THE NON-INVASIVE ANALYSIS OF ORGANIC MATERIALS IN CULTURAL HERITAGE: DEVELOPMENT OF STRATEGIES BASED ON MOLECULAR SPECTROSCOPIES

The importance of scientific analyses on artworks has become more and more crucial in the articulate processes of conservation and restoration. Moreover, the complexity of artistic materials, composed by organic and inorganic substances mixed together and often aged, requires extremely sensitive but user-friendly analytical tools. For this reason, the request of effective non-invasive diagnostic techniques is steadily increasing. In particular, a challenging task is represented by organic compounds, which can be found in the artistic field with a wide range of applications: they are binders used to spread pigments in paintings, varnishes, coatings and adhesives applied by the artist himself or during the restoration works, and colouring substances. Among the latter two different typologies can be individuated, namely natural dyes, used from the antiquity to dye textiles and in paintings in form of lakes, and a quite recent family of pigments, the synthetic organic ones. The detection of such materials can be hardly achieved by the non-invasive and in-situ appliable analytical techniques commonly employed in the field of conservation science. Briefly, X-ray fluorescence cannot provide useful information as it detects only elements with medium to high atomic number, while infrared and Raman spectroscopies are affected by the problem of fluorescence emission and of the matrix interference respectively. This Ph.D. thesis work was intended to develop innovative strategies based on molecular spectroscopies, pursuing the objective of non-invasively identifying the most common organic materials which can be found in museum objects, mainly focusing the attention on paintings (binders and synthetic organic pigments) and ancient textiles (natural dyes). In particular, spectrofluorimetry was proposed for the identification of synthetic organic pigments, a quite recent category of colourants since they were introduced by Perkin in the second half of the 19th century. The fluorescence emission upon visible excitation exhibited by several organic pigments used in contemporary artworks was successfully demonstrated and an analytical approach combining visible-induced spectrofluorimetry and visible-reflectance spectroscopy with a multivariate processing of data was established. The effectiveness of the method was finally confirmed by its application to three contemporary paintings. In addition, the possibility of exploiting fluorescence emission induced by UV radiation was considered, as well as the prospect of investigating such materials by ultraviolet fluorescence (UVF), visible and infrared luminescence (VIVF and VIL) images. The use of near-infrared (NIR) spectroscopy was proposed for the detection of organic binders and to understand the complex stratigraphy in paintings. This spectral region, in particular in the range 6500-4000 cm-1, is dominated by combination and overtone bands due to functional groups typical of the lipidic and proteinaceous materials used as binding media, therefore it can be exploited for identification purposes. Moreover, it was demonstrated that the greater penetration depth of the NIR radiation in comparison with the mid-infrared (MIR) one allows the investigation of the preparatory layers possibly applied by the artist on the painting support. In this respect, an exhaustive study was carried out at first on a consistent number of reference mock-up samples, prepared on the basis of ancient recipes, and then on eight paintings, dating from the end of the 15th century to the end of the 18th century and exemplifying different artistic techniques on different supports. Finally, the main part of the work was dedicated to surface-enhanced Raman spectroscopy (SERS) for the identification of natural dyes in textiles. This technique, which exploits the strong intensification by several orders of magnitude of the Raman scattering of organic molecules in the proximity of metal nanoparticles and nanostructured surfaces, is a well-established method for the detection of dyes, but generally requires the extraction of the target analyte from a sample. Many experiments were performed therefore performed to develop SERS-sensors suitable for performing measurements directly from dyed fibres in dry-state condition, i.e without requiring extraction protocols or the placement of a wet substrate in contact with the artefact. In this respect, a general, easy-to-apply protocol involving the deposition of colloidal nanoparticles on glass supports was developed. The importance of hot spots, areas of strong SERS intensification of the Raman signal formed between close nanoparticles or on anisotropic structures rich in tips, leads to experiment different aggregation procedures and nanoparticle shapes, ranging from nanospheres to nanostars and nanorods. The final aim was to create reproducible and homogeneous sensors, able to provide a high SERS enhancement when just put in contact in dry-state conditions with the artistic object under investigation. In this respect silver was demonstrated to be the best metal in comparison with gold and the most effective substrates, allowing the identification of four historical dyes from dyed wool yarn, were those obtained from silver nanostars. Finally, also some related issues are faced, such as the release of metal on the sample and the cleaning of the substrate surface exploiting the photochemical properties of titanium dioxide.