The technical examination & conservation of painted furniture

As a student specialising in furniture on the RCA/V&A Conservation Course, with a particular interest in painted surfaces, I welcomed the opportunity to work on an English carved and painted armchair (W18A-1931), dated about 1785, in preparation for its inclusion in the new Ornament Gallery in the V&A.

Figure 1. Before conservation, detail showing the paint losses on the foliage patera (click image for larger version)

The chair is one of a set from the Music Room at Woodhall Park, Herefordshire, and has been tentatively attributed to George Brookshaw. The main colour is pale blue with the detail of the decoration picked out in green, red, black, orange and yellow. The cane-work seat has a loose cushion covered with silk. The back is lyre-shaped with the top rail being carved with foliage and ending in lions' heads. The splat is formed of bars crossing diagonally and pierced to represent flutes. They are interlaced with ribbons and in the centre is a round medallion painted in grisaille with a female figure playing the harp. Another chair (W18-1931) of this set in the V&A is identical but for the painted medallion which shows a female figure holding a tambourine. The arms of the chair represent dolphins, the heads terminating in foliage and the bodies painted to represent scales. The spaces below the arms are filled with slender rods representing harp strings. The seat rail is curved in front and is carved with a ribbon moulding. The turned and tapered legs have a twisted ribbon decoration and square blocks at the top with foliage paterae above capitals carved with upright leaves.

From a preliminary examination of the chair with the naked eye and at low magnifications (Figure.1&2), it was evident that it had been extensively restored in the past and that many paint losses had occurred. It was clear that before making any attempt at cleaning or consolidation, it would be useful to determine the extent of the damage and repainting through technical examination. Small paint samples were therefore taken from areas adjacent to damage, mounted in clear polyester embedding resin and polished cross-sections were prepared. These were examined in incident light at x175 and x345 magnification and ultraviolet light at x250 magnification. Microchemical tests were also employed to differentiate between certain pigments where this was not possible by eye.

Many of the samples examined showed the original ground and paint layers distinctly covered with layers of filler and non-original overpaint. The pigments identified in the original layers were consistent with late eighteenth century usage. Of the white pigments, the ground was found to contain chalk but the white used in the main pale blue paint of the chair proved neither to be a carbonate nor to contain lead. It could not therefore be lead white as might have been expected. The only other common white pigment available in the late eighteenth century was zinc white. While the use of this pigment has not been confirmed by further analysis, it is assumed that zinc white has indeed been used here.

The coloured pigments positively identified in the original paint layers included Prussian blue, orange ochre, vermilion and bone black. Massicot (synthetic lead oxide) and Naples yellow (lead antimonate) were tentatively identified, along with the zinc white.

Figure 2. Before conservation, detail showing the paint losses on the dolphin head arm. (click image for larger version)

Staining of the cross sections with appropriate dyes suggests that the original paint and ground were applied in a protein medium, probably glue size. This may have been mixed with some oil to give an emulsion. The original varnish layers did not stain for protein but, in ultra violet illumination, fluoresced the characteristic yellow-green colour of an oil based varnish.

The layer of filler applied over the original layers varied considerably in thickness to hide irregularities and paint losses. It fluoresced a bright yellow in ultraviolet light, characteristic of a resin medium, and is probably shellac. The repaint, in an oil medium, contained several pigments which were introduced in the nineteenth century. These included synthetic ultramarine and Emerald green. Synthetic ultramarine was first manufactured in 1824 and so the repaint must be later than this. Emerald green, synthetic copper aceto-arsenite, was first made in 1814 and went out of general use before the end of the nineteenth century because of its high toxicity.

A first varnish layer visible over the repaint did not fluoresce in ultraviolet light, suggesting a synthetic or non-oil-based coating. A second layer of varnish, however, did fluoresce yellow-green suggesting an oil-based coating.

Having established the layer structure, pigments and binding media involved, the first stage of the conservation treatment was begun. This involved consolidation of the paint and securing a loose back splat. The flaking paint was consolidated with warm dilute isinglass (a high quality fish glue) applied with a syringe. This aqueous adhesive was found to be particularly effective. It helped to soften the paint and ground layers prior to flattening with a heated spatula, presumably because an aqueous binding medium had been used.

The next stage was to re-adhere the loose back splat, after first removing the old glue. Here the aim was to avoid the danger of damage to the original paint layers while making the repair. The heat and moisture involved in the use of a hot animal glue would have been hazardous to the original paint and ground layers, particularly in the removal of any excess adhesive. It was therefore decided to use a cold, synthetic adhesive, poly(vinyl acetate) which is readily swellable and hence reversible using acetone.

A number of solvents and soaps, both in their free states and in gel formulations, were tested for the removal of the non-original overpaint layers. Eventually, however, it was decided to use Vulpex liquid soap (potassium methyl cyclohexyl oleate) followed by acetone on cotton wool swabs. This is a method I have had much experience of and, while it is likely that some of the other formulations tried would also have been effective, it seemed best to use a technique which I could easily control. There proved to be a distinct interface between the layers of overpaint and original paint, probably again largely as a consequence of the difference in binding media, so it was possible to avoid damage to the original materials with reasonable skill and care.

Figure 3. After conservation treatment (click image for larger version)

After cleaning and further consolidation, damage and paint losses were infilled with a gesso putty made of gilder's whiting and parchment size. There was no attempt to fill all of the enormous number of losses of original paint and ground. The choice of what to fill and what to leave unfilled was made in consultation with curatorial staff. The intention was to avoid making the chair look as if losses had never been sustained. Whereas in restoring a painting it may be considered appropriate to fill and texture every loss carefully, in furniture conservation it is considered better to leave more evidence of damage as an indication of an object's history.

After filling, an isolating varnish of Paraloid B72 in xylene was applied. For inpainting, dry ground artists' pigments dispersed in Paraloid B72 were used. After inpainting, a further 'sealing' coat of B72 was applied overall. The final varnish was applied in two coats by brush. The first was of Ketone Resin N alone. This was followed by a coat of Ketone N with 3% Cosmolloid 80H (microcrystalline wax) added as a matting agent. This use of materials and succession of layers is quite commonly used by picture conservators but is still relatively new to furniture conservation. While admittedly this approach would be quite inappropriate in the treatment of most of the objects encountered by a furniture conservator, it seems reasonable for this particular object and perhaps wherever the painted surface is the main consideration. While the aesthetic decisions made for each category of object - furniture, painting, sculpture - may vary, the physical problems presented by the paint layers and the substrate remain essentially the same (Figure. 3).

Many thanks are due to Josephine Darrah of the Science Section who made the technical examination of the chair, which is now on display in the Ornament Gallery.