Étude archéométallurgique de parures hallstattiennes en alliage cuivreux du Rhin supérieur

• Laurie Tremblay Cormier ^[1] ; Benoît Mille ^[2]
  1. [1] UMR 7044. Archimède
  2. [2] Centre de recherche et de restauration des musées de France
• Localización: Bulletin de la Société Préhistorique Française, ISSN 0249-7638, T. 113, Nº. 4 (Octobre-décembre), 2016, págs. 785-808
• Idioma: francés
• Títulos paralelos:
  • Archaeometallurgical study of Early Iron Age copper alloy ring ornaments from the Upper Rhine
• Texto completo no disponible (Saber más ...)
• Resumen
  • English

    During the Early Iron Age, the Upper Rhine valley stands out by its very typical metal production. The general typology shows strong relations with the neighbouring regions of the Swiss Plateau, Württemberg and the Rhineland, composing an area of similar cultural trends set along the Rhine axis and the Upper Danube. The present work aims at improving the characterization of the Upper Rhine metal production, by completing the typological analysis with the study of metallurgical traditions. It also strives to fill a gap in research regarding copper alloys in the regions north of the Alps, where the Iron Age is still poorly known. A set of ring ornaments of regional types has been selected according to two criteria. First, the geographical distribution of the type must be limited to or, at least, clustered in the Upper Rhine valley; indeed, as the workshops are unknown, this information alone can suggest a regional origin for the objects. In addition, the objects must be thick enough to be sampled without damage, leading to the exclusion of small ornaments like brooches and pins. Consequently, the items selected for the analysis are arm rings of the Bade-Alsace type and closed, massive rings with a circular section and no decoration. They are evenly distributed in the area and cover a chronological span extending from Ha C2 to Ha D3, or from the 7th to the middle of the 5th century BC. Remains of copper alloy working are quite scarce in the Upper Rhine valley, and are mainly represented by crucible and mould fragments. Three objects from Britzgyberg have thus been analysed, as the only direct comparisons between copper alloy metallurgy and the ring ornaments: a casting sprue, and the fragments of an ingot and of an ankle ring. The procedure is based on the macroscopic study of the rings and the analysis of their elemental composition. The main goal is the development of a typology of chemical compositions and its archaeological interpretation. In order to check potential composition and fabrication relationships between the objects, several levels of association have been tested: objects composing a pair, objects with different functions from the same grave, objects from the same cemetery and different graves, and isolated objects. The typology of elemental compositions has been completed by the search for twin compositions, which can bring to light production relationships between some objects, manufactured during the same casting (crucible twins) or using the same copper ingot (ingot twins). The analyses were carried out in the Centre de recherche et de restauration des musées de France (C2RMF) using the PIXE method, on the AGLAE particle accelerator. They mainly concerned samples taken in the healthy parts of ornaments, on the inner side of the ring or on the plug, using a 1 mm HSS steel drill. Three objects were not sampled but analysed through surface analysis, without cleaning, at the request of their museum; the results are however hardly meaningful. Indeed, two objects were too corroded while the third one shows extreme values, which may be due to the segregation of chemical elements and their migration to the surface of the object during its cooling, or to surface alteration during its burial in the ground. A total of thirty-one chemical elements have been analysed. Among these, twenty are considered as significant for the study of ancient copper alloys; several of them were excluded, as they were below the detection limit (Mn, Zn, Se, Cd, In, Te, Au, Hg). The statistical processing thus concerned the following elements: copper (Cu), tin (Sn) and lead (Pb), as main alloy components, and phosphorus (P), sulphur (S), iron (Fe), cobalt (Co), nickel (Ni), arsenic (As), silver (Ag), antimony (Sb) and bismuth (Bi) as impurities. The alloy components show strong differences in the distribution of lead, which seems to correspond to two distinct tendencies in the alloys. According to this, the alloys can be classified in four groups: binary alloys where lead is an impurity (less than 1%; type 1a) or was unintentionally added (1 to 2%; type 1b); ternary alloy with a low amount of added lead (2 to 6%; type 2); and ternary alloys where lead exceeds tin and is added in a moderate (6 to 10%; type 3) or strong amount (over 10%; type 4). The tin distribution is correlated to these composition types: binary alloys have 8% or more tin, while ternary alloys have less than 8% tin, but never go under the limit of 3%. Therefore, this shows that lead did not replace tin, which remained at a level sufficient to improve the properties of the alloy (hardness, colour...), but was added to an already existing bronze mix, creating ternary alloys that can be called ‘leaded bronzes’. These results show clear variations of the chemical composition according to typology. The Bade-Alsace arm rings are made of classical tin-bronzes (1a and 1b types), with an impurity pattern that changes from one variant to the other (A1, B1 and B2); the casting sprue also falls within this group. On the other hand, many massive closed ring have a high lead content, going up to 36%, and are mainly distributed between alloy types 3 and 4, with the ingot fragment and the ‘as-cast’ ankle ring from Britzgyberg. A chronological evolution can be seen in this scheme, Bade-Alsace arm rings being older than many of the massive closed rings with a high lead content; leaded bronzes could thus appear during the second half of Ha D (6th century BC), although this needs to be confirmed with further investigations. This phenomenon raises the question of the reasons behind the addition of lead to the alloy; aesthetic and technical hypotheses can be proposed, respectively concerning the visual aspect of the ornament and the lowering of the alloy’s melting point. However, an economical reason also seems likely, as lead could have been added to bronze in order to increase the material’s mass. Leaded bronzes would thus be a response to the waning of recycled bronze stocks, showing deep changes in economic dynamics, with roots in the availability of primary resources. The search for twin compositions called on a hierarchical clustering analysis of the chemical composition, on standardized data and using the Ward aggregation method, first on the detected impurities (for ingot twins), then adding also tin and lead (crucible twins). Even with a strict similarity limit (Euclidean distance equal or inferior to 1), many crucible twins can be seen within the cluster, and even more for ingot twins. The great number of crucible twins carries important archaeological implications, among which the strict contemporaneity of the concerned objects. In this perspective, PIXE analyses need to be completed with a more sensitive and precise technique, in order to obtain even more reliable results. It is not surprising to note that archaeological pairs are also crucible twins, which confirms the strong unity of ornaments with similar shapes. But, except for object pairs, crucible twins are rarely found within the same grave: neck, arm and ankle rings of a same person were made during separate castings and using different ingots, testifying that sets of ornaments were gradually gathered. Within a same cemetery, crucible twins can be distributed between many graves; the phenomenon also extends between cemeteries, in particular between the Haguenau cemeteries but also to the whole Upper Rhine valley, adding to the high homogeneity of the material culture. These observations lead to the restitution of strong relations between different communities, sharing one common thread: the craftsman (or workshop) who cast the object, around whom bronze production thus seems to have been centralized. It will therefore be necessary to investigate this organisational trend of the metal economy, by extending the analyses to a greater number and different types of objects.

  • français

    La vallée du Rhin supérieur se distingue, au premier âge du Fer, par une production métallique caractéristique ; la présente étude a pour objectif d’en approfondir la définition, en complétant l’analyse des formes par celle des traditions métallurgiques. Elle vise également à combler un déficit de la recherche sur les alliages cuivreux, où l’âge du Fer est encore faiblement représenté. Un corpus de parures de types régionaux a été sélectionné (bracelets de type Bade-Alsace, parures annulaires massives de section circulaire, fermées et lisses), couvrant une large période allant du Ha C2 au Ha D3, soit du viie au milieu du ve siècle avant notre ère. Le protocole est basé sur l’étude macroscopique des parures et l’analyse de leur composition élémentaire, avec pour objectif d’établir différents types de composition chimique et d’en proposer une interprétation archéologique. Les analyses ont été réalisées au Centre de recherche et de restauration des musées de France (C2RMF) à l’aide de la méthode PIXE, principalement sur des prélèvements faits dans les parties saines des parures. Les résultats montrent de nettes différences de composition chimique selon la typologie ; les bracelets de type Bade-Alsace sont ainsi réalisés suivant la recette classique du bronze à l’étain et montrent une variation du spectre d’impuretés qui dépend principalement de la forme du bracelet (variantes A1, B1 et B2). À l’inverse, de nombreuses parures fermées lisses possèdent un taux élevé de plomb, allant jusqu’à 36 %. Ce phénomène pose la question des raisons qui motivent cette forte présence du plomb dans l’alliage ; un rôle plus économique que technique peut être invoqué, avec une volonté d’augmenter la masse globale de métal par l’adjonction de plomb au bronze recyclé. La pratique des bronzes au plomb trouverait ici pour explication l’appauvrissement des stocks de matière première, et traduirait ainsi des changements profonds dans les dynamiques économiques, touchant plus particulièrement aux ressources disponibles. L’analyse est complétée par la recherche des compositions jumelles, qui correspondent à des objets présentant une composition chimique identique. La mise en évidence de jumeaux de lingot et de creuset permet de proposer l’existence de liens d’ordre technologique entre certaines pièces, qui appartiennent parfois à des ensembles ou à des sites différents. Leur distribution dans l’ensemble du Rhin supérieur, autant sur ses deux rives que dans ses parties nord et sud, atteste l’existence d’un réseau de relations soutenues entre les artisans et la population. L’homogénéité des traditions métallurgiques va ainsi de pair avec celle de la typologie, ajoutant à la cohésion déjà forte de la culture matérielle du Rhin supérieur.