issues in contemporary oil paint || new approaches to surface cleaning of unvarnished contemporary...

New Approaches to Surface Cleaningof Unvarnished Contemporary OilPaintings – Moist Sponges and Cloths

Beatriz Veríssimo Mendes, Klaas Jan van den Berg, Luc Megens,Ineke Joosten, and Maude Daudin

Abstract The properties and cleaning efficacy of five different sponges and twomicrofibre moistened cloths were studied on both naturally and artificially aged andsoiled oil paint surfaces. The cleaning efficacy of the materials were compared tothat of conventional cotton swab cleaning with water and saliva. All materials testedshowed good performance in terms of cleaning compared to swab rolling with wateror saliva. The tested moist isoprene make up sponge performed best of all materialson the artificeally aged and soiled surfaces.

Keywords Surface cleaning • Aqueous cleaning • Contemporary art •Microfibre cloths • Makeup sponges

Introduction

Dirt deposition on painted surfaces is a common problem for artworks resultingin the increase of light scattering on the surface and causing effects such as de-saturation of hue and luminosity decrease for both light and dark colours (Green1990).

In the case of unvarnished objects the dirt will adhere directly to the paintedsurface and in time it becomes embedded making its safe removal a common andcomplex problem for conservators (Perry 1990).

The nature of dirt depends greatly on the environment. Size and distributionrate of airborne contaminants in a museum environment varies with factors likeinternal and external environment, ventilation systems time of year and visitor’s

B. Veríssimo Mendes (�)Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, Lisbon, Portugal

Cultural Heritage Agency of the Netherlands, Amsterdam, The Netherlandse-mail: [email protected]

K.J. van den Berg (�) • L. Megens • I. Joosten • M. DaudinCultural Heritage Agency of the Netherlands, Amsterdam, The Netherlandse-mail: [email protected]

© Springer International Publishing Switzerland 2014K.J. van den Berg et al. (eds.), Issues in Contemporary Oil Paint,DOI 10.1007/978-3-319-10100-2__25

373

mailto:[email protected]

mailto:[email protected]

374 B. Veríssimo Mendes et al.

presence (Nazaroff et al. 1990). Deposition processes depend mainly on the natureof the dirt and the painted surface, particle size, as well climate parameters suchas temperature and relative humidity (Phenix and Burnstock 1990) (Nielsen et al.1978). In addition, physical characteristics such as texture, porosity, roughness,hardness and tack have a role in deposition of dirt (Nielsen et al. 1978).

The removal of dirt from an artwork is a complicated process and damages tothe surface such as abrasion and polishing, depend on the approach and cleaningmaterials used, rather than strictly professional skills (Hackney 2010). A commonsurface cleaning technique is swab rolling with water, water and additives orsolvent combinations (Wolbers 2000). However, conservators increasingly use newmaterials such as dry or moistened sponges and cloths.1 Due to smoothness and easeof handling, these materials are pursued as a useful addition to common practicecleaning materials. There is concern, however, about their durability, performancewhen used on unvarnished painted surfaces and potential extractable components.

Studies have shown the utility and limitations of some of these materials whenapplied dry to e.g. water sensitive painted surfaces (Daudin-Schotte et al. 2011).The aim of the present study was to observe the cleaning effect and surface changescaused by sponges and cloths used moist –combined with demineralised water –on different unvarnished oil paint surfaces. The results are evaluated in termsof material characteristics, ease of handling and cleaning efficacy using visual,microscopy and electron microscopy techniques (Morrison et al. 2007; Daudin-Schotte et al. 2011).

Materials Used in This Study

Cleaning Materials

The cleaning materials chosen are two sponges, three makeup sponges of differentcomposition and two microfiber cloths of different roughness listed in Table 1. Thecomposition and additives of these materials as well as their cleaning performancewhile used dry were previously studied.2

The polyvinyl alcohol – Blitzfix® and semi-vulcanised isoprene rubber SmokeSponge® as well as the microfire polyethylene therephthalate (PET) cloths are soldby conservation suppliers – the three makeup sponges are not sold as such.3

1Verbal communication with colleagues and conservators participating in the Dry Cleaning Work-shops of January and June 2010, Instituut Collectie Nederland, Hobbemastraat 22, Amsterdam.2http://www.cultureelerfgoed.nl/sites/default/files/documenten/drycleaning%20table.pdf.3Hema is a department store and Etos is a drugstore. Both are chain stores in the Netherlands,Germany and Belgium. There are new sponges on the market. ALthough most claim to be PUsponges testing is needed to determine any additives present.

http://www.cultureelerfgoed.nl/sites/default/files/documenten/drycleaning%20table.pdf

New Approaches to Surface Cleaning of Unvarnished Contemporary Oil. . . 375

Tabl

e1

Res

ults

ofch

arac

teri

satio

nof

the

clea

ning

mat

eria

ls,e

lem

enta

lana

lysi

s,co

mpo

sitio

n,ob

serv

atio

nsw

ithSE

M,w

ater

rete

ntio

nan

dha

ndlin

gab

iliti

esW

ater

rete

ntio

nw

asst

udie

dfo

rea

chm

ater

ialb

ypl

acin

ga

moi

stw

eigh

edm

ater

ialo

nbl

ottin

gpa

per

and

wei

ghin

git

agai

naf

ter

10m

in

Cle

anin

gm

ater

ials

(com

erci

alna

me)

Ele

men

tala

naly

sis

Che

mic

alco

mpo

sitio

naO

bser

vatio

n(S

EM

)W

ater

rete

ntio

nH

andl

ing

Blit

zfix

C,O

,Si

Poly

viny

lalc

ohol

(PV

AL

)CC

Eas

y

Star

chSm

ooth

Fille

rSo

ft

Den

se

Smok

eSp

onge

C,O

,Ca,

SIs

opre

neru

bber

�E

asy

Cha

lkD

ense

Tack

y

HD

Mak

eup

Spon

geC

,O,Z

nPo

lyur

etha

neE

ther

(PU

)�

Eas

y

Tin

uvin

Smoo

th

Zin

cO

xide

?So

ft

Flex

ible

Sem

iDen

se

(con

tinue

d)


Tabl

e1

(con

tinue

d)

Cle

anin

gm

ater

ials

(com

erci

alna

me)

Ele

men

tala

naly

sis

Che

mic

alco

mpo

sitio

naO

bser

vatio

n(S

EM

)W

ater

rete

ntio

nH

andl

ing

Hem

aM

akeu

psp

onge

C,O

,Zn,

K,S

,Si

Styr

ene

buta

dien

eru

bber

(SB

R)

C/�

Eas

yB

utyl

ated

Hyd

roxy

Tolu

ene

(BH

T)

Smoo

th

Die

thyl

dith

ioca

rbam

ate

Soft

Mer

capt

oben

zoth

iazo

leFl

exib

le

Zin

cO

xide

Eto

sM

akeu

psp

onge

C,O

,S,Z

n,Si

Isop

rene

rubb

erC

Eas

yB

utyl

ated

Hyd

roxy

Tolu

ene

(BH

T)

Smoo

th

Zin

cO

xide

Soft

Flex

ible

Yel

low

clot

h(m

ore

roug

h)C

,O,A

lPo

lyet

hyle

neth

erep

htha

late

(PE

T)

(pol

yest

er)

–N

ylon

6(p

olya

mid

e)C/

�D

iffic

ult

Smoo

th

Not

flexi

ble


Blu

ecl

oth

C,O

,Al

Poly

ethy

lene

ther

epht

hala

te(P

ET

)(p

olye

ster

)–

Nyl

on6

(pol

yam

ide)

C/�

Dif

ficul

t

Smoo

th

Not

flexi

ble

a http

://w

ww

.cul

ture

eler

fgoe

d.nl

/site

s/de

faul

t/file

s/do

cum

ente

n/dr

ycle

anin

g%20

tabl

e.pd

fA

perc

enta

geof

wat

erlo

ssw

asca

lcul

ated

.Wat

erre

tent

ion

was

cate

gori

zed

very

good

CC(l

ess

than

50%

),go

odC

(fro

m50

to60

%),

aver

age

C/�

(60

to70

%),

low

�(70

to80

%)

and

final

lyve

rylo

w�(

high

erth

an80

%)

http://www.cultureelerfgoed.nl/sites/default/files/documenten/drycleaning%20table.pdf


Ease of handling, structure, softness, smoothness and flexibility were assessedduring cleaning procedure, by touching and by observations with the microscopeand Scanning Electron microscopy (SEM).

The cleaning performance of the new materials was compared with that of cottonswabs, applied with water and saliva.

Paint Surfaces

For this study two types of paint surfaces were used. The first set exists ofartificially aged and soiled paint surfaces. The second set consists of a deaccessionedunvarnished monochromatic oil painting dating from 1977.

Set (I) consisted of Old Holland tube oil paints, painted out with flat and impastoareas on canvas boards.4 The paints consisted of cadmium yellow, cadmium red andultramarine blue in oil (Mills et al. 2008). Surfaces were cured and artificially lightaged at 10,000 lux at room temperature to the equivalent of 46 years of exhibitionin recommended conditions, 8 h a day 200 lux, assuming reciprocity. The paintsand their ageing regime were also used in the previous dry cleaning study (Daudin-Schotte et al. 2011).

The red and yellow paints were artificially soiled while the panels of cobalt bluewere left unsoiled for pilot tests. In order to maximise the variety of dirt, soilingprocedures included the application of particulate dirt collected from an old cellar5

with a microfiber cloth; followed by spraying on grimy dirt, prepared after a recipefor artificial dirt (Ormsby and Phenix 2010). The soiled surfaces were further aged,at a constant temperature of 50 ºC and relative humidity variations from 30 to 80 %every 6 h, for 21 days, to increase the dirt retention of the painting.

Each reference panel was divided in four. The remaining panels were divided intotwo, separated by a reference area that remained unsoiled throughout the project(Fig. 1).

The monochromatic green painting surface dated from 1977 and had been kept instandard storage conditions at RCE Collections in Rijswijk. The surface was smoothand dirt was scarce and relatively loosely bound, except for fingerprints on themargins. Infrared Absorption Spectrometry – Fourier Transform Attenuated TotalReflectance Infrared Spectrometry (FT-ATR-IR), X-ray Fluorescence Spectroscopy(XRF) analyses and light microscopy showed on the canvas a zinc white, bariumsulphate and chalk containing ground layer followed by a grey oil paint layer

4By Polly Saltmarsh in 2008.5At the former ICN, Gabriël Metsustraat 8. Light microscopy and SEM showed that the dirt wasmainly particulate inorganic matter conaining silica, calcium, sulphur and aluminium, indicatingthat much of the dust was related to cement and building material.


Fig. 1 Layout prepared panels where areas remained soiled and unsoiled throughout the studyand remaining areas were cleaned with cotton swab in the case of the reference panel (a) or newmaterials in case of remaining panels (b). In this image it is also visible the markers that allowedobservations in the exact same spots at different stages of the project

containing titanium white and carbon black. The top paint layer contained anunidentified possibly organic green pigment in oil.

The dirt on the panels and dirt applied on carbon tape was characterised usingSEM, XRF and light microscopy. Cross sections from each kind of painted surfacewere made and studied by microscopy.

Analytical Instruments

The light microscope used was a Zeiss Axioplan 2 with Zeiss neofluar lenses.Photomicrographs were taken with a Zeiss Axiocam MR with an Axiovision 4.8program. Raking light was provided by a Leica cold light source.

25� and 100� magnifications were used.A Scanning Electron Microscope with Energy Dispersive X-ray Detector (SEM-

EDX, JEOL 5910LV) was used in the secondary electron mode to examine the paintsurface without further preparation. Due to charging of the painted surface, lowranges of accelerating voltages between 1.7 and 5 kV were used. Magnificationsranged from 100� to 750� up to 1,000� or 3,000�. A copper tape marker wasglued on the surface of each area, in order to facilitate finding the same spot duringthe various stages of the project (Morrison et al. 2007).


Cleaning Tests

Cleaning tests were performed under ambient temperature and humidity. In order tobetter understand the mechanical action of the cleaning materials, without abrasioneffects caused by the dirt, pilot tests were made on unsoiled prepared panels ofcobalt blue.

All sponges were cut into pieces of approximately 2 cm2 and were appliedonto the surface using very light pressure and then slightly turned, minimizingmechanical action. The cloths were cut into small portions, wrapped around theindex finger and then applied on the surface in the same manner as described for thesponges. Cotton swabs on bamboo sticks were rolled on the surface.

To minimise the possibility of leaving organic residues on the paint surface(Daudin-Schotte et al. 2011), all sponges and cloths were rinsed with water andpressed in between sheets of tissue paper, leaving them hardly wet to the touch.While cleaning, the materials were frequently renewed. Clearance procedures werenot performed as only water was used. Results were assessed by visual observationswith the naked eye as well as light microscopy and SEM.

Evaluation Parameters

Surfaces were compared with the reference area of each panel and recorded inevaluation tables. The deaccessioned surface has no unsoiled original area; thereforethe cleanest area on the painting is used as a reference.

The evaluation is given semi-quantitatively, assigning marks ranging from 0 to10, where 10 represents the best result possible.

The overall cleaning effectiveness is determined based on previously establishedparameters (Daudin-Schotte et al. 2011), particularly visual changes to the paint sur-face, particulate residues and cleaning results. Visual change is the most significantof the three as it has the highest impact on the appreciation of an artwork, and isirreversible.

Visual changes are indicative of surface microscopic topography changes suchas flatenning, abrasion or polishing. This can cause an increase in gloss or createa matte surface. A patched surface is usually indicative of a combination of bothabrasion and polishing. Visual changes in this study refer to gloss changes dueto polishing and abrasion and are assessed with the naked eye complemented byobservations under the light microscope when needed and measurments with aglossimeter. Scratches to the pained surface can only be observed under the lightmicroscope or SEM.

Particulate residues, only observed under the light microscope, takes into accountthe number of residues found in 1 cm2 of the surface. Sometimes the residuescould be brushed of unless they were pressed into the surface. This parameter isto determine if there are residues left behind that cannot be seen with the naked eye.


Finally cleaning level refers to how clean and also how even a surface was. It wasassessed by the observation of the surface with the naked eye, light microscopy andSEM.

Cleaning materials were also assessed in terms of their handling and waterretention properties. Water retention was studied for each material by placing amoist weighed material on blotting paper. After 10 min the cleaning materials wereremoved from the blotting paper and weighed once more. Results for water retentionand handling are presented in Table 1.

Results & Discussion

Solvent Extractable Components

The amounts of solvent extractable components in the cleaning materials weredetermined for five different solvents: tap water, demineralized water, an aqueoussolution of 1 % triammonium citrate, 2-propanol and iso-octane. The cleaningmaterials were extracted with the solvents, and the weight of the extracted materialwas determined after evaporation of the solvent.

Results for the percentage of extractable components show that tap and dem-ineralised water produced the lowest amounts of extractable residue compared totriammonium citrate (TAC), and organic solvents. The highest percentage of extractsis observed for the TAC solution. After the solvent evaporation, the TAC extractionsleave behind a translucid gel-like substance.

The exact nature and composition of the extracts will be subject of further study.

Observation and Characterisation of the Artificialand Cellar Dirt

The combination of artificial and cellar dirt forms a whitish film over the surface.Observations with the light microscope and SEM on the cadmium yellow referencearea show a superficial, dry, grainy and heterogeneous layer, which is mainlyconcentrated in the impastos and ridges of the brush strokes. Analyses withEDX identify silica calcium, sulphur and aluminium. Minor peaks of sodium andpotassium are also observed. High contents of alumina-silicates are to be expectedas both types of dirt contain a considerable amount of cement and building dust.

Natural dirt on the monochromatic green surface and cross section was limitedand was not identified.

Pilot Tests

Pilot tests were performed on the unsoiled cobalt blue paint surface.


The smoke sponge had shown high percentages of extractable components(Table 1). In the tests it showed poor water retention and released crumbles ontothe surface. Therefore the material was concluded to be unsuitable and expelledfrom further tests.

The PVA blitzfix sponge indicated its potential for polishing a surface whereasthe PU, SBR and isoprene makeup sponges showed no noticeable changes to the sur-face. Finally pilot tests for both cloths show minor signs of abrasion and polishing.

Cleaning Tests

Degree of visual change observed mainly with the naked eye, shown in Fig. 2 wasassessed by comparison to reference surfaces.

The particulate residues were caused by the cutting of cleaning materials intosmaller pieces. These particulate residues were observed only under the lightmicroscope or with the SEM. Results for particulate residues can be seen in Fig. 3.

The cleaning level was assessed mainly with the naked eye and SEM. Generalobservations with microscope showed that dirt in crevices usually remains on thepainted surface after cleaning tests suggesting the materials cannot reach them.

Results for the cleaning level obtained are observed in the bar graph of Fig. 4.Cotton swabs with water and with saliva produced similar results on all surfaces

Figs. 2, 3, and 4. The degree of visual change is comparable, but slightly better forthe cotton swab with water Fig. 2. These materials left medium amounts of fibres onthe painted surface, and they only produced medium cleaning level. Observationswith the naked eye and microscope showed areas of dirt mainly concentrated on theimpasto areas, scratches (Fig. 5) and broken impastos (Fig. 6).

Cotton swabs are easy to use but the texture of this material is sometimes toorough and open, occasionally causing the fibres to grab impasto tops and breakthem off (Fig. 6).

The following images show a surface observed with SEM, before and aftercleaning the same area with cotton swab and saliva. These ilustrate the type ofsurface damage observed.

It is noted that all the cleaning materials produce varied degrees of visual changedepending on the type of surface (Fig. 2). All cleaning materials caused lowerdegrees of change on the deaccessioned surface (Fig. 2).

The blitzfix, SBR and the PU makeup sponges caused significant amount ofvisual change to the surfaces (Fig. 2). The blitzfix and the PU makeup sponge causedgloss increase whereas the SBR sponge had results ranging from gloss increase toa patched surface. The extent of visual change is possibly related to the materialsdensity lack of flexibility and water retention capabilities, as this might influencethe mechanical interaction between material and surface (see Table 1).

The isoprene makeup sponge produced good cleaning levels and created lessvisual change than any other material (Figs. 2 and 4). This makeup sponge alsoproduced better results when compared to the cotton swab with water and saliva.


Fig. 2 Test results showing the degree of visual change observed on the painted surface for allmaterials tested on prepared (above) and deaccessioned (below) surfaces. Results were rated from0 to 10, 10 being the best result possible, in this case, the least change to the painted surface.These results were obtained mainly by observations with the naked eye and glossimeter in terms ofgloss changes, and with the light microscopy in the impasto areas to observe possible topographydamages

Microfiber cloths produced low results on both the prepared and the deac-cessioned surfaces (Figs. 2 and 4). Observations under the microscope showedscratches and flattened impastos.

For all prepared surfaces it was observed that the flat portion of the surfaceswas easier to clean than the impastos. Most materials had difficulty in reaching thecrevices where dirt is deposited. Sponges and cotton swabs caused damages to theimpasto areas such as flattening or breaking of the impastos (Fig.6).

Material residues were to be found either loose on the surface or attached onflattened or broken impastos (Fig. 6). Of the new cleaning materials, the cloths left


Fig. 3 Test results of particulate residues observed for all materials tested on prepared (above)and deaccessioned (below) surfaces. Results were rated from 0 to 10, 10 being the best resultpossible, in this case, the least amount of particulate residues. These results were obtained mainlyby observations with light microscopy in a 1 cm2 area

the highest amounts of residues behind (Fig. 3). These result mostly from loosethreads after cutting the cloths; leaving the cloths intact should reduce the amountof residues.

It is important to note that the prepared mock-up surfaces used in this studywere relatively young, artificially aged and that the dirt on them was gritty. Themature monochromatic green paint was a well-cured and aged surface. All cleaningmaterials presented better cleaning levels on the mature monochromatic surface thanon the prepared surfaces (Fig. 4).

Dry cleaning tests performed for the makeup sponges and cloths on the maturegreen surface showed that all these materials produced better results when usedmoist than when used dry. Observations revealed that using the cleaning materialsused dry the surface was polished or abraded and dirt was not successfully removed.


Fig. 4 Cleaning level results on the prepared (above) and deaccessioned (below) surfaces. Resultswere rated from 0 to 10, 10 being the best result possible, in this case, the least change to thepainted surface. Results were obtained mainly by observations with naked eye, SEM and lightmicroscopy

In terms of surface damage, particulate residues and cleaning level, the goodresults obtained with the moist isoprene makeup sponge moist, were comparable toresults obtained while used dry (Daudin-Schotte et al. 2011).

Of all materials, the cloths produce the most abrasivon and polishing results.Test results for moist microfiber cloths show a higher amount of surface damageand lower cleaning levels, than dry cleaning results (Daudin-Schotte et al. 2011).Such a difference might be caused by an alteration in the action of these clothswhen they are moist, by the combination dirt/painted surface used, or by differencesof handling of the material by the assessors.


Fig. 5 Cadmium yellow reference sample before (a) and (b) after cleaning with cotton swaband saliva. The circle indicates recognizable surface topography that helps determine that bothobservations are in the same area of the painted surface. SEM images (SE mode) �110(a) and �140 (b) magnification

Fig. 6 (a) Example of scratches on the mature monochrome surface cleaned with the PETblue cloth. Image taken with (raking) light microscopy, �25� magnification. (b) Scratches onareas tested with cotton swab and saliva on Ultramarine Blue. SEM images (SE mode) �100magnification

Conclusions

The present study shows the potential of sponges used moist for the removal ofdirt in unvarnished painted surfaces. From all cleaning materials tested, the isoprenemakeup sponge shows the best performance. This makeup sponge is easy to handle,has good water retention and is flexible, adapting well to the painted surface. Resultsfrom cleaning tests demonstrate that for the prepared panels used in this study, moistmake up sponges perform better than cotton swabs with water or saliva.

PET cloths are the least successful, as they have a tendency to affect young andmature painted surfaces due to their abrasive action. The extent of polishing of asurface cleaned with a sponge is related to the density of this material.


The research of the cleaning materials has highlighted the issue of the com-position and additives present in the best performing materials. The problem withsponges from drugstores such as the ones used in study, is that the composition ofthese materials is not guaranteed by the producer. Although not tested in the presentstudy, the potential residues containing antioxidants and the extractable componentof the sponges are of concern. The present study indicates a high percentage ofextractable components for organic solvents. Further studies must be done in orderto determine the long term effect of extractable components on a painted surface.

The possibility of the extraction of potentially harmful additives of sponges byrinsing them with solvents; or the feasibility of the production of a sponge withoutharmful additives should be considered.

The authors are aware that the combination of dirt and artificially aged paintedsurfaces used in this study are not necessarily representative when compared tothe variability of “real, natural” dirt and naturally aged painted surfaces. Testsperformed with the same cleaning material show that the well cured and ageddeaccessioned surface, suffered less abrasion and polishing than the relatively youngalbeit light aged surfaces of the prepared panels. However, cleaning test results onthese surfaces give a good insight on the potential use of the cleaning materials onsurfaces which are young, have some surface textured and a grittier dirt layer.

Practical guidelinesBased on the results of this study, a few practical guidelines could be suggested.When facing an array of sponges, the softer and least dense isoprene sponge withthe lowest amount of additives in its composition should be chosen. To minimizepotentially harmful extractable component on the surface always rinse the cleaningmaterial. For all sponges it is advised to use pre cut pieces since cut pieces tend toleave more particulate residues on the surface. Smaller pieces should be cut out onlywhen the surface has some texture and cleaning has to be more localised. Abrasioncan be minimised by avoiding dragging the materials on the painted surface, insteadit is advised to gently dab the surface.

References

Daudin-Schotte M, Bisschoff M, Joosten I, van Keulen H, van den berg KJ (2011) Dry cleaningapproaches for unvarnished paint surfaces, Post prints cleaning 2010. Smithsonian InstitutionScholarly Press, Valencia

Green T (1990) Surface dirt removal from unvarnished paint films. In: Hackney S, Townsend J,Eastaugh (eds) Dirt and pictures separated. The United Kingdom Institute of Conservation,London, pp 51–55

Hackney S (2010) The art and science of cleaning paintings. In: Laura F-L, Elena Charola A,Marion M, Teresa Doménech-Carbó M (eds) Cleaning 2010 new insights into the cleaning ofpaintings. Universidad Politecnica de Valencia, Valência, pp 15–16

Mills L et al (2008) Water sensitivity of modern artists’ oil paints. ICOM preprints of the 15thtriennial meeting New Delhi, 22–26 Sept 2008

Morrison R, Bagley-Young A, Burnstock A, van den Berg K, van Keulen H (2007) An investigationof parameters for the use of citrate solutions for surface cleaning unvarnished paintings. StudConserv 52:255–270


Nazaroff W, Ligocki M, Ma T, Cass G (1990) Particledeposition in musuems:comparison ofmodeling and measurments results. Aerosol Sci Tech 13:332–348

Nielsen H, Kornum L, Säberg O (1978) Dirt retention on painted surfaces. J Coat TechnolOrmsby B, Phenix A (2010) Artificial dirt recipePerry R (1990) Problems of dirt accumulation and its removal from unvarnished paintings: a

practical review. In: Hackney S, Townsend T, Eastaugh N. United Kingdom Institue forConservation of historic and Artistic Works, London, pp 3–10

Phenix A, Burnstock A (1990) The deposition of dirt: a review of the literature, with scanningelectron microscope studies of dirt on selected paintings. In: Hackney S, Townsend J, EastaughN (eds) Dirt and pictures separated. The United Kingdom Institute for Conservation of Historicand Artistic Works, London, pp 11–18

Wolbers R (2000) Cleaning painted surfaces aqueous methods. Archtype Publications, London

issues in contemporary oil paint || new approaches to surface cleaning of unvarnished contemporary...

Documents