Metal Soaps

The problem of metal soap aggregates in paintings put the role of metal coordinated fatty acid and diacids in the stabilisation of paintings on the map. The phenomenon was studied by the MOLART research team when they collaborated with the restorers team of Jorgen Wadum and Petria Noble at the Mauritshuis in The Hague. The Anatomy Lesson of Dr Nicolaes Tulp, a painting by Rembrandt van Rijn, had undergone many restoration treatments, especially a series of hot lining treatments.

Close examination with microscopes and microscopic study of the Xray pictures showed a very large number of protruding masses which were found to consist of lead soaps of palmitic and stearic acid. Now about 12 years later metal soap aggregates have been found to be present in thousands of paintings. Still many questions remain with respect to their origin inside the paint layers, their mobility and their stability. They remain a challenge for the restorer.

msoap nr 1

When my research group started analysing paints after direct derivatisation with silylating agents in 1993, we were stunned to discover that practically all fatty acid moieties were in a free or salt/soap state. The original ester bonds were gone. Clearly the idea that an mature aged oil paint was a viscous mass of cross-linked polyunsaturated vegetable oils was wrong. On the other hand, how could we explain that paintings could become so old and survive the centuries while their appearance was still reasonable or even good considering their age. Some self repair mechanism had to be in place. We realized that an oil painting after about one century was not an oil painting anymore but an metal coordinated network of diacids with some monocarboxylic acids attached. The metals were delivered by omnipresent transition metals like lead in traditional oil paintings.

My paper in 1997 printed in Early Italian Paintings: Techniques and Analysis suggested this idea and proposed that stationary and mobile fractions were present in paintings. After several years of research two of my graduate students and I presented what we knew about oil paint structure, metal soaps and reactivity of free fatty acids with some pigments at the AIC symposium in Portland in 2006.

Below the full text that is also printed in Volume 19 of the Postprints of the Paintings Group. It can be downloaded from www.amolf.nl/pubications:
Jaap J. Boon, Frank Hoogland and Katrien Keune, Chemical processes in aged oil paints affecting metal soap migration and aggregation In: AIC paintings specialty group postprints : papers pres. at the 34th annual meeting of the AIC of Historic & Artistic Works providence, Rhode Island, June 16-19, 2006 /ed. H. Mar Parkin. – Washington: AIC, 2007. – pp. 16-23 (AIC PSG Postprints; volume 19).

 

Chemical Processes in Aged Oil Paints affecting Metal Soap Migration and Aggregation

Jaap J. Boon, Frank Hoogland and Katrien Keune

Molecular Material Science of Art (MOLART)AMOLF, Kruislaan 407, 1098 SJ Amsterdam, The Netherlands. boon@amolf.nl

ABSTRACT – The nature of the surface of the painted picture is determined to a large extent by chemical and physical processes underneath the surface of the paint. The binding medium changes from a chemically air drying viscous mass of polyunsaturated triglycerides “the oil” via slow hydrolytic processes to a metal bound ionomer. This network system in turn appears to be vulnerable to further environmental attack especially by acidification that is postulated to disrupt the ionomeric structure. As a result monocarboxylic fatty acids can be mobilized and reorganize in the form of liquid crystalline metal soap masses within the paint layers. When they expand beyond the paint layer, they distort the paint and may erupt at the surface where they protrude or even extrude. These deformations are now recognized in thousands of paintings. Many of these metal soap masses furthermore mineralize forming minium (lead orthoplumbate) and/or lead-hydroxy/chloride-carbonates in lead soap or zinc carbonates in the case of zinc soaps. Generally the volume changes due to these processes are minimal although the metal soap may grow into larger masses of 100-200 micron diameter. This theory could be developed by an integrated molecular level approach involving GCMS (gaschromatography mass spectrometry) and DTMS (Direct Temperature resolved mass spectrometry) work and various forms of chemical microscopy using imaging SIMS (secondary ion mass spectrometry), FTIR (Fourier Transform Infrared spectroscopy), Raman spectroscopy and Scanning electron microscopy with energy dispersed Xray analysis (SEM-EDX).

Introduction

Before a painting becomes a finished picture, it has to dry physically by evaporation of water or solvent, and chemically by reaction of air with the binding medium constituents to form a non-sticky viscous mass that keeps the pigment particles in place. So when the painter is satisfied with the picture, a lot of processes are starting up to solidify the painted creation. These same processes however are also responsible for undesirable changes that take place later on when the paint ages. These changes will require the attention of conservators in charge of maintaining the quality of the visible surface and the structural integrity of the paint layers and support. Research on the nature of these materials in paintings and their compositional changes with time i.e. the molecular aspects of ageing was at the centre of the attention of the MOLARTprogramme supported by NWO and its sequel the De Mayerne Programme. Both research programs were multidisciplinary in nature and were successful because of the close collaboration of art technical historians, conservators and art scientists interested in material science aspects of works of art. This paper presents ideas on ageing oil paint and summarizes various aspects of the changing nature of the oil paint network in oil paintings. It especially points to some of the defects caused by chemical changes in the network which affect the stability of the metal coordinated structure, the mobility/reactivity of free fatty acids and the formation of metal soap liquid crystals. Phenomenological aspects of metal soaps presented by Noble and Boon in this postprint volume show the distortions in the paint layers due to metal soap aggregation and swelling, and other aspects of metal soap reactivity especially in surface layers that affect the transmission and reflection of light. An account of the chemical aspects of metal soaps in cross sections from a number of type-paintings has been given in Keune’s PhD dissertation (2005) and in a forthcoming paper in Studies in Conservation (Keune and Boon, 2007).

Dia2

Fig. 1 Light microscopy, Scanning electron microscopy with back scattered electron detection (SEM-BSE) and Fourier transform infrared (FTIR) spectroscopy of a double ground from The Anatomy Lesson of Dr.Nicolaes Tulp painted by Rembrandt van Rijn in 1632 (MH 146). The cross section represents an early stage of lead soap formation in which smaller lead white particles are seen to dissolve and transparent lead soap containing areas appear.


Reactivity of oil binding medium components

The awareness of the wide spread global occurrence of metal soap aggregates in paintings has developed in less than a decade. The restoration of The Anatomy Lesson of Dr. Nicolaes Tulp by R. van Rijn (pixit 1632; MH 146) was the start of a thorough investigation of the painting technique and its present condition (Middelkoop et al. 1998). The abundance of peculiar pustules was particularly striking. These had been reported before but they were attributed to fire damage that the painting endured in 1723 (De Vries et al. 1978). Investigation by imaging FTIR and SIMS microscopy demonstrated that these organic aggregates were composed of lead soaps (Heeren et al. 1999). The lead soap aggregates were present in large numbers as could be deduced from X-ray pictures examined under the stereomicroscope. FTIR imaging was especially useful in the detection of lead soaps in paint cross section as is demonstrated in Fig. 1 which shows a collage of light microscopic and SEM-BSE from a cross section of MH 146 in an early stage of metal soap formation and the corresponding FTIR maps of fatty acyl moieties (2929 cm-1), lead carboxylates (1520 cm-1) and carbonates (1400 cm-1). In this cross section the lead white is affected and partially dissolved leading to large transparent regions in the lead white layer of the typical double ground employed by Rembrandt in that period. In more advanced stages, where lead soaps aggregate to form larger semi crystalline structures, the upper paint layer is often flacking off (see also Noble et al, 2002). These aggregates turned out to be present in more paintings of the Mauritshuis and are now recognised in many other collections as well (Higgitt et al. 2003; Noble and Boon, 2007). A recent survey of the Tudor-Stuart collection of Tate Britain presented during the AIC meeting 2006 shows lead soap aggregates in more than 60% of the paintings (Jones et al. 2007).

AIC-Fig-2

Fig. 2 Lead soap can form at ambient temperature by reaction of lead compounds and palmitic acid. Lead soap formation was monitored by FTIR spectroscopy (Lead soap (%) = 100% * I(1510 cm-1) / [I(1510 cm-1) + I(1705 cm-1)]). Leadacetate (A); Lead white (B); Litharge (C); Minium (D);Leadtin yellow I (E)

The occurrence of lead soap aggregates points to an oil paint defect. The MOLART working model on aging of oil paint (Boon et al. 1997) proposed a transition from a cross linked plant oil to a metal coordinated paint system, because a rather extensive hydrolysis had been observed of the biological ester bonds in the cross linked polyunsaturated triglycerides. Investigations by Van den Berg et al. (1999) on oil paint models and samples from paintings and recent studies on alkyd oil paints (Schilling et al. 2007) demonstrate the extensive hydrolysis that takes place in a time frame of 50-100 years. The process of degrouping of the cross linked oil components into their biochemical components of glycerol, fatty acids, diacids and possibly some cross linked moieties with multiple acid groups is now an accepted fact in the life time of oil paintings. We expected that this process would lead to a complete failure of the paint and loss of the picture. That this usually does not happen is caused by a self-repair mechanism in which lead binds these loose ends with limited volume change into a metal coordinated network. The hypothesis that the acid groups rapidly react with available metals present as driers or metal containing pigments like lead white was supported recently by NMR studies of reconstructed and aged oil paint reported by Verhoeven et al. (2006), by earlier FTIR and SIMS of similar reconstructions (Keune, 2005) and by chemical experiments. The NMR work suggests that acid groups forming from unsaturated C18 fatty acyl moieties in the stage of oxidative cross linking of the oil readily form metal soaps. In the case of diacids forming, their biological acid group can be still linked with the original ester bond to the network, whereas the other acid group is already present as metal soap

AIC-Fig-3

Fig. 3 Secondary Ion Mass Spectra of lead white paints (coded ZD prepared and aged by Carlyle in 1999) before and after a 30 day exposure to high relative humidity (80%) at elevated temperature (50 oC). See also Keune (2005).

Acid groups in oil paint readily react to form metal soaps. Earlier Van der Weerd (2002) reported the dissolution of lead white in hot solutions of xylene to support the observation lead soap formation at the expense of lead white in ageing oil paint. More recent experiments with palmitic acid reacting with lead white in chloroform solution at room temperature – a more realistic scenario – demonstrate complete dissolution of lead white forming lead soaps (Hoogland and Boon 2005). Fig. 2 shows lead soap formation as a function of time for several lead compounds with palmitic acid as monitored by FTIR. Lead acetate (A) reacts most rapidly closely followed by lead white (B) while litharge (C), minium (D) and leadtin yellow I (E) take alittle bit longer. Most of the mineral matter has reacted away within half a month. Control experiments with tripalmitateand lead white show no lead soap formation. Four-year-old lead white paint reconstructions made by Leslie Carlyle for MOLART in 1999 formed lead soaps under high relative humidity (80% RH) and at higher temperature (50 oC) in a matter of one month of exposure (Carlyle, 2006). The SIMS data of the surface of these reconstructions in Fig. 3 demonstrates the increase in lead soap mass peaks of palmitic and stearic acid while FTIR confirms an increase in the metal carboxylate absorption at 1520 cm-1 (FTIR data not shown; analytical conditions see Keune 2005).

Dia1

Fig. 4 Proposed structure of metal coordinated oil paint, in which diacids form a 3D structure when coordinated by lead. Monocarboxylic acids are chain terminators that limit the size of the ionomeric polymer.

How metal soap structures could form a metal coordinated ionomeric network in a mixture of oil and metal containing pigments is shown schematically in Fig. 4 where the diacids as dumbbells coordinated by lead (red crosses) link the various lead white crystals in a metal coordinated 3D network. We think that azelaic acids and related diacids (C6-C10 diacids are also present in oil paints) will form a relatively stable metal coordinated network because of their chain building ability in 3D. Monocarboxylic acids can only act as chain terminating units because they have only one acid group. If such a structure is compromised by acid from atmospheric environmental sources or by anions that compete with the relatively weak fatty acid carboxyl groups, the coordination of the diacid structure could be locally lost. While a diacid might re-establish its coordination because of the remaining metal carboxylate group still attached to the network on the other end of the molecule, monocarboxylic acids may loose their connection with such a network permanently and migrate more easily away as free acids. Such free acids can form separate apolar lead soaps elsewhere, but they can also remain as free acids appearing at the surface as films and blooms. Monocarboxylic acids are therefore a potentially more mobile phase than the – on a molar basis – equally important or even more abundant diacids. The ionomeric oil paint model of Fig. 4 is our present working model for understanding the condition of an oil painting and the basis for an analysis when oil paint defects are apparent.

Dia ms 1

Fig. 5 Lead soaps form liquid crystals in which the metal coordinated carboxylic acids form raft-like structures with fatty acyl chains perpendicular to these rafts. This splayed chain structure has been proposed by Corkery (1997).


Metal soaps as liquid crystals

FTIR, Raman and XRD studies of synthetic lead soaps of longer chain fatty acids demonstrate that lead soaps organize themselves preferentially in liquid crystals as splayed chain systems (Corkery et al. 1997). This is possibly the reason why metal soaps tend to form aggregates in paintings that appear to grow as time progresses. This phenomenon of growth in stable aggregates is more typical for lead soaps than for some of the other kinds of metal soaps that occur in paintings (Corkery, personal communication in 2005). Metal carboxylates can form relatively stable rafts with long chain fatty acids closely packed by Van der Waals forces projecting above and below a rather stable metalcarboxylate plane. This is shown in Fig. 5 in a simplified form. The lead soap molecules are schematically shown as a ball (the metal carboxylate group) with two sticks attached (the fatty acyl chains). At higher temperature these fatty acyl chains transform into relatively disorganized almost liquid-like cushions. The flexibility of such metal soap rafts provides a lot of potential for movement in and between paint layers and might explain the appearance of lead soaps at the surface of paintings. The apolar cushions of the fatty acyl chain forests appear to be ideally suited to accommodate solvent molecules. Solvents can therefore potentially promote swelling and slippage between the rafts in the liquid crystals. Although such effects remain to be quantified, it might explain the solvent sensitivity of some aged oil paintings. In retrospect, Rembrandts’ Anatomy Lesson of Dr Nicolaes Tulp painting (MH146) that is so riddled with metal soap aggregates now, may have suffered quite a bit from the multiple hot relining operations in the 19th century, the accompanying solvent exposures and alcohol vapor treatments (Middelkoop et al. 1998). The key to understanding the lead soap behavior in paintings are the free reactive monocarboxylic acids. These either occur during the hydrolysis stage of oil paint or are released later on because of environmental exposure and acidification. Much needs to be learned still about these components in paint systems. Their reactivity can be understood locally within a single paint layer, but recall that paintings are multilayer systems where some paint layers may be medium rich but “underbound” with respect to coordinating metals, while others are lead (white) rich and presumably medium poorer. Lead poor paint layers are a potential source of mobile reactive fatty acids. Fatty lead white grounds are potentially an important source of lead soaps. So some layers act as reservoirs of free or liberated acids while others are sinks which makes it clear why metal soap formation and behavior in multilayer paint system is so complex. Since we have so few data on the quantitative aspects of the distribution of mono and dicarboxylic fatty acids in paint layers, we propose that the stoichiometry and relative distribution of lead and fatty acids in paint layers with and without lead soap aggregates should be studied in detail to understand the various quantitative aspects of the phenomenon.

Lead is not the only metal of importance. Apart from lead, metals like copper, zinc, iron, aluminum, earth alkali elements like calcium and alkali elements like potassium can form carboxylates, which we may encounter in paint layers. Potassium leaches from smalt (Boon et al, 2001) and potassium soaps have indeed been demonstrated recently in paintings (Spring et al. 2006). Potassium soaps are of course water soluble. The thermally relatively resilient calcium soaps may play a major role in the stabilization of chalk containing paints but their presence remains to be demonstrated. Aluminum soaps are common additions to modern oil paints but a disadvantage is their hygroscopicity leading to hydroxy-derivatives (Corkery1998). The saturated fatty acids in aluminum stearates do not contribute to the paint structure after hydrolysis. In fact residual hydroxy-aluminum soaps can form micelles that may be responsible for the recently observed water solubility of about 40-50 year old oil paints (Burnstock et al. 2007).

Dia ms 2

Fig 6 Schematic diagram of the partial dissolution of lead tin yellow I pigments by free fatty acids, the formation of lead soaps and their organisation into aggregates. Leached lead tin yellow I particulates remains are moved to the periphery of the aggregate (Boon et al. 2005).

Mineralization of lead soap structures

Metal soap aggregates undergo a process of mineralisation, which might be beneficial because it can stabilize them.Minium (lead orthoplumbate) has been identified in some aggregates (Boon et al. 2002; Van der Weerd et al. 2002; Higgittet al. 2003; Jones et al. 2007) and this mineral appears to form, in our opinion, when the aggregates are protected from a direct contact with the atmosphere. The penetration of carbon dioxide from the atmosphere or its reaction product with water (H2CO3) can thus be an important factor. Note that CO2 can be limited in some layers by chemical trapping in other paint layers. Phase diagrams (Boon et al. 2002; Garrels and Christ, 1965) suggest that minium is stable only under relatively alkaline conditions even when carbonate ions are present. The conditions that determine crystal growth within lead soap crystals are presently completely unknown and need to be studied in experimental systems. Studies of the chemical phases in lead soap aggregates in paint cross sections are however vitally important at this stage to establish the variables that need to be considered. In general, mineralization of metal soaps is less surprising when seen in the light of metal soap structures used as templates for biomineralized materials (Corkery, 1998).

Dia ms3

Fig 7 SEM-BSE (a) and the element maps of lead (b: PbM line) and tin (c: Sn L line) in a cross section of a lead tin yellow paint (LTY I) from the Sherborne Retable (Sherborne Abbey Almshouse, Dorset). Residual lead orthostannates are accumulated around the growing lead soap mass. Inside the lead soap mass new mineralization of lead white is visible with a medium BSE reflectivity. Most of the lead occurs in a finely divided form inside the lead soap structure. See also Boon et al. (2005) and Keune (2005).

Many protruding aggregates that we have observed contain lead carbonates which appear to be in the form of hydrocerussite nanocrystals (leadhydroxycarbonate) but the aggregates are often also remarkably rich in chlorides so the presence of phosgenite (leadchlorocarbonate) must be considered as well.
Phosgenite can epitaxially grow on cerusite demonstrating their structural compatibility (Pina et al. 1996), and a similar compatability is expected for leadwhite. The chlorides are most easily detected with SIMS in the negative ion mode (see Keune 2005), but the nanocrystals still remain to be crystallographically characterized. Mineralising zinc soaps in paints on a Van Gogh painting Falling leaves; LesAlyscamps were found to contain zinc carbonate (Keune 2005) and not zinc oxide as thought earlier (Van der Weerd et al., 2003). Lead soap masses can be so forceful in their crystal growth that they separate the leached lead-tin-yellow I pigments into fragments that are oriented around the lead soap mass (Boon et al. 2004). The process is schematically demonstrated in Fig. 6 showing the hypothetical original state and two subsequent stages. Remineralization leads to very complex SEM-BSE pictures of such paints where the submicron crystals of the new minerals can be discriminated from the coarser lead orthostannate residues (highest BSE signal) of the original degraded pigment that have accumulated around the lead soap aggregates (Boon et al. 2004). Fig. 7 demonstrates this in a paint cross section from a 15th C triptych painting (Sherborne retable) possibly made by the Master of Alkmaar, in which the tin (Sn in Fig. 7c) marks the distribution of the residual leached lead tin yellow particles while the lead (Pb in Fig. 7b) is present all over in dispersed and more mineralized forms. Mineralization also occurs on the surface of the paint or even on top of the varnish when an efflorescing crust of lead soaps reacts with atmospheric gases obscuring the picture. SIMS studies of efflorescent crusts on a canvas from the estate of F.E. Church (Olana Estate; Zucker1999) shows the presence of free fatty acids, fatty acylmonoglycerides and lead soaps (Van den Berg 2002; Van der Weerd 2002; Boon et al. 2006).

fig-8-aic

Fig. 8 SEM-BSE of a paint cross section from a lighter streak in the black background circle of sitter Susan Livingston (unknown Hudson River School painter c. 1850; CL1983.1). The originally dark bone black and earth pigment containing paint shows many random oriented micron-sized lead white crystals near the surface of the paint with a particularly high relative concentration of very small crystals where graying has taken place. It appears that neo-formation of lead white in the black paint is an important factor in a discoloration process that follows the canvas weave.

How the “ground staining” in the paintings by the Hudson River school is affected by the ground itself remains to be investigated in more detail. A preliminary study by SEM-BSE and EDX on cross sections in Fig. 8 from the background paint around the sitter Susan Livingston (c. 1850; CL. 1983.1) from the same period suggests that increasedremineralization of leadwhite at the surface of the paint determined by the canvas weave plays a role. Tiny lead white crystals and smaller submicron flakes appear quantitively more in the now lighter parts of the bone black and earth pigment toned background. Partially mineralized efflorescent crusts with lead soaps were found on the surface of a bone black paint on a 17th C ceiling in the Johan de Witt house in the Hague (Van Loon et al. 2005) and comparable lead soap crusts have been identified recently on a 19th C paintings by Bosch-Reitz (Keune et al. 2007). We postulate that many efflorescent crusts on paints are mineralized by further reaction of lead soaps with atmospheric gasses.

Physical effects

Lead soap formation inside paintings or near the surface has consequences for the stability of the paint layers and the reflection of light. Lead soap aggregation deforms the paint layers while protruding lead soap masses lead to a grainy sandy texture, paint loss, pitting and accumulation of dirt in local spots (Noble et al. 2002; Nobel and Boon, this volume). Dissolution of lead white leads to increased transparency of the paints and loss of reflected light, which may make the color darker (Noble et al. 2005; Shimadzu and Van den Berg, 2006; Noble and Boon, this volume). In the case of pigment mixtures, the color balance can be lost. Lead tin-yellow paints are known to form a lemon peel texture due to the abundant lead soap protrusions. Were some painters already aware of this effect and deliberately painted lemons with lead tin-yellow I paints? Lead tin yellow paints are lighter yellow than intended originally due to a loss of lead oxides from the pigment and the appearance of lead soaps that partially mineralize into a semi transparent white. The mobility of fatty acids and metal soaps is largely a physical process that is driven by gradients in temperature and moisture. Since these are slow processes, they easily span the professional life of a conservator. Documentation of paint layers preferably by X-raynanotomography would be a desirable way to monitor paintings, but we have to inventory on a much grander scale how many paintings are affected and why certain paintings do not show lead soap mobility at all.

Concluding remarks

Painters could not anticipate how much their paintings would change because of intrinsic factors such as the paints, their working methods, the instability of the oil paint system and external factors like moisture, variable temperature, light conditions and noxious gases. Self-repair mechanisms in paintings discovered in the course of the two NWO supported research programs in the last decade rescue the painting but do affect the picture. We still understand very little of the many slow processes that take place in paintings and how they change the pictures. The support of the MOLART and DeMayerne programme has enabled us to study these processes in a qualitative manner and has made it possible to estimate the magnitude of the occurrence of metal soap related defects in collections world wide. This is the tip of the iceberg. The main concern now is to find out how environmental conditions and conservation practices may accelerate or decelerate these processes that so deeply affect the quality of the picture and the properties of all the paint layers. This is a gigantic task that will need the combined efforts of conservators and art scientists in conservation research programmes and requires the support of the professional organizations like ICOMCC, IIC, AIC and money from national and international funding agencies.

Acknowledgements

The research was supported by Netherlands Organisation for Scientific Research (NWO, The Hague) via support to the DeMayerne project MOLMAP and MOLART Priority Program and by the Foundation for Fundamental research on Matter (FOM, a NWO subsidiary) via the approved programs 28 and 49. The Open Laboratory agreement between AMOLF and the Royal Cabinet of Paintings The Mauritshuis was instrumental in many studies of metal soaps in paintings. We especially thank Jorgen Wadum and Petria Noble for our fruitful collaboration. Special thanks go to Dr. Leslie Carlyle (now Tate, London, UK) for many brainstorming sessions on oil paints. We thank Jerre van der Horst for critical reading of the manuscript.

 

References

Boon J.J., S. Peulvé, O.F. Van den Brink, M.C. Duursma, and D. Rainford (1997). Molecular aspects of mobile and stationary phases in ageing tempera and oil paint films. In: Early Italian Paintings: Techniques and Analysis (Eds. T.Bakkenist, R. Hoppenbrouwers and H. Dubois). Limburg Conservation Institute (SRAL), Maastricht, p. 35-56

Boon J.J., K. Keune, J. Van der Weerd, M. Geldof, and J.R.J. Van Asperen de Boer (2001). Imaging microspectroscopic, secondary ion mass spectrometric and electron microscopic studies on discoloured and partially discoloured smalt incrosssections of 16th century paintings. Chimia, 55, p. 952-960.

Boon J.J., J. Van der Weerd, K. Keune, P. Noble, and J. Wadum (2003). Mechanical and chemical changes in Old Master paintings: dissolution, metal soap formation and remineralization processes in lead pigmented ground/intermediate paint layers of 17th century painting. In: 13th Triennial Meeting of the ICOM Committee for Conservation in Rio De Janeiro Preprints (Ed. R. Vontobel), James & James, London. p. 401-406.

Boon J.J., E. Gore, K. Keune, and A. Burnstock (2005). Image analytical studies of lead soap aggregates and their relationship to lead and tin in 15th C lead tin yellow paints from the Sherborne triptych. In: IRUG6 Conference 2004Florence (Ed. M. Picollo), Il Prato, Padova, p. 66-74

Boon J.J., K.Keune and J.Zucker (2006). Image analytical studies of lead soaps in preprimed canvas used by the Hudson River School painter F.E.Church. Microscopy and Microanalysis 11 (2), p. 444 and 445 (CD ROM).

Burnstock A., K.J. Van den Berg, S. De Groot, and L. Wijnberg (2007). An investigation of water sensitive oil paints in 20th century paintings. In: Modern Paints Uncovered (Eds. T.S. Learner, P. Smithen, J. Krueger and M. Schilling),Proceedingsfrom the Symposium in London. Getty Publications, Los Angeles. In press.

Carlyle, L. (2006). Historically accurate reconstructions of oil painters’materials: an overview of the Hart project 2002-2005. In: Reporting Highlights of the De Mayerne Program (Eds. J.J.Boon and E. Ferreira), MOLART Report 13, p. 63-77

Corkery R.W. (1997). Langmuir-Blodgett (L-B) multilayer films. Langmuir, The ACS journal of surfaces and colloids. 13, p.3591-3595.

Corkery, R.W. (1998). Artificial biomineralisation and metallic soaps. PhD dissertation Australian National University,Canberra.

De Vries A.B., Tóth-Ubbens M., Froentjes W. (1978), Rembrandt in het Mauritshuis, Alphen aan de Rijn, p. 89.

Garrels, R.M. and C.L. Christ (1965) Solutions,minerals and equilibria. Harper & Row Publishers, New York.

Higgitt, C., M. Spring and D. Saunders. 2003. ‘Pigment-medium Interactions in Oil Paint Films containing Red Lead or Lead-tin Yellow’, National Gallery Technical Bulletin, 24 75-93.

Heeren R.M.A., J.J. Boon, P. Noble and J. Wadum (1999). Integrating imaging FTIR and secondary ion mass spectrometry for the analysis of embedded paint cross-sections. In: Proceedings of the 12th ICOM-CC Triennial Meeting in Lyon, (Ed. J.Bridgland) James and James, London. p. 228-233

Hoogland F. and J.J. Boon (2005). Unpublished. Reported at the ICOMCC Scientific Research Working Group interim meeting on lead soaps in collaboration with MaSC held in the Van Gogh museum in Amsterdam on 10 September 2005.

Jones R., J.H. Townsend, K. Stonor and N. Duff (2007). Lead soap aggregates in sixteenth and seventeenth century British paintings. In: AIC Paintings Working Group Meeting 2006 Postprints (Comp. Helen Mar Parkin).

Keune K. (2005). Binding medium, pigments and metal soaps characterised and localised in paint cross sections. PhD Thesis University of Amsterdam. pp. 181. Report of MOLART Series: Molart11, AMOLF, Amsterdam.

Keune K., K. Kirsch and J.J. Boon (2007) Lead soap efflorescence in a 19th C painting: appearance, nature and sources of materials. In: AIC Paintings Working Group Meeting 2006 Postprints (Comp. Helen Mar Parkin). In press.

Keune K.and J.J. Boon (2007) Insights into the lead soap aggregate formation in oil paintings with analytical imaging studies. Studies in Conservation, in press.

Middelkoop, N., P. Noble, J. Wadum and B. Broos. 1998. Rembrandt under the Scalpel The Anatomy Lesson of DrNicolaes Tulp Dissected. Amsterdam / The Hague, 56, 57, 66.

Noble P., J.J. Boon, and J. Wadum (2002). Dissolution, Aggregation and Protrusion: Lead soap formation in 17th century grounds and paint layers. ArtMatters 1, p. 46-62.

Noble P., A. Van Loon, and J.J. Boon (2005). Chemical changes in old master paintings II: darkening due to increased transparency as a result of metal soap formation processes. In: Conference Proceedings 14th Triennial Meeting of the ICOM Committee for Conservation in The Hague (Ed. I. Verger). James & James, London. p 496-503

Noble P and J.J.Boon (2007). Metal soap degradation of oil paintings:aggregates, increased transparency and efflorescence. In: AIC Paintings Working Group Meeting 2006 Postprints (Comp. Helen Mar Parkin). In press.

Pina C.M., L. Fernandez-Diaz, and M. Prieto (1996). Topotaxy relationships in the transformation phosgenite cerusite. J Crystal Growth 158, 340-345.

Schilling M.R., J. Mazurek, and T.S. Learner (2007). Studies of Modern Oil-Based Artists’ Paint Media by Gas Chromatography/Mass Spectrometry.In: Modern Paints Uncovered (T.S. Learner, P. Smithen, J. Krueger and M.Schilling), Proceedings from the Symposium in London. Getty Publications, Los Angeles. In press.

Spring M., C. Higgitt and D. Saunders (2005), Investigation of Pigment-Medium Interaction Processes in Oil Paint containing degraded Smalt. National Gallery Technical Bulletin 26, 56-69.

Van den Berg J.D.J., K.J. Van den Berg, and J.J. Boon (1999). Chemical changes in curing and ageing oil paints. In: Proceedings of 12th Triennial ICOM-CC Meeting Lyon, (Ed. J. Bridgeland), James and James, London. p. 248-253

Van den Berg J.D.J. (2002). Analytical chemical studies on traditional linseed oil paints., PhD thesis University ofAmsterdam. Report of MOLART Series: Molart 6, AMOLF, Amsterdam.

Van Loon A., and J.J. Boon (2005). The whitening of oil paints films containing bone black. In: Conference Proceedings 14th Triennial Meeting of the ICOM Committee for Conservation in The Hague (Ed. I. Verger), James & James, London, p. 511-518

Van der Weerd J., J.J. Boon, M. Geldof, R.M.A. Heeren and P. Noble (2002), Chemical changes in old master paintings: dissolution, metal soap formation and remineralization processes in lead pigmented paint layers of 17th century paintings.Zeitschrift für Kunsttechnologie und Konservierung, 16, p. 36 – 51.

Van der Weerd J. (2002b). Microspectroscopic Analysis of Traditional Oil Paint. PhD thesis University of Amsterdam, Report of MOLART Series: Molart 7, AMOLF, Amsterdam.

Van der Weerd, J., M. Geldof, L. Struik van der Loeff, R.M.A. Heeren, and J.J. Boon (2002), Zinc soap aggregate formation in ‘Falling leaves’ (Les Alyscamps)’ by Vincent van Gogh. Zeitschrift für Kunsttechnologie und Konservierung; 17 (2) p 407-416

Verhoeven M.A., L. Carlyle, J. Reedijk and J.G. Haasnoot (2006). Exploring the application of solid-state Nuclear Magnetic Resonance to the study of the deterioration of paintings. In: Reporting Highlights of the De Mayerne Program (Eds.J.J.Boon and E. Ferreira), MOLART Report 13, p. 33-43.

Zucker, J. 1999. From the Ground Up: The Ground in 19th- Century American Pictures. Journal of the American Institute for Conservation, 38 (1), 3-20.

PDF of the paper: