Evaluation of a Principal Components-Based Searching Algorithm for Raman Spectroscopic Identification of Organic Pigments in 20th Century Artwork

Abstract

The characterization of artists' materials is of great importance for the conservation and restoration of 20th century paintings, and it provides information which is of art-historical interest. The identification of modern organic pigments by Raman spectroscopy is hampered by the large amount of different synthetic materials that exist. Therefore, an extended database of reference spectra is needed. Besides this spectral library, there is need for an accurate and fast-searching algorithm for selecting the reference spectrum that best corresponds with the unknown spectrum. A principal components-based spectral searching algorithm is proposed that largely reduces the number of data points for each spectrum. By subsequent measurement of the Euclidian distance between the unknown and the reference spectra in the principal components space, the corresponding reference can be selected. The algorithm is tested by using the results of the Raman spectroscopic analysis of different colored pencils. Different qualities of matches for the identification protocol could be distinguished. The synthetic organic pigments in micro-samples of the painting “Faubourg” by Paul Delvaux could be identified by the proposed algorithm.