PART ONE: MATERIALS 52 relatively pure native metal. The rate of this transformation increased rapidly soon after the establishment of the Tigris and Euphrates civilizations. The wealth and specialized demand provided by these urban societies must have stimulated early copper workers to prospect in the northern mountainous regions where weathered outcrops of copper were most likely to be encountered. The earliest copper workers appear to have extracted their metal from oxide or carbonate ores which, although not always rich or plentiful, could generally be smelted successfully in the primitive furnaces then available. The early smelters all appeared to understand instinctively that charcoal fires could be adjusted to provide atmospheric conditions which simultaneously reduced copper and oxidized iron. Methods were thus evolved which allowed relatively pure copper to be separated in the molten state from iron and other unwanted materials in the ore. These, when suitably oxidized, could be induced to dissolve in the slag. The primary ores of copper are invariably complex sulphides of copper and iron, and are generally disseminated in a porous rock such as sandstone which rarely contains more than 2 per cent by weight of copper. Such deposits were too lean to be exploited by primitive man, who sought for the richer if more limited deposits produced by the weathering and oxidation of primary ores. Thus, at Rudna Glava in Yugoslavia, a copper mine worked in the 6th millennia, did not exploit the main chalcopyrite ore body, but worked instead a thin, rich carbonate vein produced by leaching and weathering. This concentrated ore contained 32 per cent of copper and 26 per cent of iron. Quartz sand would have been added to such a smelting charge to ensure that most of the iron separated into the slag. At Timna, in the southern Negev, copper has been mined and smelted since the dawn of history. Extensive workings, slag heaps and furnaces have remained with little disturbance since Chalcolithic, Iron Age and Roman times. These mines, traditionally associated with King Solomon, were in fact worked by the Egyptian Pharaohs during much of the Iron Age until 1156 BC. The primary ore deposit at Timna is based on the mineral chrysocolla and is currently being exploited on a large scale. Since this ore contains only about 2 per cent of copper, it could not have been effectively smelted in ancient times. During the Chalcolithic or historical periods copper was extracted at Timna from sandstone nodules in the Middle White sandstone beds overlying the chrysocolla deposits. The nodules contain between 6 and 37 per cent of copper, which exists as the minerals malachite, azurite and cuprite. The remainder is largely silica, and the nodules contain little iron. In the fourth millennium BC copper was extracted from them in furnaces: a rough hole in stony ground, approximately 30cm (1ft) in diameter, was surrounded by a rudimentary stone wall to contain the charge, which consisted of crushed ore mixed with charcoal from the desert acacia. Controlled quantities of the crushed iron oxide haematite, was added to the charge as a flux, to reduce the NON-FERROUS METALS 53 Figure 1.3: Reconstructions based on the remains of smelting furnaces used (a) in the twelfth century BC, and (b) in the eleventh century BC, at the ancient copper smelting site of Timna in the Southern Negev region of Israel. Courtesy of the Institute of Metals. PART ONE: MATERIALS 54 melting point of the silicious material and improve the separation between the copper and the slag. The Chalcolithic smelting furnaces at Timna appeared to have had no tap hole and no copper ingots were found. High concentrations of prills and blebs of metallic copper were found in the slag, however, and it seems possible that the metal never, in fact, separated from the slag in massive form: after smelting the slag would have been broken up to remove the prills which were then remelted together in a crucible furnace. A more highly developed smelting furnace used by the Egyptians at Timna around 1200 BC is shown in Figure 1.3 (a). The slags from such furnaces contained up to 14 per cent of lime which was added to the charge as crushed calcareous shells from the Red Sea. This addition would have improved slag metal separation and allowed the reduced copper to settle to the bottom of the furnace and to solidify below the slag as plano-convex ingots. Smelting techniques appeared to have reached their zenith at Timna around 1100 BC. After the smelting operation the slag and metal appear to have been tapped simultaneously from the furnace into a bed of sand where, as recent simulation experiments by Bamberger have shown, they would have remained liquid for about fifteen minutes, providing ample time for the molten copper to sink beneath the slag to form well-shaped ingots about 9cm (3.5in) in diameter. For Bamberger’s reconstruction see Figure 1.3 (b). The rings and other small artefacts of iron found at Timna are now thought to have been by-products of the main copper refining operation. Lead isotope ‘finger-printing’ has shown that the source of the iron was the haematite used to flux the copper ore during refining. It would appear that when the ‘as smelted’ copper was remelted in a crucible furnace in preparation for the casting of axes and other artefacts, any surplus iron it contained separated at the surface of the melt to form a sponge-like mass permeated by molten copper. This layer would, in all probability, have been skimmed from the surface of the melt before it was poured. At a later stage it must have been found that the iron/copper residue could be consolidated by hot forging and worked to the shape required. The presence of copper is known to improve the consolidation of iron powder, and it would seem, therefore, that a sophisticated powder metallurgical process, utilizing liquid phase bonding, was being operated at Timna in Iron Age times. Smelting of sulphide copper ores From the presence of arsenic and other impurities in many of the early Copper Age artefacts it must be concluded that much of the copper used was extracted from sulphide rather than oxide or carbonate ores. In prehistory, as in modern times, the bulk of the world’s supply of copper appears to have been obtained NON-FERROUS METALS 55 from ores based on chalcopyrite, a mixed sulphide containing equi-atomic proportions of iron and copper. Chalcopyrite must be roasted in air to convert it to a mixture of iron and copper oxides before it can be smelted. Moreover, because chalcopyrite ores contain in general less than 2 per cent of copper, and because of the presence of large quantities of unwanted earthy material, they do not respond to simple smelting processes. The sulphide copper ores exploited at the beginning of the Copper Age appear to have been thin, localized and very rich deposits which lay some distance below the surface of a weathered and oxidized primary outcrop. Although the presence of such enrichment zones has been recognized by mining engineers and geologists for many years, their significance as ancient sources of copper has only recently been fully appreciated. Due to the atmospheric oxidation which occurs at the surface of chalcopyrite outcrops, the sulphides are partially converted to more soluble compounds which are slowly leached away. The exposed surface of such an outcrop is therefore slowly robbed of most of its heavy metals with the exception of iron which concentrates at the surface as iron oxide, generally in the form of limonite. The iron-oxide regions above rich copper deposits are known as gossans, and the German term eisener Hüt to describe a gossan led to the saying, ‘For a lode nothing is better than it should have a good iron hat,’ and extensive gossans are noteworthy features of most of the sulphide copper ore deposits which were worked in antiquity. At Rio Tinto in south- west Spain, where copper has been extracted from the earliest times, iron is so extensively exposed at the surface that the terrain resembles that of an open-cast iron ore mine. Streams such as the Rio Tinto and Aguar Tenidas which leave this gossan owe their names to the red contaminant iron oxide. The ancient mines at Oman, which provided the Sumerians with copper, are characterized by huge ferruginous gossans, and similar terrain exists at Ergani Maden in Turkey. From these gossans the copper content has been completely leached away, and transferred in aqueous solutions to lower horizons, where, as the dissolved oxygen becomes depleted, it is precipitated in sulphide form in the surrounding strata. In this way zones of secondary enrichment are formed which contain most of the metallic content which was originally uniformly distributed throughout considerable depths of rock. The average copper content of the thin secondary enrichment zones at Rio Tinto sometimes approaches 15 per cent and values as high as 25 per cent of copper in the enrichment zones at Ergani Maden have been reported. The arsenical and antimonial minerals are associated with the copper in the cementation zone, and the preponderance in the Copper Age of artefacts containing substantial quantities of arsenic is therefore a further indication of the fact that much copper of this period came from these thin zones of secondary enrichment. PART ONE: MATERIALS 56 Arsenical copper When an awareness that many early copper artefacts were in fact dilute alloys of arsenic in copper began to develop after 1890, it was thought that arsenic had been deliberately added to improve the mechanical properties. Bertholet, in 1906, was the first to demonstrate that the concept of alloying would have been unknown when these artefacts came into general use, and that arsenical copper, which was extensively used by the ancient Egyptians, was a natural alloy obtained by smelting arsenical copper ore. In the cast condition arsenical copper is only marginally harder than pure copper, although its hardness increases very rapidly as a result of cold working. This must have been a factor of great importance in the Copper Age when edge tools were invariably hardened and sharpened by hammering. Arsenic deoxidizes copper very effectively and the alloys are noted for their excellent casting characteristics, a factor which early copper workers would much have appreciated. Cast billets of arsenical copper would, however, have been more difficult to work than pure copper, and it seems evident that when not cast directly to the required shape weapons and cutting tools were fashioned by a judicious mixture of hot and cold working. Such implements would certainly have retained their cutting edges for longer periods than their pure copper counterparts. When arsenical copper ores are smelted with charcoal, the arsenic has little tendency to escape because it is held in solution by the molten copper. Similarly, when copper alloys containing up to 7 per cent of arsenic are melted in a crucible under reasonably reducing conditions, little is lost. Arsenious oxide, however, is very much more volatile than elementary arsenic, and toxic fumes must certainly have been emitted in copious quantities during the roasting of copper sulphide ores containing arsenic. Since sulphur dioxide fumes would also have been given off in vast quantities, the additional health hazards caused by arsenic would not, however, have been separately identifiable. Certainly nothing appears to have inhibited the use of these accidental arsenical copper alloys, which were extensively produced for a very long period over most regions of the ancient world. Primitive man, when he stumbled upon the thin layers of concentrated copper ores immediately below the gossan must soon have appreciated that they produced artefacts having properties vastly superior to those of the copper hitherto obtained from oxide or carbonate ores. On the basis of gradually accumulated experience he would have sought for similar or even richer ores in other localities, but in view of the absence of any concept of alloying it seems unlikely that attempts were made to control the hardness of the copper obtained by adjusting the mixture of ores fed into the smelting furnace. The beginning of the arsenical copper era is difficult to date with any certainty. The earliest Egyptian artefacts of arsenical copper were produced around 4000 BC in predynastic times. In addition to arsenic these early NON-FERROUS METALS 57 weapons contained around 1.3 per cent of nickel. All the copper or bronze artefacts found by Sir Leonard Woolley during his excavations of the Royal Graves at Ur contained similar quantities of nickel which at the time these discoveries were made was considered to be a most unusual constituent of ancient copper. It has been suggested that the Sumerians who made these artefacts came from the Caucasus and were instrumental in transferring the arts of metallurgy from the land of Elam (which now forms part of Iran) into Babylonia. The first Sumerian kingdom was destroyed by the Semitic King Sargon, and one of his inscriptions, dating from 2700 BC, indicated that the Sumerians obtained their copper from the copper mountain of Magan. Shortly after the Royal Graves excavation at Ur, the remains of extensive ancient copper workings were discovered in the vicinity of the small village Margana in Oman. The impurity spectrum of this copper ore deposit, including the nickel content, corresponded exactly with that of the artefacts of Ur. Oman, however, is a long way from Babylonia, and since nickel bearing copper ores are also found in India and in the Sinai desert, the true origins of Sumerian copper are still uncertain. Indirect evident for direct links between Ur and Oman is provided, however, by evident similarities between the smelting techniques used to produce plano-convex copper ingots at Suza in Iran and at Tawi-Aaya in Oman, during the third millennium BC. The archaeological evidence also indicates strong cultural links between Southern Iran and Oman at this time. Some of the slags found at Oman contained copper sulphide matte. Well- roasted sulphide ores can be effectively reduced to metal by techniques similar to those used to treat oxide or carbonate ores. Separation of the copper from the slag is facilitated by small flux additions such as bone ash, and this echoes the sea-shell additions made at Timna (see p.54). The essential difference between the treatment of oxide and roasted sulphide ores, however, is that small quantities of copper sulphide matte appear to be produced in that furnace when the roasted sulphide ore is being treated under atmospheric conditions which reduce the copper, but retain the iron in an oxidized state. This matte, being insoluble in the slag, separates out as a thin silvery crust on the surface of the copper ingot. The presence of copper matte in ancient slag deposits is not, therefore, conclusive evidence of matte smelting. TIN AND BRONZE The Early Bronze Age For well over 1500 years arsenical copper artefacts were extensively produced in the ancient world, which had come to expect from its metallic implements PART ONE: MATERIALS 58 standards of hardness and durability which were unattainable from pure copper. After about 3000 BC, however, the arsenical content of Middle Eastern copper artefacts begins to decline, perhaps because of the progressive exhaustion of rich, accessible arsenical copper ore deposits. Artefacts containing small quantities of tin made their first appearance during the early stages of this decline although the quantities involved rarely exceeded about 2.5 per cent. Such alloys, which first appeared in Iran around 3000 BC, did not reach regions such as England and Brittany, on the outer fringes of civilization, until about 2200 BC. These low tin bronzes must be regarded as the precursors of the later true bronzes containing 8–10 per cent of tin. They presumably emerged accidentally, either by the smelting of copper ores which contained tin minerals, or by the use of tin-bearing fluxes. If the possibilities of alloying had been grasped at this stage it seems logical to assume that they would have been progressively exploited. No evidence for such a gradual evolutionary process has been found, and virtually no artefacts containing more than 2.5 per cent of tin have been identified which antedate the sudden emergence of the 8–10 per cent tin bronzes in Sumeria between 3000 and 2500 BC. It would appear that the copper workers of Ur suddenly discovered or acquired the concept of alloying, and rapidly developed and optimized the composition of bronze. This sudden leap forward cannot be separated from the rapid expansion of trade in the Middle East around 3000 BC, since it seems improbable that the significance of tin as an additive to copper would have first been identified in a region where its ore was not indigenous. Sources of ancient tin The most abundant and significant tin mineral is cassiterite, the oxide Sn O 2 , which varies in colour from brown to black. Cassiterite is noteworthy in that its high specific gravity of 7.1 is comparable to that of metallic iron, and also because its hardness is comparable to that of quartz, so that it is highly resistant to abrasion and tends to concentrate in gravels and alluvial deposits. The name tin appears to be derived from the Chaldaean word for mud or slime, which implies that the tin originally used at Ur came from alluvial deposits. The Greek word kassiteros was taken from a Celtic term which has been literally translated as ‘the almost separated islands’, presumably the mythological tin islands. In classical times, kassiteros was loosely used to denote tin, pewter and sometimes even lead. The Greek word for Celtic, kentikov, was also used by Aristotle to describe metallic tin. Four cassiterite mines have been located in the eastern Egyptian desert, one of which was accurately dated from inscriptions associated with the Pharaoh NON-FERROUS METALS 59 Pepi II to around 2300 BC, approximately 500 years after the appearance of the first Sumerian bronze. Copper artefacts containing significant quantities of tin were not produced in Egypt before the Fourth Dynasty, around 2600 BC, and the true Bronze Age in Egypt, when artefacts containing 8–10 per cent of tin were produced, did not begin until 2000 BC. It seems logical to assume that at some time around 3000 BC, the concept of bronze manufacture was acquired by the Sumerians, either by trade or conquest, from a region outside Mesopotamia where tin was a readily available commodity. Many historians have claimed that by 3000 BC commerce in the Middle East had developed so extensively that supplies of tin might well have reached Sumeria by sea, up the Persian Gulf, from regions as remote as Malaysia or Nigeria. Support for this general idea is, of course, provided by the known importation by Sumeria of copper from Oman. From the Gulf, the tin would have been transported by overland caravan. Between 2600 and 2500 BC supplies of tin to Mesopotamia appear to have been interrupted, many of the copper artefacts dating from the Second Sumerian Revival being simple arsenical coppers containing little or no tin. This occurred, of course, in biblical times, when the Land of Sumer was devastated by floods. The shortage of tin, whether it was caused by natural disasters or by political upheavals, was not prolonged, however, and after about 2500 BC the use of 8–10 per cent tin bronzes expanded rapidly throughout the Middle Eastern world. Early Bronze Age developments have also been found in southern China, Thailand and Indonesia, where alluvial tin and copper ore are sometimes found in close association. It was not until 2000 BC that bronze, or even copper, was manufactured in northern Thailand. The bronze artefacts, including spears, axeheads and bracelets, recently found at the village of Ban Chiang were obviously made locally, since stone axehead moulds were also discovered. The alloys, which contained 10 per cent of tin, must have been produced by craftsmen well-versed in bronzeworking technology; and since no evidence of earlier or more primitive metal working has been found, it would seem that this modest agricultural community suddenly acquired a fully developed facility for bronze manufacture. It probably arrived with the peaceful integration of a group of skilled foreign workers, perhaps displaced by military disturbances in China. Chinese bronze containing 8 per cent of tin was being produced in Gansu province as early as 2800 BC, apparently independently of the emergence of high tin bronzes in Sumer. The earliest tin bronzes so far identified appear to date from the fourth millennium BC and were found in the 1930s at the site of Tepe Giyan, near Nahavand in Western Iran. This mountainous region, situated midway between the Persian Gulf and the Caspian Sea, was in the Land of Elam mentioned in the Bible, from which the Sumerians were assumed to have obtained the arts of metallurgy. Although Tepe Giyan does not appear to have a local source of tin, it PART ONE: MATERIALS 60 is possible that there were alluvial tin deposits, now exhausted. The sudden appearance of true bronzes at Ur around 2800 BC, and in Amuq, near Antioch in Turkey, about 3000 BC would be consistent with the view that the concept of bronze manufacture originated in a community such as Tepe Giyan, in the Persian highlands, during the 4th millennium, and subsequently moved southwards to Sumeria and the Persian Gulf, and westwards to the Mediterranean seaboard, during the third millennium. Bronze was first made in Italy between 1900 and 1800 BC, using tin from deposits at Campiglia Marittima in Tuscany, although it seems possible that cassiterite was also being obtained from mines in Saxony and Bohemia. Copper extraction started in Spain in Neolithic times and during the third millennium its copper and precious metal deposits were extensively worked (see p. 55) Exploitation of the Spanish tin deposits, however, did not begin until 1700 BC, when bronze artefacts were first produced at El Argar and other sites in the south-east. Most of the bronze artefacts of this period, which contain around 8 per cent tin, appear to have been cast roughly to shape and finished by forging. Evidence that the early Mediterranean cultures obtained significant quantities of tin from Cornwall, Brittany or Saxony-Bohemia is singularly lacking, although some of the tin used for Central European bronzes made between 1800 and 1500 BC appears to have come from the Saxony deposits. Apart from a tin bracelet dating from 3000 BC found at Thermi in Lesbos, very few Bronze Age artefacts made of metallic tin have been discovered. This paucity is difficult to reconcile with the vast quantities of tin which must have been used for bronze manufacture over the period. It has been suggested that bronze was originally produced by co-smelting copper ore with minerals of tin, and that at a later stage, when better compositional control was required, weighed quantities of cassiterite were reduced directly with charcoal on the surface of a bath of molten copper. This implies that cassiterite, rather than metallic tin, was the standard commodity of trade. Metallic tin, however, far from behaving at all times as a highly inert material can sometimes, when buried for example in certain types of soil, disintegrate very rapidly into amorphous masses of oxide and carbonate sludge, destroying the identity of tin artefacts. The most enduring evidence of Early Bronze Age technology is provided by a few accidentally vitrified clay tuyeres from between 2000 and 1800 BC. Many early tuyeres were ‘D’ shaped in cross-section and appear to have been used in primitive crucible furnaces. Here the crucible was heated from above by radiation from a glowing bed of charcoal, much of which must inevitably have fallen upon the surface of the melt. The crucible in the furnace used at Abu Matar for remelting impure smelted copper between 3300 and 3000 BC (see p. 51) would presumably have been extracted via a hole in the side of the vertical shaft. Many crucibles used in the Mediterranean region between 3000 and 2500 BC had, to facilitate handling, a socketed boss moulded on one side. Clay-covered wooden or metal rods were NON-FERROUS METALS 61 inserted into these sockets so that the crucible could be removed. During the Late Bronze Age, the crucibles used were thicker and more robust. The casting arrangement shown diagrammatically in Figure 1.4 was used between 1550 and 1200 BC in the Greek Islands and also in Sinai. Pouring was accomplished by rocking the crucible on its base, either by pulling with a hook from the front, or by pushing from the rear of the furnace. Bronze Age casting techniques Few plano-convex ingots of arsenical copper have been found, although the material appears to have been available at any early stage in the form of ingot torcs. These Ösenhalsringe, or neck-rings with recoiled ends, resemble in shape the later iron currency bars. Worked from cast ingots, they were traded as an intermediate product suitable for the manufacture of pins, jewellery and other small objects. No unworked Early Bronze Age ingots from which large artefacts such as axeheads could have been cast have yet been found. Figure 1.4: Rocking crucibles of this type were extensively used in the Greek Islands and also in Sinai between about 1600 and 1200 BC to ease the problems of lifting a heavy crucible of molten metal from a primitive furnace and subsequently pouring from it in a controllable manner. The rocking crucibles had thick hemispherical bases and were provided as shown with a pouring hole. When tilted they rolled away from their charcoal bed and discharged their content of molten metal simply and safely. One crucible of this type, found by Flinders Petrie at Serabit in Sinai in 1906 was large enough to contain nearly 8kg of bronze. . came from the Caucasus and were instrumental in transferring the arts of metallurgy from the land of Elam (which now forms part of Iran) into Babylonia. The first Sumerian kingdom was destroyed by the. site of Timna in the Southern Negev region of Israel. Courtesy of the Institute of Metals. PART ONE: MATERIALS 54 melting point of the silicious material and improve the separation between the copper. the fourth millennium BC and were found in the 1930s at the site of Tepe Giyan, near Nahavand in Western Iran. This mountainous region, situated midway between the Persian Gulf and the Caspian