Beyond the Great Wall: Gold of the Silk Roads and the First Empire of

Subscriber access provided by FORDHAM UNIVERSITY

Article

Beyond the Great Wall: Gold of the Silk Roads and the first Empire of the steppes Martin Radtke, Ina Reiche, Uwe Reinholz, Heinrich Riesemeier, and Maria F. Guerra Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/ac3025416 • Publication Date (Web): 12 Dec 2012 Downloaded from http://pubs.acs.org on December 22, 2012

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Beyond the Great Wall: Gold of the Silk Roads and the first Empire of the steppes Martin Radtke+*, Ina Reiche#°, Uwe Reinholz+, Heinrich Riesemeier+and Maria F. Guerra#°. +

BAM Federal Institute for Materials Research and Testing, Richard-Willstätter-Str. 11, 12489 Berlin, Germany #

present address: Laboratoire d’Archéologie Moléculaire et Structurale, UMR 8220 CNRSUPMC, 3 rue Galilée, 94200 Ivry-sur-Seine, France

°

Centre de Recherche et de Restauration des Musées de France, Palais du Louvre – 14, quai François Mitterrand, 75001 Paris, France

KEYWORDS: SRXRF, BAMline, gold, provenance, alloy, Xiongnu, platinum

ABSTRACT Fingerprinting ancient gold work requires the use of non-destructive techniques with high spatial resolution (down to 25 µm) and good detection limits (µg/g level). In this work experimental set-ups and protocols for SRXRF at the BAMline of BESSY in Berlin for the measurement of characteristic trace elements of gold are compared considering the difficulties, shown in previous works, connected to the quantification of Pt. The best experimental conditions and calculation methods were achieved by using an excitation energy of 11.58 keV, a SDD detector and pure element reference standards. A detection limit of 3µg/g has been reached. This

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

newly developed method was successfully applied to provenancing the Xiongnu gold from the Gol Mod necropolis, excavated under the aegis of UNESCO. The composition of the base alloys and the presence of Pt and Sn showed that contrary to what is expected, the gold foils from the first powerful empire of the steppes along the Great Wall were produced with alluvial gold from local placer deposits located in Zaamar, Boroo and in the Selenga River.

Text: Defining the trade routes of ancient civilizations is of particular importance when no written sources are available. The fragility, rarity and preciousness of ancient gold work require the use of non-destructive techniques and for this reason, in most cases, the study of this type of objects is restricted to the determination of their base-alloy compositions1. When sampling the objects is possible, which is hardly ever the case, isotopic analysis with TIMS or LA-MC-ICP-MS can in some particular cases accomplish fingerprinting2. However, it is the determination of minor and trace elements contents that prevails in this field of research carried out by IBA techniques (PIXE and PIXE-XRF), nuclear activation analysis, and ICP-MP in liquid or in LA mode 3. Synchrotron Radiation induced X-Ray Fluorescence (SRXRF) was applied to the analysis of a few gold items evidencing the potentiality of this technique 4. However, in the case of fingerprinting while for most minor and trace elements SRXRF is a well suited technique, the detection limits in routine conditions necessary for the quantification of one of the most important characteristic elements of secondary gold – platinum 1 – could not be reached 4f, 5. This


Page 2 of 27

Page 3 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


is due to the neighbourhood of Pt and Au: the standard method that uses an excitation energy below the absorption edge of the main element to enhance the sensitivity for the traces cannot be applied in this case, because the XRR (X-ray Resonant Raman) scattering of Au overlaps the Pt signal 6. The role of geochemical studies of gold deposits – primary and secondary – and gold metals for provenancing in archaeology has been discussed in the literature7as well as the variability of the gold grains composition in different types of deposits8.Moreover, the deposition of heavy minerals (from which are mined gold, mercury, platinum, cassiterite (tin oxide, SnO2), etc.) in stream sediments, gravels and sands,recovered by panning is a method still used nowadays 9

based on differences in densities in a same pan. The presence of Pt and Sn in ancient gold

derives in most cases from this exploration process and can be used to identify the use of placer gold whereas other heavy elements such as Hg (melting point 234 °C) are evaporated or/and absorbed by the cupel during metallurgical processing10.The melting point of cassiterite (SnO2) is 1127 °C, and of platinum is 1772 °C, while the melting point of gold is 1064.18 °C . In this work we propose a new analytical approach by optimization of the experimental setup and by processing data with a new quantification model. Instead of correcting the Raman scattering by comparing measurements above and below the Pt-L3 edge 5, pure comparator materials were used. The Pt-Lα1 line was selected as analytical signal and the excitation energy was chosen between the Au-L3 and the Pt-L3 edge. The measured spectra were fitted by a sum of the spectra obtained for pure standards. The potential of this new analytical and theoretical approach is demonstrated by the study of the gold from the Gol Mod necropolis, situated at 500 km west of Ulan Bator, excavated by the



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

French archaeological mission in Mongolia under the aegis of UNESCO (see fig. 1). In this Xiongnu necropolis the wood coffins are covered with typical Chinese motifs, for the major tombs in gold 11, as shown in figure S-1. Although all main Xiongnu settlements and complexes are located in regions where placers have always been exploited for gold 12, the question that rises for the gold found in Gol Mod is its provenance: local productions or Chinese gifts? The Xiongnu, a nomadic confederation living beyond the Great Wall, dominated in the 2nd century BC a large territory which spread from Manchuria to the western regions and to Lake Baikal 13. Only Chinese documents and chronicles refer to the Xiongnu. However, contrary to these scarce written sources, the archaeological remains revealed a complex economical structure and an intricate society 14 based on a pastoral production, tributes paid by submitted tribes 15 and trade goods, gifts and tributes from China 16. Remaining for more than 500 years the major military power in East Asia along the Chinese frontier 16, the Xiongnu emerged in Chinese history as excellent warriors who defeated the Han during their expansion into the Xinjiang . This expansion is related to Zhang Qian’s journey to central Asia (139-125 BC) 17, which is at the origin of the opening of the Silk Road. The construction of the Great Wall was the Chinese response to the nomadic aggressions 18 as well as the policy of "harmonious relations" adopted by the Han who sent large quantities of rich offerings to the Xiongnu19, among which gifts made of gold as cited in the ancient texts 20. In fact, in early China gold was not related to status 21. Till the Han gold is mostly used as foil decoration and jade and bronze were the symbol of status and wealth. Personal jewellery in gold is very rare 22 while among the pastoral tribes gold was particularly admired, especially in the Ordos 20.


Page 4 of 27

Page 5 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Except one attempt to determine the variety of the gold alloys and the origin of the Xiongnu’s gold in the case of Gol Mod 3e, f with IBA techniques, no systematic analysis was carried out on a large set of Xiongnu gold remains in order to identify the variety of the gold supplies and particularly localize the possible sources of gold. The goal of this work is to determine whether the gold used to decorate the Gol Mod coffins was imported from China or locally exploited and whether the gold sources and the gold metallurgical processing are related to the social rank of the tombs. Experimental Methods Measurements have been performed with the SRXRF setup at the BAMline at BESSY in Berlin23. The radiation source is a 7 Tesla wavelength shifter. As monochromator a double crystal with Si 111 crystals was used. Higher harmonics were suppressed by the additional use of a double multilayer monochromator. X-rays were detected with a silicon drift chamber detector (SDD). All samples were mounted on a motorized xyz-stage and positioned with an accuracy of better than 10 microns in all dimensions with a long distance microscope. The measuring time was 300s per spectra with a beam size of typically 100 x 100 µm². One point per sample was measured, due to the limited size of the samples.

Details of the method The concentrations of the main elements of the gold alloys were obtained with an excitation energy of 34 keV (figure S-2). The gold, silver, copper and tin contents have been determined by fitting the spectra with the QXAS software and calculating the concentrations using thick pure element reference materials 24.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The measurement of the Pt contents required a more complex approach. In the first step the optimum excitation energy was determined by a set of measurements at different excitation energies. The energy range was determined by the conditions to measure below the Au-L3 edge and above the Pt-L3 edge. Measurements were made on pure Pt and Au foils with a step size of 10 eV and the optimal energy was defined as the best ratio of the intensity of the Pt Lα peak to the intensity of the Au scattering in the region of the Pt Lα peak. Figure 2 shows the result of this measurement. A similar approach has been applied for the determination of Cu thin films on a thick gold substrate 6a. Due to the smaller energy difference between the Au-L3 edge and the PtL3 edge, compared to Au-L3 and Cu-K, in our case it was not possible to shift the Raman contribution totally out of the region of interest. An optimum energy of 11.58 keV was determined. All subsequent measurements for Pt have been performed with this energy (figure S-2). Additional pure element standards for Ag, Zn and Cu have been measured to model the spectra in the relevant energy range. This has been done by fitting the sum of the pure element spectra to the spectra of the sample. Some examples of the obtained fits are shown in figure 3. The presence of Pt in gold is usually associated to other platinum group elements (PGE). For this reason, when applying this method to real samples in some cases a good fit of the spectra was not possible. In our case this was due to the presence of iridium. As Ir wasn’t at hand as pure element standard, the expected spectrum was simulated, using the detector characteristics (noise, Fano factor and energy calibration) obtained from the Pt standard. To complete the possible contributions to the measured spectra a pure Os spectrum was obtained with the same procedure. When all this data was included into the fit, all spectra matched satisfyingly. To ensure, that the part of the spectrum allocated to Pt is really due to this element, the correlation between the pure


Page 6 of 27

Page 7 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Pt signal and the fit subtracted by all distributions except Pt was calculated. The correlation factor is usually above 0.99, which confirms the fact that the fit was correct. To evaluate the reliability of this method samples with known Pt content have been measured. Unfortunately the availability of standards with certified concentrations for Pt is limited and the only available certified standard, RM 8058 with a 40.8 µg/g Pt content, served as standard for our quantification. Therefore samples which have already been measured by PAA or LA ICPMS 25 were used as control materials. Results of these measurements are given in table S-1. A good agreement between expected and measured values was achieved and the MDL of 3 µg/g outreach the other methods with nearly one order of magnitude. An important aspect is the uncertainty of the obtained values. As there are no certified standards available, the uncertainty must be assessed from the measurements on the known samples from table S-1. There are two main components of the uncertainty: first the statistic of the measured spectra, second the uncertainty resulting from the fit. The second contribution can’t be determined exactly, due to the missing standards. Anyhow, a difference of a 5 percent for the Pt signal in the fit causes a difference of 10 percent in the quadratic difference of fitted and measured signal, which is the control parameter of the fit, therefore this contribution can be safely be assumed to be less than 10 percent. With the used measuring conditions the counts have been roughly 1000 counts for 1 µg/g Pt. This means for the lower Pt contents of around 50 µg/g we can assume a statistical error of 0.5 %, which is negligible compared to the fit error. The overall uncertainty can therefore be assumed to be in the order of 10%.

Comparison of different methods for the quantitative determination of Platinum in Gold



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The authors already presented two protocols to measure the Pt content in gold samples. The first was based on differential measurements above and below the Pt-L3 absorption edge5. The difference between these two measurements was allotted to Pt and quantitatively evaluated. The advantage of this method is that no additional standards are necessary to model the spectral background. With this method a MDL of 20 µg/g was reached for Pt. The drawback of this differential method is that a complex data evaluation scheme must be applied. The treatment of the Raman scattering introduces an additional uncertainty in the results. Therefore we developed a second method using the measurement of the high energetic Pt-K lines 4f. Measuring the Pt-Kα lines, due to the overlap with Os and Ir lines, which have been as well present, a MDL of 40 µg/g was achieved. While the data evaluation is straight forward in this case, the measurements at the K-edge are suffering from reduced X-ray detection efficiency and the need to use a multilayer monochromator with a poor energy resolution to obtain a reasonable flux at the necessary excitation energy. An additional problem is the influence of the sample thickness. Using the Pt-K-lines the information depth is around 0.5 mm. When this technique is applied to samplings with a thickness, which is not exactly known, such as those embedded in resin, the uncertainty in the sample thickness leads directly to an uncertainty in the Pt concentration. The method presented here avoids complicated calculations because: (1) The fitting of the measured spectra of the sample by the measured spectra of pure elements is straight forward. (2)

Comparing the spectra for pure Pt and the part of the fit allotted to Pt in the fit gives an

objective criterion for the goodness of the fit. (3) The calculation of the Pt content can easily be done by comparison with a standard.


Page 8 of 27

Page 9 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(4) A sample thickness of a few microns only is sufficient to assume an infinite sample thickness for calculations. In this situation the MDL is significantly improved. Therefore we believe, that this new protocol is at present the best available for non-destructive analysis of Pt in Au. Nevertheless, further improvement can be expected by using wavelength dispersive detection. A better energy resolution will allow separating Raman and fluorescence radiation completely.

Results and Discussion Horse chariots, bronze mirrors, jades and silk from China, turquoise from central Asia, amber from the Occident, and a profusion of other goods such as the rich belt plaques and the gold objects and decorations suggest that one of the last Xiongnu chieftains, Yu (18-46), Wudadihou (46) or Punu (46-?), could be buried in tomb 1 of Gol Mod only tomb 6 of NoinUla necropolis found in 1912

27

26

. Till the discovery of this tomb,

could be attributed to chieftain Wushuliu

who died in 13 AD 28. The finds of Gol Mod are representative of the goods traded along the Silk Roads and correspond to both the Chinese and the steppes worlds. The gold decoration foils (or the gold alloys) could be a local production or a gift from China. We note that one jade pendant from tomb 1 is a reemployment of a Chinese xi pendant, dated to the Warring States period (475-221 BC), by addition of two gold nails 11. We applied our protocol to the study of 38 gold foils from tomb 1,high rank tomb 79, and lower social rank satellite tombs 43 and 45 to determine the concentrations of the major elements (Au, Ag and Cu) of the alloys and of Sb, Sn and Pt, which are in general characteristic of alluvial gold 1, 10a, 25 recovered by panning.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 27

Discussion on data Major elements The composition of the gold alloys is given for all the analyzed foils in table S-2. Cu and Ag contents are represented in figure 4. We observe that the foils of tomb 79 are made of one single alloy containing about 10% Ag and 0.2% Cu while the foils from tomb 1 have Ag contents ranging from 1% to 6%, Cu being present at contents lower than 0.3 %. The foils from satellite tombs 43 and 45 show heterogeneous compositions, but contained in the range defined by tombs 1 and 79 except those with higher Cu contents. About 600 gold occurrences and deposits are known in Mongolia

29

; a list of placer gold

deposits is given by Dejidmaa30. Nowadays, among the major gold mining districts we find Zaamar and Boroo31 located near Gol Mod. Nevertheless, the published studies on the geochemistry of placer deposits in Mongolia only concern the composition of the gold grains. The authors are not aware of any publication dedicated to the relation between the composition of those gold grains and the composition of the gold metal derived from panning in rivers. The presence of platinum and cassiterite in the mining regions of Mongolia was discussed by several authors3233and we based our assumptions concerning the presence of Sn and Pt in placer deposits on the fact that these elements should be present in the pan together with the gold. The gold grains from the Zaamar placer and terrace deposits in the Tuul river valley contain 87% Au 34 and those from the placer deposits of the Selenga River contain from 96.8% to 99.9% Au 35. The composition of the gold from these deposits matches the composition of the Gol Mod foils. The quality of gold in one same region such as in the Boroo alluvial deposits

36

is also in

the range of the Gol Mod foils. The analysis by Murao 2006 of nuggets from 3 different deposits


Page 11 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


at Boroo showed that those from Boroo-10 have 92.6-94.0% Au, those from Tsaganchulut 82.387.8% Au, and those from Sunzigt 89.3-95.5% Au

29

. When more distant deposits are

considered, for example in Savkino where the grains contain 80-85% Au 37, the compositions are out of the range of the foils from Gol Mod. In order to check our assumption on the use of local gold in the production of the gold foils at Gol Mod, these foils were compared to a few gold fragments analysed for major element contents by XRF (Table S-3) from the Egiin Gol necropolis, situated 175 km north of Gol Mod 26

, and from the rich tombs of the Takhiltin cemetery

38

and from the Shombuuziin burial

39

,

situated in the greater Gobi-Altai region of western Mongolia on northwest region of the Silk Roads, fig. 5. We also included in figure 5 the Cu and Ag contents of the gold rivets from the Chinese xi jade pendant of tomb 1 and of four gold plaques attributed to the Han dynasty published in 3e, which could be used as a reference for “Chinese” gold. The foils from Gol Mod match the fragments from Egiin Gol, but those from Takhiltin and Shombuuziin are of lower quality and separated in two chemical groups. Placer gold in Mongolian Altai goes from high quality of nearly 100% fineness (Au content) at Ulaan Ul over Erdene Uul with 90-95% down to as low as 80.2% in Bayan Tolgoi40. The fineness of native gold in quartz vein from this deposits ranges from 75% to 100%, some deposits showing two different qualities (for example at Gobi Altai Terrane 98% and 86%), which could explain both the variety and the separation into two groups of Takhiltin and Shombuuziin gold foils. Gold of lower quality can be found but in deposits located far from Gol Mod 41. At last we must remark that the gold rivets from the jade pendant of tomb 1 and the four gold plaques attributed to the Han dynasty (designed as “Chinese” gold) form a separated group.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 27

Trace elements The very low Cu contents in the gold foils from Gol Mod and the variety of Ag concentrations are in the range of values expected for gold from the deposits located in the area of influence of Gol Mod, mainly from Zaamar and Boroo. This would prove that the gold decoration in this necropolis was a local production made by exploitation of alluvial deposits instead of gifts from China. The identification of one inclusion of cassiterite (SnO2) under the SEM

4f

reinforced the

assumption of the use of alluvial gold. However, the search for PGE inclusions under the stereomicroscope and by SEM was not successful. As expected, none could be found on the studied 30 to 60 µm thick gold foils. The hardness of PGE grains makes it difficult to producegold leafs by hammering. Therefore the use of alluvial gold in the production of the foils can only be proved by the presence of Sn and Pt in the alloys. Figure 6 reports the Sn and Pt concentrations normalized to the gold content for all the foils analysed by SYXRF. The fact, that major elements match the expected composition for alluvial deposits in the different regions of Mongolia, allowed us considering that the Sn contents in the gold from Gol Mod is only related to the presence of cassiterite in the placer deposits where gold was panned. For this reason, a change of the Sn/Au ratio could be considered as a change of the exploited deposit. Hence, we assume that the gold from tomb 1 is clearly alluvial and, in spite of a slight superposition, different from the gold used in the decoration of tomb 79. Gold from tomb 79 is contained in one chemical group of higher Sn contents. Satellite tombs 43 and 45 contained very few gold remains, which increase the difficulty of fingerprinting. Fragments from tomb 43 seems to match those from tomb 79 however, those from tomb 45 seem to be produced with gold from several origins, with different Sn and particularly Pt contents.


Page 13 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


All the gold foils from Gold Mod were produced with alluvial gold. If we except fragments from tomb 45 the Pt/Au ratio varies from 0.6 to 1 and the Sn/Au ratio from 0 to 1. The Cu and Ag contents indicate that the gold decorations of the burials of Gol Mod are in fact a local production in the Chinese tradition and not a Chinese gift. Our results reinforce the recent analytical research on Xiongnu’s metallurgical productions. Xiongnu and Han iron technologies showed up to be distinct

42

, and the possibility of local

production is reinforced by the presence of an iron site with bloomery-based iron technology located on the western shore of Lake Baikal 43. Bronze objects found in different Xiongnu burials are identically local productions and follow a technological transition from the steppe arsenicrich alloys to the high tin lead alloys characteristic of pre- and early imperial periods of China 44.

Conclusion

Major, minor and trace elements for a variety of gold samples from the Xiongnu necropolis of Gol Mod were determined. The importance of Pt for the discrimination of different gold sources made it necessary to develop a new non-destructive method with enhanced sensibility for Pt detection. This was achieved by the use of synchrotron radiation based XRF and an evaluation strategy based on pure element spectra and an elaborated fit procedure. The reached minimum detection limit of 3 µg/g accomplished the given requirements. The variety of the base alloys composition and the presence of Pt and Sn (one inclusion of cassiterite was found in one gold foil) in the gold foils from the decoration of the coffins of Gol Mod showed the use of alluvial gold for their production. These compositions are in the range expected for placer gold deposits and match the composition of the native gold in quartz veins



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 27

located in the same region of Gol Mod: Zaamar, Boroo and the Selenga River. The differences found for the Sn/Au and Pt/Au ratios were assumed to derive from panning in different sedimentary regions. The assumption that the Xiongnu use local gold from the immediate area of influence of the settlement to produce the gold decoration of their coffins is reinforced by the similar composition of the foils from Egiin Gol necropolis and the dissimilar composition of the foils from two Xiongnu necropolis in the Altai region, which are however consistent with the composition of the gold from local deposits. These results complete the overview on the Xiongnu metallurgy, showing that like bronze and iron, gold objects were local productions and not Chinese gifts as affirmed in ancient documents. The foils of tomb 79 of Gol Mod showed an uniform base-alloy composition (Au-Ag-Cu), which can be explained by the use of one single gold source by one workshop. Tomb 79 is the burial of a highly ranked person, which means that its decoration could be ordered to a single workshop. Tin concentrations in this tomb are a little dispersed, which could be explained, if we assume the exploitation of one alluvial deposit, by either the use of metallurgical processes with small differences in temperatures or the presenceof more dispersed cassiterite inclusions than in the foils from other tombs. The SR-XRF measurements with a beam of small dimensions, possibly did not allow a total averaging on the samples.. In the case of tomb 1 we observe a small dispersion for the base-alloy composition that can be explained by the importance of the tomb, which required a larger quantity of gold, and so the exploitation of several placers, and maybe the work of several gold workers. With the exception of one point in Figure 6 (considered as a particular case), tin concentrations show that the gold from tomb 1 is different from the gold used in tomb 79. Platinum concentrations show essentially


Page 15 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the variability of the gold sources in the case of the satellite tombs. This could lead to the assumption that these tombs could be decorated with the remaining gold (or gold foils) from the different important tombs decorations in one workshop or still produced in a different workshop, which was supplied with gold from independent sources.

Supporting Information Available: This material is available free of charge via the Internet at http://pubs.acs.org.”



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 27

Acknowledgements The authors are grateful to Jean-Paul Desroches and Guilhem André from the French Archaeological Mission in Mongolia for providing access to the gold objects and for guidance about the archaeological site.

Corresponding Author [email protected] REFERENCES 1. M. F. Guerra, An overview on the ancient goldsmith's skill and the circulation of gold in the past: the role of x-ray based techniques. X-Ray Spectrometry 2008, 37. 317-327, DOI: 10.1002/xrs.1013. 2. (a) S. Baron, C. G. Tamas, B. Cauuet, M. Munoz, Lead isotope analyses of gold-silver ores from Roia Montana (Romania): a first step of a metal provenance study of Roman mining activity in Alburnus Maior (Roman Dacia). Journal of Archaeological Science 2011, 38. 10901100, DOI: DOI 10.1016/j.jas.2010.12.004;(b) C. Bendall, D. Wigg-Wolf, Y. Lahaye, H. M. Von Kaenel, G. P. Brey, Detecting Changes of Celtic Gold Sources through the Application of Trace Element and Pb Isotope Laser Ablation Analysis of Celtic Gold Coins*. Archaeometry 2009, 51. 598-625, DOI: DOI 10.1111/j.1475-4754.2008.00423.x;(c) S. A. Junk, E. Pernicka, An assessment of osmium isotope ratios as a new tool to determine the provenance of gold with platinum-group metal inclusions. Archaeometry 2003, 45. 313-331, DOI: 10.1111/14754754.00110. 3. (a) L. B. Brostoff, J. J. Gonzalez, P. Jett, R. E. Russo, Trace element fingerprinting of ancient Chinese gold with femtosecond laser ablation-inductively coupled mass spectrometry. Journal of Archaeological Science 2009, 36. 461-466, DOI: 10.1016/j.jas.2008.09.037;(b) G. Demortier, F. Bodart, Complementarity of PIXE and PIGE for the characterization of gold items of ancient jewelry. Journal of Radioanalytical Chemistry 1982, 69. 239-257, DOI: 10.1007/bf02515927;(c) L. Dussubieux, L. Van Zelst, LA-ICP-MS analysis of platinum-group elements and other elements of interest in ancient gold. Applied Physics a-Materials Science & Processing 2004, 79. 353-356, DOI: 10.1007/s00339-004-2532-2;(d) A. Gondonneau, M. F. Guerra, The circulation of precious metals in the Arab Empire: The case of the Near and the Middle East. Archaeometry 2002, 44. 573-599;(e) M. F. Guerra, Fingerprinting ancient gold with proton beams of different energies. Nuclear Instruments & Methods in Physics Research Section B-Beam Interactions with Materials and Atoms 2004, 226. 185-198, DOI: 10.1016/j.nimb.2004.02.019;(f) M. F. Guerra, T. Calligaro, J. C. Dran, C. Moulherat, J. Salomon, in Geoarcheological and Bioarchaeological Studies 3, ed. H. Kars, E. Burke. 2005, pp 343-346;(g) A. Hauptmann, S. Klein, Bronze Age gold in Southern Georgia. Archéosciences 2009. 75-82. 4. (a) B. Armbruster, H. Eilbracht, A. Grüger, M. Radtke, H. Riesemeier, I. Reiche, The Vikings in Berlin: SR-XRF analyses of the Hiddensee gold jewellery. Highlights / BESSY 2005, 2004. 32-33;(b) B. Constantinescu, E. Oberlaender-Tarnoveanu, R. Bugoi, V. Cojocaru, M. Radtke, The Sarmizegetusa bracelets. Antiquity 2010, 84. 1028-1042;(c) B. Constantinescu, A. Vasilescu, M. Radtke, U. Reinholz, Micro-SR-XRF studies for archaeological gold


Page 17 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


identification-the case of Carpathian gold and Romanian museal objects. Applied Physics aMaterials Science & Processing 2010, 99. 383-389, DOI: 10.1007/s00339-010-5624-1;(d) B. Constantinescu, A. Vasilescu, M. Radtke, U. Reinholz, A study on gold and copper provenance for Romanian prehistoric objects using micro-SR XRF. Journal of Analytical Atomic Spectrometry 2011, 26. 917-921, DOI: 10.1039/c0ja00215a;(e) W. Gorner, M. Eichelbaum, R. Matschat, K. Rademann, M. Radtke, U. Reinholz, H. Riesemeier, Non-destructive investigation of composition, chemical properties and structure of materials by synchrotron radiation. Insight 2006, 48. 540-544;(f) M. F. Guerra, M. Radtke, I. Reiche, H. Riesemeier, E. Strub, Analysis of trace elements in gold alloys by SR-XRF at high energy at the BAMline. Nuclear Instruments & Methods in Physics Research Section B-Beam Interactions with Materials and Atoms 2008, 266. 2334-2338, DOI: 10.1016/j.nimb.2008.03.008;(g) E. Pernicka, M. Radtke, H. Riesemeier, C. H. Wunderlich, European Network of Competence at 1600 BC. Highlights / BESSY 2004, 2003. 89;(h) M. Radtke, U. Reinholz, E. Strub, H. Born, Neues vom Berliner Goldhut: Materialanalysen mit Synchrotronstrahlung. Acta praehistorica et archaeologica 2009, 41. 29-35. 5. M. F. Guerra, T. Calligaro, M. Radtke, I. Reiche, H. Riesemeier, Fingerprinting ancient gold by measuring Pt with spatially resolved high energy Sy-XRF. Nuclear Instruments & Methods in Physics Research Section B-Beam Interactions with Materials and Atoms 2005, 240. 505-511, DOI: 10.1016/j.nimb.2005.06.150. 6. (a) K. Peeters, K. De Wael, A. Adniaens, G. Falkenberg, L. Vincze, The influence of Xray resonant Raman scattering effects on the detection of copper(II) tetrasulphonated phthalocyanine (CuTSPc) thin-films. deposited on gold electrodes. Spectrochimica Acta Part BAtomic Spectroscopy 2008, 63. 450-454, DOI: 10.1016/j.sab.2007.12.004;(b) D. Sokaras, M. Mueller, M. Kolbe, B. Beckhoff, C. Zarkadas, A. G. Karydas, Resonant Raman scattering of polarized and unpolarized x-ray radiation from Mg, Al, and Si. Physical Review A 2010, 81. DOI: 012703 10.1103/PhysRevA.81.012703. 7. R. A. Ixer, in Metals in Antiquity, ed. S. M. Young, M. Pollard, P. Budd, R. A. Ixer. British Archaeological Reports: Oxford, 1999, vol. 792, pp 43-52. 8. R. M. Hough, C. R. M. Butt, J. Fischer-Buehner, The Crystallography, Metallography and Composition of Gold. Elements 2009, 5. 297-302, DOI: 10.2113/gselements.5.5.297. 9. M. B. McClenaghan, in Indicator mineral methods in mineral exploration, ed. B. McClenaghan, V. Peuraniemi, M. Lehtonen. Vuorimiesyhdistys: Rovaniemi Finland, 2009, vol. B92-4,1-6. 10. (a) R. K. Dube, Interrelation between gold and tin: A historical perspective. Gold Bulletin 2006, 39. 103-113;(b) P. T. Craddock, M. R. Cowell, M. F. Guerra, in Anatolian Metal III. Der Anschnitt, Beiheft 18, ed. Ü. Yalcin. Deutsches Bergbau-Museum: Bochum, 2005, pp 67-77;(c) C. J. Raub, in Prehistoric gold in Europe, ed. G. Morteani, J. P. Northover. Kluwer Academic Publishers: The Netherlands, 1995, pp 243-259. 11. J. P. Desroches, G. Forum, Mongolie: le premier empire des steppes. Actes sud: 2003. 12. D. C. Waugh, Nomads and Settlement: New Perspectives in the Archaeology of Mongolia. The Silk Road 2010, 8. 97-124. 13. N. Di Cosmo, Ancient Inner Asian Nomads: Their Economic Basis and Its Significance in Chinese History. The Journal of Asian Studies 1994, 53. 1092-1126. 14. S. S. Minyaev, Archéologie des Xiongnu en Russie : Nouvelles découvertes et quelques problèmes. Arts asiatiques 1996, 51. 5-12.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 27

15. Z. Batsaikhan, Foreign Tribes in the Xiongnu Confederation. The Silk Road 2006, 4. 4547. 16. T. J. Barfield, Hsiung-nu Imperial Confederacy: Organization and Foreign Policy. Journal of Asian Studies 1981, 41. 45-61, DOI: 10.2307/2055601. 17. B. Watson, Records of the grand historian by Sima Qiam (Translation). Columbia University Press: New York, 1993. 18. N. Di Cosmo, The Origins of the Great Wall. The Silk Road 2006, 4. 14-19. 19. A. Sheng, Why Ancient Silk Is Still Gold: Issues in Chinese Textile History. Ars Orientalis 1999, 29. 147-168. 20. E. C. Bunker, Gold in the Ancient Chinese World - a Cultural Puzzle. Artibus Asiae 1993, 53. 27-50. 21. U. M. Franklin, in The Origins of Chinese civilization, ed. D. N. Keightley. University of California Press, 1983, pp 279-296. 22. N. Barnard, in The Origins of Chinese civilization, ed. D. N. Keightley. University of California Press, 1983, pp 237-278. 23. H. Riesemeier, K. Ecker, W. Gorner, B. R. Muller, M. Radtke, M. Krumrey, Layout and first XRF applications of the BAMline at BESSY II. X-Ray Spectrometry 2005, 34. 160-163. 24. M. Radtke, L. Vincze, W. Görner, Quantification of energy dispersive SRXRF for the certification of reference materials at BAMline. Journal of Analytical Atomic Spectrometry 2010, 25. 631-634, DOI: 10.1039/b926596a. 25. M. F. Guerra, T. Calligaro, Gold traces to trace gold. Journal of Archaeological Science 2004, 31. 1199-1208, DOI: 10.1016/j.jas.2002.05.001. 26. G. Andre, J. P. Desroches, Une tombe princière Xiongnu à Gol Mod, Mongolie (campagnes de fouilles 2000-2001). Arts asiatiques 2002, 57. 194-205. 27. J. Harmatta, B. N. Puri, G. F. Etemadi, ed. UNESCO. 1994, p 572. 28. W. P. Yetts, Discoveries of the Kozlóv Expedition. The Burlington Magazine for Connoisseurs 1926, 48. 168-185. 29. S. Murao, K. Naito, G. Dejidmaa, S. H. Sie, Mercury content in electrum from artisanal mining site of Mongolia. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 2006, 249. 556-560. 30. G. Dejidmaa, in Project on Mineral Resources, Metallogenesis, and Tectonics of Northeast Asia ed. Nokleberg, Naumova, Kuzmin, Bounaeva. Asia Institute of Geology and Mineral Resources, Open-File Report, Denver: U.S. Geological Survey edn., 1999, pp 99-165. 31. W. P. Robinson, G. B. Anosova Mining and mineral development management policy in the Selenga River watershed; United States Geological Survey, Institute of General and Experimental Biology, Russian Academy of Sciences Mongolian Academy of Sciences: 2004. 32. D. R. Wilburn, D. I. Bleiwas, Platinum-group metals--world supply and demand. USGS Open-File Report 2005, 2004-1224. 33. Y. W. Wang, J. B. Wang, L. J. Wang, Y. Z. Chen, Tin mineralization in the Dajing tinpolymetallic deposit, inner Mongolia, China. Journal of Asian Earth Sciences 2006, 28. 320-331, DOI: 10.1016/j.seaes.2004.06.008. 34. B. S. Karpoff, W. E. Roscoe Report on placer gold properties in the Tuur valley, Zaamar goldfield Mongolia; RPA Inc.: Toronto, Canada, 2005. 35. H. G. Dill, Geogene and anthropogenic controls on the mineralogy and geochemistry of modern alluvial-(fluvial) gold placer deposits in man-made landscapes in France, Switzerland


Page 19 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and Germany. Journal of Geochemical Exploration 2008, 99. 29-60, DOI: 10.1016/j.gexplo.2008.02.002. 36. R. Routledge Technical report on the Gatsuurt gold deposit, mineral resource estimate in Northern Mongolia, NI 43-101; RPA Inc.: Toronto, Canada, 2005. 37. Savkino gold project, Chita region, Russian Federation Mineral resources and Reserves, Technical report NI-43-101; Micon International Co Limited: Toronto, Canada, 2010. 38. B. K. Miller, J. Bayarsaikhan, P. B. Konovalov, T. Egiimaa, J. Logan, M. Machicek, Xiongnu Constituents of the High Mountains: Results of the Mongol-American Khovd Archaeology Project. The Silk Road 2009, 7. 8-20. 39. B. K. Miller, J. Bayarsaikhan, T. Egiimaa, C. Lee, Xiongnu Elite Tomb Complexes in the Mongolian Altai. The Silk Road 2008, 5. 27-36. 40. J. Aichler, J. Malec, J. Vecera, P. Hanzl, D. Burianek, T. Sidorinova, Z. Taborsky, K. Bolormaa, D. Byambasuren, Prospection for gold and new occurrences of gold-bearing mineralization in the eastern Mongolian Altay. Journal of Geosciences 2008, 53. 123-138, DOI: 10.3190/jgeosci.025. 41. (a) Z. S. Nikiforova, B. B. Gerasimov, E. G. Tulaeva, Genesis of Gold-Bearing Placers and Their Possible Sources (Eastern Siberian Platform). Lithology and Mineral Resources 2011, 46. 17-29, DOI: 10.1134/s0024490210041017;(b) B. L. Wood, N. P. Popov, The giant Sukhoi Log gold deposit, Siberia. Russian Geology and Geophysics 2006, 47. 317-341. 42. J.-S. Park, E. Gelegdorj, Y.-E. Chimiddorj, Technological traditions inferred from iron artefacts of the Xiongnu Empire in Mongolia. Journal of Archaeological Science 2010, 37. 26892697, DOI: 10.1016/j.jas.2010.06.002. 43. N. O. Kozhevnikov, A. V. Kharinsky, O. K. Kozhevnikov, An accidental geophysical discovery of an Iron Age archaeological site on the western shore of Lake Baikal. Journal of Applied Geophysics 2001, 47. 107-122, DOI: 10.1016/s0926-9851(01)00051-9. 44. J.-S. Park, W. Honeychurch, A. Chunag, Ancient bronze technology and nomadic communities of the Middle Gobi Desert, Mongolia. Journal of Archaeological Science 2011, 38. 805-817, DOI: 10.1016/j.jas.2010.11.003. 45. L. Vincze, K. Janssens, F. Adams, K. W. Jones, A general monte carlo simulation of energy-dispersive X-ray fluorescence spectrometers .3. Polarized polychromatic radiation, homogeneous samples. Spectrochimica Acta Part B-Atomic Spectroscopy 1995, 50. 1481-1500.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. Map of the area where Gol Mod, the EgiinGol necropolis, the tombs of the Takhiltin cemetery and the Shombuuziin burial are located. These latter two sites are situated in the greater Gobi-Altai region of western Mongolia on the northwestern region of the Silk Roads.


Page 20 of 27

Page 21 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. The optimal excitation energy was determined by comparison of measurements at thick pure element foils. The relative minimum detection limit was determined by the quotient of the square root of the Au and the Pt signal for different energies between Au-L3 and Pt-L3 edge.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

b) a)

c)

d)

Figure 3.Four examples of the fit (solid black line) of measured spectra (asterisk) in the Pt–Lα region with pure element spectra of Pt, Au, Ag, Cu, Zn and Ir. The upright solid line marks the position of the Pt-Lα line. The region from channel 8 to channel 28 was taken into account for the determination of peak and background content. There was no significant contribution of Ag and Zn. a) The spectra of RM 8058, used as standard for the other measurements. The spectra can be modeled with Au (light blue) and Pt (dark blue). The diamond is marking the part of the fit which is assigned to Pt b) To model the ANG sample spectra additional Cu (yellow) must be considered. c) The spectra of the A66 sample couldn’t be modeled satisfactorily with the measured pure elements. d) A66 can


Page 22 of 27

Page 23 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


be modeled by taken into account Ir (light green). Because Ir wasn’t available as pure element, the spectrum has been simulated by the program msim5d45. With this model an excellent agreement between Pt (dark blue line) and the part of the fit assigned to Pt (diamonds) is achieved.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. Representation of the Cu and Ag contents for gold foils from tombs 1, 43, 45 and 79. The foils of tomb 79 are made of one single alloy containing about 10% Ag and 0.2% Cu while the foils from tomb 1 have Ag contents ranging from 1% to 6%, Cu being present at contents lower than 0.4 %. The foils from satellite tombs 43 and 45 show a larger variability of compositions, but contained in the range defined by tombs 1 and 79.


Page 24 of 27

Page 25 of 27

1

0.8 Tomb 1 Tomb 79 Tomb 43 & 45

0.6

Shombuuziin

Cu %

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Takhiltin Egiin Gol

0.4

Rivets jade Han plaques

0.2

0 0

5

10

15

20

25

Ag %

Figure 5. Comparison of the Cu and Ag contents for the gold foils from Gol Mod and the gold fragments from the XiongnuEgiinGol necropolis, the Takhiltin cemetery and the Shombuuziin burial. These latter two sites are situated in the greater Gobi-Altai region of western Mongolia on the northwestern region of the Silk Roads. The rivet jades from tomb 1 Chinese xi pendant and four Han gold plaques were added to the diagram and designated “Chinese” gold.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 27

Figure 6.The ratios Pt/Au and Sn/Au for foils from tombs 1,43,45 and 79. Sn and Pt are “markers” for the use of alluvial gold. Tombs 1 and 79 form separated groups where tomb 43 is contained. Only tomb 45 shows a very high variability of Pt.


Page 27 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For TOC only


Beyond the Great Wall: Gold of the Silk Roads and the First Empire of

Recommend Documents