A Multidisciplinary Approach to the Study of Polish Silver Denars Minted in Medieval Poland Concerning Their Provenance and Archaeological Aspects

Near-surface analytical techniques, such as X-ray fluorescence (XRF), high energy particle induced X-ray emission (PIXE) as well as particle induced gamma-ray emission (HE-PIGE) and neutron activation analysis (NAA), which provide an analytical data from greater depths, have been used to study more than two hundred medieval denars of the Piast dynasty. Two collections housed in the National Museum in Krakow (71 denars), Numismatic Cabinet of the National Museums Berlin (39 denars) as well as denars from several smaller collections housed in various museums and by private collectors were evaluated in the present study. All of them were minted between 992 and 1020 by two of the first Polish kings from the early Piast dynasty, namely Boleslaus the Brave and his son, Mieszko II Lambert.

An archaeometric approach for the study of these denars is fundamental to establish any possible correlation between the chemical composition of these objects and their historical context, since it concerns the beginning of a Polish state. The analysis had the objective of gaining a better understanding of the chemical composition of the alloys utilized in the early Polish coinage system.

major elements present were Ag and Cu, while minor elements such as Pb, Au, Bi, Hg and Zn were also detected. An evaluation of the results obtained with the different techniques showed that the content of Ag and Cu near the surface and in the bulk varies in some of the coins. Quantitative analysis revealed a distinction between Bolesław the Brave and Mieszko II Lambert mintage procedures. In addition, a correlation between Ag, Au and Bi was observed after evaluating the analytical results providing a first step in the determination of silver provenance.

Keywords: Coins; Silver-copper alloys; Mintage; Medieval Poland; Surface and bulk analysis; Elemental composition

The late 10th and 11th centuries are particularly important in the context of Polish state development. In the period 992-1020, during the reign of King Boleslaus the Brave and his son Mieszko II Lambert who held an important state function alongside his father, the first Polish monetary system was created and developed. For around thirty years several types of coins with various iconographies were produced and provide evidence about the importance of these rulers as well as their trading activity. The collapse of coin production came as a result of Polish state crisis during the reign of Mieszko II Lambert. The revival of coin production in Poland came 50 years later [1]. The beginning of Polish coinage has been for a long time an area of interest to researchers. Numismatic studies have allowed to enhance the knowledge of coinage origin in Poland through the investigation of new coin finds (mainly in Poland, Scandinavia but also in the former Rus region) and discoveries of new types of the oldest Polish coins. During the medieval minting process, inscriptions and patterns on coin obverse and reverse were obtained with two dies - upper and bottom. The upper die was struck with a hammer and a pattern was impressed on a metal disc placed between the dies [2]. Detailed analyses of known iconographic types and pairs of dies have allowed us to investigate the history of dies connections and create characteristic die-chains which connect particular coins types with each other. The study of this theme is important within the context of medieval development of long-distance trade and the role of Polish lands in these trades, between the Islamic world, the Northern Lands as well as with the West Europe as evidenced by numerous finds of foreign coins in Poland. The preliminary stage of Polish coin production is associated with the degeneration of the eastern silver trade and development of silver production in western Europe. Around the period 985-990 Polish markets were flooded with Saxon coins while the influx of dirhams came to an end. Moreover, the context of foreign coins proved to be extremely important for verifying the chronology of Polish coin types [3-5]. An archaeometric study of the first Polish denars can be fundamental to establish any possible correlation between the chemical composition of these artifacts and their historical context. Most probably the first Polish coins were struck in mints located in several main state centres like Gniezno, Poznań, and Mogilno as well as Giecz, from silver that had to be collected from unknown silver sources. Materials could have been imported from the foreign sources (either from Central Asia or Western Europe) and recycled. However, current archeological research has indicated that local silver sources could have been available and silver might be extracted in the areas of Upper Silesia and Little Poland. These territories should also be considered as a potential silver source when discussing archaeological objects from early medieval periods [6]. Therefore, the questions arise: what is the location of the silver source? Was one source or various sources employed? Was the first Polish coinage homogeneous or diversified in terms of its chemical composition? Is it any correlation between the chemical composition of Polish and foreign coins? The aim of this study was to correlate the numismatic knowledge with elemental composition data of Polish denars. A group of 110 early Polish coins from two collections housed in the National Museum in Krakow (71 denars) and Numismatic Cabinet of the National Museums Berlin (39 denars) were featured in previous papers [7-11]. In this work, the set of analysed coins was expanded by 104 specimens coming from different museum collections as well as from private collectors (a description of the coins is provided in Table 1). Since the number of early Piast coins is limited (only about two hundred and sixty items currently above the ground) an attempt was made to perform complementary analyses on the majority of these objects.

Table 1: Overview of studied Mieszko II Lambert and Boleslaus the Brave coins from museum collections and private collectors.

Qualitative and quantitative analyses were carried out using only nondestructive techniques. All coins were investigated with a handheld XRF spectrometer. Handheld XRF was an adequate technique for conducting in situ analysis of the coins due to multiple location of coins in various Polish museum and private collections. A summary of the results obtained using handheld XRF spectrometry is also presented in this work. In the case of coins for which it was possible to perform additional surface and core research, the results were correlated to evaluate coins surface and depth het erogeneity and to assess the suitability of non-invasive and surface techniques for studying archaeological silver objects.

The instrument (S1 Titan, Bruker, Germany) consists of Rh target X ray tube and X-Flash® silicon drift (SDD) detector. The voltage and current of the X-ray tube were set to 50 kV and 15 μA, respectively with a beam spot size of 5 mm. The acquisition time was 30 s. The quantification was based on the Fundamental Parameter (FP) method developed by Bruker with instrument average relative uncertainty for the concentration of detected elements: 0.3% for Ag, 1.0% for Cu and 4% for Pb. Moreover, for the largest denars collections housed in the National Museum in Krakow (71 coins) and the Münzkabinett Staatliche Museen zu Berlin (Numismatic Cabinet of the National Museums Berlin-39 coins) μ-XRF spectroscopy (ARTAX, Bruker, Germany) was employed. The results were evaluated and compared with those obtained using handheld XRF spectrometry [11].

Additional methods were applied due to the different location and access to individual collections. The following techniques were also used:

a) 68 denars from the National Museum in Krakow were analysed with micro-proton-induced-x-ray emission (μ-PIXE) using the proton microprobe facility at the HVEC K-3000 Van Graaff accelerator located in Institute of Nuclear Physics, Polish Academy of Science in Krakow. The average relative errors for μ-PIXE technique are 1% for Ag, Cu and Pb.

b) Neutron activation analysis (NAA) was also carried out for this group of 68 coins in the Faculty of Physics and Applied Computer Science of the AGH University of Science and Technology in Krakow. The equipment used consists of a 239 Pu-Be isotopic neutron source, a standard Canberra GC 2018 HPGe detector and electronics. Estimated relative measurements uncertainties for Ag and Cu are 3.0 % and 10 %.

c) 39 coins from the Münzkabinett Staatliche Museen zu Berlin were analysed using micro-proton-induced-x-ray emission (μ-PIXE) where the average relative errors are 0.2 % for Ag, 2.7 % for Cu and 14.0 % for Pb. Moreover, high-energy particle-induced γ-ray emission PIGE (HE-PIGE) at the Helmholtz-Zentrum Berlin für Materialien und Energie (HZB) cyclotron were also conducted. The average relative error is 30 % for Cu.

Hierarchical cluster analysis (HCA) with selected data was performed using Ward’s method and the Euclidean metric. The analyses were performed to evaluate the composition of Boleslaus the Brave coins with known estimated types, dates, and mints. Before conducting the analysis, the data was subjected to appropriate processing. For instance, records with missing values were removed, values below the detection threshold were assigned 0, while values below the quantification threshold were assigned a minimum value but different from 0. The resulting values were then standardized against the mean and standard deviation. The results are presented in the form of a two-dimensional heatmap, as it is a convenient method of presenting extensive multidimensional data. The colors of the heat map reflect the relative abundance of the element in the coin with respect to the average-a higher or lower color value means higher/lower concentration of the element than the average determined for this element for all analyzed coins.

Comparison of surface and bulk composition results obtained for coins from the National Museum in Krakow and the Münzkabinett Staatliche Museen zu Berlin using HH-XRF, μ-PIXE, PIGE and NAA techniques. The evaluation of coins indicated that the major elements present in the coins are Ag and Cu. In order to obtain reliable quantification results there are some important issues that need to be taken into account when silver-copper alloys are evaluated using non-destructive techniques. Historic silvercopper alloys are examples of eutectic type [12] with non-uniform composition resulting from element segregation during casting. Silver surface enrichment can occur with simultaneously depletion of a less noble metal like copper. This behavior has been reported and examined by several authors [13-19] and causes a certain degree of heterogeneity at the surface and core of denars affecting the depth of the analysis. This surface enrichment could be associated with coins manufacturing like the casting or chemical post-treatment and also long-term processes like wearing or corrosion. Moreover the tarnished, corroded and non-flated surface has an impact on every single measurement. If cleaning procedures are not allowed, the analysis becomes a challenge. For better estimation of elemental concentration, several measurements were conducted for each coin using (μ-PIXE) and HH-XRF techniques. They revealed that Ag and Cu concentrations are largely affected by the inhomogeneity of the coins and the average precision for Ag and Cu concentrations was estimated at 4,2% and 20.5%. Moreover, bulk analyses (both NAA and induced γ-ray emission PIGE) were carried out to identify the possible degree of heterogeneity in coin depth.

Due to the above-mentioned issues, the obtained Ag and Cu surface results show some deviations for certain coins between (μ-PIXE) and HH-XRF methods (Figure 1). The majority of Ag concentrations remain within a ± 4% range with a few specimens for which the differences are significantly higher. For Cu, the uncertainty increases up to ± 30% and sometimes even larger. Heterogeneous Cu distribution and its detection is not surprising in the case of archeological silver-copper alloys with simultaneous presence of silver surface enrichment versus copper deposition on the coin’s surfaces. All described phenomena clearly limit the usefulness of surface techniques.

Further comparison of surface techniques (HH-XRF and μ-PIXE) with bulk methods (NAA and induced γ-ray emission PIGE) reveals increasing variation between results in two different ways (Figure 2). First, higher Ag content for some coins at the surface compared with their bulk indicating silver surface enrichment. According to Beck [13] there is a possibility of direct production of silver surface enrichment depending on the initial composition of the alloy. Our results are in good agreement with the model proposed by Beck where an enriched layer (Ag > 92%) is observed at the surface of the coin with Ag >72% in the bulk. For coins with lower Ag content in the bulk, silver surface enrichment is close to the eutectic phase (72% Ag). Coins with higher fineness (Ag > 92%) show the opposite behavior consisting of similar or higher Ag amount in the bulk relative to the surface. This could be explained by the influence of the corrosion layer on surface measurements (Figure 3).

Minor elements detected in the Mieszko II Lambert’s and the Boleslaus the Brave’s denars and classification of the coins in terms of their numismatic context. Silver above 90% alloyed with copper is typical for early medieval artefacts with addition of minor and trace elements. The elemental analyses also revealed presence of Pb, Zn, Hg, Au, Bi, Br and Fe identified as minor or trace elements. These coins have a Pb content in the range of 0.05 -2.2 %, which is typical for silver that has been canceled. The presence of Zn can be explained through the addition of brass after refining. PIXE and XRF results indicate that there are trace amounts of Hg in part of the coins which could be related to surface contamination or surface treatments. Also, Br and Fe were confirmed in single cases most probable as surface contamination from the burial environment.

While none of above-described elements serves as a marker of the silver source, the significance of Au and Bi to distinguish Ag ores types has been widely recognized [20]. Au as a nobel element remains unchanged in the cupellation process. Thus, the Au content is directly related to the ore. Bi is also considered a marker, but its presence must be evaluated slightly different than Ag. According to L’Heritier, Bi does not remain in the silver and mainly oxidizes during cupellation, mostly at the end of the process [21]. The amount of Bi detected in silver coins could be used for evaluating the final cupellation process. The presence of Bi in the alloy also has an impact on the cupellation process and the silver yield, where an increase of the initial Bi content results in a decrease of the silver reaction yield. The Bi/Pb ratio in silver coins could be used to estimate the initial Bi content. However, this ratio is not directly useful for coins where additional processes like mixing different metals or remelting could have taken place. However, the Bi/Pb ratio could be used to determine ore selection or efficiency of metallurgical processes.

The identified minor and trace elements served to classify the coins into three groups. In the paper by Bolewski et al., [10], the authors focused on 69 denarii housed in the National Museum in Krakow and divided them into three groups. After increasing the number of analysed specimens to 214 and considering HH-XRF results and additional surface and bulk analyses carried out for a subgroup of the coins, the division into three groups was still adequate (Tables 2-4).

Table 2: List of Mieszko II Lambert coins analysed during the study. Concentrations obtained by HH-XRF, PIXE and NAA/PIGE techniques.

Table 3: List of Boleslaus the Brave coins analysed during the study. Concentrations obtained by HH-XRF, PIXE and NAA/PIGE techniques.

Table 4: List of Boleslaus the Brave coins analysed during the study. Concentrations obtained by HH-XRF, PIXE and NAA/PIGE techniques.

The first group contains 56 Mieszko II Lambert coins, which are presented in Table 2. The mean value of weight percent and standard deviation obtained with HH-XRF are 94.6 ± 1.8% for Ag and 4.1 ± 1.6% for Cu. For two denars, the Ag content obtained was significantly lower with an increasing amount of Cu. The quantity of Pb is < 1.2% and in 32 of the denars is < 0.5%, mostly without Au and Bi. There are two exceptions (coin VII-P-758 from the National Museum in Krakow and coin 18225385 from the Numismatic Cabinet of the National Museums Berlin) where Au was detected as a minor element. There is a high similarity between quantification results obtained with surface and bulk techniques within Mieszko II Lambert coins. However, for three coins, bulk analyses showed lower Ag content relative to surface measurements. A particular deviation was observed for coin VII-P-758 for which NAA measurements showed significantly lower Ag content than XRF or PIXE. This deficiency is not compensated for the higher amount of Cu. For Mieszko II Lambert coins, the amount of Bi remains below 500 ppm with one exception – 18225385 where the Bi content is 1900 ppm. Zn is also present in a few coins, sometimes in a significant amount (3.3%) as for the case of coin 18246642 from the Numismatic Cabinet of the National Museums Berlin. He is present as a trace element in some of the coins.

Based on numismatic studies there are 3 types to Mieszko II Lambert. All of them were minted from circa 1010 to 1020 probably in Giecz (Table 5). These coins, contain Mieszko II Lambert’s names and served as status manifestation. Mieszko II Lambert coins (type I and type II) are characterised by a uniform composition. For two denars classified as type III.4, the composition varies with significantly lower Ag content and increasing amount of Cu.

Numismatic research also revealed above twenty types of coins assigned to Boleslaus the Brave. Most of these (Suchodolski classification, 1967) are present in the analysed collections. Within these types, two essentially different groups of Polish coins minted from circa 992 to 1020 can be distinguished. The first group, served as status manifestation, contains Boleslaus the Brave name and title, and also the name of mints in particular issues. The second group of Boleslaus the Brave coins imitated foreign coins dies such as Saxon, Bavarian, Bohemian or Anglo-Saxon and was produced only for economic reasons. The dies of these two groups have been found to link and create characteristic three separated die chains (Figure 4), which presumably represent activity of various mints or one mint during different periods. It was possible to evaluate each chain grouping of coins and correlate die-links with the identified elemental composition of every specimen.

Preliminary research has revealed that at least four mints could have been used to produce Boleslaus the Brave coins. Among them, two mints for which the highest number of coins were assigned are Poznań and Gniezno. Hierarchical cluster analysis (HCA) carried out on Boleslaus the Brave coins revealed two main classes. These results coincide with a numismatic preliminary study (Table 5).

The first group (91 items) is characterized by a low amount (below 500 ppm) or absence of Au and Bi (Table 3 & Figure 5). In this set, the mean values of weight percent and standard deviation are 94.7 ± 3.3% and 3.6 ± 3.2% for Ag and Cu, respectively. As can be observed in the lower part of the heatmap, this group is characterized by a more homogeneous elemental composition. However, a higher standard deviation was observed for Ag and Cu within this group relative to those from the Mieszko II Lambert group.

It should be pointed out that there are several coins for which NAA indicated higher amount of Ag in the bulk relative to the surface. On the contrary, 4 coins from the National Museum in Krakow (VII-P-784, VII-P-786, VII-P-792, VII-P-816) measured with NAA showed lower Ag percentages compared to surface results. Interestingly, there is no increase of Cu content in bulk of the coin. Three coins from the Numismatic Cabinet of the National Museums Berlin (18246619, 18246641, 18246643) also showed a lower amount of Ag in the bulk: 86.1, 85.9, and 65.0% than at the surface and considerable increase of Cu to 12.8%, 12.3%, 33% in bulk of the coins, respectively. The quantity of Pb within this group is the largest and for 37 of the denars is > 1.0%.

Table 5: Summary of coin types and their elemental concentrations attributed to different mints based on numismatic preliminary study.

Zn is present in trace amounts, but one coin from the collection of the National Museum in Poznań showed a Zn content above 1.5%. Hg is also present in trace amounts. One coin housed in the Numismatic Cabinet of the National Museums Berlin (inventory number 18246627) contained a Hg concentration of 1.5%. In principle, however, the concentration of Au, Bi, Hg and Zn remains close to average.

Within this group several types of coins can be distinguished. The PRINCES POLONIE denars (type IX.14), are probably the most important coins in Polish history with the inscription of POLONIE, the first reference to Poland in a historical source and together with coins struck with “repaired” dies (type IX.15) and coins with heavily barbarized dies with a pseudo-legend (with a chapel) form the die-chain no. 2, which are linked to the Gniezno mint, one of Boleslaus the Brave’s capital. As can be seen in Table 5, their chronology starts from the type with a chapel (1000-1010) and examined specimens containing on average 92.9 ± 2.3% of Ag and 4.9 ± 2.0% of Cu together with some Au and Bi content (these specimens appear exceptionally in upper part of heatmap). Furthermore, corrected denars with the legend PRINCES POLONIE with a higher content of Ag (96 ± 1.5% in average) and relatively lower amount of Cu with an average of 1.8 ± 1.7% showing absence of Au and Bi. The earliest coins in this chain (1010- 1020) struck with “repaired” dies have a Ag content averaging 94.6 ± 1.6% and 3.9 ± 1,2% for Cu, without Au and Bi, but with three exceptions (VII-P-793, VII-P-795, MNP GN N5) where Au and Bi was confirmed. These three outliers appear in upper part of the heatmap.

Moreover, there is one more coin struck around the year 1000 and attributed to a mint located in Gniezno. Denar VII-P-798 with a portrait of the ruler and the legend GNEZDVN CIVITAS housed in the National Museum in Krakow is an object of great interest to numismatists because it is still a unique specimen. It contains 91.3% of Ag and 6.3% of Cu together with Au and Bi as minor elements.

High quality denars represented another coin type XII with the inscription REX. Their production has been estimated to the period 1015-1020 and they are also linked to Gniezno. For these coins, Ag and Cu content averages were determined at 96.9 ± 1.5% and 1.8 ± 1.6%, respectively. Au and Bi were not detected in these coins.

The second Boleslaus the Brave group (67 items) is characterized by the lowest fineness (Table 4). Simultaneously there are differences between surface and bulk measurements, where mainly NAA results show lower Ag values and increased amounts of Cu. The mean values of weight percent and standard deviation for Ag obtained with HH-XRF are 90.3 ± 4.3% and for Cu: 7.3 ± 3.9%. The upper part of presented heatmap Figure 5 shows considerable heterogeneity in elemental composition relative to lower part of heatmap. The coins assigned to upper part mostly were minted in Poznań and show a trend of higher concentration for Au, Bi and Zn. The quantity of Pb is however < 1 % for the majority of coins.

This group is characterised by the presence of Au and Bi at various levels. The amount of Zn is higher than in the previous groups with one coin from the National Museum in Krakow (VII-P-813) having 2.0% of Zn. Hg is present in trace amounts also within this group.

This group includes denars with the name of Boleslaus and a schematic representation of his head together with denars which are almost exact copies of Saxon patterns (types: III, III/IV, IV) and grouped together in the first die-chain have Ag and Cu content averaging to 88.1 ± 5.9% and 9.6 ± 5.7%, respectively. All of them contain Au and Bi and were minted between 995 and 1005. There is one outlier within these denars. A less fine type IV coin (18246643) from the Numismatic Cabinet of the National Museums Berlin where Au and Bi were not detected. Except this specimen all coins appear in upper part of heatmap.

Earlier types of coins (types: XI, V/VI) presumably produced in Poznań during 1010-1020 are part of the Polish-Anglo-Saxon-Bavarian die-chain no. 3 where Ag and Cu contents amounts to 92.1 ± 1.9% and 5.6 ± 1.6%, respectively are also assigned to upper part of the heatmap . One type XI.17 coin from the Numismatic Cabinet of the National Museums Berlin-(18246619) does not contain Au nor Bi. In this chain, there is noteworthy one more die that imitates a coin of Mieszko II Lambert that contains 87.2 and 10.3% of Ag and Cu, respectively.

Moreover, the mint located in Poznań was likely the place where the earliest two types of coins were struck. The denar with an arrow (in our study represented by 1 specimen from the Museum of the Middle Pomerania in Słupsk, MPŚ-NM/3202) was found to contain Ag and Cu contents of 94.5 and 4.3%, respectively. In addition, 8 denars with Saint Wenceslas name (type II) had average contents of 93.% ± 1.7% and 3.9 ± 1,7% for Ag and Cu, respectively. Almost all of them without Au and Bi, with one outlier (coin with lower fineness 2308 from the Museum of Archeology and Ethnography in Łódź)Moreover, the absence of Au and Bi in the case of type II coins remains an exception within coins produced in Poznań. All mentioned outliers appear in lower part of heatmap.

In general, Boleslaus the Brave coins produced in Gniezno as well as Mieszko II Lambert denars have a more unified composition with higher Ag content and without significant amounts of Au and Bi compared to those minted in Poznań. Moreover, the silver surface enrichment can be observed for coins signed to Poznań mint. This enrichment might be a result of deliberate action and originates directly from production process, perhaps characteristic for this particular mint.

Two types of Boleslaus the Brave coins: type I (with bust in front) and XIII with the legend in Cyrillic are associated together with an unknown mint. The amounts of Ag and Cu are 89.3% and 8.5% in the case of coin 18225143 (type I) and average to 92.1 ± 4.9% for type XIII. The silver used to produce these coins contains Au and Bi. Especially coins of the type XIII are distinguished by the highest amount of Au compared to other types. These coins are assigned to upper part of heatmap.

A different origin is possible for a denar with the legend MOGILN CIVITAS which could be struck in Mogilno. One specimen (H5-5206) from the National Museum in Prague collection was investigated and contains 98.8% of Ag, and 0.7% of Pb and a characteristic absence of Cu or any additional elements. There are five more coins containing no Cu. Three of them belong to the XI.14 type and two to the XII class.

The elemental composition of this coin is similar to coins from Gniezno.

Separate group consists of coins classified as alleged Boleslaus the Brave denars imitating Saxon, and Bavarian coins with the legends: .VIDV and +JIVA. Eight investigated specimens have a high Ag content 95.7% ± 1.9%. Cu content varies from 5.5% to 0.3%, whereas Au and Bi were not detected (all of them appear in lower part of heatmap).

The obtained results indicate diversification of silver sources for Polish coins production. These results are interesting in the context of other early medieval silver coins. Almost all Mieszko II Lambert coins do not contain Au and Bi, as well as part of Boleslaus the Brave coins mostly minted in Gniezno. Their element composition (similar proportions of main elements and absence of Au and Bi) coincides with composition of Otto-Adelheid pfenninge produced around 1010-1030 probably in Golsar, and with part of Sachsenpfennig produced during the 10th and 11th centuries. Lead isotopic and elemental analyses have demonstrated that the Otto-Adelheid are consistent with ore from Upper Harz [20]. Mining in the Harz Mountains is associated with two ore deposits. While Rammelsberg is mainly a Cu source, Upper Harz (such as Bad Grund deposit) [22] is probably more important for European silver production and could be a potential source for this group of Polish coins.

Boleslaus the Brave coins attributed to different mint located in Poznań are different and characterized bylower Ag and increasing Cu content with notable amounts of Au and Bi. This raises the following question: are we dealing again with an European silver source or perhaps oriental silver was used for manufacturing these coins? The greatest circulation of oriental coins and their influx into Polish lands are associated with the existence of the Samanid state in Central Asia. The weakening of the Samanid state in the middle of the 10th century is considered to be the main reason for stopping the influx of oriental coins to Polish lands [6]. Does the composition of Samanids coins correspond to Polish denars? Previous research has revealed that the Samanid silver is characterised by high Bi and low Au contents [23]. Moreover, the high Bi level of Samanid silver is correlated with high lead isotope ratios. Analogous conclusions were made for Sachsenpfenninge and Scandinavian coins minted in Hedeby which may indicate mixing silver from different sources where addition of non-Samanid silver could reduce the Bi content and the lead isotope ratios. As mentioned above, Bi is a non-volatile element, which finally oxidises during cupellation and remains in a very small amount in the final silver product. The obtained Bi/ Pb ratios in Polish coins indicate that the initial Bi content was low. A low Bi content vs a high amount of Au negate the use of Samanid coins.

Similar levels of Bi and Au were received for Anglo-Saxon and Hiberno-Norse pennies from Dublin. A low Bi together with a high Au content (> 0.3 %) and significant amount of zinc were previously reported by [24] and Merkel in 2016. Moreover, coin no. 18246640 (type XIII.23) has a Au content of about 1.7% which may indicate contamination by gilding. Unfortunately, the silver sources have not been identified neither for Anglo-Saxon nor Hiberno-Norse pennies. Silver with distinctive, higher contents of Au has been reported also for Carolingians silver from the reigns of Charlemagne and Charles the Bald, but these coins are from the 9th century, so they are not contemporary to Polish specimens dated to the late 10th. Furthermore, the context of native silver being used as raw material for these early medieval coins remains open [25, 26].

214 coins from the Early Piasts minted in four different locations were examined by non-destructive X-ray fluorescence (XRF) spectrometry, high energy particle induced X-ray emission (PIXE) as well as particle induced gamma-ray emission (HE-PIGE) and neutron activation analysis (NAA) analysis with the aim of better understanding the metallic composition of the coins issued between 992 and 1020. All coins were found to contain a binary alloy of Ag and Cu, together with minor amounts of Pb and trace levels of Au, Bi, Hg, and Zn.

HH-XRF spectrometry results together with additional surface and bulk analyses carried out for a subgroup of the coins, allowed to classify the coins into three groups. The first group consisted of Mieszko II Lambert coins, which have a high Ag content and generally contained neither Au nor Bi. The second group consisted of Boleslaus the Brave coins also showing high Ag concentrations and characterized by either low amounts or absence of Au and Bi. The third group contained Boleslaus the Brave coins showing lower fineness and lower Ag content in the bulk relative to the surface. This third group is also characterised by the presence of Au and Bi at various levels.

Finally, a combined approach using coin types, die-chains, historical context and chemical composition of coins allowed to associate the coins with at least five different mints, namely Poznań, Gniezno, Mogilno, Giecz, and an undetermined one. A mint located in Poznań can be distinguished for being more diverse in terms of elemental composition of the coins. Silver surface enrichment can be observed for coins assigned to Poznań. This enrichment might be a result of deliberate action and originates directly from the production process, perhaps characteristic for this particular mint.

The results obtained in the present study have revealed that the composition of the Early Piast coins is comparable to that of other European early medieval silver coins such as the Otto-Adelheid pfenning or Sachsenpfennig. The group of Boleslaus the Brave coins characterised by high Au and Zn content vs low amount of Bi is similar to Anglo-Saxon coins and Hiberno-Norse pennies from Dublin. Moreover, it was found that Bi levels are significantly lower in Polish coins relative to Samanid coins. Although it would be interesting to know the relationship between Bi content and the lead isotope ratios for this group of Early Piast coins, micro-destructive analysis such as lead isotopes was not possible due to the uniqueness of these objects. Nevertheless, the proposed approach, involving the use of XRF spectrometry in conjunction with other surface and bulk techniques, was adequate for studying these coins of great archaeological and historical value.

The authors would like to express their gratitude to the Polish Ministry of Science and Higher Education for funding this project under the framework of the National Programme for the Development of the Humanities during the years 2014-2017 (Decision No. 0100/NPRH3/H12/82/2014). The authors would also like to thank Dr. hab. Janusz Lekki from the Institute of Nuclear Physics of the Polish Academy of Sciences in Kraków for performing the PIXE analysis. The authors are also grateful to Dr. Andrzej Bolewski from the Faculty of Physics and Applied Computer Science at the University of Science and Technology in Kraków for conducting the NAA. Special thanks to Dr. Stefan Roerhs from the Rathgen Research Laboratory for carrying out the analysis on Early Piast coins currently housed at the National Museums Berlin.

No conflict of interest.

1. Suchodolski S (1981) A denarius in a wallet. Ossoliński National Institute Publishing House of the Polish Academy of Sciences, Wrocław, Poland, pp. 45.
2. Suchodolski S (1973) Polish coinage in the 11th and 12th centuries. Ossoliński National Institute Publishing House of the Polish Academy of Sciences, Wrocław, Poland, pp. 159.
3. Suchodolski S (2019) The origins of Polish coinage in the light of new research. Slavia Antiqua 60: 192-219.
4. M Bogucki, J Magira (2015) New coins of Bolesław the Brave. Polish Numismatic News 1-2 (199-200): 115-128.
5. K Jonsson, S Suchodolski (2009) A new coin type of Boleslav the Brave found in Sweden, Polish Numismatic News R 1(187): 29-39.
6. D Rozmus (2014) Early medieval district of silver and lead metallurgy in the frontier areas of Upper Silesia and Little Poland (second half of the 11^th-12^th/13^th centuries), Sztygarka City Museum, Academic Bookstore-Krakow, Poland, pp. 392
7. JM del Hoyo-Meléndez, P Świt, M Matosz, M Woźniak, A Klisińska-Kopacz (2015) Micro-XRF analysis of silver coins from medieval Poland. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 349: 6-16.
8. J Lekki, M Matosz, Cz Paluszkiewicz, E Pięta, T. Pieprzyca (2017) Comparison of PIXE and XRF in the analysis of silver denarii of the early Piast. Journal of Radioanalytical and Nuclear Chemistry 314(3): 2309-2316.
9. E Pięta, J Lekki, JM del Hoyo-Meléndez, Cz Paluszkiewicz, M Nowakowski (2017) Surface characterization of medieval silver coins minted by the early Piasts: FT-IR mapping and SEM/EDX studies. Surface and Interface Analysis 50(1): 78-86.
10. A Bolewski, M Matosz, W Pohorecki, JM del Hoyo-Meléndez (2020) Comparison of neutron activation analysis (NAA) and energy dispersive X-ray fluorescence (XRF) spectrometry for the non-destructive analysis of coins minted under the early Piast dynasty. Radiation Physics and Chemistry 171: 108699.
11. M Hrnjic, S Röhrs, A Denker, B Weisser, C Stoess (2021) Multi-technical study of silver denars from medieval Poland for an improved understanding of their archaeological context and provenance. Archaeometry 63(3): 609-626.
12. DA Scott (1991) Metallography and Microstructure of Ancient and Historic Metals. Getty Conservation Institute, Marina del Rey, CA: Getty Conservation Institute in association with Archetype Books, Los Angeles, US, p. 185.
13. L Beck, S Bosonnet, S Réveillon, D Eliot, F Pilon (2004) Silver surface enrichment of silver-copper alloys: a limitation for the analysis of ancient silver coins by surface techniques. Nuclear Instruments and Methods in Physics Research B 226: 153-162.
14. N Shugar, JL Mass (2013) Handheld XRF for Art and Archaeology, Studies in Archaeological Sciences 3. Leuven University Press, Leuven, Belgium, p. 480.
15. FJ Ager, AI Moreno-Suárez, S Scrivano, I Ortega-Feliu, B.Gómez-Tubío (2013) Silver surface enrichment in ancient coins studied by micro-PIXE. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms (306): 241-244.
16. M Hrnjić, GA Hagen-Peter, T Birch, G Hoffmann Barfod, SM Sindbæk (2020) Non-destructive identification of surface enrichment and trace element fractionation in ancient silver coins. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms (478): 11-20.
17. V Kantarelou, FJ Ager, D Eugenidou, F Chaves, A Andreou (2011) X-ray Fluorescence analytical criteria to assess the fineness of ancient silver coins: Application on Ptolemaic coinage. Spectrochimica Acta Part B: Atomic Spectroscopy 66 (9-10): 681-690.
18. R Borges, L Alves, RJC Silva, MF Araújo, A Candeias (2017) Investigation of surface silver enrichment in ancient high silver alloys by PIXE, EDXRF, LA-ICP-MS and SEM-EDS. Microchemical Journal 131: 103-111.
19. R Linke, M Schreiner, G Demortier (2004) The application of photon, electron and proton induced X-ray analysis for the identification and characterisation of medieval silver coins. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 226: 172-178.
20. SW Merkel (2016) Silver and the Silver Economy at Hedeby. Raw Materials, Innovation, Technology of Ancient Cultures RITaK 2, VML Verlag Marie Leidorf, Bochum, Germany, p. 273.
21. ML Héritier, S Baron, L Cassayre, F Téreygeol (2015) Bismuth behavior during ancient processes of silver–lead production. Journal of Archaeological Science 57: 56-68.
22. Chiarantini, I. M. Villa, V. Volpi, G. Bianchi, M. Benvenuti, et al. (2021) Economic rebound versus imperial monopoly: Metal provenance of Early Medieval coins (9th–11th centuries) from some Italian and French mints. Journal of Archaeological Science: Reports 39: 103139.
23. Eniosova, R. Mitoyan (2011) Arabic Coins as a Silver Source for Slavonic and Scandinavian Jewellers in the Tenth Century AD. I. Turbanti-Memmi (ed.), Proceedings of the 37th International Symposium on Archaeometry, pp. 579-584.
24. M. Metcalf, P. Northover (2002) Sporadic Debasement in English Coinage, c. 1009-1052. Numismatic Chronicle 162: 217-236.
25. Garbacz-Klempka, A. Łukaszczyk, D. Rozmus, J. Tokaj (2013) „Skarb hutnika” w świetle badań metaloznawczych. Argenti fossores et alii. Znaczenie gospodarcze wschodnichczęści Górnego Śląska i zachodnich krańców Małopolski w późnej fazie wczesnego średniowiecza (X–XII wiek), ed. P. Boroń, Wrocław, p. 203–224.
26. Ewelina Miśta-Jakubowska (2022) Microanalysis of early medieval archaeological objects made of silver alloy. Polish Numismatic News X, pp. 111-139.