[Note: This paper is based on the author’s research project focusing on fatty acid efflorescence]
Dit project werd uitgevoerd binnen het langlopende samenwerkingsverband tussen de Rijksdienst en het Courtauld Institute of Art. Inlichtingen: Klaas Jan van den Berg; firstname.lastname@example.org.
The term efflorescence has been adapted in conservation literature to encompass organic components and their migration process, including their final deposition on the surface of the
paint film . Organic analysis has identified the compositions of these surface crystals, but little research exists identifying the critical conditions necessary for their formation and their migration process.
During the scope of the project a number of case study paintings with characteristic surface whitening were examined. Alongside these artworks a preliminary experimental study was
carried out in order to investigate the movement of fatty acids in prepared paint samples of known compositions. This paper will discuss three of the case study paintings and give a summary from the experimental component, before providing suggestions for preventive conservation. The term efflorescence will be used interchangeably with the word ‘bloom’, as the latter term has similarly
been reserved for fatty acid deposits and their mechanisms .
These fatty acid deposits, white in appearance due to their light scattering effect, appear as ‘tiny needles’ or ‘crystals’ on the surface of paintings. They range in size from 1.5mm to 10mm and are usually easily removed with mechanical action. The density and distribution of efflorescence can be dependent on the origin of the surface deposits and the artists’ materials and techniques. Observations from existing case studies range from the whitening being most dense in the dips of the paint film to it being specific to certain paint passages or underlying compositions   .
The drying of an oil-binding medium is a complex process, but a hypothesis about the nature of oil paint and its drying processes will be helpful in understanding the phenomenon.
Fatty Acids and the Oil Binding Medium
The five most abundant fatty acids in a drying oil medium are linoleic, linolenic, oleic, palmitic and stearic acids. Linoleic and linolenic acids, di- and tri- unsaturated respectively, are the compounds principally responsible for the drying of oils as the presence of two or more double bonds makes them more prone to oxidation. Oleic acid is monounsaturated and although its role in the polymerization process is still unknown, it does gradually oxidize and become incorporated into the polymer network.
Palmitic and stearic acids are both saturated fatty acids which, due to their non-reactive state, remain uninvolved in the cross-linking process .
The esters in a drying oil are formed by the combination of individual carboxylic acids with trihydric alcohol glycerols. Polymerization, hydrolysis, oxidation, and soap formation (in the presence of metal ions), are the four major processes which occur in an oil paint film .
It is the oxidation of these triglyceride molecules at the sites of their double bonds which results in a drying oil polymerizing from a liquid to a cross-linked solid film. These reactions do not produce very long polymer chains, but this is not necessary for the three-dimensional network.
Typically, in the early phase of a drying oil film, the medium contains less than 10% saturated fatty acids, the majority of which are immobile as they form part of the triglyceride molecules. The hydrolysis of esters can occur with exposure to water alone, requiring much higher temperatures, or in the presence of an alkali or acid catalyst . It has previously been proposed, however, that the free fatty acids in efflorescence do not originate from the oil medium, even through hydrolysis, but rather from the additives in the paint.This hypothesis was tested for the experimental component of the project. It should be noted that this very simplified model did not take into account several influential factors, such as pigments and additives, the presence of light, thickness of the film or the extent of prepolymerization of the oil, all of which will affect the drying process of the oil medium.
The Portrait of Sir Edward White [Fig. 1]
The first painting investigated in the project comes from a series of portraits that was bequeathed to the Guildhall Art Gallery in 1989. The paintings depict the Chairmen of the London County Council and the first painting was completed in circa 1889, after which a portrait was commissioned every year. The 88 portraits, executed by a broad range of different artists, were surveyed last year and a total of 27 of them were found to have varying degrees of surface whitening. In a catalogue compiled by the Greater London Council 1986, most of the paintings are listed as being in a storage facility in East London, thought to be environmentally uncontrolled. Unfortunately there is very little information on the history of the paintings prior to their arrival at the Guildhall.
The Portrait of Sir Edward White, by the artist Edward Poynter, dates from circa 1911. The portrait was chosen for the interesting distribution of its bloom, which appeared crystalline under microscopic examination and was easily dislodged with a dry cotton wool swab.
The efflorescence is most dense in the interstices of the paint film, but in the more heavily bloomed
areas the surface deposits are also present along the tops of the canvas weave and appear to form a white film.
The painting is executed on an unlined commercially prepared linen canvas with an oil-bound lead white and calcium carbonate containing ground and a thin lead white priming layer [Fig. 2a and b]. The artist has applied the paint in thin layers, with washes used for the background. His medium rich palette consists largely of earth pigments and bone black, and sulphur and magnesium containing compounds may be present as additives.
The location of this painting’s bloom appears to coincide with the wooden auxiliary support, concentrating itself on the areas along the bottom stretcher bar and the lower part of the sidebars. The painting is glazed, therefore, moisture exposure at the verso of the canvas may have caused fatty acids to migrate towards the surface. Moisture will accumulate in cold hygroscopic materials and may be released upon changing temperature conditions, and this may explain the isolation of the bloom to the areas directly above the wooden support. The extensive rips along the picture plane on the right and left hand side are most likely due to the keying out of the stretcher and the possible embrittlement of the canvas due to its exposure to moisture [Fig. 3].
It may be tentatively suggested that the heavy presence of efflorescence on the lower part of the painting may be due to an upward migration of moisture within the stretcher. This may have been caused by the bottom of the auxiliary support being in a damp environment.
The paint in the background is similar to the affected areas in that it is a very medium rich matrix with a thick size layer. An absence of lead, as confirmed by EDX, and the presence of efflorescence on the tacking margin would suggest that it too should have visible surface deposits.
Figure 4 is a scanning electron micrograph taken during the examination of a varnished paint sample. Previous case studies have indicated that bloom is usually not visible in varnished areas that are still intact. The sample represents an area exhibiting characteristic surface whitening and appears to show the fatty acid crystals emerging from the thick, crack-free varnish layer.
Figure 5 is of a sample also taken from a varnished area with efflorescence, but where the varnish was removed with acetone prior to sampling. This scanning electron micrograph shows a large amount of the shard-like crystals present on the surface of the paint. The varnish was identified as an oxidized natural resin based on a greenish fluorescence in Ultraviolet light and its solubility in polar solvents. Instrumental analysis of the varnish coating would be beneficial, as the knowledge of the composition of this layer would influence the interpretation of the cause or suppression of the efflorescence. Previous case studies have found that bloom on the surface of a synthetic varnish appeared to occur synonymously with disrupted areas, therefore it may be possible that the presence of a natural resin component in the varnish may affect the ease of migration of free fatty acids through an unbroken film. A more precise analysis of the composition, porosity and degree of cross-linking of a varnish might inform the question of how migration of free fatty acids may occur through an unbroken coating.
SEM of further samples revealed the presence of shard-like fatty acid crystals on the ground of the
tacking margin [Fig. 6] as well as on the ground taken from the verso of the canvas [Fig. 7]. A painting with efflorescence by Sergei Poliakoff, presented by Laura Hinde at a previous Gerry Hedley conference, was also found to have crystals along the back of the canvas. These findings may suggest that free fatty acids may travel in both directions.
Organic analysis of the ground, paint and efflorescence from the edge of the unvarnished picture plane was carried out using gas chromatography-mass spectrometry. The fatty acid ratios in the ground layer indicate a pure linseed oil medium that has not been heat bodied, whilst the paint contains a slightly heat-bodied linseed oil. The efflorescence on this passage consists mostly of saturated fatty acids which most likely originate from the medium rich preparatory layer, but it is inconclusive as to whether they are present in free form or as metal soaps.
Male Nude [Fig. 8]
The second case study is Male Nude by the Scottish artist Herbert Wilson Foster. It dates to circa 1890 and is a painting which is currently undergoing treatment in the conservation department at the Courtauld.
The efflorescence was uniformly present over the paint surface and not localised to specific passages of paint, leading to the hypothesis that the structure and composition of the ground may play a role and be a source of free fatty acids. Analysis of the organic composition of the paint and ground layers would provide supporting evidence for this hypothesis.
Light microscopy revealed that it was present on the tops of the canvas weave as well as in the
interstices, and during treatment it was found that at least 10 swab rolls with Stoddard’s Solvent were needed for the bloom to be fully removed.
The painting is executed on an unlined commercially prepared canvas with two oil-bound lead white and calcium carbonate containing ground layers [Fig.9a and b]. The artist’s palette
consists largely of bone black and iron oxide pigments, with read earth pigments, lead white and vermillion used for the flesh, and barium sulphate may be present as an additive.
Imaging of the painting’s ground layer using SEM identified fatty acid crystals on its surface [Fig. 10] but unlike The Portrait of Sir Edward White, crystals were not found on a sample from the verso of the canvas. The back of the canvas is covered in what appears to be a proteinaceous size layer and this may have a direct effect on the possible absence of crystals along the back of the painting, as well as exacerbating the efflorescence through the absorption of atmospheric moisture.
The efflorescence on the painting’s surface is characteristic of the shard-like crystals found in fatty
acid bloom [Fig. 11]. Fungal hyphae were also discovered, indicating that the painting has been exposed to humidity levels higher than 65% RH.
Comparison of the preparatory layers on the two case study paintings shows a similar layer
structure [Fig. 2a&b and 9a&b]. The first application is a thick, oil-bound medium rich ground consisting of large proportions of chalk and a very small amount of lead, whilst the very thin opaque priming layer is largely lead white containing. The great quantities of oil in the lowest layer
would be a likely source of fatty acids, as was indicated by the organic analysis carried out on The Portrait of Sir Edward White. Lead white is known for its stabilizing effect on paint films. However, the thin priming layer in comparison to the large proportions of free fatty acids in the oily ground may not have been enough to prevent their migration to the surface.
In two separate areas on Male Nude, the efflorescence was exacerbated through the contact of the paint surface with moisture. One area had been humidified for a flattening treatment, and the other was an aqueous cleaning test carried out with de-ionised water and a micro-brush.
Imaging of the area that underwent aqueous cleaning shows an organic layer that has been broken up, with the efflorescence either re-emerging or being semi-squashed. Moisture has often been suggested as a catalyst or reactor for fatty acid migration but its exact role is unknown.
After removal of the efflorescence using stoddard’s solvent, an oily dirty layer was removed using a
2% solution of triammonium citrate at a pH of 9. Direct exposure of the dark, medium rich paint to water, caused blanching of the skin of medium in areas of these passages.
Cleaning Techniques [Fig. 12]
Topographical examination of two samples taken from Male Nude after ‘removal’ of the efflorescence highlights the differences between the two cleaning techniques employed. The removal of the bloom using a silk cloth leaves behind globules of organic matter on the surface of the paint film [Fig. 13], whilst the use of Stoddard’s Solvent leaves the surface free of crystals with visible pits where the fatty acids once were [Fig. 14].
Dry cleaning methods do not appear to sufficiently remove the fatty acid deposits and the use of
mechanical action can result in the burnishing of the paint surface and the abrasion of pigment particles in unvarnished paint films.
The use of solvents to remove these crystalline entities was found to be a more successful treatment method, but it raises causes for concern regarding the exposure of unvarnished paint films to solvents. The removal of these free fatty acids and any subsequent leaching of organic components may result in the embrittlement of the paint film. Aside from being mechanically detrimental to the paint film, this may be an exacerbating factor in the re-emergence of fatty acid crystals, as indicated by previous reports of the presence of surface deposits in areas of leached and fissure-containing paint.
Ranunculus [Fig. 15]
The third case study is a contemporary still-life by the British artist Mia Tarney. The primary support is a linen canvas that was prepared by the artist with rabbit skin glue. The oil paint is applied directly onto this proteinaceous sizing layer in a single, thin application resulting in the canvas showing through the paint along the tops of the weave. The piece is unframed and unglazed.
The painting is privately owned, and efflorescence started to become visible on the dark passages overnight. Prior to these changes the stone floor of the house, close to where the painting hangs, had been professionally cleaned using Lithofin Wexa, a slightly alkaline water-based material that contains low proportions of surfactants.
The Winsor & Newton Olive Green paint used for the background consists of carbon black and a yellow dyestuff. Elemental analysis would suggest that calcium silicates, and compounds of aluminium and sulphur are present in the paint as additive. Elemental mapping of the surface indicates the presence of sodium soaps. The bloom was found to be soluble in both water and Stoddard’s Solvent, the presence of sodium explaining its solubility in aqueous materials.
This is a material known to be added in very small quantities to paints as stabilisers and dispersion
agents. As the painting lacks a priming or ground layer the oil paint is the source of the fatty acid component of the soaps. Organic analysis using ElectrosprayIonisation – Mass Spectrometry detected the presence of free fatty acids of animal fat origin indicative of metal stearates.
During a previous Courtauld project on water sensitive oil paints, paint samples, which had been cleaned with saliva and de-ionized water were found to form efflorescence after a short period of time. The prepared paint samples contained different pigments but their common thread was in the added aluminium stearates. The stearic acid based bloom on the surfaces was found to be more extreme on samples with higher proportions of the additive.
A report from 2002 suggests that painting materials from the 19th and 20th century contain increased quantities of free fatty acids due to developments in commercial paint production. This included the use of ingredients and manufacturing methods that retarded or accelerated the oxidation process, those which promoted hydrolysis, and the introduction of new or increased amounts of materials containing fatty acids. Existing case studies, such as the first two presented today, indicate that this timeline is accountable for the majority of paintings exhibiting the
surface phenomenon, suggesting a possible correlation between developments in British paint manufacture and the formation of fatty acid efflorescence.
The samples therefore used for the experimental component of the project were paint films of known compositions that contained increased proportions of added metal stearates and linseed oil derived free fatty acids. Cold-pressed and alkali-refined linseed oils were used for the binding medium, and the ingredients were mixed using an automatic muller before being painted onto an acrylic primed canvas or a Melinex sheet. All samples were artificially light aged for the equivalent 14 to 25 years in recommended museum conditions, assuming reciprocity.
The effects of varying the environmental humidity and temperature, creating moisture gradients and the outcome of direct water application on the samples were investigated in an attempt to ascertain the conditions under which fatty acid efflorescence becomes visible. Salt solutions were used to create extreme moisture gradients in order to encourage the moisture to travel through the canvas paint samples. Paint samples were prepared on a stretcher to be hung on a cold wall in a climate-controlled room and a separate set were introduced into a climate-controlled oven. It had been anticipated that with these extreme experimental conditions and the large proportions of fatty acids in the samples, that notable results would be achievable. Longer running set-ups, which were outside the limitations of this study, would be encouraged for future testing.
Of all the different conditions, visible surface whitening was only achieved upon immersion of the samples for 75 minutes in unfiltered water at a raised temperature of 68 degrees celsius. In comparison to the de-ionised water, this water had a measurable difference in pH and ionic conductivity as well as containing naturally occurring minerals such as calcium, magnesium and sodium.
The paint samples on a Melinex substrate showed the development of significant efflorescence, reflecting the predicted one-directional migration of the free fatty acids, with canvas samples showing a less dense bloom. Another interpretation could be that the canvas absorbs some of the organic medium, reducing the source of free fatty acids. This might infer that efflorescence might be reduced in paintings or samples that have no size layer or in more extreme cases, no priming or ground layers that would otherwise prevent the absorption of oil.
An interesting result was achieved using a viridian paint film with 15% added linseed derived free fatty acids and magnesium stearates [Fig. 16]. A denser amount of bloom formed on the half of the sample that was swabbed with warm de-ionised water prior to its immersion in room temperature water for one hour and left to dry in the climate-controlled laboratory.
A cross-section from this sample shows the presence of a haze-like white layer over the surface of the paint film [Fig. 17]. The solubility of this layer was found to be unaffected by polar solvents, saliva and water. Stoddard’s Solvent caused the coating to break up, and heat tests indicated a melting point of approximately 60 degrees Celsius.
Organic analysis of the white coating using GC-MS detected oil medium components in the surface coating. The low Palmitic/Stearic ratio in the bulk paint is due to a substantial amount of metal stearates added into the paint. The efflorescence has a low P/S ratio of 0.5 due to the many fatty acids from these stearates.
SEM was carried out on a number of samples in order to characterize the surface deposits. Unlike the shard-like crystals encountered on the efflorescent areas of the case study paintings, the white
layers on the test samples were found to be curly in nature and consisted of interconnected sub-structures [Fig. 18]. The intensity of the crystalline networks appeared to be directly related to the proportion of added fatty acids in the paint samples as well as to the substrate on which they were painted.
A comparison of results from experiments using water at varying temperatures, suggests that temperature may also play a role in the migration of fatty acids. This may be due to the glass transition temperature of the paint film, the melting point of the fatty acids themselves, or the breaking of chemical bonds within the paint film. One of the migration theories in conservation literature suggests that crystallization, through the movement of chains through the matrix until they reach a nucleation site, may be encouraged if paint films are below their glass transition temperature but above the minimum melting point for the palmitic-stearic system. The melting point for this system is influenced by the proportions of each fatty acid, with a eutectic mixture of 70% palmitic and 30% stearic acid having as low a melting point as 8°C.
This migration of the liquid fatty acid molecules would explain the appearance of white, haze-like silhouettes seen on the glazing of works with efflorescence. It is the more volatile degradation components which are transferred to the glass, leaving the larger palmitic and stearic acids to solidify on the surface of the paint.
The experiments in this project were preliminary and further investigations with more sophisticated et-ups are necessary. The experiments that resulted in the visible formation of efflorescence highlight the possible role of pH and ionic conductivity of the water on the formation of acute efflorescence, as well as temperature and type of water exposure. The next step would be experiments that investigate these individual components separately in order to shed more light on the combination of critical conditions that cause the migration of fatty acids. Although not systematically investigated in this study, the individual components within a paint film, such as the oil medium, pigments and additives present, will also affect the properties of the film and its permeability to moisture.
The formation of dense efflorescence on samples containing chromium oxide green, which is a low oil absorbing pigment, strengthens the hypothesis that metal stearates are the sources of free fatty acids in these paint samples. The inorganic components from these additives have been found to remain evenly distributed within the pint matrix of an efflorescent film. urther studies on the degradation of metal stearates will therefore help to inform on the migration of the organic omponents to the surface.
Reducing the exposure of the paint films to moisture is thought to be critical in the prevention of fatty acid migration. Semi-closed systems through the glazing of a painting may result in the formation of extreme temperature and moisture gradients within the small space between the glass and the paint surface. Backboarding is therefore a sound preventive measure, but care should be taken as previous studies indicate the importance of controlling temperature, as thermal gradients will automatically result in an RH gradient. Allowing for a gap between the picture and the wall will ensure a great reduction in any temperature gradients caused by a warm or cold wall and any subsequent RH fluctuations within the system.
Until the catalysts and formation of efflorescence are better understood, steps should be taken in preventive conservation.
My thanks go to: Professor Aviva Burnstock and Dr. Klaas Jan van den Berg for their support and supervision during this project; Alexandra Walker, Nancy Wade, Vicky Leanse and Sonia Solicari at the Guildhall Art Gallery for their enthusiasm and for allowing me to examine paintings from their collection; Pearl O’Sullivan for contributing valuable information on the painting in her care; Melanie Caldwell for sharing her enlightening report on British Paint manufacture and for the use of one of her treatments as a case study painting; Jeremy Johnson and Frank Ligterink for helpful discussions; Marc Vermeulen for carrying out the organic analysis, Dr.Alex Ball for all his help with scanning electron microscopy, and Bill Luckhurst for the use of his facilities.
 For full project please see Efflorescence: An Investigation of Selected Paintings
from the 19th to the 21st Century with a Preliminary Experimental
Study of the Role of Moisture in the Development of Efflorescence. (project no.
161) held at the Department of Conservation and Technology at the Courtauld
Institute of Art
 Ordonez, E., and Twilley, J., Clarifying the Haze: Efflorescence on Works
of Art. 1998.
 Boon, J. J., Hooglund, F., and Keune, K., Chemical Processes in Aged Oil Paints
Affecting Metal Soap Migration and Aggregation, 2007; pg 516.
 Singer, B., Davenport, J., and Wise, D., Examination of a Blooming Problem in a
Collection of Unvarnished Oil Paintings, 1995.
 Koller, J., and Burmester, A., Blanching of Unvarnished Modern Paintings: A
Case Study on a Painting by Serge Poliakoff, 1990;, pg 138.
 Ibid; pg 139.
 Hinde, L., A Technical Study of Selected Paintings from the 1960s with
Characteristic Surface Deterioration from Dudmaston Hall, Shropshire. 2010;
 Singer et al (as in note 4), pg 4.
 Burnstock, A., Caldwell, M., and Odlyha,
M., A Technical Examination of Surface
Deterioration of Stanley Spencer’s Paintings at Sandham Memorial Chapel, 1993;
 Sutherland, K., An Investigation into the Relationship Between Organic Composition of
Commercially Prepared Grounds and Efflorescence, 1995.
 Singer et al. (as in note 4), pg 6-8.
 Ordonez and Twilley (as in note 2).
 Mills, J. S., and White, R., The Organic Chemistry of Museum Objects, 1987;
 Ibid, pg 30.
 Ordonez and Twilley (as in note 2).
 Koller and Burmester (as in note 5), pg
 Mills and White (as in note 13); pg 27.
 Erhardt, D., Tumosa, C. S, and Mecklenburg, M. F., Long Term Chemical and Physical Processes in Oil Paint Films, 2005, pg 148.
 Boon, Hooglund, and Keune (as in note 3),
 Unless stated otherwise, this whole paragraph is referenced from Mills and White (as in note 29), pg 31.
 Erhardt et al. (as in note 18), pg 145.
 Mills and White (as in note 13), pg 29; ‘mineral acid’.
 Hydrolysis is also effected by water catalyzed by enzymes – Gunstone, F. D., An Introduction to the Chemistry and Biochemistry of Fatty Acids and their
Glycerides, 1967; pg 127.
 Winsor & Newton’s Chief Chemist Alun Foster as cited by Caldwell, M., Some Developments in British Paint Manufacture c. 1850-2000 with Reference to Surface Deterioration on Paintings and in Particular Free Fatty Acid Efflorescence, 2002; pg 40.
 The survey was conducted by the author and Nancy Wade (Conservator at the Guildhall Art Gallery).
 Inorganic analysis was carried out on samples using Electron Dispersive X-Radiography.
 Ligterink and Di Pietro, Canvas paintings on Cold Walls: Relative Humidity
Differences Near the Stretcher, 2007, pg. 33.
 Singer et al. (as in note 4), pg 4 and 6.
 Burnstock et al., (as in note 9), pg. 235.
 Hinde, A Technical Study of Selected Paintings from the 1960s with Characteristic Surface Deterioration from Dudmaston Hall, Shropshire (2010), pg. 14.
 [P/S = 1.9
diFA9/diFA8 = 7.0] As indicated by the ratios, the oil medium consists of linseed oil (1<P/S<2) that has not
been heat bodied (diFA9/diFA8> 4.
 [P/S = 1.6 diFA9/diFA8 = 4.8] As indicated by the ratios, the oil medium consists of linseed oil (1<P/S<2) that has been slightly heat bodied (5<diFA9/diFA8> 4).
 [P/S = 2.0]
 EDX carried out on samples by P. O’Sullivan, the student treating the painting.
 Koller and Burmester (as in note 5), pg
 Information on the artist’s materials and techniques, the painting’s location of hanging, and the events preceding the acute efflorescence, were gathered through private conversation with the conservator Melanie Caldwell who carried out the treatment on Ranunculus.
 The technical information sheet for the product [http://lithofin-uk.co.uk/wp-content/uploads/tech-sheets/wexa.pdf].
 Inorganic analysis carried out by Professor A. Burnstock.
 Treatment report by M. Caldwell.
 P/S ratio of c. 0.7: K.J. van den Berg and Suzan de Groot, Investigation of overnight efflorescence-development on Mia Tarney ’s Rununculas (2007). Report Rijksdienst Cultureel
 Tempest. H., Water Sensitive Oil Paint: An Experimental Investigation with Case Studies Further Characterising the Causes of the Phenomenon, 2009.
 Caldwell, M., (as in note 24).
 The samples had been prepared during previous student projects.
 Tempest (as in note 37), pg. 11.
 [P/S = 0.7] Metal stearates have a p/S ratio of c. 0.67 whilst linseed oil has a P/S ratio of c. 1.4.
 Ordonez and Twilley (as in note 2).
 Koller and Burmester (as in note 5), pg 139.
 Wornum as cited by Singer et al. (as in note 4), pg 6.
 Tempest (as in note 37), pg. 59.
 Ligterink and Di Pietro (as in note 27).