This paper presents new developments on the provenance study of lapis lazuli started by our group in 2008: during the years a multi-technique approach has been exploited to obtain minero-petrographic characterization and creation of a database considering only rock samples of known provenance. Since the final aim of the study is to develop a method to analyze archeological findings and artworks made with lapis lazuli in a completely non-invasive way, ion beam analysis techniques were employed to trace the provenance of the raw material used for the production of artifacts. Continuing this goal and focusing the analysis on determination of more significant minero-chemical markers for the provenance study of trace elements in different minerals, the method was extended with the use of micro X-ray fluorescence (µ-XRF), to test the potential of the technique for this application. The analyzes were focused on diopside and pyrite in lapis lazuli samples of known provenance (Afghanistan, Tajikistan, and Siberia). In addition,µ-XRF data were compared with micro proton-induced X-ray emission (µ-PIXE) results to verify the agreement between the two databases and to compare the analytical performance of both techniques for this application.

The serpentine minerals lizardite, polyhedral serpentine, chrysotile, antigorite and 15‐sector and 30‐sector polygonal serpentine have been studied by micro‐Raman spectroscopy, using selected samples, that had been previously characterized. The appropriate crystal orientations were determined by optical microscopy of petrographic sections. Oriented spectra, obtained using Nd‐YAG green laser radiation (532 nm), were deconvolved, extracting wavenumber and intensity values for the peaks, possibly overlapping and forming complex bands. Relevant Raman anisotropy is common and relevant in serpentine. Depending upon the orientation of the impinging laser, significant wavenumber shifts occur (up to 10 cm⁻¹, mostly in polyhedral serpentine and lizardite). Furthermore, also, important intensity variations (up to 1 order of magnitude) occur in polyhedral serpentine, lizardite, chrysotile and antigorite as well. On the one hand, the possibly characteristic peaks have been identified and discussed. On the other hand, caution is suggested as far as the micro‐Raman characterization of polyphasic, variably oriented serpentine minerals is concerned.

This paper presents a μ-XRF analytical approach for a non-destructive study of Cultural Heritage glass finds. This technique can be used for quantitative analysis of small volumes of solid samples, with a sensitivity that is superior to the electron microprobe but inferior to an ICP-MS system. An experimental set-up with natural and synthetic glass standards is proposed here for the quantitative analyses of major and trace elements on glass objects which cannot be sampled such as small archaeological or historical artefacts from Cultural Heritage. The described method, performed by means of the commercial μ-XRF Eagle III-XPL, was applied to Islamic glass specimens of Sasanian production (III–VII century A.D.) previously analyzed by ICP-MS and SEM-EDS techniques (P. Mirti, M. Pace, M. Negro Ponzi and M. Aceto, Archaeometry, 2008, 50(3), 429–450; P. Mirti, M. Pace, M. Malandrino and M. Negro Ponzi, J. Archaeol. Sci., 36, 1061–1069; and M. Gulmini, M. Pace, G. Ivaldi, M. Negro Ponzi and P. Mirti, J. Non-Cryst. Solids, 2009, 355, 1613–1621) and coming from the archaeological site of Veh Ardasir in modern Iraq. Major elements (Na, Mg, Al, Si, K, Ca, Fe) of glass specimens show an accuracy better than 5%. Trace elements (Cr, Mn, Sr and Zr) display an accuracy better than 5% when the checked elements have a concentration >100 ppm by weight, whereas it is around 10% with a concentration <100 ppm by weight. μ-XRF is, therefore, a suitable elemental analysis technique for the non-destructive study of small glass finds due to its relatively good accuracy, reproducibility and low detection limits (∼tens ppm).

Multielemental X-ray maps collected by a remote scanning system of the electron beam are processed by a dedicated software program performing accurate modal determination of geological samples. The classification of different mineral phases is based on elemental concentrations. The software program Petromod loads the maps into a database and computes a matrix consisting of numerical values proportional to the elemental concentrations. After an initial calibration, the program can perform the chemical composition calculated on the basis of a fixed number of oxygens for a selected area. In this way, it is possible to identify all the mineral phases occurring in the sample. Up to three elements can be selected to calculate the modal percentage of the identified mineral. An automated routine scans the whole set of maps and assigns each pixel that satisfies the imposed requirements to the selected phase. Repeating this procedure for every mineral phase occurring in the mapped area, a modal distribution of the rock-forming minerals can be performed. The final output consists of a digitized image, which can be further analyzed by common image analysis software, and a table containing the calculated modal percentages. The method is here applied to a volcanic and a metamorphic rock sample.