During the past decades conservators in museums have been increasingly challenged by objects made of modern synthetic resin materials. Such objects are vulnerable towards deterioration and degradation, which in part may be ascribed to their construction from commercial products not designed to last forever. The lifetime of modern synthetic resin materials varies and it is generally considered to be no longer than about 30 years, after which defects such as discoloration, stickiness, and cracking can be observed. A natural first step in the conservation of these objects is the identification of their composition and condition. The most widely used technique for these purposes is Fourier transform infrared spectroscopy (FTIR). IR methods generally require the removal of a sample or in the case of attenuated total reflectance (ATR), it only provides information about the composition on surface of the object, which may not be adequate when understanding the degradation underneath the surface is equally important. Unilateral nuclear magnetic resonance (NMR) offers the possibility of non-invasive in situ analysis of a wide range of materials including modern synthetic resins. With the Profile NMR MOUSE®, it is possible to obtain depth profiles of materials and thereby obtain information on the composition and molecular mobility at different depths from the surface of the object. Transverse relaxation decay of protons in organic materials can be measured using the so-called Carr-Purcell-Meiboom-Gill (CPMG) NMR experiments where the time constant of the signal decay reflects molecular size and mobility. Such information is typically extracted by exponential fitting of the decay. The reliability of such procedures highly depends on the achievable signal-to-noise ratio and may be impractical when handling large datasets. Moreover, the decays are intrinsically multiexponential, and they must be approximated by a bi- or triexponential model function. Consequently, the results depend on the selected model function and are not always comparable. Here we demonstrate how multivariate data analysis of the relaxation decays can be used to provide a fast overview of large, potentially noisy, datasets with an unambiguous, model-free approach. One aim of the study is to explore the capability of principal component analysis (PCA) to discriminate data in terms of types of materials and molecular mobility. The method allows the analysis of large datasets obtained through mapping of large artifacts in all three dimensions to monitor structural changes and deterioration or obtained through the establishment of a reference database for conservation studies and analysis.
Cindie Kehlet1, Eleonora Del Federico1, Niels Chr. Nielsen2, Jens Dittmer3, Hiba Schahbaz1, Amelia Catalano1
1 Department of Mathematics and Science, Pratt Institute, Brooklyn, New York, US. 2 Interdisciplinary Nanoscience Center (iNANO), University of Aarhus, Denmark. 3 Institut des Molecules et des Matériaux du Mans (IMMM), Université du Maine, Le Mans, France.
