Non-invasive fibre optic Fourier transform-infrared reflectance spectroscopy on painted layers: Identification of materials by means of principal component analysis and Mahalanobis distance

Abstract

Fibre optic Fourier transform-infrared (FT-IR) reflectance spectroscopy has recently made it possible to perform completely non-invasive investigations on works of art and, in particular, on painted layers. The use of chalcogenide fibre optics can overcome most of the limitations due to the size of the objects under investigation, and permits the acquisition of spectra in a wide mid-IR range that includes the so-called fingerprint region (2000–900 cm −1). The non-invasiveness of the technique means that it is possible to record a large amount of spectral data for each sample. In view of the considerable dimensions of the data set, it is helpful to use a statistical treatment for the data, such as principal component analysis (PCA), in order to obtain the most significant information. As a first step in investigating actual paintings, laboratory painted layers were prepared using different pigments and binding media. These were then examined. PCA was applied to the spectral data obtained in order to identify clusters related to the different materials that made up the samples. Test samples were classified by using a Mahalanobis distance classification method in the principal component space.