Analisi archeometriche su manufatti vitrei.

Object of the present work is the archaeometric study of glass fragments found on the archaeological site of S. Andrea. The study aims at investigating possible relationships between chemical composition and type and/or production techniques; at identifying analogies with coeval glass from the Mediterranean area; at better contextualising the various phases of the site, and at extending the database on Italian Medieval glass. Archaeometric study was carried out on a selection of glass fragments from Sectors A and B for a total of 29 samples, composed of objects, glass-working residues and two glass 'cakes'. Chemical composition of glass fragments was obtained with CAMECA SX50 electron microprobe (EMPA) equipped with four wavelength-dispersive spectrometers (WDS). Textural analyses were also performed by means of scanning electron microscopy (SEM), using CamScan MX3000 instrument with LaB6 source. Results of chemical analysis show that all objects and glass-working residues are silicasoda-lime in composition, with natron as flux in accordance with production technologies used in Early Medieval times. Various compositional groups, named LP1a, LP1b and LP1c, were identified, consistent with some of the major ones identified in the western Mediterranean during the first millennium AD. Group LP1a is chemically similar to unintentionally coloured Roman glass, spanning the period between the 1st and the 4th centuries AD. The presence within LP1a group of both glass-working residues and objects points to the existence of a Late Antique/Early Medieval secondary glass-workshop in the Loppio site. Traces of at least one of the elements Cu, Sn, Sb, and Pb in quantities above 0.05 wt% in most of the LP1a objects, suggest possible recycling of older glass. Further support to this hypothesis comes from the finding in the same site of one blue glass tessera, having high contents of copper, tin, antimony and lead, and characterised by calcium antimonate as opacifier, mainly used in Roman period. Group LP1b can be compared with 'Group 2' of Foy et al. 2003, a 'weaker' HIMT (High Iron, Manganese and Titanium) group, which is datable to the 6th-8th centuries AD. Group LP1c shows chemical similarity with 'Series 3.2', a subgroup of 'Group 3' of Foy et al. (2003). In particular, 'Series 3.2', characterised by lower Al2O3 and higher Na2O than other series of the same group, with manganese as decolourant, comprises glass samples from the Western Mediterranean area dated from the late 5th to the early 6th century AD. Both groups LP1b and LP1c are entirely composed of objects. Whether groups LP1b and LP1c were obtained by recycling earlier glass or whether they are the result of a new glass batch intentionally prepared in Late Antique/Early Medieval period cannot be established with certainty. However, the bad match with previous Roman glass and the absence of glass-working residues with comparable compositions seem to invalidate the hypothesis of recycling, as also confirmed by the generally low levels of trace elements such as Cu, Co, Sb and Pb. Three outliers were also identified in the Loppio assemblage; they show similarities with Roman glass decolourised with antimony or manganese, and with that intentionally coloured in black. The two rough ellipsoidal glass 'cakes', with a concave-convex profile, show a chemical composition very different from other samples from Loppio. The main mass of the Loppio 'cakes' is composed of a green, transparent glass containing frequent, submillimetric, randomly distributed whitish inclusions. The main mass is completely surrounded by a 3-5 millimeters thick cortex of a brownish colour characterised by frequent radial, and sometime open, fractures disposed orthogonal to the contact. The contact between the main mass an