Analytical Report on a Selection of Finger-Rings from Late Iron Age Context in Bohemia

STUDIA HERCYNIA XX/1, 68–82 Analytical report on a Selection of finger‑rings from late Iron Age Context in Bohemia Romana Kozáková ABStrACt The paper presents results of technological analyses executed on selected finger ‑rings prevalently from the Stradonice oppidum. Analysed were the elementary composition of the metal parts of the finger rings; the technology employed for the manufacturing of the metal parts and for the setting of the gems in the bezels; as well as the material of the inlays. KEyWOrDS Finger ring; Roman; gem; XRF; SEM/EDS; bronze alloy; brass; amber. MAtErIAl AND OBJECtIVES The study included 28 metal finger ‑rings, the majority from the site of Stradonice and now‑ adays kept in the collection of the National Museum in Prague (Tab.1). The work focused on ascertaining the state of preservation of the single pieces and on the determination of the elementary composition of both their metallic parts and the inlays with the objective of defining the production technology and making more precise some points concerning the provenance of the rings. Moreover, imprints of the images depicted on the gems have been made in order to facilitate their iconographic and stylistic analysis. The rings featured various degrees of deterioration. All of them were affected by corrosion ranging from superficial to structurally destructive and concerning both the metallic parts and the inlays. Some of the rings were broken; some of these were reconsolidated in the past with excessive quantities of cement which disfigured their original shape. The inlays of many rings are not preserved. We took advantage of the fragmentary state of the rings using the points of fracture for measuring the elementary composition of the metal alloys not only on the surface but also in the objects’ cores. Thanks to the absence of several inlays we could study the technology of their attachment to the bezel and in some cases the materials placed beneath them. In what follows we refer to the single pieces by the catalogue numbers used in the archaeological part of the study (Kysela 2016). tHE ANAlytICAl MEtHODS XRF SPECTROMETRY This method is based on measuring the characteristic secondary X ‑ray radiation emitted by the sample. It enables the ascertainment of the elements present in the sample and also– by means of intensity analysis– of their mutual quantitative proportion (Novotná– Karhan– Pechová 2001; Hložeketal. 2005). 69ROMANA KOZÁKOVÁ Cat. n° inv. n° (NM‑H1) metal inlay analytic method K1 – Ag /Cu/Sn Glass XRF, SEM/ EDX S1 80264 Brass – XRF S2 80265 Brass – XRF S8 81575 Brass – XRF S9 81576 Brass Glass XRF S10 81577 Brass Glass XRF, SEM/ EDX S11 81578 Brass Glass XRF S12 81580 Brass Glass XRF S13 81581 Brass – XRF, SEM/ EDX S14 81582 Brass Glass XRF S15 81583 Brass – XRF S16 81599 Brass Stone XRF S17 81601 Fe – XRF S18 81602 Fe Stone/ Glass? XRF Cat. n° inv. n° (NM‑H1) metal inlay analytic method S19 81603 Fe Glass XRF S20 81604 Fe Glass XRF S21 81605 Fe – XRF S22 81606 Fe Amber XRF, IR S23 81607 Fe Glass XRF S24 81608 Fe – XRF S25 81609 Fe Amber? XRF, IR S26 105608 Brass Glass XRF, SEM/ EDX S27 105628 Brass? Glass XRF, SEM/ EDX S?2 26422 Fe Stone XRF Sx1 81579 Mixed – XRF Sx2 81600 Gilded Ag Glass XRF Sx3 105609 Bronze – XRF Sx4 212676 Mixed Amber XRF Tab.1: The analysed pieces and the analyses executed on them. The elementary composition of the metallic parts of the finger ‑rings was analysed primarily by aportable XRF spectrometer Delta Professional (produced by BAS Rudice), in the mode “analytic plus” in the case of the metallic parts and in the mode “geochem” in the case of the inlays. The measurements were taken with the use of a3 mm ray collimator in two subse‑ quent steps: in the case of the metallic parts 20 kV/20 sec > 40 kV/20 sec, and in the case of inlays, requiring alonger measuring time due to the presence of light elements, 20 kV/30 sec > 20 kV/ 90 sec the measurements were taken on the objects’ surfaces and the results may therefore be biased due to locally present corrosive products. For this reason, the quantitative measurements were taken only after cleaning the objects’ surfaces and the presented results are only average values of several measurements. In the case of the non ‑metallic parts, the presented results must be considered only roughly indicative due to the high proportion of light elements which are difficult to measure by portable analysers. 70 STUDIA HERCYNIA XX/1 SCANNINg ELECTRON MICROSCOPY WITH ENERgY ‑DISPERSIVE DETECTOR (SEM/EDS) Selected rings (cat. nos. K1, S8, S11, S17, S20, S26, S27) were further analysed by electron mi‑ croscopy with elementary composition measured by means of an EDX detector. The electron microscope JEOL JSM‑6490LV used for our analysis was fitted with the EDX detector Oxford Instruments INCA X ‑ACT. The accelerating potential was set at 15 kV. The measurements were taken on miniature samples. Sampling the artefacts was indispensable for the successful implementation of the analysis which requires the application of an irremovable conductive layer, hardly imaginable on the fragile and complexly shaped rings. The samples were placed in the device in such away as to allow the study of both the uncorroded core and of the surface. The measurements were made in cooperation with I. Macek in the Department of Min‑ eralogy and Petrography of the Natural History Museum, the National Museum in Prague. IR SPECTROMETRY The inlays of two of the studied rings (S22 and S25) seemed– according to macroscopic observa‑ tion– to be made of amber. Miniature samples were taken from crevices in the inlays in order to be analysed by means of IR spectrometry, which would not only permit one to determine if the material is actually amber, but also to ascertain its probable origin with more precision. As is well known, the geographic origin of amber can be determined by means of molecular spectroscopy. For example Baltic amber can be distinguished in the IR spectra by its specific spectral characteristic, the so ‑called “Baltic shoulder” due to the presence of the succinic acid and its esters (Giuliano 2007; Shashouna 2002). The samples were analysed by means of IR ‑spectroscopy by ATR in the FTIR spectrometer Nicolet 6700 (Thermo ‑Nicolet, USA) with the following measurement parameters: spectral range 4000–650 cm‑1, resolution 4 cm‑1, number of spectral accumulations 128, apodisation Happ ‑Genzel. The resulting IR spectrum was processed by the Omnic program and interpreted using the new methodical report for archaeological ambers (Tisucká– Ohlídalová 2013). The measurements were taken by M. Ohlídalová, National Museum in Prague. gEM IMPRINTS AND OPTICAL MICROSCOPY The majority of the preserved inlays bears anegative image. In order to facilitate their stylistic and iconographic analysis (cf. Kysela 2016) their positive casts were executed. The positive relief allows one to make details out more clearly and the white colour of the imprints su‑ presses the visually disturbing chromatic effects, reflections and superficial corrosion of the originals. The imprints were taken from cleaned inlays before the cleaning of the metallic parts and after separation by alayer of cyclododecane which prevented the casting substance from leaking deep into the body of the inlay. As the casting substance we used the synthetic rubber Lukopren N1522 (produced by Lučební závody,a.s. Kolín). The gems and their casts were studied and photographed through the stereomicroscope Olympus. The material of the inlays was determined by P.Burdová in the Department of Mineralogy and Petrography of the Natural History Museum, the National Museum in Prague by means of polarized gemologic microscope GIAGEM Instruments with Leica optics. 71ROMANA KOZÁKOVÁ ANAlySES Of tHE INDIVIDuAl PIECES K1 The inlay is made of glass. The metallic part is made of an alloy of silver (ca. 63%), tin (ca. 31%) and copper (ca. 5%). The surface XRF measurements have detected relatively high propor‑ tions of gold and also traces of iron, mercury and lead. None of these elements was, however, confirmed by the EDX analysis. Most probably, the finger ‑ring was originally gold ‑plated; the EDX measurements were probably taken in spots, where the surface deterioration removed all the remains of the gilding. The blackened surface is caused by oxidation as indicated by higher proportions of oxide. Given the absence of sulphur in the surface layer, the darkening was not caused by the typical reaction with sulphuric pollutants but with oxide and humidity reacting with copper probably producing the corrosive interlayer too (Fig. 1). Fig. 1: K1– the SEM photograph of the analysed sample. The arrows indicate the different points where measurements were taken (photo by the author). S8 The gem of this ring is not preserved. The gem setting, nevertheless, features avery well preserved underlying layer meant to enhance though its reflexion the luminosity of the gem itself (Pl. 4/1:S6a). This ground layer, into which the gem had been set, was preserved thanks to conservation through varnishing shortly after the ring was unearthed. Unfortunately, due to its general fragility, this intervention could not be removed and the exact nature of the ground layer could not be analytically investigated. The ground matter was examined under optic and electron microscope (Fig. 2) and can be determined as atwo ‑layer metal leaf with‑ out an organic support. The ground layer is made of silver, probably of partially melted silver filings covered by alayer of gilding (Pl. 4/1:S6b, Fig. 2 right). Apart from these two metals, sulphur and carbon have been detected, most probably as the result of melting. The metallic part of the ring is made of abrass alloy consisting of 90% copper and 10% zinc (excluding the elements resulting from surface corrosion, such as Al and Si). No surface treatment of the metal has been detected. 72 STUDIA HERCYNIA XX/1 Fig. 2: S8– Left: the SEM photograph of the analysed sample. The arrows indicate the different points where measurements were taken. Right: detail of the insufficiently melted silver filings (photo by the author). S11 The glass inlay is preserved firmly set in the bezel though in comparison with other rings it protrudes markedly over the level of the relatively low bezel frame (whose present height corresponds to the original shape of the object and is not due to loss of material caused by corrosion). The gem’scircumference was clearly cut in order to fit the narrow frame (Fig. 3c). The XRF surface measurements determined the alloy as brass with (apart from Cu and Zn) the presence of Fe, P and Pb, and traces of Ag and Hg. The SEM/EDX analyses found that the metallic core was plated with asurface layer on another alloy (Fig. 3a–b). The core is made of Fig. 3: S11– a) The SEM photograph of the analysed sample. The arrows indicate the different points where mea‑ surements were taken; b) The SEM photograph of the analysed sample– detail of the metal layer stratigraphy; c) Insertion of the gem into the bezel and the flaking off surface layer (a–b: photo by the author, c: photo by A. Kumstátová). 73ROMANA KOZÁKOVÁ an alloy of Cu, Zn and asmall amount of Sn (in the ratio 12:5:1) and the measured points also produced afair representation of the elements resulting from corrosion, such as Si, P, Ca and S. Silver was not detected by this method. The surface layer (also heavily corroded) is made of adifferent brass without the presence of tin, unlike the core with the presence of Sn and ahigher amount of Cl ‑ ions (3.15 wt.%). In the interlayer between the core and the surface layer elements related to corrosion damage were detected. The corrosion products of copper expand and lead to the flaking of the surface layers. Apart from these features, the surface consists only of oxidised copper, which is caused by its higher reactivity and efflorescence formation. S17 Thanks to the absence of agem, its setting in the bezel of the iron ring could be studied in detail. In the central part of the bezel, heavy corrosion is macroscopically observable, including soft corrosive products of iron and organic remains, perhaps of an adhesive by means of which the gem was glued inside the bezel (Pl. 4/1:S17a–b). The substance was not identified. Around the bezel’scircumference were detected the remains of ashiny greyish metallic layer, partly flaking off, originally lining the entire interior of the cavity (Pl. 4/1:S17a, c). The layer was identified as silver by means of the SEM/EDX analysis (Fig. 4). The other identified elements (Fe, Si, Ca and P) result from corrosive processes. Sulphur, present in the form of tiny needles, was probably employed as an admixture in the silver lining while the higher proportions of carbon in the smooth parts probably resulted from melting. The body of the ring is made of iron without admixtures and without any identifiable surface treatment. Fig. 4: S17 The SEM photograph of the analysed sample (left) and adetail thereof. The arrows indicate the different points where measurements were taken (photo by the author). S20 The gem of this ring is preserved firmly attached inside its setting. The entire ring including the gem is heavily corroded. Both the XRF measurement on the gem’ssurface and the character of its corrosion (iridising layers, secondary precipitates) prove its soda nature. Considering the condition of the artefact, there was no point in executing any further analyses. The body of the ring is made of iron and the SEM/EDX analysis (Fig. 5) excluded the pres‑ ence of asurface layer. No chloride ions were present. The white precipitates were determined as calcareous. 74 STUDIA HERCYNIA XX/1 Fig. 5: S20– The SEM photograph of the analysed sample and adetail thereof. The arrows indicate the different points where measurements were taken (photo by the author). S22 Only the amber inlay was analysed in this ring. Fig. 6 right presents the ATR ‑IR spec‑ tra of the amber sample with indicated peaks of the absorptions relevant for the study of the material’sorigin. Fig. 6 left presents adetail of the spectral area 750–1950 cm‑1 of the analysed spectrum in comparison with the standards of Baltic amber (Ukraine, 24384) and of rumanite (Ploesti, 24372). Ahint of the Baltic shoulder is present in the analysed spectrum at the point of 1203 cm‑1, but unequivocal identification of the amber as Baltic is hindered by the absence of absorption at 885 cm‑1. For this reason, the origin of the studied amber cannot be determined unambiguously. Fig. 6: ATR ‑IR spectrum of the analysed amber sample (left) and the ATR ‑IR spectrum of the analy‑ sed sample in comparison with the standards of Baltic amber (Ukraine, 24384) and of rumanite (Ploesti, 24372) in the spectral zone of 750–1950 cm‑1 (right). S25 Only the amber inlay was analysed in this ring. Fig. 7 left presents the ATR ‑IR spectrum in comparison with the spectra of bentonite and of Baltic amber (Ukraine, 24384). Fig. 7 right presents adetail of the spectral area 750–1950 cm‑1 of the analysed spectrum in comparison with the standards of Baltic amber (Ukraine, 24384) and of rumanite (Ploesti, 24372). The analysed spectrum features significant absorptions of inorganic alumino ‑silicate phase with crystalline water which significantly overlap the intensions belonging to the organic phase of the sample. Based on the absorption bands corresponding to the vibrations n (C ‑H) and d (C = O) it can only be said that the sample is anatural resin without any specifications. 75ROMANA KOZÁKOVÁ Fig. 7: ATR ‑IR spectrum of the analysed amber sample (left) and the ATR ‑IR spectrum of the analy‑ sed sample in comparison with the standards of Baltic amber (Ukraine, 24384) and of rumanite (Ploesti, 24372) in the spectral zone of 750–1950 cm‑1 (right). S26 The gem is preserved in this ring, firmly set in the bezel. Though it is set more deeply than in S11, it is likewise cut in order to fit the narrow bezel ring (Fig. 8c). Repeated surface measurements identified the alloy as brass while the punctual SEM/EDX analyses (Fig. 8a–c) failed to detect any traces of Zn.This may be caused by an insufficient homogeneity of the alloy or by the use of various alloys for the ring and bezel respectively. The measurements were taken in the transverse cut through the ring and on aflake of the surface layer from the same spot. Both samples contained Cu, numerous elements connected to corrosive processes (S, Si, Ca, Mg and ahigh amount of P and K), and also small proportions of Sn, thus excluding the possibility of an intentionally applied surface layer. The surface is thus covered by amere corrosion crust underlaid by high amount of chlorides. The composition of the gem, measured by the XRF analyser in the ‘non ‑metallic mode’, was approximately determined as Si– Al– Ca– Mn, with trace presence of K, P, S, Sr, Rb, Fe aTi, indicative of (most probably soda) glass. S27 The gem of this finger ‑ring is preserved though no more within the bezel from which it had fallen off. The lining of the bezel beneath the gems (Pl. 4/1:S27a) is well preserved including– in its entire extent– agreyish metallic layer with localised remains of gold and in its central part macroscopically visible remains of an organic substance, most probably natural glue. The nature of this substance was not identified. The metal lining was analysed by means of SEM/EDX which confirmed its assumed iden‑ tification as surface gilded silver leaf. Its dark lower level (Fig. 9) contained significant pro‑ portions of Cu and Zn while the presence of Pb, identified by XRF was not confirmed. The alloy of the finger ‑ring body was identified as brass by the XRF, while SEM/EDX de‑ tected Zn only in the bottom of the bezel beneath the lining but not in the sample from the ring body; Sn was identified in one measured point. Just like in the case of S26, this may be caused by an insufficient homogeneity of the alloy or by two different copper alloys. Unlike S26, however, no elements characteristic of corrosion have been detected. The microscopy of the gem executed in the Department of Petrography of the National Museum identified the material clearly as glass. The approximate measurements of the ele‑ mentary composition by the XRF suggest that it can be classified as soda glass. 76 STUDIA HERCYNIA XX/1 Fig. 8: S26– a) The SEM photograph of the analysed sample. The arrows indicate the different points where measurements were taken; b) The SEM photograph of the analysed sample– detail of the marginal part of the sample with the darker part more affected by corrosion while in the lighter part the metal was recently cleaned; c) Insertion of the gem into the bezel; d) Point where the sample was taken (a–b: photo by the author, c–d: photo by A. Kumstátová). Fig. 9: S27 The SEM photographs of the analysed sample. The arrows indicate the different points where measurements were taken. Left: Sample including the entire stratification. Right: Detail of the base metal (photos by the author). Centre: the point of sampling (photo A. Kumstátová).