European Geologist Journal 48

Unearthing Europe’s Bronze Age mining heritage with tin isotopes:
A case study from Central Europe

 

by W. Powell 1, R. Mathur2 , J. John3 , M. Price4, H.A. Bankoff5, M. Tisucká6 and L. Godfrey7

Department of Earth and Environmental Science, Brooklyn College – CUNY

Department of Geology, Juniata College

Department of Archaeology, University of South Bohemia

Santa Fe Institute

Department of Anthropology and Archaeology, Brooklyn College – CUNY

Department of Prehistory and Antiquity, National Museum in Prague

Department of Earth and Planetary Sciences, Rutgers University

Contact: wpowell@brooklyn.cuny.edu


Abstract

Being exclusively placer-based, evidence of prehistoric tin mining in Europe was erased rapidly in the fluvial environment. Circumstantial evidence has suggested that the tin ores of the Erzgebirge along the German-Czech border were exploited in the Bronze Age. To investigate this further, tin ores from the three Erzgebirge plutons, as well as Cornwall, were isotopically characterized and compared with the Sn isotopic composition of Bronze Age tin-bearing artifacts from the region. After accounting for isotopic fractionation associated with the smelting process, a probabilistic approach indicates that at the transition to the Middle Bronze Age, the predominant mining center was the Central Pluton, but mining activity shifted to the Western Pluton for the remainder of the Bronze Age.


Introduction

The discovery and characterization of mineral resources remains a cornerstone of applied Earth Science. Since the invention of extractive metallurgy in the Balkans 7,000 years ago, human technological and social development has been inextricably tied to the availability of mineral resources, as is evident in the fundamental three-age system of European prehistory: Stone Age, Bronze Age, and Iron Age. However, much of Europe’s ancient mining heritage remains to be discovered.

Extensive bedrock mining of European copper, dating as far back as the Eneolithic (ca. 5000 BC), has been documented in Serbia and Bulgaria. Although extensive subsurface and open pit tin mines dating to the Bronze Age have been documented in arid regions of Iran, Central Asia, and Central Turkey, such mines are absent in Europe. This is due to the contrasting climatic conditions. Weathering and fluvial processes associated with Europe’s temperate to Alpine climate produced placer tin deposits from the natural breakdown of bedrock ores. Being easier to work and requiring fewer resources, placer deposits were exploited rather than the nearby bedrock ores from which they were derived. Unfortunately, the wooden tools associated with sluicing and panning are unlikely to be preserved, and the mining process itself leaves only ephemeral scars (Tolksdorf et al., 2019). Consequently, the archaeological record of early tin mining in Europe has been largely erased in dynamic fluvial environments.

Lead isotopes have been used in archaeological studies of copper ore provenance for decades, given that Pb isotope compositions can indicate potential ore sources based on the age of the ore from which the metal was smelted. Although this approach has been used successfully to link Late Bronze Age tin ingots from the eastern Mediterranean with Variscan ores (Berger et al., 2019), Pb isotopes cannot be applied to tin provenance studies for bronze, for which the signature is derived from the volumetrically predominant copper ore. The recent development of analytical techniques for the measurement of tin isotope composition has been considered a promising means for approaching tin ore provenance and thereby determining mining practices associated with the prehistoric production of tin bronze.

Recent studies of tin ores and tin-bearing artifacts from Central Europe and the Mediterranean suggest that multiple tin sources were in use across Europe in the Late Bronze Age, including those of Cornwall and the Erzgebirge (Berger et al., 2019; Nessel et al., 2019). However, confidently assigning ore provenance has been considered problematic due to the extensive compositional overlap between metalliferous regions, with the geological processes that produced this extensive fractionation being undocumented.

Although there are 29 radioactive isotopes of tin, none have half-lives long enough for geological time frames; 126Sn has the longest half-life of 100,000 years. So, unlike Pb, there is no age dependence in the Sn system. A given mining region is likely to have similar bulk average Sn isotopic compositions inherited from the source rocks from which the tin granites were derived, and the partitioning of Sn between the magma and the residual rock during anatexis. However, isotopic variation between deposits or ore zones may result from fractionation during ore genesis.

Precipitation of cassiterite requires the oxidation of tin from Sn2+ to Sn4+, and the stronger bonding environment associated with oxidized tin favors the heavier isotopes (Yao et al., 2018; Wang et al., 2019). Thus, redox-related Raleigh fractionation results in the progressive enrichment of lighter isotopes over time (Yao et al., 2018). Thus, small, rapidly formed ores would be expected to have more homogenous isotopic compositions, whereas larger deposits in which the reaction front had advanced significantly over time would likely exhibit a more heterogeneous and zoned isotopic signature. Evolution of a vapor during ore formation also induces fractionation of tin isotopes, resulting in enrichment of heavier isotopes in the vapor, leaving the residual fluids with lower δ124Sn values (Wang et al., 2019). Thus, ores formed at shallower depths at which boiling is possible will exhibit greater degrees of Sn fractionation (Wang et al., 2019). These processes may lead to isotopic fingerprints of related ore bodies that would match that of their derived metals.

A case study in Central Europe

The Erzgebirge (Ore Mountains) of Saxony and Bohemia is a classic ore locality that hosts both copper and tin mineralization. In this region, tin mining (both bedrock and placer) has been documented as far back as the 12th century at Ehrenfriedersdorf and Altenberg, and Agricola’s De Re Metallica (1556) includes references to the mining of Erzgebirge placers. Archaeological evidence of prehistoric mining of these ores has been mostly circumstantial, based on the spatial association of placer tin sources with clusters of Early Bronze Age Únětice culture and Late Bronze Age Lausitz culture settlement sites. However, a recent geoarchaeological study documented direct evidence of placer mining near Altenberg in the early second millennium BC (Tolksdorf et al., 2019).

The Carboniferous granites (330-295 Ma) that are associated with tin mineralization in the Erzgebirge include the fault-dissected Western pluton, the poorly exposed Central pluton, and the Eastern pluton (Figure 1). The granitic magmas were produced from the late syn-collisional melting of Sn-enriched sedimentary rocks derived from the erosion of Gondwana (Romer & Kroner, 2016). Mineralization is predominantly associated with greisens and, to a lesser extent, vein systems. Tectonic tilting of the section resulted in larger, more deeply eroded, predominantly S-type plutons being exposed in the southwest, and smaller, A-type, volcanic-associated plutons to the northeast (Breiter, 2012). Such variation in geological setting would be expected to be associated with variation in Sn isotopic composition between the plutons.


Figure 1: Granitic plutons of the Erzgebirge region. Named sites on this map refer to sites from which cassiterite samples were obtained.


Here we focus on the Erzgebirge and its potential exploitation at two points in prehistory: the later phase of the Early Bronze Age and at the beginning of the Middle Bronze Age (Reinecke BrA2 to BrB1, c. 1900–1500 BC); the latter Late Bronze Age (Reinecke BrB2 to HaA, c. 1500-1000 BC). Given their proximity and size, deposits of the Erzgebirge and Cornwall are the most likely sources of tin at this time.

To characterize the ores, the isotopic composition of 43 new cassiterite samples (predominantly single crystals from the collection of the American Museum of Natural History) from the Erzgebirge and 14 from Cornwall were added to the dataset of Haustein et al. (2010) for a total of 89 samples from the Erzgebirge (38 from the West Pluton; 16 from the Central Pluton; 35 from the East Pluton), and 41 from Cornwall. Cassiterite samples were digested following the procedure of Mathur et al. (2017).

Bronze Age copper ingots in the shape of ribs (Figure 2) are widespread mainly in southern Germany, Upper Austria and South Bohemia. Specifically, more than 40 hoards of these ingots are known in the South Bohemian region. Copper ribs are found almost exclusively in hoards, which usually contain tens, but sometimes hundreds of ribs. The largest hoard so far was found in Oberding in Upper Bavaria in 2014 and contained 796 ingots.


Figure 2: Rib ingots from the hoard at Habří, South Bohemia.


The precise chronological definition of rib ingots occurrence is difficult, as hoards generally do not contain any types of chronologically sensitive artifacts. In general, their occurrence can be expected in the later phase of the Early Bronze Age and at the beginning of the Middle Bronze Age, according to the Central European chronology A2 and B1 periods. The weight of the rib bars has a relatively large dispersion, but on average it is around 100 grams. The rib ingots have a relatively varied elemental composition, sometimes including tin (up to 3%) and, in rare cases, large amounts of lead (up to 38%).

In this study, 30 rib ingots from five Bohemian hoards (Habří, Kroclov, Kučeř, Slavč, Veselíčko) with a range of 0.4 to 3.5 wt% Sn were analyzed. In addition, 17 later bronze items (BrB to HaA) including pins, wires, bracelets, and axes from museum collections in České Budějovice, Prague, and Salzburg were analyzed. Samples were obtained either by drilling with cobalt-tipped bits or cutting with cubic boron nitride cutting wheels, after the patina was removed. Approximately 0.02-0.1g of each bronze sample was digested in ultrapure aqua regia (3ml HCl, 1ml HNO3) to which 0.02ml HF was added. Samples were heated at 90°C for 6 hours in enclosed Teflon containers.

A small aliquot of each sample solution was dried and the redigested solutions were purified using the ion exchange chromatography described in Balliana et al. (2013). Samples were measured on the Isoprobe MC-ICPMS at the University of Arizona and the Neptune MC-ICPMS at Rutgers University within 24 hours of digestion. Solutions were measured and corrected for mass bias as described in Mathur et al. (2017).  Data are presented relative to the NIST 3161A Sn standard (Lot# 07033) in per mil notation defined as: 

Whole procedural 1σ errors for analysis are δ124Sn= 0.08‰.

Histograms of the Sn isotopic compositions of cassiterite from these four sites are illustrated in Figure 3 as bar graphs. Model curves for the cassiterite data were calculated using a Bayesian framework by sampling for the parameters of a three-component Gaussian mixture density using the Stan programming language (https://mc-stan.org/users/citations/) called with the rstan R package, a platform for statistical modeling and statistical computation   (https://www.R-project.org/). In Figure 3, the black curve is the 50% quantile density for each δ124Sn value across posterior samples and the blue bands mark the +/-2.5% quantile density. Although there is substantial overlap between sites, variations in mean and range are evident. As was predicted based on the geological setting of the three plutons of the Erzgebirge, the deeper level ores of the West Pluton exhibit the lowest degree of fractionation (Figure 3d), and the East Pluton tends to higher values of δ124Sn (Figure 3b). The isotopic composition of Cornwall and the East Pluton are similar (Figure 3a and b), although Cornwall tends toward slightly higher values.

The two groups of bronze artifacts from Central Europe exhibit distinct isotopic distributions. The older rib ingots have a modal peak of δ124Sn 0.65‰ and a positively skewed distribution, whereas the group of younger artifacts have a δ124Sn mode of 0.4‰ and are negatively skewed (Figure 3e). This indicates a change in tin ore source from the Early-Middle Bronze Age transition to the Middle and Late Bronze Age.


Figure 3: Histograms of δ124Sn values for cassiterite and bronze samples. (a through d) Histograms in yellow, and model curves for the cassiterite data the black curve is the 50% quantile density for each δ124Sn value across posterior samples, blue bands mark the ±2.5% quantile density. (e) Histograms of bronzes from two time periods of the Bronze Age. MBA=Middle Bronze Age; LBA=Late Bronze Age; n=number of samples.


Given that there is a compositional overlap between the four localities, a probabilistic approach was adopted. Figure 4 plots posterior probability bands as a function of δ124Sn value assuming the prior probability of all four sources are equal (0.25). The probability bands in Figure 4 apply to single observations. In order to assign provenance probabilities to assemblages of artifacts, which consist of multiple observations with differing δ124Sn values, we implemented a model in which all artifacts were assumed to originate from the same source. The likelihood that each source gave rise to the assemblage was calculated, averaged across Bayesian samples of the cassiterite densities. Assuming, as before, that each source has the same prior likelihood, the posterior likelihood is proportional to these averaged likelihoods, normalized such that the posterior likelihood sums to one.


Figure 4: (a) Posterior probability bands as a function of δ124Sn value assuming the prior probability of all four sources is equal. (b) Isotopic composition of bronze artifacts corrected for smelt-associated fractionation (-0.2‰).


Kazenas et al. (1996) noted that SnO2 evaporates at a noticeable rate above 1225°C, a process likely to induce fractionation. Furnace-based casting experiments at 1100–1200°C documented a 0.5‰ per amu shift in composition along the surface due to selective evaporation of lighter isotopes of tin (Yamazaki et al., 2014). Based on cassiterite smelting experiments, Berger et al. (2019) concluded that a negative shift of 0.25 ‰ per amu in the composition of tin-bearing artifacts is required before comparisons are made with isotopic compositions of ores in order to compensate for smelting induced fractionation. Accordingly, the provenance probability model was run using the measured δ124Sn isotopic values of artifacts, as well as measured δ124S minus 0.2‰ (124/116, 8 amu). The results are shown in Table 1.


Table 1: Probability of derivation of tin from groups of artifacts from a single ore source.

Artifact Assemblage

 

Smelt Fractionation

 

Provenance Probability

West Pluton

Central Pluton

East Pluton

Cornwall

EBA-MBA

0.0

0.0%

0.6%

0.7%

98.7%

EBA-MBA

0.2

0.0%

99.8%

0.2%

0.0%

MBA-LBA

0.0

87.4%

11.0%

1.6%

0.0%

MBA-LBA

0.2

99.9%

0.0%

0.0%

0.0%


Assuming that the tin for the Middle to Late Bronze Age artifacts from the Czech Republic and Austria was derived from one of the four tin sources studied here, there is a high probability that the ore was associated with the West Pluton of the Erzgebirge (99.9% with -0.2‰ correction; 87.4% without correction). The Sn isotopic composition of the older rib ingots indicate that the associated tin was sourced from a different site. If one includes the correction for evaporative loss during smelting, then the model indicates that there is a 99.8% probability that the tin ore was derived from the Erzgebirge Central Pluton. If, however, the uncorrected value is used, the model indicates a 98.7% probability that the tin was of British origin. Given that there is no archaeological evidence for a strong trading connection between Britain and Bohemia in the late Early Bronze Age, the Central Pluton provenance associated with the 0.25‰ per amu correction factor for smelting induced fractionation is considered more likely. This would provide empirical evidence to support the experiment-based conclusion of Berger et al. (2019), and that this correction factor should be used in future Sn provenance studies.


References

Balliana, E., Aramendía, M., Resano, M., Barbante, C., & Vanhaecke, F. 2013. Copper and tin isotopic analysis of ancient bronzes for archaeological investigation: development and validation of a suitable analytical methodology. Analytical and Bioanalytical Chemistry, 405. 2973-2986. DOI: 10.1007/s00216-012-6542-1

Berger, D., Soles, J.S., Giumlia-Mair, A.R., Brügmann, G., Galili, E., Lockhoff, N., & Pernicka, E. 2019. Isotope systematics and chemical composition of tin ingots from Mochlos (Crete) and other Late Bronze Age sites in the eastern Mediterranean Sea: An ultimate key to tin provenance? PLOS One, 14. e0218326. DOI: 10.1371/journal.pone.0218326

Breiter, K. 2012. Nearly contemporaneous evolution of the A- and S-type fractionated granites in the Krušné hory/Erzgebirge Mts., Central Europe. Lithos, 151. 105-121. DOI: 10.1016/j.lithos.2011.09.022

Kazenas, E., Bol’shikh, M., & Petrov, A. 1996. Thermodynamics of processes of vaporization, dissociation, and gas-phase reactions in vapors over tin-oxygen system. Izvestiya Rossiiskoi Akademii Nauk. Metally, 3. 23–29.

Mathur, R., Powell, W., Mason, A., Godfrey, L., Yao, J.M., & Baker M.  2017. Preparation and Measurement of Cassiterite for Sn Isotope Analysis. Geostandards and Geoanalytical Research, 41. 701-707. DOI: 10.1111/ggr.12174

Nessel, B., Brügmann, G., & Pernicka, E. 2019. Tin isotope ratios in Early and Middle Bronze Age bronzes from Central and southeastern Europe. The Journal of the International Union for Prehistoric and Protohistoric Sciences, 2. 1-11.

Tolksdorf, J., Schröder, F., Petr, L., Herbig, C., Kaiser, K., Kočár, P., Fülling, A., Heinrich, S., Hönig, H., & Hemker, C. 2019. Evidence for Bronze Age and Medieval tin placer mining in the Erzgebirge mountains, Saxony (Germany). Geoarchaeology, DOI: 10.1002/gea.21763

Wang, D., Mathur, M., Powell, W., Godfrey, L., & Zheng, Y. 2019. Experimental Evidence for Fractionation of Tin Chlorides by Redox and Vapor Mechanisms. Geochimica et Cosmochimica Acta, 250. 209-218. DOI: 10.1016/j.gca.2019.02.022

Yamazaki, E., Nakai, S., Sahoo, Y., Yokoyama, T., Mifune, H., Saito, T., Chen, J., Takagi, N., Hokanishi, N., & Yasuda, A. 2014. Feasibility studies of Sn isotope composition for provenancing ancient bronzes. Journal of Archaeological Science, 52. 458-467. DOI: 10.1016/j.jas.2014.09.014

Yao, J., Mathur, R., Powell, W., Lehmann, B., Tornos, F., Wilson, M., & Ruiz, J. 2018. Sn-Isotope Fractionation as a Record of Hydrothermal Redox Reactions. American Mineralogist, 103. 1591–1598. DOI: 10.2138/am-2018-6524


This article has been published in European Geologist Journal 48 – Geological heritage in Europe. Read here the full issue: