Cross section of a reconstructed painting (a) and UV induced fluorescence micrograph of a comparable sample stained with Sypro (b). The protein glue has a bright red fluorescence.
Experimental diagram of neutron radiography tests.
Isometric view of the reaction chamber (a). Aluminium base in black; Teflon holder in grey; sample in beige. Setup at the ICON beamline (b).
Example of a raw neutron radiography image of the sample in the sample holder.
False colour neutron radiography normalised image of the sample (upper) and of moisture distribution (lower) at time step 100 (Sample areas 1. Glue sized canvas and 2. Ground layer).
Moisture uptake profiles at increasing time intervals after exposure to high relative humidity.
Curves of moisture absorption (left) and drying (right) over the thickness of an oil painting sample. Data from magnetic resonance tests, obtained with the so-called NMR PM5 Mouse.
A link was made with the recently completed technological study of the early oeuvre of the Swiss painter Cuno Amiet (1868-1963) (Kunsttechnologische Forschungen zur Malerei von Cuno Amiet 1883–1914). The samples prepared and tested were based on materials used in late 19th century and early 20th century paintings, which had been identified in the oeuvre of Amiet.
Hydrogen has a high neutron scattering cross section which makes neutron radiography a highly suitable technique for moisture interaction studies. For the study of the spatial distribution and kinetics of such interaction, a moisture exposure chamber was purpose built for neutron radiography experiments. New collaborations with EMPA (Dübendorf) and the Technical University of Eindhoven allowed to obtain relevant material parameters and to use magnetic resonance (NMR) as an alternative technique to measure water uptake in the paint layers.
The first results on a simplified build-up system (canvas, size and ground) showed that we can measure the moisture uptake quantitatively in a time-resolved experiment and with spatial resolution of 13.5 micrometers, sufficient to distinguish the contribution of the different layers of a typical paint multiplex. The flax fibres of the canvas and the glue sizing appeared to have a much stronger water uptake than the chalk glue ground. We also found that the way the glue size was applied had an influence on the distribution of the glue, and therefore also on the location where more moisture was absorbed. The findings were published in Journal of Applied Physics A.
The observations could be clarified by the outcome of extensive material characterisation, undertaken in the labs of EMPA. The moisture capacity and permeability of the different layers within a painting was measured in cup tests and by Dynamic Vapour Sorption. The results were submitted to the Journal of Cultural Heritage, and have been published online.
During the final phase of the project the gathered results of three experimental campaigns at the neutron beamline ICON and of two sessions of NMR measurements (June 2015 at University College London and November 2015 in Eindhoven) were re-analysed in order to draw conclusions about the behaviour of different types of painting build-ups. These include samples with and without oil paint layer, with and without lining, and with various types of glue sizing. The manuscript with working title “The distribution of moisture in reconstructed oil paintings on canvas during absorption and drying: a neutron radiography and NMR study” has been submitted for publication and will hopefully become available in 2016.