The Use of Varnishes As a Final Transparent Layer to Both Protect the Surface of Paintings

Abstract

The use of varnishes as a final transparent layer to both protect the surface of paintings from dirt, light, and humidity in the air and improve their appearance, has been common practice for centuries. Undesirable yellowing and crazing of varnishes occurs in time due to oxidation and polymerisation of compounds in the natural resins they are mostly made of. They are therefore removed and replaced regularly by restorers. As varnishes react differently to the cleaning materials, classification of the resins they are made of is important.

The common analysis methods used in the field of restoration and conservation are FTIR, DTMS and GC-MS. The main advantage ESI-MS holds over these methods is that thermally labile, large and polar compounds, which especially oxidised resin components are, can be detected. The research question was: is ESI-MS a suitable analytical method for the characterisation of varnishes made from natural resins? Eleven fresh resins, being colophony, some copals (Congo, kauri and manila), sandarac, Venice turpentine, dammar, elemi, mastic, amber and shellac, were extracted with ethanol and analysed with ESI-MS in positive and negative ionisation mode. Nine partly artificially aged resins and 36 painting and furniture samples were analysed as well. Ammonium acetate was used as the ionising agent.

Many peaks in the fresh resin spectra could be assigned based on literature. The varnish samples were compared with the fresh and aged resin spectra, and the spectra showed little overlay, despite the presence of different resins and additional components as beeswax and oil. The aged spectra showed a good resemblance to their respective ‘fresh’, unoxidised counterparts. The exception was aged amber that was very different from its ‘fresh’ counterpart, making it, at least with the settings and sample preparation employed here, unsuitable for ESI-MS analysis. The presence of multiple components of one class of resins could somewhat complicate interpretation. A unique combination of two peaks can point towards one resin, whereas they might actually be originating from different resins. Though theoretically the use of ammonium acetate results in deprotonated molecules in negative mode, and ammonium adducts in positive mode, acetate adducts were seen in negative mode as well. No conclusive answer could be given as to whether these peaks were clusters or severely oxidised resin compounds. To test if these oxidised compounds are present, resin or varnish samples should be analysed using a different ionising agent or sample preparation. Performing MS/MS on these compounds can provide a definite answer as the loss of an acetate group can be seen if they are indeed clusters.

Concluding from this first exploration ESI-MS seems a very promising addition, and possibly even alternative, to the existing analysis methods that are being used in the field of restoration and conservation. The creation of a reference database through the analysis of varnish samples with known composition and age can be a logical next step in the establishment of ESI-MS as one of the standard methods in this field.