M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
1
Keywords: dry cleaning, non-contact 
cleaning, atomic oxygen, carbon diox-
ide snow, bread, soot, acrylic emulsion 
paint
AbstrAct 
Acrylic emulsion paints are sensitive to tradi-
tional cleaning with solvents and water. Dry 
cleaning methods offer an alternative, but 
there are few in-depth studies on their use. 
Selected dry cleaning methods (commercial 
erasers and sponges, and homemade bread) 
and non-contact methods (atomic oxygen 
and carbon dioxide snow) were tested on 
25-year-old acrylic emulsion paints and an 
acrylic emulsion gesso test panel from real 
and simulated fires. Water and solvents were 
used for comparison and as a link to scientific 
research on wet cleaning. Paint surfaces were 
examined visually and microscopically and 
analyzed by gloss meter and colorimeter. Wa-
ter, Absorene sponge and homemade bread 
removed soot on the acrylic gesso panel with-
out causing any visible surface alterations. 
Organic pigments in acrylic paints were more 
susceptible to removal by cleaning.
résumé
Les peintures à base d’émulsion acrylique 
sont sensibles au nettoyage traditionnel 
avec des solvants et de l’eau. Les métho-
des de nettoyage à sec constituent une al-
ternative, seulement il existe peu d’études 
approfondies sur leur usage. Une sélection 
de méthodes de nettoyage à sec (gommes 
à effacer et éponges du commerce, pain fait 
maison) et de méthodes sans contact (oxy-
gène atomique et neige carbonique) ont été 
testées sur des peintures à base d’émulsion 
acrylique âgées de 25 ans et sur un panneau 
test en gesso à base d’émulsion acrylique 
en situation d’incendies réels et simulés. De 
l’eau et des solvants ont été employés à titre 
de comparaison et pour faire le lien avec la 
recherche scientifique sur le nettoyage par 
Jia-sun Tsang*
Museum Conservation Institute, Smithsonian 
Institution
Suitland, MD, USA
TsangJ@si.edu
sara BaBo
Museum Conservation Institute, Smithsonian 
Institution
Suitland, MD, USA
sara.sbabo@gmail.com
*Author for correspondence
soot removAl from 
Acrylic emulsion pAint 
test pAnels:  
A study of dry  
And non-contAct 
cleAning
introduction 
Soot is carbon dust formed by gas-phase decomposition of hydrocarbons 
during fire. It is composed of ultrafine particles (~2.5 microns) smaller 
than most pigments (2.5 to 10 microns). Depending on the fuel source and 
composition, gas-phase soot may contain polycyclic aromatic hydrocarbons 
and aliphatic carbons. Fine soot particles can penetrate interstices and 
become embedded in painted surfaces (Figure 1). 
The challenges involved in removing soot from acrylic emulsion paint include 
the inherent softness of the paint and its sensitivity to water and solvent, 
the risk of removal of pigments, and extraction of soluble components. 
Research on the cleaning of acrylic emulsion paint has focused mainly on 
the effects of wet-cleaning (Ormsby 2009). With few exceptions (Estabrook 
1989, Saulnier 2005, Daudin-Schotte 2010), dry-cleaning methods have not 
been subjected to systematic research, perhaps because of the difficulties 
inherent in the use of erasers and sponges that require physical contact to 
remove dirt from painted surfaces. 
Recently, however, the potential benefits of emerging non-contact techniques 
have been investigated, including the use of carbon dioxide (CO2) snow 
and atomic oxygen. CO2 snow is used to clean the surface of optical and 
medical research samples. In the field of conservation, CO2 snow has 
been used to clean smoke-damaged books (Silverman 2006, 2009) and 
contemporary art objects (Shockey 2009). CO2 snow works by momentum 
transfer: the micron-sized snow particles collide with surface particulates 
and displace them. Solvency and freeze fracture can occur simultaneously 
with momentum transfer, improving the cleaning effect of CO2 snow. When 
CO2 particles hit a solid surface, the pressure produced can exceed triple-
point pressure, causing some of the solid CO2 to revert to liquid phase, 
which can absorb hydrocarbons via solvency. Temperature depression can 
also cause “freeze fractures” in the bonds between particulates and the 
surface, making particulates easier to remove (Applied Surface Technologies 
1996, Shockey 2009). 
Atomic oxygen cleaning exploits the high-reactivity of free atomic oxygen 
atoms, which can break carbon and hydrogen bonds, thereby removing 
carbon-rich substances by converting them into volatile species. Highly 
oxidized materials, such as metal oxides in pigments, are not affected. 
M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
 
soot removAl from 
Acrylic emulsion pAint 
test pAnels: A study of 
dry And non-contAct 
cleAning
2
voie humide. Les surfaces peintes ont été 
examinées visuellement et au microscope, 
et analysées à l’aide d’un brillancemètre et 
d’un colorimètre. L’eau, l’éponge Absorene et 
le pain fait maison ont décrassé le panneau 
de gesso acrylique sans causer d’altérations 
visibles en surface. Les pigments organiques 
contenus dans les peintures acryliques se 
sont montrés plus sensibles à l’élimination 
lors du nettoyage.
resumen 
Las pinturas de emulsiones acrílicas son 
sensibles a la limpieza tradicional con disol-
ventes y agua. Los métodos de limpieza en 
seco ofrecen una alternativa, pero hay pocos 
estudios detallados sobre su uso. Se hicieron 
pruebas con algunos métodos de limpieza 
en seco (borradores y esponjas comerciales, 
y pan casero) y métodos sin contacto (oxíge-
no atómico y nieve carbónica) en cuadros de 
emulsión acrílica de 25 años y en un panel 
de prueba de gesso y emulsión acrílica de 
incendios reales y simulados. El agua y los 
disolventes se utilizaron para hacer compa-
raciones y para relacionar el estudio con las 
investigaciones científicas sobre limpieza en 
húmedo. Las superficies de pintura se exa-
minaron visualmente y a nivel microscópico 
y se analizaron con un medidor del brillo y 
un colorímetro. El agua, la esponja Absorene 
y el pan casero eliminaron el hollín del panel 
de gesso acrílico sin provocar ninguna alte-
ración visible en la superficie. Los pigmentos 
orgánicos de las pinturas acrílicas eran más 
propensos a removerse con la limpieza.
Since 1994, Sharon Miller and Bruce Bank have used atomic oxygen to 
recover fire-damaged art works (Rutledge 2000, Miller 2005) and this 
technique is being further researched for new applications in museum 
conservation. The present study evaluates the effectiveness of dry cleaning 
on acrylic emulsion paint samples, using various techniques to examine 
and characterize the surfaces. 
experimentAl mAteriAls 
acrylic gesso panel 
An acrylic emulsion gesso panel was acquired from the remains of a 
1995 apartment fire. The panel contained both clean and sooty areas. 
Soot-covered strips removed from the bottom right portion of the panel 
were divided into 1 × 2 -inch (2 × 5 -cm2) pieces for systematic testing 
and evaluation (Figure 2). 
acrylic paint samples 
In 1995, sets of glass slides and gesso panels were painted with seven 
different heavy-body acrylic emulsion paints of known composition. The 
paints used were provided by Golden Artist Colors. Half the samples 
were painted with no visible brush marks, and the other half was painted 
in impasto. In 1996, one set of samples was exposed to a simulated fire, 
while the second set remained unexposed (Figure 3). The three paints 
used in this study were pigmented with pyrrole orange (NA), naphthol 
red light (PR-112), and cadmium yellow medium (PY-35). These organic 
and inorganic pigments are bright enough for dislodged matter to be easily 
detected. Only the thinly painted panels were tested because their surfaces 
were large enough to provide a separate section for testing each cleaning 
material, and for gloss meter and colorimeter evaluation. 
Commercial cleaning materials 
The authors tested seven commercial dry-cleaning products recommended 
by conservators with experience in treating smoke-damaged artworks: 
Rubgum eraser, Staedtler Mars plastic eraser, Groom/Stick cleaner, Absorene 
sponge, yellow Wishab sponge, white Wishab sponge, and Qosmedix 
make-up sponge (Table 1) (Roberts 1988, Spafford-Ricci 2000, Heydenreich 
2005, Levenson 2010).
Homemade bread 
Bread has traditionally been used by conservators to clean paintings with 
very sensitive surfaces (Tsang 2008). For each test, a fresh batch of bread was 
made, using only water, flour, yeast, and a little salt (no sugar or butter). 
Co2 snow 
CO2 snow is made by combining gaseous CO2 with dry nitrogen gas for 
moisture displacement. Tests were conducted by objects conservators Hugh 
 
Figure 1
Soot on the surface of an acrylic gesso 
panel, as seen with the Hirox microscope
M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
 
soot removAl from 
Acrylic emulsion pAint 
test pAnels: A study of 
dry And non-contAct 
cleAning
3
Shockey and Laura Kubick, using a dual gas unit located at the Lunder 
Conservation Center, Smithsonian American Art Museum (SAAM).
Table 1
Commercial dry-cleaning materials* tested
Name Type Composition
Rubgum eraser Block rubber Vulcanized vegetable oil
Staedtler Mars plastic eraser Block rubber Polyvinyl chloride, dialkyl phthalate, calcium 
carbonate, possibly titanium
Groom/Stick cleaner Moldable material Vulcanized polyisoprene, calcium carbonate
Absorene sponge Sponge Vulcanized rubber, calcium carbonate
Yellow Wishab cleaning sponge Block/Sponge Styrene butadiene rubber, vulcanized castor oil, 
antioxidant
White Wishab cleaning sponge Block/Sponge Styrene butadiene rubber, vulcanized castor oil, 
antioxidant
Qosmedix make-up sponge Sponge Latex-free polyurethane foam
*All items were acquired in 2010. Composition descriptions are based on information provided by manufacturers and 
other studies.3,4
atomic oxygen 
Atomic oxygen cleaning tests were conducted in 1996 by Sharon Miller 
and Bruce Bank at the National Aeronautics and Space Administration 
(NASA), using a small isotropic atomic oxygen chamber.
Water and solvents 
Convention cleaning materials such as water and solvents cleaning were 
tested to establish a baseline comparison for the dry-cleaning materials. 
Deionizer water, anhydrous ethyl alcohol (ethanol), and Stoddard solvent 
were used.
testing procedures 
Commercial dry-cleaning products were tested in two phases on soot-covered 
areas of the gesso panel described above. In the first phase, each sample 
was stroked ten times in one direction with an eraser or sponge, exerting 
mild pressure by finger. In the second phase, each sample was stroked 
until no further improvement could be observed. Bread was kneaded and 
tested in a similar manner. This approach allowed the authors to observe 
the efficacy of each method, employed at different pressures for varying 
amounts of time on the surfaces. CO2 snow tests were also performed 
in two phases for comparison with the dry cleaning tests with erasers 
and bread. Phase one consisted in a shorter period of cleaning, about 30 
seconds, which was considered a regular treatment for that type of surface 
by the conservators. Phase two of CO2 cleaning was aimed until no further 
improvement of soot removal was observed. 
The first phase test was repeated using flat acrylic paint samples not exposed 
to fire, to evaluate the effects of the cleaning treatments. Materials that 
crumbled during testing were collected and saved for observation. Some 
cleaning methods were tested on soot-covered impasto samples to study 
2 
 
 
              
 
 
Figure 2. Soot-covered acrylic gesso panel with diagram of cleaning materials tested. A: water. B: ethanol. C: 
Stoddart solvent. D: bread. E: Rubgum eraser. F: Groom/Stick G: Absorene sponge. H: white Wishab sponge. I: 
yellow Wishab sponge. J–N: water with several detergents. O: Staedtler Mars plastic eraser. P: Qosmedix make-
up sponge. Q: CO2 snow. R: atomic oxygen. So: soot-covered area. Wh: white area (no soot).  
  
R 
Wh So 
A D B C E F G 
N O 
P 
Q 
H I 
J K 
L M 
 
Figure 2
Soot-covered acrylic gesso panel with 
diagram of cleaning materials tested. A: 
water; B: ethanol; C: Stoddart solvent;  
D: bread; E: Rubgum eraser; F: Groom/Stick; 
G: Absorene sponge; H: white Wishab sponge; 
I: yellow Wishab sponge; J–N: water with several 
detergents; O: Staedtler Mars plastic eraser; 
P: Qosmedix make-up sponge; Q: CO2 snow; 
R: atomic oxygen; So: soot-covered area; 
Wh: white area (no soot)
Figure 3
Acrylic emulsion paint test samples on 
commercial gesso ground. Right column: 
smoothly brushed. Left column: impasto
M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
 
soot removAl from 
Acrylic emulsion pAint 
test pAnels: A study of 
dry And non-contAct 
cleAning
4
their practicality and effectiveness in cleaning uneven surfaces. Crumbs 
were cleaned with a soft brush, which was cleaned between tests. 
Atomic oxygen cleaning was tested on these same samples in 1996, as 
previously described (Miller 1997). The cleaned samples from that test 
were further investigated in this study. The samples cleaned in 1996 were 
stored in clean and stable environmental controlled museum facility. The 
RH was kept at 45–55% and temperature was kept at 68–70°F. A control 
strip (Wh area, Figure 1), where no cleaning test was done, was kept on 
the test panel and used as blank control for gloss and color measurements. 
The cleaning tests were established as relative comparison, not absolutely 
comparison. 
Water and solvent tests were done by dampening the cotton swabs in 
water and solvents, a conventional cleaning method used by paintings 
conservators. 
instrumentAtion 
All samples were examined under a Wild Heerbrugg microscope and 
photographed at x25 magnification with a Leica camera attached to the 
microscope, using Leica LAS EZ V1.5.0 software. Gloss was measured 
with a Gardner micro-TRI-gloss meter at 85°, the recommended angle 
for studying matte samples (GlossMeters 2006). However, since 85° is 
not a natural viewing angle, gloss was measured at 20° and 60° as well. 
Color was measured with a Minolta CR-300 chroma meter, using CIELAB 
color space coordinates. Both gloss and color measurements were taken 
three times in each sample/method/phase, and the average was calculated. 
Selected samples were also observed with a KH-7700 Hirox 3Ddigital 
microscope at SAAM.
results 
Visual observation and microscopy 
Surfaces of the acrylic gesso and acrylic emulsion paint samples were 
examined with the naked eye and under microscope, before and after 
cleaning, to evaluate the effectiveness of each cleaning method and any 
resulting alterations in surface integrity. Special attention was paid to 
surface defects potentially caused directly by abrasion from erasers or 
indirectly by pressure from CO2 or atomic oxygen treatment. 
The effectiveness of the cleaning methods used on the gesso panel was 
evaluated for evenness of soot removal, amount of soot removed, and shifting 
and redistribution of soot. The cleaning methods that most effectively 
removed soot from both peaks and valleys in the paint surfaces, in descending 
order, were: atomic oxygen, water, Absorene sponge, bread (Figure 4), and 
Qosmedix make-up sponge. The Staedtler Mars plastic eraser, Rubgum 
eraser, ethanol, and Stoddard solvent removed soot from peaks, but not 
from valleys. A Hirox image of sample E cleaned with a Rubgum eraser 
clearly illustrates these uneven results (Figure 5). CO2 snow removed soot 
 
 
Figure 4
3D image of gesso cleaned with homemade 
bread showing clean peaks and valleys
Figure 5
3D image of dry-cleaning with gum eraser 
showing clean peaks and sooty valleys
M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
 
soot removAl from 
Acrylic emulsion pAint 
test pAnels: A study of 
dry And non-contAct 
cleAning
5
only in the second phase of treatment. Ethanol, atomic oxygen, and CO2 
damaged the surface of the gesso panel while removing the soot. Ethanol 
removed part of the gesso layer, atomic oxygen removed the superficial 
acrylic binder, and CO2 snow created impact holes in the second phase 
of cleaning (Figure 6).
There were no visible changes in the surface of acrylic paint after cleaning 
with commercial products and bread. Crumbs from each cleaning were 
saved and examined microscopically. It was noted that PR-112 made 
with organic pigment was extremely sensitive to cleaning, and that red 
pigment was picked up and deposited in the crumb. Staedtler Mars plastic 
eraser removed more original material than other cleaners. CO2 cleaning 
caused no damage observable with the naked eye to acrylic paint surfaces; 
however, small impact holes were seen microscopically, indicating loss of 
original matter. Atomic oxygen cleaning caused visible dulling of paint 
surfaces, possibly indicating loss of binder.
Color and gloss 
Soot removal caused color changes in the gesso panel samples relative to 
increases in the L* coordinate, which corresponds to lightness (Table 2). 
Since the original color of the panel was white, greater L* values (and 
consequently greater ∆L* and ∆E) indicate whiter or, in this case, cleaner 
samples. This parameter was used to rank the cleaning efficacy of the tested 
methods, which, in descending order, were: atomic oxygen, Staedtler Mars 
plastic eraser, water, Absorene sponge, bread, and Rubgum eraser. Wishab 
sponges and Groom/Stick cleaner were less effective in soot removal. 
Samples treated with atomic oxygen surpassed the original L* value, 
becoming lighter than the original. This may be an indication of surface 
alterations, as suggested by other tests performed in the study.
Table 2
Changes in color during two cleaning phases. The greater changes are noted in ∆L values
Sample Phase I Phase II
∆L ∆a ∆b ∆E ∆L ∆a ∆b ∆E
A 3.82 0.13 0.93 3.93 10.47 -0.14 0.43 10.48
B -0.86 0.07 0.13 0.88 8.38 -0.10 -0.12 8.38
C 0.03 0.06 0.27 0.28 6.59 0.07 1.05 6.67
D 5.58 -0.08 0.54 5.61 10.18 -0.13 0.16 10.18
E 4.67 -0.04 0.44 4.69 9.75 -0.14 0.06 9.75
F 1.67 0.01 0.08 1.67 2.30 -0.01 0.13 2.30
G 5.00 0.07 1.24 5.15 10.26 -0.14 0.57 10.28
H 0.75 0.04 0.55 0.93 2.00 0.08 0.74 2.13
I 2.01 -0.07 0.33 2.04 3.26 -0.04 0.47 3.29
O 7.16 -0.08 0.36 7.17 11.44 -0.20 0.18 11.45
P 5.27 0.10 1.12 5.38 9.66 -0.02 1.04 9.72
Q 2.10 0.27 1.33 2.50 10.41 -0.12 0.86 10.45
R - - - - 13.51 -0.14 -1.33 13.58
In general, increased gloss value is an indicator of cleaning efficiency. 
In these tests, gloss changes were very low and not significant and the 
 
Figure 6
3D image of gesso panel cleaned with CO2 
showing surface holes caused by impact
M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
 
soot removAl from 
Acrylic emulsion pAint 
test pAnels: A study of 
dry And non-contAct 
cleAning
6
values obtained were sometimes inconsistent (Table 3). However, all but 
one of the gesso panel samples showed an increase in gloss after cleaning. 
The exception was the sample cleaned with atomic oxygen, where values 
measured at a 60° angle were inferior to those of the white control, possibly 
due to loss of binder during treatment.
Table 3
Gloss measurements at 20°, 60°, and 80°, after cleaning with materials described in Figure 2
Sample 20o 60 o 85 o
Bef Aft Aft2 Bef Aft Aft2 Bef Aft Aft2
A 0,80 1,00 1,17 2,17 2,70 3,23 0,50 0,87 0,90
B 0,80 0,80 1,00 2,27 2,20 2,83 0,50 0,70 0,70
C 0,80 0,80 1,00 2,30 2,50 3,40 0,70 1,10 1,90
D 0,80 1,00 1,20 2,50 3,30 3,83 0,73 1,20 1,20
E 0,80 1,00 1,20 2,53 3,30 3,90 0,83 1,40 1,60
F 0,80 0,90 0,90 2,43 2,87 2,97 0,67 1,00 1,00
G 0,80 1,00 1,20 2,57 3,57 4,17 0,63 0,97 1,17
H 0,80 0,80 0,83 2,50 2,70 2,83 0,70 0,93 0,97
I 0,80 0,93 1,00 2,37 3,00 3,03 0,77 1,13 1,03
O 0,80 1,00 1,20 2,57 3,27 3,77 1,13 1,93 2,03
P 0,80 1,00 1,10 2,50 3,40 3,90 1,20 1,83 2,17
Q 0,80 0,80 - 2,3 2.9 - 0,7 1,00 -
R - - 1,16 - - 1,97 - - 0,67
Regarding the acrylic paint samples, atomic oxygen cleaning caused changes 
in color and gloss on samples painted with organic pigments (Figures 7 
and 8). Color changes in samples of paints NA and PR-112 resulted from 
paint loss during cleaning and exposure of the canvas support surface. As 
noted previously, the decrease in gloss was probably caused by loss of binder. 
Paints with organic pigments are more sensitive to atomic oxygen cleaning 
because both binder and pigment can react with the oxygen and transform 
into volatile compounds. No significant changes in color and gloss were 
observed on any other acrylic paint samples as a result of cleaning.
 
Figure 7
Color changes (∆E) in three acrylic color samples. Significant changes were observed in the organic 
samples (NA and PR-112) cleaned with atomic oxygen
M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
 
soot removAl from 
Acrylic emulsion pAint 
test pAnels: A study of 
dry And non-contAct 
cleAning
7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Figure 8
Changes in gloss after cleaning. Note significant change in inorganic sample treated with atomic oxygen 
(values unavailable for organic samples, which were covered with varnish)
conclusions 
Dry and non-contact methods of removing soot from painted surfaces were 
evaluated by visual observation and microscopy, and by measuring gloss 
and color. Our goal was to increase conservators’ options for cleaning 
acrylic emulsion paint. Our criteria for evaluating the effectiveness of 
a particular cleaning method were preservation of surface morphology 
and homogeneity, evenness of cleaning, and ease of soot removal. Major 
considerations in selecting cleaning methods include the type of soil to 
be removed, the chemical and physical properties of the surface to be 
cleaned, the impact of mechanical forces, and maintenance of surface 
integrity. Cleaning with water, Absorene sponge and homemade bread 
delivered superior results in the removal of soot from an acrylic gesso panel. 
However, they also removed the organic pigment from PR-112 acrylic 
emulsion paint, with bread doing so more slowly than water. It is known 
that water can remove surfactants on acrylic paint, but it appears that bread 
may also remove surface components. This calls for further investigation. 
These results show that ethanol caused noticeable and significant damage 
to the paint’s surface integrity, clearly indicating that solvents should 
not be used to clean acrylic emulsion paint. It was also found that in the 
process of dry-cleaning, paint with inorganic pigments fared better than 
paint with organic pigments, which can be removed from paint by both 
dry and wet-cleaning. 
Disadvantages were discovered to both dry and non-contact cleaning 
methods. Dry cleaning is time-consuming and mostly offers uneven results. 
CO2 does not effectively remove soot from acrylic emulsion paint, probably 
due to surface chemical interactions between the paint and the polycyclic 
aromatic hydrocarbons, aliphatic carbon, and other oxygen-rich groups in 
soot. To remove soot, the CO2 had to be applied at pressures that damaged 
paint surfaces and created impact holes. Atomic oxygen cleaning removed 
soot evenly without disturbing brush strokes, but all paint samples cleaned 
with atomic oxygen showed a decrease in gloss due to removal of a portion 
M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
 
soot removAl from 
Acrylic emulsion pAint 
test pAnels: A study of 
dry And non-contAct 
cleAning
8
of the acrylic binder, which is ethically unacceptable. Nevertheless, atomic 
oxygen cleaning can potentially salvage heavily soot-damaged artworks 
previously considered unrestorable. 
Choosing the right tool can minimize the risk of inappropriate cleaning, and 
more tests will be conducted on the 25-year-old acrylic test samples. 
AcKnowledgements 
The authors thank Hugh Shockey, Laura Kubick, Sharon Miller and Bruce 
Bank for performing the CO2 snow and the atomic oxygen cleaning tests, 
respectively. Sara Babo thanks Calouste Gulbenkian Foundation and 
Luso-American Development Foundation for providing the grant for the 
fellowship at MCI.
references 
APPLIED SuRfACE TEChNOLOgIES. 1996. Carbon dioxide snow cleaning. http://
www.co2clean.com/index.html (accessed 12 October 2010).
DAuDIN-SChOTTE, M., M. BISSChOff, I JOOSTEN, h. vAN KEuLEN, 
and K. JAN vAN DEN BERg. 2010. Dry cleaning approaches for unvarnished paint 
surfaces. In Cleaning 2010 – new insights into the cleaning of paintings, ed. L. Fuster-
López, A. Charola, M. Mecklenburg, and M. Doménech-Carbó, 51–53.Washington DC: 
Smithsonian Institution Press.
ESTABROOK, E. 1989. Considerations of the effect of erasers on cotton fabric. Journal 
of the American Institute for Conservation 28: 79–96. 
gLOSSMETERS: YOuR gLOSS AND COLOR ExPERTS. 2006. http://www.
gloss-meters.com/GlossIntro.html (accessed 12 October 2010).
hEYDENREICh, g. 2005. Fire, water, air, and more than 1000 square meters of soiled 
surfaces to be cleaned. AIC Paintings Specialty Group Postprints 17: 32–38.
LEvENSON, R. 2010. Some observations: removing soot residue from contemporary 
paintings. AIC News 35(5): 4.
MILLER, S., and B. BANKS. September 1997. Results of smoke exposure and atomic 
oxygen cleaning of acrylic paints on gesso panel and slides provided by the Smithsonian 
Institution (SCRME). In-house report, SCRME, Washington, DC, and NASA Glenn Research 
Center, Cleveland, Ohio.
MILLER, S., B. BANKS, and D. WATERS. 2005. Atomic oxygen treatment and its 
effect on a variety of artist’s media. NASA/TM 2005-213434: 1–6.
ORMSBY, B., and T. LEARNER. 2009. The effects of wet surface cleaning treatments 
on acrylic emulsion artists’ paints: A review of recent scientific research. Reviews in 
Conservation 10: 29–41.
ROBERTS, B., C. vERhEYEN, W. gINNEL, S. DERELIAN, L. KROWECh, 
T. LONgYEAR, B. MILAM, L. STRAuSS, D. SILguERO, R. TANK, and J. 
gREAvES. 1988. An account of the conservation and preservation procedures following 
a fire at the Huntington Library and Art Gallery. Journal of the American Institute for 
Conservation online, 27(1): article 1. http://cool.conservation-us.org/jaic/articles/jaic27-
01-001_indx.html.
RuTLEDgE S., B. BANKS, M. fORKAPA, T. STuEBER, E. SEChKAR, and K. 
MALINOWSKI. 2000. Atomic oxygen treatment as a method of recovering smoke-damaged 
M
O
D
E
R
N
 M
A
T
E
R
IA
LS
  
A
N
D
 C
O
N
T
E
M
P
O
R
A
R
Y
 A
R
T
 
soot removAl from 
Acrylic emulsion pAint 
test pAnels: A study of 
dry And non-contAct 
cleAning
9
paintings. Journal of the American Institute for Conservation online 39(1): article 5. http://
cool.conservation-us.org/jaic/articles/jaic39-01-005_indx.html.
SAuLNIER, g., and M. ThIBAuLT. 2005. Cleaning acrylic emulsion paint: a two-part 
study. AIC Paintings Specialty Group Postprints 17: 1–8.
ShOCKEY, L. h. 2009. Blow it off: Moving beyond compressed air with carbon dioxide 
snow. AIC Objects Specialty Group Postprints (in press).
SILvERMAN, R. 2006. Fire and ice: a soot removal technique using dry ice blasting. 
International Preservation News 39: 20–29.
SILvERMAN, R., and S. IRWIN. 2009. Fire and ice revisited: A comparison of two 
soot removal techniques for books. International Preservation News 49: 31–35.
SPAffORD-RICCI, S., and f. gRAhAM. 2000. The fire at the Royal Saskatchewan 
Museum, part 2: removal of soot from artifacts and recovery of the building. Journal of 
the American Institute for Conservation online 39(1): article 3. http://cool.conservation-us.
org/jaic/articles/jaic39-01-003_indx.html.
TSANg, J. 2008. Smithsonian Museum Conservation Institute. Conservation Report 
MCI #6192.1.