Conservation Issues in Several Twentieth-Century Canadian Oil Paintings: The Role of Zinc Carboxylate Reaction Products

Abstract

In an on-going study of the materials and techniques of twentieth-century Canadian painters, similar conservation issues in oil paintings by various artists have been noted. These include delamination and lifting paint, zinc soap protrusions and surface efflorescence or accretions. Examples of these phenomena are presented. Delamination in an oil painting from 1956 was found to be related to an underlayer with a high concentration of zinc fatty acid salts (zinc soaps). In two paintings that date from 1936 and 1937, zinc soaps have aggregated and formed protrusions that have broken through the paint surface. The protrusions were analysed using a combination of SEM-EDX, GCMS and FTIR. The FTIR spectra were compared to those of synthetic zinc palmitate, stearate, azelate and oleate. The combined GCMS and FTIR results indicate that the protrusions contain primarily zinc palmitate and stearate. Peak splitting in the FTIR spectrum, which is not observed in synthetic zinc palmitate, stearate or binary palmitate-stearate salts, is likely due to structural distortion. The final example describes a disfiguring surface accretion on a 1952–1954 painting caused by the reaction of zinc with a low molecular weight carboxylic acid (2-hydroxypropanoic or lactic acid).

Appendix

Cross-section samples were mounted in polyester resin, ground and polished using standard petrographic techniques, and observed using incident light and fluorescence microscopy with a Leica DMRX microscope. Small fragments of unmounted samples were analysed using a combination of the techniques described below.

For FTIR, samples were mounted on a low pressure diamond anvil microsample cell and analysed using a Bruker Hyperion 2000 microscope interfaced to a Tensor 27 spectrometer. The samples were analysed in transmission mode by co-adding 200 scans and using a 4 cm⁻¹ resolution. Spectra were collected from 4000 to 430 cm⁻¹ using a wide band MCT detector or from 4000 to 550 cm⁻¹ using a midband detector.

GCMS analysis was performed on small fragments of the protrusion samples (approximately 20–30 μg) using an Agilent 6890 gas chromatograph interfaced to an Agilent 5973 quadrupole mass spectrometer. Samples were extracted and derivatised using two separate methods: methylation and trimethylsilylation. For silylation, the samples were treated with 30 μL of N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA; Pierce) for 30 min at 60 °C to form trimethylsilyl (TMS) derivatives. For methylation, the samples were treated with a 1:1 solution of 15 μL of Meth-Prep™ II (m-(trifluoromethyl)phenyl trimethylammonium hydroxide, TMTFTH, 0.2 N in methanol; Grace Davison Discovery Science) and 15 μL toluene for 60 min at 60 °C to form methyl derivatives. Because only micro-samples of the zinc lactate surface accretion (approximately 3 μg) were available, Py(TSP)-GCMS was employed. For this method, an Agilent TSP was used for direct sample introduction to an Agilent 7890 gas chromatograph interfaced to an Agilent 5975 triple-axis quadropole mass spectrometer. Samples were derivatised using 1.5 μL hexamethyldisilazane (HMDS; Supelco) to form trimethylsilyl derivatives. The TSP was ramped from 50 to 450 °C at a rate of 900 °C/min and held at that temperature for 3 min before cooling to 250 °C for the remainder of the run.

SEM-EDX was undertaken using an Hitachi S-3500 N scanning electron microscope integrated with a lithium-drifted silicon, light element X-ray detector and an Oxford Inca X-ray microanalysis system. The SEM was operated at an accelerating voltage of 20 kV and a working distance of 15 mm. Using this technique, elemental analysis of small volumes, down to a few cubic micrometers, were obtained for elements from boron (B) to uranium (U) with a sensitivity of about 1 %.

XRD patterns of selected paint samples were obtained with a Bruker D8 Discover with GADDS (General Area Detector Diffraction Solution) equipped with a rotating anode and cobalt target. The patterns were measured at 40 kV and 85 mA, using a 0.5 mm collimator.

The synthesis of the zinc fatty acid salts was based on a published method (Robinet and Corbeil 2003). A solution of the fatty acid in anhydrous ethanol was gently heated until dissolved and subsequently decanted into an aqueous sodium hydroxide solution held at 60 °C. The solution was heated under reflux at 98 °C overnight. The resulting sodium salt was filtered, dried and subsequently dissolved in ultrapure water and added to an aqueous zinc chloride solution, where zinc chloride was present in excess. In the case of the binary salts, the intermediate sodium salt was not isolated, and the zinc chloride was added directly to the aqueous sodium fatty acid salt solution. The purity of the synthetic fatty acid salts was confirmed by comparison of the FTIR spectra to published data (Robinet and Corbeil 2003; Price and Pretzel 2007; Taylor and Ellis 2007) and by a determination of the fatty acids present using GCMS. In the case of the 1:1 binary salt, where a published infrared spectrum was not available, the composition was confirmed by X-ray diffractometry using a Rigaku Ultima IV equipped with a scintillation counter. The diffractogram was measured in step mode from 1° to 40°, 2θ (step size: 0.02°, count time: 1 s). A single set of long spacing lines, mid-way between those of zinc palmitate and zinc stearate were obtained, consistent with the published data for this compound (Mesbah et al. 2011).