1Resin and bile soaps are aqueous products, composed of water up to 95 % and are used for the cleaning of natural resin varnishes usually found on paintings. They are composed of a weak base and acid in water, and according to the needs, other elements such as solvents, chelating agents and surfactants can be added before the mixture is gelled.
- 1 Hook, J., « The ICC Congress, Brussels, ‘Cleaning, retouching and coatings’ », in Queensland Natio (...)
2They were introduced to the conservation world in the 1980’s during the IIC Congress (Brussels, 1990)1. Although still not widely known and used, they are an alternative to traditional cleaning systems-such as solvents-for the cleaning of paintings, particularly in cases where the paint layer is solvent sensitive, or when the varnish has grime embedded or is oil based.
- 2 Wolbers, R., Aqueous Methods for Cleaning Painting Surfaces. London: Archetype Publications, 2000, (...)
3Soaps are used for the removal of natural resin varnishes off paintings, furniture or even musical instruments2. They are considered as an alternative to traditional methods as they would potentially give good cleaning results where solvents such as ketones, alcohols, aliphatic and aromatic hydrocarbons would not.
- 3 Wolbers, R., op. cit., p. 42.
4For instance, patches of old oxidized varnish, which are highly polar. The solvents that could remove them off the paint layer would be also polar, and thus a potential threat to the paint layer. Then, the need of a product that could remove the varnish while safeguarding the paint layer arose3. In the event of dirt embedded varnishes, the dirt can act as barrier and prevent the solvents from solubilizing the resin. Using an aqueous system means having to deal with pH: the combination of water molecules and the adequate pH (between 5,5 and 8,5) can overcome this particular problem by playing the solubility parameters of both the resin and the dirt.
- 4 Byrne, A., « Wolbers’ cleaning methods: introduction », in Queensland National Art Gallery
- 5 Erhardt, D., and Bischoff, J., « Resin Soaps and Solvents in the Cleaning of Paintings: Similariti (...)
5Soaps only functions on natural resin varnishes, even those containing a small part of oil and having dust embedded onto the surface. The older, more oxidized, degraded and polar the varnish, the better the soap functions: the explanation is a similar polarity4. Once again, a typical case would be a painting with very old varnish patches that are very degraded and thus very polar and that could not be removed with solvents without damaging the paint layer5.
- 6 Caretti, E., Dei, L., « Cleaning II : Applications and Case Studies », in Nanoscience for the Cons (...)
- 7 Wolbers, R., op. cit., p. 39.
- 8 Cremonesi, P., Nuovi Materiali e Metodi per la Pilitura dei Dipinti, Firenze: Phase, 1995, p.42.
6Molecular affinity between the terpene resin (in the varnish) and the acid is one of the other explanations of how soaps work. Abietic acid has a similar molecular structure to copal and sandarac while deoxycholic acid resembles dammar’s and mastic’s6. Wolbers describes soap action as “the strong affinity for one hydrocarbon structure to another that accounts for the specific solubilizing effect”7. Molecules are similar and attract each other, which dissolves the layer: similia a similibus solvuntur or “same dissolves same”8.
7Cleaning with aqueous products, especially with soaps, the action is chemical rather than physical, which means the intramolecular (inside the molecule) links are targeting, whereas with solvents, the physical action targets the intermolecular links, or the links assembling the molecules together.
- 9 Personal communication with Paolo Cremonesi during the KIK-IRPA Workshop, May 2014.
8Soaps can be used to thin varnishes. Conservators know how difficult it is to regularly thin a coat of varnish: solvents tend to swell the whole layer, whereas soaps target the most oxidized thus yellow part of the layer9. Thinning varnishes could benefit paintings as it would space out full cleanings and preserve paintings better.
- 10 Water is has a low volatibility but will eventually evaporate, although much slower than solvents (...)
9The principal difference with solvents is the non-volatility of the ingredients: an acid (abietic or deoxycholic), a base (triéthanolamine “TEA”) and water10. To these basic ingredients can be added surfactant and/or solvent and/ chelating agent…before being gelled with a cellulose ester.
- 11 Wolbers, R., op. cit., p. 39.
- 12 Erhardt, D., and Bischoff, J., op. cit., p. 142.
10A soap is called “resin” or “bile” according to the acid used: if the abietic soap is used, it is a resin soap (Fig. 1); if the deoxycholic acid is used, it is a bile soap (Fig. 2)11. The non-volatility of ingredients implies the necessity of a thorough rinse to avoid leaving potentially active residues after the cleaning12.
Fig.1 Triethanolammonium abietate soap
The reaction between abietic acid and TEA produces triethanolammonium abietate soap.
Crédits: Richard Wolbers
Fig.2 Triethanolammonium deoxycholate soap
The reaction between deoxycholic acid and TEA produces triethanolammonium deoxycholate soap.
Crédits: Richard Wolbers
11Two recipes are given.
12The first recipe is given by Wolbers.
Water: 100 ml
Abietic acid: 2 g
TEA : 5 ml
HCl concentrated at 1N
13Place the water in the jar, place the magnetic stirrer. Add the acid, then the base. Mix for 2 minutes. The acid and the base together create a salt or soap. A salt made of abietic acid and triethnaolamine is a triethanolammonium abietate soap; a salt made with deoxycholic acid and triethanalamine is a triethanolammonium deoxycholate soap.
- 13 pKa is a measure of acid strength. It depends on the identity and chemical properties of the acid. (...)
14 Filter the undissolved elements with a coffee filter. Buffer the solution with the HCl solution, until pH 8: add drop by drop the HCl solution until wanted pH is obtained, which is +/- 1 unity around the pKa13. For example, TEA has a pKa of 7,8; the solution is buffered from pH 6,8 to 8,8. Add the surfactant or solvent or chelating agent if necessary. Gel with a cellulose ether.
15Cremonesi’s recipe uses the same proportions, only the preparation differs.
16Mix the acid to the TEA with the magnetic stirrer. Heat the solution at 60 °C for 12 hours. Add the water (warmed) to avoid a thermic shock. Buffer with the HCl solution until pH 8 is reached. Add the additives if needed, then gel with Klucel H 4 %. This recipe is only for resin soap since the abietic acid is difficult to dissolve and thus a sizeable part is filtered out. Heating it in TEA accelerates the dissolution and diminishes the wasting of product.
- 14 Wolbers, R., op. cit., p. 37.
17Since the solution is aqueous, it is important to control the pH. The cleaning of paintings should be done between pH 5.5 and 8.5. Under, there are risks of hydrolysis of organic binders; above, there are risks of saponification of the oil paint layer14.
- 15 Wolbers, R., op. cit., p. 19.
18The pH of an aqueous solution plays an important role since it influence the ionization, and this the solubility of the materials to be removed. For example, a natural resin has a pKa of 6,5. Using a pH 6,5 aqueous solution start the ionization process, which means part of the resin becomes water soluble. The more the pH rises, the more groups ionize which increases the hydrophilic character. At pH 8,5, the resin is completely ionized and can be solubilized in a aqueous solution such as a soap (composed at 95 % of water)15.
19Soaps are generally at pH 8. In order to keep this measure constant and avoid sudden variations, the soap is buffered to the desired pH, which means its measure varies little despite the small addition of acid and base or even if the stock solution is diluted. To buffer a solution, equimolar quantities of base and acid are introduced. The buffer zone depends on product used, based on its pKa.
- 16 Micelles are molecules that have both polar or charged groups and non polar regions (amphiphilic m (...)
- 17 Caretti, E., Dei, L., op. cit., p.186.
- 18 Wolbers, R., op. cit., p. 158.
20Soaps are applied the same way as gelled cleaning systems: with a brush or a swab. The gel is left on the surface and worked until the targeted layer is removed. Working the gel is necessary as the soaps molecules need mechanical action to form micelles16, which gets around the molecules of varnish and remove them17. But close attention needs to be paid not to leave the gel too long on the surface of the artworks as water has unwanted effects on paint/ground/priming layers and canvas/paper/wood/cardboard supports18.
- 19 Wolbers, R., op. cit., p. 163.
- 20 Stavroudis, C., Blank, S., « Author’s Reply », in WAAC Newsletter 12, 1990, n° 2, p. 31.
21When reading studies or treatment reports where the soaps where used, one may notice that there are as many rinsing techniques as there are conservators: all the techniques are not appropriate. Wolbers19 writes about a rinse with demineralised water or saliva, then white spirit. The most complete rinsing procedure is given by Stavroudis and Blank20 and it has three steps: first rinse with a buffer solution that has the same pH as the soap: it dilutes the residues without precipitating them, as both the abietic and deoxycholic acids are in solution in a basic environment but precipitate and form crystals when the pH goes under 7.5. Second, rinse with a buffer solution pH 7 and one last rinse with an aliphatic hydrocarbon such as white spirit, Shellsol D40 or Shellsol T.
- 21 Erhardt, D., Bischoff, J., op. cit., p. 145.
22A paint layer undergoing treatment with resin or bile soaps needs to be able to stand a certain amount of mechanical action as well as successive rinses. Finally, it has to be the least porous possible and have as little open cracks as possible. The gel, even though it reduces the penetration of the liquid inside the layer, can be very difficult to remove from the cracks if pushed in during the cleaning21.
23A couple studies have been done on soaps, their working mechanisms and effects on the paint layer: Koller (1990), Ford and Byrne (1991), Bischoff and Erhardt (1993 &1994). The studies were done shortly after the introduction of soaps to the conservation world and thus they have little experience and perspective on this new cleaning system.
- 22 Koller, J., « Cleaning of a Nineteenth-Century Painting with Deoxycholate Soap: Mechanism and Resi (...)
24For instance, Koller’s study22 concludes that soap leave a great quantity of white residues after cleaning. The bile soap was rinsed only with white spirit, instead of an aqueous solution and then a hydrocarbon, as recommended by Wolbers. It is very likely the residues were caused by an improper rinse.
- 23 Erhardt, D., Bischoff, J., « The Roles of Various Components of Resin Soaps, Bile Acid Soaps and G (...)
- 24 Leaching is the extraction of low weight organic compounds presents in a polymerised film. These c (...)
25Then, Erhardt and Bischoff’s study23 used soaps which pH were around 10.3 instead of the maximum 8.5. Moreover, the soaps contained 25 % of TEA instead of 5 %, which is five times the recommended amount; the gel was left 15 minutes on the paint layer, which is quite long. It was concluded that soaps caused darkening and chromatic saturation of the paint layer after cleaning, high leaching24, changes in texture and softening, the majority of which are due to residues. These results could be explained by too much TEA in the soap and a, inappropriate rinsing procedure, causing many residues as well as leaching. Saponification and leaching could also happen due to high pH gels, long exposition to the soaps could cause softening and improper rinsing, more residues.
- 25 Ford, B., Byrne, A., « The lipid stripping potential of resin soaps gels used for cleaning oil pai (...)
- 26 Erhardt, D., Bischoff, J., op. cit., p. 16.
26Ford and Byrne’s study25 compares the leaching rates of different products used for cleaning varnishes: toluene, acetone (30 %), ethanol (30 %), DMF, ammonia (1 and 30 %), TEA and soaps. Then, the comparison was done between abietic and deoxycholic soaps. In the end, the study showed that soaps do leach, but less that solvents and much less than alkaline products; and that the abietic soap leached less than the deoxycholic. This particular difference is due to the difference in the CMC, which is twice higher in the deoxycholic soap: thus, the surface activity and detergence is twice as strong for a similar quantity of acid26. It is important to put the conclusion back in the context: are cleaning being done with pure toluene or DMF (now DMSO)? The comparisons done with ketone and alcohols at 30 % are closer to our actual practices. Checking the recipes used in the studies has been revealing and helped drawing conclusions that were consistent with the results of the studies. It is not possible to compare the results of Erhardt and Bischoff studies as the products concentration, the application and the rinses are different.
27One element that is not deniable is the fact that TEA becomes yellow/brown to light and air. This product is quite problematic is there are indeed residues left on the paint surface after cleaning.
28The author thought it would be interesting to replace the TEA used in soaps without diminishing the efficiency of the system. Observing the potential residues would be the next step but due to a restricted budget, only the uses of the fluorescent tracers and UV microscope were available. As these results were inconclusive, they will not be discussed in this article.
- 27 Personal communication with Richard Wolbers via email, October 2013.
29To replace the TEA, one must understand its role within the soap. TEA is an organic compound, formed by an alcohol and an amine. Thus, it combines the properties of both groups: amines act as a weak base and form salts (or soaps); alcohols are hygroscopic, can be esterified and can swell oil or varnish layers. TEA has a strong detergent ability due to its basicity and its surfactant properties. It is also a chelating agent. Finally, one of its particularities is its molecular structure, with both a polar and a non-polar part, enabling a rinsing with both water and hydrocarbons solvents27.
30Richard Wolbers was contacted and advised the following products to replace the TEA:
- 29 Personal communication with Dr. Spike Bucklow, informal discussion at the Hamilton Kerr Institute, (...)
- 30 Personal communication with Dr. Spike Bucklow, informal discussion at the Hamilton Kerr Institute (...)
31This last product is also used at the Hamilton Kerr Institute in Cambridge29. These products are not ideal to prepare buffer solutions as they are strong bases with respective pKa of 13, 13,5 and 9,2. Ammonia, even if considered a weak base, is highly volatile and even if the mixture is gelled, it is not stable. However, this particular aspect is what makes it interesting according to Dr. Spike Bucklow (Hamilton Kerr Institute) since the quantity of potential residues is diminished30. One last important element is the leaching properties of these ingredients, which does not make them suitable cleaning systems for oil paint layers.
32Four other products were added to the testing:
Bis Tris (2-[Bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol), weak base, tertianry organic amine, used for buffer solution in the biochemistry industry, pKa 6,5.
TrisBase (2-Amino-2-(hydroxymethyl)-1,3-propanediol), weak base, amine faible and chelating agent use in buffer solutions, pKa 8.
Ethomeen C25 (Poly(oxy-1,2-ethanediyl), a-hydroxy-w-hydroxy)( C30H63ClNO15X), tertiary amine ethoxylate, used in Wolbers’ solvent gels, pKa 7.
Di-sodium tetraborate (borax), bore mineral, used as insecticide, food preservative, detergent, pKa 9,1.
33These products were chosen according to their pKa in order to buffer the soap to the right pH. The new soaps pH ranged from 7,5 to 8,5.
34Resin and bile soaps were prepared with these new 7 bases, and an original soap with TEA was made as well for reference. Each of the 16 soaps was divided in 4 bottles (Fig. 3), to which different additives were introduced:
solvent (benzyl alcohol 3 %)
surfactant (Surfonic JL-80X 1 % replacing Triton X-100)
solvent (3 %) and surfactant (1 %)
35no additive for the last one, kept as a reference.
Fig.3 General view of the new soaps
64 soaps were prepared in total, they were then tested on different varnish layers.
Credits: Camille Polkownik.
36Soaps were then gelled with a cellulose ester. Despite many attempts, some soaps could not be prepared: the resin soap with sodium hydroxide, the resin soap with potassium hydroxide and the bile soap with ammonia. No explanation or conclusion could be drawn from these failures, even after consulting a chemist.
37Once the 64 soaps were ready, they were tested a different varnished oil paintings present at the moment at the painting conservation studio of La Cambre. After testing 20 varnish layers, only 7 reacted positively to soaps i.e. the varnish could be removed. The general observations are listed in the table underneath.
TEA, pH 7,8
Good removal. Chromatic saturation is the rinsing procedure is not thorough enough.
Bis Tris, pH 7,3
Little to no results, removal is very slow, could be due to the low pH of the soap (7,5).
TrisBase, pH 7,8
Very good results, as good or better than with the TEA soap. Easy removal without too much mechanical action.
Ethomeen C25, pH 7,6
Little to no results, removal is very slow, could be due to the low pH of the soap (7,5).
Di-sodium tétraborate, pH 8,3
Very good results, as good or better than with the TEA soap.
Sodium hydroxide, pH 8
Potassium Hydroxide, pH 8
Ammonia, pH 8,2
38The first case is an oil painting on panel, 17th century, attributed to a follower of David Ryckaert III (fig. 4).
Fig. 4 L'amateur, Follower of David Ryckaert III, 17th century
XVIIe century, Musée des Beaux-Arts de Liège, Belgium.
Credits: Militza Ganeva.
39This painting was covered with patches of old varnish, overpaints and a more recent varnish. While the upper varnish and the overpaints were removed quite easily with solvents, the old varnish bits could not, while the paint layer was sensitive to polar solvents used pure (fig. 5). Different soaps were tested and a successful cleaning was achieved with bile soaps containing TEA or borax or TrisBase, with and without additives. In the end, the borax soap was used with additives (benzyl alcohol 3 % and Surfonic JL-80X 1 %) as it gave the best results with the least amount of mechanical action (fig. 6). These results could be explained with the higher pH, and with the additives that boosted the initial action of the bile soap.
40The second case was an oil painting on canvas by Lamet, 18th century (fig. 7). Once again, old patches of varnish, but this time with dust embedded, and they could not be removed with solvents without sensitizing the paint layer (fig. 8). Moreover, the texture canvas was imprinted in the paint layer, giving a most irregular appearance after cleaning. After testing many cleaning systems (solvents, emulsions, Pemulen gels), soaps gave great results. TrisBase soap without additives was successfully used to homogenize the cleaning (fig. 9). The fact that no additives were needed was interesting as it really showed that the soaps work well and that the additives are not the active ingredients of the soap.
Fig. 5 Before cleaning patches of oxydized varnish
The remnants of varnish darken then paint layer and cause patches of unwanted gloss. They cannot be removed using traditional methods, as these sensitize the paint layer.
Credits: Camille Polkownik
Fig.6 After cleaning patches of oxydized varnish
The bile soap with disodium tétraborate (pH 8,5) removed the patches of oxidized varnish without sensitizing.
Credits: Camille Polkownik
41The second case was an oil painting on canvas by Lamet, 18th century (fig. 7).
Fig. 7 Portrait du Bourgmestre Maigret, Lamet
Œuvre datée de 1743, Musées de Verviers, Belgique
Credits: Jade Roumi.
42Once again, old patches of varnish, but this time with dust embedded, and they could not be removed with solvents without sensitizing the paint layer (fig. 8).
Fig.8 Before cleaning patches of oxydized and dirt-embedded varnish
The remnants of varnish have a layer of dirt embedded at the surface. They darken then paint layer and cause patches of unwanted gloss; and cannot be removed using traditional methods, as these sensitize the paint layer.
Credits: Jade Roumi.
Fig.9 After cleaning patches of oxydized and dirty varnish
The resin soap with TrisBase (pH 8,5) removed the patches of oxidized varnish without sensitizing.
Credits Jade Roumi.
43Moreover, the texture canvas was imprinted in the paint layer, giving a most irregular appearance after cleaning. After testing many cleaning systems (solvents, emulsions, Pemulen gels), soaps gave great results. TrisBase soap without additives was successfully used to homogenize the cleaning (fig. 9). The fact that no additives were needed was interesting as it really showed that the soaps work well and that the additives are not the active ingredients of the soap.
44In the end, the attempts at replacing the TEA were successful since two of the new products introduced were as if not more efficient than the soaps containing TEA. These products are the di-sodium tetraborate (borax) and the TrisBase. Soaps containing Ethomeen C25 and Bis Tris were inconsistent and generally very slow for a poor cleaning. This could be due to their low pH, maybe too low to break down the varnish molecules. Then, the soaps with sodium hydroxide, potassium hydroxide and ammonia never gave good results on the paintings that were part of the testing.
45Additives act as “boosters” to already efficient cleaning systems and are not the active ingredients of the soaps. They can reinforce the action of the soap is too slow or requires too much mechanical action. Benzyl alcohol as an additive is a way to lower the polarity of the mixture, which can be useful when confronted to a half degraded layer of varnish. It is of course difficult to assess with precision the degree of degradation, which is why it is interested to make a certain quantity of soap and then divide it to add different additives while keeping a bottle of untouched soap. This facilitates and accelerates the testing process and enables a targeted cleaning.