Final Report Summary - ISOGLASS (Archaeometric investigation of ancient glass - identification of the sources of Cu, Sb and Ti via isotopic analysis using (laser ablation) multi-collector ICP ...)
Nowadays, isotopic analysis is used for provenancing studies in many different fields. Archaeologists attempt to gather information on the artefacts recovered during excavations, allowing them to reconstruct, e.g. mining activities, manufacturing technologies and/or trade routes that existed in the past. Whereas studies of the provenance (geographical origin) and trade of stone and ceramics are well advanced, this is not the case for ancient glass. Moreover, several questions concerning the nature of the raw materials used and the geographical location of their transformation into finished artefacts remain unanswered.
Production sites and trade routes of Roman glass have received much attention over the past decade and glass samples have been studied from the perspective of the provenance of the sand and flux used in their manufacture. For instance, Sr and Nd isotopic analysis have been used to trace the origin of the stabiliser and source of sand, respectively and hence, to identify the exact raw materials used in glass production. Examples of provenance determination efforts using oxygen isotopic analysis can also be found in the literature. However, the provenance of Roman glass has only very rarely been investigated from the perspective of the origin of the coloriser / decolorising component.
It is assumed that raw glass was produced in primary workshops near the raw material sources, to be transported in the form of chunks or ingots to secondary glass houses, where the glass was shaped into specific objects. This suggests that the composition of formed glasses closely resembles primary centres and no distinctive composition can be attributed to the secondary working centres. In this context, isotopic analysis can help to identify the origin of the Cu (coloriser), tin (Sb) or manganese (Mn) (decolorisers) used in the manufacturing of Roman glass and, consequently, to reconstruct how such material was traded and transported, and how this can be integrated in the network of primary and secondary glass producers.
The research project has focused in two main objectives:
1. development of analytical methods for measuring the isotopic composition of Sb (decoloriser / opacifier) in antimony ores and ancient glasses.
2. Development of an analytical method for measuring the isotopic composition of Cu (coloriser) in ancient glass.
Prior to isotopic analysis using multi-collector ICP - mass spectrometry, digestion of the sample and isolation of the target analyte need to be addressed. Two digestion procedures have been optimised, one for antimony ores and one for glass and both of them consist of a two-step digestion. In the case of antimony ores, the first stage is carried out with a mixture of HF-HNO3-H2O2 at 40 ºC to get rid of the organic material, followed by treament with aqua regia at 90 ºC during approximately 10 hours. For glass, digestion is first performed with a mixture of HNO3-HF and subsequently with aqua regia. Both steps take place at 110 ºC during 2 days. Once the digestion is finished and the aqua regia evaporated at 90 ºC, the final residue is redissolved in 10 ml of 0.14 M HF. Validation of the methods developed was accomplished with an antimony ore reference material (CAN CD-1, Brammer; Standard Company, Inc., US) with 3.57 wt% Sb and NIST SRM 610 glass (US) with 396 µg/g of Sb. For copper (Cu), the digestion procedure was validated relying on NIST SRM 610 glass reference material (444 µg/g Cu).
Antimony isotopic analysis
To develop an isolation protocol for Sb, both CAN CD-1 ore reference material and NIST SRM 610 glass were used. Cation exchange chromatography using Dowex AG50-X8 resin in a 10 ml polypropylene column (Biorad) was employed first. Although this separation has been previously used, the amount of sample to be loaded and the amount of resin employed were optimised for obtaining quantitative Sb recovery (101 ± 2 %, N = 3 and 98 ± 1 %, N = 4 for the ore and the glass, respectively). In a second step, Sb can be isolated via weak anion-exchange chromatography using 16 - 50 mesh Amberlite IRA743 resin (Sigma-Aldrich, Belgium). Also in this second step, quantitative recovery of Sb was obtained for the ore (95 %) and for the glass (97 ± 6 %, N = 6) reference materials. The isolation procedure developed here was proven to be capable of separating Sb from 69 elements commonly present in glass materials.
After Sb isolation, isotopic analysis was performed using MC-ICP-MS with indium (In) as an internal standard relied on for correction for mass discrimination. In a first step, isotopic analysis of antimony ores of different origin was performed. Different mass bias correction methods have been assessed (e.g. standard bracketing, correction via the exponential law and via the revised Russell's law). Highest precision was obtained when using the revised Russell's law, such that this approach was employed further for data treatment. Results showed that the Sb isotope ratio in antimony ores varies by up to 10 units (0.1 %). It was also been observed that the isotopic composition of Sb and its spread in ores depend on the origin of the sample (e.g while Spanish and US ores exhibit Sb isotope ratios that differ from +1 to +4 & #949 units when compared to the in-house reference, the Romanian ores show variation between +3 and + 6 & #949 units with respect to the in-house reference).
Apart from the ores, also ancient glasses have been analysed. Earlier (Roman) glasses seem to show more variation in the Sb isotope ratio and the results indicate the use of at least two different sources of Sb. For opaque and colourless glasses from the same location and era, the same Sb source was used. Intermediate values (between the two hypothesised sources) could either be attributed to yet another source of Sb or to the recycling of glass, an extended practice in the Roman Empire.
Copper isotopic analysis
Some isolation methods can be found in the literature (1, 2) and an evaluation of their utility for glass has been first evaluated. The method proposed by Van Heghe et al. (3) (modified from that first described by Maréchal et al.) was finally selected, as the method proposed by Larner et al. (2) initially led to lower Cu recoveries compared to the other method. Despite the quantitative recovery of Cu, the fraction was not entirely pure (Co was co-eluting with Cu). As a result, further optimisation of the isolation procedure was attempted. By varying the acid concentration of the mobile phase used to elute the Cu, it was possible to elute Cu and Co separately.Validation of the isolation procedure was performed using two in-house standards and two glass reference materials (NIST SRM 610 and NIST SRM 612). All Cu recoveries were quantitative.
Similarly to Sb, first Cu isotope ratio results for in Roman glasses have proved not to exhibit a wide variation, which supports the hypothesis of the use of only few sources of raw materials for glass production in antiquity.
Although the project focuses on ancient glass (an important object of art in Europe's culture throughout many centuries), the methods developed could be also applicable in other applications: archaeometrical investigation of other objects of art / types of material, investigation of animal and human migration behaviour and studying the isotopic variation in these elements contained in 'chronological archives', such as ice cores, sediment layers or corals, for using the corresponding isotope ratios as palaeoproxies, revealing variation in the prevailing conditions (temperature, pH, redox potential, etc) over time.
1. C.L. Maréchal, P. Télouk, F. Albarède, Chem. Geol. 1999, 156, 251 - 273.
2. F. Larner, M. Rehkamper, B.J. Coles, K. Kreissig, D.J. Weiss, B. Sampson, C. Unsworth, S. Strekopytov, J. Anal. At. Spectrom., 2011, 26, 1627 - 1632.
3. L. Van Heghe, E. Engström, I. Rodushkin, C. Cloquet, F. Vanhaecke, J. Anal. At. Spectrom., 2012, 27, 1327 - 1334.
Production sites and trade routes of Roman glass have received much attention over the past decade and glass samples have been studied from the perspective of the provenance of the sand and flux used in their manufacture. For instance, Sr and Nd isotopic analysis have been used to trace the origin of the stabiliser and source of sand, respectively and hence, to identify the exact raw materials used in glass production. Examples of provenance determination efforts using oxygen isotopic analysis can also be found in the literature. However, the provenance of Roman glass has only very rarely been investigated from the perspective of the origin of the coloriser / decolorising component.
It is assumed that raw glass was produced in primary workshops near the raw material sources, to be transported in the form of chunks or ingots to secondary glass houses, where the glass was shaped into specific objects. This suggests that the composition of formed glasses closely resembles primary centres and no distinctive composition can be attributed to the secondary working centres. In this context, isotopic analysis can help to identify the origin of the Cu (coloriser), tin (Sb) or manganese (Mn) (decolorisers) used in the manufacturing of Roman glass and, consequently, to reconstruct how such material was traded and transported, and how this can be integrated in the network of primary and secondary glass producers.
The research project has focused in two main objectives:
1. development of analytical methods for measuring the isotopic composition of Sb (decoloriser / opacifier) in antimony ores and ancient glasses.
2. Development of an analytical method for measuring the isotopic composition of Cu (coloriser) in ancient glass.
Prior to isotopic analysis using multi-collector ICP - mass spectrometry, digestion of the sample and isolation of the target analyte need to be addressed. Two digestion procedures have been optimised, one for antimony ores and one for glass and both of them consist of a two-step digestion. In the case of antimony ores, the first stage is carried out with a mixture of HF-HNO3-H2O2 at 40 ºC to get rid of the organic material, followed by treament with aqua regia at 90 ºC during approximately 10 hours. For glass, digestion is first performed with a mixture of HNO3-HF and subsequently with aqua regia. Both steps take place at 110 ºC during 2 days. Once the digestion is finished and the aqua regia evaporated at 90 ºC, the final residue is redissolved in 10 ml of 0.14 M HF. Validation of the methods developed was accomplished with an antimony ore reference material (CAN CD-1, Brammer; Standard Company, Inc., US) with 3.57 wt% Sb and NIST SRM 610 glass (US) with 396 µg/g of Sb. For copper (Cu), the digestion procedure was validated relying on NIST SRM 610 glass reference material (444 µg/g Cu).
Antimony isotopic analysis
To develop an isolation protocol for Sb, both CAN CD-1 ore reference material and NIST SRM 610 glass were used. Cation exchange chromatography using Dowex AG50-X8 resin in a 10 ml polypropylene column (Biorad) was employed first. Although this separation has been previously used, the amount of sample to be loaded and the amount of resin employed were optimised for obtaining quantitative Sb recovery (101 ± 2 %, N = 3 and 98 ± 1 %, N = 4 for the ore and the glass, respectively). In a second step, Sb can be isolated via weak anion-exchange chromatography using 16 - 50 mesh Amberlite IRA743 resin (Sigma-Aldrich, Belgium). Also in this second step, quantitative recovery of Sb was obtained for the ore (95 %) and for the glass (97 ± 6 %, N = 6) reference materials. The isolation procedure developed here was proven to be capable of separating Sb from 69 elements commonly present in glass materials.
After Sb isolation, isotopic analysis was performed using MC-ICP-MS with indium (In) as an internal standard relied on for correction for mass discrimination. In a first step, isotopic analysis of antimony ores of different origin was performed. Different mass bias correction methods have been assessed (e.g. standard bracketing, correction via the exponential law and via the revised Russell's law). Highest precision was obtained when using the revised Russell's law, such that this approach was employed further for data treatment. Results showed that the Sb isotope ratio in antimony ores varies by up to 10 units (0.1 %). It was also been observed that the isotopic composition of Sb and its spread in ores depend on the origin of the sample (e.g while Spanish and US ores exhibit Sb isotope ratios that differ from +1 to +4 & #949 units when compared to the in-house reference, the Romanian ores show variation between +3 and + 6 & #949 units with respect to the in-house reference).
Apart from the ores, also ancient glasses have been analysed. Earlier (Roman) glasses seem to show more variation in the Sb isotope ratio and the results indicate the use of at least two different sources of Sb. For opaque and colourless glasses from the same location and era, the same Sb source was used. Intermediate values (between the two hypothesised sources) could either be attributed to yet another source of Sb or to the recycling of glass, an extended practice in the Roman Empire.
Copper isotopic analysis
Some isolation methods can be found in the literature (1, 2) and an evaluation of their utility for glass has been first evaluated. The method proposed by Van Heghe et al. (3) (modified from that first described by Maréchal et al.) was finally selected, as the method proposed by Larner et al. (2) initially led to lower Cu recoveries compared to the other method. Despite the quantitative recovery of Cu, the fraction was not entirely pure (Co was co-eluting with Cu). As a result, further optimisation of the isolation procedure was attempted. By varying the acid concentration of the mobile phase used to elute the Cu, it was possible to elute Cu and Co separately.Validation of the isolation procedure was performed using two in-house standards and two glass reference materials (NIST SRM 610 and NIST SRM 612). All Cu recoveries were quantitative.
Similarly to Sb, first Cu isotope ratio results for in Roman glasses have proved not to exhibit a wide variation, which supports the hypothesis of the use of only few sources of raw materials for glass production in antiquity.
Although the project focuses on ancient glass (an important object of art in Europe's culture throughout many centuries), the methods developed could be also applicable in other applications: archaeometrical investigation of other objects of art / types of material, investigation of animal and human migration behaviour and studying the isotopic variation in these elements contained in 'chronological archives', such as ice cores, sediment layers or corals, for using the corresponding isotope ratios as palaeoproxies, revealing variation in the prevailing conditions (temperature, pH, redox potential, etc) over time.
1. C.L. Maréchal, P. Télouk, F. Albarède, Chem. Geol. 1999, 156, 251 - 273.
2. F. Larner, M. Rehkamper, B.J. Coles, K. Kreissig, D.J. Weiss, B. Sampson, C. Unsworth, S. Strekopytov, J. Anal. At. Spectrom., 2011, 26, 1627 - 1632.
3. L. Van Heghe, E. Engström, I. Rodushkin, C. Cloquet, F. Vanhaecke, J. Anal. At. Spectrom., 2012, 27, 1327 - 1334.