Egyptian Middle Kingdom Copper: Analysis of a Crucible from Buhen Christopher J. Davey 1,2 , Brunella Santarelli 3 , Thilo Rehren 3 (1) University of Melbourne, Parkville Vic 3052, Australia (2) Australian Institute of Archaeology, La Trobe University, Vic 3086, Australia (3) Science and Technology in Archaeology and Culture Research Center, The Cyprus Institute, Nicosia, Cyprus Introduction Zone A, ceramic body with quartz and small iron oxide inclusions, and evidence of chaff temper. SiO 2 : Al 2 O 3 ratio 3:1, ~10 wt% FeO and 4-5 wt% CaO. Zone B, bloated layer of heavily vitrified ceramic, large porosity. Zone C, dense and fused ceramic, enriched in lime. Zone D, crucible slag, ~100-300 µm thick; rich in FeO (25-65 wt%) and SiO 2 (20-40 wt%). Inclusions of silicates (quartz, pyroxenes and fayalite), spinels (magnetite, hercynite), other iron oxides, and copper-iron oxides (delafossite), with a thin layer of Cu- and Fe- rich phases on the exterior edge of the slag, containing also several copper prills (ranging in size from a few microns to the largest ~60 µm). Results Analytical Methods Sampling and Sample Preparation: A small section of a crucible from Buhen held at the UCL Petrie Museum (UC 21748, Fig. 2) was removed from the base of the crucible, where temperatures were the highest during use. Sampling was done by the UCL staff following consultation with C.J.D. and Th.R. The sample was then prepared as a cross-section by Andreas Ludwig at the Deutsches Bergbau-Museum in Bochum and sent to CyI for analysis. The polishing step at the DBM includes lapping the mounted block on a lead-based plate, which results in widespread contamination of porous samples with metallic lead particles trapped in pores and cracks. Analysis: The cross-section was first studied using a Zeiss Axio Imager microscope in reflected light. Images were collected with Zeiss Zen 2 Core software. SEM-EDS analysis was then performed with a Zeiss EVO 15 SEM equipped with an Oxford Instrument Ultim Max EDS with a 65 SDD detector. EDS analysis was carried out in high vacuum at 20 kV, 1nA, with a 30 µm aperture and 8.5 mm WD. Lead was excluded from the element list of all analytical points due to the known contamination of the sample with this element as a result of the sample preparation procedure. Fig. 3: Reflected light image of crucible fragment at 2.5X mag Imaging Fig. 4: Reflected light images of crucible slag layer showing copper chloride corrosion in pores (a) and copper prills (b) Fig. 5: Backscattered electron images of crucible slag layer showing multiple phases (including magnetite, delafossite and pyroxenes) and copper prills Interpretations The crucible slag is a combination of molten ceramic fluxed with fuel ash, and remains of the charge. Thus, the composition of the slag layer and the metal prills trapped within it reflects the metallurgical operation that took part in this crucible. Though there is wide variation in the composition of the slag layer, as is typical in crucible slags, the bulk consists of iron oxide, which does not derive from the ceramic nor from the fuel ash but from the charge in the crucible. The remaining composition of the slag resembles the composition of the ceramic, but with a significantly increased silica content relative to alumina, indicating an additional source for silica in the slag. The increased lime content is again most likely due to fuel ash. There are two possible metallurgical techniques that this crucible could have been used for: the smelting of copper ores to extract copper metal, or the re-melting of pre-existing (raw) copper metal to refine and prepare it for casting. The increase in iron oxide in the crucible slag, which is reflective of the metallurgical charge could be explained by both the refining of iron-containing copper under oxidizing conditions, or by the smelting of a copper ore that contained a small amount of iron-bearing minerals. However, the increased content of silica in the crucible slag cannot be explained by the re-melting of raw copper, as it would not include silica. The composition of the copper prills within the crucible provides further indication regarding the process. Besides the dominant copper, these prills contain also several wt% each arsenic, iron and nickel. Particularly, the presence of metallic iron in copper indicates that this metal was liquid under strongly reducing conditions typical of early copper smelting, and is not consistent with the more oxidising conditions of re-melting pre- existing (raw) copper. The high levels of nickel and arsenic resemble a large group of As- and Ni-rich copper known from the wider region. Therefore, we can conclude that this crucible was most likely used for the smelting of malachite containing minor amounts of iron- and silica-rich gangue minerals, as well as nickel- and arsenic-containing accessories. This smelting took place under relatively strong reducing conditions, as implied by the high iron content in the copper prills. This is consistent with the documented smelting practices described elsewhere during the Middle Kingdom. Fig. 2: Crucible UC 21748, (a) location of removed sample and (b) profile drawing The ancient city of Buhen was located immediately below the second cataract on the west bank of the Nile in the Sudan and near the border with Egypt. It was often the southern frontier of the ancient Egyptian state. Prior to its inundation in 1964, it was excavated by the University of Pennsylvania’s Eckley B. Coxe Expedition in 1909-10 under the direction of David Randall-MacIver (1873-1945) and Charles Leonard Woolley and by Walter Emery of the Egyptian Exploration Society after 1957. Emery found what he claimed was a copper ‘factory,’ making the site important to Egyptologists researching metal working. While some of Emery’s metallurgical features are questionable, the site yielded numerous crucibles and other metal working equipment and debris. The crucible styles are typical of other Egyptian crucibles of the period, few of which are fragmentary and available for analytical study. A sample of the fabric from one Middle Kingdom crucible fragment in the Petrie Museum (UC 21748) has been investigated at The Cyprus Institute’s STARC, using optical microscopic techniques and SEM-EDS analysis to investigate the processes for which this crucible was used. Average Composition of Cu Prills (wt%) Cu Fe Ni As 85 3.3 2.3 8.5 Average Bulk Composition of Major Oxides in Slag (wt%) MgO Al 2 O 3 SiO 2 CaO FeO CuO 2 7 30 10 43 4 Thin surface layer of Cu- and Fe-rich phases Magnetite and delafossite Copper prills Fig. 1: Map of Egypt and Sudan showing location of Buhen (a) (b) (a) (b) (a) (b) (c) (d) Average Bulk Composition of Ceramic Zone A (wt%) Na 2 O MgO Al 2 O 3 SiO 2 K2O CaO TiO 2 FeO 1.3 2.5 17 55 2.6 4.6 2.3 11 References Claes, W., Davey, C.J. and Hendrickx, S., 2019. An Early Dynastic Crucible from the Settlement of Elkab (Upper Egypt), Journal of Egyptian Archaeology, forthcoming. Craddock, P.T. 2000. From hearth to furnace: evidences for the earliest metal smelting technologies in the Eastern Mediterranean. Paleorient, 26 (2), 151-165. Davey, C.J. 2009. A metalworking servant statue from the Oriental Institute, University of Chicago, Bulletin of the Australian Centre for Egyptology 20, 37-46. Emery, W.B., et a.l 1979. The Fortress of Buhen: The Archaeological Report, London: Egypt Exploration Society. Tadmor, M., et al. 1995. The Nahal Mishmar Hoard from the Judean Desert: Technology, Composition, and Provenance. Atiqot 27, 95–148.