Toward the comprehensive spectrochemical imaging of painted works of art: a new instrumental approach

This paper introduces novel technologies that will be instrumental in the development of new methods for the Raman spectrochemical imaging of painted works of art. We show that second‐derivative preprocessing diminishes interference from fluorescence. Normalization and mean‐centering of such second‐derivative spectra cast them in the form of unit vectors, the displacement of which from library standards serves as a basis for the instantaneous on‐line identification of artistic pigments and binders. Multivariate measures of displacement are shown to function well in difficult cases such as spectrally similar fatty acid media and pigments diluted to varying degrees in uncharacterized binders. A self‐calibrating technique for background subtraction extracts pure component spectra from pigment–acrylic mixtures.