gems, ancient gemology, ancient jewelry, Egypt, Egyptian jewelry, turquoise, XRF

One puzzle from the ancient Mediterranean world is why there are not more examples of jewellery set with turquoise. The mines of the Sinai Peninsula and Iran are well known today, and we are aware that the Sinai mines were worked in antiquity by the Egyptians because there are texts telling us so (Aston et al. 2000, pp. 62–63; Mansour 2014). There is some surviving ancient Egyptian jewellery set with turquoise, but perhaps not as much as we might expect. When it comes to Greek and Roman jewellery, however, turquoise is noticeably absent. The present author knows of very few examples of turquoise set in ancient jewellery (for a discussion of this, see Ogden 1982, pp. 112–113). This justifies documenting the ring shown in Figure 20. It has been in a private collection for some 40 years since the author first saw and illustrated it (in black and white; Ogden 1982, figure 5.13). Recently, the author had an opportunity to re-examine this ring and take better photos. It dates to around the first century AD, and is probably from the Eastern Roman Empire, quite possibly Asia Minor, in what is now Turkey. It is small—20.5 mm across—and quite delicate. The low-cabochon turquoise is oval shaped and measures 7.9 × 4.9 mm. The ring itself is hollow, the gold being a thin sheet. This elegant, hollow form was popular in the early Roman world, but it is still unclear how such metalwork was formed. It is almost certainly filled, as was usual for Roman jewellery, with sulphur, which acted as both a support for the thin gold and a bedding for the gem (Ogden 1982, p. 40). The gem was previously identified as turquoise based only on its visual appearance. This has now been confirmed by portable EDXRF analysis using an Oxford Instruments X-MET8000 unit, which revealed that the primary elements present were Cu, Al and P. Turquoise is a hydrated phosphate of copper and aluminium, with the theoretical chemical formula CuAl₆(PO₄)₄(OH)₈ • 4H₂O. In practice, Zn often replaces some of the Cu, and in this case traces of Zn were detected, as well as significant Fe and a small amount of As. Recent research has shown the potential for determining turquoise sources on the basis of relative Fe, As and Zn contents (Carò et al. 2017; Mousavipak 2020). In this case, measuring the chemical composition of the turquoise with EDXRF was complicated by the relatively large beam diameter (about 3 mm), and precise analysis of small cabochons is difficult because of their surface topology and the inevitable analytical overlap with the metal setting. Nevertheless, even allowing for a reasonable margin of error, the proportions of Fe, Zn and As suggested that the gem’s composition is consistent with an Iranian origin. This certainly shows the potential for portable EDXRF technology for gem research, as long as its limitations are understood. It is hardly surprising that such instrumentation is increasingly being used in archaeogemmology, alongside the typically slightly less-portable Raman spectroscopy. Portable EDXRF analysis is non-invasive and can quickly identify the elements present (i.e. Na and heavier), which is usually sufficient to confirm the identity of a gem and in some cases make further deductions.