A preliminAry study into the swelling behAvior of

MODERN MATERIALS AND CONTEMPORARY ART

Elli Kampasakali Conservation Department, Tate London, UK [email protected] Bronwyn Ormsby* Conservation Department, Tate London, UK [email protected] Alan Phenix Getty Conservation Institute Los Angeles, CA, USA [email protected] Michael Schilling Getty Conservation Institute Los Angeles, CA, USA [email protected]

Tom Learner Getty Conservation Institute Los Angeles, CA, USA [email protected] *Author for correspondence

Keywords: acrylic, swelling, paint, cleaning, thermal, conductivity, pH

Abstract The swelling behavior of dried acrylic emulsion (dispersion) paint films during immersion in a series of wet cleaning systems (both aqueous and hydrocarbon solvents) was explored using thermal analysis techniques and light microscopy. Differences in the degree of swelling were observed with the type of solution or solvent, solution pH, solution conductivity, paint type (including pigment(s) used, acrylic copolymer type and paint brand) and the degree of aging. In general, the greatest swelling was observed with alkaline aqueous solutions and for paints containing synthetic organic pigments. Light aging caused a decrease in swelling capacity and aliphatic hydrocarbon solvents were confirmed as very low swelling systems. Despite the clear need for further research in this area, results from this preliminary study will help to inform conservators about the potential swelling behavior of acrylic emulsion paints during wet-cleaning treatments.

Résumé Le gonflement de films de peinture à base d’émulsion acrylique (dispersion) secs lors de leur immersion dans une série de systèmes de nettoyage humides (dans des solvants aqueux et hydrocarburés) a été exploré au moyen de techniques d’analyse thermique et par microscopie optique. Des différences dans le degré de gonflement ont été observées en fonction du type de solution ou de solvant, du pH et de la conductivité de la solution, du type de peinture (y compris le ou les pigment(s) utilisé(s), le type de copolymère acrylique et la marque de peinture) et le degré de vieillissement. De manière générale, le plus fort gonflement a été constaté avec les solutions aqueuses alcalines et pour les peintures contenant des pigments organi-

A preliminary study into the swelling behavior of artists’ acrylic emulsion paint films

Introduction

Swelling constitutes a change of physical and chemical state for any paint film, progressing from a rigid and dense structure to a more open and flexible one. Sometimes the phenomenon can be of real assistance to conservators, for example in the reduction of cupping or tenting in paint films. However, it is more commonly something to avoid. Some of the immediate consequences of swelling in a paint film include: a build-up of internal stresses within the film as it attempts to expand in size; the softening of the paint film, which can then imbibe dirt and surface deposits more readily; and the redistribution – or even extraction – of mobile components from within the paint film or at its surface (Wolbers 2000). Many conservation procedures for painted surfaces, in particular wetcleaning treatments, clearly carry a risk of swelling, so it is important for research to be conducted into ways in which the potential for swelling can be minimized. However, swelling is a difficult phenomenon to measure, and is influenced by a large number of variables. The swelling potential of dried, unpigmented acrylic latex films has been studied by a number of researchers (Snuparek 1976, Agarwal et al. 1999, Zumbühl et al. 2007), and found to be influenced by factors such as the type and amount of water-soluble materials present, polymer elasticity, copolymer composition, the particle size of the original latex, as well as film drying conditions and film age. More recently, Ormsby et al. (2007) explored the relative swelling responses of several acrylic paint (i.e. pigmented) films with dynamic mechanical analysis (DMA) and found significant differences in behavior between different brands of acrylic paint. A range of techniques have been used to measure swelling in other types of paint. For example, Phenix (2002) used a microscopy-based image analysis method to study the swelling behavior of oil paint films in organic solvents, where the lateral swelling of paint film fragments was recorded. This methodology was later modified by Zumbühl (2005) for studying the cross-sectional swelling of paint films. Ploeger et al. (2007) applied thermomechanical analysis (TMA) at room temperature to study the swelling of alkyd paints in toluene, hexane, n-butanol and distilled water, and concluded that the paints tested followed similar swelling trends, although to a lesser extent, to those exhibited by traditional oil paints. 1



ques synthétiques. Un léger vieillissement a entraîné une diminution de la propension au gonflement et il a été confirmé que les solvants d’hydrocarbures aliphatiques étaient des systèmes entraînant très peu de gonflement. Bien que des recherches additionnelles soient évidemment nécessaires dans ce domaine, les résultats de cette étude préliminaire aideront à informer les restaurateurs sur le gonflement potentiel des peintures à base d’émulsion acrylique lors des traitements de nettoyage humides.

Resumen La hinchazón de las películas de pintura de emulsión acrílica (dispersión) seca durante la inmersión en una serie de sistemas de limpieza en húmedo (tanto con disolventes acuosos como de hidrocarburos) se estudió por medio de técnicas de análisis térmico y microscopía óptica. Se observaron las diferencias en el grado de hinchazón en función del tipo de solución o disolvente, el pH de la solución, la conductividad de la solución, el tipo de pintura (incluyendo los pigmentos utilizados, el tipo de copolímero acrílico y la marca de pintura) y el grado de envejecimiento. En general, la mayor hinchazón se observó con soluciones acuosas alcalinas y con pinturas que contenían pigmentos orgánicos sintéticos. Un ligero envejecimiento provocaba una disminución de la capacidad de hinchazón, y se confirmó que los disolventes de hidrocarburos alifáticos son sistemas con muy baja hinchazón. A pesar de la clara necesidad de una mayor investigación en este campo, los resultados de este estudio preliminar aportarán información a los conservadores sobre el posible comportamiento de las pinturas de emulsión acrílica ante la hinchazón durante los tratamientos de limpieza en húmedo.

In this study, the swelling behavior caused by (mainly aqueous) cleaning systems commonly used by conservators to remove soiling from acrylic emulsion paint films was explored using TMA and the microscope-image analysis method developed by Zumbühl (2005). Experimental Materials

Two main brands [Winsor and Newton (W&N) and Liquitex] of professional quality artists’ acrylic emulsion paints were used; both containing the n-butyl acrylate/methyl methacrylate p(nBA/MMA) copolymer typical of paints made since the mid-1980s. These brands represent the highest (Liquitex) and lowest (W&N) swelling paints from previous studies carried out by the authors. Films were applied onto Teflon-coated stainless steel plates, using a Sheen instruments adjustable film caster, producing dry films of 110 ± 20 µm thickness. The paints tested contained single pigments: an inorganic pigment, titanium white (PW6); and two synthetic organic pigments, azo yellow (PY3) and phthalocyanine green (PG7). Talens azo yellow was also tested to explore the ethyl acrylate/methyl methacrylate p(EA/MMA) acrylic copolymer type. At the time of testing the naturally aged paint films were ~6 years old. Selected films were light aged at 15,000 lux for 16 weeks under Philips TLD 58W/840 daylight tubes, with the UV component filtered by Perspex acrylic sheet. The liquids used to examine the swelling responses of the above paint films were selected to represent a broad range of chemical types from within the general spectrum of liquids employed by conservators for cleaning acrylic paintings (see Table 1); both aqueous solutions and hydrocarbon solvents were included. These included water, pH-adjusted water, water + nonionic surfactant, and water + chelate (triammonium citrate, TAC), plus a series of TAC solutions (0.25-1.0% in water – see Table 2) prepared with their pH adjusted with HCl or NaOH to four different pH values (4, 5, 6 and 8). Two hydrocarbon solvents were also tested: Shellsol D60 (boiling range 193–207°C; c. 0.1% aromatics), and a mixture of 17% Shellsol A100 in Shellsol D60 (to simulate a white spirits-type solvent). Table 1 Cleaning systems tested for their swelling effect Cleaning system

pH

Conductivity (µS/cm)

Deionized water

6.5

4

Shellsol D60 (hydrocarbon solvent, 0.1% v/v. aromatics)

-

-

Shellsol D60 + 17% v/v aromatics

-

-

5.5

105

1% v/v. Triton® XL-80N in deionized water 1% w/v. triammonium citrate (TAC) in deionized water

8

9000

Acetic acid solution

4.5

108

NH4OH solution

9.5

160

NH4OH solution

11

580 2



Analytical methods

Table 2 Concentration and conductivity of TAC solutions Concentration of triammonium citrate (TAC) (% w/v.)

Conductivity (mS/cm)

0.25

2.8

0.5

5.1

0.75

7.3

1

9.0

TMA swelling experiments were performed using a Mettler Toledo TMA/SDTA841e instrument. Data was collected using Mettler STARe software v8.10, using a similar procedure to Ploeger et al. (2007). Small pieces (~0.5 × 0.5 cm) of each film were cut, left to relax for 24 hours, and then placed in a silica cup with slotted Teflon lid (to slow solvent evaporation). The cup was positioned under a glass, flat-faced probe (3 mm diameter; set to 0.02 N applied force), to make contact with the middle of the paint film. 300 µL of each solvent/solution (Table 1) were injected directly into the cup using a micro-pipette. All experiments were conducted at room temperature (~22°C) and at 25–35% RH. Data were recorded every second with Seiko Instrument software. For most tests, a sudden peak was observed in each swelling curve when the test solvent/solution was added, due to the buoyancy of the sample when liquid was introduced. In calculating the percentage swelling of each sample, the first data point after this peak was considered the starting point of each experiment, with an endpoint at 180 minutes (or 120 minutes if swelling had reached a plateau by then). Microscopy/image analysis swelling experiments were carried out using a Wild Makroskop M420, 6.3–32 x stereo zoom microscope fitted with a Diagnostic Instruments ‘Spot’ RT Color digital camera, operated by DI ‘Spot Advanced’ software. Paint films were cut into thin strips, and clamped at both ends by two horseshoe-shaped glass pieces, held together by a metal spring running around the outside of the overall assembly, as shown in Figure 1. The whole system was then placed under the microscope for imaging. Each cleaning system was added to the ring via a dropping funnel, taking care not to disturb the paint film. Images were captured at regular intervals for five minutes and the curve height was measured using Adobe Photoshop 7.0 software. Each paint sample surface was also analyzed with Fourier transform infrared-attenuated total reflection spectroscopy (FTIR-ATR), using a Nicolet Avatar 360 spectrometer with a Germanium ATR crystal (200 scans at 4 cm-1 resolution), and processed with Omnic 6.2 software. Results

Figure 1 Sample set-up for microscopy based swelling experiments: a) Liquitex PG7 sample clamped between two glass pieces before swelling; b) same sample after swelling in water

Figure 2 illustrates the range of TMA swelling curves for Liquitex titanium (Ti) white films, which show typical swelling profiles in terms of absolute values of thickness change, i.e. a rapid initial swelling rate that decreases over time until a plateau is reached. It should be noted that this swelling response overrides a natural tendency for the TMA probe to continue compressing the paint sample, which was seen in all samples prior to the addition of the test solvent/solution (although not shown in Figure 2). The main exception to this typical profile was obtained with the Shellsol D60 solvent, where no swelling was noted for any of the films tested, and in fact for D60 the curve is dominated by the natural compression of the 3



sample. The other exception was with the highest swelling ammonium hydroxide (NH4OH) solutions, where a second hump was observed in the Liquitex paint samples between 10 and 15 minutes, which corresponded to the temporary curling of the paint film edges. This feature was clearly related to the strong alkaline environment; since this response did not occur in the corresponding swelling curves of the W&N films, it may reflect differences in paint formulation. Figure 3 shows the percentage swelling (via TMA) for all paint films immersed in the range of aqueous solutions, plotted as a column graph against solution pH. The graph illustrates that the Liquitex paint films consistently swelled more than the W&N or Talens films. Of the cleaning solutions tested, the NH4OH solutions (pH 9.5 and 11.0) resulted in the greatest swelling, while the TAC (pH 8.0) solution had the least effect – probably in part due to the relatively high ionic strength of this solution. Within the acidic range, the acetic acid solution (pH 4.5) tended to result in slightly more swelling than the Triton® XL-80N (pH 5.5) solution or the deionized water (pH 6.5). It was also observed that the films containing synthetic organic pigments (azo yellow or phthalocyanine green) swelled more than the corresponding titanium white films.

Figure 2 TMA: swelling curves for Liquitex titanium white free films in selected solvents and aqueous solutions Figure 3 TMA: swelling effect of the cleaning systems on several acrylic emulsion paint films: acetic acid solution pH 4.5; 1% Triton XL-80N solution pH 5.5; deionized water pH 6.5; 1% TAC solution pH 8; NH4OH solutions pH 9.5 and pH 11

The behavior of the Talens azo yellow paint film resembled that of the W&N films, where the degree of the swelling was generally low. In this case, the TAC solution also caused the least swelling, but it was interesting to note that for this film, the mildly acidic Triton® XL-80N solution caused a similar degree of swelling to the alkaline NH4OH solutions. This may indicate the presence of an interaction between the Triton® XL-80N surfactant with similar surfactants present in the paint films (e.g. Triton ® X-405), which requires further investigation. The non-polar solvent Shellsol D60 did not swell any of the samples tested including the Talens azo yellow sample. However, increasing the aromatic content of the solvent to 17% v/v. (by adding Shellsol A100) induced a slight swelling of the white and yellow Liquitex paint films, although the degree of swelling remained lower than even the lowest swelling TAC solution. The alkaline pH solutions are likely to affect any acrylate salts present in the films, causing them to ionize, which in turn would lead to a higher affinity for aqueous solvents and thus higher swelling in the film (Peppas 2008, Zagorodni 2007). However, the distinguishably higher swelling caused by the NH4OH solutions and the curling observed in the Liquitex films indicated an additional action, possibly involving the partial hydrolysis of the acrylic polymer. A degraded polymer would provide less resistance to deformation, facilitating faster penetration of the solution (Brown et al. 1953). It is likely that the fast swelling rate witnessed in the first 10–15 minutes (as seen in Figure 2) is due to soluble components being washed away by the aqueous cleaning systems, and as these are gradually removed, the swelling rate decreases. In a study of water penetration through waterborne acrylic coatings, Whitmore et al. (2007) suggested that the removal of 4



water-soluble components reduces the subsequent water-permeability of the film. Surfactants are known to be extracted from bulk paint films during water immersion, and the removal of surfactant from the surfaces of the acrylic paint films with polar solvents has been demonstrated repeatedly (Ormsby et al. 2007, Ormsby et al. 2009). It appears that the amounts of surfactant present on paint film surfaces did not affect the relative degree of swelling. FTIR-ATR spectra were taken from the surface of each paint film before immersion (Figure 4), where the characteristic polyethoxylate (PEO)-type surfactants bands at 2890, 1343, 1110 and 964 cm-1 are clearly visible. High levels of surfactant were detected on the Talens azo yellow film, but much less so on the W&N films, and only trace amounts on the Liquitex films. In other words, while the Liquitex films had almost no surfactant on their surface, they exhibited the greatest swelling potential (indicating that there may be significant amounts of water-extractable surfactant remaining in the bulk film). This was also assessed by testing the swelling capacity of waterswabbed paint samples, which were found to be indistinguishable from the untreated controls.

Figure 4 FTIR-ATR spectra of the untreated azo yellow acrylic emulsion paint films. The arrows indicate the characteristic surfactant peaks Figure 5 TMA: Effect of solution conductivity on the swelling behavior of Liquitex azo yellow film

To begin to investigate the complex relationship between aqueous solution conductivity, pH and swelling, the highest swelling film (Liquitex azo yellow) was immersed in a series of TAC solutions over a pH range from 4 to 8, as described in Table 2. The results shown in Figure 5 indicate that there was generally a slight decrease in swelling with increased solution conductivity at all pH levels and that pH appears to have a more pronounced effect on swelling behavior than solution conductivity. A pH of 4.0 appears to have caused the least swelling for all levels of ionic strength tested (which incidentally was not found to be the lowest swelling pH for the range of solutions explored in Figure 3). Within the range of conductivity values (2–9 mS/cm) tested, the lowest swelling combination for this paint film was found to be pH 4.0 and a conductivity of 7.0 mS/ cm; the reason for the low swelling action of this particular solution warrants further exploration. The effect of ageing on swelling behavior was also investigated via the TMA method. Samples from a light aged Liquitex titanium white film were immersed in water (a “medium-swelling” system), in 1% w/v. TAC solution (least swelling) and in the pH 11 NH 4OH solution (most swelling). Assuming reciprocity, the aging was equivalent to ~50 years exposure under normal museum conditions. The aged samples followed the same trend for all solutions tested; however, the degree of swelling was less than for the equivalent unaged samples (see Figure 6). Decreased swelling in light-aged oil paint films was attributed by Phenix (2002) to a lower organic content after the loss of volatile scission products, and to an increase in paint density for the aged samples. The reduced swelling exhibited by aged Liquitex films could also be due to these factors, in addition to an increase in cross-linking of the polymer (Learner et al. 2002, Smith 2007). It is known that cross-linked acrylic films are more 5



rigid and swell less (Whitmore et al. 1995, Zagorodni 2007) and that water-soluble materials are harder to extract from dense, more highly cross-linked films (Snuparek 1996). In parallel with the TMA swelling experiments, Liquitex phthalocyanine green films were imaged (in cross-section) during immersion via the previously described microscopy method established by Zumbühl (2005). The characteristic response was a curving/bowing of the film cross-section during immersion, which varied according to the solution used. Figure 7 depicts the increase in curve height (distortion of the sample as illustrated in Figure 1) over time. Although the order of the aqueous solutions – as dictated by the curve height – is not exactly the same as the order obtained via the TMA experiments, image analysis confirmed that the alkaline solutions had the strongest swelling effect on the phthalocyanine green films, as was also observed with TMA, and once again (although not shown) no swelling was observed for the Shellsol D60 aliphatic hydrocarbon solvent. Conclusion

Figure 6 TMA: Comparison of swelling percentage between unaged and light aged Liquitex titanium white films: deionized water pH 6.5; 1% TAC solution pH 8; and NH4OH solution pH 11 Figure 7 Microscopy: cross sectional changes of Liquitex phthalocyanine green films during the immersion in different aqueous solutions

The TMA and microscopical methods for measuring paint film swelling both demonstrated the significant differences in swelling power across the range of aqueous and organic liquids tested. Swelling potential proved greatest with aqueous alkaline solutions, which may have implications for the removal of emulsion-based surface coatings – which often require high pH systems. Somewhat encouragingly, the equivalent aged acrylic emulsion paint samples tended to swell less – presumably due to the natural loss of low molecular weight material from the bulk film over time, increased polymer coalescence and/or chain entanglement and small amounts of cross-linking. The low aromatic content aliphatic hydrocarbon solvents tested were confirmed as very low swelling in all cases. The group of aqueous systems caused similar, moderate amounts of swelling, with the TAC solutions inducing the least swelling. The role of the ionic strength of aqueous solutions requires further exploration, as swelling potential appears to decrease with increased cleaning solution conductivity. On the whole, swelling behavior did not appear to be overly influenced by differences in acrylic copolymer type; hence, detectable differences in swelling rates were probably due to paint formulation differences, including the amounts of hydrophilic additives present. It was noted that paints containing synthetic organic pigments were slightly more prone to swelling (in aqueous systems) than titanium white equivalents, which may be due to the typically higher medium content of these films. The highest swelling organic pigmented paint film was found to swell the least in aqueous solutions at pH 4.0, with a solution conductivity of ~7 mS/cm.

6



Although further research is required to explore the effects of pH and solution conductivity on the swelling and extraction of material from bulk acrylic paint films, it is possible that slightly acidic aqueous systems with moderate solution conductivity may reduce swelling potential during wet surface cleaning treatments, particularly when aliphatic hydrocarbon solvent options are not appropriate. Acknowledgements

The authors would like to thank AXA Art Insurance for funding the Tate AXA Art Modern Paints project, in addition to Dr. Rebecca Ploeger (ex-University of Torino) and Dr. Stefan Zumbühl (University of Applied Sciences, Bern). References Agarwal, N., and R. Farris. 1999. Water absorption by acrylic based latex blend films and its effect on their properties. Journal of Applied Polymer Science 42: 1407–1419. Brown, G., and J. Scullin. 1953. Water penetration of emulsion polymer films. Industrial and engineering chemistry 743–745. Learner, T., O. Chiantore, and D. Scalarone. 2002. Ageing studies of acrylic emulsion paints. In ICOM-CC 13th Triennial Meeting, Preprints, Rio de Janeiro, 22–27 September 2002, ed. R. Vontobel, Vol. II, 911–919. London: James & James. Ormsby, B., T. Learner, G. Foster, J. Druzik, and M. Schilling. 2007. Wet-cleaning acrylic emulsion paint films: an evaluation of physical, chemical and optical changes. In Modern paints uncovered, eds. T. Learner, P. Smithen, J. Krueger, and M. Schilling, 187–198. Los Angeles: Getty Conservation Institute. Ormsby, B., E. Kampasakali, C. Miliani, and T. Learner. 2009. An FT-IR based exploration of the effects of wet cleaning artists’ acrylic emulsion paints. e-Preservation Science 6: 186–195. Peppas, N., and C. Bures. 2008. Glucose – responsive hydrogels. In Encyclopedia of biomaterials and biomedical engineering, 2nd edition, Vol. 2., eds. D. Wnek and G. Bowlin, 1163–1173. New York: Informa Healthcare. Phenix, A. 2002. The swelling of artists’ paints in organic solvents. Part 1, a simple method for measuring the in-plane swelling of unsupported paint films. Journal of the American Institute for Conservation 41(1): 43–60. Ploeger, R., A. Murray, S. Hesp, and D. Scalarone. 2007. Morphological changes and rates of leaching of water-soluble material from artists’ acrylic paint films during aqueous immersions. In Modern paints uncovered, eds. T. Learner, P. Smithen, J. Krueger, and M. Schilling, 201–207. Los Angeles: Getty Conservation Institute. Smith, G.D. 2007. Aging characteristics of a contemporary acrylic emulsion used in artists’ paints. In Modern Paints Uncovered, eds. T. Learner, P. Smithen, J. Krueger, and M. Schilling, 236–246. Los Angeles: Getty Conservation Institute. Snuparek, J. 1976. Some factors affecting the water absorption of films from synthetic lattices. Part II: Particle size and latex stability. Journal of the Oil and Colour Chemists Association 59: 19–21. Snuparek, J. 1996. Some aspects of water absorption in free films from non-pigmented copolymer latex binders. In XXIII FATIPEC Congress Brussels, Vol. B: B232–B244. 7



Whitmore, P., H. Morris, and V. Colaluca. 2007. Penetration of liquid water through waterborne acrylic coatings. In Modern paints uncovered, eds. T. Learner, P. Smithen, J. Krueger, and M. Schilling, 217–226. Los Angeles: Getty Conservation Institute. Whitmore, P., and V. Colaluca. 1995. The natural and accelerated aging of an acrylic artists’ medium. Studies in Conservation 40: 51–64. Wolbers, R. 2000. Cleaning painted surfaces: aqueous methods. London: Archetype. Zagorodni, A. 2007. Ion exchange materials, properties and applications. Elsevier. Zumbühl, S. 2005. Illusion mit System – das Lösemitteldreieck in der Praxis, Aspekte zur Charakterisierung der Wirkung von binären Lösemittelmischungen. Zeitschrift für Konservierung und Restaurierung 19(2): 253–262. Zumbühl, S., F. Attanasio, N. Scherrer, W.Müller, N. Fenners, and W. Caseri. 2007. Solvent action on dispersion paint systems and the influence of morphology – changes and destruction of the latex microstructure. In Modern paints uncovered, eds. T. Learner, P. Smithen, J. Krueger, and M. Schilling, 257–268. Los Angeles: Getty Conservation Institute.

8