PART ONE: MATERIALS 112 Deville, at Salindres, found that only after very carefully purifying his alumina was he able to produce aluminium containing less than 2 per cent impurities. The technique he employed, the Deville-Pechiney process, was developed by his associate Paul Morin; it is still occasionally used for the treatment of bauxites which contain a lot of iron although it was not too effective in removing large quantities of silica. The Deville process was, however, superseded in 1887 by a cheaper and simpler approach devised by the Austrian chemist Karl Joseph Bayer which is now almost universally employed. Unlike the expensive Deville technique, it depended entirely upon wet chemistry and involves no fusion process. Bauxite was digested under pressure by a caustic soda solution in an autoclave at temperatures between 150° and 160°C. This reaction produced a soluble solution of sodium aluminate, the major impurities, such as iron oxide, titania and most of the silica, being left behind as a red mud. Alumina was precipitated from the caustic soda solution when it was cooled and diluted with water. After calcination it was then suitable as a feedstock for the electrolytic cells. The British Aluminium Company (BAC) Limited was floated on 18 December 1894 to acquire the British rights to the Bayer and Héroult processes and others including the patents and factory site of the Cowles Syndicate at Milton in Staffordshire where a rolling mill was installed. Lord Kelvin was appointed as Scientific Adviser to the company and in 1898 he joined the board of directors. The progress of this company, however, is inevitably associated with William Murray Morrison, who was appointed as chief engineer at the beginning of 1895 and served BAC for half a century. Over this time the world output of aluminium increased from 200 tonnes per annum in 1894, to 5000 tonnes in 1900, and in 1945 to well over two million tonnes. The world output of primary aluminium in 1980 reached a peak in the vicinity of 16 million tonnes per annum. The first BAC plant was established at Foyers, close to the Falls of Foyers on the southern side of Loch Ness, in 1895. It produced about 3.7MW (5000hp) of hydroelectric power and by June 1896 it was extracting about 200 tonnes per year of aluminium, most of which could not be sold. In 1896 about half the power generated at Foyers was sold to the Acetylene Illuminating Company which made calcium carbide by fusing lime and carbon in an electric furnace, a process which had been invented by Moissan in 1892. When after the turn of the century the world demand for aluminium began to increase significantly, BAC started to build an additional hydroelectric plant at Kinlochleven. This took about five years to build. The Lochaber scheme, which commenced in the mid-1920s, was far more ambitious, since it took water from a catchment area covering more than 300 square miles. Waters draining from the Ben Nevis mountain range and the waters of Lochs Treig and Laggan were collected and fed to a powerhouse situated only a mile from Fort William. Further hydroelectric plants were NON-FERROUS METALS 113 established in other regions of the Western Highlands as the demand for aluminium increased after the 1930 period. Magnesium Magnesium was first isolated in 1808 by Humphry Davy, who electrolysed a mixture of magnesia and cinnabar in naphtha. The magnesium liberated was absorbed into a mercury cathode to form an amalgam. The first chemical reduction was accomplished in 1828 by Bussy, who used electrochemically produced potassium to reduce anhydrous magnesium chloride. Magnesium, however, remained a chemical curiosity until 1852 when Bunsen, at Heidelberg, devised a method of producing it continuously by the electrolysis of fused anhydrous magnesium chloride. In 1857 the metal was produced for the first time in quantities large enough to allow its properties to be evaluated by Deville and his colleague Caron. Using the technique he had already perfected for aluminium production (see p. 103), Deville reacted sodium with fused magnesium chloride to obtain magnesium, which was found to be a very light reactive metal. It was also volatile, and could readily be separated, by distillation in hydrogen, from the mixture of fused sodium and magnesium chloride left behind in the reaction vessel. By this time it was also known that magnesium was a metal which burned readily to produce a very intense white light. Bunsen, who studied this effect, found that the light emitted had powerful actinic qualities which rivalled those of sunlight, so that the metal could therefore be of value to the new science of photography. In 1854, Bunsen and a former pupil of his, Dr Henry Roscoe, published a paper on the actinic properties of magnesium light in the Proceedings of the Royal Society and this stimulated a great deal of interest in the metal. In 1859, Roscoe became Professor of Chemistry at Owens College, Manchester. The paper stimulated the inventive genius of Edward Sonstadt, a young English chemist of Swedish descent who is known today largely because he was the first analyst to determine accurately the concentration of gold in sea water. Between November 1862 and May 1863, Sonstadt applied for patents which covered an improved process for producing magnesium, and for purifying it by distillation. By the summer of 1863 Sonstadt was able to claim that his ‘labourer and boy’ had produced several pounds of magnesium metal. During 1863, Sonstadt met Samuel Mellor, from Manchester, who made his living in the cotton industry. Mellor was an enthusiastic amateur chemist, who had, as a part-time student at Owens College, worked under Roscoe and had established close personal contact with him. Sonstadt and Mellor became partners, and Sonstadt moved his laboratory from Loughborough to Salford. In Manchester, Mellor introduced Sonstadt to the pharmaceutical chemist PART ONE: MATERIALS 114 William White, who had developed an improved method of making sodium metal, needed by Sonstadt for his magnesium reduction process. On 31 August 1864 the Magnesium Metal Company was incorporated with a working capital of £20,000. Land for a production plant was acquired at Springfields Lane at Patricroft, with a frontage on the River Irwell. One of the directors appointed at this time was William Matther who also acted as chief engineer of the company. At a later stage, as Sir William Matther, he headed the well-known firm of Matther and Platt. At Salford he developed improved methods of drawing magnesium wire which was then rolled into ribbon. Magnesium ribbon was a major innovation, since it was simpler and cheaper to produce than fine round wire. A very important advantage from the photographic viewpoint was that it burned far more consistently than wire of circular crosssection. Magnesium production at Salford commenced in 1864 and in 1865 the plant produced 6895 ounces (195.5kg) of the metal. A peak output of 7582 ounces (215kg) of metal was reached in 1887. In 1890 magnesium was being imported into Britain from Germany at 1s 6d per lb: Salford could no longer compete and the works were closed. Just before the First World War, Britain had no indigenous source for the magnesium it required, and in 1914 Johnson Matthey, in association with Vickers, set up a magnesium plant at Greenwich. Using the sodium reduction process originally developed by Sonstadt, this plant produced most of the magnesium needed by the Allies for pyrotechnical purposes. After it was closed down in 1919, no more magnesium was produced in the United Kingdom until 1936. German magnesium Bunsen had appreciated as early as 1852 that the cheapest and most favourable production route would involve the electrolysis of a fused anhydrous magnesium salt. Large deposits of carnallite were found in the salt beds of Stassfurt, and most of the earliest German magnesium production operations attempted to utilize this readily available compound. Carnallite, which is found in many evaporite deposits is extremely hygroscopic. Magnesium metal was first obtained on a production scale by the electrolysis of fused carnallite in 1882 by the German scientists Groetzel and Fischer. In 1886 the Aluminium- und Magnesiumfabrik Hemelingen established a plant for the dehydration and electrolysis of molten carnallite using cell designs based almost completely on Bunsen’s original conceptions. Much of the magnesium originally produced at the Hemelingen plant was used for the manufacture of aluminium by a variant of Deville’s process (see p. 106). NON-FERROUS METALS 115 Shortly after the turn of the century, the Hemelingen plant was acquired by Chemische Fabrik, Griesheim Elektron, who transferred it to Bitterfeld in Saxony which soon became the centre of the magnesium world. Before this time magnesium had been regarded merely as an inflammable metal which had applications in photography and pyrotechnics. Griesheim Elektron, however, was led by the energetic and far sighted Gustav Pistor who was one of the first to appreciate the engineering possibilities of a metal with a density of only 1.74g/cm 3 (0.063lb/in 3 ). Aluminium, the only comparable alternative, had a density of 2.70g/cm 3 (0.098lb/in 3 ). Magnesium could be of great value, it was felt, particularly in the aeronautical field. The first hurdle to be overcome was to develop a magnesium free from the corrosion difficulties which had hitherto inhibited its industrial application. Pistor, supported by the very able chemist Wilhelm Moschel, concentrated initially upon the existing carnallite process and evolved an electrolytic bath based on the chlorides of sodium, calcium and magnesium, mixed in such proportions that the eutectic mixture formed melted at approximately 700°C. This type of bath, although melting at a lower temperature than the cryolite/ bauxite mixture used for aluminium production, was and still remains far more difficult and expensive to operate. The main difficulty is that the magnesium metal obtained, being lighter than the electrolyte, floats to the surface of the bath where it must be collected without shorting the electrodes. Furthermore, the gas liberated at the anode is chlorine rather than the oxygen given off in the aluminium cell. The chlorine liberated in magnesium production is collected in bells surrounding the anode and utilized in the production of fresh magnesium chloride. These disadvantages were soon appreciated and after 1905 determined attempts were made to evolve a magnesium bath using an analogy of the Hall/ Héroult method in which magnesia was dissolved in a fused salt which, like cryolite, did not participate in the electrolytic process. A process developed in 1908 by Seward and Kügelgen utilized magnesia which was dissolved in a mixture of magnesium and lithium fluorides. A cathode of molten aluminium was used and the cell produced a magnesium-aluminium alloy. A variation of this approach was introduced by Seward in 1920 (see Figure 1.9) and was used in the 1920s by the American Magnesium Corporation: magnesia was dissolved in a bath of fused sodium, barium and magnesium fluorides and electrolysed to obtain magnesium of 99.99 per cent purity. The fundamental obstacle in such processes is the low solubility of magnesia in fluoride baths, so that it is difficult to ensure its constant replenishment. A problem encountered at Bitterfeld was caused by the reluctance of the magnesium globules produced to coalesce. To improve this, small quantities of calcium fluoride were added to the bath. This addition had no beneficial effect, however, upon the hygroscopic tendencies of the carnallite-containing baths, PART ONE: MATERIALS 116 and special techniques were required to dehydrate the fused salts so that magnesium oxychloride was not produced. It was against this difficult electrochemical background that Pistor and Moschel developed the first practical magnesium alloys, which contained aluminium and zinc for strengthening purposes and were introduced by GriesheimElektron to the general public at the International Aircraft Exhibition at Frankfurt in 1909. These alloys were used in considerable quantities by Germany during the First World War, by which time Griesheim Elektron had merged with IG Farben Industrie. After struggling with the carnallite process until 1924, Pistor and Moschel decided to synthesize a truly anhydrous magnesium chloride using the chlorine which was readily available from IG Farben. Chlorination was accomplished in large retorts where magnesite was reacted with chlorine in the presence of carbon at temperatures above 708°C, the melting point of the chloride. This accumulated as a liquid in the base of the reactor, and was transferred, still in the molten state, to the electrolytic cells. Since 1930 most of the world’s magnesium has been produced by the electrolytic route of Pistor and Moschel, now known as the IG electrolytic process. Although some technical improvements have been introduced, the general philosophy of this approach Figure 1.9: Electrolytic magnesium. A sectional view of the cell subsequently introduced by Seward in 1920 (US Pat. 1408141, 1920). In this process, operated by the American Magnesium Corporation, pure magnesia was dissolved in an electrolyte consisting of the fluorides of magnesium, barium and sodium. The cell shown used currents of 9000–13,000 amperes at emfs of 9–16 volts. The Dow process by then was producing purer and cheaper magnesium, and the American Magnesium Corporation ceased production in 1927. NON-FERROUS METALS 117 has never been seriously challenged in situations where chlorine and cheap electric power are readily available and continuous magnesium production over a long timescale is envisaged. Revival of magnesium production in Britain After the First World War, British interest in the future of magnesium was kept alive by two firms, F.A.Hughes and Co. and British Maxium, later Magnesium Castings and Products. F.A.Hughes was led by Major Charles Ball, who originally established contact with IG Farben and Griesheim Elektron while he was still in the post- war Control Commission. F.A.Hughes and Co. gradually developed the market for magnesium, although the demand was not great until the mid- 1930s when industry began to revive. By that time it was obvious that Britain would soon require an indigenous source of magnesium, and in 1935 F.A.Hughes, supported by the British government, acquired the British and Commonwealth rights to the IG patents covering the Pistor/Moschel process of magnesium production, and then, in partnership with IG and ICI, set up Magnesium Elektron Ltd (MEL) as an operating base. The site of the works at Clifton Junction, near Manchester, was determined by its proximity to the ICI works from which the large quantities of chlorine needed for the production of anhydrous magnesium chloride could be obtained. The Clifton Junction plant began production in December 1936. It was initially intended to produce 1500 tons of magnesium a year, although by government intervention in 1938 this capacity was increased to 4000 tons per annum. In 1940 another unit of 5000 tons per annum was added, and a further plant, capable of producing 10,000 tons per annum of magnesium commenced production at Burnley in 1942. The MEL production process soon began to use sea-water-derived magnesia as a raw material. Before 1938 it had been dependent upon imported magnesite. The sea-water magnesia was produced by the British Periclase Company at Hartlepool, by treating sea water with calcined dolomite so that both constituents precipitated magnesia. During the period 1939–45, 40,000 tonnes per annum of magnesia were produced in this way for the UK. Magnesium in North America Magnesium appears to have first been manufactured in the United States by an offshoot of Edward Sonstadt’s Magnesium Metal Company in Salford. The American Magnesium Company (AMC) of Boston, Massachusetts, was incorporated by Sonstadt on 28 April 1865, shortly after the grant of his US PART ONE: MATERIALS 118 Patent, to which the company was given an exclusive licence. Sonstadt himself appears to have been the head of this company, which continued to produce magnesium by the sodium reduction process until it ceased operations in 1892, two years after the demise of the parent company in Lancashire. Magnesium production in the United States then ceased until 1915 when a number of prominent American companies, such as General Electric, Norton Laboratories and the Electric Reduction Company, were persuaded to produce some part of the magnesium requirements of the Allies. American production in 1915 totalled 39 tons. Three producers were involved in the production of magnesium stick and ingots which sold for $5 per lb. The Dow Chemical Company started to produce magnesium in 1916. By 1920 only Dow and the American Magnesium Corporation remained in operation. The AMC, then a subsidiary of the Aluminium Company of America, operated the oxide/fluoride process developed by Seward and Kügelgen. AMC went out of business in 1927, leaving Dow as the sole primary producer of magnesium in the United States. The Dow magnesium process Dr Herbert Dow initially established his company to extract the alkali metals sodium and calcium, together with the gases chlorine and bromine, from the brine wells of Michigan. No outlet was found for the magnesium chloride solutions which were originally run to waste. As the demand for magnesium began to increase, however, Dow devised a very elegant method of making anhydrous magnesium chloride in which the hydrated chloride was dried partially in air and then in hydrochloric acid gas, which inhibited the formation of the oxychloride. The anhydrous chloride thus obtained was then electrolysed in a fused salt bath. This magnesium extraction process, which ran as an integral part of the existing brine treatment operation, produced magnesium so cheaply that Dow soon emerged as the major US manufacturer. In 1940, Dow moved his magnesium plant from Michigan to Freeport in Texas where anhydrous magnesium chloride was obtained from sea water by processes similar to those used by MEL in Britain. The Texas site had the advantage of unlimited supplies of natural gas, which allowed it to produce magnesium very cheaply indeed. The strategic significance of magnesium as a war material had become very evident by 1938, and between 1939 and 1943 the United States Government financed the construction of seven electrolytic and five ferrosilicon plants so that any foreseeable demands for the metal could be satisfied. The biggest plant, Basic Magnesium Inc., was jointly built and managed by Magnesium Elektron from Clifton Junction and by the US Company Basic Refractories. This plant, in the Nevada Desert, used power from the Boulder Dam, and had NON-FERROUS METALS 119 a rated capacity of 50,000 tonnes per annum of magnesium. All these government-financed plants ceased production in 1945, leaving the Dow Chemical Company unit at Freeport as the sole US manufacturer, although some were reactivated for short periods during the Korean war. Thermochemical methods of magnesium production Magnesium, unlike aluminium, is a fairly volatile metal which at high temperatures behaves more like a gas than a liquid. It can, therefore, be obtained from some of its compounds by thermochemical reduction processes to which alumina, for example, would not respond. Because of its volatility, magnesium is able to vacate the vicinity of a high temperature reaction zone as soon as it is liberated, and this provides the reduction mechanism with a driving force which is additional to that provided by the reduction capabilities of carbon or of a reactive metal. Some of these processes are capable of producing high quality magnesium directly from magnesite without consuming vast quantities of electric power, although the total energy requirements of such processes are usually greater than the electrolytic technique, and they tend also to be highly labour intensive. During the war years, when magnesium was urgently required, the economics of thermal reduction processes were of less importance than the simplicity of the plant required and its ability to produce metal quickly, using indigenous resources without the benefit of hydroelectric power. Several processes were pressed into service and some proved surprisingly effective, although few survived into the post war era. The ferro-silicon reduction process was developed at Bitterfeld by Pistor and his co-workers in parallel with electrolytic routes to magnesium production. As with the electrolytic process, British and Commonwealth rights to the IG Ferrosilicon process were acquired by MEL at Clifton Junction in 1935. The process depended upon the reaction which occurs between dolomite and silicon, in which, as indicated below: 2 Mg O Ca O+Si → 2 Mg+(CaO)2 Si O 2 In this reaction, the tendency of silicon to reduce magnesia is assisted by the high affinity of lime for silica. Because the product of the reaction, calcium silicate, is solid, it has little tendency to reoxidize the liberated magnesium, which escapes from the reaction zone as a superheated vapour and is then rapidly condensed directly to the solid state. Rocking resistor furnaces were used at Bitterfeld to obtain magnesium from dolomite in the 1930s. The coaxial resistance element provided temperatures in the reaction zone of 1400°C and the furnace was run 1n a vacuum. Ferro- silicon was found to be the cheapest and most effective reduction agent, and the magnesium liberated from the reduction zone condensed directly to the solid state in a cooled receiver at the end of the furnace. Furnaces of this type, PART ONE: MATERIALS 120 which produced about a tonne of magnesium a day, were built by IG Farben for the Italian government and used at Aosta until 1945. Magnesium is still produced from dolomite in Italy by variants of the ferrosilicon process. The ferrosilicon process was also used, very successfully, by the Allies during the Second World War, more particularly in North America. The process introduced by M.Pidgeon in Canada in 1941 produced high purity magnesium from dolomite by reacting it with ferrosilicon in horizontal tubular metal retorts which in the first pilot plant were only about 10cm in diameter. These retorts were disposed horizontally, as shown in Figure 1.10, in a manner which resembled closely that of the Belgian zinc retort process (see p. 93). A plant built by the Dominion Magnesium Company at Haley, Ontario, began to produce magnesium at a rate of about 5000 tonnes per annum in August 1942, and five similar plants were subsequently established in the United States. Figure 1.10: One of the earlier experimental retorts used by M.Pidgeon in Canada in 1941 for his thermo-chemical magnesium production process. In this approach dolomite was reduced by ferrosilicon, in nickel-chromium retorts, in vacuo, at temperatures around 1100°C. The magnesium vapour which distilled from the reaction zone condensed, as indicated, on the inside of a tube maintained at 450°C. Gas pressure inside the retorts were kept below 0.2 mm of Hg. From L.M.Pidgeon and W.A.Alexander, Magnesium, American Society for Metals, 1946, p. 318. With permission. NON-FERROUS METALS 121 Magnesium can be obtained by the direct reduction of magnesite with carbon, although at atmospheric pressures the reduction does not commence at temperatures below about 2000°C. The magnesium escapes from the reaction zone as a superheated vapour, and to prevent reoxidation it must then be rapidly cooled, when a fine pyrophoric powder is obtained. The reaction involved in the carbo-thermic process is as indicated by the equation below completely reversible. MgO+C Mg+CO If, therefore, the magnesium vapour released from the reaction zone is allowed to establish contact with carbon monoxide at a lower temperature, it will reoxidize. In this respect, therefore, the vapour of magnesium resembles that of zinc. After carbo-thermic reduction the magnesium vapour must be rapidly cooled and a variety of approaches to this difficult problem have been explored. Most of the plants which attempted to operate the carbo-thermic process used variants of the approach patented by Hansgirg in Austria in the 1930s. This involved the reaction of magnesite with carbon in a carbon arc furnace at temperatures which exceeded 2000°C. The magnesium vapour leaving the furnace was rapidly cooled by a recirculating curtain of gas or by an appropriate organic liquid, and was obtained in metallic form as a fine pyrophoric powder. A good deal of magnesium was made this way in Austria by the OsterrAmerick Magnesit AG before and during the Second World War. In 1938 the Magnesium Metal Corporation was jointly established by Imperial Smelting and the British Aluminium Company to produce magnesium by this approach. The factory, at Swansea in Wales, produced magnesium at a high cost and was shut down in 1945. The magnesium powder produced was handled under oil to avoid pyrophoric combustion, and required pelleting before it could be melted down. A large carbo-thermic plant built privately by Henry Kaiser during the Second World War at Permanente, California, had a rated capacity of nearly 11,000 tonnes of magnesium per annum. The magnesium vapour leaving the reaction furnaces at this plant was condensed in a curtain of oil. The lethal potentialities of this type of mixture were soon appreciated. Under the trade name of Napalm it was soon in great demand. Although the bulk of the world’s requirement of magnesium is still produced electrochemically, thermochemical reduction processes continue to attract a good deal of research effort and it is possible that a cheaper and more viable alternative to electrolysis will eventually emerge. AGE HARDENING ALLOYS At the beginning of the twentieth century, steels were the only alloys which were intentionally strengthened by heat treatment. In 1909, however, Dr Alfred ← → . tons per annum. In 1940 another unit of 5000 tons per annum was added, and a further plant, capable of producing 10,000 tons per annum of magnesium commenced production at Burnley in 1942. The MEL. required, the economics of thermal reduction processes were of less importance than the simplicity of the plant required and its ability to produce metal quickly, using indigenous resources without the. explored. Most of the plants which attempted to operate the carbo-thermic process used variants of the approach patented by Hansgirg in Austria in the 1930s. This involved the reaction of magnesite