thomas KÖBERLE 2 , Johannes WEBER 1
1 University of Applied Arts Vienna, Institute of Arts and Technology, Section of
Conservation Sciences, Salzgries 14/1, A-1013 Wien, Austria
2 Geologie-Denkmalpflege-Bauforschung, Nordstrasse 39, D-01099 Dresden, Germany
Abstract
Attempts to evaluate the efficacy and harmlessness of a consolidation treatment for
porous mineral materials have to deal with the task to measure relevant properties at
sufficient in-depth resolution. Non- to low-invasive methods, such as drill resistance or
ultrasound velocity measurements, prove useful in this context, but need eventually to
be complemented by other means of analysis which provide more precise topographic
and micromorphologic information. In view of this, the present study, performed in
frame of the EU-project STONECORE, aimed to assess some of the relevant features
related to a stone consolidant based on nano-lime by methods of microscopy. Focus is
put on the in-depth distribution and the bonding properties of the consolidant after
evaporation of the solvent, an issue of specific interest for the final result of a treatment.
The treatments tested were performed on a sieve fraction of crushed stones and
mortars from various sources. Thus, textural characteristics of the different materials
could be largely eliminated, so that emphasis was put on the precipitation of the
consolidant in dependence to the chemical and mineralogical nature of the substrates at a
given mode of treatment. The analyses included polarising microscopy (PL) on thinsections
and scanning electron microscopy (SEM). By use of these methods, the
consolidant could be well traced in the pore system of all samples, and quantitative data
by digital image analysis on the rate of pore filling at varying depth from the surface of
treatment could be calculated. The results show that, despite full penetration of all
samples, the precipitation of the consolidant was partly governed by its backward
migration towards the surface. Reducing the rate of evaporation could significantly
contribute to achieve a more even distribution. For the given mode of treatment,
substrates rich in quartz had especially high gradients of pore filling from the surfaces
inwards. No clear impact of the zeta potentials on this effect could be established.
In addition to the above, since the rate of carbonation of the consolidating lime was
another issue of interest, an approach is presented to identify, again at high spatial
resolution, the conversion of the calcium hydroxide to carbonate. Both PL and SEM
proved useful in that respect. They can be employed to replace, in a more precise and
significant way, the usual check of pH by liquid indicators. It was thus shown that
moisture plays a key role in the formation of calcium carbonate from the hydroxide.
Keywords: consolidant distribution, nano-lime, SEM, microscopy, image analysis
1. Introduction
In the course of conservation of decayed stone and mortar, the impregnating
consolidation primarily aims at bridging those contacts between single grains or grain

fragments, which had been weakened by the action of weathering. The size of those
gaps can widely vary from a few micrometers up to several millimetres. In a similar way,
the shape, location and hence the accessibility of the defects to be consolidated differ
significantly between objects and even within one object, depending on its petrographic
nature and the specific decay symptoms. This range of different conditions can only be
met with a variety of carefully selected consolidants, be it of inorganic or organic nature.
Another issue of relevance is the capacity of a product in its liquid state to penetrate
a given pore structure in a sufficient way, and to react or precipitate in the right place
with no major backwards migration of the consolidant upon evaporation of the liquid. In
that context, parameters like the molecule or particle size of the consolidant, the
properties of the solvent or the suspending agent in respect to the pore system of the
substrate are of importance. It is generally believed that small molecules or finer
particles penetrate better than coarse ones. In addition, the electric potential of all phase
surfaces involved in the consolidation must be taken into consideration, though no
precise prediction of its impact on the above parameters can be given.
Amongst all possible consolidants, inorganic systems are frequently preferred
because of their presumed better compatibility with the mineral substrates. One of the
most traditional inorganic consolidants is lime in the form of Ca-hydroxide, applied
either as a solution – lime water - or as aqueous dispersions – diluted lime wash. Several
drawbacks of these systems are due to either a low concentration of the active
component (Ettl & Wendler, 2005), or to its tendency to agglomerate to particles too
large to penetrate the pore system of the substrate. The latter problem is sometimes met
by using alcoholic dispersions of Ca-hydrate (Giorgi et al., 2000) or by using specific
dispersion devices to obtain smaller agglomerates in the range of a few micrometers
(Jägers, 2000). In order to obtain Ca-hydroxide in significantly smaller size stable in
suspensions, recent developments have focussed on nano-sized systems of lime in
organic solvents, mostly alcohols. They are synthesised in different ways and yield
particles in the size range of approx. 20 to 200 nm. (Daniele & Taglieri, 2010,
Ziegenbalg, 2011). The success of such treatments for the consolidation of stone and
mortars is varying, depending, amongst others, on the above mentioned substrate
parameters.
One group of nano-sized lime systems traded under the brand name of CaLoSiL ®
was researched in the frame of the EU-funded “STONECORE” project between 2008
and 2011. This project has dealt with the development and test application of calcium
hydroxide nano-particles with sizes in the range between 50 and 250 nm, stably
dispersed in different alcohols (ethanol, n-propanol, iso-propanol). Amongst the main
issues of interest were the penetration behaviour of the systems into the porous structure
of several natural and artificial mineral materials, as well as the evaluation of their
consolidating effect, their micromorphological aspects, and the carbonation of the
Ca(OH)2-consolidant (Ziegenbalg, 2011).
The efficacy of consolidation treatments was assessed through partly well
established, partly novel methods of measuring the bulk mechanical properties of the
stone or mortar before and after treatment, such as e.g. the drill resistance in depth and
the ultrasound transmission velocity along with other low-destructive approaches
(Valach et al., 2011, Ziegenbalg, 2011). These methods were complemented by
topography-related investigations by means of optical and electron microscopy
supplemented by X-ray microanalysis. The mechanical methods proved generally useful
yielding quantitative results, but by their nature they failed to provide clear information
about the exact depth of penetration and precipitation of the consolidant, its precise

topographic location in the substrates’ pore system, and the micromorphological
features. It was the latter which used to compliment the information on the kinetics of
carbonation obtained by measurements of the pH and X-ray diffraction.
The paper presents the most relevant tests and measurements of microscopy aimed
at the evaluation of the nano-lime consolidation effect in laboratory tests. In-situ
applications were also evaluated, they are however not included in this contribution.
Main aim is to present and discuss the approach of combined methods of light and
electron microscopy as an interesting alternative to mechanical methods for case studies
of conservation in the practice.
2. Topographic and morphologic studies of nano-lime consolidants in different
porous solids
Aim of the study was to assess the properties in terms of penetration and
precipitation of CaLoSiL ® in different substrates – stones and mortars. In that context,
the in-depth distribution and the bonding of the consolidant to the grain surfaces after
evaporation of the solvent is of specific interest for the final result of a treatment. While
measurements of the physico-mechanical parameters before and after a given treatment
yield important information on the bulk effect, the above mentioned properties can be
best assessed by methods of microscopy. In this context, Pintér et al. (2008) presented
the usefulness of a combined approach for ethyl silicate consolidants, while its
suitability to study lime-based systems was so far not well established.
The present study is focussed on detecting possible impacts of the mineralogical
and chemical composition of a variety of substrates on the performance of CaLoSiL ®
treatments. To that end, it was decided to conduct the test on a single sieve fraction of
crushed fragments of those materials. Thus, the impact of petrographically and/or decayrelated
differences in pore and grain sizes could be largely eliminated.
The distribution of the consolidant was analysed by scanning electron microscopy
(SEM) on polished sample sections, and by polarising microscopy (PL) of thin sections.
An attempt was made to interpret the results against the zeta potential of the
fractions.
2.1 Substrates
In earlier experiments it was found that penetration behaviour and distribution of
CaLoSiL ® is neither significantly controlled by particle size and corresponding pore size,
nor by the variety of available solvents, but rather by the mineralogical nature of the
different materials.
Seven rocks and four mortars were used to achieve a broad variety of substrates.
They are listed in Table 1.
2.2 Sample preparation and laboratory treatment
The rocks and mortars were reduced to small pieces in a laboratory crusher. A
fraction of 0.09 to 0.160 mm was obtained by dry sieving. The sands were filled into
plastic cylinders placed on air-permeable disks (glass-frits), as illustrated in Figure 1. No
further compaction of the sands was accomplished before treatment. The amount of sand
was determined gravimetrically. The samples were then treated with CaLoSiL ® E25, a
calcium hydroxide suspension in ethanol with an average particle size of approx. 150
nm, at a concentration of 25 g/L. The treatment was performed by dropping the product
onto the top surface of each sample until full penetration to the bottom was observed.
No Material Short description Main

1 Drachenfels trachyte volcanic rock with mainly sanidine in a dense microcrystalline
groundmass
silicates
2 Römer tuff volcanic tuff, mainly of feldspar and lithic fragments in an
altered, formerly vitreous groundmass
silicates
3 Leuben mortar historic dolomitic lime mortar with quartz aggregates quartz + Mgcarbonates
4 Lab-made lime
mortar
5 Lab-made dolomitic
lime mortar
Table 1. Substrate materials tested in this study
weakly bound lime mortar with quartz aggregates quartz + Cacarbonate
weakly bound dolomitic lime mortar with quartz
aggregates
6 Greek limestone compact limestone (grainstone) with fossil and lithic
fragments
7 Maastricht limestone fine-grained, highly porous biocalcarenite consisting of
uniform lime particles
Samples were freely left to dry under laboratory
conditions - only in the case of one sample from the
“Greek limestone” was the surface covered to reduce
the rate of evaporation. The treatment was repeated
several times wet-in-wet, until the fractions appeared
saturated. After 24 hrs of curing under laboratory
conditions, the treatment procedure was repeated. The
experiment was finished after another 72 hrs. Then the
sections were prepared for microscopic observation.
2.3 Section preparation and methods of analysis
The whole compounds including the frit supports
quartz + Mgcarbonates
Ca-carbonate
Ca-carbonate
8 Sterzing marble coarse-grained white crystalline marble Ca-carbonate
9 Kremersand Mori commercial non-natural mixture of aggregates from
yellow marble and fossil limestone
Ca-carbonate
10 Aachen marl marl, a clayey, mainly micritic carbonate rock Ca-carbonate +
silicates
11 Dahlen stucco historic white gypsum plaster without aggregate gypsum
Figure 1. Illustration of sample
treatment
were vacuum-embedded in a blue dyed epoxy resin (Araldite ® 2020). Petrographic thin
sections of about 25 µm thickness as well as polished sections were produced
perpendicular to the surface of treatment.
The thin-sections were observed by PL (Jenavert), photographs were taken with a
microscope camera (Leica DFC290). The polished cross sections were coated with
carbon and studied by SEM (Philips XL 30 ESEM, 20 KV, high vacuum, back-scattered
electron detector-BSE), equipped with an energy-dispersive X-ray analyser (Link-ISIS).
The SEM-micrographs, taken at low magnification of 100 times for the full depth of a
sample, were assembled to composite images by use of a Photoshop ® software. Pores,
aggregates and consolidant were edited in different pseudocolours to ease their visibility
and to allow for digital image calculation. The latter was performed using a Leica QWin
Plus ® software.
Exemplary SEM micrographs are shown in Figures 2 and 3.

Figure 2. Consolidant from CaLoSiL ®
treatment, filling the pore systems
Figure 3. Consolidant from CaLoSiL ®
treatment, forming bridges between grains
2.4 Topography and morphology: results and conclusions
By SEM/BSE as well as by PL techniques, the consolidant revealed well visible in
all substrate materials. Given its higher resolution, SEM/BSE additionally allows for a
detailed characterisation of crystal shape and orientation for each of the studied
substrates.
In view of the above, pseudocolour editing of the SEM-micrographs proved a
practicable, though time consuming task. In such way, the basis for a number of useful
digital image calculations could be laid, enabling to assess the specific performance of
the consolidant for a given material in a quantitative way.
The decision to use loose aggregates rather than compact material for the test
treatments afforded a representative comparison between mineralogically different
specimens, and compensated the lack of real pore geometries.
It was shown that the liquid consolidant had penetrated through the full depth of all
samples. In contrary to the materials in their compact state which sometimes show
difficulties of impregnation with CaLoSiL ® , the penetration of the liquid into the loose
substrates was no limiting factor at all.
The decisive factor was found to be the precipitation of lime on top and in
subsurface areas of both surfaces of the specimens. This caused an uneven distribution
of CaLoSiL ® through the depth of impregnation for virtually all samples. The resulting
gradient, an important factor in context with a successful and harmless consolidation
treatment, is more pronounced for materials containing high amounts of quartz (samples
No. 3 and 5) than for the rest of the specimens. Another effect is due to the presence of
smaller grains in some samples, an unexpected technical consequence of dry sieving.
The zeta potential of the substrates, measured for their granulations in ethanol and
in water, respectively, revealed of no clear significance for the success of a treatment,
neither in terms of total average pore filling nor in respect to the gradients of
precipitation. It must be mentioned in this context that even no clear dependence on the
mineral composition and their zeta potential could be detected.
For the case of the “Greek limestone”, the different conditions of curing – free
evaporation of the solvent (sample No. 6a) vs. prolonged curing by covering the sample
surface (sample No. 6b) – proved of high significance for the development of gradients:
the latter sample showed no increased concentration of the consolidant at or near the
treated surface. This is an indication that, at least in some cases, the frequently observed
formation of a white haze on treated surfaces, as e.g. reported by Dähne (2011), is
caused by migration and should be prevented by suitable measures of protection.

Figure 4
Figure 5a Figure 8a
Figure 5b Figure 8b
Figure 5c Figure 8c
Figure 6
Figure 7
Figure 9
Figure 4. SEM-BSE image of sample 3, full sample depth with frit support on the bottom
Figures 5a-c. Details of Figure 4. (5a) top of sample, (5b) central area, (5c) bottom area
Figure 6. Distribution of consolidant in aggregate samples of historic lime mortar, sample 3. Bold
line refers to this sample
Figure 7. SEM-BSE image of sample 6a, full sample depth with frit support on the bottom
Figures 8a-c. Details of Figure 7. (8a) top of sample, (8b) central area, (8c) bottom area
Figure 9. Distribution of consolidant in aggregate samples of “Greek limestone”, uncovered
(sample 6a). Bold line refers to this sample

Table 2. Performance of the nano-lime consolidant for the different substrates
No.
Substrate
material
1 Drachenfels
trachyte
Zeta potential
of substrate
(mV)*
in water
in ethanol
total porosiy (area-%)
in depth distribution
of consolidant
Agglomeration of consolidant Degree of pore filling
by consolidant
on treated surface
(white haze)
in subsurface
area
in bottom area
on bottom
surface
avg. across full
depth (area-%)
in subsurface
area
factor of surface
excess
consolidation
-23,7 -2,6 55,38 good +++ + + ++ 16 24 1,55
2 Römer tuff -22,24 -4 52,35 good + - - ++ 10 13 1,29
3 Leuben
mortar
4 Lab-made
lime mortar
5 Lab-made
dolomitic
lime mortar
6a Greek
limestone
6b -- “ --
covered
7 Maastricht
limestone
8 Sterzing
marble
9 Kremersand
Mori
10 Aachen
marl
11 Dahlen
stucco
-17,35 3,8 48,43 poor ++ +++ ++ ++ 7 55 7,79
-17,85 3,05 48,18 fair ++ + + ++ 12 21 1,77
-7,1 7,95 48,80 poor + +++ ++ ++ 7 41 5,44
-18,2 -1,05
54,78 good + ++ - + 22 41 1,83
49,65 fair - - +++ +++ 20 14 0,67
-16,45 3,75 53,59 fair + - ++ ++ 10 15 1,49
-9,15 5,75 50,09 fair + + - + 13 28 2,14
-14,25 2,05 54,23 good ++ +++ - + 32 72 2,27
-19,2 -1 56,50 good +++ + - - 35 62 1,77
-5,15 -3,9 53,44 good +++ +++ - + 24 56 2,34
* The values for the zeta potential of the substrate materials were supplied by IBZ Freiberg. CaLoSiL ® E 25 has
a zeta potential of 38 mV (IBZ Freiberg)
3. Carbonation
Apart from the visualisation of the presence of the consolidant, a big advantage of
both SEM and PLM lies in the possibility to evaluate the state of carbonation of the
consolidant in a more precise way than measuring the pH which is usually carried out by
the phenolphtalein test: by means of SEM, the micro-structure of the consolidant gives a
clear indication of the presence of carbonate, while by PLM this is proved through its
optical properties.
The distribution of the consolidant in the pore structure of the treated material was
evaluated by the two most common microscopic methods in material research, i.e. the
high-resolution scanning electron microscope (SEM) and the petrographic polarising
microscope (PLM).

In order to find out the limits and advantages of the two different methods for the
detection of CaLoSiL ® , a special survey was carried out.
Polished thin sections of differently treated materials were carefully investigated.
They were produced with a water-free lubricant to avoid sample disturbance by water
and carbon dioxide reactions. This type of preparation is very convenient for both
microscope procedures. By use of PL and SEM, a defined sample area was examined,
and the presence of the consolidant was documented by micrographs. These records
could then be compared with each other.
Both methods show reliable results, as shown in Figures 10, 11 and 12. The
consolidant is detectable in more than 95 % of the cases with both microscopic
approaches. For the remaining samples, sometimes SEM seams to yield more significant
results, sometimes PL is advantageous.
Apart from the visualisation of the presence of the consolidant, a big advantage of
both SEM and PLM lies in the possibility to evaluate the state of carbonation of the
consolidant in a more precise way than measuring the pH: by SEM, the micro-structure
of the consolidant gives a clear indication of the presence of carbonate, while by PLM
this is proved through its optical properties.
CaLoSiL ® , which is deposited in the pores by evaporation of the alcohol, consists of
Ca(OH)2. The birefringence (Δ) of portlandite appears moderate, with a value of 0,029
(Tröger, 1967), but due to the small particle size the interference colour show a first
order grey.
The particles than convert in CaCO3 by carbonation, that means CO2 uptake in
presence of small amounts of water. The precipitated calcite show a much higher
birefringence with Δ = 0,1719 (Kordes, 1960). Even in the smallest areas of the section,
submicron-sized particles can be clearly observed. The increase in interference colours
due to the formation of the carbonate makes these particles visually stand out.
In Figure 11, the non carbonated area of CaLoSiL ® can be seen (area b1, compare
to area b in Figure 10), while the outermost layer towards the pore appears carbonated
(b2 in Figure 11).
Figure 10. PLM photograph of
CaLoSiL ® (area b) in a loose
sand of carbonate rock
fragments (a). C is a pore filled
with blue dyed resin. Plane
polar light
Figure 11. PLM photograph of
carbonated CaLoSiL ® (b2) and
not carbonated CaLoSiL ® (b1)
in crossed polar light
Figure 12. Same area as in
Fig. 10 and 11, seen by SEM

4. Discussion and conclusions
The approach to use the SEM to detect consolidants in the pore system of solids is
not novel, though it is not frequently used in conservation studies. A prerequisite to
produce photographs suitable for further digital image analysis is a distinct grey value of
the consolidant in the SEM-BSE image. Whenever this applies, a number of
micrographs must usually be produced, in order to cover reasonable areas of a sample
section if one wants to study the full depth of penetration. The effort is considerable, but
indispensable to fully understand the in-depth distribution of the consolidant on test
samples, an important factor which has produced many unsatisfactory or even harmful
results in the practice of conservation.
An alternative approach can be seen in polarising microscopy on thin sections. This
method, well known by geologists and scientists from related fields, needs not only
optimal section preparation, but also relatively high skills in using the microscope in the
best possible way to recognise structures and phases, and to interpret observations. Its
advantage lies in the low magnification, allowing to study and record larger areas of a
sample section when compared to SEM. Disadvantages of PL are the limits of resolution,
and the fact that micrographs suitable for digital image analysis can be produced just in
rare cases. Thus, PL can be viable to check for the presence of a consolidant, however
without quantifying its amount within certain areas of the section.
It may be advisable to use polished thin sections which can be studied by both
methods, PL and SEM, in order to achieve the best possible results.
When it comes to the assessment of the progress of carbonation of consolidants
based on calcium hydrate, however, both tools are equally useful.
The selected way of specimen preparation for the tests presented in this contribution,
i.e. by using a sieve fraction of crushed stone material, limits the significance of the
observations to some extent. Nevertheless, some conclusions useful for the application
of consolidants in the practice can be drawn. Thus, it was shown that the most critical
step in the course of impregnation of an open porous solid may be the instance when the
liquid phase evaporates, which might cause a backwards migration of the precipitate.
Reducing the rate of evaporation can greatly help minimise this effect, thus preventing
the formation of over-consolidated subsurface areas or even surface layers.
It was shown that, for a given application procedure of a consolidant to a specfic
pore system, as experimented in this study, the mineral nature of the solid seems to be of
relevance to the above effect of migration – the backwards migration proofed most
pronounced for substrates rich in quartz. The assumption that the zeta potential of the
solid would yield a measure of significance, however, was not supported by the data.
One could rather assume that the surface morphology of the grains – naturally differing
for different types of minerals, especially when they underwent crushing – might be a
decisive factor for the degree of “trapping” the precipitate by chemo-physical forces.
More investigations are needed to support this hypothesis.
Finally, it can be stated that the nano-sized lime used in the study can yield a good
consolidant for mineral materials, provided that care is taken for its proper mode of
application. The amount of shrinkage of the precipitate in the pore space is limited,
especially when compared e.g. to most ethyl silicate systems. The slow rate of
carbonation in the absence of moisture is probably due to the dense packing of the nano
particles. This needs not be considered a significant risk, however, since measurements
of mechanical strength have proofed the efficacy of the consolidant even in the
hydroxide state. Whenever moisture would enter the pores, it will result in the
immediate, isotopic crystallisation of calcium carbonate.

Acknowledgements
The present study was financially supported by the European Commission through
its 7 th FP research project EU-213651 STONECORE. Thanks are due to the project
partners who supplied the materials and provided fruitful discussions.
5. References
Dähne, A. 2011. ‘Evaluation of lime nano-sols for conservation of wall paintings,
mortars and stucco’, EWCHP-2011, Proceedings of the European Workshop on Cultural
Heritage Preservation, Fraunhofer IRB Verlag, 262 – 266
Daniele, V., Taglieri, G. 2010. ‘Nanolime suspensions applied on natural lithotypes: The
influence of concentration and residual water content on carbonatation process and on
treatment effectiveness’, Journal of Cultural Heritage, 11, 102 - 106
Ettl, H., Wendler, E. 2005. ’Strukturelle Putzfestigung mit Kalkwasser? Grenzen und
Alternativen’- Beiträge zur Erhaltung von Kunst- und Kulturgut, 1, 129 – 133
Giorgi, R., Dei, L., Baglioni, P. 2000. ‘A new method for consolidating wall paintings
based on dispersions of lime in alcohol’, Studies in Conservation, 45, 154
Jägers, E. 2000. Dispergiertes Weisskalkhydrat für die Restaurierung und
Denkmalpflege: Altes Bindemittel - Neue Möglichkeiten, Michael Imhof Verlag,
Petersberg, 140
Kordes, E. 1960. Optische Daten zur Bestimmung anorganischer Substanzen mit dem
Polarisationsmikroskop, Verlag Chemie GmbH, Weinheim, 54
Pintér, F., Weber, J., Bajnóczi, B. 2008. ‘Visualisation of Solid Consolidants in Pore
Space of Porous Limestone Using Microscopic Method’, Proceedings of 11th
international congress on deterioration and conservation of stone, Torun, Poland, 15 –
20 September 2008, Volume I, edited by Jadwiga W. Lukaszewicz, Piotr Niemcewicz,
473 - 480
Tröger, W. E. 1967. Optische Bestimmung der gesteinsbildenden Minerale, Teil 2
Textband, Schweizerbart´sche Verlagsbuchhandlung, Stuttgart, 69
Valach, J., Hasníková, H., Dobrzynska-Musiela, M. et al. 2011. ‘Influence of material
and consolidant type on aggregate strengthening’, EWCHP-2011, Proceedings of the
European Workshop on Cultural Heritage Preservation, Berlin 2011, Fraunhofer IRB
Verlag, 267 – 272
Ziegenbalg, G. 2011. ‘Stone conservation for refurbishment of buildings (STONECORE)
- A project funded in the 7 th framework programme of the European Union’, EWCHP-
2011, Proceedings of the European Workshop on Cultural Heritage Preservation, Berlin
2011, Fraunhofer IRB Verlag, 240 – 245

CONSOLIDATION OF POROUS LIMESTONE WITH NANOLIME
LABORATORY STUDY
Zuzana Slížková 1 and Dita Frankeová 1
1 Institute of Theoretical and Applied Mechanics, Academy of Sciences of the CR
slizkova@itam.cas.cz frankeova@itam.cas.cz
Abstract
Effects of a lower and a higher concentrated nanolime product CaLoSiL after its
multiple applications on porous limestone have been studied and presented in the paper.
The assessment of the consolidating efficiency of the products was based on
performance of both the destructive and the non-destructive laboratory tests: bending
and compressive strengths, ultrasound velocity measurements and porosity
determination of stone specimens. Concluding the test results performed for the depth
profiles of specimens, two applications of the CaLoSiL E 25 or IP 25 products seem to
be an optimal procedure for treatment of the studied porous limestone. Further
applications of the product especially in the case of the higher concentrated CaLoSiL E
50 resulted in an uneven distribution of the measured characteristics (strength and
porosity) due to an accumulation of the consolidating product near the surface. The
appropriate amount of the product (both the concentration of the effective compound
and the sum of the product applications) should be chosen in relation to the stone
porosity and the treatment condition in order to achieve a regular distribution of the
product in the substrate.
Keywords: nano-lime, calcium hydroxide, consolidation, strengthening, CaLoSiL, stone
1. Introduction
The recent outburst of nanotechnologies has influenced the conservation science
and has brought nanoparticles of calcium hydroxide dispersed in alcohols as a new form
of treatment product based on Ca(OH)2. Nanolime products are an attractive choice
mainly for the consolidation of substrates containing calcite considering the intended
chemical compatibility between the treated substrate and the treatment product.
CaLoSiL® (introduced in October 2006) has been the first commercially available stone
treatment product based on calcium hydroxide sol (Ziegenbalg 2008; Drdácký et al.
2009). The efficiency assessment and other research of this product in the field of
building materials treatment consolidation were carried out within the European project
STONECORE (2009-2011). Another commercial product Nanorestore® was developed
at the University of Florence (CSGI Consortium) and has been also studied by
conservation scientists (e.g. López-Arce et al. 2010). Consolidation action depends on
the products' characteristics but also on procedures adopted for their application
(Ferreira Pinto et al. 2008). This paper presents the influence of the repeated application
(1 to 6) of CaLoSiL products of the lower and the higher concentration of Ca(OH)2 on
the stone with high and large porosity.

2. Research aim
The treatment product CaLoSiL® is available in various concentration of Ca(OH)2
from 15 to 50 g/l in different alcohols. However this product brings a much more
effective alternative in relation to lime water (Drdácký and Slížková 2008), the
concentration of the effective substance in the product is low when compared to the
products based on silicic acid ester (100-500g/l). Multiple applications of the CaLoSiL
product can be considered when a higher consolidation effect is required (Drdácký and
Slížková 2011). This can be achieved either by multiple application of a low
concentrated CaLoSiL or using a high concentrated product. The presented research
studied differences in the both approaches when applied on a porous limestone. The
assessment of the consolidating efficiency of the products was based on performance of
both the destructive and the non-destructive laboratory tests: bending and compressive
strengths, ultrasound velocity measurement and porosity determination. Except of the
compressive strength the other characteristics were investigated in the depth profile of
stone specimens in order to find the distribution of the treatment product within the
substrate. The testing aimed at a quantitative evaluation of the product strengthening
effect and at a recommendation of an optimal treatment procedure for the selected type
of stone and treatment condition (dry condition T 20-25°C, RH 30-40 %).
3. Materials and Methods
3.1. Treatment products
Three variants of the commercial product CaLoSiL were used in the present study.
The products contain Ca(OH)2 particles dispersed in different alcohols (ethanol and
isopropyl alcohol types were investigated). The particles size ranges from 50 to 150 nm.
The producer of the agent is IBZ-Salzchemie GmbH & Co. KG, Germany. In the study
an ethanol based type with two different concentrations of dispersed lime particles
(25g/L, 50g/L respectively) and an isopropyl alcohol based type with concentration 25
g/L were examined. Trade names of the selected types are CaLoSiL E 25, CaLoSiL E 50
and CaLoSiL IP 25. The behaviour of differently concentrated CaLoSiL and the
influence of a different alcohol medium were studied at consolidation of a highly porous
limestone aiming improvement of material (calcite crystals) cohesion.
3.2. Stone
The laboratory experiment was carried out on the Maastricht limestone (Table 1),
which is a material with low mechanical characteristics, a high open porosity and large
pore sizes. These stone characteristics seemed to be suitable for such a consolidation
study focused on the stone strength enhancement. Upper Cretaceous Maastricht
limestone ("mergel") outcrops in the provinces of Dutch and Belgian Limburg.
Formation represents one of the few native Dutch natural stones used for building and
construction. The Maastricht limestone is very homogeneous, and layering is rarely
observed. The only location where the Maastricht limestone can be quarried today is in
Sibbe, Holland. Under the optical microscope (Rescic et al. 2010) the Sibbe variety
shows a good sorting with grain dimension of about 100µm. The grains, subangular in
shape, are constituted mainly by sparitic calcite (shell fragments and skeletons of sea
organisms) and secondarily by micritic calcite. There are also rare silicatic grains. The
binder is scarce and constituted by sparitic calcite. The porosity is high (≈ 50%) and

mainly constituted by macroporosity (most of all pores have a diameter ≈48µm). The
petrographic classification is grainstone and intrasparite. The decay of historic stone
develops mainly through the detachment of the crust.
Table 1 Physical characteristics of Maastricht limestone
Stone
Maastricht
limestone
Water
absorption
by immersion
Water absorption
coefficient
Bulk
density
Open
porosity
Main pore
diameter
% wt. kg/m 2 min 1/2 g/cm 3
% vol. µm
34,4 20,2 1,3 50 48
4. Application procedure
Stone specimens were conditioned in the laboratory before the experiment. The
treatment was performed on 5x5x3cm 3 prismatic specimens and only one face of the
specimen was treated with a certain volume of the product (30ml). The product was
applied by syringe pouring the stone surface until the total volume of 30 ml of the
product has been soaked. The volume (30ml) was chosen as a result of a preliminary test
when the amount of the product needed for the wetting of 75cm 3 (5x5x3cm 3 ) stone
specimen had been determined. The selected volume (30ml) does not refer to the
volume needed for full saturation of the stone but for its impregnation by capillary
absorption when the front of the product (visible by darkening of the impregnated stone
material) reaches the bottom of the stone specimen. The time needed for the first and for
the successive applications (2-6) was measured.
Specimens for the bending test, US and porosity investigated in depth profile were
treated on the face 5x3cm 2 . Since the multiple application of the products was studied,
two specimens were impregnated twice, another two specimens four times and the last
two specimens six times. The break between every single impregnation represented 4
days. This time was sufficient for drying out of impregnated stone under the laboratory
condition which had been found on the basis of an individual preliminary test. Stone
specimens were not covered by any impermeable coating which should control
evaporation of alcohol from the impregnated material and so the drying process of the
stone went through all specimen faces and was relatively fast thanks to alcohol medium
of the nanolime and also to a high and large stone porosity. Stone specimens were able
to absorb 30ml of the nanolime product even in the case of the sixth application but the
time needed for the repeated impregnation increased for some types of the CaLoSiL
product.
Specimens for the compressive test had the same parameters (5x5x3cm 3 ). These
specimens were impregnated by the same manner like the previous specimens with one
difference that the product was applied through the face 5x5 cm 2 to the depth of 3 cm.
The maturing of impregnated stone specimens occurred in dry conditions
(laboratory environment, 20-25 °C, 30-40 % RH) without any special care. The testing
of impregnated specimens started one month after the consolidation treatment.
The stone specimens were analysed with several techniques in the following order:
propagation of US velocity, bending test and open porosity distribution by MIP - all in
the depth profile, and water accessible porosity and compressive strength.

Results from the performed tests are expressed as an average value gained from
individual measurements of two specimens treated with the same procedure.
5. Methods
The US velocity was measured with a portable instrument USG 20 (Krompholz
Geotron Elektronik, FRG) with a 250 KHz transmitter (USG-T) and receiver (USE-T).
The measurement was taken in a direct transmission/reception mode, across opposite
parallel sides of the specimen. In a first step, the untreated stone specimens were
measured in all three spatial directions (one measurement for the one axis) so that the
difference between X, Y and Z axes was found for each specimen and an unified
geometric orientation for all specimens could be set before the treatment. Only one
geometric plane was selected for a more detailed study. Then the depth profile of
ultrasound velocities was measured with a step of 7 mm in the direction from the treated
top surface to the bottom (Figure 1.a). The measurement points were marked on the
stone specimens and the comparative test after the specimen treatment could be taken
using the same points and tracks. The diameter of the flat contact area on the transducer
was 2 mm.
The bending strength values in the depth profiles of not treated and treated
specimens were tested on thin plates. The 50 mm depth was cut to slices and so 8
rectangular thin plates of the depth 3,7mm±3mm with the top side parallel to surface
were created (Figure 1.b). Slides were marked 1 to 8 and their specific positions relating
the original surface were registered. Bending strength and the Young’s modulus of
elasticity were then tested on these small-size partial specimens, which resulted in the
knowledge of strength values in relation to the location of a stone layer in the original
stone specimen. This way of testing enabled to determine the strength values
distribution along a depth profile perpendicular to the surface. The specimens were
tested in three point bending in a special rig after long term conditioning in laboratory
(20-25°C /RH 30-40%) using load cell Lucas 100 kN for the load measurement, LVDT
1 mm sensor for the deflection measurement, at the cross head speed of 0,15mm/min.
Even tough the destructive character of this method does not allow the performance
of these tests before and after the treatment on identical specimens, a quite good
homogeneity of stone along the selected axis enabled the evaluation of the strengthening
effect by means of comparing the strength values obtained on different not treated
specimens and different treated specimens of the same type of stone.
Figure 1. a (left) Layout of the US velocity measurements from top to bottom, b (right) Layout
of a specimen cutting to slices for detailed bending strength tests.

The open porosity and its distribution in the studied materials were determined by
Mercury intrusion porosity, using a Quantachrome porosimeter, model Poremaster PM-
60-13 with pressure range of 0,005 – 413 MPa. The mercury parameters were set to
values of 480 erg/cm 2 for the surface tension of mercury and 140 degrees for the contact
angle. The samples were dried out before the test and a penetrometer of 1 cm 3 was used
for the measurement. After performing the bending test the broken stone slices were
used for porosity characterization.
The hydric behaviour through the stone was studied on specimens 50x50x30 mm 3
(treated surface 50x50mm 2 ). The determination of the porosity accessible to water was
carried out following RILEM recommendation (25-PEM norm 1980). The compressive
strength was tested on the same specimens like hydric characteristics using load gauge
Lucas 50 kN, sensor Megatron 25mm, load speed 0,45mm/min.
6. Results and Discussion
The time needed for the first and for the each successive application (2-6) indicates
a possible change of transport characteristics of the CaLoSiL within the stone in case of
the repeated applications. The best results were obtained for the CaLoSiL E 25 which
needed roughly the same time (around 300 s) to penetrate the specimen from the top to
the bottom for the first and for the following applications. Simple measurement showed
that even in case of the 6 th application the penetration coefficient of this product did not
change significantly. Figure 2 illustrates that a different behaviour was found for
CaLoSil IP 25 and E 50. After the first three applications the time needed for specimens
impregnation increased as the product penetration ability got worse especially in case of
the highly concentrated CaLoSiL E 50. The accumulation of these products in some
parts of the stone specimen is probable the reason for the rise of the penetration time in
cases of the 4 th to the 6 th application.
Figure 2. Times needed for wetting of a stone specimen with 30 ml of CaLoSiL in relation to the
successive number of application.
The determination of the ultrasonic (US) velocity profiles seems to be a suitable
tool for the estimation how an even or uneven distribution of the product's effective
compound (calcium carbonate in this case) has developed within the stone specimen
after the treatment. The US velocity in the untreated specimens was 1.7-1.8 km/s and the
profile was quite homogenous (Figure 3). Most of the treated specimens showed an

increased velocity in the surface area (top and bottom), probably as a result of the
cumulation of the consolidation product during the evaporation period.
The increased US velocity in the bottom region gives an evidence that products
penetrated the all depth of the specimen (5 cm). The rise of US velocity due to the
treatment represents about 5 % of the original value in a central part of a specimen for
the lower concentration and about 13% for the higher concentration after 2 applications.
In the surface area the increase of velocity was 2-3 times higher compared to the central
part of the stone. The consolidating effect (US velocity) increased more after successive
applications (4-6) but more intensively in the area near the surface. It seems from the US
velocity measurements that multiple applications of the CaLoSiL products lead to
cumulation of the product within the surface region and this trend is more significant in
case of the higher concentrated studied type - CaLoSiL E 50. The optimal procedure for
consolidation of the investigated stone seems to be 2 applications of the lower
concentrated types - CaLoSiL E 25 or IP 25. Different behavior of the products E 25 and
IP 25 has not been found in case of the US velocity tests: the differences between the
velocity profiles are quite low for the both products with a different alcohol medium.
The US velocity corresponds with the stone modulus of elasticity and characterizes
its change resulting from the consolidating treatment very sensitively. Bending strength
profiles (Figure 4) brought similar results but in detail these data have a higher scatter,
which is probably influenced by local defects. Considering the distribution of the
strength in the profile the better results were found on stone consolidated by 2
applications only. The higher number of applications went to an uneven distribution of
the product in the stone which was followed by an increase of the strength in the surface
part.
The strength rise considering overall stone specimens is illustrated in Figure 5 for
both the bending and the compressive strengths. The distribution of the strength in the
stone specimen profile is not considered in these graphs: the average values of the
specimen strength are shown (for the bending the average is calculated from individual
slices strength of 2 specimens and the compressive strengths were determined by testing
2 specimens 5x5x3cm 3 ). The increase of the average bending strength value of a
specimen represented 45% after 6 applications for E 25, 63% for IP 25, and 100% for E
50. The rise of the compressive strength was 93% for E25, 47% for IP 25, and 126 % for
E 50 after 6 applications. After 2 applications (which seems to be a suitable treatment
amount relating the even distribution of the product in the stone) the bending strength
increase represented 20% (E 25), 18% (IP 25) resp. 54% (E 50) and the compressive
strength increase 50 % (E 25), 23% (IP 25) resp. 73% (E 50). The presented test results
support conclusions that two applications of the CaLoSiL products seem to be the
optimal procedure for the treatment of porous limestone. When a stronger strengthening
is needed the higher concentrated product E 50 applied with care may be useful.
The porosity investigations correspond with previous results. The values of the
open porosity obtained on specimen 5x5x3cm 3 were quite optimistic: the porosity
decreased due to the treatment (table 2) but not dramatically. MIP method used for the
testing of porosity in the depth profile showed its uneven distribution coming from a
bigger influence on pores near the surface (Figure 6) even after 2 applications.
Concerned the stone color the local white hazes appeared on the surface of some
specimens (IP 25 and E 50 after 6 applications of the products). White staining was not
found for specimens treated with the CaLoSiL E 25.

CaLoSiL E 25
Figure 3. Ultrasound velocity in depth profile for the Maastricht limestone before and after
treatment with CaLoSiL IP 25, CaLoSiL E25 and CaLoSiL E 50.

Figure 4. Bending strength of stone slices in the depth profiles for not treated specimens and the
specimens treated with CaLoSiL E 25, CaLoSiL IP 25 and CaLoSiL E 50.

Figure 5. Increase of bending strengths (left graph) and compressive strengths (right graph) of the
Maastricht limestone due to consolidation with multiply applied the CaLoSiL products.
Table 2. Reduction of water accessible porosity of the Maastricht limestone due to consolidation
with 6 applications of the CaLoSiL products
Maastricht limestone
Water accessible porosity
(RILEM I/1)
Decrease of porosity
% % abs. % rel.
untreated 50,3
treated by 6 cycles E 25 47,4 2,9 5,8
treated by 6 cycles IP 25 47,4 2,9 5,8
Figure 6. Distribution of open porosity values (MIP) within the treated specimen depth profile
(CaLoSiL E 25, 2 applications).
7. Conclusion
The appropriate amount of the product (both the concentration of the effective
compound and the sum of applications) has to be chosen in relation to the stone porosity

and the treatment condition in order to achieve the regular distribution of the product in
the substrate. The optimal procedure for consolidation of investigated stone seems to be
2 applications of the lower concentrated types CaLoSiL E 25 or IP 25. Application of
the products with a different alcohol medium (CaLoSiL E 25 and IP 25) generates rather
low observed differences measured by the US velocity and the strength in the depth
profiles. The higher sum of performed applications resulted in an uneven distribution of
the product in the stone specimen which was followed by the increase of the strength in
the surface part. The local white hazes appeared on the surface of some stone specimens
after 6 applications of the products IP 25 and E 50.
8. Acknowledgement
The authors acknowledge the support of Ministry of Culture of the Czech Republic
under the project NAKI DF11P01OVV012 entitled “New materials and technologies for
conservation of monuments´ materials and preventive conservation”. The authors are
grateful to Dr. Claire Moreau who carried out the measurement of water accessible
porosity and prof. Miloš Drdácký for helping with mechanical tests. Special thanks to
IBZ –Salzchemie GmbH & Co.KG, Freiberg, Germany for supplying the consolidating
products CaLoSiL.
9. References
www.stonecore-europe.eu
Drdácký M., Slížková Z., Ziegenbalg G. 2009. ”A Nano Approach to Consolidation
of degraded Historic Lime Mortars”. Journal of Nano Research, 8 : 13-22.
Drdácký M., Slížková Z. 2008. ”Calcium hydroxide based consolidation of lime
mortars and stone”, in Delgado Rodrigues J., Mimoso JM (editors) Proceedings of the
<strong>International</strong> symposium Stone consolidation in culture heritage, Lisbon, pp.1109-15.
Drdácký M. Slížková Z. 2012. ”Lime Water Consolidation Effects on Poor Lime
Mortars”. APTI Bulletin, XLIII (1): 31-36.
Ferreira Pinto A.P., Delgado Rodrigues J. 2008. “Stone consolidation: The role of
treatment procedures”. Journal of Culture Heritage, 9 (1): 38-53.
López-Arce P., Gomez-Villlalba L.S., Pinho L., Fernández-Valle M.E., Álvarez de
Buergo M., Fort R. 2010. “Influence of porosity and relative humidity on consolidation
of dolostone with calcium hydroxide nanoparticles: Effectiveness assessment with nondestructive
techniques”. Materials Characterization, 61 : 168-184.
RILEM Commission 25-PEM, Tentative Recommendations, 1980. Materials and
Structures, 13 (75).
Rescic S., Fratini F., Tiano P. 2010. "On-site evaluation of the mechanical
properties of Maastricht limestone and their relationship with the physical
characteristics", in Smith B.J., Gomez-Heras M., Viles H:, & Cassar J. (eds.)
"Limestone in the built environment: Present-day Challenges for the Preservation of the
Past", Geological Society, London, Special publication, (331): 203-208.
Ziegenbalg G. 2008. “Colloidal calcium hydroxide – a new material for
consolidation and conservation of carbonatic stones”, in Lukaszewicz J., Niemcewicz,
P. (eds.) Proceedings of the 11th <strong>International</strong> <strong>Congress</strong> on deterioration and
conservation of stone, Poland. Nicolaus Copernicus University Press, pp. 1109-1115.

AN EXPERIMENTAL ASSESSMENT OF 3 INORGANIC CONSOLIDANTS FOR
USE ON MUSEUM ARTEFACTS IN COMPARISON TO ORGANO SILANES
Jennifer Booth 1,2 jennifer.booth@ouce.ox.ac.uk, Heather Viles 1 , Philip Fletcher 2
1 School of Geography and the Environment, University of Oxford, 2 Department of
Conservation and Scientific Research, The British Museum
Abstract
Organo silanes are consolidants that have been praised for their effectiveness on
stone both in the internal and external environment. Their success on limestone however
has been debated and there are also significant drawbacks with the health and safety
requirements for working with Silanes. Inorganic consolidants can address compatibility
issues with limestone, and are often safer to use and dispose of. In the past, inorganic
consolidants have been considered less effective to their organic counterparts. With the
advent of new technology and the development of nano-limes, there has been a
resurgence in interest in inorganic consolidants.
Here are presented the findings from an experimental investigation into three
inorganic consolidants being considered for use within The British Museum; Calcite Insitu
Precipitation System (CIPS); CaLoSil; and Ammonium Oxalate treatment. A range
of measures of the effectiveness of these consolidants, in comparison with a previously
used Silane; SILRES® BS OH 100 (Wacker OH), have been taken and assessed on 40,
artificially pre-weathered, limestone replicates.
Keywords
Carbonate Stones, Consolidation treatment, Effectiveness, Inorganic
1. Introduction
The British Museum has a large collection of stone artefacts some of which have
reached a state of deterioration such that application of a stone consolidant would be the
only way to prevent further disaggregation.
Over the last few decades, emphasis in conservation has been on the importance of
the reversibility of treatments (Rodgers 2004). However, consolidation treatments for
stone are rarely truly reversible. Therefore, stone conservation has started to focus more
on compatibility of a conservation treatment with the stone substrate (Sasse and
Snethladge 1996). Along with compatibility, a consolidant usually undergoes testing to
check how it measures against a range of performance requirements including
durability, depth of penetration, effect on stone porosity, moisture transfer and effect on
appearance (Cnudde et al 2004).
Organo silanes are consolidants that have been fairly popular in the past, and
indeed are still used by many stone conservators. Used in the British Museum in the
1980’s, there are a number of treated artefacts in their collections that demonstrate the
success of silanes as stone consolidants. However, there are some stone artefacts, mainly
limestone objects, where deterioration has worsened since being treated with organo
silanes. This phenomenon has been seen elsewhere and effectiveness of silanes on
carbonate stones is a debated topic (Wheeler 2008). One suggested reason for the issues
sometimes experienced with silanes and carbonate stones, is a lack of affinity with the

stone substrate (Favaro et al 2008). Other difficulties associated with silanes include
safety of application and environmental sustainability.
Inorganic consolidants address compatibility issues with limestone artefacts, but
have been considered less effective than silanes, predominantly due to poor penetration
depth and deposition within pore spaces (Favaro et al 2008). Recent scientific
advancements, in particular the development of nano-limes, have led to a resurgence in
interest in the possibility of inorganic consolidants.
The British Museum has instigated an investigation into potential new inorganic
consolidants to be used on their stone collections, in particular on limestone. Presented
here are results from an experimental assessment of three potential new inorganic
consolidants for use on the limestone collections: Calcite In-situ Precipitation System
(CIPS); CaLoSil; and Ammonium Oxalate treatment, in comparison to a previously used
and well-known silane: SILRES® BS OH 100 (Wacker OH). 40, 50x50mm cubic
limestone replicates were cut to give 10 replicates per consolidant investigated. All 40
replicates were artificially weathered using a furnace. The four consolidants were then
applied as directed by the manufacturers and after discussion with BM stone
conservators. Replicates were tested before and after weathering and before and after
treatment, and results analysed to determine effectiveness and which (if any) consolidant
had the potential to be used in the future in the museum. This paper focuses on results
from the non-destructive testing of the replicates, destructive testing is ongoing.
2. Materials and Methods
2.1 Stone Materials
The stone types for the replicates used in these experiments were Elm Park and
Hartham Park. These two types of stone are both Bath stones from quarries near
Corsham in Wiltshire. Pale buff in colour, with occasional blue flecks or veins, they are
oolitic limestones of middle Jurassic age. The 40 replicates were freshly cut into
50x50mm cubes, after which artificial weathering was undertaken. The replicates were
heated to 250˚C and left for half and hour before being removed and held under cold
running water for 5 minutes in a method similar to those used in McCabe et al 2007
(McCabe et al 2007). This was repeated for 5 cycles.
2.2 Consolidants
2.2.1 CIPS
The ‘Calcite In-situ Precipitation System’ or CIPS is one new inorganic
consolidant under consideration by the British Museum. The CIPS method (developed
by Lithic Technology Pty Ltd) is based upon precipitation of calcite within granular
materials- from loose sand to disaggregated stone, in order to create high strength
bonding between the grains (Price 2010). The calcite is precipitated from a solution that
is mixed and then permeated through the material to be consolidated. Calcite crystals
form around the edges of grains, creating bridges between them. Along with
compatibility, particular advantages of CIPS are its ease of use and low toxicity.
The CIPS fluid is formed from two solutions mixed in equal proportions just prior
to application. The exact composition of the solutions, and their production methods, are
proprietary to Lithic Technology Pty Ltd who can be contacted for more details.

Multiple applications of CIPS can be undertaken to further strengthen the stone; 6
are recommended for sufficient consolidation without inducing a highly noticeable
change in the surface appearance (Price 2010). Of the ten replicates for CIPS analysis,
eight were treated by spraying with the mixed solutions; two of the 10 replicates were
kept as controls. The replicates were treated every 24hrs for 6 applications.
2.2.2Ammonium Oxalate
Calcium oxalate films are naturally occurring patinas on calcareous stones (Cariati,
et al. 2000). Oxalate films provide a thin, compact, non-porous shell highly resistant to
acid attack and atmospheric pollution. Due to this, there has been a large amount of
research conducted into methods of producing this layer artificially (Doherty et al. 2006,
Doherty et al. 2007) including a method involving the application of ammonium oxalate.
The ammonium oxalate reacts with the stone substrate to form calcium oxalate.
Whilst the majority of recent research has focused upon the use of oxalate films as
protective layers, there have been suggestions they could also have consolidation
properties (Matteini 2007). To test this, Ammonium Oxalate Monohydrate was mixed
with water at 5%w/v and then applied to 8 of the ammonium oxalate replicates using a
cellulose poultice as recommended by Cezar (1998). A layer of Japanese tissue paper
was placed between the poultice and the stone surface as protection. After application,
the poultice was covered with cling film and left for 24hrs before removal. After 3
weeks the process was repeated for a second time.
2.2.3 CaLoSil
CaLoSil is a new, commercially available (from IBZ Frieberg), nano-lime stone
consolidant developed specifically for consolidating calcareous historic material
(Campbell, Hamilton et al. 2011). Nano-particles of Ca(OH)2 are suspended in different
alcohols. After application, the alcohol evaporates to leave solid calcium hydroxide
behind. This then reacts with atmospheric carbon dioxide to form calcium carbonate.
Due to the small size of the particles, CaloSil can penetrate deep into the stone. CaLoSil
E25 (calcium hydroxide particles dispersed in ethanol at 25g/L) was applied by spraying
to 8 of the 10 replicates (with two remaining as untreated controls). After 3 weeks a
second application was undertaken.
2.2.4 Wacker OH 100
Organo-Silanes have been commonly used since the 1960’s (Wheeler 2008), and
have been a very popular choice for consolidation, especially in Europe (Price 2006).
The British Museum began using them in the 1980’s but to a lesser extent than
institutions on the continent, and has not used them for several years.
Silanes can be alkoxy silanes or silicone esters (Bradley 1986), the main difference
for conservation purposes being that alkoxy silanes possess hydrophobic properties so
strengthen and prevent entry of water (De Muynck et al. 2010), whereas silicone esters
are solely used to improve internal cohesion (Bradley 1986).
The silane selected to use in this experiment is SILRES® BS OH 100, also known
as Wacker OH 100. Produced by Wacker Chemie AG, SILRES® BS OH 100 is a
solventless ethyl-silicate that penetrates into the stone forming a glass-like silica gel
binder (SiO2 . aq.)(Wacker Chemie AG MTDS). The SILRES® BS OH 100 was applied

y pipette (again to 8 of the 10 replicates with 2 left as untreated controls) until
saturation and left to cure. After three weeks a second application was undertaken.
2.3 Testing methods
To investigate, and compare the four different consolidants, a variety of analytical
methods were used; some non - (or minimally) destructive, others destructive. Table 1
summarises the methods used and the property of the replicate that they assess.
Method Data Provided Destructive?
Equotip Surface hardness of stone; Leeb Minimally
rebound method
destructive
Pundit Ultrasonic Pulse Velocity to
calculate E Modulus
Non-destructive
Scanning Electron Microscopy Depth of penetration, deposition
within pores. Elemental mapping of
consolidant distribution
Destructive
Karsten Tubes Water penetration depth into stone Minimally
destructive
Drilling Resistance Measurement Internal strength and, depth of Destructive
System
penetration of consolidants
Mercury Intrusion Porosimetry Pore size and distribution Destructive
Thin-Section Petrography Distribution of consolidant, depth of
penetration and how pore spaces
affected
Destructive
Spectrophotometry Colour change Non-destructive
Other Physical Properties
Table 1: Analytical methods used
3. Results and Discussion
Dimesions, weight, density etc Non -destructive
3.1 Weight
All replicates were weighed before artificial weathering, after artificial
weathering/before treatment, and after treatment. This was used to help calculate the
change in density at both stages. The density was further used in the calculation of the E
Modulus.
Replicate
Group
Mean ΔM (g)
Weathering
Mean Δ ρ (kg/m 3 )
Weathering
Mean ΔM (g)
Treatment
Mean Δ ρ (kg/m 3 )
Treatment
CIPS -0.22 -1.714732 3.14 24.54
CaLoSil -0.57 -4.656352 0.32 2.63
Wacker -0.42875 -3.351183 1.56 12.22
AmOx -0.52125 -4.344182 0.61 5.07
Table 2: Change in weight and density after artificial weathering and after consolidation treatment
As can be seen from table 2, the CIPS treatment increased the weight (and hence
the density) the most out of the four treatments, at nearly double the next highest weight
increase; the Wacker OH replicate group. This in turn was more than double the next

highest increase, which was the replicate group treated with ammonium oxalate. The
lowest weight and density increase was seen with the CaLoSil replicates.
3.2 Surface Hardness - Equotip
Surface hardness was determined using a Proceq Equotip (Equotip 3); a dynamic
hardness test using rebound technique that calculates Leeb hardness using change in
velocity (Viles, Goudie et al. 2010). 10 measurements were taken per 50x50mm area on
the topmost surface of each replicate. Measurements were taken before artificial
weathering, after artificial weathering/before treatment, and after treatment.
(a) (b)
(c)
Figure 1: Graphs (a) – (d) to show change in surface hardness as measured by Equotip . (a) CIPS
replicates; (b) AmOx replicates; (c) CaLoSil replicates; (d) Wacker replicates.
Figure 1 shows before treatment (u) and after (l) treatment Equotip values for each
replicate, divided into consolidant groups. (a) Shows that some CIPS replicates
demonstrated a noticeable increase in surface hardness, however, some did not and
overall there is not a statistically significant difference (paired t-test P = 0.237; 95%
confidence interval). Similarily, some CaLoSil replicates visually show an increase (c)
but overall there is no statistically significant difference (paired t-test P = 0.519; 95%
confidence interval). All Ammonium Oxalate (b) and Wacker (d) replicates showed a
clear increase in surface hardness. The increase in surface hardness demonstrated by the
Ammonium Oxalate replicates was quite dramatic with many experiencing an increase
of more than 90 HLD Leeb D.
(d)

These results clearly show that the ammonium oxalate treatment was the
consolidant most effective at increasing surface hardness. However, this does not
necessarily mean it is the most effective treatment – it will be important to see using
other methods such as DRMS and SEM how far into the stone this consolidant has
penetrated. How the consolidant has formed is also important – a solid surface layer or
crust might be more detrimental than positive if there is a sudden interface with the
stone substrate, or if it does not allow vapour to penetrate and allow the stone to
‘breathe’.
3.3 Internal Strength - Pundit
The Pundit Lab is an ultrasonic pulse velocity (UPV) test instrument, which when
used with 250kHz Sheer Wave Transducers can be used to calculate the Elasticity
Modulus of a material. This in-turn can be used as an indicator of internal strength. The
control, untreated replicates were measured, along with the treated replicates posttreatment.
Each replicate was measured five times and an average calculated.
Unfortunately, there was no observable difference in the E Modulus between the
control and the treated replicates. This could be due to only one surface being treated,
the consolidant may not have penetrated far enough into the replicate to affect the
overall UPV. Future experiments looking at the change in UPV and E Modulus when all
sides of a replicate are treated could look into this factor.
3.4 Colour Change - Spectrophotometer
To assess the effect of consolidation treatments on the appearance of the stone,
change in the colour parameters of the limestones (CIE L*a*b*) were measured using a
spectrophotometer (Konica Minolta CM 700d/600d) under illuminant D65 at a 10
degree of observation and using colour data software SpectraMagic xs. 10
measurements were taken on the 50x50mm area on the top treated surface to give a
mean colour difference. Overall change in colour (∆E00) was examined, along with a
more detailed look at change in colour, hue and saturation.
Before analysis, replicates were observed qualitatively. Colour change between
CIPS replicates and untreated blocks was difficult to detect, however, the surface had
changed in appearance with treated replicates acquiring a sheen. Colour difference
between the ammonium oxalate blocks and the untreated blocks was also difficult to
detect. The Wacker replicates were slightly darker in appearance, but only on very close
inspection. The biggest change was seen with the CaLoSil replicates. A white haze or
‘bloom’ was visible on nearly all treated replicates.
Replicate Group Group Traits ΔL* Δa* Δb* ΔC* ΔE00
CIPS SCI -3.15 0.31 2.65 2.67 2.72
CIPS SCE -3.30 0.32 2.76 2.76 2.85
AmOx SCI -3.74 1.09 -0.24 0.05 3.58
AmOx SCE -3.71 1.08 -0.23 0.05 3.57
CaLoSil SCI 5.99 -1.47 -12.75 -12.59 9.50
CaLoSil SCE 5.94 -1.47 -12.77 -12.60 9.50
Wacker SCI -5.22 1.03 3.61 3.81 4.47
Wacker SCE -5.20 1.29 3.65 3.85 4.46
Table 3: Summary of mean colour change per replicate group

(a)
(c)
As can be seen from table 3, all replicates demonstrated an overall colour change
(ΔE00) greater than 1; the value generally accepted as the level at which colour change
is noticeable to the naked eye. The replicates that demonstrated the lowest overall colour
change were ones treated with CIPS; ΔE00 of 2.72 SCI and 2.85 SCE. The Ammonium
Oxalate replicates had the next lowest colour change with ΔE00 of 3.58 SCI and 3.57
SCE. Wacker replicates had an even greater overall colour change and most
dramatically, CaLoSil replicates had a ΔE00 of 9.50 SCI/SCE. This CaLoSil result was
to be expected with the significant white haze that appeared on the surface. SCI and
SCE values can help to show a difference in the gloss on the surface. For all groups
except the CIPS these two values are almost identical. The CIPS replicates showing a
difference is likely due to the sheen now seen on the surface.
If purely considering the colour change, the choice for the preferred consolidant
would be between the CIPS and the ammonium oxalate; since these two altered the
colour the least. Whilst overall colour change is lower with the CIPS treatment, the
change in the gloss of the surface is definitely noticeable and so the ammonium oxalate
might be the better choice.
3.5 Water Penetration - Karsten Tubes
Water penetration capacity (WPC) was measured before and after treatment using
Karsten Tubes. A glass tube containing water is bonded to the surface of the replicate
with putty. The drop in water level is measured over time. This was repeated 10 times
for each replicate and the mean ml/minute/area of water in contact with the stone
calculated.
(b)
(d)
Figure 2(a-d): Graphs to show change in mean ml/min water penetration of replicates before (u)
and after (l) treatment. (a) CIPS replicates; (b) AmOx replicates; (c) CaLoSil replicates; (d)
Wacker replicates.

As can be seen in Figure 2, all replicates treated with CIPS, Ammonium Oxalate,
and Wacker OH demonstrated a dramatic decrease in the average millilitres of water
penetrating per minute. The Wacker replicates had the most severe decrease with hardly
any water penetrating after treatment. The CIPS and Ammonium Oxalate replicates
demonstrated a fairly similar decrease; after treatment water penetration was low, but
some was still going in. Some individual CaLoSil replicates demonstrate a decrease in
water penetration, although very slight and overall the differences are not statistically
significant (paired t-test P = 0.055; 95% confidence interval).
None of the treatments used were designed with hydrophobic properties in mind,
however some decrease in water penetration capacity is to be expected as the
consolidant forms new bonds in the spaces between the individual grains of the stone.
Completely reducing the water penetration can be undesirable however as this can
prevent the stone from dealing with changes in relative humidity, and can cause stress
between the untreated substrate and the treated zone near the surface of the stone.
3.6 Destructive analysis
The next stage in this investigation will be to undertake the destructive testing of
the replicates. This will include observing thin sections on a petrological microscope,
and cross sections using Scanning Electron Microscopy (SEM), to look at depth of
penetration of the consolidants, interface between treated and non-treated areas, and
distribution of the consolidants within the stone. Drilling Resistance Measurement
System (DRMS) will be used to look at depth of penetration and internal strength of the
stone. Mercury Intrusion Porosimetry (MIP) will be used to look at the pore spaces
within the stones.
4. Conclusions
There are many consolidants available to treat deteriorated stone artefacts and the
best one to use will depend upon a variety of factors including the stone itself, its
location, deterioration problems and history. Preliminary testing looking at treatment
effectiveness, compatibility with the stone and how it changes the stone appearance
(along with other parameters), can help identify which consolidant would be best suited
to the artefact.
Presented within this paper are the results from the non-destructive analysis of 3
inorganic consolidants being considered for use on British Museum limestone artefacts,
in comparison with a previously used organic treatment, Wacker OH (SILRES® BS OH
100).
Surface hardness testing using an Equotip 3 showed the ammonium oxalate
treatment provided the greatest increase in surface strength, far larger than that
demonstrated by the organic consolidant Wacker OH, although the increase shown with
the Wacker was statistically significant. Both the CIPS and the CaLoSil had some
individual replicates which demonstrated an increase in surface hardness, but overall
neither of these treatments provided a statistically significant increase in surface
hardness. The most effective consolidant in this instance is the ammonium oxalate
treatment, however these results should be used in conjunction with analysis of depth of
penetration and the interface between consolidated areas and inner stone substrate; a
very solid surface crust could pose many problems for an artefact.

Overall colour change for all replicates was at a level detectable by the human eye.
The lowest values were for those treated with CIPS, although this treatment did change
the gloss of the surface in a noticeable way. Next lowest were the replicates treated with
ammonium oxalate, which were below the values for the replicates treated with Wacker
OH. The replicates treated with CaLoSil developed a very obvious white bloom on the
surface; shown by the extremely high ΔE00 values. The colour change was so severe
that further experiments to try and reduce/prevent this white bloom would be needed
before any recommendation for trialling on museum artefacts could be made.
Karsten Tube tests to analyse water penetration showed replicates treated with
CIPS, ammonium oxalate, and Wacker OH, all had a significant decrease in water
penetration with virtually no water penetrating into the Wacker OH replicates. Some
individual CaLoSil replicates had a decrease in water penetration, however overall there
was not a statistically significant difference. These treatments were all designed as
consolidants, not as hydrophobic coatings. Whist some decrease in water penetration
could be expected, blocking all water penetration could be detrimental to an object.
No one parameter can be used on its own to make a decision about whether any of
these three inorganic consolidants hold potential for use on limestone artefacts in the
British Museum, or are a more preferable option than previously used Wacker OH. This
decision will be made once results from the destructive tests are in. However,
information from these non-destructive tests can be used to focus on certain areas during
the destructive testing. The results here seem to show the ammonium oxalate treatment
succeeding more than the other two inorganic consolidants and indeed outperforming
the Wacker OH. However, it will be really important to analyse if this treatment is
consolidating the stone substrate, or just has just formed a very hard superficial layer.
Acknowledgements
This research has been made possible due to a CDA studentship from the Arts and
Humanities Research Council, between the School of Geography and Environment;
University of Oxford and the Department of Conservation and Scientific Research;
British Museum. Many thanks to Nic Lee and Tracey Sweek from Stone Conservation at
the British Museum, for their kind help and advice. A big thank you also to Mona
Edwards and Hong Zhang, for assistance in the lab.
The CIPS technology was used under licence from Lithic Systems Pty Ltd. Advice
about the preparation and use of CIPS was given by Dr Graham Price.
Bibliography
Bradley, S.M., 1986. An Introduction to the Use of Silanes in Stone Conservation. The
Geological Curator, 4(7), pp. 427-432.
Campbell, A., Hamilton, A., Stratford, T., Modestou, S. and Ioannou, I., 2011. Calcium
Hydroxide Nanoparticles for Limestone Conservation: Imbibition and Adhesion
Adhesives and Consolidants for Conservation: Research and Application ,
17/10/2011 - 21/10/2011 2011.
Cezar, T., 2011. Calcium Oxalate: A Surface Treatment for Limestone. Journal of
Conservation and Museum Studies, 4.
Cnudde, V., Cnudde, J.P., Dupuis, C. and Jacobs, P.J.S., 2004. X-ray micro-CT used for

the localization of water repellents and consolidants inside natural building stones.
Materials Characterization, 53(2-4), pp. 259-271.
De Muynck, W., De Belie, N. and Verstraete, W., 2010. Microbial carbonate
precipitation in construction materials: A review. Ecological Engineering, 36(2),
pp. 118-136.
Doherty, B., Pamplona, M., Miliani, C., Matteini, M., Sgamellotti, A. and Brunetti, B.,
2007. Durability of the artificial calcium oxalate protective on two Florentine
monuments. Journal of Cultural Heritage, 8(2), pp. 186-192.
Doherty, B., Pamplona, M., Selvaggi, R., Miliani, C., Matteini, M., Sgamellotti, A.
and Brunetti, B., 2006. Efficiency and resistance of the artificial oxalate protection
treatment on marble against chemical weathering. Applied Surface Science,
253(10), pp. 4477-4484.
Favaro, M., Tomasin, P., Ossola, F. and Vigato, P.A., 2008. A novel approach to
consolidation of historical limestone: The calcium alkoxides. Applied
Organometallic Chemistry, 22(12), pp. 698-704.
Matteini, M., 2007. Conservation of Stone Monuments and Artifacts: New possibilities
offered by the Ammonium Oxalate Based Treatment, <strong>International</strong> Meeting on
Science and Technology for Cultural Heritage, 7/02/07 - 10/02/07 2007.
McCabe, S., Smith, B.J. and Warke, P.A., 2007. Preliminary observations on the impact
of complex stress
histories on sandstone response to salt weathering: laboratory simulations of
process combinations. Environmental Geology 52, pp.251-258.
Price, C., 2006. Consolidation. In: A. Henry, ed, Stone conservation : principles and
practice. Shaftesbury: Shaftesbury : Donhead, pp. 100-117.
Price, G., 2010. Calcite In-situ Precipitation System- CIPS- Technical Description-
Chapter 2: Calcite Crystallisation. Book Chapter edn. Australia: .
Rodgers, B.A., 2004. The Archaeologist's Manual for Conservation: A Guide to Non-
Toxic, Minimal Intervention Artifact Stabilization. New York: Kluwer
Academic/Plenum Publishers.
Sasse, H.R. and Snethlage, R., 1996. Evaluation of stone consolidation treatments.
Science and Technology for Cultural Heritage, 5(1), pp. 85-92.
Viles, H., Goudie, A., Grab, S. and Lalley, J., 2010. The use of the Schmidt Hammer
and Equotip for rock hardness assessment in geomorphology and heritage science:
a comparative analysis. Earth Surface Processes and Landforms, , pp. n/a-n/a.
Wheeler, G., 2008. Alkoxysilanes and the Consolidation of Stone: Where We Are Now,
J. Delgado Rodrigues and J. Manuel Mimoso, eds. In: Stone Consolidation in
Cultural Heritage: Research and Practice, 06/05/08 - 07/05/08 2008, Laboratorio
Nacional De Engenharia Civil (LNEC), pp. 41-52.
Wacker Chemie AG Material Technical Data Sheet retrieved from
http://www.wacker.com/cms/en/productsmarkets/products/product.jsp?product=10094
(20/06/2012).

DAMAGE ASSESSMENT OF THE FERRUGINOUS SANDSTONE MASONRY
– STRENGTHENING EFFECT OF CONSOLIDATION WITH TEOS –
THE VIRGIN TOWER OF ZICHEM
Hilde De Clercq 1 , Roald Hayen 1 , Michiel Dusar
1 Royal Institute for Cultural Heritage, Jubelpark 1, 1000 Brussels, Belgium
2 Geological Survey of Belgium, Jennerstraat 13, 1000 Brussels, Belgium
Abstract
The Virgin Tower of Zichem is a 14 th century dungeon of iconic meaning for the
Hageland, a natural landscape in east-central Belgium. The poorly consolidated Diestian
ferruginous sandstone of the dressed stone masonry shows severe damage for which a
conservation strategy had to be established, based on an investigation of the
deterioration phenomena. This sandstone is characterised by the presence of a black
crust serving as natural protection. Recently, the black crust tends to break off, leaving
behind a weakened material exposed to natural weathering.
An investigation was carried out on the characteristics and the origin of the black crust
by microscopical analyses. The mechanical properties of the surface layers and the
possible strengthening effect of a consolidation treatment with ethylsilicate (TEOS)
were determined by means of the portable drilling device DRMS (Drilling Resistance
Measurement System).
The black crust is the result of internal and external processes. Iron (hydr)oxide
mobilisation towards the outer surface results in a compact superficial layer overlaying a
depleted, hence weakened layer. The crust is covered by a thin black veneer of arguably
biological origin but also containing airborne pollution particles showing activity during
the period 1860-1960. Severe material loss due to slabs of crust breaking off must
postdate this period since affected stones did not regenerate a black crust.
The strengthening with ethylsilicate of deteriorated stones was generally not effective on
stones of poor strength while sufficient on better quality stones. Laboratory tests showed
that clay content of the sandstone makes the difference in durability and inhibits an
effective consolidation.
Keywords: ferruginous sandstone, damage, consolidation, TEOS
1. Introduction
Diestian ferruginous sandstone is a regionally important building material of the
Hageland, to the north-east of Brussels in Belgium. The Hageland is characterised by a
typical landscape built by rows of longitudinal hills, underlain by partly consolidated
ferruginous sands. Diestian sandstone is a rather young geological material, originally
deposited as coast-parallel sandbars of poorly sorted glauconite rich marine sand in
Miocene times. Sands became lithified near the earth’s surface by limonitic - goethitic
cements, after weathering of the mineral glauconite. Diagenesis occurred in Quaternary
times, a process that started after peneplanation, resumed after incision of the drainage
network and is still going on along exposed slopes and roadcuts (Bos and Gullentops

1990). This irregular, low pressure, low temperature lithification resulted in weak
strength and varying durability for the rusty brown sandstones. Nevertheless, its local
use as building stone links the architectural heritage to the landscape; both forge the
combined natural-cultural identity of this region. Diestian ferruginous sandstone is the
typical building stone of the Hageland and was widely used at the height of the
economic and demographic expansion in the 14 th to 16 th centuries, after which its
popularity faded. All emblematic gothic monuments of the region are constructed in this
material whose limited capacity for sculpturing led to a local variant of the Gothic style,
known as ‘Demer Gothic’. Further stone production and use declined till the early 20 th
century and became generally restricted to vernacular architecture and to the ground
levels of larger buildings. No quarries have survived today. Moreover, the durability of
the sandstone was not a limiting factor for its application, contrary to its regularity and
ease of working. As a consequence, exposure to natural weathering processes often
causes decay and material loss. Stocks of replacement stone become depleted and
attempts to discover new quarry sites yielding quality stones so far were not successful
and are actually under investigation. Hence, restoration of monumental constructions
built with Diestian ferruginous sandstone for which the state of preservation could imply
replacement is a complex issue, as is the case of the Maagdentoren (freely translated as
“Virgin Tower”). This dungeon (Figure 1) having an original height of nearly 25 m and
a diameter of 15 m was erected in the 14 th century at the outskirts of the former city of
Zichem and left in state of abandonment two centuries after its construction in a now
rural environment (Doperé and Corens 2002). Moreover, the urgent conservationrestoration
concept was highly modified after its partial collapse in 2006 due to the
effect of general weathering, severe water infiltration and thick soil deposit at its top
level.
Figure 1. Virgin Tower at Zichem, before its partial collapse in 2006

The sandstone masonry suffers from several deterioration phenomena for which an
appropriate conservation strategy was required. Generally, the Diestian ferruginous
sandstone is characterised in this monument by the presence of a thick ferruginous crust
under a thin black surface layer serving as natural protection of the stone. This crust
tends to break off, leaving behind a weak surface to natural exposure (Figure 2).
Figure 2. Superficial crust of a Diestian ferruginous sandstone covered by a black veneer.
Damage is caused by the crust breaking off in large slabs, exposing a weakly cemented sandy
layer.
Deterioration of sandstone proceeds according to a pattern which is quite comparable to
the weathering process of natural exposures. Although the outer layer of weathered
sandstone contains atmospheric pollutants, biological processes could play an important
role in stone decay as well as in stone protection (Guilitte 1995). The weak layer
underlying the crust is comparable in composition, but contains a lower amount of iron
(hydr)oxides, which migrated outward followed by a re-deposition near the surface. The
weak layer underneath functions as source for the cementing minerals and is therefore
characterised by a reduced cohesion and a lower mechanical strength.
The relation between the ferruginous crust and its black cover remains elusive.
Numerous studies deal with the formation of black crusts on lime stones, whereby
gypsum crystallisation is the main cause for decay. Several investigations demonstrated
that the black crust on sandstones generally consists of algae and fungi, soot, gypsum,
iron (hydr)oxides, lead, organic and some phosphor compounds and fly-ash (Nijland et
al 2003). The black colour results from the deposition of organic matter and/or pollutant
particles.
Black crusts can also have a biological source, through the biogenic deposition of iron-
and manganese oxides and hydroxides, the so called ‘black varnish’ or ‘desert varnish’
(Krumbein and Jens 1981). Especially cyanobacteria and to a lesser extent lichens
stimulate the mineral transformation processes of sand(stone)s subjected to wet-dry
cycles. Being very thin, its thickness may be just a few µm, this black crust resembles a
patina and serves as its natural protective veneer. The formation of black crusts of the

type ‘desert varnish’ occurs especially on sand(stone)s serving as an unfavourable food
medium for organisms and having an evaporation-transpiration factor of at least 1, as is
frequently the case on walls. ‘Desert varnish’ can be distinguished from other black
crusts by the absence of sulphates, a low content of organic matter and an enrichment of
manganese compared to iron (Gaylarde et al 2007; Galletti et al 1997). Despite its
protective function resulting in a decrease of the wet – dry cycles, the black crust on
ferruginous sandstone can be bypassed along the edges of stones or through internal
channels and does not prevent further leaching and transformation processes deeper in
the stone (Thomachot and Jeannette 2004), as is evidently observed for the ferruginous
sandstone of the Virgin Tower. It was therefore important to investigate whether the
black crust was recently formed due to the exposure to a polluted environment or
whether it is of the type ‘desert varnish’ and hence much older.
A further step consists of an evaluation of the effect of a strengthening treatment of the
weathered ferruginous sandstone of which the protective black crust had been largely
removed. The in situ applications are carried out using ethylsilicate (tetraethoxysilane or
TEOS). Of the many products used over the past century, the consolidant that best
seems to be standing the test of time in terms of efficiency, durability and lack of side
effects on damaged historical masonry, is TEOS (Grissom 1981). The vast current use is
also due to its easy application and versatility, for it has proven efficiency on limestone,
sandstone and brick masonry (De Witte et al 1977, 1985; Delgado Rodrigues et al 1998).
This contribution describes the results of an investigation of the outer black crust, the
mechanical properties of the superficial layer and the strengthening effect of
consolidations with ethylsilicate.
2. Experimental
2.1 Petrographic analysis of the ferruginous sandstone – characteristics and origin
of the black crust
Petrographic analyses of the ferruginous sandstone were carried out on thin sections
(Logitech PM5 Autolap) analysed with an optical polarised light microscope (Nikon).
Samples of the surface layer of ferruginous sandstone having a black crust were
embedded in an epoxy resin. Polished samples were analysed by means of the optical
microscope (ZEISS AXIOPLAN) and scanning electron microscope-energy dispersive
X-rays system (SEM-EDX, JEOL).
2.2 Consolidation properties
The mechanical properties of the outer layer of the ferruginous sandstone of the Virgin
Tower were investigated by means of the portable drilling device Drilling Resistance
Measurement System (Sint Technology DRMS Cordless). Surrounding scaffolding
provided access to all levels. After evaluation of the DRMS results, six areas of varying
mechanical strength properties were selected for an in situ treatment with TEOS (75%
solution, dry weight 51%). The applications were done in two- or threefold. Each
application was carried out wet-in-wet as to obtain a run-off of at least 10 cm. The time
between successive treatments was at least 24 h. One month after the last application,
the consolidation effect was evaluated by means of repeated DRMS measurements.
Consequently, the hardness profile of the treated stone could be compared to that of the

untreated one serving as reference. On each test area, at least 3 measurements were
carried out.
In order to verify the effect of the clay content in the sandstone on the effectiveness of
the TEOS application, the treatment was also conducted on a sandstone powder
produced from grinded ferruginous sandstone samples. The test material was obtained
by softly grinding two different sandstone blocks from the collapsed part of the tower
and sieved for the determination of the average grain size distribution. The grinded
samples were divided in two parts; one part was used as such while from the other part
the fraction smaller than 75 µm was removed. A plastic cylinder (3 cm diameter, 4 cm
height), of which the bottom side was protected by a cotton tile, was filled with the
sandstone powders. The consolidation by TEOS was carried out by capillary absorption
through the bottom side till the consolidation product reached the upper surface. The
treatment was repeated twice. The time between two successive treatments was at least
24 h. One month after the last treatment, the plastic cylinder was removed and DRMS
measurements were carried out on the consolidated artificial sandstone.
3. Results and discussion
3.1 Characteristics of the black crust
Figure 3 presents the SEM-image of a cross section of the surface zone of the
ferruginous sandstone revealing a rather dense superficial layer of some tens of µm
thickness. The black layer is in this sample thin and hardly visible, hence more
resembling a patina. The dense outer layer, showing some thin cracks perpendicular to
the surface, is composed of the elements silicon, iron, potassium and some aluminium
and magnesium, referring to the natural mineralogical composition of the stone. The
underlying layer (thickness about 10 µm) contains as main elements iron and some
silicon, and is mainly composed of limonitic cement as binding medium as observed on
the thin sections.
Figure 4 presents the optical microscopical image of a detail of a much thicker black
crust (thickness up to 100 µm) on top of the brown limonite layer. The SEM-image of
the marked detail clearly shows round particles consisting of silicon, aluminium and
some iron and potassium, typical for fly-ash. No gypsum but some phosphor compounds
are detected in the surface layer, unlike the black sandstone crusts described by Nijland
et al (Nijland et al 2003). There is no evidence for a changing ratio manganese/iron in
the crust to conclude a biogenic transformation process.
Apparently, different phenomena contribute to the formation of the crust. Firstly, the redeposition
of iron (hydr)oxides leached out from deeper in the stone working at different
scales and rates, resulting in case hardening that decreases in strength towards the inside
and becomes easily detachable from the weakly cohesive underlying sandy layer.
Secondly the atmospheric encrustation by organic matter and the trapping of airborne
pollutants, highlighted by the fly ash, that result in the black crust.
The first one is a natural process stimulated by wet-dry cycles by which stone material
gets dissolved and migrates to the surface where it re-deposits, lowering the porosity of
the surface layer. Recent research by means of X-ray tomography confirms the lower
porosity (Cnudde et al 2011). The higher the content of crystalline goethite in
comparison to limonite the more shiny black its colour becomes (Bos 1989). The dense

outer layer of some tens of microns in thickness lowers the rate of moisture uptake by
the stone and can be considered as its natural protection. This process started in
principle from the original building phase. The formation of thick crusts, which is
detrimental for stone conservation in the long-term, could also be considered as the
result of insufficient seasoning of the extracted stone, before emplacement in the
building (Bos and Gullentops 1990; Bos 1989). This would mean that either the blocks
were extracted too quickly from the quarry site or were left soaked with water as was
common practice for porous lime stones. However, drilling cores of the sandstone of the
dressed masonry revealed that material loss under the black crust until now has been
rather minimal, more or less preserving the original thickness of the stones, thus
suggesting that the loss of thick crusts is a fairly recent and rapidly progressing process.
The black layer has locally a thickness of only a few micron, rendering it difficult to
analyse. Its texture, welding sediment grains and pore filling, resembles bacterial
encrustations. In zones where the black crust has a thickness up to 100 µm, fly-ash
particles can be detected. Fly-ash is generally linked to coal burning activities, in this
particular case to the construction of a railway nearby the monument. Because of a
polluted environment the formation of the black crust might date from the mid 19 th
century to the first half of the 20 th century corresponding to the period when steam
locomotives were used. Except for the railroad traffic, the Virgin Tower is situated in a
highly rural environment. Therefore, it is questionable whether atmospheric pollutants
trapped within the black layer play a discernable role in the stone decay.
Figure 3. SEM image of a cross section of the outer layer of the ferruginous sandstone, EDXanalysis
of the top layer (top right) and mapping of silicon (Si), calcium (Ca), aluminium (Al),
potassium (K), magnesium (Mg) and iron (Fe) of the area presented top left.

Figure 4. Optical microscopic image (left) of a locally thick black crust and SEM-image of the
marked zone (right). The arrows mark fly-ash particles.
3.2 Consolidation properties
From the evaluation of the DRMS hardness profiles, a wide range in mechanical
properties and hence quality is observed, sometimes within a same stone. Figure 5
presents the hardness profile of a sandstone containing a black crust (thus including both
the black toplayer and underlying iron (hydr)oxide-cemented layer). A rather hard
surface layer is observed on top of a sandy section characterised by poor mechanical
properties followed by the healthy stone material. Irregularities in the profile are
probably due to either limonite cement strings or to the effect of bioturbation
(synsedimentary animal burrows of which the interior is poorly consolidated). This
hardness profile confirms the hypothesis of a superficial crust formed in situ on the
monument. The low cohesive layer underneath probably served as the leaching zone to
the (hydr)oxide-cements deposited in the superficial crust. A strengthening of such a
mechanically weak layer is hardly possible since the consolidation product needs to
cross over the harder surface layer to completely impregnate the weak zone.
Figure 6 presents the DRMS profile of good quality ferruginous sandstone before and
after two applications with TEOS. The untreated sandstone shows a weathered zone of
about 9 mm thickness. Two applications with TEOS, with a consumption of 2,3 l.m -2 ,
result in a successful strengthening effect. On the contrary, effective strengthening
properties were hardly obtained for poor quality sandstones representative for the largest
group of tested building stones of the Virgin Tower (Figure 7). A markedly low
mechanical strength of the entire untreated sandstone in comparison to a good quality
one (Figure 6) is noticed. Up to 15 mm depth, the poor quality stone possesses almost no
cohesion at all. Moreover, hardly any strengthening effect is noticed after two
applications with TEOS for which the consumption is 3,8 l.m -2 . A third application did
not improve the mechanical properties of the superficial layer any further. Apparently,
positive results are obtained for good quality stones showing superficial weathering
whereas no strengthening effect on poor quality stones could be discerned. These results
confirm the variety in stone quality already observed by visual inspection and
petrographic analysis of thin sections, despite their common origin and composition.
From (Wheeler 2005, 2008; De Clercq 2009) it is suggested that the effectiveness of the
consolidation effect of TEOS is influenced by the grain size of constitutive minerals and
by the clay content. More specifically, the maximum bridging capacity for TEOS is
50 µm, corresponding to the pore diameter between uniform round particles with grain

size of 325 µm. Powder consisting of round grains of a size larger than 325 µm cannot
be consolidated. Concerning the clay effect, however, no straightforward results are
available from literature. The average grain size distribution of the constitutive minerals
of two ferruginous sandstone blocks is presented in Figure 8. In the ferruginous
sandstone of the Virgin Tower, most grains are larger than 150 µm with an average size
around 250 µm. In the tested sandstone blocks not more than 12% of the grains pass the
limit of 325 µm, although the maximum size can exceed 1 mm. As the sandstone is
poorly sorted, the grain size of the constitutive minerals and hence limiting dimension of
the pore throats, cannot explain the poor strengthening effect. The presence of clay
might therefore play a decisive role. The clay hypothesis could be tested by the DRMS
results of the artificial ferruginous sandstone samples. The samples composed of the full
range of grain sizes, hence including clays and silt, revealed poor strengthening effects,
while better results were obtained for the samples from which the fraction smaller than
75 µm had been removed and hence considered as clay-free (Figure 9). These results
confirm that clays can inhibit the strengthening effect of TEOS on ferruginous sandstone.
Since the majority of the ferruginous sandstones is of a poor quality, generally showing
an insufficient strengthening effect, the usefulness of a consolidation treatment for the
entire monument is questionable.
16
14
12
10
8
6
4
2
0
0 10 20 30 40
Force (N)
Depth (mm)
Figure 5. Typical DRMS-hardness profile of untreated ferruginous sandstone showing a case
hardening effect covered by a black veneer and underlain by a weakly cohesive layer.
Force (N)
80
70
60
50
40
30
20
10
0
0 10 20 30 40
Depth (mm)
Figure 6. DRMS-hardness profile of good quality ferruginous sandstone with superficial
weathering, before (solid line) and after (dashed line) 2 applications with TEOS.

Force (N)
4
2
0
0 10 20
Depth (mm)
30 40
Figure 7. DRMS-hardness profile of poor quality ferruginous sandstone, before (solid line) and
after (dashed line) 2 applications with TEOS.
%
50
40
30
20
10
0
500-300 300-150 150-75 75-0
Figure 8. Grain size distribution after grinding of ferruginous sandstone samples from the
collapsed part of the building.
Force (N)
4
3
2
1
0
maaswijdte (µm)
0 10 20 30 40
Depth (mm)
Figure 9. DRMS profile of artificial sandstone samples with (grey) and without (black) the
fraction

lack superficial crust serving as natural protection. This crust tends to break off leaving
behind a weakened surface exposed to natural weathering and further material loss.
An investigation was carried out on the characteristics and origin of the black crust, the
mechanical properties of the surface layer and the strengthening effect of a consolidation
treatment with ethylsilicate.
The case hardening results from iron (hydr)oxide mobilisation and re-cementation
towards the surface of the stone by wet-dry cycles, probably initiated immediately after
emplacement in the building. Strengthening of the outer layer and weakening of the
underlying zone was proven by means of the portable drilling device DRMS (Drilling
Resistance Measurement System). This ferruginous superficial dense layer of internal
origin is covered by a black veneer. Its origin remains elusive. Examination of the black
crust with microscopical techniques did not reveal determining criteria for either a
biological origin (‘black varnish’) or environmental pollution. Stones protected by a
black crust did not experience noticeable loss of material during the lifetime of the
monument. Nevertheless, the trapping of fly ash shows that the black crust has been
activated during the last 150 years, based on supply of fly ash by the nearby railroad
during 1860-1960. Rapid deterioration due to the loss of the superficial crust must
postdate the same period since affected stones do not show a re-growth of a black crust.
The effect of strengthening with ethylsilicate (TEOS), evaluated by means of the
portable drilling device DRMS, was generally poor for deteriorated stones while
sufficient for stones of better quality. Further laboratory tests to explain the different
strengthening effect on stone materials of similar type and origin revealed that the clay
content and not the grain size was the determining factor, inhibiting consolidation in
more clayey sandstones. As most of the ferruginous sandstones display a clay-related
weathering profile, the usefulness of a consolidation treatment is questioned.
5. References
Bos, K. , Gullentops F. 1990. ‘IJzerzandsteen als bouwsteen in en rond het Hageland’.
In Bulletin de la Société belge de Géologie, 99: 131-151.
Bos, K. 1989. ‘The Saint Peter’s church of Langdorp. The ferruginous sandstone in
northern Belgium’. M.Sc., Centre for the Conservation of Historic Towns and Buildings,
KU Leuven.
Cnudde, V., Dewanckele, J., Boone, M., De Kock, T., Dusar, M., De Ceukelaire, M., De
Clercq, H. 2011 ‘High-resolution X-ray CT for 3D petrography of Belgian
ironsandstone with crust formation’, Microscopy Research and Technique.
DOI: 10.1002/jemt.20987
De Clercq, H. 2009. ‘De grenzen aan een consoliderende behandeling’, Geological
Survey of Belgium Professional Paper 2009/1,N 305: 129-136
Doperé, F., Corens, K. ‘Huizen in torens. De Zichemse Maagdentoren en andere
donjons’, Openbaar Kunstbezit in Vlaanderen, 2002
Delgado Rodrigues, J., Ferreiro Pinto, A.P., Charola, A.E., Aires-Barros, L., Henriques,
F.M.A. 1998. ‘Selection of Consolidants for Use on the Tower of Belem’, in Int
Zeitschrift für Bauinstandsetzen, 4. Jahrgang: 653-666

De Witte, E., Huget, P., Van Den Broeck, P. 1977. ‘A Comparative Study of Three
Consolidation Methods on Limestone’, in Studies in Conservation 22
De Witte, E., Charola, A.E., Sherryl, R.P. 1985. ‘Preliminary tests on Commercial Stone
Consolidants’. Proc. of the 5 th Int <strong>Congress</strong> on Deterioration and Conservation of Stone,
September 25-27, Lausanne, France, ed Felix G. , Furlan V.: 709-718.
Galletti, G.C., Bocchini, P., Cam, D., Chiavari, G.,Mazzeo, R. 1997. ‘Chemical
characterization of the black crust present on the stone central portal of St. Denis abbey’.
In Journal of Analytical Chemistry, 357 (8): 1211-1214.
Gaylarde, C., Ortega-Morales, B., Bertolo-Perez, P. 2007. ‘Biogenic black crusts on
buildings in unpolluted environments’.In Current Microbiology, 54-2: 162-166.
Grissom, C. A. 1981. ‘Alkoxysilanes in the Conservation of Art and Architecture: 1861-
1981’, In IIC Art and Archaeology Technical Abstracts, 18 (1)
Guilitte, O. 1995. ‘Bioreceptivity: a new concept for building materials ecology studies’.
In The Science of the Total Environment, 167: 215-220.
Krumbein, W.E. , Jens, K. 1981. ‘Biogenic rock varnishes of the Negev Deseert (Israel),
an ecological study of iron and manganese transformation by cyanobacteria and fungi’.
In Oecologia, 50: 25-38.
Nijland, T.G., Dubelaar, W., van Hees, R.P.J. , van der Linden, T.J.M. 2003. ‘Black
weathering of Bentheim and Obernkirchen sandstone’. In Heron 18-3: 179-195.
Thomachot, C., Jeannette, D. 2004. ‘Effects of iron black varnish on petrophysical
properties of building sandstone’. In Environmental Geology, 47(1): 119-131.
Wheeler G. 2005. ‘Alkoxysilanes and the Consolidation of Stone’, The Getty
Conservation Institute
Wheeler, G. 2008. ‘Alkoxysilanes and the Consolidation of Stone: Where are we now?’.
In Proc. of the Int. Symposium Stone Consolidation in Cultural Heritage, research and
practice, Ed. Rodrigues J.D., Mimoso J.M., May 6-7, Lisbon, Portugal,41 – 52

EFFECT OF PRETREATMENT FOR TEOS BASED STONE CONSOLIDANTS
Jiyoung Choi 1 , Sungjin Park 1 , Jongok Won 1,* , Yong Seok Choi 2,3 , Yong Soo Kang 2 and
Myeong Seong Lee 3
1 Department of Chemistry, Sejong University, Seoul, 143-747 Korea
(jwon@sejong.ac.kr)
2 WCU Program Department of Energy Engineering, Hangyang University, Seoul 133-
791, Korea
3 Conservation Science Divisions, National Research Institute of Cultural Heritage,
Daejeon 305-380, Korea
Abstract
Consolidants based on alkoxysilanes, such as tetraethoxysilane (TEOS) have been
applied widely in Korean stone heritages since they penetrated inside of the decayed
stone and they polymerize within the porous structure of the decaying stone by means of
a classic sol-gel process, significantly increasing the cohesion among the decayed stone
grains. The study presented here is aimed to provide the interaction between the TEOSbased
gels and the surface of the decayed grains, which give improved properties for the
preservation of cultural stone monuments. We developed the interaction between the
stone and the consolidants by the pretreatment with the functional alkoxysilanes, and
have characterized them for the application of Korean granite stone consolidants. The
effect of the pretreatment with the functional alkoxysilanes was compared with TEOS
based solution including the commercial products such as Wacker OH 100. For the
model study, functional alkoxysilanes were treated with SiO2 particles and we obtained
crack-free dried gel for all TEOS based consolidants. The surface attachment test for
Korean granites showed improved interaction with the TEOS based consolidants
including commercial products. The properties and the applicability of the developed
pretreatment process for the decayed Korean granites are also investigated.
Keywords: consolidants, tetraethoxysilane, pretreatment, granite, sol-gel
1. Introduction
Wacker OH based on alkoxysilane such as TEOS, are commonly used for
consolidating weathered stones in Korea (Cnudde 2007; Tsakalof 2007; Wheeler 2005),
specially located outdoors. TEOS-based consolidants polymerize in situ inside the
weathered stone heritage, by means of a sol-gel reaction with environmental moisture,
providing strength to the structure of the heritage. While TEOS increases the cohesion
of the grains of the weathered stone via a sol-gel reaction, the gel obtained from
commercial materials based on TEOS has two drawbacks: crack formation during the
drying process and formation of dense fragments of gel inside the stone (Wheeler 1992,
Mosquera 2002, Mosquera 2003), which would block the pores of the stone. This may
promote a significant reduction in the water vapor evaporation through the stone,
preventing the desirable removal of the moisture condensed inside the stone.

Cracking occurs as a result of a differential capillary pressure produced within the
gels during its drying phase (Mosquera 2008) inside and/or outside of the pores of the
decayed stone. Because the capillary pressure is inversely proportional to the radius of
pores filled with the gel network, it is obvious that the dense microstructure having very
small pores promotes high pressures and subsequent cracking of the network of the gels.
Several trials have been carried out to obtain crack-free consolidating materials
through the addition of nanomaterials such as silica nanoparticles (Escalante 2000, 2002;
Agglakopoulou 2002; Alessandrini 2000; Zendri 2007; Mosquera 2005) because larger
pores form in the presence of nanomaterials which reduce the capillary pressure. The
other approach is to enhance the ductile properties of the network, as in the new
perfluorurate polymers (Puterman 1996; Toniolo 2002), organosilicone-modified
polyurethanes (Mazzola 2003) and organic segments, to obtain additional flexibility
(Kim 2007, 2009) which leads to a smoother transition in the drying phase.
On the other hand, there is no specific interaction between the surface of the decayed
stone grains and the consolidants, which would also reduce the consolidation effect of
TEOS-based solution. The study presented here is aimed to provide the interaction
between the TEOS-based gels and the surface of the decayed grains, which give
improved properties for the preservation of cultural stone monuments. We chose (3aminopropyl)
triethoxysilane (ATEOS) as a pre-treatment reagent which would be suit
to the Korean granite. The formation of the gel and characteristics of the treated Korean
granite stone were used to evaluate their efficacies as conservation materials.
2. Experimental
2.1 Materials
TEOS, (3-glycidyloxypropyl)trimethoxysilane (GPTMS), ATEOS,
dibutyltindilaurate (DBLT) were purchased from Sigma-Aldrich Co. Inc. Aerosil® OX-
50 (SiO2, 40 nm) was purchased from Degussa Chemical Co. Ethanol was purchased
from Samchun Pure Chemical Co. Ltd. DBLT, which is also included in the commercial
consolidants, was used as a catalyst for the gelation reaction at a neutral pH. A red dye
(DyeRed 33, Hyundai Chemical) was used for visual observation. Naturally weathered
granite samples from Namsan, Korea were used, and all stone samples were obtained
from one rock specimen for consistency. The specimens were rinsed with deionized (DI)
water for 1 h with ultrasonic agitation and dried in an oven at 100 °C in order to get the
reproducibility. Commercial consolidation agents (Wacker OH 100) were used as
received.
2.2 Preparation methods
ATEOS was applied into the SiO2 nanoparticles (1.21g of ATEOS/1g of SiO2
nanoparticles), which is a mimic model for the Korean granite in order to check the
interaction between the stone and the consolidants. 35wt% of 1:1 TEOS:GPTMS (1T1G)
solution was prepared; DI water was added for a stoichiometric content with respect to
the hydrolysable alkoxy groups of TEOS and GPTMS. Then, ethanol was added to

each the final silicate solid content. 0.08 wt.% of silicate weight of DBLT was added as
catalyst. The solutions were mixed by ultrasonic agitation and then magnetically stirred
for 24 h at room temperature.
Interaction between the modeled SiO2 nanoparticles and consolidants were tested by
the formation of the gel from the treated SiO2 and consolidant solution; various
concentrations of treated silica nanoparticles were added into 1T1G solution and/or
Wacker OH solution. The gels were prepared at room temperature by pouring the
solution into polypropylene dishes. The gelation time was similar for all samples (Kim
2007, 2009; Son 2009), and the drying time was determined gravimetrically when the
treated stone specimens reached a constant weight (ΔM< 0.001 g); it was completed
after 10 days at room temperature. In order to see the adhesion interaction between the
pretreated fresh granite and the consolidants, the ATEOS solution was coated twice to
attach the surface of the granite and the adhesion interaction of the consolidants on the
surface of the pre-treated granite was macroscopically investigated using the ISO 2409
cross-cutting test after 1 month; a red dye was added for visual observation of the
consolidants.
2.3 Characterization
The viscosity of the solution was measured with a vibro viscometer using sine-wave
vibro viscometer SV-10 (A&D Co. Ltd.) at 25°C. Fourier-transform infrared (FT-IR)
spectra of the gels were measured on a Spectrum 100 (Perkin Elmer, Inc.) instrument
equipped with attenuated total reflectance at a resolution of 4 cm −1 . The cross-sectional
morphology of the dried gel was investigated using a scanning electron microscope
JSM-6369 (JEOL Ltd.). The adhesion interaction of the developed consolidants on the
surface of the pre-treated granite was macroscopically investigated using the ISO 2409
cross-cutting test
3. Results and Discussion
In order to optimize the consolidation effect, the consolidants must penetrate deeply
into the pores of the weathered stone. Unless the consolidant penetrates into the
weathered stone until it reaches the intact stone, an internal weakness will be created,
which will eventually lead to further decay phenomena, such as detachments and scaling.
Although the TEOS-based consolidants show excellent properties in the solution phase
since the extremely low viscosity of alkoxysilane monomers allows them to penetrate
deeply into porous stone, sol-gel reaction is in progress in the presence of water, which
increases the viscosity of the solution with time. It is generally accepted that the
penetration of a consolidant depends on its viscosity, therefore the viscosity of the
solutions were measured with time and the result is shown in Figure. 1.
The viscosity of the 35 wt.% of 1T1G solution is lower than that of commercial
Wacher OH 100 in all time period, which is reasonable since the solid content of
Wacker OH 100 is higher than that of 1T1G solution. The viscosity of two solution
increased only by ca. 1mPa·s for a period of time of up to three months, showing no
significant phase change in solution.

Figure 1. Viscosity change of consolidants with storage time.
It is well known that the dried gels obtained from the TEOS based solution
containing Wacker OH were brittle and many cracks formed in the dried gels (Kim 2009,
Son 2009), while the one obtained from 1T1G solution is clear and crack-free. Crackfree
dried gel obtained from 1T1G solution is considered the help of the flexible chains
of the GPTMS. Small amount of silica nanoparticles would provide the crack-free dried
gels due to the increased the size of pore of the network, the high amount of addition of
silica nanoparticles develop the cracks in the gels. Images of the gel obtained from the
Wacker OH 100 and 1T1G solutions with 7wt% of silica nanoparticles are shown in
Figure 2. Cracks are clearly visible with the addition of 7 wt% of silica nanoparticle for
both dried gels.
(a) (b)
Figure 2. The dried gel obtained from (a) 1T1G and (b) Wacker OH 100 solution with 7 wt% of
pristine SiO2.
From the previous research for the decayed granite located in Namsan, Korea, which
is a major stone material for Korean cultural heritage built during the Silla Dynasty (BC
57-AD 935), it is known that highly weathered Namsan granite consists mainly of

orthoclase and quartz (Kang 2007), so well-controlled size of SiO2 nanoparticles were
chosen as a model for the Korean granite.
With the addition of 7wt% of ATEOS treated SiO2 particles the cracking clearly
disappeared, which is shown in Fig. 3a and 3b. However the color of the dried gels
changed to be yellowish and it is significant for the dried gels obtained from 1T1G
containing 7wt% of ATEOS treated SiO2 particles. It is considered to be the reaction
between the epoxide group of GPTMS and amine of ATEOS which attached to SiO2
particle, as confirmed by FT-IR.
(a) (b)
Figure 3. The dried gel obtained from (a) 1T1G and (b) Wacker OH solution with 7 wt% of
ATEOS-treated SiO2 nanoparticles.
For all samples studied here, there was a very pronounced band appearing at
1033cm −1 , along with a less pronounced band at 797 cm −1 which corresponded to the
vibration absorption of Si-O-Si groups, indicating that all samples were mainly
composed of a silica network (Fidalgo 2005). Due to the organic oxirane groups, which
are not involved in the sol-gel reaction, a band lying at 906 cm −1 (Innocenzi 2001) is
shown in 1T1G solution, however it decreases with the addition of ATEOS-treated SiO2
nanoparticles, implying the reaction between the oxide and amine of ATEOS.
The cross-section of the dried gel obtained from Wacker OH 100, 1T1G solution
with different amount of ATEOS treated SiO2 particles, are shown in Figure 4. As can
be seen in the figure, any significant heterogeneous segregation inside of the gel matrix
is not shown. ATEOS treated SiO2 nanoparticles was covered roughly with the matrix,
showing there is an interaction developed between the SiO2 nanoparticles and gel matrix,
implying positive possibility of the adhesion on the surface of the stone with the gel
matrix.
In order to see the effect of the pretreatment of ATEOS on the granite, the adhesion
interaction of the consolidants on the surface of the granite was macroscopically
investigated with the ISO 2409 cross-cutting test. The detached tapes after the adhesion
test of ISO 2409 cross-cutting test are shown in Figure 5. We observed that the adhesion
interaction increased with the ATEOS-treated granite for both of consolidant solutions.

Wacker OH 100/5 wt% of ATEOS-SiO 2 Wacker OH 100/7 wt% of ATEOS-SiO 2
1T1G/5 wt% of ATEOS-SiO 2 1T1G/7 wt% of ATEOS-SiO 2
Figure 4. SEM image of cross-section of dried gel obtained from Wacker OH and 1T1G solution
with different amount of ATEOS-treated SiO2 nanoparticles .
Wacker OH 100 (a) (b)
1T1G (c) (d)
Figure 5. Tape after the adhesion test of Wacker OH 100 (a, b) and 1T1G solution (c, d) with
different amount of ATEOS pretreatment on the granite.

The effect of the pretreatment is not clearly seen for the case of 1T1G since this
consolidant has an intrinsic ductile properties and showed good interaction with the
surface of the granite (Kim 2007, 2008). The adhesion interaction of Wacker OH 100
changed with the addition ATEOS on the surface of the granite, implying that the
functional group of ATEOS has an effect on the adhesion with the granite and Wacker
OH 100 solution. The interaction between the ATEOS-treated granite and the gel
formed by Wacker OH (or 1T1G) is strong enough that there is not significant
detachment from the surface of the granite.
In order to optimize the consolidation effect of the stone consolidants based on
TEOS, the interaction between the consolidants and the stone grain should be strong
enough; otherwise, the independent brittle gel debris formed inside the stone will block
the condensed water evaporation from the inside of the stone, which is responsible for
the secondary decay of the stone. In this research, we applied the interaction by the
pretreatment of granite with functional reagents and the applicability of the developed
pretreatment for the decayed Korean granite is under study in the laboratory.
4. Conclusions
The effect of the pretreatment with the ATEOS was compared with TEOS/GPTMS
solution and Wacker OH 100 commercial product. For the model study, SiO2 particles
were treated with ATEOS and we obtained crack-free dried gel for two TEOS based
consolidants (1T1G and Wacker OH 100). The surface attachment test showed
improved interaction with the TEOS based consolidants including with commercial
products with granite. The increased interaction between the grain and the TEOS-based
would improve the consolidation on the decayed granite, implying its potential for the
preservation of granite heritage in Korea.
5. Acknowledgment
This study was done with the support of the Conservation Technology Research and
Development project, which is hosted by the National Research Institute of Cultural
Heritage Administration. We express our gratitude to this organization.
6. Reference
Aggelakopoulou, E., Charles, P., Acerra, M. E., Garcia, A. I., Flatt, R. J. and Scherer, G.
W. 2002. ‘Rheology optimization of Particle Modified Consolidants’. Materials
Research Society Symposium Proceedings, 712:15-20.
Alessandrini, G., Aglietto, M., Castelvetro, V., Ciardelli, F., Peruzzi, R. and Toniolo, L.
2000. ‘Comparative evaluation of fluorinated and unfluorinated acrylic copolymers as
water-repellent coating materials for stone’. Journal of Applied Polymer Science,
76(6):962-977.

Cnudde, V., Dierick, M., Vlassenbroeck, J., Masschaele, B., Lehmann, E., Jacobs, P.
and Hoorebeke, L. V. 2007 ‘Determination of the impregnation depth of siloxanes and
ethylsilicates in porous material by neutron radiography’. Journal of Cultural Heritage,
8(4): 331-338.
Escalante, M. R., Flatt, R. J., Scherer, G. W., Tsiourva, D. and Moropoulou, A. 2002. In
Protection and Conservation of the Cultural Heritage of the Mediterranean Cities,
Gala´n, E., Zezza, F., (eds). Lisse: A. A. Balkema.
Escalante, M. R., Valenza, J. and Scherer, G. W. 2000. In Proceedings of the 9th
<strong>International</strong> <strong>Congress</strong> on Deterioration and Conservation of Stone, Venice, Italy, 19-24
June 2000. Fassina, V., (ed).; New York: Elsevier.
Fidalgo, R., Ciriminna, L.M., Iharco, M. and Pagliaro. 2005. ‘Role of the alkylalkoxide
precursor on the structure and catalytic properties of hybrid sol-gel catalysts’. Chemistry
of Materials, 17 (26):6686–6694.
Kim, E. K., Park, S. Y., Cho, H. D., Won, J., Do, J. and Kim, S. D. 2007. ‘Development
o alkoxysilane mixed solution as stone preservation and consolidation materials’
Journal of Conservation and Science, 21: 21-32.
Kim, E.K., Son, S., Won, J., Kim, J.-J. and Kim, S. D. 2008. ‘study of adhesion
phenomena of alkoxysilane-type consolidants on fresh granites’ Journal of
Conservation and Science, 23: 1-10.
Kim, E. K., Won, J., Do, J., Kim, S. D. and Kang, Y. S. 2009. ‘Effects of Silica Nano-
Particle and GPTMS Addition on TEOS-Based Stone Consolidants’. Journal of Cultural
Heritage, 10(2):214-221.
Kang, Y. S., Kim, J. J., Won, J. and Kim, H. J. 2007. Development of weathering
mechanism and conservation technology: development of consolidants for weathered
stone, Korea: National Research Institute of Cultural Heritage of Korea.
Mazzola, M., Frediani, P., Bracci, S. and Salvini, A. 2003. ‘New strategies for the
synthesis of partially fluorinated acrylic polymers as possible materials for the
protection of stone monuments’. European Polymer Journal, 39(10):1995-2003.
Mosquera, M. J., Pozo, J., Esquivias, L., Rivas, T. and Silva, B. 2002 ‘Application of
mercury porosimetry to the study of xerogels used as stone consolidants’. Journal of
Non-Crystalline Solids. 311(2):185-194.
Mosquera, M. J., Pozo, J. and Esquivias, L. 2003.’ Stress During Drying of Two Stone
Consolidants Applied in Monumental Conservation’. Journal of Sol-Gel Science and
Technology, 26(1-3), 1227-1231.

Mosquera, M. J., de los Santos, D. M. and Montes, A. 2005. Proceedings of the
Materials Research Society Symposium, Boston, 2005. Warrendale: Materials Research
Society.
Mosquera, M. J., de los Santos, D. M., Montes, A., Valdez-Castro, L. 2008. ‘New
nanomaterials for consolidating stone’. Langmuir, 24(6):2772-2778.
P. Innocenzi, A., Sassi, G., Brusatin, M., Guglielmi, D., Favretto, R., Bertani, A., Venzo,
F. and Babonneau, A. 2001. ‘A novel synthesis of sol-gel hybrid materials by a
nonhydrolytic/hydrolytic reaction of (3-Glycidoxypropyl) trimethoxysilane with TiCl4’.
Chemistry of Materials, 13(10):3635–3643.
Puterman, M., Jansen, B. and Kober, H. 1996. 'Development of organosiliconepolyurethanes
as stone preservation and consolidation materials'. Journal of Applied
Polymer Science, 59(8):1237-1242.
Son, S., Won. J., Kim, J.-J., Jang, Y. D. and Kang, Y.S. and Kim, S. D. 2009. ‘Organic-
Inorganic Hybrid Compounds Containing Polyhedral Oligomeric Silsesquioxane for
Conservation of Stone Heritage’. ACS Applied Materials & Interfaces, 1(2): 393-401.
Tsakalof, A., Manoudis, P., Karapanagiotis, I., Chryssoulakis, I. and Panayiotou, C.
2007 ‘Assessment of synthetic polymeric coatings for the protection and preservation of
stone monuments’. Journal of Cultural Heritage, 8(1):69-72.
Toniolo, L., Poli, T., Castelvetro, V., Manariti, A., Chiantore, O. and Lazzari, M. 2002. ’
Tailoring new fluorinated acrylic copolymers as protective coatings for marble’.
Journal of Cultural Heritage, 3(4):309-316.
Wheeler, G. S., Fleming, S. A. and Ebersole, S. 1992. In Proceedings of the7th
<strong>International</strong> <strong>Congress</strong> on Deterioration and Conservation of Stone, Lisbon, Portugal,
15-18 June 1992; Delgado, J., Henriques, F., Jeremias, F., eds. 1992. Lisbon:
Laborato´rio Nacional de Engenharia Civil.
Wheeler, G. 2005. ‘Alkoxysilanes and the Consolidation of Stone’. The Getty
Conservation Institute in Los Angeles.
Zendri, E., Biscontin, G., Nardini, I. and Riato, S. 2007. ‘Characterization and reactivity
of silicatic consolidants’. Construction and Building Materials, 21(5):1098-1106.

A CLEANING METHOD BASED ON THE USE OF AGAR GELS: NEW TESTS AND PERSPECTIVES
A. Sansonetti 1 , M. Casati 1 , J. Striova 2 , C. Canevali 3 , M. Anzani 4 , A. Rabbolini 4
1 Institute for Conservation and Valorization of Cultural Heritage, ICVBC, CNR, Milan – Florence, Italy
2 National Institute of Optics, INO, CNR, Sesto Fiorentino, Italy
3 Dept. of Materials Science, University of Milano-Bicocca, Milan, Italy
4 Aconerre Conservation Studio, Milan, Italy
Abstract
For the peculiar artworks such as stucco plasterworks cleaning is troublesome for different reasons: weak
mechanical behavior, high porosity and resulting water absorption, partial water solubility in the case of gypsum
plaster. The choice of the cleaning method, especially as concerns gypsum based stucco materials, is crucial in
assuring the lowest harmfulness possible together with a good level of efficacy. With the term agar we name a
powder product composed mainly of polysaccharides and extracted from red algae species. When boiled with water,
in a percentage range in between 0.5 and 5, it produces a colloidal solution that gellifies towards 35 °C. It can be
used gelified and cold placed on a plan stucco surface, tepid and fluid poured on a surface relief or otherwise milled
till a snow consistency, then pressed as a pad onto any surface; it has a high content of water which is slowly
released into the porous substrate system. Hence the water-soluble components of soiling present on the surface are
extracted and removed with the gel. Conservators should tune the following parameters to adjust the cleaning mode
to a specific situation: agar powder concentration in water, application method (gel, tepid solution, “milled”) and
contact time. At the moment, agar gel is successfully used in cleaning gypsum plaster objects with grey coherent
soiling deposits.
In our study agar was used in cleaning a dolostone capital. The power of agar in extraction of soluble salts was
compared for different application methods and concentrations. The ions in the gel were analyzed with the aid of IC
(Ionic Chromatography) and ICP-AES (Inductively Coupled Plasma-Atomic Emission Spectrometry).
Keywords: agar cleaning, salt extraction, ionic chromatography, ICP-AES
1. Introduction
Agar cleaning of artificial stone materials has been already applied on various artworks after its validation by several
studies and lab surveys, in which its efficacy has been discussed and demonstrated (Anzani et al. 2008). At the
moment, Agar gels have been successfully applied even on mural paintings, wooden ceilings and marble sculptures.
Agar Agar is a brand-name for a polysaccharide extracted form red algae of Gracilaria order composed mainly by
Agarose and Agaropectine. Agarose is, due to its high molecular weight, responsible for the gel properties of Agar.
The gel preparation starts by heating the mixture of the agar powder in water to 80-93 °C because of agar insolubility
in water at room temperature. The fluid mixture (sol) is than allowed to cool slowly and at around 30-40 °C it
gellifies. The sol-gel process, in which agar passes from the random coil structure to organized three-dimensional
network, is thermoreversible for several cycles. The melting and gel point temperatures may vary slightly due to the
natural diversity in the Agar chemical composition influenced by the algae growing conditions and
extraction/preparation technique. At room temperature the gel is quite rigid due its macro-reticulate structure based
on the hydrogen bonds among the Agarose helices. This structure also allows to entrap enormous quantities of water
either bonded to the polysaccharide molecules or free.
A general procedure for preparing agar starts by mixing 0.5/5 % in weight agar powder and de-mineralised water;
then it is heated to 90 °C and cooled at room temperature. A solution of agar at 1.5 % is in the range 6.5/7.5 pH units.
Agar gels have been used as aquagel cleaning systems with the ability to confine water action to outer surface of the
substrate on which the gel is applied. Water is released in a controlled manner. The specific action and great
advantage of agar gels, respect to traditional aqueous methods, dwells in its ability to entrap removed soiling into
the gel structure itself.
It can be used in the fluid form, applying it while it is at a temperature in the range 40/50 °C directly on artwork
surfaces; moreover it can be used in gel foils or “unstructured” after having it milled into a snow consistency,
applying it with a spatula.
Agar gel allows to carry out cleaning operation on large or small areas, undercuts and surfaces not easy to reach, on
vertical plan, moldings and ornaments. The formerly described performances are fruitfully used on gypsum plaster,
since the slow and controlled release of water is especially useful when high water absorption capability, porosity
and solubility features of gypsum are considered. Soiling is removed by agar gels by means of both a physical and
chemical action. In fact agar gel, applied as a viscous fluid, forming an adherent thin film on a complete cooling,
exhibit the ability to peel off soiling after its partial solubilization. The advantage consists in avoiding any

mechanical stress of the treated surface taking into account that gypsum plasters are very sensitive to any kind of
scrape, when wet. Still, the water release is limited even in case of agar in fluid and dense form (Figure 1).
Moreover agar gels allow to calibrate the cleaning levels varying agar concentration, time of application and
application technique (if fluid, milled or rigid gel). If the cleaning degree is not sufficient, the whole cleaning
operation can be repeated. As to author experience no residues of organic substances are released onto the substrate
(Anzani et al., 2008)
The conservator is helped in obtaining the desired cleaning level by agar gel transparency; in fact this feature allows
a good control of the cleaning effect and at the same time the surface texture and morphology are magnified by a sort
of “lens effect” which saturates the colors and puts in evidence micro details not visible at the naked eye.
The conservator should choose the correct gel concentration, the temperature of the gel, the time of application;
preliminary tests should address the various parameters.
The viscosity of the solution is strongly dependent on temperature; gel is fluid for a longer time and consequently it
can be applied at lower temperature, if it is prepared at lower concentrations. Agar gel is easy to remove and
apparently no residues remain on the surface. On the contrary other types of gels demand a water wash after the
removing. The rigid gels exhibit the lower adhesion force to the substrate, when compared to fluid and milled ones.
This point should be taken into account when a pulverised surface or a scarce coherent material is under
conservation.
The “milled” or “unstructured” gel is useful in case of sensitive surfaces with a quite high level of relief, where the
adherence and possible air bubbles at the interface could be a problem, giving a not homogeneous cleaning.
When the pure aquagel is not enough, a specific formulae could be tested containing surfactants, or polar solvents as
ethanol or acetone. This choice could be successful when a hydrophobic material should be removed. Even chelating
agents have been tested, for example on iron rust staining.
Currently is under test the use of agar gel as intermediary agent in laser cleaning. This innovative method combines a
laser source (Nd:YAG λ=1064 nm and 532nm; Q-switch or Long Q-switch pulse regimes) with the agar gel. The
laser radiation is addressed through the gel. As a consequence the laser energy is diminished. Advantages regarding a
possible use on laser sensitive pigments such as lead white and cinnabar are under investigation. Moreover an
increase in saturation is obtained when the artwork surface is wet with the agar gel, and this result is supposed to
increase the substrate absorbing power and to limit the increase in surface temperature during laser ablation (Striova
et al. 2012).
In order to test the effectiveness of the agar cleaning, the quantification of several ions (Na + , K + , NH4 + , Mg ++ , Ca ++ ,
NO2 - , NO3 - , F - , Cl - , C2O4 -- , SO4 -- ) present in the aqueous solution extracted from stone materials by agar gels was
performed by ion chromatography (IC) analysis, which in general detects and quantifies ionic species at µg/ml (ppm)
level. The Mg ++ and Ca ++ concentrations present in solutions extracted from stone materials were confirmed by
inductively coupled plasma-atomic emission spectrometry (ICP-AES), which allows to quantify atoms and ions,
mainly metals, down to ppb (µg/l) level with high precision and accuracy.
Fig. 1 Removing agar gel from a gypsum relief

2. Aims of the research
The present research aims at understanding variables that regulate the cleaning process of stone materials using
agar gels. In particular, this study tries to determine the soluble salts extraction performances of agar gels applied
with increasing concentrations and various methods. The cleaning ability is estimated by quantifying the soluble
salts extracted from a natural stone substrate by agar gels. A point of interest is also the comparison of the
concentration of extracted soluble salts with two different analytical methods: Ionic Chromatography (IC) and ICP-
AES Analysis. ICP-AES is mainly used in the field of cultural heritage in geochemistry studies and in the
characterization of metals artworks. As to author’s knowledge, this is the first time that ICP-AES is used with the
aim of detecting a concentration of metals in order to evaluate the effectiveness of a cleaning operation (Sansonetti et
al. 2008).
3. Materials and Methods
In the present text the term “extraction” is used to indicate the uptake of soluble salts from the porous stone
substrate by means of agar gels; while the term “recovery” refers to the quantitative removing of extracted salts from
agar gels, giving a clear solution suitable for quantitative analyses.
3.1 Agar gels
With the term agar we name a powder product composed mainly of polysaccharides and extracted from red
algae species of Gelidium and Gracilariae. Agar is composed of two polymer fractions: agarose and agaropectin.
Agarose is a linear polymer (molecular weight 100.000 – 150.000) whose repeating unit is agarobiose, dimer -(1
3)-β-D-galactopyranose-(1 4)-3,6-anhydro-α-L-galactopyranose (Armisen, R. Galatas, F., 2000). Agarose is
responsible for the gelling properties of agar. Agaropectin is a complex mixture of low-weight saccharide molecules
and contains all the charges units, such as sulfate, pyruvate and carboxilate. Agaropectine has no gelling properties.
The gel network of agarose contains double helices formed from left-handed threefold helices. These double helices
are stabilized by the presence of water molecules bound inside the double helical cavity. The hydroxyl groups are
arranged outside and allow for the aggregation of double helices into higher order assemblies termed suprafibers
(Labropoulos et al, 2000). The agar powder can be dissolved in water at the boiling temperature, in a percentage
ratio by mass in the range 0,5 - 5, to form a colloidal solution. This colloidal solution approximately gelllifies at
35°C, forming a rigid thermoreversible gel that can be re-liquefied by heating it to about 80°C. Melting and regelification
of the gel improves homogeneity and transparency. Gellified agar can be portioned by cutting, and
applied on the surface to be cleaned in pads of appropriate form. It is also possible to apply the solution not yet
gelified at the temperature of 50-60°C. Finally as already said, a further method of application is also to apply the gel
slightly milled, which thus assumes a consistency similar to snow.
3.2 Angera stone capital
Cleaning tests with agar gels were performed on a fragment of a capital in Angera stone that belongs to the collection
of the Civic Museum of Milan. The capital was kept in a warehouse at the Sforza’s Castle in Milan, where a
collection of stone elements coming from dismantled buildings of the town, is on display. The capital is affected by
surface decay phenomena such as pulverization and efflorescences, probably due to incorrect conservation
conditions (Figure 2). Angera stone is a Triassic sedimentary rock, extracted in the southern part of lake Maggiore,
chemically composed by a fine grain dolomite (total open porosity in the range 16-21%) (Alessandrini G., 1993).
Agar gel, as already said, is an effective cleaning system on surfaces with the projection of figures or forms from a
flat background, because of its ability in adhering perfectly to any detail of the profile; in this study a stonework was
chosen as substrate for the tests because of its regular and plain external shape in order to better apply the agar pads
in different areas, with comparable initial conditions (Figure 3). Moreover the aim of this study is to evaluate the
capability of soluble salts extraction by agar, therefore a dolostone represents a good substrate because of the
dolomite very low solubility in water at room temperature. Solubility product of dolomite, expressed in terms of
activities, ranges from 10 -17 to 10 -20 (Jinghwa Hsu K., 1963); hence every ion detected during analyses should come
from decay mechanisms. The pulverization phenomena which affect some part of the capital surface making it
particularly fragile; so the agar gels cleaning method’s advantages may emerge more clearly. A particular care was
taken in selecting test areas that were chromatically as uniform as possible and that did non contain surface
inhomogeneities. X ray diffraction analyses carried out on efflorescences sampled from the surface of the Angera
dolostone capital, detected nitratine (NaNO3) and gypsum (CaSO4 2H2O) as main components. Sampling in different
points was carried out in order to check the homogeneity of soluble salts presence over the stone surfaces.

Fig. 2 The capital at the end of cleaning tests.
3.3 Cleaning tests
The first series of tests was designed in order to compare the extraction ability of gels prepared at equal
concentration of 1% by mass, but applied in three different methods: fluid (colloidal solution at 50-60°C), gelified
(rigid gel at room temperature) and milled (slightly whipped with a mixer). A second series of tests was based on the
equal method of application (gelified pads) but four different concentrations were tested (1, 2, 3 and 4 w % of agar in
water) . Agar gels were prepared as follows: a weighted amount of powder was put into a suitable closed vessel filled
with a known volume of water, then the vessel was heated using a microwave oven until the water began to boil.
After cooling, the gel thus formed was heated again, then it was applied according to the three chosen methods. Agar
was applied in test areas of 25 cm 2 (square 5x5 cm) in order to allow a comparable surface coverage (Figure 3). To
obtain equal pad surfaces, fluid and milled agar was cast inside a suitable mold, once gelified it was cut with a blade
at the correct size. With this method all the pads can be considered uniform in terms of cleaning area. Thickness was
equal as well (1 cm). The contact time was the same for all the tests (3 hours). After this time lapse the pads were
removed from the capital surface and they were sealed in a plastic envelope to be transported to the laboratory. A
blank sample consisting of a 1% gelified pad, in size identical to those applied on the stone, but not used for salt
extraction was also prepared and analysed as comparison.
Fig. 3 Capital of Angera Dolostone during cleaning tests with agar gels 1,2,3 % concentration in water

3.4 Preparation of samples for ion analysis
After stone cleaning, the agar samples were divided in two equal parts (Figure 4); The first part was suspended
in water for 24 hours on a vibrant platform; then the aqueous solutions containing ions recovered from agar was
analyzed by IC (Na + , K + , NH4 + , Mg ++ , Ca ++ , NO2 - , NO3 - , F - , Cl - , C2O4 -- , SO4 -- ) and ICP-AES (Mg ++ and Ca ++ ). A
blank sample consisting of a 1% agar gel, in size identical to those applied on the stone, but not used for salt
extraction was also prepared and analysed as comparison.
The second portion of agar gels were fragmented, filtered under vacuum, then suspended in a small volume of Milli-
Q water (about 3 ml), filtered and thoroughly pressed on the filter, with the aim of recovering of the solution ions.
The suspension in Milli-Q water and the subsequent filtration and pressing were repeated 5 times, in Milli Q water,
in order that ion recovering from agar was quantitative as possible. Then the aqueous portions containing ions were
collected and diluted to the final volume of 25 ml with Milli-Q water and analysed by means of ICP-AES.
Fig. 4 Subdivision scheme of agar pad
3.5 IC analyses
Ionic chromatography analyses were carried out on the water solution extracted from the agar pads. Analyses
were performed with a Dionex ICS-1000 ionic chromatograph. For the anions separation an AS19 4mm column with
a ASRS ultra II self-regenerating suppressor has been used. The elution took place using a NaOH solution as eluent,
with a concentration gradient from 10 to 60 mM and a flow rate of 0,25 ml/min. For the cations separation a CS12
4mm column with a CSRS ultra II self regenerating suppressor has been used. The isocratic elution took place using
a 20 mM methanesulfunic acid solution as eluent with a flow rate of 0,25 ml/min. Injection was made with an
autosampler and the injection volume was 25 µl. For quantitative analysis a specific multi elementary standard for
both cations and anions has been used.
3.6 ICP-AES analyses
Magnesium and calcium concentrations of aqueous solutions extracted from stone materials and recovered from
agar samples were determined by ICP-AES analysis.
Table I provides information on the apparatus and experimental conditions for the determination of Mg ++ and Ca ++
Table I. ICP-AES apparatus specifications and analytical conditions for the determination of Mg ++ and Ca ++ .
Spectrometer Instruments SA, Jobin-Yvon 38 Sequential (France)
Monochromator Czerny-Turner mounting
Nebulizer Meinhard pneumatic
RF power 1.0 kW
Argon flow auxiliary flow:14 l/min
sheath flow: 0.15 l/min
aerosol flow: 0.35 l/min
Nitrogen flow 1 l/min
Sample feed 1.2 ml/min
Analyte line [Mg] 280.270 nm
Analyte line [Ca] 393.366 nm
Reference line [C] 193.091 nm

Calibration was performed using Mg and Ca standard solutions (Aldrich Chemical Co., Milwaukee, WI, USA) in
Milli-Q water.
4. Results
4.1 IC results
Comparing the results obtained for the agar gels at 1% (percentage by mass) applied with different methods,
allows some fruitful considerations (Figure 5). In particular the sulphate content (as regards anions results) and
calcium (as regards cations results) are considered the most significant. These two ions are deemed as particularly
meaningful because they are connected to the sulphation phenomena that affects carbonatic materials such as Angera
stone. The obtained results on sulphate and on calcium content suggest that both the milled and the fluid gels have a
higher extraction force compared to the gelified application method (Fig. 5). Low values of ions detected in the
blank confirmed the absence of soluble species in the agar raw powder used for the preparation of gels. Other ions,
in particular nitrates, show trends more difficult to interpret. This can be explained by considering that these ions are
usually connected to very soluble salts that can migrate and concentrate locally within a inhomogeneous substrate
such as the Angera stone.
The second series of tests concerns gelified pads with increasing concentration of agar powder in water. In this
case the cations concentration trend suggests that agar extraction power improves with increasing concentration.
This consideration is valid also in the case of chlorides and sulphates; on the contrary, nitrites and nitrates show very
high values locally due to the aforementioned high solubility of the salts to which they are possibly associated.
Fig. 5 Concentrations of anions and cations detected with IC analysis.
4.2 ICP-AES analysis
Figure 6 shows the magnesium and calcium concentrations of aqueous solutions extracted from stone materials
and recovered from agar pads prepared with different concentrations of agar (1-4 %) and in different modes (gelified,
milled and fluid).
ICP-AES data for Mg ++ and Ca ++ confirm those already observed by IC, since they show that the gelified agar
sample at 1 % contains the smallest concentration of both the metals, while milled and fluid samples containing agar
at 1 % display the highest concentrations of both ions. These results suggest that milled and fluid agar are able to
extract higher amounts of ions from stone with respect to the gelified agar, By increasing the agar concentration in
the gelified agar, the concentrations of both ions increase almost linearly, suggesting that the higher the agar
concentration, the higher level of extracted ions.

Fig. 6 Concentrations (µg/ml) of Mg ++ and Ca ++ in aqueous solutions extracted from stone materials and recovered from agar
samples prepared with different concentrations of agar (1-4 %) and in different modes (gelified, milled and fluid). Blank sample
was used as comparison (see text).
5. Conclusions and perspectives
The survey focused on the attempt to compare the soluble salts extraction ability of various agar gels applied at
different concentration and with various methods. This kind of survey is very troublesome in real cases because of
salt distribution inside the stone substrate, that we are not able to evaluate precisely as regards its homogeneity.
Anyway the checks by means of XRD on powders sampled from different areas, gave very similar diffraction
patterns.
As regards the various application systems, the better ability in extracting ions by means of milled and fluid agar
is probably due to an higher water release onto the surface, an higher specific surface area and because they can
better adhere to the stone surface compared to the gelified agar pads.
As regards the agar gels with increasing concentration, the ability of salt extraction is proportional to the gel
concentration. A possible chemical role of the polysaccharides has to be taken into account in attempting to explain
these data.
IC and ICP data display a comparable trend as concerns calcium and magnesium (Figure 7); a significant aspect
of this study was to verify the IC reliability in studying salts concentration coming from a stone substrate, compared
with another analytical technique.
Work is in progress in order to understand the transport phenomena of water release at the interface agar/stone
substrate.

Fig. 7 Comparison between results of calcium and magnesium detected by means of IC and ICP-AES.
6. References
Alessandrini G., 1993, ‘Le pietre del monumento’ in La Ca’ Granda di Milano. L’intervento conservativo sul cortile
Richiniano, Silvana Ed. 173-205
Anzani, M., Berzioli, M., Cagna, M. et. al. 2008., ‘Use of Rigid Agar Gels for Cleaning Plaster objects’. Quaderni
del Cesmar7, 6, Il Prato Ed. 35-53
Armisen, R. Galatas, F. 2000, Agar, in Handbooks of Hydrocolloids, Phillips, G.O., Williams, P.A. Eds, CRC Press,
Boca Raton, Florida, 22-40.
Jinghwa Hsu K. 1963 ‘Solubility of dolomite and composition of Florida ground waters.’ J. of Hydrology; Vol. 1. 4.,
288-310.
Labropoulos, K.C., Niesz, D.E., Danforth, S.C. et al. 2002 ‘Dynamic rheology of agar gels: theory and experiments.
Part I development of a rheological models’ Carbohydrate Polymers 50, 393 – 406
Sansonetti, A., Mecchi, A., Poli, T., et. al. 2008., ‘Problems in drawing up standards to evaluate effectiveness and
harmfulness in cleaning operations’paper presented at Conservation Science 2007 Int. Conf. Milan, 10-11 may
2007.Archetype Publ. London (2008), 35-53.
Striova, J., Anzani M., Borgioli L. et al. 2012 ‘A novel approach: laser cleaning intermediated by agar gel’ paper
presented at Lacona IX conference, London, 7 th - 10 th September 2011, in press Taylor & Francis,

THE DEVELOPMENT OF A NEW PREVENTIVE TREATMENT COMBINING WATER
REPELLENT AND ANTI-COLONIZATION PROPERTIES FOR THE CONSERVATION OF
STONE MONUMENTS
Stéphanie Eyssautier-Chuine* 1 , Claire Moreau 1 , Maxime Gommeaux 1 , Céline Thomachot-Schneider 1 , Gilles
Fronteau 1 , Jessica Pleck 2 , Benoit Kartheuser 2 .
1
Groupe d’Étude sur les Géomatériaux et les Environnements naturels Anthropiques et Archéologiques EA
3795 (GEGENAA) – Université de Reims Champagne-Ardenne, CREA, 2, Esplanade R. Garros, 51100
Reims, France.
2
Centre de Ressources Technologiques en Chimie (CERTECH) – rue Jules Bordet, Zone industrielle C,
7180 Seneffe, Belgique.
*Corresponding author : E-mail : stephanie.eyssautier@univ-reims.fr.
Abstract
Water is known as being the main deteriorating agent of building materials. Due to the presence of
water the alteration phenomena observed on calcareous stones are transportation of soluble salts, dissolution
of calcite and development of microorganisms. In restoration works, several chemical treatments can be used
such as biocides as curative treatments to remove existing biocolonization, and water repellents to prevent
the ingress of water and its deteriorating effects.
In the Hybriprotech project, the idea was to combine these two properties (anti-colonizing and waterrepellency)
in one product that would be applied at the end of the restoration work to prevent further
deterioration due to water, and inhibit the development of microorganisms as well. The aim of the project
was to develop products with protective properties close to those of conventional formulations but based on
natural products and safer for the users and environment.
The products developed by the project were tested on a bioclastic limestone both in laboratory and outdoor
conditions, and are compared to commercial water repellents and a preventive biocide. Samples have been
exposed for two years in a forested environment, while the hydrophobicity and the biocolonisation were
evaluated every six months through microdrop test and assessment of chlorophyll content by
spectrophotometry. In the laboratory, anti-biocolonization effects were determined through algae growth
measurements. The product that confirmed an efficient hydrophobicity and anti-biocolonization effect will
be applied after all in a restored site in order to test it in real conditions.
Keywords: limestone, water-repellent, biocolonization, sol-gel, chlorophyll.
Introduction
Several treatments exist that aim at improving the durability of building materials. In the case of manmade
materials, such as concrete, a variety of compounds can be added to the mixture to improve its bulk
properties. In the case of stone, only treatment of the surface (including a varying penetration depth) is
possible. Such surface treatments have in many cases been successfully used (their application being
generally part of broader renovation works). Due to growing concerns as regards the environment and human
health (including that of the workers), there is currently a strong trend towards replacing organic-solvent
treatments with more eco-friendly water-phase treatments, while retaining their protective properties.
The Hybriprotech project, which is a sub-program of the Interreg European Program and gathers French and
Belgian scientists, is part of this effort. This project attempts to create a new product to limit the deterioration
of stone while satisfying sustainable development criteria. Such product is to be less pollutant as well as
equally performing as existing ones.
A high variety of sedimentary rocks, especially limestones, have been traditionally used as building stones in
the Northern part of France and the Southern part of Belgium. Within the Paleozoic basement, dark and nonporous
stones such as the Givetian limestones can be found. Many stones from the Mesozoic and Cenozoic
sedimentary cover of the Paris Basin are light-colored, porous and capillary limestones. Among famous
examples are the Lutetian limestones used for the Reims Cathedral and the Cretaceous limestones (chalks,
russet limestones…) used in a lot of monuments as well.
The climatic conditions in this region, with rainy and cold weather in winter and warm summer generate
wetting-drying cycles and induce damages like cracking blocks, powdering and flaking of stone. The
seepage of rain into the stone enhances the development of biological colonization.
1

One limestone used in buildings of the Belgium-France was chosen for the tests. Different commercially
available treatments used to limit such damages were tested on that stone including one preventive biocide
and water-repellents. Results were used to compare their performance with one of treatments developed over
the course of the Hybriprotech project. However the most advanced product developed in the project had a
combination of two functions: a hydrophobic effect and an anti-biocolonization effect. It has to be applied on
clean stone at the last stage of restoration. Moreover we used the anti-biocolonization properties rather than
biocide ones mostly used to clean soiled stones at the first stage of restoration. The Hybriprotech product
should be therefore a preventive treatment rather than a curative one. This work reports the results of outdoor
and laboratory studies to assess the efficiency of commercial and Hybriprotech treatments in order to
improve the latter and make it as competitive as the commercial ones.
Materials
Stone :
The natural stone on which the products were applied is the Dom stone. It is a Bajocian limestone
(Middle Jurassic, 180 My) mainly used for buildings and monuments, both in the French northern Ardennes
district and in the Belgian southern district of Luxembourg. This limestone is characterized by a high content
of iron oxide (0.5%) giving the russet color. Many quarries were located around the town of Dom-le-Mesnil.
The stone extraction was intense until the end of the Second World War in the French Ardennes district and
its use for building was common from this area to the eastern Belgium like the Liège city. Stone blocks were
collected from the underground quarry which is closed now.
Petrography: The Dom stone is a bioclastic stone [1]; it has been selected to apply the treatments because of
its good homogeneous composition. The facies in thin section revealed calcitic debris: numerous echinoderm
ossicles in a syntaxic cement, secondary shells fragments and some Foraminifera (Figure 1). A few quartz
grains were sparse in the rock. The cementation of the stone is made of two diagenetic stages: a spar calcite
developed around elements and a last blocky calcite that can be partial.
Petrophysics : The petrophysical properties of the Dom stone were measured according to norms EN-1925
[2] and EN-1936 [3]. The Dom stone has a high total porosity (32 ± 3 %), the capillary coefficient C1,
relative to the weight increase per surface and per time unit was 146 g/m²/s ½ and the capillary coefficient C2
relative to wet fringe migration was 26.8 m/s ½ . Such high coefficients showed that the intergranular porosity
observed in thin section was well connected and homogeneous.
Figure 1: Thin section view of the
Dom stone in natural light showing
echinoderm ossicles in a syntaxic
cement.
Products:
Different products were used for the outdoor and the laboratory tests: three commercial products obtained
from the manufacturers and four Hybriprotech products developed over the course of the project.
The commercial products were:
- An alkylpolysiloxane oligomer, the Rhodorsil H224 from BlueStar Silicones (thereafter referred to as
H224), which has water-repellent properties. It was diluted at 10% in an organic solvent (white-spirit).
- An organofunctional silane, the Protectosil SC Concentrate from Evonik (thereafter referred to as Prot.C),
with water-repellent properties. It was diluted in water at 10%.
- A product based on photocatalytic Titanium dioxide, the Photocal, made by Nanofrance technologies
(thereafter referred to as Photocal). It was directly applied to the stones as sold. This product is advertised
self-cleaning and bactericide properties.
2

Several products developed by the Hybriprotech project were tested. They are produced by sol-gel process,
based on the polymerizing ability of tetraethoxysilane (TEOS). Several additives were used, depending on
the desired properties of the final product:
- Chitosan (low molecular weight, Sigma-Aldrich), for antimicrobial property,
- Silver Nitrate, for antimicrobial property,
- Fumed silica (Evonik industries), the surface of which being modified to give a hydrophobic
property.
Four products were developed by the Hybriprotech project:
- TEOS basis with chitosan to get an anti-biocolonization effect (thereafter referred to as T-Chi),
- TEOS basis with silver nitrate to get an anti-biocolonization effect (T-Ag),
- TEOS basis with silver nitrate and hydrophobic silica to get both anti-biocolonization and
hydrophobic effects (T-Ag-Si),
- TEOS basis with silver nitrate, chitosan and hydrophobic silica to get both anti-biocolonization and
hydrophobic effects (T-Ag-Chi-Si).
Experimental procedures
The outdoor test
The test consisted in natural ageing of control and treated stones to reveal the real efficiency of the
commercial products and of a new Hybriprotech product. Surface treatments were applied to series of twelve
samples (dimensions:10x10x5 cm). Samples were exposed in the forested park of Laclaireau castle in Ethe -
Southern Belgium (Coordinates 49°35’12”N - 5°36’08”E, Figure 2) for two years (December 2009 to
December 2011) on a galvanized steel platform, 1 m above the ground and 20° tilted to the South to limit the
stagnation of water and maximize sun-shine (Figure 3).
Before setting up stone samples in the outdoor exposure platform, each block was individually described,
from visual observation as well as using photos. Surface color measurements were conducted as described
below, on the bare surface of the stone as well as after application of the products (if applicable), to ensure
that the products did not induce important change of the stone’s color. Micro-drop test was performed on the
top surface of samples treated with water-repellent products as described below.
Every six months (i.e. after 6, 12, 18 and 24 months of exposure), three samples from each series were
collected. One of them was immediately analyzed for chlorophyll-a content by spectrophotometry. The
remaining two were put in a dry-keeper at 50% relative humidity and 20°C until weight stabilization. After
the conditioning of samples, the surface color was measured and the microdrop test was carried out on the
stones treated with water-repellents.
At the beginning of the test, there were (Figure 4):
- 12 un-treated control samples (1),
- 12 samples treated with H224 by brush (2),
- 12 samples treated with Prot.C by brush (3),
- 12 samples treated with Photocal by brush (4).
One year after the beginning of the test:
- 3 samples treated with a first Hybriprotech formula (T-Chi) by spray.
Figure 2: Laclaireau castle located in
Ethe (close to Virton in the
Luxembourg district, Southern
Belgium.
Figure 3: Outdoor weathering test
with treated limestone samples in a
forested park (Laclaireau castle).
1 2 3 4
Figure 4: 1:12 control samples. 2: 12
samples treated with the waterrepellent
H224 diluted in an organic
solvent. 3: 12 samples treated with the
water-repellent Prot.C diluted in
water. 4: 12 samples treated with
Photocal.
3

The laboratory test of stone biocolonization:
In order to appraise the anti-biocolonization effect of the Hybriprotech formulae, stone samples are
inoculated with a freshwater algae broth and are incubated under neon lights (Sylvania Gro-Lux) for four
weeks at room temperature (18-22°C) [4].
The algal broth is made of demineralized water with BG11- 50 times concentrated medium from Sigma-
Aldrich and Chlorella vulgaris from the CCAP (Culture Collection of Algae and Protozoa, Dunstaffnage
Marine Laboratory, Scotland; strain reference CCAP 211/12).
The algal broth was grown in sterile flasks and an aliquot was transferred every month into fresh medium. It
was diluted to get a similar algal concentration for every test by the chlorophyll-a absorbance control at 665
nm by spectrophotometry. Stone samples (dimensions:5x5x1 cm) were placed in a small cup in which algal
suspension was added until the level reached 2-5 mm above the stone surface. Gravitational settling of the
algae was allowed by letting the samples stand for twenty four hours to obtain a homogeneous seeding. Then
the broth was carefully taken off and demineralized water was added up to 0.5 cm from the bottom of each
sample. Water was added regularly over the whole incubation period to ensure permanent wetness of the
stones due to capillary absorption. Color measurement was performed as described below, before incubation
and after every week of incubation.
Analysis methods
Colorimetry
Color measurements were performed using a Chroma Meter CR-400 by Konica-Minolta with a light
projection tube CR-A33c of 11 mm diameter (corresponding to the measurement zone). Calibrations were
carried out with a white ceramic plate CR-A43. Values are given in the CIE-L*a*b* color space [5]. These
three dimension-less parameters determine the color location in color space: L* indicates lightness (0=
absolute black, 100 = absolute white); a* and b* are the chromaticity coordinates. a* > 0 is the red direction,
-a*< 0 is the green direction; b* > 0 is the yellow direction and -b* < 0 is the blue direction.
Twenty five measurements were performed on 10x10x5 cm blocks (outdoor exposure test) and nine
measurements were performed on 5x5x1 cm blocks (laboratory test). A frame was used to ease the placing of
the colorimeter’s head. The global color variation (ΔE) was calculated from the three color parameters with
the formula: ΔE = . The ΔL*, Δa* and Δb* were the difference between the
measurement on treated stones before exposure for the outdoor test or before incubation for the laboratory
test, and that on stones exposed or incubated, every six months for the outdoor test and every week for the
laboratory test. The ΔE data were the average of ΔE from every measurement.
Spectrophotometry analysis: estimation of the chlorophyll-a concentration
The upper surface of one stone for each series was cut and ground (25 cm 2 ). Then the chlorophyll was
extracted with 95% ethanol (50 ml) at 70°C for three hours. After overnight incubation at 4°C, chlorophyll
concentration estimation was performed by measurement of the absorbance at 665 nm wave-length after
centrifugation (spectrophotometer Thermo Fisher Scientific Genesys 10-S). The conversion factor was
72.3 mg chl/ml/a.u. (absorbance unit).
Water micro-drop contact angle measurement:
The test was used to assess hydrophobicity degree of the surface of the stones, and thus the efficiency of
the water-repellent treatments. For one measurement, a water drop (5 µl) was deposited on the treated
surface of the stone. The angle made by the water drop and the stone surface was measured at once and ten
minutes later. The measurements were made by computer analysis of digital images acquired and processed
by the Advex Instruments system (including the software See System). An average was calculated from
sixteen angles measured on each stone [6].
Outdoor test results:
Colorimetry
After comparison between un-treated and treated un-exposed stones, all commercial treatments fulfilled
the generally accepted requirement (i.e. no visual color change [7], assessing at ΔE ≤ 5 for the Dom stone ),
4

ut the Hybriprotech T-Chi product did not. However, being the first Hybriprotech formula, treated stones
were set up on the outdoor platform in order to evaluate the efficiency of chitosan against biocolonization.
ΔE data were the variation of the stone color before the exposure and after six, twelve, eighteen and twenty
four months of setting in the platform. The results are given in Figure 5.
After six months, the ΔE was 7.4 for control stones that suggested a clear change color. The ΔE of H224
treated stones was 3.1 being the lowest value. The color of the stones treated by the water-repellent H224 has
not changed significantly during the first six-month exposure. The ΔE for Prot.C treated stones was 7.4 and
7.9 for Photocal treated stones. Therefore ΔE obtained with these two treatments were really close to the ΔE
of control stones. The ΔE of Hybriprotech treated stones was 9.9, therefore higher than the others values.
After twelve months, the ΔE of the control stones was 12.0; the change color still increased. The ΔE was
10.9 for H224 treated stones that was close to the ΔE of the control stones, indicating a more important color
change than over the previous six months. The ΔE of Prot.C treated stones was 14.38, being the highest
value. The ΔE of the stones treated by Photocal was 11.2. Therefore the stones treated by commercial
products had an important change color after one year of outdoor exposure.
The ΔE value of stones with Hybriprotech treatment was 7.6 thus lower than the control stones value at the
same exposure time. Moreover, it was lower than the ΔE at six-month exposure as well but the decrease of
ΔE needs to be confirmed by the next analysis after eighteen-month exposure.
After eighteen months, only data with stones with commercial treatments were available to date. The ΔE of
control stones kept increasing, reaching 14.3. ΔE for the water-repellents were 9.8 for H224-treated stones
and 13.9 for Prot.C-treated stones. The ΔE for Photocal was 15.9, much higher than in the control stones.
At twenty four-month ageing, ΔE was 18.1 in the control stones, 12.0 for the H224-treated stones and 15.1
for Prot.C-treated stones. For the Photocal-treated stones, ΔE was 17.0 that was still higher than before.
ΔE
25
20
15
10
5
0
Control H224 Prot.C Photocal
Hybriprotech
Exposure time :
6 months
12 months
18 months
24 months
Figure 5: Global color variation ΔE calculated from data before and
after 6 and 12-months exposure for three samples per each stone
series treated by different commercial products and a Hybriprotech
Spectrophotometric estimation of chlorophyll-a content
The estimation of the extent of biological colonization in tested stones was carried out just after the
collection. Results come from the analysis in spectrophotometer previously described.
After six months of outdoor exposure, the chlorophyll-a content was 1.1 µg/cm² for the control stone (Figure
6). The value from the H224- treated stone was 0.1 µg/cm² whereas the value was 1.2 µg/cm² with the Prot.C
that was very different than the H224 one. The chlorophyll-a content for the Photocal treated stone is 1.3
µg/cm². The result regarding the Hybriprotech treatment was 0.6 µg/cm² which was lower than the control
and Photocal values.
After twelve months, the Chlorophyll-a content for the control stone was 2.4 µg/cm², 0.7 µg/cm² for the
H224-treated stone, 1.3 µg/cm² for the Prot.C-treated stone and 1.7 µg/cm² for the Photocal-treated stone.
The Hybriprotech-treated stone had 1.8 µg/cm² of chlorophyll content. All treated stones had weaker
chlorophyll-a content than the control stone but the values had strongly increased above all the Hybriprotech
and Photocal-treated stones.
After eighteen months of ageing, data came only from the commercial treatments. The chlorophyll-a content
reached 3.5 µg/cm² on the control stone and 1.5 µg/cm² on the H224-treated stone. For the Prot.C-treated
stone, value was 1.5 µg/cm² and 4.9 µg/cm² for the Photocal-treated stone.
After twenty four months of ageing, the chlorophyll content was 3.5 µg/cm² for the control stone, 1.5 µg/cm²
for the H224 treated stone, 2.5 µg/cm² for the Prot.C treated stone and 5.2 µg/cm² for the Photocal treated
stone. Results were still higher at the end of the test except for the H224 treatment where the chlorophyll
content decreased.
5

Chlorophyll a content (µg/cm²)
5,5
5,0
4,5
4,0
3,5
3,0
2,5
2,0
1,5
1,0
0,5
0,0
Control H224 Prot.C Photocal
Hybriprotech
Exposure time:
6 months
12 months
18 months
24 months
Water micro-drop contact angle measurement:
The hydrophobicity effect was evaluated on stones treated with both water-repellents H224 and Prot.C.
Before the exposure in the platform, contact angles were 129.6° and 138.0° respectively on the stones treated
with H224 and Prot. C (Figure 7). These data were compared to requirement [7]: a water-repellent has a
good performance with a contact angle superior or equal to 90°. The contact angles obtained with H224 and
Prot. C treatments were much higher than the required limit. These products had a good hydrophobic effect
before exposure.
At six-month ageing, contact angles were 127.1° and 127.3° then very close to the initial value for the H224
treatment but still 10° smaller for the Prot. C treatment. At twelve-month ageing, the values were
respectively 116.9° and 116.6°, the hydrophobicity had decreased again. At eighteen months of exposure,
contact angle reached 112.9° for H224-treated stones, which showed a good hydrophobicity performance.
Nonetheless, the value for the Prot. C treatment was only 96.4°. Thus the hydrophobicity effect had much
decreased and the value was close to the required limit. At twenty four-month ageing, the data from the last
removing stones, were much higher (124.5° for H224 and 120.6° for Prot.C) whereas we would expect the
contact angles to decrease again.
Average of contact angles (°)
Laboratory test results:
Figure 6 : Chlorophyll-a content measured by spectrophotometry from the
top surface of a control stone and treated stones for two years of exposure
except with Hybriprotech treated stone which had one year of exposure.
150
140
130
120
110
100
90
80
70
60
50
40
0 months 6 m. 12 m. 18 m. 24 m.
Treatments :
Δa* was calculated as the variation of the a* parameter between the value before incubation and after
every week of incubation. Results are given in Figure 8.
After the first week of incubation, the Δa* was -3.0 for the control stones, -2.1 for the T-Ag treated stones, -
0.2 for T-Ag-Si treated stones and -0.3 for T-Ag-Chi-Si treated stones. At this first step of the test, Δa*
values were still different according to treatments.
After the second week, the Δa* was -3.8 for the control stones, -4.1 for the T-Ag treated stones, 0.04 for T-
Ag-Si treated stones and -0.4 for T-Ag-Chi-Si treated stones. Therefore Δa* followed the same evolution
than in the first week with an important decrease of Δa* for the control stones and the T-Ag treated stones
and Δa* values close to zero for the T-Ag-Si and T-Ag-Chi-Si treated stones.
H224
Prot. C
Figure 7: Average of contact angles from treated surface of stones
with water-repellent, H224 and Prot. C, before natural exposure
and after 6, 12, 18 and 24 months exposure.
6

After the third week, the Δa* was -2.0 for the control stones, -1.2 for the T-Ag treated stones, -0.3 for T-Ag-
Si treated stones and -0.2 for T-Ag-Chi-Si treated stones. Data of the control stones and T-Ag treated stones
were less negative than before whereas stones with the other treatments had still a low Δa*.
After the fourth week, the Δa* was -1.9 for the control stones, -1.5 for the T-Ag treated stones, -1.1 for T-
Ag-Si treated stones and -0.4 for T-Ag-Chi-Si treated stones. The values were similar to values after three
weeks of incubation except for the T-Ag-Si stones where Δa* increased.
Δ a*
0
-1
-2
-3
-4
-5
-6
Control
TEOS,
AgNO 3
TEOS,
AgNO 3
Hydro. silica
TEOS,
AgNO 3,
Chitosan,
Hydro. silica
Incubation time :
1 week
2 weeks
3 weeks
4 weeks
Figure 8: Variations of parameter a* calculated from initial
measurements before incubation and measurements after one,
two, three and four weeks of incubation.
Control stone sample Treated stone sample
with TEOS and AgNO3
Treated stone sample
with TEOS and AgNO3
and Hydro. Silica
Treated stone sample
with TEOS and AgNO3,
chitosan and Hydro.
Silica
Figure 9: Stone samples after two weeks of
incubation with algae broth on the top face.
Discussion outdoor:
In this study, treatments were applied to freshly-cut stone samples. In conservation works, treatments
would rather be applied to stone after cleaning or on replacement stone. Different products were tested, but
in no case were two products applied on the same stone. It is well-known that combination of treatments can
have negative side-effects, products interfering with each other and losing part of their efficiency [8].
Instead, one product developed by the Hybriprotech project combined water-repellent and antibiocolonization
functions.
The measurement of stone’s color with Chroma Meter has the advantage of being both quantitative and more
sensitive than naked-eye observation. A few studies even applied this technique to estimate the degradation
state of chlorophyll [9]. This is especially useful in studies dealing with early-stage stone discoloration and
the onset of biological colonization.
Color measurements after outdoor exposure revealed a progressive change of the stones’ color for every
series. The change was significant with exposure times as short as six months. At this date, samples clearly
divided in two groups: (1) stones treated with the H224 water-repellent, with low ΔE value and no change
visible to the naked eye and (2) stones untreated or treated with the water repellent Prot.C or the bactericide
treatments Photocal or Hybriprotech, with higher ΔE values and discoloration visible to the naked eye.
From this time on, the values of ΔE increased on average, but the details vary between treatments. For the
un-treated and Photocal-treated stones, the ΔE values increased steadily until after 24 months of exposure,
suggesting biological colonization was still developing. For the H224- and Prot.C-treated stones, no
important increase of the ΔE values occurred after 12 months of exposure.
After 6, 18 and 24 months, the ΔE values were lowest for the stones treated with the H224 water-repellent.
Only after 12 months did the Hybriprotech formula perform better. This result will be compared to the
situation after 18 and 24 months when the data will be available.
Overall, the results of chlorophyll-a content showed a gradual increase on the control stones. That suggested
a progressive fouling of stones because of biocolonization. Biological colonization was clear as early as after
six months. At this date, only the H224-treated stones had a chlorophyll-a content close to zero, related to a
lower colonization. This is a side-effect of the very efficient water-repellent property of this product. Until
12 months of exposure, the chlorophyll-a content was highest in the un-treated stones, but after 18 and 24
months the chlorophyll-a content in the Photocal-treated stones was the highest value. This suggests that this
treatment did not prevent biological colonization efficiently in this study’s exposure conditions. This result
7

indicates that, despite the high interest that photocatalytic treatments currently arouse [10], this range of
products is not efficient in any situation.
The Hybriprotech T-Chi treatment had good results after six months but after twelve months the chlorophyll
content was the highest value among the treated samples. The two water-repellents had the lowest
chlorophyll content and the H224 treatment had the best result assuming a weak bio-development.
Consequently the H224 water-repellent had an anti-biocolonization effect thanks to its hydrophobicity effect
which avoids stagnation of water on the surface of the sample.
That was confirmed with micro-drop results. The contact angles of H224 and Prot.C decreased slowly but
were well above the 90° required limit. Nonetheless at 18 months of exposure, result of Prot.C was close to
this limit. The un-expected increase of contact angles for Prot. C-treated stones at 24 months was associated
with a steep rise in the chlorophyll content. Thus, the rise of hydrophobicity is likely due to the biofilm
development rather than to the water-repellent itself. Without this biocolonization on Prot.C-treated stones,
the water-repellent wouldn’t have been efficient any longer.
The anti-biocolonization effect of several Hybriprotech formulae developed later in the course of the project
was tested from the artificial growing of algae on stones. To our knowledge, no standard method for
laboratory testing of the efficiency of anti-colonization coatings exists. Most studies dealing with this issue
develop their own method [11] [12]. Based on existing literature, an in-house protocol was developed. This
protocol was aimed at seeding the algae on the surface on stone samples as homogeneously as possible,
resorting to gravitational settling of the algae for seeding and to capillary rise of water to maintain a moist
surface.
Negative values of Δa* showed up the decrease of a* during the incubation of stones and therefore a color
evolution which went from red color to green one. The algal growth can be assessed with a green color
detected by a* measurement.
Δa* data of the control stones revealed a great and progressive color change going to the green until two
weeks of incubation. An algal fouling was achieved by means of water and light on natural stones. During
the two last weeks, the algal development is still negative but less significant probably because of a
temperature increasing to 24°C that decreased the optimal conditions of algal growing.
Results from the stones treated with the TEOS and silver nitrate formula were close to the control stones’
one. The treatment didn’t avoid the algal development and the efficiency needs to be improved by increasing
the silver nitrate concentration.
Both treatments with TEOS, silver nitrate and hydrophobic silica have a good result with a very low Δa*,
that assumes a low algal fouling during the four weeks of incubation. In consequence it has a real antibiocolonization
effect, inhibiting the algae growth. Water is supplied to algae on the top of the sample by
seepage from the bottom nonetheless the hydrophobic silica limits the water on the surface and the algae
growth. Furthermore, the data with chitosan added to the formula shows color variations more stable in time
than without it. So the addition of chitosan, which has a bacteriostatic effect, seems to stabilize the algae
growth over time.
Conclusion
Several products were tested on stone either in outdoor exposure or in artificial incubation of algae in
laboratory. The outdoor exposure tested the real efficiency of three commercial products and the first
Hybriprotech formula during two years in natural conditions of weathering.
Results reveal a good hydrophobicity of the water-repellents H224 and Prot.C. Nonetheless after eighteen
months of exposure Prot.C is less efficient though hydrophobicity is carried out by the biofilm naturally
developed on the stones.
The exposure conditions in a forested park were made to promote a quick biocolonization of stones. The color
and the chlorophyll-a data attest of such a development in the control natural stones as a common result. The
surprising result occurs with the Photocal in which the photocatalytic activity by TiO2 content is not efficient.
The Hybriprotech treatment exposed during one year needs to be more efficient. The bacteriostatic effect of
chitosan is probably not enough to inhibit the biocolonization. The laboratory test showed that the adding of
AgNO3 and hydrophobic silica to chitosan contributes to slow down the algal growth. According to those
results, the Hybriprotech formulae are progressively improved to have a real efficiency either in laboratory
conditions or in outdoor conditions.
Acknowledgments
This work is funded by Interreg IV European grant (Hybriprotech project), co-financed by FEDER,
Région Champagne-Ardenne, Conseil général des Ardennes, Conseil général de la Marne and Région
Wallonne. The authors would like to thank: Eric Hance from Acanthe society (sprl) who allowed the
8

installation of the plat-form in the park of Laclaireau castle. Véronique Vergès-Belmin from LRMH and
Hilde De Clercq from KIK-IRPA for their advices.
References
[1] G. Fronteau. 2000. Comportements télogénétiques des principaux calcaires de Champagne-Ardenne, en
relation avec leur faciès de dépôt et leur séquençage diagénétique, Ph.D, University of Rheims Champagne-
Ardenne, France.
[2] European Committee for Standardization EN 1936, May 2007. Natural stone test methods - Determination
of real density and apparent density, and of total and open porosity. AFNOR Edition.
[3] European Committee for Standardization EN 1925, July 1999. Natural stone test methods -
Determination of water absorption coefficient by capillarity. AFNOR Edition.
[4] W. De Muynck, A. M. Ramirez, N. De Belie and W. Verstraete. 2009. Evaluation of strategies to prevent
algal fouling on white architectural and cellular concrete, <strong>International</strong> Biodeterioration and Biodegradation,
Volume 63, Issue 6, p.679-689.
[5] European Committee for Standardization EN ISO 11664-4, July 2011. Colorimetry - Part 4: CIE 1976
L*a*b* Color space. AFNOR Edition.
[6] European Committee for Standardization EN 15802, January 2010. Conservation of cultural property -
Test methods- determination of static contact angle. AFNOR Edition.
[7] H.R. Sasse, R. Snethlage. 1996. Methods for evaluation of stone conservation treatments, in: N.S. Baer,
R. Snethlage (Eds), report of the Dahlem Workshop:”Saving Our Heritage: the Conservation of Historic
Stone Structures”, Berlin, Germany, March 3-8, pp. 223-243.
[8] C. Moreau, V. Vergès-Belmin, L. Leroux, G. Orial, G. Fronteau, V. Barbin. 2008. Water-repellent and
biocide treatments: Assessment of the potential combinations; Journal of Cultural Heritage, 9, p.394-400
[9] P. Sanmartin, F. Villa, B. Silva, F. Cappitelli, B. Prieto. 2011. Color measurements as a reliable method
for estimating chlorophyll degradation to phaeopigments. Biodegradation, 22, p.763-771.
[10] M.Z. Guo, T.C. Ling, C.S. Poon. 2012. TiO2-based self-compacting glass mortar: Comparison of
photocatalytic nitrogen oxide removal and bacteria inactivation. Building and Environment, 53, p.1-6.
[11] H.-L. Alakomi, M. Saarela, A.A. Gorbushina, W.E. Krumbein, C. McCullagh, P. Robertson, K.
Rodenacker. 2006. Control of biofilm growth through photodynamic treatments combined with chemical
inhibitors: in vitro evaluation methods. In : Heritage, Weathering and conservation : proceedings of the
international conference, HWC-2006, 21-24 June 2006, Madrid, Spain / ed. by R. Fort, M. Alvarez de
Buergo, M. Gomez-Heras & C. Vazquez-Calvo, p. 713-717.
[12] F. Gladis, R. Schumann. 2011. A suggested standardised method for testing photocatalytic inactivation
of aeroterrestrial algal growth on TiO2-coated glass. <strong>International</strong> Biodeterioration & Biodegradation, 65,
p.415-422.
9

COMPARISON OF SALT DECAY SUSCEPTIBILITY OF NHL REPAIR MORTARS
UNDER DIFFERENT TESTING CONDITIONS
Davide Gulotta 1 , Margherita Bertoldi 1 , Cristina Tedeschi 2 , L. Toniolo *1
1
Dipartimento Chimica, Materiali e Ingegneria Chimica “Giulio Natta”, Politecnico di Milano,
Milan, Italy
2 Dipartimento di Ingegneria Strutturale, Politecnico di Milano, Milan, Italy
Abstract
In the present work, the salt decay susceptibility of commercial products for the preparation of
restoration mortars is presented. Four commercial NHL ready-mixed mortars and two commercial
NHL binders have been selected among the most diffused product in the Italian market for the
preservation of the architectural heritage. The salt susceptibility has been tested according to two
standard protocols by RILEM TC: salt resistance of the hardened mortars on cubic specimens
(modified version of MS-A2); and durability test of three-leaf wallettes (MS-A1). In both cases, a
sodium sulphate solution was employed to contaminate the specimens (steady T and RH regime).
The damage evolution was recorded during the tests and the final decay of the mortar cubic
specimens was evaluated as mass variation, while the wallettes degradation was measured by means
of a laser profilometer. The results of the two methods applied on the same mortars are not always
in accordance for what concerning damage rate and extent, confirming the crucial importance of
choosing the most suitable crystallization test conditions for a reliable assessment of durability. The
evaluation of salt susceptibility is confirmed to be a complex issue which depends on a number of
concomitant parameters but the results of the study showed the relevance of the microstructural
characteristics in the damage occurrence.
Keywords: salt decay; crystallization test; natural hydraulic lime; repair mortar; damage evaluation.
1. Introduction
Salt crystallization is widely recognized as a major cause of decay of porous materials of the
architectural heritage (Goudie and Viles, 1997, Charola, 2000, Steiger and Siegesmund, 2007).
When historic masonries are considered, surface scaling, cracking, granular disintegration,
powdering and loss of adhesion of the mortar joints are the main degradation patterns related to the
presence of soluble salts (Doehne, 2002). In such cases the restoration of the bricks/mortar system
integrity is required. In recent years, natural hydraulic limes, NHL according to UNI-EN 459-1
technical standard (UNI-EN, 2002), have been widely employed in the preservation of the built
heritage activity. NHL are commercially available both as anhydrous binders and a as components
of ready-mixed mortars. These latter, in particular, represent a particularly advantageous alternative
to traditionally prepared mortars, as they only require to be added with water followed by simple
mixing operation.
On the other hand, when new materials are introduced in heterogeneous and aged systems, as
historic masonries usually are, they are subjected to the same potentially harmful environment
affecting the original ones and can suffer of similar decay mechanisms. Despite their increasing
* Reference author: lucia.toniolo@polimi.it

diffusion, commercial NHL products are often not adequately supported by a complete knowledge
of the specific compositional features and of the performances at work of the hardened mortars. In
such way, the fulfilment of the compatibility requirements (Maravelaki-Kalaitzaki et al., 2005,
Rodrigues and Grossi, 2007) and the evaluation of the resistance to damaging agents, especially in
term of salt decay behaviour, can hardly be achieved. As a consequence, the overall durability of
the conservative intervention cannot be properly assessed.
Four commercial NHL ready-mixed mortars and two commercial NHL binders for the
conservation of historic masonries were selected among the most diffused product in the Italian
market. All the products are classified by the manufacturers as totally cement-free, with a minimum
content of soluble salts, based on NHL binders and designed for the preparation of bedding mortars.
The present study reports a comparison of the durability results of the hardened mortars respect to
salt decay as indicated by two standard protocols by RILEM TC (RILEM, 1998). Uni-directional
sodium sulfate crystallization tests have been preliminary performed according to a modified
version of the RILEM MS-A2 on cubic mortar specimens. The RILEM MS-A1 three-leaf wallettes
have been prepared and studied as well, in order to take into account the influence of a porous
substrate and to better simulate the real application condition of a restoration mortars in a masonry
system. Mass variation has been employed as a damage indicator of the mortars’ cubic specimens
while the damage evolution of the wallettes has been recorded an monitored by means of laser
profilometry. The durability results of the two protocols have been compared. Moreover, an attempt
has been made in order to correlate the salt susceptibility with specific compositional,
microstructural and mechanical characteristics of each mortar in order to identify the most
significant parameters influencing the damage.
2. Materials and methods
2.1 Commercial products
Four ready-mixed commercial mortars (named M1, M2, M3, M4) and two commercial binders
(named B1 and B2) specifically designed for restoration purpose have been selected. All products
are defined as based on NHL, classified as cement-free and with a minimum content of soluble
salts. The commercial materials are supplied by Basf, Italcementi, Kerakoll, Kimia and Tassullo.
The aggregate used for mortars preparation (mixed with B1 and B2) is a standard quartz-siliceous
sand.
2.2 Mortars preparation
Mortars preparation has been performed according to the indications of the technical data sheet
provided by the suppliers. NHL anhydrous binders have been mixed with the standard aggregate
(B/A ratio 1:3; weight/weight) and a fixed amount of water. Ready-mixed mortars powders have
been simply added with the required amount of water. Mixing operations have been performed with
a mechanical mortar mixer according to European technical standard (UNI-EN, 2005).
4x4x16 cm prismatic specimens have been casted in demountable steel mould, compacted by
mechanical vibration and stored at 20°C - 90% RH for 48 hours. Specimens have been then
removed from the mould, cured at 20°C - 90% RH for 60 days, and divided into 4 cm cubes by
means of a diamond wheel cutter.
Three-leaf wallettes specimens (approx. 250x200x120 mm) have been prepared with traditional
fire-clayed red bricks as porous substrates (Fig. 1a). Each wallette consists of three courses of
bricks with two horizontal bed joints and a vertical one The mortars have been laid between the
brick’s surface, let to harden at 20°C - 90% RH for 48 hours and the wallettes have been finally
stored at 20°C - 90% RH for 60 days.

2.3 Crystallization test
Cubic mortar specimens have been dried at T = 60°C until constant weight and the initial weight
has been recorded. The four lateral faces of the specimens have been sealed in order to allow
evaporation through the upper surfaces alone. A saturated Na2SO4 water solution (anhydrous
Na2SO4 reagent grade, Fluka) has been added until an imbibitions depth of 1 cm from the bottom of
the specimens. The imbibition phase has been carried out for 2 hours at 20°C and followed by the
storage of the specimens in a dessicator for 22 hours at t = 20°C and RH 80% to promote mirabilite
crystallization (Flatt, 2002). The complete imbibition/crystallization cycle (2 hours imbibitions + 22
hours crystallization) has been repeated four times a week (week cycle) and then specimens have
been dried at 60°C. Debris and loose particles have been removed by brushing the upper surface
and the mass variation of the specimens has been measured. The week cycles have been repeated
until a significant damage of the evaporation surfaces has been has been developed by most of the
mortars.
The wallettes have been imbibited through their lower side with a 10% (w%) Na2SO4 solution
(anhydrous Na2SO4 reagent grade, Fluka), stored over a layer of dry gravel and sealed in a plastic
container. In such way the upper side alone has been available as an evaporation surface exposed to
the laboratory environment (20°C and 50% R.H.). Each crystallization cycle has been prolonged for
four weeks. At the end of the cycle the upper surface of the wallettes has been brushed to remove all
the efflorescences and detached materials prior to the profilometry measurements; de-mineralised
water was added inside the sealed box in order to promote further crystallization (Fig. 1b) and a
new 4-weeks cycle has been started.
Fig. 1 - Three-leaf wallette before the crystallization test (a) and stored in a plastic container at the
beginning of a new crystallization cycle (b).
3. Results and discussion
3.1 Commercial products characterisation
The complete characterisation of both the anhydrous commercial products and the related
hardened mortars has been reported in a previous work (Gulotta et al., 2009) and here summarised
in Tab. 1.
The hydraulic behaviour of the commercial products mostly relies on the presence of di-calcium
silicate as larnite, which is the main hydraulic compound in M2, M3, M4 and in binder B2. In
binder B1 larnite is present together with portlandite while in M1 it cannot be traced and portlandite
is the only binding compound detected by XRD. As far as the aggregate fraction is concerned, the
one of M1 and M2 is the most heterogeneous and includes both quartz-siliceous sand and

carbonatic minerals. M3 aggregate fraction has calcite prevalent respect to dolomite and minor
quartz; while the one of M4 is eminently carbonatic and almost entirely composed of dolomite.
The total porosity values of the hardened mortars range between 21,42% of B2 and 34,26% of
M4. The four ready-mixed mortars show a limited variation, with the lower porosities belonging to
M2 and M3. If the median pore radius is considered, M1 is characterised by a particularly fine
porosity with a significant pore concentration in the micro-pores and meso-pores range. The mortars
prepared with the commercial binders (B1 and B2) have a rather mutually comparable porosity,
which is slightly lower than those of the ready-mixed ones, and a quite similar median pore radius.
Tab. 1 - Summary of the main compositional, microstructural and mechanical characteristics of the
initial anhydrous commercial products and of the final hardened mortars.
ANHYDROUS COMMERCIAL PRODUCTS HARDENED MORTARS
Binder
Sample NHL
class
P L
XRD results
F C D M Q
Porosity
[%]
Median pore Compressive
radius strength
[�m] [MPa]
M1 5 ++ - +++ + - ++ +++ 30,37 0,05 9,57
M2 5 - ++ + +++ - + ++ 26,48 0,16 10,06
M3 3,5 - ++ - +++ + - ++ 27,08 0,43 8,84
M4 5 - + - ++ +++ - + 34,26 0,66 4,39
B1 3,5 +++ +++ - + - - ++ 22,59 0,27 4,70
B2 5 - ++ - +++ ++ - ++ 21,42 0,20 7,38
+++ = dominantly present, ++ = present, + = traces, - = not detected
P = portlandite, L = larnite, F = feldspar, C = calcite, D = dolomite, M = muscovite, Q = quartz
The average compressive strength values of the samples after 60 days curing allow the define
three groups: the highest strength is shown by M1, M2 and M3 which all reach at least 8,5 MPa;
M4 and B1, on the other hand, have the lowest final resistance; B2 has an intermediate behaviour.
The results highlights that for all the ready-mixed mortars a clear correlation between the NHL
binder class and the final mechanical behaviour cannot be established, while in the case of B1 and
B2 the 60 days strength results are rather in accordance with the binder classes they belong to.
Finally, it has to be pointed out that mortar M2 showed an unexpected presence of slag fragments
within the hardened binder’s matrix after the petrography analysis. Despite the manufacturer’s
indication, it cannot be defined as a proper natural hydraulic lime based product.
3.2 Salt crystallization test: cubic specimens
Seven week cycles have been performed in order to obtain significant damage in most of the
mortars. The salt crystallization over the mortars’ surfaces takes place according to two main
patterns which can be observed during the initial cycles: elongated and powdered crystals can be
formed all over the evaporation surfaces (M1), as well as a non homogeneous white veil due to the
salt accumulation just beneath the mortars surface (M4). As the imbibition/crystallization cycles
proceed, the former pattern gives rise to massive granular disintegration. The salt accumulation of
the latter one leads to the progressive cracking of the mortars, with delamination and scaling of the
external material. In all cases, further analysis have been conducted in order to characterize the
efflorescences coming out from the substrate. Sodium sulfate is confirmed to be the only soluble
salt present, thus excluding any potential contribute of the mortars’ composition itself to salt
formation.

The damaging rate is reported in Fig. 2, in which the percentage mass variation (100%
corresponds to the initial mass of the samples) is plotted against the crystallization cycles. The first
two cycles corresponds to the initial salt accumulation within the mortars’ porosity. No damage still
occurs during this stage and in all cases a mass increase is detected ranging from extremely limited
(M2) to significant values (M3). During the subsequent cycle, three samples, namely M1, M4 and
B2, start to show a mass decrease due to the ongoing mortar damage. B2, in particular, is
characterized by a rather steady degradation rate which tends to slightly slowing down during the
final cycles. M1 and M4, on the contrary, show an increase in the damaging rate in the last cycle,
which is particularly evident for M1 and determines the worst final performance in terms of
durability. The initial major salt uptake of mortar M3, respect to all others, does not give rise to
significant mass variation during the subsequent stages. This means that after a certain amount of
salt solution is penetrated within the porous substrate, this last is able to let it crystallize without
damage and the specimen remains almost stable during the remaining cycles. Minor variations, if
compared to those of M1, M4 and B2, can be observed in B1, which only shows a slight mass
decrease due to limited damage since the third cycle. Among all mortars tested, M2 is the only one
showing no variations at all, neither during the first stage nor as the crystallization/dissolution
proceeds. This indicates that the salt solution penetration within the mortar is almost completely
inhibited.
Fig. 2 - Damaging rate of mortars’ cubic specimens during the uni-directional salt crystallization
test
The photographic documentation of all specimens at the end of the test is shown in Fig. 3. A
clear correlation between the mass variation and the final damage level to the naked eye can be
observed. The main degradation features of the mortars are described as it follows, according to
decreasing extent of the final damage:
� the massive granular disintegration of M1 has lead to a dramatic loss of material;
� M4 and B2 both show significant damage and the loss of material is mostly located on the
edges of the specimens;

� B1 has minor damage mainly occurring on the edges; M3 only shows a minimal loss of
material from the upper side edges, while the central area of the evaporation surface is
almost undamaged;
� M2 does not show any damage at all.
Fig. 3 - Photographic documentation of the damage level of mortars’ cubic specimens at the end of
the uni-directional salt crystallization test (seven weekly cycles).
3.3 Salt crystallization test: wallettes
Eight crystallization cycles have been performed in order to obtain significant damage in most of
the wallettes. The damage evolution has been monitored by means of a laser profilometer: the
profile pattern of each wallette has been recorded at the end of every cycle along a fixed direction
(transversal direction, crossing three bricks and two horizontal bed mortar joints); profiles of the
same wallette collected at subsequent cycles have been then compared and analysed; the differences
between the initial profile and the following ones has been ascribed to the damage occurrence.
As a significant example, in Fig. 4 is reported the graphical elaboration of a selection of the
profile results of mortar M1. The blue profile describes the surface pattern prior to any
crystallization (indicated as T0) and shows an irregular cross-section of the wallette, in which the
two mortar joints can be easily identified being slightly in relief respect to the contiguous bricks.
After the first crystallization cycle the first degradation already occurs to both the lateral bricks and
the mortar joints (height decrease), while the central brick highlights a height increase. This is due
to a blistering effect induced by salt crystallization taking place just below the brick surface. As a
matter of fact, as it can be observed at the end of the following cycles (cycle III), a massive loss of
material takes place in this location due to delamination and detachment of brick scales. The
damage of the remaining section proceeds at a lower but more constant rate respect to the central
area and blistering can also be detected in joint 2. The major damage of the mortar joints takes place
at the end of the fifth cycle (corresponding to a five months test duration). At the end of the test,
after eight cycles, joint 2 doesn’t show any further loss of material, while joint 1 appears extremely
decayed due to massive granular disintegration and scaling occurred in the meanwhile.

Fig. 4 - Graphical elaboration of the profilometer results of mortar M1 and (below) photographic
documentation of the superficial area of the wallette on which the profiles have been acquired. The
upper border of the image indicates the measuring line followed by the profilometer.
The areas defined by the T0 profile (upper limit) and by the subsequent profiles collected after
each crystallization cycle (lower limit) have been calculated in the two regions indicated by the
vertical dotted line in Fig. 4. These values have been used to monitor the damage rate of the mortar
joints. The initial salt accumulation phase which preceded the damage of the cubic specimens
previously discussed, does not take place. As a matter of fact, the damage of the joint occurs since
the first crystallization cycle in most of the wallettes.M2 and M3 are the only exceptions as they
both begin to show some degradation from the subsequent cycle. M2, M3, M4 and B1 are subjected
to a rather constant damaging rate, while M1 reaches a sort of “critical point” at the end of the fifth
cycle, when massive degradation takes place due the granular disintegration and scaling of the
mortar.
The final damage extent values allow to classify the mortars according to three classes of
resistance to salt decay:
� M1 shows the worst final result, with significant loss of material from the joint which
dramatically increases after the fifth crystallization cycle;
� M2 and B2 are characterised by a medium-level damage and by quite mutually similar
damaging rates especially during the first cycles of the test;
� M3, M4 and B1 show the best durability result, as they only develop very limited damage.
M4 and B1, in particular, develop most of the damage during the first cycles. After the salt
has reached the evaporation surfaces of the wallette, the subsequent
crystallization/dissolution cycles seem to be much less effective in inducing further damage.
As a consequence, the damaging rates tend to slow down and to become quite steady.

4. Conclusions
The results of the durability test of mortar cubic specimens highlight quite different behavior of
the commercial products for what concerning damaging rate, decay patterns and damage location.
At the end of the test, granular disintegration and delamination were observed on most of the
specimens, while some of them only showed superficial crystallization and in one case no damage
at all has been observed. The mass variation seems to be a reliable damage indicator and provides
results which are supported by the naked eye observation of the final damage extent. The
compositional heterogeneity of the commercial products gives rise to quite different microstructural
and mechanical characteristics of the hardened mortars. It is therefore difficult to identify a single
specific parameter which mostly affects the salt susceptibility. However, the presence of a
particularly fine porosity is confirmed to be a weakening factor respect to salt decay, while, on the
other hand, the presence of a high amount of hydraulic compounds seems to generally improve the
mortar resistance.
The salt susceptibility test of wallettes are more representative of the real application conditions
of the restoration materials, but the presence of the brick porous substrates indeed introduces a
further heterogeneity factor. In particular, the brick porosity strongly influence the migration of the
salt solution as well as the evaporation front, which is no longer taking place on the mortar surface
alone. Consequently, part of the damage occurs at the expenses of the brick and this contribution
must be properly considered within the test result. Moreover, the use of a 10% Na2SO4 solution
provided once instead of a saturated one with continuous feeding determines a much slower damage
evolution respect to the cubic specimens. The laser profilometer seems to be able to record the
variations of the wallettes as a result of the salt crystallization effect. Being the testing condition
very different respect to those of the cubic specimens test, the final results are not always in
accordance with those. In particular: whereas the extremely low resistance of M1 is always detected
by both the test procedures; the very good result of M2 is not confirmed when the mortar is coupled
with the brick substrate; both M3 and B1 always show only limited damage while the general
performance of M4 slightly improves in the wallettes test, due to the bricks contribution.
Crystallization test still remains fundamental procedures to preliminary assess the behavior of
hardened materials at work with respect to durability. The test methodology should be properly
chosen according to final application conditions and considering that the more heterogeneous the
materials/systems to be tested are, the more potential variation in the final results obtained with
different procedures have to be expected. In all these cases, a large number of specimens is highly
desirable in order to obtain reliable results.
5. References
Charola, A. 2000. Salts in the deterioration of porous material: an overview. Journal of the
American Institute for Conservation, 39, 327-343.
Doehne, E. 2002. Salt weathering: a selective review. In: Siegesmund, S., Weiss, T. N. &
Vollbrecht, A. (eds.). Natural Stone, Weathering Phenomena, Conservation Strategies and Case
Studies. Geological Society, 51-64.
Flatt, R. J. 2002. Salt damage in porous materials: how high supersaturations are generated.
Journal of Crystal Growth, 242, 435-454.
Goudie, A. & Viles, H. 1997. Salt weathering hazards, Chichester John Wiley.
Gulotta, D., Toniolo, L., Binda, L., Tedeschi, C., Van Hees, R. & Nijland, T. G. 2009.
Investigation of commercial ready-mixed mortars for architectural heritage. Proceedings of the 11th
Conference on Structural Repairs and Maintenance of Heritage Architecture. Tallinn, Estonia, 22-
24 July. WIT Press, 231-241.

Maravelaki-Kalaitzaki, P., Bakolas, A., Karatasios, I. & Kilikoglou, V. 2005. Hydraulic lime
mortars for the restoration of historic masonry in Crete. Cement and Concrete Research, 35, 1577-
1586.
RILEM TC 127-MS. 1998. Tests for Masonry Materials and Structures. Materials and
Structures, 31, 2-19.
Rodrigues, J. D. & Grossi, A. 2007. Indicators and ratings for the compatibility assessment of
conservation actions. Journal of Cultural Heritage, 8, 32-43.
Steiger, M. & Siegesmund, S. (eds.). 2007. Special issue on salt decay. Environmental Geology,
52, 185-420.
UNI-EN 459-1. 2002. Building Lime - Definitions, specification and conformity criteria.
UNI-EN 196-1. 2005. Methods of testing cement - Part 1: Determination of strength.

THE POSSIBILITY OF USING ROMAN CEMENT AS A BINDER OF REPAIR
MORTARS TO RESTORE POROUS STONES
Jadwiga W. Łukaszewicz 1 , Iwona Michniewicz 1
1 Institute for the Study, Restoration and Conservation of Cultural Heritage
Nicolaus Copernicus University, Toruń, Sienkiewicza 30/32, 87-100, Poland
Phone: +48 56 6113832, Fax +48 56 6113852
Abstract
Repair mortars, which could be successfully employed to restore sculptures and
architectonical details made of porous stones, such as the sandstone or light limestone,
were the subject of numerous research, which examined both mineral and organic
binders (Domasłowski, W., 1966; Łukaszewicz, J.W., 2002; Obajtek, M. 1993).
However, until present this issue remains unsolved. The following study covers the
problem of the potential use of Roman cement, the technology of which was redeveloped
in the frames of the European project ROCEM (Adamski, G., Bratasz, Ł.,
Mayr, N. et al. 2009; Weber, J., Gadermayr, N., Kozłowski, R. et al. 2007), as a binder
of repair mortars to restore stone monuments with a porous structure, particularly of
light limestone.
The scope of the research includes a determination of the influence of the following
factors on the qualities of the mortar: type, size and amount of the aggregate used to
formulate a repair mortar. Moreover, the influence of citric acid employed as the setting
retarder was examined. Ready-to-use casting mortar produced in The Division of Glass
and Building Materials in Cracow, Pińczów limestone and Żerkowice sandstone were
employed as reference materials for comparison. In the frames of the study following
properties of the formulated mortars were examined: workability, linear contraction,
setting time, the capability of capillary water transport, density, water absorption,
permeability, compressive and tensile strength, resistance to deterioration mechanisms,
therein to water and soluble salts.
As a result of those investigations, the possibility of employing Roman cement in
art conservation as a binder of repair mortars, especially for the light limestone, can be
concluded.
Keywords: mineral binders, repair mortars, Roman cement, porous stones, sandstone,
light limestone
1. Introduction
Filling losses in stone monuments, especially in those made of porous stones being
exposed in external conditions, should be considered as a treatment of a special
significance, determining the final effect of the whole conservation process. Two
methods of filling losses are commonly in practice: employing natural stone and repair
mortars based on organic or mineral binder. Such mortar should be characterised by the
following features: proper application qualities (workability, plasticity, consistency,
capability of adhesion and binding in a layer of any thickness, adequate working time)
and proper aesthetic qualities (colour and texture). While setting, repair mortar should
show no contraction and produce no harmful substances for the stone (such as acids or

soluble salts). After setting, water absorption, permeability and primarily the capability
of capillary water transport of the repair mortar should be comparable to the original
material and slightly higher. Also tensile and compressive strength of the mortar should
be comparable to the stone, but slightly lower. High corrosion resistance of the repair
mortar is desirable.
Due to the re-appearance of the Roman cement on the conservation market,
researches aiming to determine whether this binder fulfils above-mentioned criteria were
carried out.
Roman cements were commonly employed in the architecture of European
Historicism and Art Nouveau (second half of the 19 th century/ beginning of the 20 th
century) as binders in stucco mortars and elements decorating façades of many buildings.
During years, in places where Roman cement mortars were being exposed to prolonged
performance of rain water, the deterioration of the material was becoming more and
more advanced and conservation treatment became a necessity. However, the problem
of obtaining the original binder occurred, due to the decline of the manufacturing of
Roman cements in 1950s. The need for carrying out restoration work of the built
heritage from this period was underlying the researches and it was performed in the
frames of European projects ROCEM and ROCARE (Hughes, D.C., Weber, J.,
Kozłowski, R. 2010), which aimed at re-developing the technology of producing high
quality Roman cements and at formulating the composition of mortars, putties and
paints. Currently, those are being produced in the Division of Glass and Building
Materials in Cracow, Institute of Ceramics and Building Materials in Warsaw.
2. Materials and methods
2.1 Materials
Mortar: ready-to-use casting mortar, self-formulated repair mortars
Binder: Roman cement
Aggregates: quartz sand (size : 0.125-0.25 mm, 0.25-0.5 mm, 0.5-1.25 mm, 1.25-
2.0 mm), pulverised Pińczów limestone (size: 0.125-0.25 mm, 0.25-0.5 mm)
Additives: citric acid (0.4% of the weight of the cement)
Water-to-cement ratio: 0.6, 0.7, 0.8, 0.9, 1.0, 1.1
Cement-to-aggregate ratio: 1:2, 1:3, 1:4, 1:5, 2:1, 1:1, 1:2
Stones (as the reference materials): Żerkowice sandstone, Pińczów limestone
Both, ready-to-use casting mortar and Roman cement binder were produced in the
Division of Glass and Building Materials in Cracow. Ready-to-use casting mortar was
used to formulate samples to determine optimal curing conditions. The paste was
produced according to technical specifications at the following proportion: 0.35l of
water to 1 kg of ready-to-use dry casting mixture.
Samples used to examine the influence of the sort, size and amount of the aggregate
on repair mortars were formulated from Roman cement binder and respective aggregate.
The pastes were produced at the water-to-cement (w/c) ratio varying from 0.6 to 1.1.
The value of w/c ratio was determined by the workability of the pastes. The quartz sand
of the following sizes: 0.125-0.25 mm, 0.25-0.5 mm, 0.5-1.25 mm, 1.25-2.0 mm and the
pulverised limestone Pińczów of the sizes: 0.125-0.25 mm, 0,25-0,5 mm were used as
aggregates and the cement-to-aggregate (c/a) ratio was from 1:2 to 1:5 and from 2:1 to

1:2 by weight for the quartz sand and the limestone respectively. Furthermore, the
influence of the citric acid employed as setting retarder was examined in relation to
the mortar with limestone aggregate. The percentage was 0.5% in water, which
corresponds to 0.4% related to the weight of cement (Bayer, K., Gurtner, Ch., Hughes,
D.C. et al. 2006; Szeląg, H., Garbacik, A., Pichniarczyk, P. et al. 2008). Pińczów
limestone and Żerkowice sandstone were employed as the reference materials for
comparison.
2.2 Methods
For all measurements, prismatic specimens of 16x4x4 cm were cast in steel
modules in accordance with PN-EN 196-1:2006 standard. All the samples were
demoulded after 48 hours and cured for 14 days. The specimens with quartz sand
aggregate were cured in a cotton cellulose compress, which was being alternately wetted
in distilled water and left for drying. After curing period the samples were cut into
4x4x4 cm cubes and left for 28 days in laboratory conditions to reach constant weight.
The specimens with limestone aggregate were cured in a climatic chamber allowing the
vapour transport, i.e. near 95% relative humidity. After curing period the samples were
cut into 4x4x4 cm cubes and left for 21 days in laboratory conditions to reach constant
weight. Curing period and methods were chosen basing on preliminary tests on readyto-use
casting mortar.
Following properties were determined for the mortar samples: workability, linear
contraction (in accord with PN-55/H-104189 standard), setting time (PN/B-04300
standard), the time of capillary water transport, density, water absorption and
permeability (PN-54/B-04100 and PN-54/B-04101 standards), tensile strength (using
a RMU Testing Equipment SRL M052), compressive strength (using a ToniPRAX
Zwick/Roell, in accord with EN 196, ISO 679 and ASTM C 109, C 349 standards),
freeze-thaw resistance (PN-54/B-04102) and resistance to soluble salts. The workability
of mortars was determined basing on the observation of the freshly formulated pastes.
The time of capillary water transport was measured on 4x4x4 cm mortar cubes
immersed in water at the depth of 1 cm. The resistance to soluble salts was examined on
the specimens which were alternately immersed for 24 hours in 10% Na2SO4 x 10H2O
water solution, then placed in a laboratory drier at 378.15K for 24 hours and finally
placed in desicator with silica gel for 24 hours. Those cycles of examination were being
conducted repeatedly until the samples were destroyed.
3. Results and discussion
The following features of the mortars formulated according to a given technology
after curing period (in the aforesaid conditions) and reaching constant weight were
examined: the time of the capillary water transport, water absorption, permeability,
tensile and compressive strength. Moreover, the freeze-thaw resistance and resistance to
the soluble salts were determined. The results are presented in charts and discussed
therein under.
3.1 Influence of the size of the aggregate on the qualities of Roman cement mortar
Four types of mortars formulated using the quartz sand aggregate of the following
sizes: 0.125-0.25 mm, 0.25-0.5 mm, 0.5-1.25 mm and 1.25-2.0 mm were examined. The
cement-to-aggregate ratio was 1:3 by weight and w/c ratio varied from 0.6 to 0.8.

Table 1. Influence of the size of the aggregate on the qualities of Roman cement mortar
Type of
sample
Roman
cement
repair
mortar
Size of
quartz sand
aggregate
[mm]
Den
sity
[g/c
m 3 ]
Time of
capillary
water
transport
[3 cm]
[min]
Water
absorp
tion
[%]
Permeab
ility [%]
Tensile
strength
[MPa]
Compress
ive
strength
[MPa]
0.125-0.25 1.51 38 16.69 25.05 0.65 1.97
0.25-0.5 1.54 25 18.15 27.82 0.67 2.06
0.5-1.25 1.59 19 15.43 24.55 0.52 1.63
1.25-2.0 1.55 15 14.56 22.53 0.65 1.60
Limestone - 1.71 25 14.24 24.84 - 10.48
Sandstone - 1.97 8 6.68 12.48 - 25.31
The best workability was observed for the mortar with the quartz sand 0.25-0.5 mm
and the worst – for the 1.25-2.0 mm mortar. A linear contraction of mortars was in the
range of 0.23% (for 1.25-2.0 mm mortar) to 0.8% (0.125-0.25 mm mortar). Thereby, the
linear contraction increases pro rata with the increase of the w/c ratio. A setting time
was between 1.5 hour for 1.25-2.0 mm mortar and 27 hours for 0.125-0.25 mm mortar.
This correlation could also be explained by lower w/c ratio for the mortars formulated
with the aggregate of a bigger size.
Water absorption of the examined samples (14.56-18.15%) and limestone (14.24%)
were comparable, as well as their permeability (22.53-27.82% for mortars and 24.84%
for the limestone). Both values, water absorption and permeability were the lowest for
the mortar with the biggest size of the aggregate employed. Thereby, an inversely
proportional correlation could be observed between size of the aggregate and value of
water absorption and permeability of examined Roman cement mortars. All four types
of tested samples were characterised by good capability of capillary water transport (the
time of capillary water transport at the height of 3 cm was 15-38 min, while for the
limestone the time measured was 25 min).
The examined mortars were obtaining relatively low tensile (0.52-0.67 MPa) and
compressive strength (1.63-2.06 MPa), with the highest values for the mortar with 0.25-
0.5 mm aggregate, while the limestone was characterised by a compressive strength of
10.48 MPa.
3.2 Influence of the amount of the aggregate on the qualities of Roman cement
mortar
Four types of mortars formulated using quartz sand aggregate of the 0.125-0.25 mm
size and following c/a ratios: 1:2, 1:3, 1:4, 1:5 by weight were examined. The value of
w/c ratio was determined by workability of the pastes and varied from 0.7 to 1.1 (for 1:2
and 1:5 mortars respectively).

Table 2. Influence of the amount of the aggregate on the qualities of Roman cement mortar
Type of
sample
Roman
cement
repair
mortar
C/a
ratio
Density
[g/cm 3 ]
Time of
capillary
water
transport
[3 cm] [min]
Water
absorpti
on [%]
Permea
bility
[%]
Tensile
strength
[MPa]
Compress
ive
strength
[MPa]
1:2 1.56 53 17.25 26.77 0.80 4.08
1:3 1.43 2 23.17 33.00 0.45 1.36
1:4 1.42 4 22.64 31.95 0.35 1.15
1:5 1.53 5 18.82 28.72 0.40 0.02
Limestone - 1.71 25 14.24 24.84 - 10.48
Sandstone - 1.97 8 6.68 12.48 - 25.31
The best plasticity and workability of the pastes were noticed for samples with the
smaller amount of the quartz sand (1:2 and 1:3 by weight). Those properties were
declining with an increase of the amount of the aggregate employed. The linear
contraction of the examined mortars was in the range of 0.35% to 0.47% and setting
time between 25 hours for 1:2 mortar to 51 hours for 1:5 mortar.
Both, water absorption and permeability of the tested mortars were higher than in
the previous group and amounted to 17.25-23.17% and 26.77-33.0%, respectively.
Among specimens examined in this group 1:2 mortar was indicating capillary properties
the most compatible with the limestone and sandstone tested as the reference material.
Nonetheless, its time of capillary water transport at the height of 3 cm was 53 min, while
for sandstone and limestone it was adequately 8 min and 25 min.
Depending on the amount of the aggregate mortars were indicating tensile strength
between 0.4 MPa to 0.8 MPa and compressive strength in the range of 0.02 MPa to 4.08
MPa. The highest values revealed 1:2 mortar and the lowest were observed for 1:5
mortar.
3.3 Influence of the sort of the aggregate on the qualities of Roman cement mortar
Pulverised Pińczów limestone was chosen as an aggregate due to its common
occurrence in architecture monuments and aesthetic considerations (the mortar based on
Roman cement and pulverised limestone could successfully imitate natural stone). Four
types of mortars were prepared, three of them were formulated with the aggregate of
a mixed size (1 part by weight 0.125-0.25 mm and 1 part by weight 0.25-0.5 mm) and
fourth was formulated with 0.25-0.5 mm limestone. It was expected that using the
aggregate of a mixed size could effect in the increase of mechanical strength of
the mortar. The cement-to-aggregate ratio was 2:1, 1:1, 1:2 by weight for the samples
with limestone of a mixed size and 2:1 by weight for the sample with the 0.25-0.5 mm
aggregate. Due to a porous character of the pulverised limestone, pastes required bigger
amount of water to obtain desirable workability and w/c ratio was from 0.6 to 0.9.

Table 3. Influence of the sort of the aggregate on the qualities of Roman cement mortar
Type
of
sampl
e
Roma
n
ceme
nt
repair
morta
r
Limes
tone
Sands
tone
Sort
of
aggreg
ate
C/a
ratio
Size of
aggregat
e [mm]
Dens
ity
[g/c
m 3 ]
Time of
capillary
water
transport
[3 cm]
[min]
Water
absorp
tion
[%]
Perme
ability
[%]
Tensil
e
strengt
h
[MPa]
Compre
ssive
strength
[MPa]
pulver 2:1 0.25-0.5 1.36 295 24.67 33.15 1.40 5.84
ised 2:1 1 part 1.37 30 24.82 32.88 1.33 6.71
limest 1:1 0.125- 1.56 245 20.54 31.09 1.85 8.47
one 1:2 0.25
1 part
0.25-0.5
1.35 30 27.66 33.96 0.88 1.94
quartz 1:2 0.125- 1.56 53 17.25 26.77 0.80 4.08
sand
0.25
- - - 1.71 25 14.24 24.84 - 10.48
- - - 1.97 8 6.68 12.48 - 25.31
A rapid decay of plasticity was observed for all the pastes after few minutes. The
linear contraction was between 0.06-0.67% and the setting time was in a range of 20
min for the 2:1 mortar with the mixed-size aggregate to 6 hours for the 2:1 mortar with
0.25-0.5 mm limestone.
The examined mortars revealed higher water absorption (20.54-27.66%) and
permeability (31.09-33.96%) than mortars with quartz sand. As predicted basing on
a porous character of a limestone, samples containing less aggregate were showing
lower values of water absorption and permeability. The time of capillary water transport
at the height of 3 cm was between 30 minutes and almost 5 hours.
The tensile strength of the tested mortars was in a range of 0.88 MPa to 1.85 MPa
and their compressive strength between 1.94-8.47 MPa. The best mechanical resistance
was noted for the 1:1 mortar with the mixed-size aggregate, which was the highest result
throughout the research. The samples with the higher c/a ratio were obtaining better
tensile and compressive strength test outcomes.
3.4 Influence of the citric acid employed as setting retarder on the qualities of
Roman cement mortar
The last part of the research was an attempt to improve workability and plasticity of
a Roman cement mortar with the pulverised limestone. The 1:2 mortar with the mixedsize
aggregate (1 part by weight 0.125-0.25 mm and 1 part by weight 0.25-0.5 mm) was
chosen as the one providing relatively good results in earlier tests. Citric acid was
employed as setting retarder at the percentage of 0.5% in water, which corresponds to
0.4% related to the weight of the cement and w/c ratio was 0.9.
The setting retarder considerably improved workability and plasticity of a paste and
the setting time was prolonged to 25 min (for the non-modified mortar it was 20 min).
The linear contraction raised from 0.06% to 0.18%.

The water absorption was slightly lower (25.15% for modified and 27.66% for nonmodified
mortar), while its permeability increased (from 33.96% to 35.72%). The time
of capillary water transport was prolonged from 30 to 43 minutes.
The citric acid influenced substantially mechanical properties of a mortar. The
tensile strength declined from 0.88 MPa to 0.58 MPa, while compressive strength varied
from 1.94 MPa to 1.11 MPa.
3.5 Resistance to the decay mechanisms (freeze-thaw resistance and resistance to
soluble salts)
In the first group of examined samples (1:3 mortars with the quartz sand aggregate
of a size varying from 0.125-0.25 mm to 1.25-2.0 mm) the best freeze-thaw resistance
was observed for the mortar with 0.25-0.5 mm aggregate. The first changes in the
appearance of the samples of this type were noticed in the fifth cycle of the test and the
destruction of the specimens occurred in the last, 20 th cycle. A general regularity could
be concluded that mortars with the bigger size of the aggregate are more prone to low
freeze-thaw resistance. No changes were observed throughout the examination time
neither for the limestone, nor for the sandstone. All types of mortars were characterised
by general good resistance to soluble salts, comparable to the limestone, which revealed
first changes in the appearance of a sample in the second cycle of examination. The
mortars with bigger size of the aggregate employed revealed a slightly better resistance.
This could be explained by the bigger diameter of their pores, which permit longer
crystallisation of the salts without exerting pressure on the walls of the mortar itself.
It was observed that the cement-to-aggregate ratio was determining freeze-thaw
resistance of the tested mortars. The samples formulated with the lower content of
quartz sand were indicating better resistance while examined. The 1:2 mortar revealed
the first changes in the appearance in the sixth cycle of the test and did not undergo
destruction throughout 20 cycles of the examination, while for the 1:5 mortar the decay
processes started in the fourth cycle and samples survived only until 16 th cycle of the
test. The resistance to soluble salts of the examined mortars with the quartz sand and c/a
ratio varying from 1:2 to 1:5 is comparable with limestone. During the test binding
qualities of the samples were decreasing and specimens started to disintegrate. No direct
correlation between the amount of the aggregate in the sample and its resistance to the
soluble salts could be concluded.
The sort of the aggregate was also influencing the freeze-thaw resistance of mortars.
The samples with pulverised limestone were more prone to the destructive action of
water than the specimens formulated with quartz sand, which could be concluded basing
on the porous character of the aggregate. However, mortars with a higher c/a ratio (2:1,
1:1) did not get destroyed throughout 20 cycles of the test. All sort of samples with the
limestone aggregate showed relatively good resistance to soluble salts. General good
density of mortars was noticed during the examination. Deep cracks (Figure 1.) were
observed on the surface of all the samples in the fourth cycle of examination.

Figure 1. Deep cracks observed on the surface of the samples of the Roman cement repair mortar
with the pulverised limestone, the fourth cycle of the resistance to the soluble salts examination
Good freeze-thaw resistance of the mortar with setting retarder was observed. Only
in the 13 th cycle of examination first changes appeared on the surface of the sample,
while for the non-modified mortar they were noticed already in the third cycle.
Specimens with setting retarder did not get destroyed throughout the 20 cycles of the
examination. However, lower resistance to soluble salts was observed. The samples get
damaged in the second cycle of the test in a very characteristic way: walls of the cubes
were spalling and the inside of the sample was dilapidating (Figure 2.). This mechanism
could be the effect of the non-homogenous fast binding process.
Figure 2. The destruction of the samples of the Roman cement repair mortar with setting retarder,
the second cycle of examination of resistance to soluble salts

4. Conclusions
The results obtained in the frames of this research allow to state that following
factors: sort, size and the amount of the aggregate used to formulate a repair mortar
influence its later qualities, as well as employing the citric acid as a setting retarder.
Among all examined types of the Roman cement repair mortars, two of them seem to
represent the most promising qualities: the 1:3 mortar with 0.25-0.5 mm quartz sand and
the 1:2 mortar with the mixed-size pulverised limestone (1 part 0.125-0.25 mm and 1
part 0.25-0.5 mm, by weight). Presented test results lead to conclusion, that both
mortars mentioned could be employed to restore sculptures and architectonical details
made of light and porous stones, particularly of limestone characterised by low
mechanical strength and high water absorption. However none of those mortars could be
considered as a material which fulfils all the requirements that conservators impose on
the ideal repair mortar. One can hope, that further research will help to eliminate those
negative properties of mortars, which could be an obstruction to their common use in
stone conservation.
5. Acknowledgements
The authors would like to thank The Division of Glass and Building Materials in
Cracow, Institute of Ceramics and Building Materials in Warsaw for providing Roman
cement for tests. We also wish to thank Wiesława Topolska, Barbara Kraska, Andrzej
Podgórski and Krzysztof Lisek from The Institute for the Study, Restoration and
Conservation of Cultural Heritage for technical support while carrying out the research.
6. References
Adamski, G., Bratasz, Ł., Mayr, N. et al. 2009. ‘Roman cement – key historic material
to cover the exteriors of buildings’. In Proceedings Pro067 <strong>International</strong> RILEM
Workshop Repair Mortars for Historic Masonry, Groot, C. (ed.) 2-11. Delft:
RILEM Publications SARL.
Bayer, K., Gurtner, Ch., Hughes, D.C. et al. 2005-6. Roman Cement: Advisory Notes (5
volumes), vol. 5 (2006) of a series EU-project ROCEM - Roman Cement to Restore
Built Heritage Effectively, Bradford-Cracow-Litomysl-Vienna.
Domasłowski, W., 1966, Badania nad technologią materiałów do kitowania i
rekonstrukcji kamiennych rzeźb i detali architektonicznych, Zeszyty Naukowe,
Konserwatorstwo i Muzealnictwo II, Toruń
Hughes, D.C., Weber, J., Kozłowski, R. 2010. ‘Roman cement for the production of
conservation mortars’. Paper presented at the 2 nd Historic Mortars Conference &
RILEM TC 203-RHM Repair Mortars for Historic Masonry Final Workshop,
Prague, 22-24 August 2010.
Łukaszewicz, J. W., 2002, Badania i zastosowanie związków krzemoorganicznych w
konserwacji zabytków kamiennych, UMK Publisher, Toruń
Obajtek, M., 1993, Lisek. K., Badanie możliwości zastosowania cementu glinowego w
konserwacji kamiennych obiektów zabytkowych, Naukowe podstawy ochrony i
konserwacji dzieł sztuki oraz abytków kultury materialnej, Strzelczyk, A. Skibiński
S., (ed.) UMK, Toruń.
Szeląg, H., Garbacik A., Pichniarczyk, P. et al. 2008. ‘Cement romański i jego
właściwości cz. II’. Surowce i Maszyny Budowlane, 2/2008: 94-100.

Weber, J., Gadermayr, N., Kozłowski, R. et al. 2007. ‘Microstructure and mineral
composition of Roman cements produced at defined calcinations conditions’.
Materials Characterisation, 58/2007: 1217-28.

A STUDY OF ACRYLIC-BASED MORTAR FOR STONE REPAIR
Thibault Demoulin 1 , Fred Girardet 2 and Robert J. Flatt 1
1 ETH Zürich, Institut für Baustoffe (IfB), Physical Chemistry of Building Materials,
HIF B 60.2 and E 11, Schafmattstrasse 6, CH-8093 Zurich (Switzerland).
2 Rino S.A.R.L, Ruelle Belle Maison 14, CH-1807 Blonay (Switzerland).
Abstract
Repair of altered stone using mortar is an interesting approach, if it avoids a
replacement and extends the lifetime of the original stone. In this paper we examine the
properties of an acrylic dispersion-based mortar that had been developed in the late
seventies for stone repair. The key features of this type of mortar are an easy adaptation
to the stone color, a good workability and reversibility in an organic solvent. A mortar
was prepaired from stone powder and a dispersion of methacrylic ester-acrylic ester
copolymer, and the flexural strength, capillary water absorption, hydric swelling and
thermal expansion were measured and compared with the stone it aims to repair. The
mortar has a low elastic modulus and a relatively low thermal expansion that avoid this
repair material from damaging the original substrate. The high hydric swelling is
compensated by the low elastic modulus in wet conditions.
Keywords: repair mortar, acrylic dispersion, reversibility
1. Introduction
In historical buildings, repair of altered stone using mortar is an interesting
approach, if it avoids a replacement and extends the lifetime of the original material. In
this paper we examine the properties of an acrylic-based mortar that had been developed
in the late seventies in the Ecole Polytechnique Fédérale de Lausanne in Switzerland,
and applied on the townhall. After more than thirty years, the durability of these mortars
and their property of reversibility have relaunched the interest of stone carvers.
The repair mortars (thereafter also called artificial stone) are made by mixing a stone
powder, in which the smaller grain fraction has been removed, with a dispersion of
methacrylic ester-acrylic ester copolymer. The dispersion represents around 20% of the
weight of the stone powder. The hardening of the mortar is due to the drying of the
dispersion and the subsequent coalescence of the polymer particles. Due to the drying
time, it is better indicated for repairs of degradations in the range of centimeters.
Alterations in this size range are commonly found in the weathering pattern of molasse
sandstone, which was widely used in the construction of historical buildings in
Switzerland. The molasse is a sedimentary stone mainly composed of quartz and
felspaths cimented by calcite and clays (Kündig 1997), and is very sensitive to wetting
and drying cycles. These cycles commonly lead to spalling of flakes of 0.5 to 3 cm
(Furlan 1983 ; Jimenez-Gonzalez 2008). The stone is however in good conditions above
this limit. A repair of these weathered stones with natural ones would imply a removal
of potentially sound original material on at least 8 cm, to ensure a good placement,

while the use of an artificial stone would save the original material. This approach is
particularly relevant when applied to finely worked stone that can be repaired instead of
being replaced by a new stonework. Examples of application are presented in Fig. 1.
Fig. 1 (a) Common dimension of spalling in a molasse sandstone - (b) and (c) Application of the
artificial stone as a repair material.
An interesting feature of this mortar lies in its ability to be worked like a stone, and thus
makes its integration possible by the normal tools of a stone carver. It also has the
peculiarity of being solvable in an organic solvent (see Fig. 2). This makes the mortar
easily removable without damaging the underlying original material and facilitates
further treatments that could be applied to the stone for its
conservation. It is interesting to point out that these
features are still observable after thirty years of use in the
townhall of Lausanne However, in some places the two
materials show a differential erosion that could be
attributed to a different durability or to a stress due to the
mortar on the natural stone. This calls for an investigation
of the underlying mechanisms controlling the
compatibility of such mortars with the original stones they
mimic. This article presents the first results made on this
artificial stone, through the study of the elastic modulus,
the capillary absorption, the hydric swelling and the
thermal expansion of a natural stone and on the mortar
made with it.
Fig. 2 The mortar is still
reversible after thirty years.
2. Experimental
The experiments were aimed at examining the possibilities of repairing a molasse
sandstone coming from the region of Bern, the blue Ostermundigen. We thus performed
work both on this stone and on the artificial stone considered to be used as repair
material. The later is made of powder obtained by grinding the natural stone. This
powder is sieved and only the fraction higher than 0.125 mm is used. By doing this it is
intended to obtain a similar granular skeleton to the stone but remove fines that would

pack in between the larger grains and cause a much denser and smoother surface
appearance than the molasses typically have.
The sieved powder is mixed with a commercially available aqueous dispersion of
methacrylic ester-acrylic ester copolymer. The samples were then let to dry at room
temperature during at least four weeks. This allows for film formation of the dispersion
and causes the hardening of this type of artificial stone.
3.1 Color measurement
The color of the wet and dry stones are measured by a spectrophotometer Konica
Minolta CM-700d with illuminant D65 and 10° angle, through an aperture of 8 mm,
L*a*b* space is calculated. The samples were plates of natural and artificial stone of
5*20 cm 2 . It was determined that a minimal number of six measurements should be
taken to obtain a representative color of the samples (Giachetto 1994).
After drying, the difference between the artificial
and the natural stone is hardly distinguishable by
naked eye. However, a slight darkening of the
artificial stone can still be perceived in the dry state.
This can be explained by the fact that the finer
fraction of the grains, which is removed, is lighter
than the higher fraction, and thus induces a
darkening of the mortar. This difference is
attenuated when the stones are wet, suggesting it is an issue of scattering. This tendency
is shown in Table 1; a just noticeable color difference is defined as a difference of 2.3.
3.2 Capillary water absorption
The tests were performed using at least three cylindrical samples of 4.5 cm diameter
and 2.5 cm height for each material, dried to constant mass at 105°C beforehand. The
samples were hung from electronic balances having a resolution of 0.01 g and placed in
contact with water in a closed container to avoid the evaporation of water. The water
uptake was monitored continuously by a computer and the capillary absorption
coefficient (in mg.cm -2 .min -0.5 ) was calculated from the slope of this initial water uptake
(see Fig. 3) and are reported in Table 2.
Table 2. Sorptivity coefficients of the natural and artificial stones. For the natural stone the
orientation of the bedding with respect to the direction of rise is given.
Sample
Natural ⊥
Sorptivity
mg.cm -2 .min -0.5
73.7
Natural // 82
Artificial 9.86
Table 1. Color difference ΔE*
(CIE1976) between the natural and
the artificial stone.
dry 5.6
wet 2.8

The two materials have very different absorption kinetics. The natural stone is fully
impregnated in the range of minutes while the artificial stone needs days.
Fig. 3 Capillary sorption of the artificial stone (full line) and
the natural stone (dash line).
3.3 Hydric expansion
The linear hydric expansion was measured by a LVDT displacement sensor from
TESA with a resolution of 0.1 μm, placed on aluminium tips on both ends of the sample
to avoid any movement of the sensor. Three samples for each material of dimension of
approximately 6 cm were prepared for the test. They were immersed in water and their
expansion continuously recorded.
Both materials swell when they are in contact with water. The swelling of the stone is
attributed to the presence of swelling clays (Steiger 2011) while the swelling of the
mortar is related to the swelling of the polymer. These processes probably have different
time frames, but the difference in sorptivity certainly plays a key role in the difference
of swelling kinetics. It can be noticed that the artificial stone swells more than the
natural stone and the consequences of this are examined further in the paper.
First of all there is a big difference in the swelling kinetics (see .). Only a few minutes
are needed for the stone to swell up to 1.5 mm/m, a few days are needed for the mortar
to reach a value of 2.1 mm/m. The values are reported in Table 3. This value is due to
the spontaneous water absorption, but when water is forced to enter (by vacuum
impregnation) the swelling can go up to 5 mm.m -1 (not shown here).
Table 3 Values of free swelling strain.
Sample Hygric swelling (mm.m -1 )
Natural ⊥
1.46
Natural // 0.69
Artificial 2.1

Fig. 4 Hydric swelling of the natural
stone (dash line) and the artificial
stone (full line).
Fig. 5 Thermal expansion of the artificial
stone (full line) and the natural stone (dash
line). The dependency of the thermal
expansion with the bedding for the natural
stone is reported.
3.4 Thermal expansion
The thermal expansion was measured for both the artificial and the natural stone by
Dynamic Mechanical Analysis in a Perkin Elmer DMA 7e, on cubes of 1 cm 3 . The
thermal range was -30 to +90°C with a heating rate of 1°C.min -1 under nitrogen
atmosphere. Dilation curves are reported in Fig. 5 and the average expansion
coefficients are given in Table 4.
Table 4 Thermal expansion at 20°C.
Sample
Natural ⊥
Thermal expansion
10 -6 K -1
12.9
Natural // 8.9
Artificial
-20 – 10°C 8.6
10 – 25°C 19.5
25 – 60°C 4.9
While acrylics are reported to have high thermal expansion values (Brady 2002),
the expansion of the artificial stone is only slightly higher than the Ostermundigen stone.
This is because most of the mass of this artificial stone is made of stone powder and not
of polymer. However, due to the glass transition (29°C) the thermal expansion is not
linear over the temperature range studied, and shows its higher value, 19.6 °K -1 in the
range of 10 to 25°C.

3.5 Flexural strength test
The flexural strength test provides the flexural moduli, the flexural strengths and the
critical strains of our materials. For this we performed a three-point bending test on a
Zwick 1454 10kN, with a span of 12 cm, for both dry and wet artificial and natural stone.
The dimensions of the samples were 1*5*16 cm 3 , giving a span to thickness ratio of
12:1. Due to the very different plasticity behavior of the materials, the load rate was 1
mm.min -1 for the artificial stone and 0.5 mm.min -1 for the natural stone. At least five
samples for each material and conditions (dry, wet, perpendicular and parallel to the
bedding for the natural stone) were used. The flexural strengths refer to the maximum
strengths developed by the samples during loading. The elastic moduli are calculated
using the slope of the initial elastic part of the stress-strain curves, using the following
relation:

Table 5. Mechanical characteristics of the natural and artificial stone determined by three point
bending.
Flexural modulus
(GPa)
Flexural strength
(MPa)
Critical deformation
(%)
Ostermundigen blue Artificial stone
Bedding Conditions Conditions
⊥
dry
wet
1.87
0.43
dry 0.75
//
dry
wet
1.37
0.36
wet 0.01
Bedding Conditions Conditions
⊥
dry
wet
2.44
0.94
dry 2.82
//
dry
wet
1.87
0.66
wet 0.18
Bedding Conditions Conditions
⊥
dry
wet
0.28
0.36
dry 1.92
//
dry
wet
0.30
0.29
wet 7.23
3. Discussion
The present part focuses on the possible mechanical damage provoked by the
differential response to a hydric or thermal stimuli. The question to be answered is
whether the expansion of the artificial stone can damage the original material. We
examine mainly this case as it would put the natural stone in tension, which is much
more critical than putting it in compression.
The stone after repair can be seen as a juxtaposition of two layers that expand. In the
simplest case, we consider that the stress is homogenous in each layer. In this simplified
situation the equilibrium position in case of differential dilation is satisfied by the
following equation:

1

it is low. This means that very large temperature differences between the artificial stone
extending beyond this temperature range lead to a reduction in differential thermal
dilation. The most critical situation is therefore if the underlying stone is at 10°C and the
mortar overlay is heated to 25°C due to the sun appearing.
Owing to the low elastic modulus in the artificial stone when wet, the worse situation for
the dry stone is if the artificial stone is dry. Whether the underlying stone is dry or wet,
we once again find that the strains do not exceed the elastic limit (about 0.2 mm.m -
1 -1
versus 0.6 mm.m ). While the difference is significant it is less important than in the
case of hydric swelling.
Concerning the artificial stone, the worst situation would be for it to be fully confined.
Here we find that the temperature difference taken above would not lead to straining it
beyond its elastic limit. Therefore we also expect the artificial stone not to be affected
by differences in thermal dilation coefficients.
4. Conclusion
The low elastic modulus of the artificial stone, associated to a relatively small
thermal expansion avoids this repair material from damaging the original substrate.
Moreover, despite a large hydric swelling, its low elastic modulus especially in wet
condition, once again avoids subjecting the original stone to critical stresses. It is also
worth noting that the large elastic limit of the artificial stone together with its low
modulus allow it to accommodate deformations of the stone without being damaged or
damaging the natural stone.
These assumptions obviously do not consider the phenomenon of fatigue due to
repetitive stress cycles that should be studied by experimental means. Another situation
needing consideration is damage at the interface between both materials. While this may
be an issue for the longevity of the repair, it is however not expected to be a cause that
would drive substantial damage to the underlying original stone. These first results are
encouraging parameters for the long term behavior as a repair material. However, since
laboratory tests are not able to completely reproduce the passage of time, a monitoring
of the artificial stone in a newly-repaired building is also ongoing.
Acknowlegments
The authors acknowledge the contribution of Mr Pierre Lachat and Mr Daniel Lachat for
kindly providing the stones used in this study, as well as for helpful discussions and
access to job sites. Mr Richner is also warmly thanked for his constant help with the
equipment and Dr. Busato for his help and patience when we used his DMA.
5. References
Brady S. G., CLauser R. H, Vaccari A.J.. 2002. ‘Materials, their properties and uses’. In
McGraw Hill Materials Handbook 15 th edn, McGraw Hill ed, New York : 1-1066.

Furlan V., Girardet F. 1983. ‘Pollution atmosphérique et dégradation de la molasse’.
Chantiers/Suisse 14 : 989-994.
Giachetto J.Ch., 1994. ‘Caractérisation chromatique de pierre par photocolorimétrie’.
Master thesis. Ecole Polytechnique Fédérale de Lausanne.
Kündig R. 1997. Schweizerische Geotechnische Kommission, ‘Die Mineralischen
Rohstoffe der Schweiz’. ETH-Zentrum, Zürich.
Steiger M, Elena Charola A. 2011. ‘Weathering and Deterioration’. In Stone in
Architecture, 4 th edn., Siegesmund, S. and Snethlage, R. (eds.) : 227-316. Berlin:
Springer.

LINOSTONE: A SANDSTONE COATING PRODUCT INVESTIGATED
A. B. Custance-Baker 1
1 Scottish Lime Centre Trust (research carried out at the University of Edinburgh,
funded by Historic Scotland and supported by the British Geological Survey)
Abstract
A year-long study was carried out to provide the first documented investigation of the
sandstone coating product Linostone to understand the material characteristics,
processes of ‘failure’ and to identify a means by which to remove it from sandstone
buildings. Originally designed for aesthetic improvements on sandstone buildings
(c.1960), Linostone has since been adapted and redesigned to repair spalling or friable
sandstone masonry by incorporating an additional ‘primer coat’. The coating is a resinbased
material incorporating sand and other solid components applied as a paste onto
the masonry in several coats in total no more than c.2mm thick. Sand is cast onto the
wet paste and joints are marked out to give the appearance of ashlar sandstone masonry.
The composition and application procedure, however, is likely to have been frequently
adapted in terms of composition and application procedure by individual contractors.
A block-by-block survey was carried out on a tenement building in Edinburgh both
before and after the removal of a full ‘Linostone’ coating. This documentation allowed
the comparison of the condition of the coating with that of the underlying
masonry. Coating removal was carried out by the use of a combination of superheated
steam and a solvent-based paint stripper/remover. This was found to be generally
successful, but much care was needed in the removal technique.
‘Failure’ of the coating, in the form of blisters, is a common feature of Linostone and
this study identified that ‘failure’ does not generally occur at the coating-sandstone
interface, but within the sandstone. This therefore implies that detachment of a
Linostone coating will damage, or enhance the pre-existing damage, of underlying
stonework. ‘Failure’ is enhanced on face-bedded stones and at the edges of cement
patches. The ability of Linostone to undergo physical change in fluctuating
environmental conditions is believed to influence its detachment and blistering.
Keywords: Linostone, sandstone, Scotland, coating, face bedding, cement
1 Introduction
Scotland is well known for having suffered the ill effects of the stone cleaning boom in
the second half of the 20 th Century (Historic Scotland, 2007a), where inappropriate
methods were used for the removal of thick black pollution crusts.
At the same time other options were being developed for improving the appearance of
blackened buildings, including a sandstone coating product called Linostone. Patented
between 1957 and 1960, Linostone was designed as a thin coating to replicate the

appearance of new ashlar sandstone masonry. The use of Linostone, or similar products
collectively referred to as ‘Linostone’ by the industry, has been widespread throughout
Central Scotland and is still in use today. Research was carried out between November
2007 and January 2009 as the first known investigation into these coating products to
gain a greater understanding of their properties and applications. Observations of
‘failure’ of these coatings are relatively common, mainly in the form of blisters and for
this reason the research aimed to understand the processes of ‘failure’, the effect that it
has on the condition of the masonry and whether it was possible to remove a defective
coating.
1.1 Linostone
At the time this research was undertaken, Veitchi Ltd, a Glasgow-based company, were
the owners and manufacturers of the Linostone product having purchased the rights to it
around 16 years earlier. The in situ coatings that were investigated could not be
confirmed to have been applied using the official product or the official application
procedure but did have the general characteristics of the official product and would have
been referred to by the industry as ‘Linostone’, hence the use of the term in inverted
commas.
The patented Linostone is composed of 42.3 parts polyvinyl acetate resin blend, 6.0
parts water, 20.7 parts lithopone (pigment), 10.3 parts talc and 20.7 parts quartz sand.
The manufacturers say that Linostone can be set apart from other products by its high
solid content of 65-70 per cent, where others are said to have 33-35 per cent solid
content.
The stone substrate should be clean and firm prior to application of the coating. Firstly a
thin primer layer of the Linostone is applied to the stone by brush or spray and then
pressure sensitive tapes are placed over this in the position of the mortar joints. Once
dried a second coat of Linostone mixed with c.5-10 per cent by weight of water is
sprayed or brushed directly on top. Whilst still damp a translucent layer of sand is
applied by spray or hand over the surface and then the tapes are removed. A polyvinyl
acetate resin blend product referred to as Linostone Finish is then applied over the
surface. The finished coating is just 1-2mm thick and has the appearance of finely
jointed ashlar sandstone.
The modern Veitchi Ltd application procedure is similar although now the initial step is
to treat all surfaces prior to Linostone Application with a product known as Linotol
Stabiliser, an acrylic resin liquid that is designed to provide ‘a microporous protective
surface with a key to receive most decorative coatings’ (details from product data sheet).
Two Linostone layers are still specified, although without the alteration to the water
content and the use of the tapes and cast sand has not changed. No finish layer is applied
unlike in the original specifications. The finished colour of the coating comes from a
combination of the pigment included in the Linostone and the colour of the cast sand.
The original Linostone product is understood to have been designed for aesthetic
improvements, whereas the current product is marketed as ‘especially useful where

spalling stonework is required to be made good’ implying that it is a physical repair
product. The introduction of the Linotol Stabiliser probably saw the change in the use
for the product as this would enable the coating to be applied onto more friable surfaces.
The patented Linostone (1960) was designed as a ‘waterproof coating’ that ‘breathes’.
In addition the patent states that ‘during excessive rain [the Linostone] softens slightly
and when drying out, re-hardens. This ensures that the coating is sympathetic to
changing atmospheric conditions.’ When ‘Linostone’ coatings ‘fail’ they tend to do so
by forming blisters indicating that the coating has pulled away from the substrate.
2. Aims and objectives of the research
A broad study was carried out looking at application variability and how the
composition of the sandstone substrate may have affected the subsequent deterioration
of the coating. This paper focuses on the surveys that were carried out on a case study
building in central Edinburgh to understand the processes of failure of this type of
coating. In addition the removal process that took place on the case study building was
assessed for its suitability and success.
3 Methods for the block-by-block surveys
A four storey, mid-terrace, late nineteenth century, ashlar sandstone tenement building
in central Edinburgh acted as a case study during this research. This building had a full
‘Linostone’ coating applied in 1988 which was in a state of deterioration and was to be
removed from the full façade. Several removal methods were trialed and thin sections of
the stone examined to assess the effect this has on the sandstone and if there could be
any potential issues with back-diffusion of the material into the stonework. This could
potentially cause further issues (this work was carried out prior to the research). The
most successful method was found to be the use of the DOFF superheated steam system
to remove the thickness of the coating down to a thin skin and then to apply a paint
stripper/solvent with the trade name Stonehealth No. 6 to soften the coating. This was
left overnight and then washed off at lower pressures with the DOFF system the
following day.
A central section of the building was assessed block-by-block from a scaffold at eye
level and then documented prior to removal of the coating, with a particular record of
where the coating was seen to blister or ‘fail’. Each course of stone was given a number
(from one at ground level to 45 at cornice level) and then each block in each row was
given a letter from left to right so that each block could be easily identified. Prior to the
erection of the scaffold, observations were also made of water dispersal issues on the
building following heavy rainfall. Following removal of the coating, the condition of the
underlying stonework was then assessed to identify the level of damage to each stone,
the presence of ordinary Portland cement (OPC) repairs, the bedding direction of the
stone and the blocks that had been identified for replacement by the project architect.
The surveys produced a set of data that could be compared qualitatively as graphic
representations referred to as ‘block diagrams’ and quantitatively using the statistical chi
squared (χ 2 ) method to identify whether there were any correlations between the
condition of the coating and the condition of the masonry.

4 Results of the block-by-block surveys
Fig. 1. Water dispersal issues Fig. 2. Blistering coating Fig. 3. Coating removal pitting
Fig. 4. Cement patches Fig. 5. Natural bedding direction Fig. 6. Stone replacements
Figures 1-6. Block diagrams displaying the block-by-block survey results.
4.1 Water dispersal issues
During heavy rainfall, prior to erection of the scaffolding, the case-study building
showed four main areas of enhanced water saturation; a point at the right-hand side of
the gutter directly below the chimney where water appeared to be directed toward the
wrong end of the gutter and was subsequently overflowing and draining down the front
of the building; a point half-way down the left-hand side of the façade beside the down

pipe and adjacent bay window ledge; the two string courses between the ground and first
floors; and a significant rising damp issue at ground floor level. This data was not
collected as part of the block-by-block surveys and could therefore not be accurately
correlated with the other data sets from which block diagrams were produced, but a
diagram was produced to represent the observations.
4.2 Blistering coating
In several places on the case study building the coating had a blistered appearance. The
blisters varied from tens of centimetres in diameter down to less than one millimetre. If
a small amount of pressure was applied, the majority of the blisters could be broken in a
brittle manner revealing a cavity behind, either hollow or containing loose sand grains
derived from the stone substrate. Coating material that had detached from the stone
substrate was commonly observed to have several millimetres of sandstone still attached
to its inner surface. The block diagrams for blistered coating and the observations of
water saturation were compared and indicated that blistering is more likely to occur in a
water saturated area.
4.3 Damage to the stonework caused by the coating removal technique
The apparent damage caused by the coating removal technique was recorded as; no
pitting, minimal pitting, slight pitting and strong pitting. The blocks that might require
replacement as a result of damage following coating removal were identified as ‘slightly
pitted’ and those identified as ‘strongly pitted’ were considered to almost certainly
require replacement. Sandstone in close proximity to a cement patch seemed to have a
high chance of suffering from severe pitting.
470 blocks of stone were surveyed following removal of the coating and of these 243
were recorded as having no pitting or minimal pitting, 110 with slight pitting and 84
with strong pitting. 33 blocks were not assessed because either a cement patch was
covering the entire block or no attempt had been made at removal because it was already
clear that the stone would need replacing.
4.4 Cement repairs
Cement on the façade of the case-study building came in two main forms; either as repointing/infilling
of recessed joints or as larger patches of varying size and thickness in
any location. The cement observed around the edges and joints of the masonry blocks is
believed to be a method used for preparing the stone (possibly infilling mortar joints)
prior to the application of the coating. This is not part of the Linostone application
specification given by Veitchi Ltd., but it is a technique that was observed to be in use at
the time of this research on a site where a ‘Linostone’ type coating was being applied.
The larger cement patches on the case-study building pre-dated the application of the
coating, although by how long is not clear. The patches varied in thickness, some
several centimetres thick, which indicated that cement had been used to repair
stonework that was already damaged or decaying, whilst others were only a few
millimetres thick and were more likely to be associated with surface preparation prior to

the coating application. The condition of the cement was variable across the façade with
the majority of the larger patches cracked and often fairly easily separated from the
sandstone substrate by hand and the underlying stone was deteriorated and friable.
Most cement patches did not cover the entire block and observations tended to show that
the coating blistered over the sandstone at the margins of the cement patches rather than
overlying the cement (see figure 7).
Figure 7: The same block on the case study building that showed blistering of the coating prior to
removal (left) and cement patches either side of the area of blistering coating following its
removal (right). A stone core was taken from the block, visible at the base of the right-hand image.
The block diagram for cement shows just the locations of cement patches covering more
than ten per cent of a block’s surface, but does not display the widespread use of cement
along the joints (generally less than ten per cent). The notable areas of cement patching
on the case-study building include a cluster at the first floor level on the left-hand side
and at the base of the building; both areas identified as having high levels of water
saturation following heavy rainfall.
Almost 20 per cent of the masonry blocks on the case-study building had a cement patch
covering at least ten per cent of the surface area. One third of these were located on
rows two to six at ground floor level (row one was not documented). Just over a quarter
(27 per cent) of the masonry blocks adjacent to a window or door were found to have a
cement patch covering more than ten per cent of the surface area. Only 13 per cent of
blocks not adjacent to an opening were found to have a similar cement patch.
4.5 Face bedding
Sandstones have a natural bedding plane that should be placed perpendicular to the
direction of maximum compression in a structure. In ashlar masonry the sandstone
should be placed with the bedding planes horizontal. However, sandstone can often be
found with the bedding planes vertical and either parallel with the outer face of the
building, face bedded, or perpendicular to the outer face of the building, end bedded.
These stones are inherently weaker within a structure. From here on face bedded and
end bedded stones, will be referred to collectively as ‘incorrectly bedded’, which applies
to more than 40 per cent of the blocks. It was not possible to assess this where the
coating had not been removed, where there was a large cement patch or where a dark
pollution layer obscured the stone’s surface; this was the case for 84 of the 470 masonry
blocks. The main indicator of incorrectly bedded stones was the presence of the platey
mineral mica lying parallel with the surface of the stone.

4.6 Stone replacement
It is standard practice on sandstone tenement buildings in Scotland to remove and
replace defective masonry. The number of stone replacements that would be required
could not be identified prior to removal of the coating. One significant area of damage
that was already visible, however, was the cornicing at high level where there had been a
loss of material. In addition, the high number of cement patches visible at ground floor
level indicated that a number of replacement stones would be necessary in this location,
particularly because there were also significant associated internal damp issues.
Following removal of the coating, the project architect marked up the masonry blocks
that would be requiring replacement, as displayed on the block diagram. Not all of the
damaged blocks were to be fully replaced, but where possible, if the damage was
localised at an edge beside a stone being replaced, then this could be replaced as one.
One third (32.5 per cent) of the total blocks on the case study building were marked up
for full replacement and a further six per cent for partial replacement. Many of these
partial replacements were intended to be replaced with a neighbouring stone enabling
replacement as a single block. Ten of the 13 lintels on the case study building were
marked up for replacement. Of these, nine were found to be face bedded; the three
lintels not requiring replacement were all correctly bedded.
4.7 Chi squared (χ 2 ) analyses of block diagram data
Figures 8. & 9. The case
study building prior to
any work being carried
out in 2008 (left) and
following completion of
the work in 2009 (right).
Statistical analyses using the chi squared (χ 2 ) method were carried out between the
different sets of data collected from the block-by-block surveys to identify correlations.
The factors compared against one another were blistered coating, the higher levels of
pitting of the sandstone following removal, cement patches, natural bedding direction
and stone replacements. Null hypotheses and alternative hypotheses were set up for each
statistical analysis. In each case the Null Hypothesis is that there is no relationship
between the two factors studied and the Alternative is therefore that a statistical
relationship exists.

For all data sets the threshold value of χ 2 for one degree of freedom at the 5% level is
3.84 and for 0.1% level is 10.83.
Table 1: Results of the chi squared (χ 2 ) analyses
Slightly or
strongly
pitted
Cement patches
(covering a
surface area over
10% of the block)
Incorrect
bedding
direction
Full stone Partial stone
replacement replacement
Blistering coating 18.2 11.7 22.3 35.0 29.9
Slightly or
strongly pitted
Cement patches
19.0 41.2 65.0 51.5
(covering a
surface area over
10% of the block)
29.8 179.6 154.7
Incorrect bedding
direction
84.4 79.2
Based on the threshold values, as given above, these are all statistically significant
results and all of the Null Hypotheses can be rejected and all of the Alternative
Hypotheses must be accepted. The results therefore show, with 99.9% confidence, that
each of the relationships shown in the table occurred with frequencies that cannot be
explained by random chance. These factors are therefore all positively related.
5 Discussion and conclusions
The information collected from the block-by-block surveys of the case study building
were found to be highly significant as clear correlations between the sets of data were
demonstrated. The observations of water dispersal issues also contributed to an
understanding of the deterioration of the ‘Linostone’ coating in many key areas.
The case study building acted as a good example of a visually typical ‘Linostone’ type
coating with the standard blister features of a deteriorating coating. The presence of
cement patches on the case study building, believed to pre-date the coating, are an
indication that there was already a level of stone deterioration on this façade prior to
application of the coating. The survey results showed that a high level of incorrectly
bedded stones were present on the case study building adding an inherent weakness to
the masonry. This may well be due to the lack of a distinct bedding structure, however
this stone type still had a direction of preferential mineral orientation due to the manner
in which it was formed geologically. The strong correlation between cement patches and
incorrectly bedded stones indicates that these incorrectly bedded stones were weak in
relation to correctly bedded stones within the building.
The condition of the individual blocks of stone on the case study building prior to
application of the coating was unknown. The use of cement as a repair material is likely
to have provided a temporary solution but ultimately to have exacerbated any previous
problems, namely the inherent issues associated with face bedding. The strong presence
of cement patching in the area of rising damp indicates that this has been a long-term

problem, which had most likely been compounded by these repairs. Ease of access for
applying the cement may have influenced the amount of repairs carried out here. The
same could be true for the slightly above average number of applications of cement
around window and door openings as they are easier to access and therefore easier to
repair. It is likely that many, if not all, of the numerous cement patches at high level
were applied in a single phase because access would be difficult without a scaffold.
It is unknown why the decision was taken to apply the coating material, but it is possible
that the high level of cement patches was considered unsightly and a ‘Linostone’ type
coating provided a solution for improving the appearance of the building. Additional
cement that is believed to have been used to prepare the surface of the stone prior to
application of the coating material will have added to the decay issues that are related to
the use of impermeable mortars. It is not known how common this practice is when
applying this type of coating (it is not a practice promoted by the manufacturers), but
this could certainly affect the condition of the underlying sandstone. Although Linostone
is believed to have a low level of breathability, it would still limit the movement of
liquid water and water vapour into and out of the stone. The stonework is likely to have
remained damp for long periods following heavy rainfall because the moisture would
not be able to easily evaporate from the surface due to the coating barrier, and this
would not have been possible at all where cement was present. The coating on the case
study building is considered to have been applied using the modern Veitchi Ltd
procedure including a Linotol Stabiliser layer or similar product. This is a thin layer
which in thin section does not appear to penetrate past the outer layer of mineral grains,
but has a strong bond with them.
Linostone is said to soften when wet (this was demonstrated in simple laboratory tests)
and also was found to increase in flexibility with increasing temperature. If these
adaptations occur where the stone is in a sound state then it may well have no effect. If
however, the stone is either in a deteriorated state or has an inherent weakness, such as
face bedding, then the adaptation of the coating material could well add additional stress
onto the stone’s surface and allow it to pull away from the stone substrate. The strong
bond between the coating and the substrate means that the coating would pull away with
some of the substrate attached which, as identified, could be several millimeters of the
stone’s outer surface. Hall & Hoff (2002) illustrate exactly this, that if the bond between
a surface finish and the underlying substrate is stronger than the backing surface alone,
then this will lead to failure of that backing surface not the applied material. This ties in
well with the observation that blistering coating is correlated with the presence of a
cement patch, because the sandstone at the margins of the cement patch are likely to be
in a state of deterioration. It could therefore be argued that the failure of the coating
increased the deterioration of the sandstone.
The coating removal method of a combination of a paint stripper/solvent and
superheated steam (DOFF) appeared to be generally successful. The DOFF system
appeared to be able to take the coating down to a thin white layer without reaching the
stonework in most places, allowing the solvent to then work through the remaining thin
layer of coating before being washed off at lower pressure with the DOFF system.
Where the stonework was in a good condition the removal was generally successful with

only a low level of surface pitting, possibly as a result of the nozzle for the superheated
steam being too close to the substrate. A stone that is already in an enhanced state of
decay will have a low tolerance to pressure due to the breakdown of its binding
components and any attempt to remove the coating is likely to result in damage. This
was particularly the case, and shown to correlate, in areas where the coating had
blistered and at the margins of cement patches. Each block was individually treated,
however, the latter may be due to higher pressures being used where it appeared that the
substrate of a block was stable (over the cement patch) but that when the hose moved to
the non-cement coated part of a block the pressure was too high and cut through the
outer few millimeters of the stone. The paint stripper/solvent does not work below
temperatures of 5 o C, which may have impacted on its success in this case because it was
used in lower temperatures at some stages of the work.
A large proportion of the stonework on the case-study building is being replaced
following removal of the coating due to its poor condition which is related to the factors
as laid out above. This information gives an insight into the areas of the highest levels of
deterioration but does also include blocks that may have cracked or suffered mechanical
damage, which were not dealt with in this study.
One aim of this study was to create an understanding of what the appearance of the
coating might say about the state of the underlying substrate. The statistical analyses
could therefore be used as a basis for mathematical modeling to predict the probable
level of repair on buildings with similar features of coating deterioration.
We cannot say that ‘Linostone’ type coatings necessarily cause the initial deterioration
of the stonework, but that where they fail they will enhance any underlying stone
deterioration due to the strong bond between them and the sandstone substrate.
This study leads to the question; in what circumstance could you effectively use
Linostone on a sandstone building? If the stonework is sound then there would be no
need to coat it, and if it was in a state of decay then this type of coating may be a
temporary fix but ultimately is likely to exacerbate the deterioration. Many such
methods have been tried over the past century for preserving the structure of our built
heritage, but the end result is that they do not seem to stand the test of time and may
result in more harm than good (Caroe & Caroe, 1984).
6 References
Caroe, A.D.R. & Caroe, M.B. 1984. Stonework: Maintenance and surface repair. CIO
Publishing, London.
Hall, C. & Hoff, W.D. 2002. Water transport in brick, stone and concrete. Spon Press,
London & New York.
Historic Scotland. 2007a. Inform Leaflet; Cleaning Sandstone: Risks and consequences.
Historic Scotland, Edinburgh.

PHILLYSEAL R- TO GOOD TO BE TRUE?
A LESSON FOR THE FUTURE
Stephanie de Roemer
Conservator (3-D Art) Sculpture/Installation Art, Glasgowlife, Glasgow Museums,
Glasgow Museums Resource Centre, 200Woodhead Road, G53 7NN, Glasgow ,
Scotland
Abstract
Phillyseal® R, a two part epoxy putty developed by ITW Polymer Technology as a sea
faring putty has been utilized in the practice of mount making and heritage conservation.
Experience in its use for structural fixings, mounts, and supports in the conservation and
restoration of monumental, fragile and diverse stone sculptures from the collection of
Glasgow Museums, Scotland provided informative data of a material study and
furthered practical skills in the mount making and structural stabilization of varying
large stone artifacts.
Phillyseal® R’s ease of handling in addition to its working, and finishing qualities led to
the resin’s employment as a casting material for the replication of a Viking grave stone
for outdoor display of the newly opened Museum of Transport, Glasgow, in June 2011.
In the first part of this paper three conservation programs involving the utilization of
Phillyseal® R for structural repair and support of stone sculpture are presented.
The second part will focus on aspects of sustainability and permanence in the
availability of materials and how these differ for their assessment of natural materials in
comparison to modern manufactured materials such as Phillyseal® R.
Keywords: structure, Epoxy, conservation, stone, sustainability, availability,
permanence
1. Introduction
Phillyseal® R
The two part Epoxy resin supplied by Philadelphia Resin, as a sea faring putty is also
known in the trade as ‘rat seal’.
As a proven high performance gap filler employed on marine vessels in marine
environments it applied itself to be used and explored practically by mount makers and
conservators for creating high spec structural supports to facilitate safe handling,
installation and display of fragile structures and large scale objects (1, 2).

Figure 1. Phillyseal® R Epoxy resin
The white resin and the black hardener, supplied in 2.5 l tubs are mixed by hand in a 1:1
ratio, similar to mixing two colors of plasticine.
The grey pale blue homogenous coloration of mixed resin serves as a visual reference,
when sufficient mixing for optimum resin performance is achieved. The material retains
its malleability for approximately 1 hour, gradually becoming more and more ridged.
During the various stages of setting different methods and tools can be employed to
achieve high precision surface decoration through detailed modeling, carving and
sanding of the Epoxy putty.
The overall consistency of the mixed resin is similar to plasticine and does not sag or
run off vertical surfaces, another desirable quality for many applications in mount
making, molding and casting.
Phillyseal® R was also found to have excellent adhesive properties; a fresh batch of the
resin can be added to a cured part of the material and will form a solid bond. This allows
a level of care and precision in its application over a long period of time for particularly
large areas to be built up or filled when creating precision and integral supports for a
stone object.
The Epoxy’s compatibility to almost all materials, such as metal, stone, and wood is
specifically useful in its application for structural fixing mechanisms facilitating an
external handling and display system that often included other materials such as wood
and metal screws.
2. Case studies
Battle of Langside Memorial stone, Kelvingrove Art Gallery, 2006
The monumental memorial sandstone was originally erected on the Court Knowe, south
of Cathcart Castle commemorating the spot where Mary Queen of Scots viewed the
downfall of her hopes at the Battle of Langside, May 1568 (3).
For its proposed open display in the newly refurbished Kelvingrove Art Gallery in 2006,
a mount and support base had to be prepared to allow a stable upright positioning of the
stone on its uneven bottom surface.

Figure 2. Uneven bottom surface of memorial stone
Phillyseal® R with its ease of preparation, modeling qualities and chemical and physical
long term strength was selected to build up a profile base as part of the overall mounting
system with metal brackets secured to the wall to provide additional safe and secure
mounting on open display.
The profile of the bottom part of the stone was initially recorded with permanent marker
pen on Melinex, as a template for building up the resin base inside a plastic crate.
Prior to lifting the stone onto the uncured resin, cling film was placed between resin and
stone surface as a physical barrier.
After 48 hours of resin curing time the stone was lifted off and revealed a precise mold
of the bottom face. The resin base was forwarded to the external mount maker to
measure and adjust the metal bracket components and returned to the conservation
studio for coloring the pale blue resin surface.
Acrylic paints were particularly good for achieving a color match resembling the surface
coloration of the stone which rendered the resin base almost invisible.
Figure 3) Memorial stone on display
Roman Distance Slab; St Mungo’s Museum of Religious Life and Art, Glasgow, 2010.

For the temporary exhibition ‘Digging up the Past’ in the St Mungo’s Museum of
Religious Life and Art, a Roman distance slab stone had to be structurally stabilized and
restored.
The local buff sandstone originates from the Antonine Wall, and displays incised Latin
inscription testifying the completion of 3666 and a half paces of the Antonine wall,
between Summerston and Castlehill, by a detachment of the Sixth Legion (4).
Figure 5. Roman distance slab in old mount Figure 6. recording of inscription and surface
Initial conservation condition assessment of the stone object resulted in proposing the
removal of the old wooden frame crate and cement adhering the fragile and porous
sandstone to the wooden backing and to replace the old mount with a system that would
also facilitate safe and secure handling installation and storage of the heavy yet fragile
stone.
15 % Paraloid B72 in acetone was selected as a consolidant for the porous stone to
strengthen the individual stone fragments prior to employing Phillyseal® R as a
structural filler and adhesive to join the two halves and secure the distance slab onto its
purpose made wooden plinth.
To meet the criteria for Phillyseal® R’s reversibility contact surfaces of the stone were
consolidated with two further applications of Paraloid B 72 in Acetone at increasing
concentration. In some cases the prepared surface was further covered with cling film,
particularly where the set resin had to be removed for working the resin component
itself.
If the resin repair is to be removed soaking cotton wool in Acetone and placing it over
the area of stone-adhesive-resin and covering it with cling film softens the Paraloid B72.
With the aid of mechanical tools such as scalpel blades and wooden tooth picks and
localized brushing of acetone the resin component can be safely removed from the stone
surface.
Tests were carried out on samples of Phillyseal® R in achieving surface finishes and
color matching to blend the repair in with the original stone surface in regards to texture
and coloration.
Good results in surface fastening of added materials such as sand, dust and pigments to
the resin and the ease of modeling and carving of the Epoxy during its curing process
allowed for detailed texturing and coloring of the surface.
Stainless steel metal bolts were set into the resin gap fill and fixed the overall restored
stone against a wooden backing board. The stone rests on a wooden plinth on top of a

Phillyseal® R resin mount fixed with screws to a plinth and backing board, providing an
overall integral physical support mechanism.
Figure 7. Mount system with stone Figure 8. Treated stone on display
Viking hog backed replica: Riverside Museum, 2011
As part of the permanent outdoor display for the opening of the Riverside Museum in
2011 a cast of one of the so-called Govan gravestones was to be created.
The 9 th Century ‘hogbacked’ gravestone is one of a number of gravestones and Viking
king’s sacophogae found in the graveyard of Govan Old Parish Church.
A replica of this stone was to be permanently displayed outside the new Museum of
Transport to provide a reference to Govan across the river Clyde, a center for
shipbuilding and for the industrial success of Glasgow in the late 19 th and early 20th
Century.
Prior to molding and casting various plasters and cements were considered and tested to
assess the proposed material’s performance in regards to durability of shape, decoration
and structure in an uncontrollable wet, cold and physically demanding outdoor
environment as found at the West Coast of Scotland.
Through consultation with one of the Riverside Museum project conservators the
remaining amount of Phillyseal® R ordered in bulk for its use in mount making for
specific exhibits meant the usually high cost of the resin did not apply and led to
considering the resin as a potential casting material.
The Epoxy displayed the most desirable performance during preliminary comparable
material testing of various traditional and modern casting materials in regards to
permanence of homogenous and saturated surface coloration and retained good overall
cohesion even after long periods of being submerged in water.
For the actual cast about 3% of black pigment paste was added to the two components
prior to mixing. The decision to color the resin itself instead of painting the surface
afterwards was based on achieving a more permanent stability of the surface color in an
exposed and severe outdoor environment. Additionally potential damage through human
activity such as using the replica as a skate boarding ramp are realistic considerations.

With the installation of the completed replica outside the Riverside Museum- this resin
sculpture also serves as a study of the long term performance of this material.
Figure 9. South face of resin cast in situ Figure 10. North face of resin cast in situ
Shortly after the installation of the replica outside the Riverside Museum a substantial
crack appeared along the horizontal surface of the width of the cast.
The mounting and installation of the replica was carried out employing self-expanding
foam to fill the large volume inside the cast prior to adding cement for allowing strong
hold to stainless steel bars set into the concrete plinth.
The installation had to be rushed to fullfill managerial demands on having the replica on
site for the opening and meant some of the internal materials had not fully set before
being permanently mounted and sealed.
The crack in the outer resin layer was assumed to be a result of physical expansion of
the internal self-expanding foam pushing the two resin halves apart along the seam lines
and filled employing more Phillyseal® R and pigment.
After an unusual spell of sunny and warm weather 11 month later, further cracks
appeared on the south face of the replica. With the cracking occurring along the seam
lines of individually prepared batches the cause for this is thought to be a combination
of thermal expansion of the black resin and a potential decrease of adhesive qualities of
the resin through addition of the pigment paste. The mixing by hand of the two part
epoxy by only 1 member of staff slowed down the application of batches into the rubber
mold achieving only a small number of batches in one working day.

Figure 11. Crack along joining seems of separately cast halves
Excessive heat was the least anticipated scenario for the outside display environment in
Glasgow and tests in assessing the performance of the colored resin in sustained periods
of heat were not considered.
2.1 Epoxy resins as modern materials and challenges relating to their future preservation
Revisiting the ITW Polymer Technology/ Philadelphia Resins web page to order further
quantities of the material for repair work on the replica- it was discovered that the
manufacturer stopped its production of Phillyseal® R in 2004.
After further enquiries via Email an advisor for the company recommended two other
products with similar performance properties to Phillyseal® R and indicated the reason
for its discontinuation in following reply…’The unfortunate nature of Phillyseal “R” is
the fact that a little went a long way. Nobody ever needed 20 gallon kits. We
discontinued it about 7 years ago. Unfortunately, there is just not enough market
demand for this product…’ (5).
The large quantities of Phillyseal® R purchased from a Scottish supplier in 2006 were
the remaining stock that lasted for a further two years after discontinuation of the
product and supported the statement that ‘its demise is the fact that a little went a long
way. Nobody ever needed 20 gallon kits’ (5).
The increase in the resin shell cracking and splitting at the same time of this Email
correspondence appeared to somewhat emphasizethe end of Phillyseal® R.
With the Epoxy ceasing to exist and therefore no longer an available material to the
conservator the wealth of information gathered from practical study and use of the
Epoxy to support its use as a conservation material for stone became somewhat obsolete.
Additionally the task was now to find a gap filling material for the repair of the cracks
that would be compatible with the resin in regards to surface coloration and adhesion.
Introducing further materials to the damaged resin shell is also thought to impact on the
overall long term stability and surface appearance of the replica and the investment of
time, cost and equipment into what is likely to become an ongoing maintenance and care
program is not in proportion to its long term benefits.
This has led to a complete re-think. The reviewed conservation recommendation of
preparing a new replica from the existing mold is informed by investigation into varying
traditional and modern casting materials and identification and assessment of aspects for
sustainability and permanence to assure the materials availability rather than aspects
such as reversibility.
There has been much debate on the suitability of Epoxy Resins in the various disciplines
of conservation in general and in stone conservation specifically (6).
Particularly the conservation concept of ‘reversibility’ of materials and treatments
employed in conservation practice meant the utilization of an ‘irreversible’ Epoxy is
likely dismissed without a critical comparison of its properties with ‘reversible’
materials such as acrylics, thermoplastic resins, and plasters.

In some ways the ethic of reversibility and the general use of the term ‘irreversible’ for
characterizing Epoxies can be misleading if the term ‘irreversible’ is equated to the
concept of ‘permanence’.
Stone as a material in sculpture intrinsically implies permanence due to its age of when
formed as igneous, sedimentary or metamorphic rock and its utilization through human
history as a structural and building material.
Particularly in 19 th and early 20 th century sculpting practices stone would be selected by
the sculptor as the final medium of the artistic expression for its associated character of
permanence.
While conservation assessment of aspects for sustainability and permanence in regards
to availability of natural resources in creating or manufacture of cultural heritage is a
mechanism for interpretation of material culture and their technologies of the past,
assessment of sustainability and permanence of availability of manufactured modern
materials may be a mechanism to identify suitable and potentially sought after materials
in and for the future.
Anticipating the shortage of materials that are currently readily available may aid
preparation to assure the future preservation of not only such materials, but also the
required knowledge and skills for their maintenance and conservation care.
The detrimental effect the loss of natural resources and associated knowledge and
working skills pose to the future preservation of cultural heritage can be seen in the
example of Glasgow’s stone heritage.
The wealth created by the ship building and associated industries at the river Clyde in
the late 19 th and early 20 th Century was expressed in Glasgow’s architecture, sculpture
and monuments of the readily available resource stone and related expertise in required
working skills.
Tenement blocks for example, build during this time, obtained the building stone from
approximately 400 active stone quarries all within the area of Glasgow. The diversity of
the many available stone quarries led to rows of stone houses with distinct coloration for
individual streets.
Today there are only 3 active quarries within Scotland left and a history of industrial
pollution and stone cleaning of resulting pollution crust’s, has left much of the build
stone heritage in poor condition (7).
The lack of availability of specific stones for potential replacement or restoration is only
one part of the problem to carry out adequate maintenance and restoration care.
Additionally, today after the decline of the many industries and workshops there is a
shortage of relevantly skilled and experienced practitioners in the field of stone heritage
and building conservation.
This leaves the wealth of Glasgow’s19th and 20 th Century cultural stone heritage such as
architecture, sculpture and monuments threatened to become permanently lost. A
scenario that would have not been considered by the sponsors and makers involved in
the creation of this heritage in times of plenty.

Figure 12. ‘Sculpture’, outdoor sculpture, Kelvingrove Art Gallery
Conclusion
Epoxy resins are important materials with a wide range of use in the field of stone
conservation. The possibility of designing and manufacturing a polymer with very
targeted and specific functions to some extent assures availability of specialist products
for the conservation practitioner.
Their chemical and physical strength make them popular materials in the creation of
structural supports for heavy yet fragile stone monuments or structures and allow to
create decorative and permanent mounting systems for stone structures and sculptures.
The focus of conservation ethics on ‘reversibility’ resulted in often critical views of the
‘irreversible’ Epoxy resin, with the justification of their employment based on a
somewhat misleading understanding of the definition of their chemical ‘irreversibility’
with ‘permanence’.
The overall experience led to viewing modern manufactured materials such as Phillyseal
R as temporary materials, their existence being determined by economic and
technological circumstances. It highlighted a need to consider preservation of physical
and intellectual access to the creative process, working skills and required resources of
modern manufactured materials since a substantial amount of this temporary media
makes up our present day environment and culture.
For the future care of our contemporary cultural heritage preservation of the technical
knowledge and skills of the manufacturing process of temporary materials will
determine the level of its conservation care and survival in the long run.

References
1. BJ Farrar; J. Maish; M. Shiro: A preliminary Review of some alternatives to
Phillyseal R Epoxy for Conservation and Mountmaking.
cool.conservation-us.org/waac/wn/wn31/wn31-1/wn31-103.pdf
2. http://cool.conservation-us.org/waac/historical/pastmeetings/2008-losangeles/images,
19/06/2012, BJ. Farrar; J. Maish: Phillyseal R- is there an alternative?
3. Source: Multi Mimsy Museum database. Entry record 1927.78, 15/06/2012
4. Source: Multi Mimsy Museum database. Entry record A.1942.18, 15/06/2012
5. Personal Email correspondence with ITW Polymer Technology, 02/06/2012
6. G. G. Amoroso; 1983, Stone decay and conservation, Materials Science Monographs,
11, Elsevier, Amsterdam, Lausanne, Oxford, New York
7. Personal notes from 1 day symposium ‘Stone cleaning in Kelvingrove Art Gallery’,
Kelvingrove Art Gallery, 6 th October, 2006.

REQUIREMENTS FOR REPLACEMENT STONES AT THE COLOGNE
CATHEDRAL – A SYSTEMATIC APPROACH TO GENERAL CRITERIA OF
COMPATIBILITY
B.Graue 1 , S.Siegesmund 1 , B.Middendorf 2 and P.Oyhantcabal 3
1 Department of Structural Geology and Geodynamics, Geoscience Center of the
University of Goettingen, Goldschmidtstr. 3, 37077 Goettingen, Germany; e-mail:
info@graue.org
2 Faculty of Architecture and Civil Engineering, Department of Building Materials, TU
Dortmund University, August-Schmidt-Str. 8, 44227 Dortmund, Germany
3 Department of Geology, Faculty of Sciences, Universidad de la República Igua, 4225
CP11400 Montevideo, Uruguay
Abstract
The Cologne cathedral is one of the most outstanding monuments in Northern
Europe. Its construction history began in 1248 and spreads over a 600 year building
period. The main building material, the Drachenfels trachyte was extracted from
quarries at the Siebengebirge. Due to the different construction phases, over 50 different
building stone materials were used to build the cathedral.
The cathedral suffers severe stone deterioration, which endangers the structure of
the building. The Drachenfels trachyte shows pronounced deterioration phenomena such
as contour scaling, flaking and structural disintegration to crumbling. Other main
building stones e.g. sandstones, carbonates, and volcanic rocks, show significant
degradation as well. For the preservation of the monument it is crucial to find an
appropriate replacement material, which is not only suitable to the original Drachenfels
trachyte but is also compatible to the other rocks.
In the present paper, the mineralogical, petrophysical and moisture properties of
the eight main rocks from the Cologne cathedral are determined and discussed in
correlation to each other. The strong divergence of the ascertained parameters (i. e
mineral composition, porosity, water absorption and saturation, drying characteristics,
moisture and thermal dilatation, strength properties, etc.) shows, that the constraints for
a replacement material makes it almost impossible to find an ideal stone. The sum of the
properties is ranked and their relevance for replacement criteria is determined in view of
the observed deterioration phenomena and processes. This evaluation leads to a
systematic approach for developing general criteria of compatibility in selecting
replacement materials for historical monuments, which comprise more than one natural
building stone material.
Keywords: stone decay, Cologne cathedral, fabric, mineralogy and petrophysical
properties, compatibility of building materials, requirements for replacement stones
1. Introduction
The Cologne cathedral has a very long construction history, in which different stone
materials were used. The Drachenfels trachyte from the quarries of the Siebengebirge, a
natural building stone for construction in the city of Cologne since the Roman period

(Wolff 2004), was the construction material for the cathedral during the medieval period.
At the beginning of the 16 th century the construction was halted and recommenced at the
beginning of the 19 th century. At that time the Drachenfels trachyte was no longer
available. Initial renovations were carried out with latite from the “Stenzelberg” and a
few other materials from the quarries of the Siebengebirge. In the middle of the 19 th
century the second construction phase used sandstone from “Schlaitdorf”, southern
Germany. Later on the “Obernkirchner” sandstone from Lower Saxony and from 1918
until 1940`s the “Krensheimer Muschelkalk” were implemented. In the 1950’s, the
decay resistant basalt lava from “Londorf” was used. The materials presently applied are
the trachyte from “Montemerlo” (Italy) for the replacement of deteriorated Drachenfels
trachyte as well as the Czech sandstone from “Bozanov”, which is in use to replace the
weathered Schlaitdorfer sandstone (Scheuren 2004; Schumacher 2004).
2. Decay phenomena
The different building stones of the Cologne cathedral show a large variation of
deterioration. The main deterioration phenomena observable in the Drachenfels trachyte
are surface deterioration and back-weathering coexisting with flaking and structural
disintegration to crumbling. Back-weathered areas often display stronger further decay
in terms of microcracks and crumbling to total collapse. Scaling is observable von
Plehwe-Leisen et al. (2007) and very often shows a granular disintegrated zone on the
reverse side. Formation of cracks and fissures may also propagate many centimeters in
depth into the stone. The Drachenfels trachyte is characterized by large phenocrysts of
sanidine – up to 7 cm in length. These may cause a different weathering behavior
between the matrix and the phenocrysts; e.g. loss of sanidine phenocrysts or loss of
matrix. The weathering behavior of the building stones is also controlled by the
orientation of the rock fabric. In the Drachenfels trachyte the deterioration is more
intense when the magmatic foliation, defined by the preferred orientation of sanidine
phenocrysts, is parallel to the visible surface of the building stone. A number of
breakouts, as a result of the mechanical impacts of bombing during WW II, can be
observed in the Drachenfels trachyte. The flaking can occur in a very pronounced
fashion, which eventually leads to structural disintegration and total fabric collapse.
There are strong indications that the decay of the Drachenfels trachyte is especially
critical in the direct neighborhood of carbonate replacement stones (Kraus 1985; von
Plehwe-Leisen et al. 2007). In many places the decay starts from the joints, which is
indicated by gypsum crusts, flaking and scaling.
At Cologne cathedral, the generally very deterioration resistant Obernkirchner
sandstone (Grimm 1990) shows some forming of grayish to black crusts as well as the
formation of gypsum crusts in sheltered areas. Further severe damage is visible along
joints, where the sandstone shows breakouts due to spalling, especially on the decorative
parts, e.g. pilaster strips.
Already Kraus (1985) and Grimm (1990) mentioned the characteristic degradation
of the Schlaitdorfer sandstone at the Cologne cathedral. Due to the carbonate binder
(app. 14%) gypsum formation is observed leading to massive scaling and flaking as well
as granular disintegration. Deterioration in form of rounding and notching is also typical
for the Schlaitdorfer sandstone.
At present little is known about the deterioration behavior of the Montemerlo
trachyte at the Cologne cathedral, since this stone has only been implemented in recent

years. However, Lazzarini et al. (2008) report of exfoliation and flaking, powdering and
alveolic weathering for the Montemerlo trachyte in Venice, Italy.
The Stenzelberg latite and Londorfer basalt lava have high weatherability. Due to
the high porosity, the Londorfer basalt lava is susceptible to microbiological action.
Stenzelberg latite shows a typical formation of scales with a thickness of 2 – 3 mm.
The Bozanov sandstone shows spalling along edges when mounted, which is
problematic for masonry works. Přikryl et al. (2010) report on granular disintegration,
scaling, flaking, crust formation as well as blistering, fracturing, salt efflorescences and
alveoli formation for the medium grained Bozanov sandstone.
The Krensheimer Muschelkalk is a deterioration resistant rock. This carbonate
building stone shows massive gypsum crust formations in rain protected areas. On
surfaces exposed to rain solution phenomena (e.g. microkarst) can be observed leading
to surface roughness and loss of shape of figural areas.
3. Building stones of Cologne cathedral
3.1 Petrography
The Drachenfels trachyte is a light gray, partially yellowish porphyritic trachyte
with phenocrysts of sanidine up to 7 cm in size enclosed in a fine-grained matrix
composed mainly of feldspar and quartz. The phenocrysts can show a preferred
orientation in the matrix tracing the magmatic foliation. The rock comprises sanidine
(50%), plagioclase (24%), quartz (13%), augite (5%), biotite (5%), ore (2%) and apatite,
zircon and sphene (1%). In places calcite occurs accompanied by iron oxides, indicating
the influence of hydrothermal fluids. Interstitial volcanic glass, partially recrystallised
and altered to montmorillonite is observed between the feldspar laths of the matrix.
Pyrite and aggregates of pyrite (partly altered to hematite and limonite?) are found in
cavities (Grimm 1990).
The trachyte of Montemerlo shows a quasi-isotropic and homogenous fabric.
Feldspar crystals (0.5-10 mm), biotite (< 2mm) and amphibole float in a gray to
yellowish microcrystalline matrix composed of anorthoclase, sanidine, plagioclase and
seldom quartz. The rock’s composition shows K-feldspar (53%), plagioclase (15%),
quartz (8%), amphibole (8%), biotite (5%), pyrite (7%) and calcite (4%) (Koch 2006).
The Stenzelberg latite is a medium gray, porphyritic, and in part porous latite. The
micro- to cryptocrystalline matrix (77%) is mainly composed of plagioclase and
sanidine. Plagioclase (14%), hornblende with individual grain sizes up to 10 cm (5%),
augite (1%), and biotite (1%) occurs as phenocrysts. Accessory minerals are apatite,
sphene, zircon and ore minerals (Grimm 1990).
The Obernkirchner sandstone is a medium-grained, moderately to well-sorted
quartz arenite of white to orange color with maximum grain size of 300 microns. The
rock is composed of monocrystalline quartz (98 %), muscovite, zircon, tourmaline,
rutile and opaque minerals. The fabric is grain-supported with sutured grain contacts.
The matrix (ca. 5%) consists of authigenic kaolinite, rare silica and iron oxide patches
(Dienemann and Burre 1929; Grimm 1990).
The Schlaitdorfer sandstone is a coarse-grained and well-sorted arenite, whitish to
yellowish with often parallel, angular and cross-bedding. The detrital fraction (65%) is
represented by quartz (72%), rock fragments (12%), feldspar (2%) and cement (14%).
The cement consists of coarse-grained dolomite, in parts silica and rarely illite and

kaolinite. The main accessories (< 1%) are apatite, zircon, tourmaline and opaque
minerals (Grimm1990).
The weakly cemented Bozanov sandstone is a coarse- to medium-grained arkosic
arenite of light gray to yellow color. The mineralogical composition is given by quartz
(79%), rock fragments (10%), feldspar (5%), clay minerals (smectite and kaolinite,
around 5%) and accessories like biotite, zircon and opaque minerals (Koch 2001; Přikryl
et al. 2010).
The Krensheimer Muschelkalk is a light-colored, brownish-grayish, fine porous
limestone very rich in shell fragments. It is classified as a densely packed
bio(micro)sparite after Folk (1962). Mussel- and brachiopod shells of 5-7 mm sized are
densely packed in a fine calcite matrix. The components are oriented parallel to the
bedding, showing a moderate sorting. The composition consists of 75% biogenic
components mainly with micritic rendering, 5% cement and 20% pores (Grimm 1990;
Siegesmund et al. 2010).
The Londorfer basalt lava is a brownish to bluish gray basalt. The fabric is fine- to
medium-grained with hyaloophitic fractions. The rock is composed of plagioclase (50%),
clinopyroxene (25%), olivine (15%) and ore minerals (ilmenite and magnetite around
10%). Cryptocrystalline accessories (< 1%) and glass (up to 50%) also occur. The rock
is highly porous and the pores are often coated by light-gray bluish secondary zeolites
(Grimm 1990; Steindlberger 2003).
3.2 Petrophysical properties
The investigated stones show medium porosities from 11.8% (Montemerlo
trachyte) to 19.9% (Schlaitdorfer sandstone), except the Stenzelberg latite, which
belongs to low porosity stones with a porosity of 8.5% (Tab. 1). The Drachenfels
trachyte has a porosity of 11.9%. Their densities range from 2.10 g/cm 3 to 2.52 g/cm 3
(Tab. 1).
Table 1 Bulk and matrix density, porosity, mean and mode pore radius, capillary water uptake,
saturation coefficient, vapour diffusion resistance and sorption.
Rock type
Bulk
density
Matrix
density
Effective
porosity
(g cm -3 ) (g cm -3 ) (vol.-%)
Mean
pore
radius
(µm)
Mode
pore
radius
(µm)
Capillary
water
uptake
(w-value)
(kg/m 2 √h)
Saturation
degree
(Svalue)
Vapor diffusion
resistance
factor (µ)
Sorption
(max.
moisture
content)
(wt-%)
Z X
Drachenfels trachyte 2.33 2.64 11.92 0.414 1.334 0.55 0.74 37.38 17.91 1.877
Stenzelberg latite 2.46 2.69 8.53 0.017 0.013 0.3 0.76 56.39 50.2 2.778
Schlaitdorfer sandstone 2.1 2.63 19.91 2.891 33.497 6.68 0.64 20.56 16.93 0.38
Obernkirchner sandstone 2.16 2.65 18.58 0.821 3.35 1.26 0.64 15.89 14.95 0.716
Krensheimer Muschelkalk 2.25 2.68 16.03 1.501 8.414 1.3 0.59 69.45 51.25 0.291
Londorfer basalt lava 2.52 2.92 13.13 0.621 21.135 0.39 0.59 37.78 38.6 1.615
Bozanov sandstone 2.17 2.63 17.79 5.953 21.135 6.9 0.65 16.51 19.69 0.746
Montemerlo trachyte 2.35 2.66 11.76 0.108 0.211 0.99 0.71 40.49 31.82 1.114
The pore size distributions (PSD) of the investigated stones are unimodal – except of the
Krensheimer Muschelkalk and Londorfer basalt bava (Rüdrich and Siegesmund 2007).
Schlaitdorfer and Bozanov sandstones have a broader distribution of pores ranging from
0.0064 – 82 µm with a clear peak of pores at > 10 µm. Obernkirchner sandstone has a
narrower distribution in the range of 0.0064 – 64 µm and Drachenfels trachyte in the

ange of 0.0082 – 28 µm. The PSD of the Montemerlo trachyte is limited from 0.0064 –
1 µm and that of the Stenzelberg latite from 0.0064 – 0.28 µm with a relatively
narrowed pore radii maximum. Krensheimer Muschelkalk and Londorfer basalt lava
show a bimodal PSD.
3.3 Moisture properties
The Drachenfels trachyte, Stenzelberg latite and Londorfer basalt lava show low
capillary water absorption (w < 0.5 kg/m 2 √h). The Montemerlo trachyte, Obernkirchner
sandstone and Krensheimer Muschelkalk have a medium value (1-1.5 kg/m 2 √h);
Schlaitdorfer and Bozanov sandstones show a high capillary water absorption (6.5-7
kg/m 2 √h) (Snethlage 2005; Siegesmund and Dürrast 2011). The values are listed in
Table 1.
Based on the measured data of pore radii and capillary water absorption, the stones
can be divided into three groups (Snethlage 2005): 1) Stenzelberg latite, Londorfer
basalt lava and Drachenfels trachyte have small mean pore radii and low capillary water
absorptions (w-values); 2) Montemerlo trachyte, Krensheimer Muschelkalk and
Obernkirchner sandstone have medium pore radii in the lower to medium range of
capillary active pore sizes and medium capillary water absorptions; 3) Schlaitdorfer and
Bozanov sandstones with large mean pore radii have high water absorbing coefficients.
The values for the water saturation (s-value) of the investigated stones range from
0.59 – 0.76 (Tab. 1). The Krensheimer Muschelkalk and Londorfer basalt lava show the
lowest s-values. The Schlaitdorfer, Obernkirchner and Bozanov sandstones are in a
medium range. The Drachenfels and Montemerlo trachytes as well as Stenzelberg latite
are stones with higher s-values.
In terms of hygroscopic water adsorption Stenzelberg latite shows the highest mass
increase of 2.78 wt.-% at 95% RH and Krensheimer Muschelkalk the lowest value (0.29
wt.-%) (Tab. 1). The Drachenfels trachyte and Londorfer basalt lava also show a
relatively high water adsorption, whereas the Montemerlo trachyte, Obernkirchner and
Bozanov sandstones have a medium water adsorption. Schlaitdorfer sandstone only
shows a small mass increase (Tab. 1). Stenzelberg latite, Londorfer basalt lava and
Drachenfels trachyte show a hysteresis in their sorption-desorption-behavior, indicating
that the stone material dries slower with descending relative humidity and still contains
a residue of moisture as a possible indication of capillary condensation.
In respect to the water vapor diffusion resistance the Krensheimer Muschelkalk
and Stenzelberg latite have a high resistance. Drachenfels and Montemerlo trachyte as
well as the basalt lava show a medium resistance. The sandstones have the highest
permeability of the investigated stones (Tab. 1). The Drachenfels trachyte shows a
remarkable directional dependence of water vapor diffusion resistance which could
mainly be controlled by the flow fabric (Tab. 1). A higher resistance correlates with a
higher amount of micropores: capillary condensation takes place in micropores, which
holds back water due to solvent water diffusion. This leads to capillary suction
(retention), which is much slower than water vapor diffusion (Snethlage 1984). Only the
Krensheimer Muschelkalk does not fit this correlation.
Hygric dilatation (in the range between 0% and 95% RH) and hydric dilatation
(water saturated) were measured. In general, moisture dilatation is low. Montemerlo
trachyte has the highest hydric dilatation (0.316 mm/m). The values for the hygric
dilatation of the Montemerlo trachyte are somewhat higher than that of the Drachenfels

trachyte. High moisture swelling was measured as well in the Drachenfels trachyte,
Londorfer basalt lava and Stenzelberg latite. The later shows a significant hygric
dilatation (0.231 mm/m) due to its high content of micropores. The Obernkirchner
sandstone has medium moisture related swelling. Schlaitdorfer and Bozanov sandstone
have low moisture dilatation. The length change of the Moisture expansion in the
Krensheimer Muschelkalk is negligible (Tab. 2). With increasing relative humidity a
sharper increase of hygric expansion can be observed at around 80 - 85% relative
humidity.
Table 2: Thermal expansion coefficient and hygric dilatation.
Rock type
x z anisotropy x z
(10 -6 K -1 ) (10 -6 K -1 Thermal dilatation coefficient Hydric dilatation
) (%) (mm/m) (mm/m)
Drachenfels trachyte 5.32 6.05 12 0.253 0.236
Stenzelberg latite 9.41 7.36 21.7 0.196 0.23
Schlaitdorfer sandstone 9.65 11.96 19.3 0.025 0.025
Obernkirchner sandstone 11.6 12.17 4.6 0.089 0.06
Krensheimer Muschelkalk 4.75 6.82 30.3 0 0.005
Londorfer basalt lava 5.32 5.78 8 0.226 0.186
Bozanov sandstone 8.78 8.65 1.5 0.027 0.013
Montemerlo trachyte 6.25 4.65 25.5 0.291 0.316
3.4 Thermal dilatation
In terms of thermal dilatation only for the Obernkirchner and Schlaitdorfer
sandstone high thermal expansion coefficients are detected. The other stones show low
to medium thermal expansion and no residual strain (Tab. 2).
3.5 Strength properties
The compressive strength varies between 45.1 N/mm 2 and 126.4 N/mm 2 (Tab. 3).
The Stenzelberg latite has the highest compressive strength (126.4 N/mm 2 ), while the
strength values for the Krensheimer Muschelkalk and the Schlaitdorfer and Bozanov
sandstones are less than 50 N/mm 2 . Only the Stenzelberg latite belongs to the high
strength rocks, while rocks with compressive strength values between 55 N/mm 2 and 70
N/mm 2 belong to the low strength rocks, as there are the Drachenfels trachyte and
Londorfer basalt lava (Mosch 2008). Obernkirchner sandstone and Montemerlo trachyte
display higher uniaxial compressive strength values. A directional dependency is
detected for the Obernkirchner sandstone and the Montemerlo trachyte (Tab. 3).
The flexural strength values of the investigated rocks cover the range between 4.0
N/mm 2 (Bozanov sandstone) and 15.7 N/mm 2 (Stenzelberg latite) (Tab. 3). Londorfer
basalt lava has a high flexural strength, whereas the Schlaitdorfer sandstone, the
Drachenfels and Montemerlo trachytes show medium flexural strength values. The
Obernkirchner sandstone and the Krensheimer Muschelkalk display somewhat higher
flexural strength values (Tab. 3).
The tensile strength measured using the Brazilian Test varies between 3.1 N/mm 2
(Drachenfels trachyte) and 9.7 N/mm 2 (Stenzelberg latite), depending on the sample and
direction of load with respect to the rock fabric. Londorfer basalt lava shows a medium

tensile strength. Obernkirchner, Schlaitdorfer and Bozanov sandstones, Krensheimer
Muschelkalk and Montemerlo trachyte have a lower tensile strength (Tab. 3).
Table 3 Uniaxial compressive strength, tensile strength and flexural strength of the
investigated rocks in non-weathered condition.
Rock type
Compr.strength
(N/mm
Z X Z X Z X
Drachenfels Trachyte 65.54 66.59 3.087 3.674 5.999 6.1
Stenzelberg Latite 126.41 120.02 9.735 8.621 15.707 9.881
Schlaitdorfer Sandstone 47.59 51.44 3.256 3.343 6.492 5.733
Obernkirchner Sandstone 86.72 76.29 4.594 4.669 7.992 6.825
Krensheimer Muschelkalk 48.35 52.74 4.498 4.544 8.467 6.763
Londorfer Basalt Lava 63.1 72.18 5.099 5.921 12.567 12.749
Bozanov Sandstone 45.1 52.08 3.462 3.343 3.975 4.41
Montemerlo Trachyte 75.53 84.75 3.439 3.68 6.725 8.195
2 )
Tensile strength
(N/mm 2 )
Flexural strength
(N/mm 2 )
4. Correlation of fabric, mineralogical, petrophysical properties, and decay
processes
The Drachenfels trachyte has a porphyritic fabric with a strong magmatic foliation.
Large phenocrysts of sanidine are embedded with preferred orientation in a matrix with
strongly aligned microcrystalline feldspar laths. The fabric can be divided in three
structural components: the large phenocrysts, secondly the microcrystalline matrix,
composed mainly of feldspar and third a mesostasis consisting mainly of recrystallized
interstitial volcanic glass. As mentioned by Grimm 1990 this is partially altered to
montmorillonite.
In contrast to the relatively high porosity of 12% and the high ratio of capillary
active pores (84%), the stone shows a low capillary water uptake (0.55 kg/m 2 √h). This
may indicate a lack of connectivity of the pore space. The sorptive water uptake and the
saturation degree measured were high, analogous to the values of Snethlage 2005.
Larger mineral grains show a lot of cracks and breakages, which are to be considered as
part of the pore space. The Drachenfels trachyte shows medium water vapor diffusion
resistance and drying is retarded. Kraus 1985 reports that within 15 days the tested stone
samples still contain rest moisture. This water has to be released via water vapor
diffusion. The strength properties of the stone are medium to low: low compressive
strength, medium flexural strength, very low tensile strength. Moisture dilatation is high
compared to the other eight investigated stones, thermal dilatation is little.
The Drachenfels trachyte is very inhomogeneous not only in respect of grain sizes
but also in terms of the mineralogical composition of the structural components
(phenocrysts, matrix and mesostasis) and their specific properties. Phenocrysts and
matrix consist of chemically fairly stable components in comparison to the mesostasis.
The recrystallized glass in the mesostasis is liable to chemical decay impact and the
fractions already altered to montmorillonite enhance deterioration processes due to the
swelling properties of these clay minerals. Even though capillary water uptake is low,
the high porosity, saturation degree and water adsorption as well as the retarded drying
suggest a susceptibility to moisture related processes. These parameters, which are

significant for the Drachenfels trachyte, indicate that wetting-drying cycles are not very
pronounced, but the stone stays humid over long periods of time. This involves a
significant capacity for adsorption and transportation of pollutants and guarantees
sufficient water supply for the degradation processes. In these terms, direct mechanical
material reaction, e.g. moisture dilatation, can be considered minor, but pollution impact
and salt weathering become more important.
crit.
fabric
pore space
moisture properties therm. mechanical strength
min.
A B C D E F G H I J K L M N O P
Drachenfels trachyte
A components 3 2 1 1 1 1 1 1 1 2 1 2 2 3 2 24 A
B matrix 0 3 2 3 3 3 3 3 3 3 2 2 2 2 2 39 B
C meso stasis 0 0 3 3 3 2 3 2 2 3 1 2 1 1 1 32 C
critical mineral components
D 0 1 3 1 2 1 2 1 2 3 2 2 1 1 1 25 D
(calcite, clay minerals)
E porosity 1 3 3 1 3 3 2 2 2 2 2 1 2 2 2 31 E
F PSD 1 3 2 2 3 3 3 2 3 3 3 2 2 2 2 37 F
G capillary water uptake 1 3 2 1 3 3 1 3 1 3 3 0 2 2 2 30 G
H sorptive water uptake 1 3 2 2 3 3 1 3 2 3 3 0 1 1 1 29 H
extrinsic impact correlation of the parameters (decay index)
I saturation degree 1 3 2 1 3 2 3 3 0 3 2 0 2 2 2 28 I
J vapor diffusion 2 3 2 2 3 3 1 2 0 1 3 0 1 1 1 23 J
K moisture expansion 1 2 3 3 2 3 3 3 2 0 1 0 1 2 1 31 K
L drying 1 2 1 2 2 3 3 3 3 3 2 0 1 1 1 26 L
In a matrix analogous to Visser and Mirwald 1998, the afore described fabric and
pore space parameters as well as the petrophysical properties of the Drachenfels trachyte
are evaluated from 0 to 3 (rating numbers) in terms of their significance to each other for
the material behavior (Fig. 1). This material-intern correlation points out the
significance of distinct parameters in respect to the characteristic properties of this stone,
M thermal dilatation 2 2 2 1 2 2 0 0 0 0 0 0 1 1 1 14 M
N compresive strength 3 3 2 1 3 3 1 1 1 0 1 1 0 19 N
O flexural strength 3 3 2 1 3 3 1 1 1 0 1 1 0 2 21 O
P tensile strength 3 3 2 1 3 3 1 1 1 0 1 1 0 2 2 19 P
material-intern correlation of the parameters (material index)
20 34 28 22 38 39 27 27 26 21 27 28 11 24 24 24
A B C D E F G H I J K L M N O P
Figure 1. Correlation of fabric, mineral and petrophysical parameters of the Drachenfels trachyte in terms of
their significance for material behavior (lower left triangle) and decay characteristics (upper right triangle)

e.g. the water up take strongly correlates with porosity and PSD (rating number 3),
whereas thermal dilatation is not interrelated to moisture properties (rating number 0).
The sum of these rating numbers of one parameter is the degree of the material-intern
correlation of the parameters, the “material index” of each parameter. This is shown by
the lower left side of the matrix. In the upper right part of the matrix the parameters are
correlated to each other in terms of their significance for the deterioration processes,
again from 0 to 3, giving the “decay index” for each parameter. The correlation of the
material parameters shows, that in the case of the Drachenfels trachyte mainly fabric
parameters, characterized by the specific features of the matrix and mesostasis, and pore
space properties such as PSD and porosity as well as moisture properties, e.g. drying
and capillary water uptake, determine the behavior of the stone (Fig 2). In respect to the
decay of the stone mainly fabric and pore space parameters as well as moisture
properties control the deterioration processes. Strength and thermal properties are of
minor impact (Fig 2).
rank material
index
parameter parameter
decay
index
1 39 PSD matrix 39 1
2 38 porosity PSD 37 2
3 34 matrix meso stasis 32 3
4 28 drying porosity 31 4
5 28 meso stasis moisture expansion 31 5
6 27 capillary water uptake capillary water uptake 30 6
7 27 sorptive water uptake sorptive water uptake 29 7
8 27 moisture expansion saturation degree 28 8
9 26 saturation degree drying 26 9
10 24 flexural strength calcite, clay minerals 25 10
11 24 tensile strength components 24 11
12 24 compresive strength vapor diffusion 23 12
13 22 calcite, clay minerals flexural strength 21 13
14 21 vapor diffusion tensile strength 19 14
15 20 components compresive strength 19 15
16 11 thermal dilatation thermal dilatation 14 16
material-intern correlation decay significance
Figure 2. Ranking of fabric, mineral and petrophysical parameters of the
Drachenfels trachyte in terms of their significance for material behavior and
deterioration, indicating eight “key parameters” for replacement criteria.
5. Relevance for replacement criteria of potential building stones
The correlation of the parameters and their ranking in terms of material behavior
and deterioration impact (Fig. 2), indicate the relevance for replacement criteria. The
material behavior of the Drachenfels trachyte is determined by fabric and pore space
parameters as well as moisture properties especially pore size distribution, porosity,
capillary and sorptive water uptake. In terms of deterioration besides the mentioned
parameters also matrix, mesostasis and moisture dilatation become more pronounced.
The eight parameters characterize the Drachenfels trachyte and are significant for the
behavior of the stone in terms of extrinsic impact and decay. These are the “key
parameters” a replacement stone for the Drachenfels trachyte should match.
rank

In terms of restoration and conservation any import of potential harmful substances
by a new material, e.g. critical mineral components in a replacement stone has to be
avoided. Further, the optical properties of the replacement rock should be similar to the
original material considering aging and patination. In respect to petrophysical criteria,
the replacement stone for the Drachenfels trachyte should have a comparable PSD and
porosity as well. Moisture dilatation should not be pronounced and water uptake –
capillary as well as sorptive – should be low. Generally the s-value should be less than
0.75. Although strength and thermal properties play a minor role in deterioration
processes in the Drachenfels trachyte, the replacement stone should be in a range of 80-
120% of the strength values (Snethlage 2005) and thermal dilatation should be less than
the original stone.
The actual replacement stone for the Drachenfels trachyte at Cologne cathedral is
the Montemerlo trachyte from Italy. If the two stones are compared in respect to the
mentioned constraints, it is to ascertain that the mineralogical composition and optical
properties match perfectly. The porosity (11.8 %) is similar; the pore size distribution
shows a higher ratio of micropores (37%) than capillary active pores (63%). In the
Drachenfels trachyte the ratio of micro to capillary pores is 16:84 (Graue et al. 2011).
Moisture dilatation is slightly pronounced (Tab. 2); capillary water uptake is higher but
water absorption and s-value are lower (Tab. 1). In terms of strength properties the
Montemerlo trachyte is a slightly stronger stone, averagely 112%, which is in the range
of constraints. Thermal dilatation of the Montemerlo trachyte is comparable to the
Drachenfels stone, as well is drying.
In general, the parameters of the Drachenfels and Montemerlo trachyte are in a
close comparability. The higher ratio of micropores, the higher capillary water uptake
and the slightly pronounced moisture dilatation can be critical. In resemblance to the
observation by Lazzarini et al. (2008), the Montemerlo trachyte shows little resistance to
salt deterioration experiments. It was the first of the eight investigated stones losing 50%
of its weight after 19 cycles; Drachenfels trachyte is the second after 30 cycles.
6. Conclusions
The different building stones employed at Cologne cathedral show a diverse
petrography and mineralogical composition as well as a broad variety of petrophysical
properties. To understand possible interactive deterioration processes it is important to
determine the basic petrophysical data. For the Drachenfels trachyte the parameters are
correlated to each other in terms of their significance for the behavior of the stone as
well as their relevance in respect to deterioration processes. The determined “key
parameters” mark the critical factors. A replacement stones has to be suitable to these
key factors within a certain range.
This evaluation leads to a systematic approach for developing general criteria of
compatibility in selecting replacement materials for historical monuments, which
comprise more than one natural building stone material.
7. Acknowledgements
This work is supported by Deutsche Bundesstiftung Umwelt (DBU-AZ-28253-45).
Special thanks go the colleagues of the Cologne cathedral maintenance department as
well as to T. Schumacher and master builder B. Schock-Werner for supporting our work.

8. References
Dienemann W, Burre O 1929. Die nutzbaren Gesteine Deutschlands und ihre
Lagerstätten mit Ausnahme der Kohlen, Erze und Salze. Stuttgart: Enke.
Folk RL 1962. ‘Spectral subdivision of limestone types’. In Classification of carbonate
rocks, Ham WE (ed) 62–84. Tulsa: AAPG.
Graue B, Siegesmund S, Middendorf B 2011. ‘Quality assessment of replacement stones
for the Cologne Cathedral: mineralogical and petrophysical requirements’. Environ
Earth Sci 63:1799–1822.
Grimm WD 1990. Bildatlas wichtiger Denkmalgesteine der Bundesrepublik Deutschland.
München: Bayer. LAD.
Koch R 2001. ‘Zur Geologie und Fazies eines Sandsteins aus Nordost-Tschechien’.
Unpubl. report. University of Erlangen.
Koch R 2006. ‘Trachyte aus dem Colli Euganei, Norditalien’. In Modellhafte
Entwicklung von Konservierungskonzepten für den stark umweltgeschädigten Trachyt
an den Domen zu Köln und Xanten, Schock-Werner B (ed.) 9-42. Köln: Domverlag.
Kraus K 1985. ‘Experimente zur immissionsbedingten Verwitterung der Naturbausteine
des Kölner Doms ’. Ph.D. dissertation. University of Cologne.
Lazzarini L, Antonelli F, Cancelliere S, Conventi A 2008. ‘The deterioration of
Euganean trachyte in Venice’. In Proc. 11th Int. Congr. Deterioration and Conservation
of Stone. Lukaszewicz J, Niemcewicz P (eds.) 153–162. Torun: Copernicus Univ. Press.
Mosch S 2008. ‘Optimierung der Exploration, Gewinnung und Materialcharakterisierung
von Naturwerksteinen’. Ph.D. dissertation. University of Göttingen.
Přikryl R, Weishauptová Z, Novotná M, Přikrylová J, Št’astná A 2010. ‘Physical and
mechanical properties of the repaired sandstone ashlars in the facing masonry of the
Charles Bridge in Prague’. Environ Earth Sci. 63:1623–1639.
Rüdrich J, Siegesmund S 2007. ‘Salt and ice crystallisation in porous sandstones’.
Environ Geol. 52:225–249.
Scheuren E 2004. ‘Kölner Dom und Drachenfels’. In Steine für den Kölner Dom,
Schock-Werner B and Lauer S (ed.) 22-45. Köln: DomVerlag.
Schumacher T 2004. ‘Steine für den Dom’. In Steine für den Kölner Dom, Schock-
Werner B and Lauer S (ed.) 46-77. Köln: DomVerlag.
Siegesmund S, Grimm WD, Dürrast H, Rüdrich J 2010. ‘Limestones in architecture’. In
Limestone in the built environment: present day challenges to preserve the past, Smith
B, Gomez-Heras M, Viles H, Cassar J (eds) 37–59. vol 331. London: Geol. Society.
Siegesmund S, Dürrast H 2011. ‘Physical and mechanical properties of rocks’. In Stone
in architecture, Siegesmund S, Snethlage R (eds) 97–225. Berlin: Springer.
Snethlage R 1984. Steinkonservierung 1979–1983. München: Bayer. LAD.
Snethlage R 2005. Leitfaden Steinkonservierung. Stuttgart: Fraunhofer IRB.
Steindlberger E 2003. Vulkanische Gesteine aus Hessen und ihre Eigenschaften als
Naturwerksteine. (vol. 110). Wiesbaden: Geol Abhdl. Hessen.
Visser H, Mirwald P 1998. ‘Baumberger Kalksandstein – Materialeigenschaften und
Schadenspotential’. In Die Steinskulpturen am Zentralbau des Jagdschlosses
Clemenswerth/Emsland, LAD Nieders. (vol. 15) 26-45. Hannover: Selbstverlag.
von Plehwe-Leisen E, Leisen H, Wendler E 2007. ‘Der Drachenfels-Trachyt – ein
wichtiges Denkmalgestein des Mittelalters’. ZDGG 158/3: 985-998.
Wolff A 2004. ‘Steine für den Dom’. In Steine für den Kölner Dom, Schock-Werner B
and Lauer S (ed.) 8-21. Köln: DomVerlag.

The Ince Blundell composite marble statue of a man with an ivy wreath –
‘Marcus Aurelius’: revisited/restored
Nicolas Verhulst 1 and Lottie Barnden 2
1Conservation Trainee, specialising in stone, University of Amsterdam, Ateliergebouw,
Hobbemastraat 22, 1071 ZC Amsterdam, The Netherlands (contact author:
verhulst.nicolas@gmail.com).
2 Head of Sculpture Conservation, National Museums Liverpool, Conservation Centre,
1 Peter Street, Liverpool L1 6HZ, United Kingdom (supervisor of the project).
Abstract
In the eighteenth century Henry Blundell bought an antique over-life-sized statue of ‘Marcus
Aurelius’ to add to his collection at Ince Blundell Hall, north of Liverpool. His private collection of
antique sculpture became the second largest in the United Kingdom, after the Townley collection
(British Museum). As a ‘composite’ sculpture ‘Marcus Aurelius’ represented the fashion of eighteenth
century restoration ethics. During the condition report the sculpture was divided into three different
zones: the plinth with the feet were separated from the main body and another zone existed of 58
pieces. The treatment proposal focused on a series of complex issues. A range of pins and clamps
needed to be removed from lead and eighteenth century resin. After cleaning the surface dirt, the
disturbing staining of the white Carrara marble by old resin needed to be tempered or removed.
Considering different cleaning methods, a superficial impregnating gel worked by dissolving 4% agar
in deionised water. Grey dirt layers and sulphation could be removed with a Nd:YAG laser. To
reinstate the structural integrity of the statue a ‘piston fit pin sleeve’ armature was devised, besides the
use of common pins. This will create an easier disassembly in the future, as the stainless steel pin
slides into stainless steel sleeves that are fixed inside the holes with a bulked epoxy. For choosing the
right colour fill seven adhesives with seven fillers were tested. This ‘colourfill chart’ worked as a
reference and each recipe could be adjusted by adding more or less fillers when searching for the right
properties. Finally a protective coating of microcrystalline wax was applied to protect the surface from
surface dirt ingression in the future.
KEYWORDS: This paper is about a composite statue and represents ideas for the removal of pins,
cramps and lead; removal of surface dirt and stains with Laser and agar; the choice of appropriate
adhesives according to its function; the development of a ‘piston fit pin sleeve’ armature system,
creating a colour fill to imitate white Carrara marble and the application of a protective coating.
Introduction
The Former head of the sculpture studio at the Conservation Centre in Liverpool (UK):
Samantha Sportun examined the white Carrara marble composite statue with an ivy wreath – ‘Marcus
Aurelius’ (210 x 85 x 50 cm) in 1995-1996. She started with the disassembly and removal of corroded
armature. Since fifteen years this project was left unfinished in the studio consisting of two main zones
and another one about 58 fragments. Over the years surface dirt accumulated on top of the disturbing
resin staining and sulphation. Under supervision of the current head of the sculpture studio: Lottie
Barnden, it became the multifaceted restoration project during my conservation trainee internship in
2011-2012, as part of the University of Amsterdam. It was necessary to get familiar with the object’s
background and current condition while writing a treatment proposal. Different options were
considered and will be explained in the following article. An important idea that was kept in mind
during the whole interdisciplinary process was that of a MINIMAL INTERVENTION AND EASE OF
DISSASEMBLY.

Historic background
The Ince Blundell collection
The antique over-life-sized statue was given the label: ‘Marcus Aurelius’ (Ince 569) by the
wealthy aristocrat Henry Blundell (1724-1810), who gathered an impressive collection of antiquities
in his garden temples at the Ince Blundell hall in Lancashire, situated north of Liverpool (UK). In one
of the greenhouses of his estate he displayed the statue in a niche (ill. 1). He probably bought the
statue at the end of his life in England after the sculpture supply in Italy dried up, around 1800.
Blundell’s catalogue mentions that the sculpture was found in the yard of a sculptor. Thomas Banks
(1772-1779) is suggested, however London based sculptor Joseph Nollekens (1760-1770) also sold
‘antiques’. Blundell must have been familiar with the contemporary restoration approaches as he was
aware that he had accumulated an eclectic collection of uneven quality. Most of his collection,
including ‘Marcus Aurelius’ became partially alienated from its eighteenth century setting, because
the contemporary owner could not safeguard his collection. He entrusted the majority of it to the city
of Liverpool, where the National Museums Liverpool became responsible for its conservation and
partially displays it to the public for free.
The label of ‘Marcus Aurelius’ revised
Two key questions appeared to be crucial to understand the iconography of this statue: What
type of statue are we dealing with and how is it possible to typecast a composite sculpture of ancient
and modern pieces, with a distinction in style between body and head? By scanning through literature
one could describe the ‘Ince 569’ as an over-life-size freestanding statue, with a bare torso and ‘hip
mantle’ held together by his left advancing arm, which recalls the Claudian timeframe. However his
stance is already inspired by the early canon from Polykleitos’ Doryphoros. His physiognomy bears
more similarities with the second type of ‘Marcus Aurelius’ than the first type. But it was still
uncertain whether that this portrait represents ‘Marcus Aurelius’. The scholar Fejfer considered that
the ivy wreath around his head may point into the direction of a Claudian (Antonine) or Hadrianic
type, maybe as part of the Liber Pater cult.
Ill. 1 Ill. 2 Ill. 3
A restoration history backed up by its condition research
From Antiquity, over the Renaissance to the end of the nineteenth century the integrity of
antique fragments was often compromised drastically as the fashion of the period was not for broken
fragments. As a ‘composite’ sculpture ‘Marcus Aurelius’ represents the ‘enlightened fashion’ of
eighteenth century restoration ethics. One of the best-known workshops during the time was the studio
of Bartolomeo Cavaceppi (1717-1799), who wrote a three-volume work on the principles of
restoration. In one of these volumes he gives an illustration of ‘Augusto’, resembling to the Ince

‘Marcus Aurelius’ (ill. 2). Some authors attributed the ‘Ince 569’ to Cavaceppi, while others think that
the piece was assembled in England rather than in Rome. Similarities in some of the restored parts of
drapery to the joint surfaces that were worked relatively smoothly and then chiselled with quite fine
marks to anchor the adhesive may point into the direction of Cavaceppi. The ‘purist’ approach in the
1960’s and 1970’s was to remove all previous restorations, which did not happen in this case.
The early additions to the ‘Ince 569’ represented a valuable insight into the methods and
techniques of the restorers of the period and must be valued and examined as such wherever possible.
Sportun was the first one to highlight the different interventions in a drawing, while discussing them in
detail (ill.3). The original parts are the face, except for the nose, the drapery and probably the central
part of the torso (separated from the drapery), as is a part of the tree trunk and the right leg and ankle.
The plinth, with both feet and the left leg, is supposed to be a classical repair. Eighteenth century
additions are the right arm, the left arm with overhanging drapery, the left shoulder, part of the breast,
the neck with the back of the head, including the right ear and the rim of the left one. Some other
undated repairs are the small pieces inserted in the back of the drapery, the back of the left ankle and
the front of the left knee, as a larger part of the drapery, in between the original and later addition. The
back of the tree trunk and the back corner of the plinth underneath can also be attributed to this phase,
according to Sportun.
Condition
Ill. 4 Ill. 5 Ill. 6
In general
Research about its assembly, ‘patination’ and used resins was executed in the past. For the
current condition research the statue was divided into three different zones: the plinth with the feet
were separate from the main body and another zone existed of 58 pieces, plus some extra pieces that
could not be located as part of the statue (ill. 4-6). They were approached by visual means under
daylight, ultra violet light and raking light. What happened to the statue exactly when it came into the
museum was unclear, except that most of the statues were sprayed for a couple weeks continuously.
This has more than likely accelerated sugaring, forming of corrosion and staining due to the old resins.
After fifteen years lying unfinished in the studio dust accumulated on of the surface of the marble
fragments, as they were not covered properly. The different marbles showed different forms of
weathering: the plinth and arms had still a smooth finish, while the head, torso and drapery were
rough. The nose, left ear and right eye, to the right side of the drapery, a piece in the back of the
drapery and some edges of the piece inserted in the back of the left ankle tended to sugar. Some of the
disassembled pieces have smooth joint surfaces, like the ones along the legs and the corner of the
plinth. But the joint surfaces of the tree trunk and top of the left leg bear marks of fine chisel incisions
for better adhesion. The broken edges of the recarved ribbons are smooth. Some of these were
removed in the past and some got lost.
Armature
In total seventeen small corroded pins can be detected visually, some causing iron staining: in
the drapery, nose and left ear and hand. To the back of the right foot and in the tree trunk a corroded

cramp was still in place, which did not cause any structural damage. The same was true for the slightly
corroded square pin, fixed to the back of the right leg and tree trunk. The corresponding hole in the
tree trunk was filled with lead, surrounded with resin. It looked like the cramp at the bottom of the
plinth was the cause of a hair crack. Next to it was a round pouring hole, filled with lead. The pin from
the corner of the plinth was sawn in two. It was fixed with lead and resin. The resin around the lead
probably sealed the bond in the past, but is crumbling now. The left foot and the hole in the right leg
and tree trunk were filled with lead. Both pins were removed and both rough holes clearly widened
when going deeper. In front of the left ankle an incision was made in the smooth break surface that
looked like a pouring channel for lead. Another half of a big square pin was still imbedded in lead in
the tree trunk, behind the right leg of the main body, and appeared to have flared, barbed ends, as the
pin became wider towards the back in a hole that also became wider towards the back. The large
corroded iron cramp left from the tree trunk at the back was causing brown-orange iron staining and
was removed in 1995. The common rule was to hide the armature as much as possible. Several hair
cracks were visible, of which the long one that runs over the drapery looked the most unstable. A
couple of them were consolidated with darkened facings of Japanese or fibreglass paper in 1995.
Staining and residues
It is remarkable that in Ashmole’s publication of 1929 the statue still has all its fingers, they
have since broken off, were lost and the exposed edges became rough. The repaired middle and index
finger of the right hand were broken of together with the scroll. They were patinated with a yellowish
substance. The nose is broken in three pieces and was patinated. The old residues caused orangebrown
staining of the white marble. They are not supposed to be patination. Sportun explained that
two kinds of resin were used: based on two kinds of binding media, maybe to facilitate pouring: a dark
one with a softwood tar (a distillation product of pine) became sometimes powdery. The other one was
more yellow due to mixing in a wax or oil and was not brittle. Marble powder and gypsum were added
to provide bulk and strength. Under ultra-violet light these resins fluoresced. The resin stains were
most persistent along the joints of the marble. Why the drapery also showed similar staining became
clear while removing the sugaring part from the back and removing the resin underneath. This
confirmed that the torso and part of the drapery were put into the drapery as a separate piece. The
marble appeared to be only 2-3 mm thick around the hole. The open space was filled completely with
resin to fix the torso. Due to a damp environment or wet cleaning this resin could migrate to the
surface. More greyish stains – entrapped dirt, were also visible and were slightly soluble in water.
Sportun (Laser) cleaned most of zone II, zone III and the front of zone I.
Tool marks and mechanical damages
By making use of raking light different patterns, in first instance tool marks, on the surface
became more visible. The marks of tooth chisels and flat chisels were visible, some fine scratches
point to the use of rasps, files or fine stone powders to polish the surface like pumice powder. A tooth
chisel was used on the original flat part of the plinth in between the feet. The rest of the restored plinth
has a very smooth, almost polished surface finish, still baring traces of a very fine rasp or stone
powders. The part of the tree trunk between the feet has been carved with a flat chisel in contrary to
the outer side of the tree trunk where traces of a tooth chisel are visible. Besides the tool marks several
scratches and impact damage could be noted, especially on the back of the torso and on the sides of the
plinth. A strange marking appeared across the two pieces of marble on the chest. It is assumed that
some kind of clamp aligned the two pieces together during assembly. Raking light also revealed the
burial marks on the drapery better, which weren’t removed in the past. The roughened surface of the
face for example was also clearly visible in this way. The head isn’t fixed completely anymore as it
moves a tiny bit.

Cleaning
Ill. 7 Ill. 8 Ill. 9 Ill. 10
Ill. 11 Ill. 12 Ill. 13
Dry cleaning
Working according to a phased approach the process of cleaning would develop slowly,
allowing the conservator to adjust the visual balance of the sculpture as he worked, as John Larson
phrased it in the past. The dust that had gathered over the years on the surface of the statue hid the
white Carrara marble underneath. In first instance the removal of this loose surface dirt was achieved
mainly with a soft chip brush and a museum vacuum cleaner (ill. 7). However the ‘Wishab’ sponge
was a good cleaning method for removal of more ingrained surface dirt, it could also damage sugared
marble not visible due to the surface dirt. In the end this technique was used only on the whole surface
of zone II, which was in a good and safe condition.
Mechanical removal of cement, lead and armature
Some cement residues on the back of the plinth were removed with a flat chisel, leaving a thin
layer that was removed with a scalpel. The use of ultrasound abrading tools would be nice to test for
this last step. But the more invasive question was developing a technique for the removal of the small
pins, cramps and lead. Both the chisels and drill bits tended to get stuck in the lead and snapped. All in
all the high steel drill bits seemed to work best for drilling out the lead around the pins (Ill. 8). While
removing the bottom clamp it could be seen that the perpendicular hair crack was not caused by the
corrosion of the clamp, as the crack did not run through the marble underneath the clamp. But the
clamp was probably put in the marble to preventing it from cracking further.
If the corrosion had not expanded the pin in the hole so much, the removal of the small-corroded
pins could be achieved in first instance with a needle, loosening the resin around it. At this stage they
could be removed with pliers, however some pins snapped. Then a three-millimetre diamond core drill
was applied to drill out the pins with as little removal of the marble as possible (ill. 9). A couple of
pins were left in place at the drapery as it might have caused more damage putting pressure on them
during the removal of the ribbon. Where a pin or clamp was left in the tree trunk, the corrosion was
first removed mechanically with a scalpel and sanding tool on a ‘Dremel’. Afterwards a rust inhibitor
was applied. During this tannin process the metal turned black. It is important not to spill any of the
inhibitor on the marble as it may cause staining but also attacks the marble, as its pH is around 3.5. An
interesting technique for future reference would be to test dry ice blasting for the removal of corrosion.
Mechinal and chemical removal of resins

For the removal of paint and thick resin residues a viscous paint stripper: ‘X-TEX’, was used
after the marble was pre-wetted with deionised water. After a minute the blue paste was removed with
a wooden spatula and a swab. Any residues were removed with acetone. Next step was to remove all
the resin residues on the restored edges of the marble. All the removed residues were labelled and kept
separately in the object’s conservation file, for any further research. For the removal different steps
were necessary. After the broken edges were cleaned with a chip brush, the residues were removed
mechanically with wooden pickers, glass fibre brush and scalpels. Brushing this clean was followed
with an acetone swab. Then cotton wool was applied over the surface and dampened with acetone to
soften the residues for about an hour, covered with cling film (ill. 10). With cotton wool swabs of
acetone most of the traces could be removed afterwards. Finally the surface was cleaned with bristle
brushes and low-pressure steam. The darkened Japanese tissue facings consolidating the hair cracks or
fractures could not be removed with deionised water, but with acetone. This indicated that the facing
was not adhered with a water based adhesive such as the ‘Primal WS-24’ emulsion, but with a solvent
based ‘Paraloid’ co-acryllic polymer.
LASER cleaning
Due to the risks of material loss and creating white hazes when using steam to remove the grey
dirt layer on marble that is not always in a good condition Sportun also preferred to use the Qswitched
Nd:YAG (infrared: 1064 nm through a fibre-optic cable) Laser to clean (ill. 11). With a
fluence of around 50 a smooth cleaning was established, similar to the one of Sportun. Laser cleaning
systems can be used on dry marble or one can pre-wet the marble to achieve a certain kind of micro
superficial steam effect. Cleaning the back with or without a pre-wetted surface did not seem to
produce a different result, however the dirty parts on the fronts that were pre-wetted did not yellow.
Strangely enough the Laser cleaning of the back with the same energy seemed to look a bit more
yellow than the front. The application of agar afterwards did not seem to reduce this. When Laser
cleaning the broken off scroll a part of the surface turned black. This appeared to be a small fill from a
former restoration containing lead white.
Agar application
Cleaning with water did work for certain kinds of surface dirt. But the risk in applying deionised
water, through swabs or moreover steam, is the migration of the orange-brown resin stains along the
broken edges and on the drapery, which are difficult to control. By applying a superficial impregnating
agar(ose) gel, a technique was established however with a minimum of water contact to the marble
surface, but still enough to dissolve the stain and pull it back into the gel. The results of a 4% Agar
solution in deionised water (w/v) were very satisfying. The pH of the solution was around 7, like the
marble surface. For the removal of iron stains in plaster agar works very well, but the influence it had
on the removal of iron stains from white Carrara marble was, even after 5 applications, minimal. Its
application directly onto the corroded metal proved however that it was capable of taking up the
corrosion product. The literature states that agar doesn’t leave any residues. This might be true for a
smooth surface, but on a rough marble surface this appeared not easy to achieve when the gel was
completely dried out (ill. 12). Its application on sugared marble needed to be avoided as the dried up
gel took up loose crystals of marble.
The difficulty in applying agar is to get an even layer of about 1 centimetre (also along the
edges) on a vertical surface without creating drips, so it is easier to remove afterwards. Pouring or
brushing did not always work out when the liquid was too warm and it took too long to form a fixed
gel. A thin gel layer tended to form quickly whilst he bulk of the liquid ran off the vertical surface. A
reapplication to created one gel layer created two or more layers with peeling. When dried out
completely the residues can be clearly seen under UV-light. A full application to the torso did not
reduce the possibility of creating stain tide lines, as could be seen also under UV-light. On smooth
surfaces like the plinth the dried out gel was very easy to remove, as it was lying loose on the surface.
But on the coarser surface it got stuck and light pressure steam seemed to be the best method to
dissolve and remove it with a bristle brush, with migration of the stains as a consequence. In that sense
the decision was made to apply the agar only locally and tests pointed out that the agar was able to
remove the stains after only 10 to 30 minutes. After searching for the right application methodology
quite a quick technique was developed in the end. By working in dammed sections a still flexible gel

of about one centimetre could be removed with wooden spatulas, while trying not to leave any
residues (ill. 13 and 19). Before the back of zone one could be cleaned in the similar way, and after the
Laser cleaning had taken place, the assembly of the statue needed to be completed.
Re-assembly of an over-life-size marble statue
Introduction
With the pins removed and the old cavities cleaned the next phase of the project could start
reinstating the structural integrity of the statue. Different stainless steel pins were used depending on
their location and specific purpose, different types of adhesives and armature needed to be selected
according to their function (consolidation, structural or not and time available during application), all
the time bearing in mind the leitmotiv of minimal intervention and ease of disassembly in the future.
In that sense the meaning of reversibility: the complete undoing of an adhesive bond with no or
minimal effect on the substrate is not always easy to achieve; because the resistance of epoxies and
polyesters towards a more common solvent is considered to be one of their best characteristics. The
choice of adhesives was discussed with Norman Tennent and the internal forces that this kind of
armature need to withstand was with Bill Wei (RCE) and Jerry Podany (Getty); but could not be
calculated within the timeframe. Their knowledge was very helpful during the design of piston fit pin
sleeve system.
Ill. 14 Ill. 15 Ill. 16
Selecting appropriate adhesives
It was necessary to apply barrier coatings on areas where epoxy otherwise would come into
contact with the marble. First a 4% Paraloid B-72 in acetone was brushed on the surface, followed by
a coating of 10% the next day. Both coatings were left to dry for a couple of days. The 4% solution
was also used to consolidate the thin shattered left knee piece. Caution was taken that the solution did
not reach and possibly darken the surface that would be visible. Pieces (for example the nose –
including small stainless steel pin, left ear, drapery pieces, small pieces near the tree trunk, the odd
rectangular piece at the back of the plinth), which did not have any structural function, were glued
with a 50% solution of Paraloid B-72 in acetone, bulked-out with marble powder (grade 100) (ill. 14).
This was done to reduce shrinkage caused by the evaporation of the solvent and to increase the
compressive strength (if necessary). The fingers were fixed with ‘General – Vertical Transparent –
Polyester’ after the marble received a barrier coating.
For the structural adhesions the first choice was the use of a viscous high-grade epoxy, like
‘Araldite 2011’. However, although this adhesive is already viscous, extra Fumed Silica was added to
tweak its viscosity (to reduce dripping) before application, but also to create a more flexible cured
adhesive that can withstand greater tensions during future movements.
Selecting appropriate armatures
The broken pieces of the corner of the plinth were glued together with ‘General – Vertical
Transparent – Polyester’. As a whole it was glued with dots of ‘Araldite 2011’ epoxy and a keyed in
stainless steel pin, bulked-out with Fumed Silica. Besides the classical solid pin a ‘piston fit pin sleeve

system’ was developed. Inspiration was drawn from other semi-bonded systems, similar to the ones
developed by the J. Paul Getty Museum. In this system the pin slides into two separate loose sleeves,
each closed from one side, which are fixed with the thixotropic epoxy.
The torso was strapped around the drapery, avoiding any movement of the two separate pieces.
After the first sling was applied and the statue was lifted up diagonally a second sling was applied.
Bearing in mind its centre of gravity the correct position was reached to lift the statue according the
corresponding alignment of zone two. During lifting all the hair cracks were monitored. A dry test run
was completed to check the procedure. First the left leg was fixed with a solid pin (keyed in along the
surface and rounded at the edges) with quick curing polyester: ‘General – Vertical Transparent –
Polyester’ (ill. 15). This choice was made, because the leg could be held in place while zone one was
hanging into the correct position over zone 2. After the polyester cured the statue could be lifted with
the left leg. The big sleeves were then placed into the right leg and the small sleeves were placed into
the left ankle. Both of them were placed into a bed of the bulked-out ‘Araldite 2011’ epoxy with
Fumed Silica. Using this slow curing epoxy allowed the time to place everything into its exact position
(ill. 16). The pins were first sprayed with a release agent to reduce friction before they were slid into
the sleeves. Now zone one could be lowered in the correct position onto zone two. The slings were
loosened but left in position, because the epoxy needed to cure over night. As the statue was leaning a
bit forward on the plinth (which narrows down to the front) it was decided to raise the front of the
plinth by 12 mm, to correct the centre of gravity, creating a more stable object. On the back of the
statue a new cramp, following the contours of the surfaces was fixed with the bulked-out epoxy, where
the old one next to the tree trunk had held zone one and two together as an extra support. The two
pieces of the tree trunk were fixed with dots of the bulked-out epoxy to give extra support over epoxy
around the tarnished cramp. The hole in the top fragment of the tree trunk was filled with this epoxy as
it slid over the old tarnished square. The statue was leaning a bit more forward. The chin of the right
leg was glued with the bulked-out Paraloid B-72 with marble powder as it almost stayed in position by
itself and had very little structural support.
Loss compensation for indoor white marble sculpture
Introduction
Several criteria were involved in considering materials to be used for loss compensation on
white marble sculpture. Choices were based largely on former research when selecting materials for
making a ‘colour fill chart’ as a reference (ill. 17). Through that selection procedure (of seven bonding
media and seven aggregates) the main criteria during application needed to be its visual appearance
and its workability, which were described. The environment in which the sculpture will be stored or
displayed plays a significant role in the range of materials to be used. Objects on display outdoor have
a more limited range of materials, which are preferably sacrificial lime based mortars. Interior objects,
like this case, have a wider range of possibilities – also making use of synthetic materials. Each
sample recipe in the chart could be adjusted by adding more or less fillers to a certain binding media
when searching for the right properties. It became clear that not all combinations were a success. Some
of them work very pleasantly while others were more difficult to work with, or don’t give you much
working time. Remarkable was that a couple of the test samples failed completely as they crumbled
while shrinking, produced bubbles or had the wrong colour.

Ill. 17 Ill. 18 Ill. 19
Application procedure
On the bottom of the plinth a test was conducted in the gap left by the removed clamp. After
applying barrier coatings (4% and 10% Paraloid B-72 in acetone) and filling the gap with ‘Polyfilla’
almost completely, the last final layer was applied with the colour fill. Into a high-grade epoxy of
‘Epotek 301-2’, Fumed Silica was mixed to create a paste. A combination of alabaster- and marble
powder was mixed into this to obtain a similar colour and translucency as the marble, while no
pigments were added (ill. 18). Pigments can be added according to the surrounding stone. Because the
gap was quite big the fill tended to slump after a while. To avoid this more Fumed Silica should be
added in advance. In this case the filling was kept in place by covering it with a ‘Melinex’ sheet. If the
filling is too high it can be cut back with rifflers, files or scalpels, without creating any sort of
mechanical damage to the marble. For the other fills another high-grade epoxy, ‘Hxtal-NYL 1’, was
chosen as it has better light ageing properties.
Protective coating
After treatment the procedure at the Conservation Centre is to apply a protective coating to
prevent any dust, dirt or grease (and any unwanted additions by the public) getting trapped into the
surface or cracks. It both seals the open porous structure and must be reversible. The natural result of
such treatments is to enhance the translucency of the marble. Some argue that this would soften the
effect of the tooling (‘sfumato’) giving the sculpture a more ‘finished’ appearance than in reality. The
‘Renaissance micro-crystalline wax polish’ was applied with a soft brush or a cloth onto the surface of
the marble and left to dry for at least 30 minutes. When the wax had partly dried, the excess wax was
buffed away with a clean cloth or brush. After this a second layer was applied to impart sheen to the
surface. By this method a uniform finish was achieved over the whole object that enhances the ‘patina’
without the appearance of an applied coating.
Conclusion
One could describe the composite statue, formerly labelled: ‘Marcus Aurelius’ (Ince 569), as an
over-life size free-standing statue, with a bare torso and a ‘hip mantle’ held together by his left
advancing arm, which recalls the Claudian timeframe. However his stance was already inspired by the
early canon from Polykleitos’ Doryphoros. His physiognomy beared more similarities with the second
type of ‘Marcus Aurelius’ than the first type. But it was not certain that this portrait really represented
‘Marcus Aurelius.’ Especially considering the ivy wreath around his head, which may have pointed
into the direction of a Claudian (Antonine) or Hadrianic type, maybe as part of the Liber Pater cult.
Henry Blundell’s private collection of antique sculpture became the second largest in the United
Kingdom, after the Townley collection (British Museum). As a ‘composite’ sculpture ‘Marcus
Aurelius’ represented the fashion of the eighteenth century restoration ethics. However the origin of
this statue remained vague, Blundell acquired it at the end of his life. He displayed it within a niche in
the greenhouse of his Ince Blundell Hall just north of Liverpool. Until 1959 the statue stayed on his
estate. Afterwards the collection was trusted to what became National Museums Liverpool, who still
conserve of a large part of the collection. What happened to the statues, besides a thorough wet
cleaning, remains unknown. It was only until 1995 when Samantha Sportun undertook research into
this statue, started disassembly and laser cleaned different parts that more information about this

sculpture became available.
Sportun did not complete the treatment of the sculpture and it was left to rest within the
sculpture studio for at least fifteen years. In the fall of 2011 the author started writing a treatment
proposal during his internship for the University of Amsterdam. The sculpture was divided into three
different zones: the plinth – with the feet (2) was separated from the main body (1) and another zone
consited of 58 loose fragments, which could be located back on the object (3). A range of pins and
clamps needed to be removed from lead and/or eighteenth century adhesive. After cleaning the surface
dirt with a soft brush, the disturbing brown orange resin stains could be removed or tempered from the
white Carrara marble. A superficial impregnating gel worked by dissolving 4% agar in deionised
water. As the resin stain was soluble in water this gel was applied in separate zones and could be
removed after already fifteen minutes with a wooden spatula. The rough surface made it not always
easy to remove all residues of this gel. Grey dirt layers and sulphation could be removed with a
Nd:YAG Laser. The only downside was a yellow ‘discoloration’ of the back, which was not prewetted
in advance, for avoiding migration of resin stains. Small pieces on the front of the drapery and
the leg were pre-wetted and had the same colour as it surrounding marble which was laser cleaned by
Sportun years earlier.
This project also allowed for the development of a technical approach to re-assembly,
addressing issues like stability and reversibility of its adhesives and armature. Adhesives were being
used according to their function: 4% Paraloid B-72 in acetone for consolidation and as a barrier layer
(4% followed by a 10%) before applying a quick setting polyester of ‘General – Vertical Transparent –
Polyester’ or a slow curing epoxy, ‘Araldite 2011’, bulked-out with Fumed Silica to increase viscosity
and strength. A 50% of Paraloid B-72 in acetone bulked-out with marble flour to reduce shrinkage was
chosen for non-structural adhesion. Stainless steel pins were used and a ‘piston fit pin sleeve’-system
for the right leg and the left ankle was designed to enable an easier disassembly, if necessary in the
future.
After testing several materials for loss compensation of white marble, gathered in a reference
‘colour fill chart’, the choice was made to use the high grade epoxy: ‘Hxtal Nyl 1’; together with
Fumed Silica, marble powder, cooked alabaster powder and if necessary pigments to imitate the
surrounding marble. This ‘colourfill chart’ worked as a reference and each recipe could be adjusted by
adding more or less fillers when searching for the right properties. After the application of a barrier
coating of Paraloid B-72 in acetone (4% followed by a 10%) and a bulk fill of ‘Polyfilla’ if the gap
was considerable, the epoxy mixture was applied. By closing cracks or missing parts (without
remodelling parts) the chances of the accumulation of dust and other surface dirt was reduced.
Applying a protective coating of microcrystalline wax further enhanced this and its look.
Acknowledgments
Many thanks are due to the numerous people who helped me with the realisation of this project:
Martin Cooper and Annemarie La Pensée from Conservation Technologies for answering Laser
related questions and to be able to be part of the workshop on Laser cleaning. More over I want to
thank my supervisor and head of sculpture conservation, Lottie Barnden. Her patience, focus and
sensitive approach towards our field will be remembered. Also a big thank you to Bill Sillitoe and
especially Dave Whitty from Technical Services. I don’t want to forget the people from the other
disciplines providing me with materials and discussing subjects. I could always approach the curator
of antiquities, Gina Musket, with questions relating to ‘Marcus Aurelius’ his background. Furthermore
I want to thank specialists in the field: Prof. Norman Tennent (UVA), Deborah Carty (private
restorer), Bill Wei (RCE, Amsterdam) and Jerry Podany (J. Paul Getty Museum) for having e-mail
contact about adhesives and armatures. Last but not least my lecturer Lisya Biçaçi, who insisted on me
going to Liverpool – a decision I will never regret.
References
- Cooper, Martin, ‘Laser Cleaning of Sculpture, Monuments and Architectural Detail’, Journal of

Architectural Conservation, November 2005: 201-119.
- Fejfer, Jane, The Ince Blundell Collection of Classical Sculpture. Vol. 1 - The Portraits, Part 2
The Roman Male Portraits, Liverpool, 1997.
- Fotakis, C., e.a, Lasers in the Preservation of Cultural Heritage. Principles and Applications
(Series in Optics and Optoelectronics), New York – London, 2006.
- Garland, Kathleen M. and Rogers, Joe C., ‘The Disassembly and Reassembly of an Egyptian
Limestone Sculpture’, Studies in Conservation, vol. 40, no.1 (1995): 1-9.
- Griswold, John and Uricheck, Sari, ‘Loss compensation Methods for Stone’, Journal of the
American Institute for Conservation, vol. 37, no. 1 (1998): 89-110.
- Jorjani, Mersedeh, e.a., ‘An evaluation of potential adhesives for marble repair’, in Janet
Ambers, e.a., Holding it all together - Ancient and Modern Approaches to Joining, repair and
Consolidation, London, 2009: 143-149.
- Kleiner, E.E., Diana, Roman Sculpture, New Haven and London, 1992.
- Nagy, Eleonore E., ‘Fills for White Marble: Properties of Seven Fillers and Two Thermosetting
Resins’, Journal of the American Institute for Conservation, vol. 37, no. 1 (1998): 69-87
- Podany, Jerry, ‘Faked, flayed or fractured? Development of loss compensation approaches for
antiquities at the J. Paul Getty Museum’, Abstract of papers at the Twenty second annual
meeting, Nashville, Tennessee, June 6-11, 1994, Washington, 1994: 84-85.
- Podany, Jerry, et al., ‘Paraloid B-72 as a structural Adhesive and as a Barrier within Structural
Adhesive Bonds: Evaluations of Strength and Reversibility’, Journal of the American Institute
for Conservation, vol. 40, no. 1 (2001): 15-33.
- Podany, Jerry, Risser, Eric and Sanchez, Eduardo, ‘Never forever: assembly of sculpture guided
by the demands disassembly’, Holding it all together - Ancient and Modern Approaches to
Joining, repair and Consolidation, London, 2009: 134-142.
- Rhabar, Nima, e.a., ‘Mixed mode fracture of marble/adhesive interfaces’, Materials Science and
Engineering A, 527 (2010): 4939-4946.
- Rockwell, Peter, Rosenfeld, Stanley and Hanley, Heather, The Complete Marble Sleuth, Rome,
2004.
- Sportun, Samantha, Case study: An 18 th century restored classical male statue, MA.
Conservation of Historic Objects, Durham University, Department of Archaeology, 1996.
- Sportun, Samantha, ‘The Investigation of Two Male Sculptures from the Ince Blundell
Collection’, History of restoration of Ancient Stone Sculptures - Papers from a symposium,
October 2001, Los Angeles, 2003: 127-136.
- Wolfe, Julie, ‘Effects of Bulking Paraloid B-72 for Marble Fills’, Journal of the American
Institute for Conservation, vol. 48, (2009): 121-140.
Illustrations
1) Bartolomeo Cavaceppi, Raccolta d’Antiche statue busti bassirilievi ed altra sculture Restaurate
da Bartolomeo Cavaceppi sculpture Romano, Rome, Vol. 1, 1768, Rome: Augusto, ill. 33.
2) Photo of the Ince 569 at the Ince Blundell Hall, in a nice of the greenhouse.
Ill.: Ashmole, John, Catalogue of Ancient Marbles at Ince Blundell Hall, Oxford, 1929: 50.
3) Colour scheme by Sportun, Highlights of different parts.
Ill.: Samantha Sportun, History of restoration of Ancient Stone Sculptures - Papers from a
symposium, October 2001, Los Angeles, 2003: 128 and 171, plate VIII: Drawing of ‘Marcus
Aurelius’ figure, p. 128 based on an engraving by Bernard de Montfaucon (1655-1741), from
H. Blundell, Engravings and Etchings, pl. 35.
4) Zone I in the studio after 15 years, ill. by author.
5) Zone II in the studio after 15 years, ill. by author.
6) Zone III in the studio after 15 years, ill. by author.
7) Dry cleaning with brush - half of the statue, ill. by author.
8) Removed of pin and lead, ill. by author.
9) Drilled out small pins with 3 mm diamond coredrill, ill. by author.
10) Acetone poultice on the surface to soften the adhesive and take it of with cotton swabs, ill. by

author.
11) Laser cleaning the back with a fluence around 50, ill. by author.
12) Residues of agar fluorescence under UV-light after complete covering, ill. by author.
13) Local application of agar - 30 minutes, ill. by author.
14) Application of 50% Paraloid B-72 in acetone bulked-out with marble powder for nonstructural
fillings, ill. by author.
15) ‘General – Vertical Transparant Poyester’ for quick fix structural fixings of a solid pin, ill. by
author.
16) Bulked-out ‘Araldite 2011’ with Fumed silica for slow setting structural fixing of the piston fit
pin sleeve system, ill. by author.
17) Representative ‘Colour fill chart’, ill. by author.
18) Colourmatching the colourfill: ‘Epotek 301-2’ with Fumed Silica, Alabester powder and
marble powder, ill. by author.
19) Man with an ivy wreath – ‘Marcus Aurelius’, Ince 569, National Museums
Liverpool, after reassembly, ill. by author.

CONSERVATION OF ARCHAEOLOGICAL MEGALITHIC SITES UNDER
MARINE ENVIRONMENT: EXAMPLE OF THE GRANITIC MENHIRS AT I
STANTARI (CAURIA PLATEAU, CORSICA ISLAND, FRANCE)
Jean-Marc Vallet 1 , Philippe Bromblet 1 , Emilie Heddebaux 1 and Nicolas Bouillon 1
1 CICRP 21, rue Guibal F-13003 Marseille, France.
Abstract
Granitic menhirs from the Neolithic to Bronze Age period are present in few
places of Corsica Island (France) such as I Stantari alignments on Cauria plateau. The
emblematic alignment of statue-menhirs of Stantari shows various degradation features.
They are characteristic of granite stones alteration and lead to the disappearance of the
fine sculpted relief. A first campaign of conservation work using acrylic resins as
consolidant was performed in situ in 1993. Ten years after this intervention, the
archaeologists who were working on the site, asked for an evaluation of the efficiency of
the treatment. The performance of this impregnation of the whole monoliths has been
evaluated by means of visual examination, ultrasonic velocity measurements and
microanalyses performed on several samples in the laboratory. The investigations have
shown that the intervention didn’t succeed to improve the conservation of the sculptures.
The properties of the selected consolidant product, the application mode and the
deficient diagnosis were involved in this failure. An experimental procedure was then
set up to test new products that could be applied on the granitic stones. Two acrylic
resins including the one yet used, ethyl silicate and ethyl silicate combined with a silane
resin were tested as strengthener. None of these products are efficient enough. The
causes of the degradation and the adequacy of the product application are discussed in
regard to the specific properties of the stone material.
Keywords: megaliths, granite, degradation patterns, ultrasonic velocities, consolidants,
experimental ageing.
1. Introduction
I Stantari menhirs and statue-menhirs alignment is one of the main megalithic sites
of Corsica Island (France). They are located near Sartène on Cauria plateau at a few
kilometres from the seaside (figure 1) and were discovered during archaeological
excavations (figure 2). Only 7 were still visible during the XIXe century as Adrien de
Montillet described it. Robert Grosjean's team dug out others in 1964 and 1968 (In
d'Anna et al, 2004). In 1975, some menhirs belonging to the alignment were set upright
again in order to re-establish the alignment of 12 raised monoliths that are currently
visible. Their orientation is north-south. They were probably initially dressed up at the
beginning of the final Bronze Age, towards 1000BC.
Deterioration issues were first described after Grosjean's excavation campaign in
1968. The degradation partially erased the relieves on several menhir-statues. A first
study was made in 1993 in order to establish a degradation diagnosis and to propose a

treatment protocol to stop the degradation (Casta, 1993). A strengthening treatment was
then applied in situ on most of the raised megaliths (Casta and Poli, undated).
During the 2002-2004 periods, d’Anna et al. (2004) who was starting new
archaeological excavations on the site considered that a new diagnosis was required.
They were wondering what had been the effects of the treatment. They asked for a new
diagnosis of the granite degradation including a condition report of the raised megaliths.
The current study deals with the evaluation of the 1993 intervention, the diagnosis
of the stone degradation, and laboratory testing of new products that could be used to
consolidate the granite.
Figure 1. Location of the site of I Stantari (Corsica Island, France).
a
b
Figure 2. View of I Stantari’s site during the excavation campaign in 2007 and its erected and
strengthened monoliths (11 are visible and are referenced in the insert).
2. Description of the raised megaliths and their state of conservation
Megalithic granite is a monzogranodiorite. Microgranites and aplitic sills locally
cross it. It is constituted in quartz, feldspars, biotite and amphiboles. The potassic and
calco- sodic feldspars are about 1 to 2 cm in size (J.-B. Orsini in d'Anna et al., 2004).

Granite megaliths of I Stantari’s site show degradation patterns. Several networks
of fissures affect them: the largest ones mark the stone foliation and interganular
microfissures lead to the granular disintegration of the stones. The cracks network is
very important. The main ones are vertical and parallel to the granite foliation. Flashings
used in 1993 plug them. The other are inter or intra minerals and mainly affect feldspars.
The statue-menhirs present faces and several armed warriors facing east. The
surface of the stone is eroded. The carved faces, swords etc. are difficult to see. The
exposed West face of the erected stones to the prevailing wind and the rain is rougher
than the opposite one. Granular disintegration affects also the West faces of the
megaliths. East faces and mostly their bottom show desquamations and blistering.
Scaling and locally blistering are visible on some monoliths (M2 and M4). The scales
are few millimeters thick and bases of blistering are 4 to 5 millimeters in diameter. M5
monolith shows little white craters few millimetres to one centimetre in diameter. These
degradations locally lead to a loss of materials.
A red colouration locally affects the erected stones except in their excavated parts.
This reddening is probably due to fires.
Past and recent biological colonisations (mostly mosses and lichens) are locally
present on the monoliths, mainly in their upper part. Lichens are grey or have a green to
yellow colour. Some are under the consolidant film, other developed themselves on it.
All of these alterations affect also the granitic outcrops which would be the quarry.
3. Evaluation of the 1993 intervention
According to Casta (1993) and Casta and Poli (undated), biological colonisations
described as lichens were cleaned using hydrogen peroxide. Scales were stuck using an
Acryl TM AC261K (Röhm and Haas) and the large fissures were filled by application of a
flashing with the same resin. The strengthening impregnation using this water dispersed
acrylic resin had been performed by infiltration from the upper part of the stones in two
steps, first 40% in dilution, then at 60%. This method was preferred to brush application
to have the best penetration of the resin (Casta and Poli, undated). It was expected to
give a homogeneous repartition of the resin; a long contact with the stone which should
allowed making easier the absorption.
The observations in 2006 and 2011 showed that the degradation is still active but it
seems to be slow. Because of the age of these cultural heritage objects and the shallow
residual engravings, the degradation has to be stopped and a precise conservation policy
has to be carrying out (Vallet et al., 2008). First, Casta’s consolidant has been
characterised using spectrometries Raman and FTIR and the nature of the product was
able to be determined as a poly(nButyle Acrylate- MethylMethacrylate) which is
equivalent to the current one (Rogalle, 2006). FTIR spectra revealed also the presence of
polyethylene glycol which shows a degradation of the resin.
The surfaces aspect did not change a lot. No new fissure and no new coloration
appeared. Lichens size was quite constant. Some monoliths (M2, M4, M5, M7 and M12)
show new degradation patterns mostly on their oriented East surface. Chatain (2010)
underlined that these monoliths and their East surface were already the most sensitive to
the degradation. Scaling and blistering were in progress, mainly on the lower part of the
East oriented faces. One can see the discontinuous film on the surface (figure 3). There
are locally lichens, green algae under it. The surfaces of the monoliths seem to be

washed. Their base and the upper part of the surrounding soil have been strongly
hardened. In 2008, the surface was so fragile that it was impossible to make small cores.
Figure 3. Detail of the surface showing that the acrylic film does not cover the entire surface and
that there are some lichens which grow.
Ultrasonic velocities measurements were performed in 2005 and 2011 to study the
cohesion of the 12 monoliths and to evaluate the residual consolidant effects. An
AU2000 from CEBTP (54 kHz) which measures the times that the ultrasounds take to
cross the stone has been used in transmission. They were performed from bottom to top
on the West-East faces and also along the North- South direction on the same spotting
points. 3 measures have been done on each point.
Megaliths show sound velocities less than 3000m/sec. These velocities (megaliths
2 and 7 excluded) are equivalent to those measured on walls from the nearest granitic
outcrops (from 1962 to 2729 m/sec; see figure 4). Compared to the velocities value on
fresh granite (6000 m/sec), these ultrasonic velocities are very low.
Some variations occurred from 2005 to 2011 campaigns (figure 4). Most of the
megaliths show the same average velocity except M5, M6, M7, M11 and M12 where the
ultrasonic velocity seems to diminish a little. The deviation of the measures is lower in
2011 except in case of M1, M6 and M7. The ultrasonic velocities confirm the
observations on degradation. They have not decreased significantly. The variations of
the measure for most of the monoliths (except M2 and M9) have diminished. This can
be linked to the loss of scales which leads to a better cohesion of the surface.
The degradation patterns affecting I Stantari’s monoliths (cracks, scaling, granular
disintegration, and lichens as biological colonisations) are classical on granites (Silva et
al., in Sanmartin et al., 2008). The degradation depends on several intrinsic and external
factors. The monoliths have been sculpted in weathered granites as the comparison of
the ultrasonic velocities and the observation of degradation patterns on both the near
possible extraction site and megaliths shows.
Alvarez and al. (2008) indicate that granular disintegration depends on the rainwater
spray and capillarity. Rising damps effects and splashing can also explain the location of

the decays which mainly affect the bottom part of the monoliths. The climate induces
differential erosion of the monoliths: the West side suffered from prevailing wind,
sunshine and drastic dry- humidity cycles. Degradation leads then to a coarse roughness
of the granite surface. Another cause would be the partially buried position of the
monoliths (M2, M3 to 5, M7, M8 to M12) before their erection in 1968 (Grosjean’s
excavations). But none available comparative element could confirm this. More, the
presence of saline spray combined to dry periods leads also to the preferential granular
and mineral decohesion as the observation and characterisation of powder collected on
experimental samples showed. Phyllosilicates, clay minerals and quartz are mainly
detached from the surface of the stone in presence of salts which are supposed to come
from marine sprays. Collective (2008) and Sebastian et al. (2008) note that the presence
of both soluble salts and swelling clay minerals leads to the scaling.
ultrasonic velocities (m/s)
3500
3000
2500
2000
1500
1000
500
0
-500
6
7
7
6
7
3 3 3
4
4 7
M1 M2 M4 M5 M6 M7
monolith
M8 M9 M10 M11 M12
vmoy.2005
vmin2005
vmax2005
v.moy2011
v.min2011
v.max2011
Figure 4. Variation of medium ultrasonic velocities for each monolith from 2005 to 2011. (The
grey straight lines correspond to the measurements on the nearest granitic outcrops in 2005. x is
the number of measurements that we did along each monolith).
Because the degradation is still active, and because of the lichens development
since 1993, one can consider that Casta’s consolidation using the Acryl product is no
more efficient. More, the diagnosis (Casta, 1988) appeared to be incomplete because the
causes of the degradation were not taken into account before this treatment. In another
hand, no explanation about the choice of the product was given. Last, the kind and the
period of treatment were inappropriate because a rainfall was supposed to wash partially
the product (Casta and Poli, undated) and the soil around megaliths is strengthened.
4. Laboratory testing of new consolidants
4.1 Presentation of the studied granite
Because of the difficulties to obtain some Cauria granitic samples coming from the
nearby outcrops, experimental samples were cut in damaged area of the Sidobre's
granite (near Castres, France). The mineralogy of the Sidobre's monzogranite is closed
to the megalithic one. It is constituted in quartz, potassic and calco- sodic feldspars and
biotite, which are respectively 36%, 33%, 34% and 6% in volumetric modal
composition; from Didier, 1991). Some potassic feldspars are few centimeters in size.
The measured ultrasonic velocities are quite equivalent to those coming from I Stantari’s
monoliths; they vary from 1000 to 3000 m/sec.

50 samples were cut as cubes from 10 cm in size in the peripheral area of the
exploited blocks in order to have damaged part of the granite. These areas are weakly
yellowish because of the presence of iron oxides coming from the partial biotite
transformation. The granite shows thin fissures.
4.2 Experimental treatments and accelerated ageing protocols
Four consolidant products have been applied on 11 samples each in a classical way
of 3 applications until saturation point is reached and using a paintbrush (Begonha and
Fojo 2008, Sanmartin and al., 2008):
- The acrylic resin ACRYL TM AC261K as the reference;
- The acrylic resin Paraloïd TM B-72 (Röhm and Haas);
- The ethyl silicate SILRES® BS OH 100 (Wacker) which is largely used in
France and seems to give quite good results (eg Alvarez et al., 2008; Beghona
and Fojo, 2008; Costa and Rodrigues, 2008; Rojo et al., 2008);
- The same ethyl silicate associated with the water repellent SILRES® BS290
(Wacker) which is a copolymer silane/ polysiloxane without solvent because
the use of silanes (Begonha and Fojo, 2008; Wheeler, 2008) and this kind of
association seem to give interesting results (Kim et al., 2008).
Both climatic and salt spray accelerated ageing chambers were performed on the
samples. 5 samples for each treatment and for one ageing chamber have been studied
and 1 other has been kept as the reference. The climatic ageing was performed using a
VC4034 chamber from Vötsch. The samples suffered 282 cycles (table 1). The first 114
cycles reproduced climatic variations closed to South Corse climate ones then
temperature constrains were increased during the 168 last cycles.
Cycle 1 Cycle 2
T (°C) RH (%) D (h.) T (°C) RH (%)
25 65 2 25 65
45 20 2 85 17
10 90 2 10 90
-10 0 2 -10 0
Table1. Details of the climatic accelerated ageing cycle. (T, temperature; RH, relative humidity;
D, duration in hours).
Because of the close distance to the sea, soluble salts effects of on strengthened
granite had to be investigated. Soluble salt have not been clearly detected except on the
surface. Then, non salty samples have been aged using a saline spray chamber (SC450
chamber from Weiss Technik). The samples suffered 210 cycles in presence of halite
(NaCl 10% in mass). The protocol of the first 56 cycles was: 2 hours at 25 °C, 5 minutes
to increase the temperature until 40°C then the decrease of the temperature during 55
minutes until 25°C. But because of an insufficient drying, the samples were then dried
every 6 cycles at 18

effects of strengthening and of ageing as well. In order to compare the ageing results, a
balancing coefficient has been applied in order to take into account the differences of
velocity between the untreated stones.
Weight measurements were done on samples in order to determine the consumption
of the applied consolidant and the loss of material.
The collected powder on the samples was first cleared out of halite using distilled
water then crushed. The 0-4 µm granulometric fraction has been sampled. Minerals were
identified using a X- Ray Diffractometer D8 Advance from Bruker (Cobalt tube, θ-2θ
configuration, 40 KV- 35mA) by way of the powders method and applying the Bragg's
law. The nature of clay minerals was also determined by way of oriented deposits, then
ethylene glycol and heat (500°C during 2 hours) treatments of the deposits.
4.4 Results
The consumption of the different products is low (table 2). The products enter
weakly in the stones. The variations between the cubes are high in case of acrylic resins.
Consolidant Kg/m² variations
Acryl 0,027 28,9%
Paraloid 0,010 44,8%
Ethyl silicate 0,145 14,1%
Ethyl silicate+water repellent 0,074 18,2%
Table 2. Average consumption of the consolidants applied on 11 samples.
If we compare the ultrasonic velocity before and after consolidation, it increases for
all the treatments. Nevertheless, the strengthened treatment induces variations in terms
of ultrasonic velocity (figure 5):
- the velocity increases more in presence of consolidant containing ethyl silicate
than in presence of acrylic resins;
- the treatment led to differences depending on the direction of measurement
(parallel or perpendicular to the foliation);
- Acryl treatment shows the lowest increase of velocity.
After the climatic ageing, all the ultrasonic velocities are reduced and are quite
equivalent, excepted Paraloïd. The strengthened treatments seem to have a little residual
effect because the difference is a little bit lower than in case of untreated stones.
In the case of saline spray ageing, all the treated stones show a positive difference
of the velocities between before and after while this difference is negative in case of
untreated ones. The best results concern the ethyl silicate- base ones.
Powder appeared on cubes after ageing. Very few grains were collected using a
paintbrush on the samples which suffered the climatic ageing. The powder was collected
in the case of saline spray ageing. It was weighted after dissolution of halite in order to
determine the contribution of the stones. Table 3 recapitulates the results. Untreated
cubes loss more material than the treated ones. The ageing of the treated cubes with

ethyl silicate+ water repellent treatment led to a quite important loss of material
compared to the other treatments. Paraloïd B72 and Ethyl silicate are the less affected.
Difference of velocities (m/s)
1400
1200
1000
800
600
400
200
0
Acryl Paraloid Ethyl silicate
Treatments
ethyl silicate
+
water repellent
perpendicular to the foliation
parallel to the foliation(1)
parallel to the foliation(2)
Figure 5. Evolution of the ultrasonic velocity depending on the nature of the applied treatment
Difference of velocities (m/s)
400
300
200
100
0
-100
-200
-300
-400
-500
-600
Non treated Acryl Paraloid Ethyl silicate ethyl silicate
+
water repellent
Treatments
saline srpay ageing, perpendicular to the foliation
saline srpay ageing, parallel to the foliation(1)
saline srpay ageing, parallel to the foliation(2)
climatic ageing, perpendicular to the foliation
climatic ageing, parallel to the foliation(1)
climatic ageing, parallel to the foliation(2)
Figure 6. Comparison of the ultrasonic velocity variations vs the applied experimental ageing
The collected powder is mainly composed of biotite, chlorite, quartz and feldspar.
Interstratified mineral containing a swelling clay mineral has been detected in very few
quantities. This is an illite- smectite mixed layer. The content of smectitic layer is low if
we consider the XRD positions of dome- shaped peak depending on the preparation
(1.28 nm deposit, 1.33 nm ethylene glycol treated, 1.23 nm (Ir weak) heated).

The accelerated climatic ageing affects all the strengthened cubes. The differences
between the products are not significant. Ethyl silicate seems to give more resistance to
the stones than acrylic ones face to a saline spray accelerated ageing. But none treatment
gives satisfaction as the equivalent ultrasonic velocities measured on treated and
untreated samples suggested.
Table 3. Collected mass (g) of powder on aged cubes depending on the nature of the
treatment and the kind of experimental ageing they suffered.
The collected powder on samples after ageing under salt spray contains more
phyllosilicates and quartz in the Paraloïd sample than in the Ethyl silicate sample. This
shows a higher alteration of the Paraloïd strengthened samples.
The measured consumptions of all the experimental samples are low. In another
hand, the increase of the ultrasonic velocity after treatment was weak and not so
different from the velocities measured on untreated weathered granite. Therefore the
penetration depth is probably low because of the porosity (Sanmartin et al., 2008). A
better evaluation of the durability of these kind of treatments applied on a granite needs
then a higher penetration depth. This last would also favour a better resistance to
hygroscopic salt degradation (Sanmartin et al., 2008). Despite the employed application
technique of the products is the usual one (eg Begonha and Fojo, 2008) and the most
adapted to the isolated sites, the application method must be modified. Different
techniques can be applied: immersion gives good results in laboratory (Sanmartin et al.,
2008) but it is totally not adapted to such big and heavy artworks the monoliths are; the
gravity impregnation did not lead to good results on I Stantari's megaliths because the
acrylic resin did not penetrate enough in the porosity. Last, vacuum impregnation as the
Ibach method could be a good way for both in field and in laboratory treatments.
The two families of resins are different in terms of their properties. Acrylic resins
have quite elastic properties compared to ethyl silicate ones but they are film- forming.
They do not give good results in terms of durability. The first results concerning ethyl
silicate and the combination of ethyl silicate with silane are not also convincing.
Silicate- based resins are not elastic and can induce crack networks in the resin and the
treated stone as well. The presence of clay minerals increases the issue (Félix, 1995) but
the addition of micro or nano- particles of silica seem to be a good way of investigation
according to Escalante et al. (2000) and Kim et al. (2008).
5. Conclusion
The strengthening treatment applied in 1993 did not give any satisfaction if we
consider the evaluations done in 2006 and 2011. In order to find a durable and efficiency
consolidant, treatments tests have been performed taking into account the environment
(climate and marine environment), the nature of the cultural heritage objects (which do

not allow every kind of treatment application) and the degradation state of the granite.
They are also unsatisfactory because they do not increase the resistance of the stone in
front of accelerated ageing. The reason seems to be inherent to the employed technique
to apply the products. The consumption was too low and the products should penetrate
deeply to ensure a better efficiency. A quite low porosity and supposed weak capillary
properties lead to the difficulty to treat such kind of materials.
New experimental tests are therefore needed to know if it is possible to treat on site
I Stantari’s monoliths using products which would be efficient through time. Vacuum
impregnation could be a good way to test them even this treatment is difficult to perform
in such archaeological sites. In front of their apparent inadequacy over time, some other
strengthened products such as nano- consolidant products must be then studied.
Complementary tests and analyses such as experiment on samples containing
soluble salts, pores distribution study and spectrocolorimetry will be also perform before
and after climatic and saline spray experimental ageing. All the results will be able to
precise what the best conservation surveys are for each monolith.
Meanwhile obtaining convincing results, emergency operations have been done
locally, consolidating scales using Paraloid resin (Chatain, 2008, 2010).
6. Acknowledgements
We thank the Corsican Local Authority for their support. We also thank the
archaeologists Dr. A. d’Anna and Dr. P. Tramoni.who advertised the Corsican Authority
on the degradations and had a very constructive contribution helping us in our work.
7. References
Anna (d’), A., Guendon, J.-L., Pinet, L. and Tramoni, P. 2004. 'Les alignements de I
Stantari, Cauria, Sartène, Corse-du-Sud'. Fouilles programmées pluriannuelles –
Rapport 2002-2004.
Begonha, A. and Fojo, A. 2008. 'Application and efficiency of water repellents and
consolidating products on two two-mica granites used in the matriz de caminha
church'. In 11 th <strong>International</strong> congress on deterioration and conservation of stone,
Torun, Poland, volume II, 15-20 September 2008: 785-793.
Casta, L. 1988. 'Nature pétrologique des statues-menhirs de Corse, processus
d’altération et de dégradation'. In actes du colloque Les Statues-menhirs de Corse,
intérêt archéologique, problème de mise en valeur, Sartène 1988: 93-108.
Casta, L. 1993. 'Opération de sauvetage du site archéologique de Cauria Sartène-Corse
du Sud'. Rapport de fin de mission, Première tranche, CNRS, laboratoire de géologie
quaternaire, Marseille, 21 : 91.
Casta, L. and Poli, E. undated. 'Site archéologique D’I Stantari, Sartène-Cauria (Corse
du Sud). Opération de sauvetage du site'. Rapport final, CNRS, LAPMO-UA 164
CNRS, Aix-en-Provence : 27p.
Chatain, S. 2008. 'Site d’I Stantari-Sartène- Alignement de Statues- menhirs'. Rapport
d’intervention : 22 p.
Chatain, S. 2010. 'Site d’I Stantari-Sartène- Alignement de Statues- menhirs'. Rapport
d’intervention : 22 p.
Collective 2008. 'Glossaire illustré sur les formes d’altération de la pierre'. Monuments
et Sites XV, ICOMOS.

Didier, J. 1991. 'The main types of enclaves in the Hercynian granitoids of the Massif
Central, France'. In Didier J. et Barbarin B., Enclaves and granite petrology,
Developments in Petrology 13, p. 47-61, Le granite du Sidobre (Tarn), Aissaoui D.
et Perrier R, Mines et Carrières, vol. 76, août-sept. 1994 : 8.1-89 http://www.rochesornementales.com/Site-Sidobre/SIDOBRE.HTML
(2012-05-30).
Escalante, M. R., Valenza, J. and Scherer, G. W. 2000. 'Compatible consolidants from
particle- modified gels'. In Proc. 9 th <strong>International</strong> <strong>Congress</strong> on Deterioration and
Conservation of Stone, 2: 459-465.
Félix, C. 1995. 'Peut-on consolider les grès tendres du Plateau suisse avec du silicate
d’éthyle ?'. In ‘Conservation et restauration des biens culturels’, actes du Congrès
LCP, Montreux 24-29 septembre 1995 / Laboratoire de conservation de la pierre,
Département des matériaux, Ecole polytechnique fédérale de Lausanne ; ed. Renato
Pancella: 267-274.
Kim, E.K., Won, J., Kim, J.J., Kang, Y.S. and Kim, S.D. 2008. TEOS/GPTMS/Silica
nanoparticle solutions for conservation of Korean heritage stone'. In 11 th
<strong>International</strong> congress on deterioration and conservation of stone, Torun, Poland,
volume II, 15-20 September 2008: 915-923.
Rogalle, D. 2006. 'Études d’un produit de traitement de protection appliqué en 1993 sur
les statues-menhir d’I Stantari'. Rapport de stage de 1e année IUP Chimie analytique
et démarche qualité, Université de Poitiers: 24.
Rojo Alvarez, A., Mateos Redondo, F. and Valdeon Menendez, L. 2008. 'Consolidation
of granite building stones used in continental climates: San Juan de Los Reyes
Church in Toledo, Spain'. In proceedings <strong>International</strong> symposium stone
consolidation in cultural heritage, Lisbon, 6-7 May 2008: 213-222.
Sanmartin, P., De Los Santos, D.M., Rivas, T., Prieto, B., Mosquera, M.J. and Silva, B.
2008. 'Different behavior of TEOS-based consolidant applied to granite and
biocalcareous sandstone'. In proceedings <strong>International</strong> symposium stone
consolidation in cultural heritage, Lisbon, 6-7 May 2008: 409-417.
Sebastian, E., Cultrone, G., Benavente, D., Fernandez, L. L., Elert, K. and Rodriguez-
Navarro, C. 2008. ‘Swelling damage in clay-rich sandstones used in the church of
San Mateo in Tarifa (Spain)’. Journal of Cultural Heritage, 9: 66-76.
Vallet, J.-M., Bromblet, P. & Bouillon, N. 2012. 'Du diagnostic à la conservation du
Patrimoine archéologique en pierre en milieu isolé : quelle démarche adopter ?
Exemple des statues- menhirs en granite d’I Stantari'. In actes de la table ronde
'Conserver, étudier, protéger le patrimoine en milieu isolé', Mariana, 9-11 Octobre
2008: in press.
Wheeler, G. 2008. 'Alkoxysilanes and consolidation of stone: where are we now?'. In
Proceedings <strong>International</strong> symposium stone consolidation in cultural heritage,
Lisbon, 6-7 May 2008: 41-50.

CLIMATIC EXPOSURE AND PROTECTION OF THE RUNE STONES IN
JELLING, DENMARK
Poul Klenz Larsen 1 , Susanne Trudsø 1
1 The National Museum of Denmark, Department of Conservation, Brede, DK-2800
Lyngby. E-mail: poul.klenz.larsen@natmus.dk
Abstract
The rune stones in Jelling had been exposed to the natural climatic conditions for
centuries, and there was a growing concern for their state of preservation. A condition
survey concluded that there was an urgent need to protect the stones from further
degradation. A monitoring program was initiated in 2006 to determine the influence of
temperature and humidity and establish requirements for a climatic protection. Solar
heating to the surfaces was significant, but the temperature fluctuations were not likely
to cause damage. Frost in combination with wetness was identified as the most
important risk factor for sudden degradation. Rainfall was frequent and ample, due to
the spill from the treetop of a large lime nearby. Frequent episodes of condensation
occurred at the stone surface at all times of year. A wet surface enabled the growth of
moss and lichens, agents of slowly progressing decay. An open shelter would prevent
direct rain to the stones, but it would not keep the surface dry. The winner of a design
competition devised two separate, permanent enclosures made of glass and bronze. The
project was realized in the autumn of 2011. The first results of the climate monitoring
indicate that the climatic enclosure performs well.
Keywords: Rune stone, climate monitoring, condensation, frost damage, climatic
enclosure
1. Introduction
The rune stones in Jelling are among the most important monuments in Scandinavia.
Located next to the 11 th century church and two large barrows they are part of a
UNESCO World Heritage Site (fig 1). They were erected in the 10th century by Gorm
the Old and his son Harold Bluetooth, ancestors to the Danish royal family. Gorm’s
stone has the first known reference to the name ‘Denmark’ and Harold’s stone marks the
transition to Christianity. Gorm’s stone has text on two sides, whereas Harold’s stone
has both text and decorative reliefs on three sides. Harold’s stone was presumably
erected at its present position, whereas Gorm’s stone was moved next to Harold’s stone
around 1630.
Both stones were exposed to the natural climatic conditions for centuries. Several
visual inspections during the 20 th century identified an ongoing deterioration and there
was a growing concern for their state of preservation. A condition survey was
undertaken in 2006 - 2007 by the National Museum of Denmark and several external
experts (Trudsø, Jensen & Madsen, 2008). The investigation concluded that Gorm’s
stone was in poor condition and there was an urgent need to protect it from further

degradation. Harold’s stone was in a better state of preservation, but there was evidence
of progressing decay. The question was if the degradation was related to the climatic
exposure, and how to reduce environmental influence.
The use of microclimate monitoring to select appropriate protection of monuments
in situ is recommended by many authors. Thorn (1994) measured surface temperature
and wetness on sandstone tombstones in Australia, and found that thermal and hygric
fluctuations were a major cause of damage. Maekawa & Agnew (1996) conducted a
monitoring campaign at the Sphinx in Giza, Egypt. The decay of the limestone related to
large variations in RH, caused by solar heating and radiant cooling. In both cases an
open shelter would prevent the decay. Klenz Larsen (2008) measured several climate
parameters on the ruin of the Cathedral in Kírkjubøúr in the Faroe Islands and found that
salt accumulation increased if the walls were sheltered. The walls of basalt blocks would
be best preserved without protection. An open shelter does not always provide sufficient
climatic protection.
Figure 1. View from the south of the world heritage site in Jelling. The rune stones are located at
the south side of the church outside the main entrance, next to the large lime tree.
2. Environmental decay phenomena
Both stones have different types of decay related to climatic, chemical, biological or
human factors. Some minerals have dissolved from the gneiss rock over the centuries,
leading to gradual loss of substance from the stone surface. Water is the main agent for
this process, possibly enhanced by airborne pollutants or acid exudation from moss or
lichens. Mosses were present only in small areas, but have previously covered larger
parts of the stones, according to photos and written reports. This type of slowly
progressing decay has blurred, but not yet erased the carvings.
A variety of cracks and fissures are visible in most areas of the stone. Some clearly
predates the carving of the stones and relates to the gneiss geology, whereas some have

developed in recent time. Microscopic fissures arise from differential thermal and hygric
expansion or contraction, and relates to fast or large variations in temperature and
humidity. They give access for moisture below the gneiss surface, which is otherwise
dense and rather impermeable to water.
Loss of the surface by flaking has occurred at all faces on both stones. Some of the
flaking is possibly related to human intervention, but a substantial part is due to
exfoliation of the stone. Numerous areas with cavities below the surface were identified.
It is only a matter of time before these delaminated areas are irreversibly detached.
Exfoliation arises from the structure of the gneiss and may develop due to freeze-thaw
cycles.
3. Microclimate monitoring
The monitoring program for the Jelling project faced a serious challenge. It was
not permitted to install equipment in front of the stones, or to attach any sensors to the
stones surface. The stones had to be accessible for inspection and for the public. To
overcome this limitation, a metal rig for the instruments was erected over the large stone
(fig. 2). The rig consisted of four vertical posts, each approximately 3 m high. An
ordinary foundation was not allowed because of possible archaeological remains, so
each post was fixed in a cast concrete foot above ground. The four posts were connected
by four lintels to form a cubic geometry around the stone. Only Harold´s stone was
monitored, because Gorm’s stone was in such a poor condition, that it would have to be
protected in any case.
Figure. 2 View of the rune stones from the southwest. A rig for the instrumentation was raised
around Harold’s stone. Gorm’s stone is in the foreground.
The surface temperature on each of the three faces was monitored with infrared
sensors. The sensors were mounted on the lintels, pointing at an angle down at the stone
surface. The sensors detected the infrared radiation from a circular area with a diameter
of half the distance between surface and sensor. Due to the angled position of the sensor,

the actual field of temperature monitoring was elliptic with a major axis of
approximately one metre. The wetness sensor was mounted in a vertical position on the
south facing lintel. The air temperature and relative humidity was measured 3 m above
ground with the sensor mounted in a radiation shield. The rainfall was measured with a
tipping bucket rain gauge mounted on the rig at the southwest side of Harold’s stone. In
this position, the rain gauge would detect the influence of the tree nearby, and give a
realistic picture of the actual rainfall on Harold’s stone. All sensors were connected to a
single data logger, which was powered by a solar panel. Measurements were made every
2 minutes, and the hourly average was stored. Further information about the
instrumentation is given in the appendix.
4. Environmental exposure
The results of one year’s monitoring are shown in selected periods with one week of
data on separate charts. The temperature trend is represented with four curves,
corresponding to the three surface measurements and the air temperature. The first
diagram (fig. 3) shows a typical week in early summer, where the temperature ranged
10-15 °C over the day. The strongest surface temperature changes was during day 161
when the temperature on all three sides of the stone rose from 15 °C to 33 °C within 8
hours. At the same time the temperature of the ambient air increased only to 28 °C, but
this difference became apparent only in a few hours in the afternoon. Most of the time,
the stone had the same temperature as the air. There was a time lag in solar heating, with
the east and south face warmed 2-3 hours before the west side. But the west face
temperature rose above ambient long before the sun could possibly have hit this surface.
This is evidence that the entire stone heated up by the sun within an hour or two. This is
certainly possible because gneiss is a quite good heat conductor. The temperature
difference between the stone surface and the interior was therefore never more than 5 °C.
Differential temperature expansion is not likely to cause any damage in this situation.
Figure 3. The climate record for one week in the early summer (week 23) with daily solar heating

Figure 4 shows a week in midsummer with a different type of weather. The daily
temperature variation was less regular, indicating that the sun was not always shining.
There were several episodes where the wetness sensor gave a signal. On day 178 the
rain gauge had simultaneous records of rainfall. This was repeated again the next days, a
little more erratically. These incidents show that the wetness sensor responded quite
precisely to rainfall. The total precipitation recorded in one year was 2500 mm, which is
more than three times the rainfall on open land. The excess of moisture came from the
treetop of the large lime tree nearby, which acted as a canopy. This means that Harold’s
stone got much more water from above than Gorm’s stone. But rain was not the only
source of surface moisture on the stone.
Figure 4. The climate record for one week in the mid summer (week 26) with much precipitation
Figure 5 shows a week later in summer with consecutive events of wetness, even
though there is no rainfall at all. The episodes coincided with the situation that the
stone's surface temperature met the dew point (black curve). The cause of wetness was
condensation formed when the temperature dropped so much that the RH was 100%.
The condensation primarily occurred at night, so this phenomenon had little visual
impact. But the condensation was a source of moisture for moss or lichens in dry
periods, and could promote microbial growth on the stone surface. In one year, the total
time of surface wetness was 4000 hours. There were 1000 hours of precipitation,
whereas condensation occurred in 3000 hours. It is evident that condensation was the
main cause of wetness to the stone.
Figure 6 shows a week in late winter with much precipitation. On day 52 and 53
there was constant rainfall for almost 24 hours. As the rain stopped, the temperature
dropped from around zero to -2 °C. The surface was soaking wet at this point, so a
period of frost for two days could cause damage to any area with entrapped water.
Freeze-thaw cycles are a powerful decay process, which can disrupt the gneiss structure,

if water is contained in cracks or voids. Crack formation can be aggravated by the
formation of ice, or it may rupture pieces of the surface. The prerequisite for this to
happen, is that there actually is water present in the areas that are prone to freezing.
Figure 5. The climate record for one week in late in summer (week 32) with much condensation
Figure 6. The climate record for one week in late winter (week 8) with a freeze-thaw cycle
During day 54 there was a steady but small record of precipitation, which was
probably snowfall. This may explain why the wetness sensor did not respond. Snow
itself was not a problem, but any accumulation of snow or ice on the stone surface could
slowly deliver water that seeps into cracks and then freezes. This was only harmful if the
temperature actually went below zero in the depth where water may collect, but this

cannot be determined by the measurements. Based on the results of solar heating, it is
assumed that it took 1-2 hours for the stone to reach temperature equilibrium. During the
winter of 2006-2007 there was a total of 18 freeze-thaw cycles. Some cycles took a few
hours whereas some lasted a full day and reached below -5 ° C. Most of the freeze-thaw
cycles were therefore considered to be a major risk factor.
5. Climatic protection
There were three possible solutions for climatic protection. To move the stones
inside the main hall of the adjacent museum building, to install temporary protective
shelters in winter or to erect a permanent enclosure for climatic control. The winner of a
design competition devised two separate enclosures made of glass and bronze. The
project was realized in the autumn of 2011(fig. 7). The encapsulations have moderate
heating in winter to avoid frost and the relative humidity is controlled by
dehumidification. The two separate enclosures were established in the autumn 2011, and
the climate has been monitored both inside and outside since January 2012.
Figure 7. The rune stones are now protected against climatic exposure in two separate display
cases made of glass and bronze. The climate is controlled by heating and dehumidification.
(Design by Erik Nobel Architects)
The effect of the climatic protection is shown in fig. 8. The diagram displays the
temperature and dew point inside the enclosure over 6 weeks in February and March
2012. The temperature never dropped below 5 °C, so there was obviously no risk of
frost damage to the stone. The dew point was well below the surface temperature, which
excluded any risk of condensation at the stone surface. The daily rise in temperature was
due to solar heating to the enclosure. The course of such a sunny day is displayed in
figure 9. The temperature rise started from 5 °C at nine in the morning and continued
steadily until four in the afternoon, when the temperature reached 16 °C. The stone´s
surface temperature followed the temperature rise with a little delay, but the decline was

slower due to the large thermal capacity of the gneiss. Such temperature variations are
not likely to cause significant thermal differential expansion. The conditions may be
different in summer. The dew point inside the enclosure is lower than outside most of
the time due to the effect of dehumidification.
Figure 8. The climate record for six weeks in February and March 2012 showing the effect of the
climatic controlled enclosure of Harold’s stone. The temperature was always above 5 °C and the
dew point was lower than the surface temperature.
Figure 9. The climate record during 24 hours on 6 March 2012. The difference in dew point
temperature was due to the effect of dehumidification. Heating at night and solar radiation in
daytime controlled the stone’s surface temperature.

6. Conclusion
1) The surface temperature was close to air temperature most of the time. When
exposed to solar radiation, the surface temperature of the unprotected stone rose 4-6 °
above ambient. The heat diffusion through the stone was fast. Thermal differential
expansion was not considered to be a main cause of decay.
2) There were 18 recorded freeze-thaw cycles during winter 2006-2007. Some
were preceded by episodes of wetness due to rain or condensation. This combination
could cause severe damage due to ice formation in fissures or cavities. The combination
of frost and surface wetness was possibly the main cause of surface deterioration.
3) The stone surface was only in very few cases colder than the surrounding air.
Frost loads or condensation resulting from thermal radiation at night was therefore
minimal.
4) Surface wetness was recorded in roughly half the time evenly distributed over
all seasons. Precipitation was responsible for the surface wetness at approximately 1/3
of the time, and the rest was due to condensation or fog. Condensation occurred mainly
during the night or morning.
5) An open shelter would not protect the stone from frost and only partially
against surface wetness. A shelter may increase deposition of airborne contamination,
since the surface was not washed clean by rain. The rune stones would be best protected
in a controlled environment to avoid subzero temperatures and condensation or
precipitation. The first climate monitoring results confirm that the enclosures fulfill their
purpose.
7. Acknowledgements
The condition survey was supported by the Velux Foundation. Lars Aasbjerg Jensen
helped installing the equipment for climate monitoring.
8. References
Larsen, P. K. 2008 ‘Weathering at the Cathedral in Kirkjubøur, The Faroe Islands’. In
Proceeding of the conference Salt Weathering on Buildings and Stone Sculpture, in
Copenhagen 22-24 October 2008, pp 249-256
Maekawa, S. & Agnew, H.H. 1996 ’Investigation of environmentally driven
deterioration of The Great Sphinx and concepts for protection’. In Archeaological
conservation and its consequences. Preprints of the IIC conference in Copenhagen 26 –
30 August 1996, pp 116-120
Thorn, A. 1994 ‘Direct measurements to determine sandstone deterioration’. In
Preprints of the IIC <strong>Congress</strong> in Ottowa, 12-16 September 1994, pp 85-89
Trudsø, S., Jensen, K.S., Madsen, P. K. 2008 ‘How bad is it – and why. Study of the
preservation of the Catheads Door in Ribe Cathedral and the Rune Stones in Jelling’. In
Nationalmuseets Arbejdsmark 2008, pp. 239 – 256. (English summery)

9. Appendix
Instrumentation
Parameter Instrument Sensor Accuracy
Surface temperature IRTS-P thermocouple Voltage -
Surface wetness 237 Resistance -
Rainfall CS 52202 Tipping bucket +/- 1%
Air temperature Vaisala HMP45C PT 100 +/- 0,3 (0-20°C)
Relative humidity Vaisala HMP45C HUMICAP 180 +/- 2% (0-90%RF)
Datalogger CR10X -
Power supply PS100E-LA (12 V) -
Solar panel SOP10 (10W) -
Supplier: Campbell Scientific Ltd.

THE NEAPOLITAN YELLOW TUFF AND THE VICENZA STONE:
EXPERIMENTAL INVESTIGATIONS ABOUT EFFECTIVENESS OF
ANTISWELLING TREATMENT.
Claudia Di Benedetto 1 , Sara Bianchin 2 , Piergiulio Cappelletti 1 , Abner Colella 1 , Maurizio
de Gennaro 1 ,Monica Favaro 2 , Arianna Gambirasi 2 , Alessio Langella 3 , Giovanni Luca 4 ,
Maria Soranzo 4
1 Dipartimento di Scienze della Terra, Università Federico II, Via Mezzocannone 8,
Napoli, Italy.
e-mail: claudia.dibenedetto@unina.it
2 Istituto di Chimica Inorganica e delle Superfici, C.N.R., Corso Stati Uniti 4, Padova,
Italy.
3 Facoltà di Scienze, Università del Sannio, Via Port’Arsa 11, Benevento, Italy.
4 FILA Industria Chimica Spa, Via Garibaldi 32, San Martino di Lupari (Pd), Italy.
Abstract
In this study, two relevant geomaterials, traditionally employed as building stones,
the Neapolitan Yellow Tuff (a volcanoclastic rock; NYT) and the Vicenza Stone (a
porous sedimentary rock; VS), were consolidated under laboratory conditions. Although
with different genesis and composition, the two stones display similar pore radii
distribution and this allows to better evaluate the mechanisms of stone consolidation.
Goal of this research is to test the performance of a consolidation with Ethylsilicate
based product (PRC100; ES) on these stone materials previously treated with an
antiswelling commercial products (Antihygro; AH); the use of AH is intended to reduce
the swelling ability of some minerals contained in NYT and VS.
Keywords: Neapolitan Yellow Tuff, Vicenza Stone, consolidation, durability,
antiswelling.
1. Introduction
The use of the stone as building and decoration material has ancient origin and
Italy offers several examples of stone used over the centuries in sculpture, decoration,
road paving, etc.
Building stone are often affected by deterioration caused by weathering which
sometimes causes irreparable damage.
During the past intervention, replacing damaged blocks was a cheap solution, but
nowadays it’s almost difficult to due to the closure of many quarries.
In the recent years many efforts have been made to search products able to
consolidate and stabilize weathered material and to reduce stone decay, protecting it
from the causes of degradation.
The choice of a conservative intervention is a very risky phase that requires a
thorough knowledge of the stone.

In this study a Campanian tuff (NYT) and a biodetrital carbonate (VS) were treated
with conservative products, an ethyl silicate and an antiswelling protective, to evaluate
their effectiveness against weathering mechanism.
NYT was used as a building material since ancient times because of its peculiar
colour, light weight, easy workability and good insulating property.
VS was used as building stone mostly in Vicenza, Verona and Padova, with
architectural and decorative function. Both the stones display high porosity that makes
them extremely vulnerable to deterioration caused by weathering.
The effectiveness of conservative treatments was evaluated by mineralogical and
physical-mechanical investigation and by accelerated ageing test.
2. Materials
NYT is the most important volcanic product in the Neapolitan area, linked to a
phreatomagmatic eruption in the area of Campi Flegrei (15,000BP, Deino et al., 2004).
It is characterized by an altered ashy matrix containing zeolites and very subordinate
clay minerals (smectite); the most abundant zeolitic phase is phillipsite, along with
chabazite and analcime in minor amounts (de’ Gennaro et al., 2000). Primary
phenocrysts are represented by sanidine, pyroxene, mica, plagioclase and magnetite (de’
Gennaro & Langella, 1996; de’ Gennaro et al., 2000). Its chemical composition ranges
from trachyte and alkali-trachyte to phonolite (Scarpati et al., 1993; Orsi et al., 1995).
The VS is a biodetrital carbonate formed at back-reef in the tidal channels during
Oligocene (33M y.b.p ; Mietto, 1988).
It is a limestone characterized by high calcium carbonate content (90-95 per cent)
and low content of both silica and clay components and with non-negligible percentage
of aluminum oxides and iron (Marchesini et al., 1972; Cattaneo et al., 1976).
The main biogenic components are foraminifera such as Amphistegine and
Lepidocicline, Briozoi, red algae Melobesoidea, corallinae algae.
Frequently, in the case of monumental buildings, the ornamental stones were used
facciavista and this makes them more sensitive to weathering processes. The basic agent
of this phenomena is mainly represented by water that plays a decisive role in the
weathering of stone surface: its action is manifested both directly (through hydrolysis
processes in silicates and dissolution in carbonates), and indirectly (through processes of
hydration-dehydration,freeze-thaw cycles, crystallization of salts).
The most frequent weathering typologies of NYT are: alveolization, mainly due to
detachment of particles such as lithic clasts or zeolitized pumices; scaling and
exfoliation both due to the combined action of damp waters and subsequent
recrystallization of soluble salts; disaggregation due to infiltration of water and the
consequent dissolution of the constituent phases of the stone; patinæ, stains,
efflorescences due to the evaporation of damp waters and consequent deposition of
soluble salts on stone surface.
It should be noted that most of these forms such as scaling and exfoliation are
promoted by the presence of minerals with swelling properties such as zeolites and/or
clay.
The weathering forms visible on the surfaces in VS are mainly due to dissolution
and pulverization processes and biological activity; humidity traces, and black crusts are
often present.

2.1 Products
Two commercial products have been used both with NYT and VS: a silicate-based
consolidant (ES) and a protective having antiswelling properties, which was applied
before ethyl-silicate (AES; table 1)
Table 1. Characteristics of tested products
Product Company Composition Abbreviation
Density
(gr/cc)
PRC110 Fila Etylsilicate ES 1.060
Antihygro Remmers
Tetramethylenediammonium
dichloride
AH 1.000
The first one (ES) is a one component-system based on organic compound of
silicon, produced marked by Fila Industria Chimica Spa (S. Martino di Lupari –
Padova). The consolidation effect is primary due to deep penetration into the pores and
furthermore its ability to react with the silicate matrix (as in the case NYT).
AH is a tetramethylenediammonium dichloride in aqueous solution. Its property is
to reduce the hygroscopic swelling of stone, leaving unchanged their physicalmechanical
characteristics.
3. Methods
The NYT was collected from Chiaiano quarry, in the northwestern part of Naples.
The VS was quarried in Nanto (VI), located on the Colli Berici and was provided
to us by Marmi Sgambaro SNC company (S. Martino dei Lupari-PD).
Consolidation has been achieved by total immersion. The specimens were divided
into the following test groups: untreated, treated with consolidant (ES) and treated with
antihygro and consolidant (AES). The samples treated with AH were air-dried for one
week, those treated with ES for three weeks (in agreement with its polymerization
time).
Laboratory tests (mineralogical and petrographic analyses, physic-mechanical
determinations and accelerated ageing test) were carried out before and after application
of each treatment according to European-suggested standards (UNI EN, NORMAL).
3.1 Mineralogical investigation
Mineralogical composition was carried out by X-ray powder diffraction (XRPD -
Panalytical X’Pert Pro; CuKα radiation, 40kV, 40mA, 4-80º 2θ scanning interval, 0.017º
equivalent step size, 60 sec per ste equivalent counting time, RTMS X’Celerator
detector), optical microscopy (Leica Wild MZ 8) and SEM observations (Jeol JSM 5310
Oxford Inca- CISAG, Federico II University of Naples).
Quantitative mineralogical analyses were performed by XRPD using "Reference
Intensity Ratio” (RIR - Chipera e Bish, 1995) technique. Powders with grain size less
than 10 µm were obtained using a McCrone micronizer mill; such a particle size allows
several problems to be overcome, such as: particle statistics, primary extinction,
microadsorption and preferred orientation. (Klug & Alexander, 1974)

The EDX analysis (FEI Quanta 200 scanning electron microscope) were performed
to evaluate the penetration depth of consolidating products into the stone by measuring
the concentration of silicon in untreated and treated samples and were carried out at the
ICIS Istituto di Chimica Inorganica delle Superfici di Padova.
3.2 Physical-mechanical investigation
The physical-mechanical investigation was carried out by swelling test, thermal
expansion and ageing tests.
The influence of water on the behavior of both untreated and treated stones was
evaluated by swelling tests following the procedures suggested by Nascimento et al.
(1968); particularly the swelling strain was measured on cubic specimens (side of 5 cm)
using an apparatus suitably realized by Lonos Test S.r.l. (Monza) and able to measure a
volumetric swelling strain by five digital micrometer arranged along the three axes
(x,y,z).
The thermal beahviour was carried out on prismatic specimens (25 cm x 5 cm x 2,5
cm); according to the NORMAL UNI EN 14581 the tests were performed using a
mechanical device which allows to record variation in length of specimen with
increasing temperature.
Ageing tests included salt crystallization tests, freeze-thaw and salt spray test. All
these tests were performed according to standard procedures (UNI EN, 12370 UNI EN
12371 and UNI EN 14147 respectively).The tests were chosen taking into account the
weather conditions of the areas where the two stones are used. In order to assess the
change of physical-mechanical features, weight, open porosity, ultrasonic wave and
UCS were measured before and after treatment.
Open porosity was calculated by means of apparent and real volume with an He
pycnometer (Micromeritics Multivolume Pycnometer 1305) on cylindrical specimens.
Ultrasonic tests were carried out according to NORMAL 22/86, using an
Ultrasonic Concrete System RCL 34/200 with a pair of 54 kHz transducers, in direct
arrangement.
Uniaxial Compressive Strength test were performed according to UNI EN 1926
(Controls C5600), with maximum load of 3000 kN and a load constant rate of 1 ± 0.5
MPa).
Freeze-thaw cycles were carried with a Binder MK 53 climatic chamber. Saturated
specimens were placed in climatic chamber for 6 hours at a temperature ranging from
20º to -12º C and then immersed in water at room temperature for 6 hours.
Salt crystallization cycles were carried out by immersing cubic specimens (side 4 cm) in
a saturated solution of sodium sulfate dechaydrate for 2 hours. Specimens were then
dried in oven at 105 ± 5 ° C at least 16 hours, cooled for about 2 hours and re-immersed
in a fresh solution. The cycle is repeated for 15 times.
Salt spray test was carried out at the Fila Industria Chimica Spa (S. Martino di Lupari -
Padova). It was performed on cubic specimens (side 5 cm) by an ERICHSEN
Corrotherm 610 climatic chamber (400 litres), capable to alternate cycles of treatment
with salt spray (sodium chloride) and drying. The test consists of 60 cycles.
4. Results and discussion
NYT is characterized by altered ashy matrix containing zeolites and subordinate
clay minerals; the authigenic phase are phillipsite (55 per cent), chabazite (3 per cent),

analcime (4 per cent) and smectite (4 per cent).The pyrogenic minerals are represented
by sanidine (23 per cent), pyroxene (2 per cent), mica (traces), hydrated iron oxides
(traces). The volcanic glass and/or the amorphous component is roughly 11 per cent.
The diffraction analysis performed on the VS show calcite as prevailing phase (98
per cent) and subordinately quartz (1 per cent), clay mineral (1 per cent) and iron oxides
(traces).
The penetration capability of AH into the pores is more difficult to detect by SEM
or EDS techniques, mainly due to low chlorine content. On the contrary the
effectiveness of ES treatment was highlighted by means of EDS analyses of the Si
concentration along the cross section of treated and untreated specimens (side 5 cm;
figure 1). First of all, it is notable the deep penetration of the ES in the entire section of
the specimens for both lithotypes. Furthermore the pretreatment with AH does not seem
to influence the ES penetration except for NYT (figure 1, on the left) in which is clearly
observed a lowering of the silica percentage, in the range from about 200 to 300 mm,
close to the untreated samples.
Figure 1 Profiles of the Si concentration (in atomic percent) along the cross section of NYT (left)
and VS (right) specimens both untreated and treated.
Expansion or shrinkage as a result of water imbibition or temperature variations
play an important role in the stone deterioration especially for tuffs (Steindlberger,
2004). This parameter are strongly influenced by the mineralogical composition of the
stone and their evaluation can provide significant informations in all those contexts
characterized by constant presence of water or major climatic excursions, both daily and
annual. The expansion coefficient is negligible when the stone can expand; on the
contrary, when even expansion is prevented, tensions that are created within the rock
can go beyond the value of resistance to compression, or be likely to cause flexing and
bending of the buildings (Primavori, 1999).
Changes in the dimension of the specimens due to rock swelling after water
immersion, were measured (table 2). The hygric expansion of the untreated stone mainly
depends on the content of phases with swelling capability, such as clay and zeolites.
This explain the greater tendency to dilatation of the NYT compared to VS.

A strong decrease of expansion (more than 60 per cent) was recorded for NYT
after pretreatment with AH (figure 2). A strong decrease of expansion (more than 60 per
cent) was recorded for NYT after pretreatment with AH (figure 2).
Table 2. Comparison of some physical-mechanic parameters for NYT and VS before and after
conservative treatments (means value).
NYT VS
Untreated AES ES Untreated AES ES
Samples number 5 5 5 5 5 5
Amount of applied
product (%)
― 20 18 ― 5 6
Hygric dilatation (%) 0.649 0.199 0.417 0.033 0.086 0.086
Linear coefficient of
thermal dilatation
(°C -1 )
- 26 - 12 - 11 3 2 2
Figure 2 Volumetric dilatation of NYT and VS, both untreated and treated.
The NYT thermal behavior with increasing of temperature is characterised, in
accordance with literature (Marino et al., 1991; Colella et al., 2009), by a dimensional
contraction due to dehydration of the zeolitic cement (table 2).
The linear coefficient of thermal expansion of untreated NYT is about -0.26ºC -1 ;
the consolidated samples show a lower contraction, which is about -0.12 ºC -1 and -0.11
ºC -1 , for AES and ES samples, respectively.
In contrast, the VS shows an average coefficient of thermal expansion of 3 ºC -1 ;
This dilatation is related to the response of the calcite to the increase in temperature, the
crystal shrinks in one direction and expands in the direction orthogonal to the previous
(Rota Rossi-Doria, 1987). Both AES and ES effectively contrast VS dilatation with
lower values by 33 per cent (table 2).

In order to evaluate durability of consolidated stones is necessary to carry out test
able to reproduce the environmental condition in which the stone is used.
Therefore stones have been tested by ageing test such as salt crystallization, freezethaw
cycles and salt spray.
The variation of the physical-mechanical properties was reported in table 3.
Salt crystallization is one of the most powerful weathering agents for the stone.
This test resulted very aggressive for NYT for both untreated and consolidated samples
(AES and ES). This behavior is probably due to decrease in pore spaces that resulted in
a more effective pressure of crystallization. Consolidated samples are broken down
completely before the end of test with the same times of the untreated (table 3).
Different behavior is observed for the VS: the salt crystallization cause a
considerable loss of material for the untreated samples. On the contrary, all consolidated
samples show good resistance to weathering induced by repeated salt crystallization
cycles, as evidenced by the low weight loss (table 3).
Table 3 Variation of the main petrophysical properties in NYT and VS (untreated and treated)
after ageing tests.
NYT VS
% sc ft ss sc ft ss
weight loss
compressive
strength
P-wave
porosity
Unt disag. -2 -2 -73 -6 -1
AES disag. -4 -4 -1 -1 -2
ES disag disag. -4 -1 -1 -2
Unt disag. not mes.
AES -17 0,5
ES disag -7
Unt -54 -15 not mes. +2
AES -1.6 -6 -27 +10
ES disag 12 -26 +10
Unt disag 4 3 2
AES disag 13 4 13
ES disag 17 9 30
sc= salt crystallization; ft= freeze-thaw cycle; ss= salt spray; disag=disaggregated; not mes= not
measurable.
The behavior of tuff during freeze-thaw cycles is changed by pretreatment with
AH. The unconsolidated NYT shows a strong degradation with intense fracturation in
the early cycles. The stronger degradation is showed by NYT treated with ES: in this
case, in fact, the freeze-thaw cycles lead up to disaggregation before the end of the test.
NYT treated with AES, on the contrary, does not show, except for a single test
specimen, fractures at end of cycles.

The behavior of NYT to the frost action is closely related to the kinetics of water
absorption: fast saturation of porous system causes a more effective pressure into the
stone when water pass to solid state.
As regards VS, freeze-thaw cycles mainly act on some portions, brownish in color,
clearly distinguishable and well localized.
The EDS analysis performed on these portions indicate the presence in addition of
calcite, quartz, iron oxides and a micaceous component.
As a result of water absorption this portions swell creating pressure within stone.
This phenomenon leads to physical degradation with loss of material due to
pulverization.
The treated samples (both with AES and ES) do not show any weathering forms as
confirmed by negligible weight loss (1 per cent)
Finally to evaluate the resistance to weathering that affected the building stone, in
case of aggressive atmosphere, such as those typical of coastal areas, samples have been
exposed to salt spray chamber.
Salt spray test preserved the original shape of the specimens both for the NYT and
VS, as shown by the slight loss in weight (table 3).
As general remarks ageing tests produced an increase of total porosity and a
reduction of the compressive strength either in untreated as treated stones; in other
words, the durability was considerably reduced (Winkler, 1997; Goudie, 1999) and this
is probably true for NYT, regardless of treatment method.
Moreover, alteration mechanism such as salt crystallization and freeze-thaw
resulted more aggressive in stones with a high percentage of mesopores (r < 1 μm; Rossi
Manaresi, 1976).
5. Conclusion
The collected data represent a contribution to understanding of material behavior
after conservative treatments in conditions close to those in which material will be used
as building materials.
The choice of NYT and VS, silicatic and carbonatic in composition, respectively,
allows to evaluating how the different nature of constituents influence the consolidation
mechanisms.
Behaviour of NYT and VS during the thermal stress is affected by consolidation
especially when the antiswelling is applied.
The consolidation with ethylsilicate showed poor compatibility with the NYT; by
contrast, the consolidation was effective for VS especially in terms of durability.
A pretreatment with antiswelling was very effective for NYT, due to its high
percentage of expanding phases. The VS did not appear to be influenced by this specific
treatment.
The results indicate that the consolidation treatments for the NYT need a new
approach. In this light, the good petrophysical properties shown by the TGVT (Tufo
Giallo della Via Tiberina; PRIN 2008-2011), probably due to the presence of a calcitic
cement, suggest that the future use of inorganic products (possibly inducing the
formation of a similar cement) could improve the resistance to degradation of zeolitized
lithotypes.

6. References
Bish D.L., Chipera S. J., 1988. Problems and solution in quantitative analysis of
complex mixture by X ray powred diffraction, Adv. X ray Anal., 31, 295-307.
de’ Gennaro M., Cappelletti P., Langella A., Perrotta A. et al. 2000. Genesis of zeolites
in the Neapolitan Yellow Tuff: geological, volcanological and mineralogical
evidence. Contrib Mineral Petrol. 139:17-35.
Cattaneo A., De Vecchi G.P., Menegazzo Vitturi L., 1976. Le pietre tenere dei Colli
Berici. Soc. cooperativa tip., Padova.
Colella A., Calcaterra D., Cappelletti P. et al. 2009. I tufi zeolitizzati nell'architettura
della Campania. In: Cuzzolin, M., (Ed.), La diagnostica per il restauro del
patrimonio culturale, Napoli, pp. 327-341.
Deino, A.L., Orsi, G., de Vita, S., Piochi, M., 2004. The age of the Neapolitan Yellow
Tuff caldera-forming eruption (Campi Flegrei caldera – Italy) assessed by 40 Ar/ 39 Ar
dating method. J. Volcanol. Geotherm. Res. 133, 157-170.
Klug H.P., Alexander L.E., 1974. X-ray Diffraction Procedures for Polycrystalline
and Amorphous Materials, 2. Auflage. John Wiley & Sons, New York-Sydney-
Toronto 1974, 966 Seiten.
Marchesini B., Biscontin G., Frascati S., 1972. Alterazione delle pietre tenere dei colli
Berici. Atti XXVI <strong>Congress</strong>o A.T.I., Roma, pp 1-23.
Marino O., Mascolo G., Cioffi R. et al. 1991. Tufi vulcanici: meccanismi di
deterioramento chimico-fisico e tipologia di intervento. L’Edilizia, pp. 523-535.
Nascimento U., Oliveira R., Graca R., 1968. Rock swelling test. Proc. Int Symp. On
Determination of the properties of rock masses in foundations and observations of
their beahviour, Editorial Blume, Madrid-Barcelona, 363-365.
NORMAL UNI EN 12370, 2001. Metodi di prova per pietre naturali: determinazione
della resistenza alla cristallizzazione dei sali. Ente Nazionale Italiano di Unificazione,
Milano.
NORMAL UNI EN 14579, 2005. Metodi di prova per pietre naturali: determinazione
della velocità di propagazione del suono. Ente Nazionale Italiano di Unificazione,
Milano.
NORMAL UNI EN 12371, 2003. Metodi di prova per pietre naturali: determinazione
della resistenza al gelo. Ente Nazionale Italiano di Unificazione, Milano.
NORMAL UNI EN 14581, 2005. Metodi di prova per pietre naturali: determinazione
del coefficiente di dilatazione lineare termica. Ente Nazionale Italiano di
Unificazione, Milano.
NORMAL UNI EN 1926, 2000. Metodi di prova per pietre naturali: determinazione
della resistenza a compressione. Ente Nazionale Italiano di Unificazione, Milano.
Orsi G., Civetta L., D’Antonio M. et al. 1995. Step-filling and development of a three
layer magma chamber: the NYT case history. Journal of Volcanology and
Geothermal Research 67, 291-312.
Primavori P., 1999. Pianeta pietra. Giorgio Zusi Editore S.A.S, Verona.
Rota Rossi-Doria P., 1987. Il problema della porosità in rapporto al degrado ed alla
conservazione dei materiali lapidei. In: Materiali lapidei : problemi relativi allo
studio del degrado e della conservazione / Ministero per i Beni Culturali e
Ambientali, Ufficio Centrale per i Beni Ambientali, Architettonici, Archeologici,
Artistici e Storici. Bollettino d'arte ; 41.1987 : Supplemento. pp 11-14.e

Scarpati C., Cole P., Perrotta A. 1993. The Neapolitan Yellow Tuff - A large volume
multiphase eruption from Campi Flegrei, Southern Italy: Bull. Volcanol. 55, 343-
356.
Steindlberger E., 2004. Volcanic tuff from Hesse (Germany) and their weathering
behavior. Special issue: stone decay hazards.Environ Geol 46:378-390.
UNI EN 14147 Metodi di prova per pietre naturali: determinazione della resistenza
all'invecchiamento mediante nebbia salina.

BIO-POLYMERS AS STONE PROTECTIVES
Barbara Sacchi 1 , Emma Cantisani 1 , Giulia Giuntoli 1,2 , Silvia Salvini 1 , Costanza
Scopetani 1 , Luca Rosi 2 , Marco Frediani 2 , Piero Frediani 2
1 CNR-ICVBC Institute for the Conservation and Valorization of Cultural Heritage,
Sesto Fiorentino (Florence), Italy
2 Department of Chemistry, University of Florence, Sesto Fiorentino (Florence), Italy
Abstract
In the last decades, natural organic products have been substituted by synthetic
products in the field of Conservation of Cultural Heritage. Unfortunately synthesized
products have sometimes some drawbacks, due to the fact that they have high costs of
synthesis, or they are soluble in toxic solvents for man or environment, or they become
irreversible once applied on the substrate. As a consequence, it appears clear that the use
of new polymers derived from natural sources as protectives for lapideous materials
would be very welcome.
Lactic acid is produced starting from 100% annually renewable resources.
Moreover, under appropriate conditions, polymers of lactic acid can be completely
biodegraded. They are widely used in packaging and medical devices, and in this work
they were proposed as stone protectives. They were tested on samples of two Apuan
Marbles having different conservation state: one of them was specifically quarried for
this experimentation, while the other one was obtained by an ancient front quarry,
probably used in Roman Age.
Samples of the two marbles were characterised and treated with an homopolymer
of lactic acid and two co-polymers between lactic acid and Fluorolink D-10H, a low
weight perfluoropolyether, with the aim to improve hydrorepellence properties.
Treated materials were subjected to artificial ageing, both with thermohygrometric
cycles and UV exposure.
Colour changes and protective efficacy were evaluated after treatment and
monitored during the artificial ageing cycles.
Keywords: poly(lactic acid), renewable resource, protective treatment, marble
1. Introduction
In the last years, several treatments have been used to protect and/or consolidate
deteriorated stones, using both inorganic and organic products (Dohene and Price 2010;
Borgioli 2002; Amoroso 2002). Among the latter ones, polymer-based materials
obtained from animals or plants were substituted in course of time by synthetic
polymers, mostly petrochemical-based, which have unfortunately some disadvantages,
like high costs of production, often high toxicity for man or environment, and so on.
Nowadays, in the approach of the green chemistry, polymers from renewable
sources would be very attracting potential substitutes petrochemical-based materials,
even in the field of Conservation Science like in many other.

Among these compounds, poly(lactic acid) (PLA) has raised particular attention.
The starting material, lactic acid, is produced using 100% annually renewable sources
and the polymer is biodegradable and compostable (Stevens 2002). Furthermore
structure and molecular weight of the polymer can be easily controlled and modified
through the synthesis to obtain tailored products. PLA is widely employed in packaging
and biomedical devices (Ljungberg and Wesslen 2005), and it may also represent an
appealing alternative to synthetic polymers commonly employed in Conservation of
Cultural Heritage. Recent studies showed indeed that PLA may have a good long-term
photo-stability, making it particularly appealing for Conservation applications.
In this study PLA and fluorinated PLA (in order to enhance the water repellence
effect) were tested on two different marbles. Stone properties were tested before and
after application by means of water absorption and colour variation measurements, in
order to evaluate the performance of the polymers as protective products. Moreover,
treated stones were subjected to accelerated ageing in Climatic Chamber and in Solar
Box, with the aim to verify the durability of such treatments behind thermo-hygrometric
conditions variations and UV irradiation. During ageing cycles, stone properties were
tested at scheduled intervals.
2. Experimental procedures
2.1 Stone materials and products
The stone samples used for testing PLA and modified PLA were two calcitic marbles
characterised by different total open porosity and water accessible porosity. The first,
named MVB, is a medium-fine grained (150-200 μm) white marble with many gray
veins. The marble become from an ancient front of Gioia quarry (Carrara, Tuscany,
Italy), from a block probably extracted in the Roman age. This lithotype is characterised
by a total open porosity of 2.8% and a water accessible porosity of 1.1%. The second
one, named MG, is a white commercial calcitic marble, medium-fine grained, with total
open porosity < 1% and water accessible porosity < 0.5%.
The samples were cut into 5 x 5 x 2 cm specimens.
Three different polymers of lactic acid were studied and tested: an homopolymer of
LL-lactide (named PLLA), and two A-B-A block copolymers between lactic acid (A)
and Fluorolink D-10H (FLK) (B), that is a low weight perfluoropolyether. One of them
was a copolymer of LL-lactide and FLK (named PLLA-FLK), and the other was a
copolymer of LD-lactide, that is the rac-lactide, and FLK (named PDLA-FLK). All
polymers were synthesized through ROP (Ring Opening Polymerization) technique, that
permits a good control of structure and molecular weight of obtained products. The Tg
(glass transition temperature) of the products were determined through DSC
measurement technique, while their Mw (molecular weight) were evaluated through
GPC technique.
GPC (Gel Permeation Chromatography) was performed using a GPC Waters system
equipped with a pump Waters model Binary HPLC 1525, a refractive-index detector
Waters model 2414 and three columns Shodex KF-803 (length: 300 mm; diameter: 8.0
mm). Analysis was performed at 30°C using chloroform as eluent, with a flow rate of
1.0 mL/min. Chloroform solutions (1 mg of polymer in 1 mL of solvent) were injected.
Weight-average molar mass (Mw)/retention time was calibrated against polystyrene
standards.

DSC (Differential Scanning Calorimetry) was performed with a Perkin Elmer
instruments Pyris 1 DSC model equipped with an Intracooler 2P cryogenic system. A
heating rate of 20°C/min was used. Traces were recorded in the temperature range from
0 to 200°C under a nitrogen atmosphere. To eliminate any effect of thermal history,
measurements were made from a second heating cycle, after heating the sample to
200°C at 20 °C/min, followed by quenching to 0°C.
The stability of the polymers was studied in previous works, and they proved an
interesting ageing behaviour, with a very slight decrease of molecular weight after 750
hours of UV irradiation and FT-IR spectra generally unchanged up to 1,000 hours of
irradiation (Giuntoli et al. 2012a, 2012b).
2.2 Treatment procedure and performance evaluation
The products were applied on samples of the two marbles. Two ml of a 13% (w/w)
solution of the polymer in chloroform were applied on one of the 5 x5 cm 2 surface of
each sample. The amount of product applied for each treatment was determined by
weighing the specimens before and after the treatment (after complete evaporation of the
solvent). The amount of product is expressed as the weight of product per surface unit
(g/m 2 ).
The treatments were evaluated by measuring colour variations and capillarity water
absorptions.
Colour variation measurements were performed with a portable instrument
MINOLTA Mod. Chromameter CR200 comparing the colour of the samples before and
after treatment, according to the method adopted by the UNI-EN commission (UNI EN
15886:2010).
The water absorption tests were carried out before and after treatment, according to
the capillarity test method adopted by the UNI-EN commission (UNI EN 15801:2010).
2.3 Artificial ageing
The durability of the performance of PLA and fluorinated PLA was tested with
accelerated ageing.
Treated and untreated samples of both marbles were submitted to
thermohygrometric cycles in an Angelantoni Challenge 500 Climatic Chamber.
Temperature and Relative Humidity changed according to the cycle represented in the
Figure 1. This cycle has been repeated for 15 times and colorimetric measurements were
repeated after 5, 10 and 15 cycles, while water absorption tests were carried out after 5
and 15 cycles.
In order to evaluate the performances of products after UV irradiation, treated and
untreated samples of both marbles were submitted to artificial ageing in a
CO.FO.ME.GRA Solar Box model 3000e up to 1,000 hours, according to the adopted
by the UNI-EN commission (UNI EN 10951:2001). Irradiance was kept at 500 W/m 2 .
Colorimetric measurements were carried out after 250, 500, 750 and 1,000 hours of
irradiation, while water absorption tests were repeated at the end of the 1,000 hours (750
hours for MG).

Figure 1. Temperature and Relative Humidity variation cycle
3. Results and discussion
3.1 Chemical-physical characteristics of polymers
The most important chemical-physical parameters for a polymer to be used as a
protective of a stone material are reported in Table 1.
Table 1. Principal chemical-physical parameters of polymers
Mw
(g/mol)
PD
Tg
(°C)
PLLA 68,000 2.0 60
PLLA-FLK 34,000 1.5 46
PLDA-FLK 18,000 1.1 30
Data on Table 1 show that the co-polymers containing a fluorinated moiety have
smaller molecular dimensions, in addition to a lower Tg, especially for PLDA-FLK: this
last datum is coherent with the supposed stereochemistry of the sample obtained using
rac-lactide, which should be amorphous. Moreover, PLDA-FLK has a low
polydispersity, near to the ideal value (PD = 1), that means that all the molecules of the
polymer have the same Mw.
3.2 Amount of product absorbed
Data reported on Table 2 show that in all cases, the amount of product absorbed by
the stone is lower than the expected one (16 g/m 2 ), this is due to the general low porosity
of a lithotype like an Apuan Marble. Anyway, a greater amount of the three polymers is
able to penetrate inside the porous structure of MVB, while the lower porosity of MG
makes the penetration of the polymers more difficult. In contrast with expectations,
PLDA-FLK is the least absorbed product on both marbles, although it has the smallest
molecular dimensions.

Table 2. Amount of product absorbed by stone samples
Amount of product absorbed (g/m 2 )
PLLA PLLA-FLK PLDA-FLK
MBV 12.8 14.2 11.8
MG 12.4 9.6 8.8
3.3 Colour variation and water absorption
In Figure 2 the colour variations for the stone samples treated with the polymers are
reported.
Figure 2. Colour variations (in terms of ∆E) of the treated marble samples
For all treated stones, ∆E values are lower or very close to the limits of the human
eye (~ 2), except for PLDA-FLK treatment applied on MBV (3.75). The more relevant
contribution to ∆E comes from ∆L, that is lightness. The samples were darkened (a
decrease in L) after treatment.
The presence of the fluorinated moiety in the polymers seems to give to the
treatment the desired increase of the hydrorepellent effect (Fig. 3): Protective Efficacy
of PLLA-FLK and PLDA-FLK treatments is consistently higher than the one of PLLA.

Figure 3. Protective Efficacy (P.E.%) of the treated marble samples
3.4 Artificial ageing
In Figure 4 the colorimetric variation of treated samples subjected to
thermohygrometric cycles are reported. The measures were repeated after 5, 10 and 15
cycles.
In Figure 5 the Protective Efficacy of the treated samples is reported: in this case
the water absorption tests were carried out after 5 and 15 cycles.
Figure 4. Colour variations (in terms of ∆E) of the treated marble samples after 0, 5, 10 and 15
cycles of thermohygrometric ageing

Figure 5. Protective Efficacy (P.E.%) of the treated marble samples after 0, 5 and 15 cycles of
thermohygrometric ageing
For all products applied on both marbles, the thermohygrometric cycles don’t entail
an aggravation of the colour variation of the stone surfaces. On the contrary, L
parameter (lightness) in all cases tends to come closer to the initial values, reducing ∆L.
The low Protective Efficacy given by PLLA to both marbles, further decreases after
5 and 15 thermohygrometric cycles. Regarding fluorinated polymers, instead, they
maintain very good values of P.E.% on MVB. On the other marble (MG), PLDA-FLK
even improves protective performance with thermohygrometric ageing, while P.E.%
values for PLLA-FLK slightly decrease going along with cycles.
These behaviours are probably due to the fact that an increase of the temperature
can carry to a reorganisation of the polymer structure in the first layers of stone material
and a better distribution on the surface: in this way, the colour turns to resemble to the
initial one, and the Protective Efficacy is improved or approximately unchanged. In the
case of PLLA treatment, the reduction of ∆L, together with the drastic decrease of
Protective Efficacy, can suggest the detachment of the product from the surface.
Figures 6 and 7 show respectively the colorimetric variations and the Protective
Efficacy of the treated samples subjected to artificial ageing in Solar Box: colour
measurements were carried out after 250, 500, 750 and 1,000 (only for MBV) hours of
irradiation, while water absorption tests were carried out at the end of the complete
ageing period: 1,000 hours for MBV and 750 hours for MG.

Figure 6. Colour variations (in terms of ∆E) of the treated marble samples after 0, 250, 500, 750
and 1,000 hours of UV irradiation
Figure 7. Protective Efficacy (P.E.%) of the treated marble samples after 0 and 750 (for MG) or
1,000 (for MBV) hours of UV irradiation

As well as for thermohygrometric ageing, UV irradiation brings to a decrease of ∆E
on the treated samples too. Again, L parameter (lightness) turns back towards the initial
values, except for PLLA-FLK on MBV. In this case, it can be noted an increase of ∆E
after the first 250 hours of irradiation and then a slight decrease up to 1,000 hours.
Protective effect of PLLA treatment on both marbles decreases after 750 or 1,000
hours of UV irradiation in Solar Box. The fluorinated polymers have instead a different
behavior on the two marbles: both polymers maintain a very good P.E.% on MBV after
1,000 hours of irradiation, while they drastically decrease their protective performance
when applied on MG.
This different trend of treated MG subjected to UV irradiation can be due to its
lower open porosity: the polymers irradiated can be more easily lost from a very
compact stone like MG, because they don’t penetrate in deep but remain on the surface.
4. Conclusions
The application of polymers from a monomer produced starting from renewable
resources in the field of Cultural Heritage represents an attracting alternative to
traditional petrochemical-based materials.
In this work, the properties of innovative homopolymers and block fluorofunctionalised
co-polymers of lactic acid for the protection of stone having a low open
porosity have been reported. As wished, the presence of a fluorinated group in the
polymer chain gives an enhancement of the water-repellent effect with respect to the
unfluorinated PLAs.
Such property is maintained even when the treated stone is subjected to
thermohygrometric variations cycles.
On the contrary, treated samples submitted to UV irradiation show a different
behaviour: the marble with the lower porosity (MG) seems to lose P.E.% as a
consequence of the detachment of the applied product. The other marble (MVB) instead
maintains unaltered the water-repellence given by the fluorinated products.
Taking account of the promising results of such fluorinated co-polymers as stone
protective products, it can be useful to test different applying methods (ie more diluted
solutions, iterated application, etc.) in order to reach a better penetration of the product
even in lithotypes having a very low porosity.
5. Acknowledgements
Authors would like to thank Solvay Solexis for kindly supplying Fluorolink-D10H
and Regione Toscana Project TeCon@BC, POR-FESR 2007–2013 for financial support.
6. References
Amoroso, G.G. 2002. Trattato di scienza della conservazione dei monumenti, Firenze,
Alinea Ed.
Borgioli, L. 2002. Polimeri di sintesi per la conservazione della pietra, Padova, Il Prato
Ed.
Dohene, E., Price, A.C. 2010. Stone conservation: an overview of current research, Los
Angeles, The Getty Conservation Institute.

Giuntoli, G., Frediani, M., Pedna, A. et al. 2012a. In Polylactid Acid: Synthesis,
Properties and Applications, Piemonte, V. (ed.) 161-189. New York: Nova Science
Publishers.
Giuntoli, G., Rosi, L., Frediani, M. et al. 2012b. Fluoro-functionalized PLA polymers as
potential water-repellent coating materials for protection of stone. Journal of Applied
Polymer Science, 125: 3125-3133.
Ljungberg, N., Wesslen, B. 2005. ‘Preparation and properties of plasticized poly(lactid
acid) films. Biomacromolecules, 6(3): 1789-1796.
Stevens, E.S: 2002. An introduction to the new Science of biodegradable plactics,
Princeton, Princeton University Press.
UNI EN Protocol 15886/2010. Conservation of cultural property, Test methods, Colour
measurement of surfaces.
UNI EN Protocol 15801/2010. Conservation of cultural property, Test methods,
Determination of water absorption by capillarity.
UNI EN Protocol 10951/2001. Cultural heritage, Natural and artificial stones, Method
for artificial solar light test.

THE INFLUENCE OF OSMOTIC PRESSURE
ON POULTICING TREATMENTS
Leo Pel 1 , Victoria Voronina 1 , Alison Heritage 2
1 Transport in Permeable Media, Department of Applied Physics,
Eindhoven University of Technology, P.O. Box 513,
5600 MB Eindhoven, The Netherlands
2 <strong>International</strong> Centre for the Study of the Preservation and Restoration of Cultural
Property (ICCROM), Via di San Michele 13, 00153 Rome, Italy
Abstract
The crystallization of salts is widely recognized as one of the most significant
causes of irreversible damage to many cultural objects such as wall paintings, stone
sculptures, and historic buildings. The removal of salts from these objects is however
difficult and often poultices are used. In these methods a wet poultice is applied to the
surface of the substrate to be treated and is kept in place for some period of time before
being removed. Many studies up to now on poulticing have focused on the salt and
moisture transport solely in terms of advection and diffusion. The objective of this study
is to demonstrate the potential contribution of osmotic pressure to salt extraction during
poulticing treatments. To this end we have conducted a series of experiments where we
have measured the moisture transport during poulticing for some well defined materials.
Here we have used Nuclear Magnetic Resonance (NMR) to measure non-destructively
the moisture transport during these experiments. This study shows that osmotic pressure
can exert a significant influence on salt extraction by poulticing methods during drying.
Importantly, as salt is transported from the substrate and into the poultice, this results in
a build-up of osmotic pressure within the poultice decreasing the effective pore-size of
the poultice. Therefore the build-up of osmotic pressure enhances the salt extraction and
thus increases the efficiency of the poulticing treatment.
Keywords: poulticing, osmotic pressure, NMR
1. Introduction
The crystallization of salts in porous media is widely recognized as one of the
primary causes of irreversible damage to many cultural objects such as wall paintings,
sculpture, historic buildings, and other artworks (Arnold and Zehnder 1991, Goudie and
Viles 1997). Moreover, contemporary buildings and civil constructions also suffer from
salt-related deterioration processes. Salt crystallization can therefore be regarded as a
common deterioration problem with significant cultural and economic implications.
Poulticing is a common method used in conservation to reduce the salt content of the
affected object (Verges-Belmin and Siedel 2005). The methodology of application is
relatively simple: the wet poultice is applied to the surface of the substrate to be treated
and is kept in place for some period of time before being removed. The desalination
treatment by poultice includes two main phases. The first is the wetting phase: water is
transported from the poultice into the wall where it starts to dissolve the salts. Water

may penetrate the substrate due to water vapor transport by means of water vapor
diffusion or due to water capillary suction. The second phase is the salt extraction. The
dissolved salt ions travel in the form of an aqueous saline solution from the substrate
into the poultice. This salt migration can be the result of two different processes. The
first is generated by the existence of a concentration gradient between the substrate and
the poultice. In this case the salt ions diffuse through the solution. Other one is realized
by the capillary water flow from the substrate to the poultice (generally due to drying)
and is accompanied by ions advection within the solution.
The present research focuses on salt extraction by drying poultices, during which
advection is the main mechanism for salt removal. In the case of advection based salt
extraction the efficiency of salt extraction is strongly dependent on the relative pore-size
range of the substrate and the poultice (Pel et al 2010). This extraction process is
however potentially considerably faster than diffusion based methods. This can to some
extent be achieved through the inclusion of clay minerals (such as kaolin) in poultice
mixtures (Lubelli and van Hees 2010)
Up to now most studies on poulticing have focused on the salt and moisture
transport solely in terms of advection and diffusion. However in salt extraction, as there
are salt gradients present so too will there be osmotic pressure gradients which could
have an influence. The aim of the work was to investigate the influence of osmotic
pressure on poulticing. First the fundamental aspects of capillary transport and osmotic
pressure relating to a combination of two porous materials drying will be discussed first.
The NMR method and setup for measuring non-destructively the moisture distribution,
during the experiments on the influence of the osmotic pressure in poulticing drying is
explained in Section 3. Finally the results of the experiments showing the influence of
the osmotic pressure on the desalination process will be discussed
2. Theory
The term 'advection' refers to the transport of mass by a moving medium. In the
case of salt extraction by advection, the dissolved ions can be transported by the
moisture flow from substrate into the poultice. Advection is generally more rapid than
diffusion, and so desalination treatments based on advection can be much faster.
However, in order for advection from the substrate into the poultice to take place,
certain requirements regarding the pore size distribution of the poultice and of the
substrate need to be fulfilled; in particular the poultice must contain a sufficient quantity
of pores that are smaller than the majority of those in the substrate (Pel et al 2010).
2.1 Influence of the capillary pressure
In general the driving force for the transport of water by advection is drying.
During drying, the largest pores will empty first, where the capillary pressure (Pc) is
lowest, as can be seen from the following equation:
2γ
Pc
= cos( ϕ)
(1)
r
m

In this equation rm is a pore radius that discriminates between the pores filled with
water (r < rm) and the empty pores (r > rm), γ [Nm -1 ] is the surface tension of the
liquid/vapor interface and φ is the contact angle between the liquid/air and liquid/solid
interface.
In most porous materials the pores are not uniform, and therefore there is a pore size
distribution. Hence in this case the overall macroscopic capillary pressure ψc of the
material is a function of its pore size distribution. For any moisture content θ [m 3 m -3 ],
there will be a critical pore radius, rm, corresponding to a certain capillary pressure that
discriminates between the pores filled with water and the empty pores. Hence the
macroscopic capillary pressure ψc is a function of the moisture content θ which can be
described thus:
= ψ (θ )
(2)
ψ c c
If we assume a perfect hydraulic contact at the poultice/substrate interface, the
capillary pressure will be continuous at this interface, i.e.:
p
( θ ) ψ ( θ )
ψ = (3)
p
s
s
Where ψp is the capillary pressure of the poultice, ψs the capillary pressure of the
substrate, and θp and θs the moisture content of the poultice and substrate at the interface.
Hence due to differences in the porosity and pore size distribution between the two
materials there will be a jump in moisture content across the interface. The relationship
between the moisture contents either side of the interface can be described thus:
p
−1
p
s
( θ ) f ( θ )
θ = ψ ψ =
s
s
Hence in general there will be a jump in moisture content across an interface.
Figure 1. The pores size distribution as obtained by mercury intrusion porosimetry for Migne
limestone and Bentheimer sandstone reflecting a fine and coarse porous material.
(4)

As an example reflecting a poultice/substrate combination we can have a look at the
drying of a combination of two materials, i.e., Bentheimer which is a material with
coarse pores and Migne limestone which is a material with fine pores. The pore size
distribution as measured by MIP for these materials is given in figure 1.
In figure 2 the measured moisture profiles are given for the water saturated
fine/coarse material combination. As can be seen the measured moisture profiles reflect
the pore size distributions. In both cases the material with the largest pores dries first, i.e.
the Bentheimer sandstone. These results conform to the general idea that in order to
extract salts by advection using drying poulticing methods, the poultice should have
pores smaller than that of the substrate (Pel et al 2010, Sawdy et al 2010).
Figure 2. The moisture profiles as measured during drying at various times for a combination of a
coarse and fine porous material. The samples are dried at the left side where as the right side is
sealed off. As a fine porous material Migne limestone is used and as a coarse porous material
Bentheimer sandstone is used.
2.2 Influence of the osmotic pressure.
When the substrate/poultice system contains a saline solution there will be an
additional contribution to the macroscopic capillary pressure due to the osmotic pressure
for each material, i.e.:
c
( θ ) ψ o
ψ ψ +
= , (5)
where the osmotic pressure, ψo, is given by:
RT
ψ o = ln( aw
) , (6)
Vw
where R is the universal gas constant, T the absolute temperature and aw the water
activity (for pure water aw=1 and hence the osmotic pressure is zero).

When the well known Pitzer's activity coefficient model is applied the osmotic
pressure can be calculated as (see e.g. Englezos and Bishnoin 1988):
( 2ν
ν )
⎛
⎜
⎛ 2ν
⎛
mν
n ⎞ 2
ψ
⎜
o = νRTm
l 1+ z+
z−
A1
+ ml
⎜ ⎟A2
+ ml
⎜
⎜
⎝
⎝ ν ⎠ ⎝
where
3 / 2 ⎞ ⎞
m n ⎟ ⎟
⎟
β 2 (7)
ν ⎟
⎠ ⎠
1/
2
Aφ
I
A1
=
1/
2
1+ 1.
2I
(8)
2
0
1
1/
2 ( 2 )
A = β + β exp − I
(9)
AΦ is the Debye-Huckel coefficient, ml the molality of the solution, z+ and z- are
ion charges, I the ionic strength of the solution, νm and νn are the number of moles of
ions produced by one mole of the electrolyte (i.e., 2 for NaCl) and β0, β1, and β2 are the
given parameters for Pitzer's activity coefficient model.
The osmotic pressure of a NaCl solution is given in figure 3 as an example as
calculated using the Pitzer model. As can be seen the osmotic pressure reaches almost
350 bar for a saturated sodium chloride solution. Hence due to the presence of soluble
salts in a porous material, its effective pore size (i.e., the equivalent pore size for a water
saturated system) will decrease as the total macroscopic capillary pressure is increased
by the osmotic pressure of the salt. Hence due to the salts absorbed in the poultice its
effective pore size will decrease.
Figure 3. The osmotic pressure of NaCl solution as a function of the concentration as calculated
from the Pitzer model.

3. NMR
In this study we have used Nuclear Magnetic Resonance (NMR) to measure nondestructive
and quantitative the moisture in the samples while drying. NMR is based on
the principle that in a magnetic field, nuclei have a specific resonance frequency and can
be excited by a radio frequency field. The resonance frequency f (Hz) depends linearly
on the magnitude of the magnetic field:
0
2 B f
γ
=
π
(10)
Where γ/2π (HzT -1 ) is the gyromagnetic ratio, B0 (T) is the main magnetic field. For 1 H
γ/2π is 42.58 MHzT -1 and 23 Na is 11.26 MHzT -1 . Therefore, by using a specific
frequency the method can be made sensitive to a particular type of nucleus, in this case
either hydrogen or sodium. The signal intensity S of a spin echo as used in the
experiment is given by:
⎡ ⎛ T ⎞ ⎛ ⎞⎤
r Te
S = ρ ⎢1
− exp ⎜
⎜−
⎟
⎟exp
⎜
⎜−
⎟
⎟⎥
(11)
⎣ ⎝ T1
⎠ ⎝ T2
⎠⎦
Where S is signal intensity, ρ is the density of the hydrogen nuclei, Tr and T1 are the
repetition time of the pulse sequence and spin-lattice relaxation time, Te and T2 are the
spin echo time and spin-spin relaxation time. To measure the maximum signal, i.e. from
all pore sizes, Te should be as short as possible as T1 and T2 are proportional to the pore
size. For the presented experiments, a home-built NMR scanner with a static magnetic
field of 0.78 T and gradient up to 0.3 T/m is used (Petkovic et al 2007). To perform
quantitative measurements a Faraday shield is placed between the coil and the sample.
Figure 4. A schematic diagram of the NMR setup for drying experiments. The Teflon holder with
the sample and the standard is moved in the vertical direction by means of a step motor.

The experimental set-up is given in Fig. 4. The sample, which has a cylindrical
shape with a diameter of 20 mm and a length up to of 80 mm, is moved vertically
through the magnet with the help of a step motor. It is sealed at all sides, except for the
top over which air with a relative humidity of 5% is blown. In this way a onedimensional
drying process is created. After each measurement the sample is moved in
the vertical direction by the step motor. The measurement time for the moisture content
at one position is in the order of 1 min. This procedure is repeated until a complete
moisture profile has been measured. A time stamp is given to each measurement point.
4. Osmotic pressure influence
In order to show the effect of purely the osmotic pressure we have conducted an
experiment where desalinate a Bentheimer with a poultice with the same pores size
distribution. If we would only take the capillary pressure into account we would not
expect any advection and therefore no desalination effect. In this experiment we have
dried a Bentheimer saturated with 5M NaCl solution on top of a 2 M NaCl solution
saturated Bentheimer. In this case due to the higher salt concentration the effective pore
size of the Bentheimer with 5 M NaCl solution will be smaller because of the osmotic
pressure. The resulting measured moisture profiles during drying are given in figure 5.
As can be seen indeed the Bentheimer with 5 M NaCl solution dried more slowly
indicating that its effective pore size is smaller due to the osmotic pressure.
Figure 5. The measured water profiles in a Bentheimer/Bentheimer system for several times
during drying. Initially the Bentheimer on the top was saturated with 5M NaCl solution and the
substrate Bentheimer was saturated with 2M NaCl solution.
The corresponding moisture content at the interface for this drying experiment is
plotted in figure 6.

Figure 6: The water saturation of the 2M NaCl solution saturated Bentheimer at the interface as a
function of water saturation of the 5M NaCl solution saturated Bentheimer at the interface (see
also figure 5). The solid line represents the relation when both materials are water saturated
whereas the dashed line represents the relation as determined from the capillary pressure curve
and the calculated osmotic pressure
As can be seen there is a clear deviation from the curve as expected from pure
capillary effects. After some time, a deviation can be seen from the relationship as
predicted on basis of the calculated osmotic pressure. This is due to the salt transport in
the Bentheimer saturated with 2M NaCl solution towards the interface, as a result of
which the concentration difference decreased at the interface and the osmotic pressure
will change.
5. Conclusions
These drying experiments demonstrate the effect that osmotic pressure has on salt
and moisture transport within porous materials. A saline solution in a porous material
will exert an osmotic pressure that will to reduce the effective pore size of the material.
This study shows that the contribution of osmotic pressure can exert a significant
influence on salt extraction using poultices during drying. Importantly, as salt is
transported from the substrate and into the poultice, this results in a build-up of osmotic
pressure within the poultice, thereby enhancing the extraction.
These findings have potential practical implications for the optimisation of
poulticing treatments. While the pore size requirements for advection to take place at
the start of the process remain, these constraints need not be quite so severe. They are
gradually overcome by the build up of an osmotic pressure due to the ongoing migration
of salt from the substrate to the poultice. As a result the longer a poultice stays in
contact with the substrate, the more it will accumulate salt, and thereby the osmotic
pressure is increased and its effective pore size will become smaller enhancing the
desalination process.

6. References
Arnold, A. and Zehnder, K. 1991. ‘Monitoring wall paintings affected by soluble salts’,
in The Conservation of Wall Paintings S. Cather (ed) , 103-136, Los Angeles: The Getty
Conservation Institute.
Englezos, P. and Bishnoin P.R. 1988. ‘Prediction of gas hydrate formation conditions in
aqueous electrolyte solutions’. AIChE Journal, 34: 1718-1721.
Goudie, A. and Viles, H. 1997. Salt weathering hazard, Wiley, Chichester
Pel, L. Sawdy, A. and Voronina, V. 2010 ‘Physical principles and efficiency of salt
extraction by poulticing’. Journal of Cultural Heritage 11: 59–67.
Lubelli, B. and van Hees, R.P.J. 2010, ‘Desalination of masonry structures: Fine tuning
of pore size distribution of poultices to substrate properties’, Journal of Cultural
Heritage 11: 10-18.
Petkovic, J. Huinink, H.P. Pel, L. Kopinga, K. and van Hees, R.J.P. 2007. ‘Salt
transport in plaster/substrate layers’. Materials and structures, 40: 475-490.
Sawdy, A. Lubelli, B. Voronina, V. Funke, F. and Pel, L. 2010, ‘Optimising the
extraction of soluble salts from porous materials by poultices’. Studies in Conservation,
55: 26-40.
Verges-Belmin, V. and Siedel, H. 2005 ‘Desalination of masonries and monumental
sculptures by poulticing: A review’. Restoration of Building and Monuments, 11: 1-18.

CRUST FORMATION PROCESSES ON ARCHAEOLOGICAL MARBLE,
CONSERVATION METHODOLOGY AND TREATMENT EVALUATION
Nikolaos-Alexis Stefanis, Panagiotis Theoulakis
Technological Educational Institute of Athens, Department of Conservation of
Antiquities and Works of Art, Athens, Greece
Abstract
This work presents the findings of a study that was conducted in order to reveal the
geological processes that lead to the formation of different crust layers on the surface of
buried archaeological marble. The case study is a group of about 3,000 marble
fragments, belonging to the highly decorated altar constructed by Octavian Augustus in
31BC in Nikopolis at the western coast of Greece. The excavation lies at an area, which
displays a hilly terrain with gentle slopes. The hydrographic network is composed of
small streams with seasonal flow. The area of the monument consists of geological
formations of the Neogene, of alternating conglomerates, sands, clays and marls which
are covered by quaternary deposits and deposits of artificial materials.
The diverse burial environment led to the formation of different crust layers on the
surface of the marble, which depended on the location and depth that the fragments were
found. Representative samples were examined under the petrographic and scanning
electron microscope, coupled with energy dispersive X-ray analyser (SEM/EDX).
Mineralogical composition of the encrustations was determined by X-ray diffraction
(XRD). The results of the study revealed different formation mechanisms that created
three different types of encrustations, including loose depositions and impregnations
with metal oxides and hydroxides, calcareous and siliceous encrustations.
At the second part of this work, the conservation methodology that was developed
and the techniques tested in order to remove the encrustations from the marble surfaces
are presented. Before the application of any cleaning treatment, the surface of the marble
was thoroughly studied with microscopic techniques. To determine the depth of cleaning,
a set of criteria was set, with a specific number of parameters, which were used during
the treatments in order to control the quality of cleaning and to evaluate the result. The
tested cleaning techniques included mechanical, chemical and laser cleaning. The results
of each technique were evaluated in order to propose the most appropriate one for each
case of encrustations.
Keywords: archaeological marble, crust formation, cleaning, treatment evaluation
1. Introduction
The archaeological site of Nicopolis is situated in the plain of the peninsula of
Preveza, between the Ionian Sea and the Ambracian Gulf. A little to the south, opposite
the tip of the peninsula of Preveza, lies the promontory of Actium, known for the
famous sea battle between Gaius Julius Caesar Octavian, Cleopatra and Antonius in
31BC In the centre of this area, spread the ruins of the city that Octavian built in
commemoration of his victory, a victory that constitutes a turning point in the history of

the ancient world. North of Nicopolis lies the monument that Octavian erected after the
sea battle of Actium. The building, a trophy as well as a sanctuary, was both a symbol of
Octavian’s victory and power and also a monument of political and religious
propaganda.
The monument was built on two terraces, of which the lower is defined by two
retaining walls. An inner court is formed measuring 38x38 m, in which stood a
monumental altar (6x22 m) and three pedestals for over-life size statues. The entire
complex formed a kind of open-air sanctuary (Figure 1). According to the votive
inscription in the upper part of the façade of the monument, it was dedicated to Mars
and Poseidon. Apollo’s sacred hill and the Victory monument were already visible from
the city’s north gate.
From the first excavations at the aera, the Greek Archaeological Service marked the
site as one of immediate interest. During the 1960s, interest was focused on salvage
excavations and small-scale reconstructions. Since 1995, an initiative of the 12 th
Ephorate of Antiquities has launched the financing of the excavations in the
archaeological site of Nicopolis. In the layers above and around the foundation,
numerous smashed marble architectural fragments were found. In some cases the
fragments were found in piles, which suggest deliberate dismantlement, rather than
abandonment and natural collapse. More than 43,000 fragments have been found
including a variety of moldings and fragments of relief sculptures from the altar’s
decoration. Of this total, most are undecorated, some of which are of small size (data
acquired from the 12 th Ephorate of Prehistoric and Classical Antiquities).
The burial environment caused the formation of different encrustations and
impregnations on the stone surface. The aim of this project was to conclude to the most
suitable method of cleaning the stone fragments. In order to reach to a conclusion, the
stone was mineralogically characterised, the depositions on its surface were extensively
studied and analysed in terms of their composition and physical characteristics, and a
variety of different cleaning methods were tested and finally evaluated.
Figure 1. Reconstruction of the altar
(12 th Ephorate of Prehistoric and Classical Antiquities)
2. The burial environment
The altar of Octavian Augustus occupies an area of approximately 6,300 m 2 . The
area displays hilly terrain with gentle, natural slopes. The hydrographic network, which

the rainwater is drained through, is composed of small streams with seasonal flow. The
area is not susceptible to intense erosion phenomena.
The area of the monument consists of geological formations of the Neogene which
are covered by quaternary deposits and deposits of artificial materials. The Neogene
formations consist of alternating conglomerates, sands, clays, mixtures and marls. These
formations present a characteristic irregular form resulting to lateral transitions from one
formation to another.
Specifically, the geological formations that structure the area, from the oldest to
the youngest, are:
• Geological formations that are encountered at a depth of 30-35 m. They consist of
marly clays and humus ash horizons. These formations contain organic materials. In
between the marly formations, thin layers, of about 1-2 m, of sandy marls can be
found.
• Marly formations that are covered by fine sand with thin conglomerate horizons.
The fine sand is converted laterally into yellowish-brown sandy-clayey formations
that are exposed at the archaeological site and at the area north of it.
• Fine sands that are covered with conglomerates with clayey binder.
• Conglomerates that are covered by a horizon of marls.
• Artificial deposits that are found in the superficial layer.
The monument has been fully excavated at various depths. Stratigraphic
observations were made during the excavation, whilst the stratigraphy was documented
at the deep sides of the trenches. The basic stratigraphic sequence is as follows:
• OM1 layer: surface (humus) with plants having a brownish colour.
• OM2 layer: yellowish sandy clay, either pure or strongly conglomerated.
The survey concluded that the foundation of the monument was constructed on
clean soil of the OM2 layer. On these two main layers, several changes can be detected,
depending on the use and the interventions to the monument, such as destruction layers
and layers of deposits that consist of geological and archaeological material (data
acquired from the 12 th Ephorate of Prehistoric and Classical Antiquities).
The soil at that location is slightly alkaline with a pH of 7.4 to 8.0. Its chemical
composition is mainly characterised by the presence of carbonates, while nitrates and
magnesium ions are absent. In some samples, small quantities of sulphates and chlorides
were identified.
The mineralogical analysis of the soil, carried out by XRD, showed the presence of
the clay minerals montmorillonite (Al2O3.4SiO2.nH2O), illite (K2O.3Al2O3.6SiO2.2H2O)
and kaolinite (Al2O3.2SiO2.2H2O), as well as a high content of calcite (CaCO3) and
quartz (SiO2), and a few feldspars.
The analysis of the excavation data, show that 36 per cent of the fragments studied,
were found at a depth of 2.5 to 2.9 m, 26 per cent of the fragments were found at a depth
ranging between 1.0 and 1.5 m, and the rest were excavated from a depth between 0.0
and 0.5 m. Sampling was carried out in different fragments from different depths in
order to obtain representative results.
3. The construction material
The stone used in the construction of the monument is marble. All the samples that
were examined have the same general petrographic characteristics with slight variations
concerning mainly the existence of impurities and veining in the parent rock. The

impurities were detected mainly at fractured surfaces and not on sculptured surfaces,
indicating the careful selection and placement the marble.
Mineralogical analysis of the stone was carried out by XRD, light microscopy (LM)
and SEM/EDX. The analyses showed that calcite (CaCO3) is the dominant mineral but
dolomite (CaCO3.MgCO3) was also identified in relatively small quantities. The small
quantity of dolomite is not able to characterise the marble as dolomitic. Many impurities
and negligible mineralogical components were also identified. These are mainly
phyllosilicate minerals of the mica group, such as muscovite Αl2(Si3Al)O10(OH,F)2,
quartz crystals (SiO2) and in a few fragments ankerite Ca(Fe,Mg,Mn)(CO3) was
identified in marble veining.
Regarding the size of the calcite crystals, the marble of the monument can be
characterised as fine-grained with granoblastic tissue.
4. Study of the depositions on the marble surface
The macroscopic examination of representative fragments, in combination with
mineralogical analysis with XRD, LM, and SEM/EDX, of the depositions on the marble
surface, concluded to the categorisation of the depositions into three main groups:
• Dendritic depositions,
• Coherent depositions (encrustations) with either pure calcitic composition or with
calcitic-aluminosilicious-ferric composition,
• Loose depositions and impregnations with either ferric and aluminosilicious
composition of dark brown to black colour, or with aluminosilicious composition of
light brown colour.
4.1 Dendritic depositions
The main characteristic of these depositions is the form in which they appear
(Figures 2 and 3). They were found in several fragments after the removal of loose soil
deposits. In some cases two layers were identified. The upper layer with thickness of 50-
100 μm, which is more compact, and the lower layer with thickness of 150 μm
containing confined grains of calcitic, silicious composition along with aluminosilicates
(Figures 4 and 5).
Figure 2. Photomacrograph of the
surface of the marble.
Figure 3. Photomacrograph of the
deposition layer.

Figure 4. Light Microscopy //Nichols.
The two layers of the depositions can
be distinguished.
Figure 5. SEM photomicrograph of the
deposition with the confined grains in
the lower layer.
4.2 Coherent depositions (encrustations)
Case 1: These are hard, compact crusts of mainly calcitic composition with
silicious material. Their main characteristic is their hardness and the relief in which they
appear (Figure 6). Their thickness varies from 150 to 900 μm. In some cases two
characteristic layers can be distinguished. The upper layer that is compact and has a
thickness between 150 and 300 μm. In most cases a characteristic stratigraphy can be
identified with alternating layers of calcite and quartz (Figures 7 and 8). The lower layer,
where grains of calcite and quartz are confined (Figure 9). Its thickness ranges from 150
to 600 μm. In some cases, the trapped grains are large, measuring up to 500 μm. In
general, the boundaries between the layers described and the stone substrate are quite
clear.
Case 2: These are encrustations of ferric and aluminosilicious composition with a
maximum thickness of 80 μm with an alternating calcitic layer, which is relatively
compact. They present a maximum thickness of 200 μm and in most cases the limits
with the stone substrate are very clear (Figures 10 and 11).
Figure 6. Photomacrograph of the
encrustation (case 1) on the surface of
the marble.
Figure 7. Light Microscopy //Nichols.
Alternating deposition layers of calcite
and quartz.

Figure 8. SEM/EDX mapping of Ca
and Si of the same area shown at
figure 7. Alternating layers of
siliceous and calcareous deposits.
Figure 10. Photomacrograph of the
encrustation (case 2) on the surface of
the marble
Figure 9. Light Microscopy //Nichols.
Two deposition layers can be identified
as well as the grains of calcite and
quartz.
Figure 11. SEM photomicrograph of
the encrustation. Small Al-Si grains
can be identified. Clear limits with the
substrate.
4.3 Loose depositions and impregnations
Case 1: They mainly consist of silicon, aluminium and iron. In some cases, they
appear as a layer of loose deposits of moderate consistency (Figure 12). The thickness of
this layer does not exceed 100 μm.
Case 2: Loose depositions and impregnations of aluminosilicious composition.
They penetrate between the crystals of the stone, to a depth of 500 μm (Figure 13).
Figure 12. Photomacrograph of the
deposition (case 1) on the surface of
the marble
Figure 13. Light Microscopy
//Nichols. Impregnations (black stains)
of ferric and aluminosilicious
composition.

5. Cleaning
Cleaning is one of the key stages of conservation interventions, aiming at removing
the substances that promote stone weathering as well as to restore the aesthetic value of
the work by revealing the original surface and its details. The most important step is to
determine the cleaning depth (Papakonstantinou et al., 2007), which is defined by the
preservation state of the original surface of the substrate, protecting at the same time the
relief of the sculptured surface (Webster, 1992). In order to evaluate the cleaning
treatments, clear criteria were set with a number of parameters that were used during and
after cleaning treatments, to control the quality of each treatment (Moropoulou et al.,
2002).
Acceptable cleaning limits were determined by the study of the morphology of the
depositions and the substrate. For each deposition category the desired cleaning depth
was determined. After the application of the cleaning methods the treated surface was
examined in order to evaluate the result.
5.1 Cleaning depth
Dendritic depositions: The aim of cleaning was to remove the top layer of the
deposition having thickness of 200-250 μm. The interface of this layer to the substrate
was the cleaning limit. Since the outer layer of the stone displayed an intense relief, of
40 μm, particular care was taken in order not to affect it during cleaning (Figure 13).
Coherent encrustations: The scope was to remove the coherent calcitic depositions
of 200-1,000 μm thickness. The interface of this layer to the substrate was the cleaning
limit. The outer layers of the stone presented a relief of 20-50 μm, and particular
attention was paid in order not to affect it during cleaning. The original surface of the
stone was protected by preserving a very thin layer (20-30 μm) of the deposition. This
layer is translucent and it does not affect the aesthetic values of the stone surface (Figure
14).
175μm
40μm
Figure 13. Cleaning limits for
dendritic depositions
250μm
Figure 14. Cleaning limits for coherent
depositions
5.2 Cleaning methods tested
Based on the chemical-mineralogical composition and the morphology of the
deposition layers, the tested cleaning methods were:
• Poultices, based on aqueous solutions of ethylenediaminetetraacetic acid (EDTA)
with sodium bicarbonate (NaHCO3) and ammonium bicarbonate ((NH4)HCO3),
mixed with sepiolite.

• Mechanical methods: hand tools, micro-sandblast, and ultrasonic scaler. In the case
of micro-sandblast, aluminium oxide (Al2O3) with grain size ranging from 25-250
μm was used as the abrasive material, at an operating pressure ranging from 100 to
600 kPa.
• Two types of laser systems, developed at the Institute of Electronic Structure and
Laser – FORTH, Greece (Pouli et al., 2003): The Q-switched Nd:YAG @1,064 nm
(ElEn) and the SL805 Nd:YAG @1,064 nm (Spectron). Both laser systems were
tested with different fluence (J/m 2 ) and pulse duration.
• Water-based methods: Wetting, vapour blasting.
All cleaning methods were tested on each of the three categories of the deposition
layers. After the treatment and the removal of the depositions, the stone surface was
studied by the use of SEM/EDX and LM, in order to evaluate the result.
6. Results
The results from the cleaning tests performed are presented by category of deposits.
The evaluation of each method was based on the results of laboratory examination and
analyses as well as on the macroscopic characteristics of the stone surface after the
application of each cleaning method. The results were then classified based on the
following scale:
• “Effective method”: The cleaning was successful. Depositions are removed without
disturbing the substrate and the patina layer.
• “Method partially effective” due to one of the following reasons:
- The thickness of the depositions is reduced but the cleaning is not fully achieved.
- A complementary method is required to achieve the result
- Can be applied to specific sub-cases.
• “Method ineffective” due to one of the following reasons:
- Cleaning is not achieved and the depositions are not removed.
- The method is aggressive and causing damage to the substrate.
6.1 Dendritic depositions
The method of micro-sandblast with small grain size of Al2O3 was excluded
without further testing, because it was not possible to apply it to the limited areas of this
type of depositions. The method was evaluated as “ineffective”.
The method of poultices showed no satisfactory results. The depositions were not
removed, but their thickness was reduced. The results were good in cases where, after
the application of the poultice, ultrasound scraper was applied. In these tests, the
dendritic deposits were removed to a greater extent. The method was evaluated as
“partially effective”.
The method of the ultrasonic scaler at low ultrasound intensity on a wetted stone
surface produced satisfactory results but the depositions were not completely removed.
The method was evaluated as “partially effective”.
The method of vapour blasting had no effect in removing the deposits. The method
was evaluated as “ineffective”.
The method of laser (Q-switched Nd: YAG laser@1,064 nm) had a very good
result with optimal operating conditions of fluence at 0.62 J/cm 2 , pulse rate at 10 Hz and
a variable number of pulses. The substrate is not disturbed while the time required to
remove the deposits is small. The method was evaluated as “effective”.

The method of laser (Spectron laser system SL805 Nd: YAG laser@1,064 nm) had
better results with optimal operating conditions of fluence at 1.7 J/cm 2 , pulse duration 10
ns and a variable number of pulses. The substrate is not disturbed while the time
required to remove the deposits is small. The method was evaluated as “effective”.
6.2 Coherent depositions (encrustations)
The method of micro-sandblast with small grain size of Al2O3 did not give
satisfactory results. The substrate surface appeared rough and it the method was likely to
cause damage to the substrate in areas where the encrustation is less thick. The method
was evaluated as “ineffective”.
The method of poultices did not produce good results. Although application times
were quite large, the depositions were not removed nor was their thickness reduced. The
method was evaluated as “ineffective”.
The method of the ultrasonic scaler at low ultrasound intensity on a wetted stone
surface removed the crusts of aluminosilicious composition and reduced the thickness of
the calcitic ones. The method was evaluated as “partially effective”.
The method of vapour blasting had no effect in removing the deposits. The method
was evaluated as “ineffective”.
The method of laser (Q-switched Nd: YAG laser@1,064 nm) had satisfactory
results with optimal operating conditions of fluence at 1.43 J/cm 2 , pulse rate at 10 Hz
and a variable number of pulses. After the application though, remains of the crust on
the stone surface were observed which were removed by the use of ultrasonic scaler.
The method was evaluated as “partially effective”.
The method of laser (Spectron laser system SL805 Nd: YAG laser@1,064 nm) was
the most effective of all methods tested. The optimal operating conditions were 3.8
J/cm 2 of fluence, with pulse duration of 10 ns and a variable number of pulses. The
substrate was not disturbed and the crust was completely removed. In the case where the
operator persisted in a specific area, the high operating power could possibly disturb the
substrate. The method was evaluated as “effective”.
6.3 Loose depositions and impregnations
The method of micro-sandblast with small grain size of Al2O3 was excluded
without further testing, since the impregnation does not create a surface layer. Its use
may have caused damage to the substrate. The method was evaluated as “ineffective”.
The method of poultices did not achieve adequate cleaning. Further tests with
poultices of hydrogen peroxide (H2O2) and ammonia (NH3) gave very good results. The
method was evaluated as “effective”.
The method of the ultrasonic scaler at low ultrasound intensity on a wetted stone
surface was excluded without further testing, since the impregnation does not create a
surface layer. Its use may have caused damage to the substrate. The method was
evaluated as “ineffective.
The method of vapour blasting had good results on flat surfaces with no or little
relief. The impregnations were removed without disturbing the substrate. Particular care
was needed in relief surfaces where the highest points received greater amount of vapour
resulting in uneven cleaning. The method was evaluated as “partially effective”.
The method of laser (Q-switched Nd: YAG laser@1,064 nm) had satisfactory
results with operating conditions of fluence 1.43 J/cm 2 , pulse rate at 10 Hz and a

variable number of pulses. The impregnation was completely removed without
disturbing the patina layer and the surface of the substrate. The method was evaluated as
“effective”.
The method of laser (Spectron laser system SL805 Nd: YAG laser@1,064 nm) is
the most effective with optimal operating conditions of fluence at 2.8 J/cm 2 , pulse
duration 10 ns and a variable number of pulses. The substrate was not disturbed and the
deposition was completely removed. The method was evaluated as “effective”.
7. Conclusions
By comparing the results of the cleaning tests, some general remarks can be made.
The marble surface after cleaning with micro-sandblast appeared rougher. According to
the degree of control of each method they were classified as follows: laser, ultrasonic
scaler, micro-sandblast, vapour blast, poultices. Poultices and ultrasonic scaler produced
the slowest cleaning times of the all the methods tested.
The SL805 Nd:YAG laser@1,064 nm laser system proved to be the most suitable
for the removal of all three types of depositions since the quality of the final result was
high and the possibility of damaging the original surface was minimum. The settings of
the laser system though, are different for each case of depositions. Other cleaning
methods gave satisfactory results and the factor of cost-effective treatment has to be
carefully considered.
8. Acknowledgements
The authors would like to thank the President of the Scientific Committee for the
Restoration of Ancient Nicopolis, Dr. K. Zachos for providing the archaeological
material and Dr. Paraskevi Pouli, researcher at the Institute of Electronic Structure and
Laser, FORTH for her collaboration with the laser cleaning tests. The study was funded
by the S.S. Niarchos foundation.
9. References
Moropoulou, A., Tsiourva, Th., Bisbikou, K., et al. 2002. “Evaluation of cleaning
procedures on the facades of the Bank of Greece historical building in the center of
Athens”. Building and Environment, 37: 753–760.
Papakonstantinou, E., Panou, A., Franzikinaki, K., et al. 2007. “The surface
conservation project of the Acropolis monuments: Studies and interventions”. Paper
presented at the XXI <strong>International</strong> CIPA Symposium, Athens, Greece, 01-06 October
2007.
Pouli, P., Zafiropulos, V., Balas, C., et al. 2003. “Laser cleaning of inorganic
encrustation on excavated objects: evaluation of the cleaning result by means of multispectral
imaging”. Journal of Cultural Heritage, 4: 338s–342s.
Webster, R.G.M. 1992. Stone Cleaning and the Nature, Soiling and Decay Mechanisms
of Stone. London: Donhead.

CONSERVING STONE IN TURBULENT TIMES: IMPACT OF ARMED WARFARE ON IMMOVABLE
HERITAGE
Lisa Mol, PhD 1
1 Oxford Rock Breakdown Laboratory, School of Geography and the Environment, University of Oxford
Abstract
Little is known about the impact of armed warfare, such as bullets and shrapnel, on heritage sites. The long-term
geomorphological consequences of a high- energy event such as a bullet impact into a stone surface are still
unknown, despite a large number of monuments getting caught in the cross-fire during the recent uprisings in the
Middle East. In addition, the IPCC marks this area as likely to experience increased temperatures and changing
precipitation patterns, which may exacerbate the weathering rates of damaged surfaces.
In addition to the physical threat of armed warfare, the social problems associated with unstable areas can include a
lack of access to heritage sites, low priority attitudes towards conservation and a lack of allocated funds for research
and conservation methods. These problems call for effective, simple and low-cost methodologies for assessing
damage to stonework.
Laboratory studies have been conducted to investigate simulated ‘war victims.’ Using an assortment of bullet holes
created in sandstone of known strength and characteristics, the impact, deformation, los of surface material and
reduction in strength were investigated using a variety of geomorphological methods including photography, surface
hardness (Piccolo), and ERT moisture measurements. The preliminary results of these studies are presented here.
Keywords: weathering, armed warfare, sandstone, conservation, moisture regimes, surface hardness
1. Introduction
Damage to immovable heritage frequently occurs during armed conflict and can have a significant detrimental
impact on the future preservation of these sites. Widespread use of weapons in countries such as Iraq, Egypt, Syria
and Libya, all of which contain large numbers of heritage sites, is a grave cause for concern and while the unrest
continues it is difficult to obtain an accurate assessment of the damage incurred. Aside from the red tape usually
encountered around heritage sites, health and safety risks form even greater obstacles when it comes to expert access
to in these areas. Any pre-existing local protection structures and organizations are also at risk of rapid deterioration,
as illustrated by the State Board of Antiquities and Heritage (SBAH) in Iraq, which closed down in 1980 at the start
of the First Gulf War (against Iran) when it lost a large proportion of its budget and many of its employees were sent
to the war front and others left the country (Ghaidan, 2008 p. 87).
The destruction of heritage during armed warfare is by no means a recent problem; the burning of the library at
Alexandria is one of the earliest known victims of deliberate destruction by an invading party likely because of a
clash in religious and intellectual believes – the exact cause and time of destruction have never been confirmed but
the library certainly has not been in existence for centuries (Thiem, 1979). Indeed, one only has to look at the San
Marco Basilica in Venice, Italy, to see the result of centuries of looting and destruction of the original settings of the
various statues, mosaics and pillars.
However, the nature of warfare has drastically changed over the past century. The introduction of highly destructive
weaponry such as grenades, bombs and machine artillery has increased the risk of damage or destruction of heritage
in areas of conflict. Vivid scars of WWI and WWII are still visible in London, in particular on the east side of the
Tate Britain (see figure 1) while the bullet holes in the General Post Office in Dublin bear testament to the execution
of the leaders of the Easter Rising in 1916. Increasing velocity and effectiveness of the projectiles used in warfare
have resulted in a far greater potential for damage. These high-impact weapons can leave historic materials’ surfaces
weakened and vulnerable to accelerated deterioration processes.
While a significant body of engineering science research exists investigating the impacts of alien objects on surfaces,
very little is known about the subsequent deterioration of the material (over decades to centuries). Equally, within
conservation science very little is known about the physical impact of armed warfare on heritage sites subjected to
weathering processes. Bullet and shrapnel holes are mentioned only in passing in stone damage classification
schemes. The ICOMOS-ISCS (2008) classifies weapon damage as ‘impact damage’ without further detail or

classification or distinguishing between the nature and impact of the projectile. Other classifications, such as that
used by Fitzner et al (2003), omit armed warfare entirely.
The physical impact of warfare on heritage can have damaging social and cultural effects. For example, failure to
protect heritage has ‘contributed to the breakdown of social order, and alienation of much of the Iraqi population’
(British Academy, 2010). Preservation of heritage, and thereby cultural identity, in unstable times can greatly
enhance the cultural well-being of a population. Despite protective measures, such as The Hague Convention (1954),
heritage can be deliberately targeted to destroy cultural identity, such as during the Croatian Independence Wars
(1991-1995). Thus there is a clear need for a better understanding of the threat of armed warfare to heritage
conservation.
To address this need for a greater understanding of the geomorphological impact of armed warfare, preliminary
laboratory tests were performed in the Oxford Rock Breakdown Laboratory which simulate stone surfaces damaged
by bullet impacts and subjected to fluctuating environmental circumstances. To this end, artificial ‘war victims’ were
created by firing bullets at sandstone samples from a known distance which were analysed using rock surface
hardness, moisture distribution and material loss as indicators of aggravation. In addition, half the samples were
treated with a surface consolidant (Wacker) to mimic case hardening as often found in building stone that has been
exposed to the elements for a number of decades. This article reports the first findings.
2. Methodology
2.1 Stone samples
A well-consolidated sandstone was selected for the samples to reduce the risk of severe damage. Previous laboratory
tests showed that this is a mesoporous sandstone (average pore size between 40 and 70µm) and a WAC of 1.8%. The
blocks were cut to 15 x 15 x 7.5 cm, the larger 15 x 15cm surface was used to conduct the tests. All blocks were
weighed before impact, after impact and after environmental cabinet treatment. The block originate from a quarry in
the Huesca Province (northeast Spain).
2.2 Pre-mortem tests
Before any tests were carried out an Proceq Piccolo was used to map the rock surface hardness. In the Equotip, a 3
mm diameter spherical tungsten carbide test tip is mounted in an impact body and impacts under spring force against
the test surface from which it rebounds (Verwaal and Mulder, 1993). The velocity before impact (V,) and after
impact (V2) are measured automatically and displayed as a ratio (V2/V, x 1000) which is denoted by the unit “L”, or
Leeb unit (Hack et al, 1993). Each block was divided into 16 squares measuring 3.75 x 3.75cm. Within each square
four measurements were taken to obtain a total of 64 measurements per block surface.
To simulate the effects of cementation and case hardening, as can often be found on stone exposed to weathering
processes (McAlister et al, 2003; Rossi-Manaresi and Tucci, 1991) samples 5-8 were treated with Wacker OH. This
compound is known to have a bonding effect quantifiable by non-destructive methods like ultrasonic velocity (da
Costa and Rodrigues, 2011). The treatment was applied with a pipette to the front surface of the block only and left
to infiltrate and cure over 10 days in a fume cupboard. All non-destructive tests were therefore conducted both
before and after the Wacker treatment to assess the influence of the consolidation technique.
2.3 Projectile impact
Samples 2-4 and 6-8 were taken to the Whitney Rifle Club (Oxfordshire, UK). Bullets of .22 calibre were shot from
20 yards distance in the centre of each block, clearly marked during the pre-mortem tests, using a standard shouldersupported
rifle. Each centre was hit three times to increase the impact of the stones. In all cases the impact areas
overlapped, as demonstrated by sample 6 shown in figure 1. This figure also shows the sampling strategy for the
Piccolo (as indicated by the crosses, 4 measurements per square section) and the ERT measurements (moisture
movement, see post-mortem tests).

Figure 1: Impact area at sample 6, also showing sampling strategy for the Piccolo and ERT.
2.4 Post-mortem tests
A number of tests were carried out after the impact. Firstly the samples were re-weighed to determine volume loss
during the impact. The Piccolo tests were repeated to assess stones surface strength loss.
To further the investigation into the deterioration of the samples under more extreme environmental conditions the
samples the were placed in an environmental cabinet (Sony-FE 300H). A programme was designed for 6 consecutive
days with alternating 6 hour cycles of 15ºC and 65ºC. Half the samples (3,4,7 and 8) were placed in 1.25cm of water
to mimic groundwater conditions whereas samples 1,2,5 and 6 were placed directly within the cabinet without
further access to water other than the humidity within the cabinet which was kept low to account for the evaporating
water of the immersed samples.
For further clarification of the samples and their treatment, see table 1 for an overview.
Sample no Wacker treatment Bullet holes Water immersion in cabinet
1 No No No
2 No Yes No
3 No Yes Yes
4 No Yes Yes
5 Yes No No
6 Yes Yes No
7 Yes Yes Yes
8 Yes Yes Yes
Table 1: Overview of individual sample treatment
The immersion of half the samples in water fulfilled a dual role; firstly to test the hypothesis that a known weakness
in the surface would lead to increased evaporation of moisture from that particular area, leading to an increase in

deterioration around the area. Previous work by Mol and Viles (in press) showed that increased temperatures lead to
higher capillary pull through a mass of stone and high surface evaporation. In addition, the internal movement of
moisture is known to deteriorate a stone surface (see Mol and Viles, 2010; 2012). This gives a theoretical
justification for the assumption that a weakness such as a bullet hole would under high temperature conditions
experience higher deterioration rates. However, this has never been tested, nor is the influence of case hardening
known in this scenario.
To monitor the movement of moisture through samples 3,4, 7 and 8 a GeoTom (Geolog2000) ERT equipment was
used, which can measure profiles involving up to 100 electrodes. Miniaturized cabling attachments with fixed
spacing have been developed in the OxRBL which use spring-loaded permanently 0.5cm spaced electrodes to
measure resistivity at up to 2.5cm depth, giving a very high resolution 2D distribution of resistivity. The basic
principle of ERT is to pass an electrical 12V current between two designated potential electrodes. By increasing the
spacing between the two designated electrodes the current penetrates the stone further, building up a 2D transect of
point specific resistivity which can then be recalculated into a most probable resistivity distribution using the
inversion programme RES2DInv (see Sass and Viles, 2006 and Mol and Preston, 2010 for more information). Each
sample was measured along three vertical transects,
3. Results
3.1 Material loss
Material loss from the surface after the bullet impact was almost negligible, as it averaged 2.57 grams per sample
(see figure 2). Interestingly, there was no noticeable difference reported between Wacker treated surfaces and nontreated
surfaces. However, this illustrates that while material loss at the time of impact may be nominal the impacts
of the object could have more far reaching consequences for strength loss and subsequent weathering, as illustrated
in the following sections.
Figure 2: Weight of samples used in bullet experiment before and after impact
After the environmental cabinet treatment all blocks appeared heavier than before the cabinet. This is due to the
uptake of water, whether directly from the pooled water around samples 3,4,7 and 8 or from the evaporated moisture
in the air. Further tests are conducted at the time of writing to determine actual weight changes in the samples, but
could not be incorporated in this manuscript.
3.2 Rock Surface Hardness
The repeated measurements of the sample surfaces are shown in figure 3. The initial Wacker treatment shows a
significant increase in the measured surface hardness at samples 5-8. This fits into the original hypothesis that one
treatment with Wacker could be used to simulate the cementation of the surface associated with case hardening.

However, this increase in hardness appears to be rapidly reversed at the time of impact, while the overall surface
hardness in samples 1-4 does not appear to reduce as dramatically, staying either roughly even or in the case of
sample 4 strangely increasing a fraction. To interpret this data the elasticity of the rock face needs to be taken into
account; while Wacker can protect the surface from overall weathering processes such as abrasion and infiltration of
moisture, due to its hydrophobic nature (Rodrigues da Costa and Rodrigues, 2011), it can also be a deterioration
factor as the loss of elasticity within the stone surface can severely decrease the overall stability of the surface upon
impact. This phenomenon has been previously applied by Momber (2004) who concluded that harder surfaces (as
simulated by the Wacker treatment) are prone to developing fracture networks upon impact whereas in softer
materials (as simulated here by the untreated surfaces) crack formation was dampened by plastic flow of the impact
resonance.
Figure 3: Overview of mean Piccolo measurements per sample stage. Note that samples 5-8 were treated with Wacker and
samples 1 and 5 were treated as control samples and therefore not shot at.
However, one has to be careful with analysis of mean sample data when investigating the specific impact of an
object within a stone surface. Therefore, the Piccolo data of samples 3 and 7 have been plotted in figure 4 to show
change during the various stages of the experiment. Here, two interesting trends emerge. Firstly, sample 3, which
was not treated with Wacker, does not show a clear impact area of weakening. This follows the trend of the
absorption of the impact by the more elastic nature of the surface (as compared to the Wacker treated surface).
However, after the environmental cabinet treatment there is a clear weakening of the impact area, as represented by
the lighter coloured area and relatively stable measurements around the impact area. This trend is reversed in sample
7 where initially the impact area is clearly identifiable post-mortem (the very light grey area that corresponds with
the location of the impact), accompanied by a general deterioration of the surface strength across the surface of the
block. This impact area is much less visible after the environmental cabinet treatment. Instead there appears to be a
general deterioration of the surface.
As can be seen in figures 3 and 4, all samples were negatively affected by the environmental cabinet treatment (EC1).
The presence or absence of a fracture network developed at the time of impact, as based on the research set out by
Momber (2004), therefore appears to be crucial in post-impact behaviour of the impacted surface. The fractured
surface appear to be more vulnerable to temperature changes overall, whereas a more plastic surface appears to cope

far better with the thermal stresses placed upon the samples in the environmental cabinet but suffered noticeable
degradation of the impact area, leaving it vulnerable to potential cavernous weathering. The pre-impact condition of
the surface is therefore instrumental in determining the impact of a projectile, such as a bullet, and projecting future
deterioration patterns.
Figure 4: Piccolo data of samples 3 (top) and 7 (bottom) in Leib value.
3.4 Moisture movement
Little is known about the relationship between internal moisture and impact areas of armed warfare, even though
internal moisture has been shown to be a driving weathering process (Mol and Viles, 2012). The surface hardness
experiment was therefore extended to include the behaviour of internal moisture under the conditions mapped out
previously. Both samples 3 and 7 were immersed in water at the foot to a depth of 2cm and will continue to be used
here as examples to illustrate the processes found. The photos used in figure 4 show the capillary water flow, as the
dark area at the foot of the stone indicates the rising wetting front. The research questions driving on this part of the
experiment are (i) how does the presence of a weakness within the surface of the stone affect moisture uptake and
evaporation and (ii) does the case hardening and ensuing network of fractures influence the movement of internal
moisture. It was already established that the environmental stress placed on the samples in the environmental cabinet
affected the surface strength of the impact area. Shockey (1975) conducted a study on the impact of projectiles on
steel and found that underneath the impact area a hemispherical area of deformation was visible. Since not many
impact tests have been performed on stone surfaces this model will be used, and initially the assumption will be
made that underneath the direct impact area a weakened hemispherical area will have developed that could influence
moisture movement, as its higher permeability would encourage moisture flow into this area.
Resistivity measurements were taken at three transects across the stone sample (for sampling strategy, see figure 1).
The middle transect covers the impact area and it was expected that the most noticeable changes would be seen in
this area. Figure 5 shows the distribution of moisture in sample 3 after the environmental cabinet treatment. The
areas of low resistivity (light greys and whites) assumed to indicate higher moisture content, whereas the dark grey
and black areas indicate high resistivity, assumed to be lower moisture contents. Moisture is seen to rise up until the
impact area where it evaporates from the surface. The impact area below the surface can also be distinguished as
moisture ingress appears to be more intense, forming a wedge shape of low resistivity into the stone body. At the
middle transect, the direct impact area shows a higher moisture content than in the surrounding area, suggesting that
the evaporation is indeed concentrated in this area, as previously suggested.

Figure 5: ERT transects of sample 3 after the environmental cabinet treatment
Sample 7 (see figure 6) shows a very different pattern; moisture is drawn through the face of the stone much more
evenly than the moisture in sample 3. The impact area, while visible, is less clearly defined. However, the impact
area behind the surface impact area clearly draws moisture in to a much greater degree than that seen in sample 3.
Towards the foot of the block the moisture appears to be drawn up rather than into the block, as visible in sample 3,
indicating more rapid capillary rise along the surface. Considering the noticeable decrease in surface hardness after
the impact it is likely that the fracture network at and near the surface facilitates this capillary rise. In addition, with
the lack of plasticity near the surface, as the Wacker can increase flexural strength within a stone mass (Pinto and
Rodrigues, 2008) which is useful when the stone is under physical pressure but decreases its ability to transmit
shockwaves with minimal damage, it is possible that the area directly behind the impact can severely weaken.

Figure 6: ERT transects of sample 7 after the environmental cabinet treatment
4. Discussion
The surface hardness and ERT measurements indicate that the impact of a projectile, such as a bullet, can greatly
alter the behaviour of a stone under changing environmental circumstances. While the Wacker was successful in
raising the surface hardness, the increase in strength appeared to be a disadvantage after the projectile impact as
surface strength was substantially lowered again. The non-treated samples did not show such surface-wide
deterioration but instead showed a noticeable lowering of surface hardness at the actual bullet impact site. This
resulted in very different moisture regimes as the treated sample appeared to conduct capillary rise of moisture along
the surface whereas the non-treated sample showed further ingress of moisture into the body of the stone. Placing
this in the wider context of fracture formation, figure 7 shows a simple conceptual model that indicates the
development of fractured, weakened, areas in both the treated and non-treated samples. Assuming that moisture

follows the available pathways, i.e. the fracture networks, a continued deterioration of the surface over time due to
increased evaporation and capillary rise could potentially be
Figure 7: A simple conceptual model showing the relationship between surface treatment and weakened areas after impact.
This brings with it a number of potential implications for stone conservation. Firstly, any monuments which have
been treated with consolidation substances or have developed case hardening during years of exposure could be at
greater risk of deterioration after a projectile impact due fracturing of the surface as a result of decreased plasticity.
While the damage on the surface may only be visible at the direct impact point, the actual weakening of the stone
may be far more widespread through the material than previously thought. In addition, the subsequent dramatic
difference in internal moisture behaviour could spell significant problems for the conservation of heritage as
previously dominant weathering processes could be altered to suit the new moisture regime. While the impact of
projectiles on surfaces and the subsequent formation of fracture networks is not a new concept, this research places it
in a new context by combining weathering studies and the impact studies that tend to be restricted to engineering
studies. As stated in the introduction, armed warfare is becoming an increasing threat to heritage as the availability
of weapons and their impact potential are increasing. These tests are based on relatively small bullet impacts and the
surface material loss was minimal yet the geomorphological impact was far greater than the eye would have
suspected. This research provides a small first step in understanding these mechanisms.
5. Concluding remarks
While this report gives a preliminary indication of the influence of projectile impacts on a stone’s response to
temperature and moisture ingress, further research is needed to comprehend this rather complex issue. Additional
factors such as salt weathering and loss of tensile strength throughout the stone mass need to be investigated further
before we draw firm conclusions regarding the influence of projectile impacts on conservation strategies in armed
warfare zones.
6. References
British Academy (2010) Fighting a looting battle? http://www.britac.ac.uk/news/news.cfm/newsid/16
Da Costa, D.M.R., Rodrigues, J.D. (2011) The effect of water on the durability of granitic materials consolidated
with Ethyl Silicates. Proceedings of CCI Symposium ICC 2011 – Adhesives and Consolidants for Conservation:
Research and Applications, Ottawa Canada 17-21 October
.
Fitzner, B., Heinrichs, K., La Bouchardiere, D. (2003) Weathering damage on Pharaonic sandstone monuments in
Luxor – Egypt. Building and Environment 38 (9-10) 1089 – 1103

Ghaidan, U. (2008) Damage to Iraq’s wider heritage. In: Stone, P.G., Bajjaly, J.F. (eds) The destruction of cultural
heritage in Iraq. Woodbridge, Boydell Press.
Hack, H.R.G.K., Hingira, J., Verwaal, W. (1993) Determination of discontinuity wall strength by Equotip and ball
rebound tests. <strong>International</strong> Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 30 (2) 151
– 155
ICOMOS-ISCS (2008)
http://www.international.icomos.org/publications/monuments_and_sites/15/pdf/Monuments_and_Sites_15_ISCS_Gl
ossary_Stone.pdf
McAlister, J.J., Smith, B.J., Curran, J.A. (2003) The use of sequential extraction to examine ion and trace metal
mobilisation and the case-hardening of building sandstone: a preliminary investigation. Microchemical Journal 74 (1)
5-18
Mol, L., Preston, P.R. (2010) The writing’s in the wall: a review of new preliminary applications of Electrical
Resistivity Tomography within Archaeology. Archaeometry 52 (6) 1079 - 1095
Mol, L., Viles, H.A. (2010) Geoelectric investigations into sandstone moisture regimes: Implications for rock
weathering and the deterioration of San Rock Art in the Golden Gate Reserve, South Africa. Geomorphology 118 (3
– 4) 280 – 287
Mol, L., Viles, H.A. (2012) The role of rock surface hardness and internal moisture in tafoni development in
sandstone. Earth Surface Processes and Landforms 37 (3) 301-314
Mol, L., Viles, H.A. (in press) Exposing drying patterns: using Electrical Resistivity Tomography to monitor
capillary rise in sandstone under varying drying conditions. Environmental Earth Sciences
Momber, A.W. (2004) Deformation and fracture of rocks due to high-speed liquid impingement. <strong>International</strong>
Journal of Fracture 130: 683 - 704
Moropoulou, A., Haralampopoulos, G., Tsiourva, Th., Auger, F., Birginie, M. (2003) Artificial weathering and nondestructive
tests for the performance evaluation of consolidation materials applied on porous stones. Materials and
Structures 36 (4) 201 – 217
Pinto, A.P.F., Rodrigues, J.D. (2008) Stone consolidation: The role of treatment procedures. Journal of Cultural
Heritage 9 (1) 38 - 53
Rodrigues da Costa, D.M., Rodrigues, J.D. (2011) The effect of water on the durability of granitic materials
consolidated with Ethyl Silicates. Proceedings of CCI Symposium ICC 17-21 October 2011, Ottawa, Canada.
Rossi-Manaresi, R., Tucci, A. (1991) Pore structure and the disruptive or cementing effect of salt crystallization in
various types of stone. Studies in Conservation 36 (1) 53-58
Sass, O., Viles, H.A. (2006) How wet are these walls? Testing a novel technique for measuring moisture in ruined
walls. Journal of Cultural Heritage 7 (4) 257 – 263
Shockey, (D.A.) Damage in steel plates from hypervelocity impact. I. Physical changes and effects of projectile
material. Journal of Applied Physics 46 (9) 3766 - 3775
Thiem, J. (1979) The Great Library of Alexandria Burnt: Towards the History of a Symbol. Journal of the History of
Ideas 40 (4) 507 – 526
Verwaal, W., Mulder, A. (1993) Estimating rock strength with the Equotip hardness tester. <strong>International</strong> Journal of
Rock Mechanics and Mining Science & Geomechanics Abstracts 30 (6) 659 – 662