A whole spectroscopic mapping approach for studying the spatial distribution of pigments in paintings

We propose a non-invasive approach for the identification and mapping of pigments in paintings. The method is based on three highly complementary imaging spectroscopy techniques, visible multispectral imaging, X-Ray fluorescence mapping and Raman mapping, combined with multivariate data analysis of multidimensional spectroscopic datasets for the extraction of key distribution information in a semi-automatic way. The proposed approach exploits a macro-Raman mapping device, capable of detecting Raman signals from non-perfectly planar surfaces without the need of refocusing. Here, we show that the presence of spatially correlated Raman signals, detected in adjacent points of a painted surface, reinforces the level of confidence for material identification with respect to single-point analysis, even in the presence of very weak and complex Raman signals. The new whole-mapping approach not only provides the identification of inorganic and organic pigments but also gives striking information on the spatial distribution of pigments employed in complex mixtures for achieving different hues. Moreover, we demonstrate how the synergic combination on three spectroscopic methods, characterized by highly different time consumption, yields maximum information.