FIGS 4–7 include the curves for temperature and RH% values just above the surface of the canvas and provide an indication of the conditions close to the painting. The temperature values given in the 'listing of events' are from the thermocouple placed temporarily on the painting surface. Tension in the graphs is expressed in Newtons (N).
FIGS 4a, b and c show the change in tension in the warp and weft directions during each of the three stages of the wax-lining process for the lining canvas (4a), painting (4b) and final lining (4c). Each process causes similar tension changes as the adhesive melts or solidifies during the application of vacuum pressure. Rapid solidification of the liquid wax occurs once it has been applied to the lining canvas. On heating, the wax starts to melt again causing a reduction in canvas tension (fig.4a); this occurs throughout all the stages of the wax lining as the adhesive begins to melt at around 40-45°C. Over a 22-minute period the tension is lowered from 50N and 45N, to 16N and 6N in the warp and weft directions respectively. There are two mechanisms that contribute to the tension reduction. Firstly, there is thermal expansion of the canvas caused by the direct application of the hot liquid wax. Secondly, lubrication by the liquid wax of the yarns in the linen canvas reduces friction at the points where the yarns cross over one another in the weave. This enables the yarns to move easily, thereby allowing a relaxation of the stresses created during the pre-wetting and pre-stretching process and possibly some residual stress induced during manufacture. On cooling, with the vacuum on, there is an initial and immediate increase in tension that is associated with a visible and rapid solidification of the wax. Both tension response and fall in temperature are approximately exponential. As room temperature is reached, after 75 minutes, tension stabilises to 48N in the warp and 100N in the weft. Tension is reduced by approximately 10N when the vacuum pressure exerted on the painting is removed.
As the wax-resin, applied to the reverse of the painting, melts there is a similar decrease in tension, from 48N and 43N before wax application to 5N and 7N in the warp and weft respectively (see fIg. 4b). This occurred more quickly than for the lining canvas, taking only 11 minutes because the painting was heated more rapidly. The tension eventually stabilised to 61N in the warp and 78N in the weft after approximately 70 minutes.
FIG 4c shows the tension response during the final lining process. The initial peaks in tension occur as the painting is placed on top of the prepared lining canvas tensioned in the tester. As heat is applied and the adhesive (wax-resin) melts, the tensions drop from 38N and 53N to 14N and 4N in warp and weft respectively. Tension stabilises at 40N in the weft and 95N in the warp after approximately 70 minutes. This large discrepancy between the warp and weft final tension, when compared to the preparation of the painting, is probably because the behaviour of the lined painting was dominated by the characteristics of the lining support rather than that of the painting.
For all of the above stages there was an increase in tension once the wax-resin had solidified. This is caused by a combination of contraction of the wax-resin with decreasing temperature and an associated locking of the canvas weave which will reduce residual creep.
Glue-paste lining process
FIG 5a shows tension change in the warp and weft directions during the lining operation. A new piece of linen without adhesive was tensioned and the painting, with the wet paste having been applied to its reverse, was positioned on top of the lining canvas. An initial decrease in the tension of 11N in the warp and 6N in the weft was observed. After rolling the surface and application of heat the tension reached 34N in the warp and 54N in the weft. (Although a temporary loss of vacuum occurred, the tension stabilised at around these values.)
FIG 5b shows tension change in the warp and weft directions during the application of glue-paste to the lining. The sample was used for comparison in the environmental tests. As the glue-paste is applied to the lining canvas there is an initial but small loss in tension of 10N in both warp and weft, followed by a steady rise in tension as the adhesive dries in room conditions. Tension stabilised at 113N in the warp and 127N in the weft after 40 minutes.
The increase in tension on drying of the glue-paste results from the contraction of the glue and, as with the wax-resin, a locking of the canvas weave.