**5. Discussion**

The current research, as part of an ongoing study on the early indigenous southern Levantine coinages, employed an archaeometallurgical NDT approach in order to analyze die-linked Late Persian period and Macedonian period Yehud silver coins. In the case of the current coins, based on their shiny metallic appearance, it is evident that the coins have been cleaned. Yet such cleaning procedures were not recorded and remain unknown. Nevertheless, even in such circumstances, surface characterization of silver coins can be performed without any additional polishing of the exterior when the coins' surfaces are shiny and the remaining oxides and corrosion products layers are thin enough [19,30,31].

All the examined Yehud coins were shown to be in a good state of preservation, based on VT, multi-focal LM, and SEM SE and BSE observations. Such preservation can be explained by the excellent corrosion resistance of silver, due to the formation of stable Ag2O on the metal surface [32,33]. Nonetheless, the SEM BSE mode observation combined with EDS analysis revealed the bright areas to constitute exposed silver metal regions and the dark areas as covered with oxides and some corrosion products. The EDS analysis of the surface of the Yehud coins also revealed the presence of the elements O, Si, Cl, Sn, Au, Pb, Al, Ca, Fe, Mg, P, and S (Supplementary Materials, Tables S1–S8). The presence of oxygen and chlorine can arise from both use and storage of the coins and is therefore of little use in shedding light on the metallurgical aspects related to the coins. The absence of lead in most of the examined areas of the coins (lead was only detected by EDS in coin IMJ 34620 (reverse, area 3) of Type 5 O1/R2 and in coin Edom hoard no. 2 (reverse, area 4) of Type 16 O2/R2) indicated that all coins were produced by a successful cupellation refining process. The detection of Cl and S on the surface of the coins was predictable since silver is sensitive to chloride and sulfide ions, leading to the formation of silver chlorides (AgCl) and sulfides (Ag2S) as the main contamination products [19,33]. The presence of the elements O, Si, Al, Ca, Fe, Mg, and P is related to the existence of corrosion products and soil remains [6,19,20,33], yet the presence of Fe could also be related to natural impurities in the silver ores [9]. In aggressive environments, silver alloy objects often also contain other corrosion products related to base metals, such as cuprite and litharge [33]. No presence of the elements Cr, Ni, and Cd was detected. These elements are typical of modern forgeries [24], indicating the authenticity of all examined silver coins.

The presence of gold in silver alloys was apparently unfamiliar to the ancient minters. Yet ancient silver coins often contain less than 1 wt% Au as an impurity associated with the silver ore. Therefore, the gold content in silver coins can assist to distinguish between silver sources used for coinage [34].

Good agreement was observed between the SEM-EDS analysis results of the surface and the bulk after grinding the surface, and between the SEM-EDS analysis results and the XRF analysis results of Type 5 O1/R1, O1/R2, and O1/R4 coins' surfaces, where with both methods, high-purity silver was detected, with an average copper content of less than 5 wt% Cu, as well as other elements, including, O, Si, Sn, Au, Pb, Al, Ca, Fe, Mg, P, and S. The elements Mn, As, Bi, and Ti were also detected by XRF analysis on the coins' surface but were not detected by EDS analysis, whereas the presence of Cl was detected in the EDS analysis but not by the XRF analysis.

Because of the importance of silver as a representative of the material cultural heritage of different populations throughout history, numerous studies of ancient silver objects, including coins, exist in the literature, studying the chemical composition, microstructure, provenance, manufacturing processes, and the corrosion products of silver items. These studies usually combine NDT and destructive testing methods, for example, LM and SEM observation of metallographic samples, particle-induced X-ray emission (PIXE) analysis, inductively coupled plasma (ICP) analysis, X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS) analysis [19,22,23,27,28]. Some possible NDT methods that can be applied for elemental bulk analysis of silver coins are the neutron activation analysis (NAA) and the prompt-gamma activation analysis (PGAA) [34–39]. Yet these methods have some limitations resulting from the high cross section for neutron capture by silver itself, and therefore the silver data would prevent revealing trace elements that could be useful fingerprint variations in the minting process. There are certain rare-earth metallic elements with large neutron cross sections that could provide interesting data concerning differences in the raw silver source. Another NDT technique that could be used to provide additional

information concerning the coins' bulk material composition is the muon induced X-ray emission (MIXE) technique [40].

Ancient silver was commonly produced from silver-rich galena lead sulfide ore using the cupellation process [10]. The presence of Pb and S in some of the measurements is thus probably related to the cupellation process [11,21]. The cupellation technique is very effective for the production of high-purity silver metal, with more than 95 wt% Ag [11], which may explain the high purity of the discussed Yehud silver coins. While the presence of Au and Sn might be related to the ore deposits that were used, the presence of Sn might also be related to the addition of copper to the alloy.

Copper was a main alloying element in ancient silver coins and the addition of copper to silver was often performed to depress the melting point as well as to improve the mechanical properties of the alloy [12,18–20,22]. Since the presence of Sn is quite rare in galena ores, its presence might point to alloying the silver with bronze instead of pure Cu [21]. According to Brocchieri et al. (2020), the detection of high-purity silver coins probably indicates that during the production process of these coins, the mint did not suffer from economic constraints [5]. In addition to the Ag and Cu alloy elements and the corrosion products and soil elements that were detected in all Yehud series coins, other elements were also detected. For example, up to 4.7 wt% Sn was detected in a specific 300 μm × 300 μm scanned area of IAA 153975 (Type 5 O1/R1, Table S1). Up to 1.1 wt% Au was detected in the reverse of coin IMJ 34553 and up to 1.2 wt% Au was detected in the obverse of this coin (Type 5 O1/R2, Table S2). In addition, up to 1.7 wt% Pb was detected in a specific 300 μm × 300 μm scanned area of IMJ 34620 reverse (Type 5 O1/R2, Table S2). This may also support the assumption that each series was manufactured by using a specific composition of silver–copper alloy.

Although SEM-EDS is a valuable NDT tool for surface analysis of ancient silver coins, if the detected objects are covered with a thick oxide layer and contain massive corrosion products, the analysis may not provide a representative characterization of the bulk alloy composition of the object [19,30,41]. Ancient silver objects such as coins sometimes exhibit silver enrichment of the surface and, as a result, a considerable reduction of copper on the surface [33,42]. In silver objects retrieved from a soil environment after a long burial period, a local selective galvanic corrosion attack on the copper-enriched areas may occur, resulting in the diffusion of Cu from the bulk of the Ag object to its surface, causing the formation of a cuprite layer on the external surface of the coin [43,44]. Moreover, under some circumstances, unsuitable conservation methods may cause damage to the ancient coins and affect their surface composition. Nevertheless, when the oxide layer is thin and the shiny metal is exposed beneath, the limitations of SEM-EDS analysis should not prevent the use of this method as a valuable tool for the characterization of ancient silver coins. However, it is essential first to determine whether the composition of the surface layer is representative of the bulk composition of the object [18,19,32,42,45]. In order to do so, seven Yehud silver coins were locally ground in several areas to expose their bulk metal and their composition was examined before and after grinding. Good agreement was achieved between surface and bulk compositions of coins with low copper content (Supplementary Materials, Table S9).

In order to obtain reliable results from the EDS surface analysis, only bright areas with a shiny silver metal appearance were examined here for the calculations of the average value and SD of the alloy composition of each group of coins (Table 1). Our results, after eliminating the peaks of oxides and soil elements, revealed that five of the die combination issues of the Yehud series Type 5 (O1: R1, R2, R3, R4, R5), Type 16 (O2/R2), Type 24 (O1/R2), and Type 31 (O1/R1) are composed of high-purity silver with a small percentage of copper (Table 1).


**Table 1.** The average alloy composition (Cu wt% content) of the different Yehud Types 5, 16, 24, and 31, the late addition Yehud coins, and the Samaria and Nablus Hoards coins, according to SEM-EDS analysis.

The copper content in the Type 5 coin alloy is between 0.3 ± 0.8 wt% Cu (for O1/R5) and 3.6 ± 2.5 (for O1/R2). The copper content in the Type 16 coin alloy is 0.1 ± 0.4 wt% Cu (for O2/R2); the copper content in the Type 24 O1/R2 is 0 ± 0 wt% Cu; and the copper content in the Type 31 O1/R1 is 1.7 ± 3.7 wt% Cu (Table 1). Four coins (IAA 153976 of Type 5 O1/R1, IAA 154383 and IMJ 27387 of Type 5 O1/R2, and IAA 153981 of Type 5 O1/R4) were not included in the average composition values and SD calculations of the main group of each coins. The fact that these four coins contain a high amount of copper (20.7 ± 19.3 wt% Cu, Table 1) may indicate that towards the end of the minting process of each series there was a shortage of raw materials and therefore recycled silver was used.

The manufacturing processes of the coins from all the examined groups were similar (casting and striking a blank flan), with some slight differences (between pure silver alloy for the case of Type 24 O1/R2 coins up to 3.6 ± 2.5 wt% Cu for the case of Type 5 O1/R2 coins). Three additional die-connected Yehud coins of the late addition type (a female {?} head to the right with the Aramaic letter yod in the left field on the obverse and an owl standing right, head facing on the reverse) were studied previously (Table 1) [2] and are used here as a reference group with an average composition of 0.4 ± 0.8 wt% Cu, presenting the similarities and differences in terms of metallurgical composition compared with the above-mentioned coin types.

For comparison, the alloy composition of 80 selected Persian period Samarian silver coins from the Samaria and Nablus Hoards, as well as other coins from the Israel Museum collection, was 95.9 ± 2.5 wt% Ag and 4.1 ± 2.5 wt% Cu (SEM-EDS analysis after omitting the peaks of oxides, corrosion products, and soil elements) [19]. Moreover, the average alloy composition of the Persian period jewelry from the Samaria Hoard was 93.4 ± 1.65 wt% Ag and 6.6 ± 1.6 wt% Cu, while that of the jewelry from the Nablus Hoard was 94.9 ± 1.9 wt% Ag and 5.1 ± 1.9 wt% Cu (excluding the joining areas). In order to determine with sufficient certainty whether the current studied die-linked silver coins were produced from the same metal batch throughout the minting of each group, the copper concentration distribution of each series is shown based on the SEM-EDS analysis results after omitting the peaks of oxides, corrosion products, and soil elements (Supplementary Materials, Figure S8), presenting the Cu wt% concentration range vs. the relative no. of measurements (%). Based on the average values and SD of the alloy composition of each series and the copper concentration distribution of the different groups, each series (including the dielinked issues of Type 5 O1/R1, O1/R2, O1/R3, O1/R4, and O1/R5 subtypes, as well as Type 16 O2/R2, Type 24 O1/R2, and Type 31 O1/R1) was most probably manufactured by using a controlled specific composition of silver–copper alloy.

Based on the current findings, a four-step methodology is suggested for the study of ancient silver coins: (first) a VT should be performed on the objects; (second) the coins should be examined by multi-focal LM in order to establish the areas where shiny silver metal is exposed and whether there are corrosion products on the surface of the coin (the silver alloy composition should be measured only in these shiny areas); (third) initial examination of the coins' surface should be conducted by XRF; and (fourth) the coins should be examined by SEM-EDS analysis in both SE and BSE modes in order to determine the state of preservation of each coin and its alloy composition. Only bright areas observed in the BSE mode should be subjected to EDS analysis in order to determine the alloy composition of each series of coins.

The elaborate iconographic designs on these tiny Persian period and Macedonian period Yehud silver coins demonstrate high artistic and technological abilities (Supplementary Materials, Figures S9–S16). Based on the average alloy composition values and SD of the examined series of the five die combination issues of the Yehud series Type 5 (O1 linked with R1, R2, R3, R4, and R5), each series was struck from a different and controlled specific composition of silver–copper alloy. Approximately 10% of the coins revealed a different alloy composition, however, with a much higher amount of copper and a heterogeneous composition. This implies that at a certain stage of the minting process, a different batch of, possibly recycled, alloy was used rather than the standardized alloy that was recorded for all the other coins of the same die connection.
