264 Proceedings - XV. International Numismatic Congress Taormina 2015 Introduction During the late 1980s, X-ray Fluorescence (XRF) spec- troscopy played an important role in developing our understanding of the composition of ancient bronze coins in the Iberian Peninsula 1 , but the studies involved addressed only individual mints and issues and did not aim at developing a wider picture. Between 1995 and 1999, Abascal and Ripollès, building on previous work, undertook a series of metallographic analyses, also us- ing XRF spectroscopy , published in four papers 2 . Their goal was to produce an initial databank of information on the composition of coinages minted in the Iberian Peninsula during the Republican period and the early Empire, and map this (Fig. 1). This remains the only systematic treatment of this material so far completed, though some further analyses were undertaken by other scholars 3 . The aim of this paper is to add yet more data to the existing databank of metallographic analysis of Iberi- an coinages, and present the results of an analysis and comparison of the composition of the bronze coins of the Punic mint of Ebusus and the Iberian mints of NE Spain, during the 3rd-1st c. B.C.E.: 339 coins were ana- lyzed using XRF spectroscopy. In the case of Ebusus, the sample was large enough to make possible a dia- chronic study of the metal composition during the life of this mint. The decision to use XRF spectroscopy makes possible comparison with the earlier analyses, though we recog- nize that the reliability of such comparative studies is lessened by the variability that results from using dif- ferent equipment and operators, but it still is accurate enough for the needs of this project. 1. X-ray luorescence analyses X-ray luorescence (XRF) spectroscopy is useful in per- forming elemental analyses of various materials, with- out destructive sampling. This makesit very suitable for studies in the ield of Cultural Heritage. A portable energy dispersive X-ray luorescence spec- trometer (ED-XRF spectrometer) was used. It comprises a low power X-ray generator, an X-ray detector, a pulse ampliier, a multichannel analyzer, and a computer. The X-ray generator has a palladium anode target and a be- ryllium window 150 µmthick, powered by an accelerat- ing potential diference of 40 kV and an electric current of 0.1 mA. The detector is a Peltier cooled silicon drift (Amptek 123-SDD), which, for every photon detected, 1 Chavez 1897; Olcina-Ripollès 1987-1988. 2 Abascal-Ripollès 1995; 1998; 1999; Gozalbes-Abascal-Ripollès 1996. 3 Chavez-Pliego 1999; Montero et al. 2004. provides a current pulse of amplitude proportional to the photon energy. The detector has a surface of 25 mm 2 ,500 µm thick, and a beryllium window of 12.5 µm. Its en- ergy resolution is 150 eV full width half maximum of the manganese K α line at 5.9 keV. The pulses are ampliied and sent to the multichannel analyzer, which divides and counts the pulses in 1024 diferent channels, on the basis of their amplitude. This results in is a spectrum (number of photons by photon energy) that is sent to the computer for storage and further analysis. The instrument used cannot detect oxygen present in ox- ides, or carbon and nitrogen, all present in organic com- pounds and in corrosion products. K lines of all elements from aluminum to barium can be detected and L lines of all chemical elements from zirconium to uranium. After their acquisition, a best it of the spectra is per- formed to obtain, for each characteristic line, the num- ber of counts. Counts less than three times the value of the statistical error are discarded. The best-it software used was Canberra’s WinAxil 4 . The number of counts of each characteristic line is re- lated to the concentration of the corresponding element but, to ascertain this concentration, it is necessary to perform a complex procedure that takes into account the sensitivity of the experimental apparatus at the ener- gy of the characteristic lines, the geometry of the set-up, and the eiciency of each excited atom for the X-ray emission 5 . Canberra’s WinFundsoftware was used for this purpose. 4 Van Espen 2002. 5 De Vries 2002. Alejandro G. Sinner, Giacomo Pardini, Mario Piacentini, Anna C. Felici, Margherita Vendittelli Analysis of the metal of the coins of Ebusus and Northeastern Spain (3 rd – 1 st C. B.C.E.) Fig. 1 Mints and metals in Spain 2nd-1st c. B.C.E., Ripollès 2005, ig. 6.2. Circled the territories and mints analyzed.