Organic colorants have been used for millennia as textile dyes and as pigments for paintings and other kinds of polychrome works of art. Even today, natural and synthetic organic dyes are of great interest in the forensic field as food colorants, ink components and textile dyes. As they are often present in art objects in minute amounts due to their high coloring power, their identification is really challenging. High-performance liquid chromatography (HPLC) and UV-visible spectroscopy have been routinely employed to detect dyes, but they require sizable samples. During the last few years, a lot of research in this field has been carried out by surface-enhanced Raman scattering (SERS), which has proven to be an extremely effective technique because of its great sensitivity. Although the development of several micro-invasive analytical procedures has recently expanded the applicability of SERS spectroscopy in the art conservation and forensic fields, the necessity of obtaining a reliable match between spectra from unknowns and reference materials in a fast and convenient way is still to be fulfilled. In this context, the availability of extensive spectral databases as well as the accessibility of library search methods designed to match a query spectrum to the closest library references would significantly increase the spectral recognition rate, making the entire identification process considerably easier. In the present work, a number of pure colorants, extracts from plants and insects, fragments from dyed paper and textiles, as well as lakes and glazes from several works of art were characterized by FT-Raman, ordinary Raman and SERS spectroscopies in different experimental conditions, and the spectra thus obtained were then used to build a searchable library. Preliminary attempts to classify query spectra against the library clearly showed that the search process is strongly affected by spectral variations which are completely unrelated to the identification goal, such as the amount of fluorescence, the background shape and the signal resolution. Improved results were achieved by searching against a second library, in which all the spectra were subjected to baseline correction and selective exclusion of certain spectral regions. Multivariate analysis methods, such as principal component analysis, will be applied to the spectra acquired thus far and evaluated to further improve the classification process and provide an easy, fast and reliable way to identify unknown dyes.
Searchable databases for the identification of organic colorants using surface-enhanced Raman scattering (SERS) : critical evaluation and possible approaches / F. Pozzi, S. Porcinai, M. Leona. ((Intervento presentato al 6. convegno International Conference on Advanced Vibrational Spectroscopy tenutosi a Sonoma County nel 2011.
Searchable databases for the identification of organic colorants using surface-enhanced Raman scattering (SERS) : critical evaluation and possible approaches
F. PozziPrimo
;
2011
Abstract
Organic colorants have been used for millennia as textile dyes and as pigments for paintings and other kinds of polychrome works of art. Even today, natural and synthetic organic dyes are of great interest in the forensic field as food colorants, ink components and textile dyes. As they are often present in art objects in minute amounts due to their high coloring power, their identification is really challenging. High-performance liquid chromatography (HPLC) and UV-visible spectroscopy have been routinely employed to detect dyes, but they require sizable samples. During the last few years, a lot of research in this field has been carried out by surface-enhanced Raman scattering (SERS), which has proven to be an extremely effective technique because of its great sensitivity. Although the development of several micro-invasive analytical procedures has recently expanded the applicability of SERS spectroscopy in the art conservation and forensic fields, the necessity of obtaining a reliable match between spectra from unknowns and reference materials in a fast and convenient way is still to be fulfilled. In this context, the availability of extensive spectral databases as well as the accessibility of library search methods designed to match a query spectrum to the closest library references would significantly increase the spectral recognition rate, making the entire identification process considerably easier. In the present work, a number of pure colorants, extracts from plants and insects, fragments from dyed paper and textiles, as well as lakes and glazes from several works of art were characterized by FT-Raman, ordinary Raman and SERS spectroscopies in different experimental conditions, and the spectra thus obtained were then used to build a searchable library. Preliminary attempts to classify query spectra against the library clearly showed that the search process is strongly affected by spectral variations which are completely unrelated to the identification goal, such as the amount of fluorescence, the background shape and the signal resolution. Improved results were achieved by searching against a second library, in which all the spectra were subjected to baseline correction and selective exclusion of certain spectral regions. Multivariate analysis methods, such as principal component analysis, will be applied to the spectra acquired thus far and evaluated to further improve the classification process and provide an easy, fast and reliable way to identify unknown dyes.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.