PART ONE: MATERIALS 122 Wilm noted that an aluminium alloy he was working on had the remarkable ability to harden slowly at room temperature after having previously been quenched from a temperature just below its melting point. The effect was observed, quite fortuitously, while Wilm, working on a contract from the Prussian government, was attempting to develop alloys for cartridge cases which were lighter than the 70/30 brass which was normally employed. Between 1909 and 1911 Wilm filed several Patent Applications, and assigned his rights in this ‘age-hardening’ invention to the Durener- Metallwerke in Duren, who subsequently marketed the alloys under the trade name Duralumin. During the First World War large quantities of age-hardened aluminium alloys were used by the combatants, first for Zeppelins and then for other types of aircraft. It was found that the strengthening reaction could be slowed down, very significantly, by refrigeration. This made it possible to quench aluminium alloy rivets, and store them in the soft condition in a refrigerator. The age hardening process began only after the rivet head had been closed after insertion into the aircraft structure. A great deal of the metallurgical research stimulated by the introduction of Duralumin was co-ordinated in Great Britain from 1911 by the Alloys Research Committee of the Institution of Mechanical Engineers. The 11th report of this committee, published in 1921, summarized eight years of work by the National Physical Laboratory (NPL). The outstanding result of this work was the development of ‘Y’ alloy (4 per cent Cu, 2 per cent Ni, and 1.5 per cent Mg), which retained its strength to moderately high temperatures and was extensively used for pistons and cylinder heads. Y alloy was originally used in the cast form. High Duty Alloys was established in 1928 by Colonel W.C. Devereux to produce variants of Y alloy in the forged condition for aircraft engine use. From these experiments stemmed the RR series of light alloys, jointly developed by Rolls Royce and High Duty Alloys. The first rational explanation of the age hardening process in light alloys such as Duralumin and Y alloy was provided in 1919 by Paul Merica and his colleagues Waltenberg and Scott at the National Bureau of Standards in Washington. They found that when an alloy such as Duralumin was heated to temperatures close to its melting point most of the alloying constituents were taken into solution by the matrix. Quenching retained the dissolved metals in this supersaturated solid solution which was, however, somewhat unstable at room temperature: minute crystals of various intermetallic compounds were slowly precipitated. Provided these crystals were below a certain critical size, invisible to the optical microscope, they strained and distorted the aluminium lattice and acted as mechanical keys, which inhibited plastic flow in the alloy and increased its strength. The aluminium alloys which behaved in this way were unique only in the sense that the precipitation effects manifested themselves, quite fortuitously, at room temperature. A wide range of other alloys were soon identified in which precipitation could be induced by ageing NON-FERROUS METALS 123 at elevated temperatures and new precipitation hardening alloys of all types were rapidly developed. In 1934 Dr Maurice Cook of ICI was able to review the precipitation hardening characteristics of 37 copper alloy systems. Beryllium and beryllium alloys Beryllium, first identified as an element by Wöhler in 1828, was at first regarded merely as a chemical curiosity. Soon after the introduction of the incandescent gas mantle by Auer von Welsbach in 1885 it was found that the strength and durability of the ash skeleton could be greatly improved by adding small quantities of beryllium nitrate to the mixtures of thorium and cerium nitrates used to impregnate the fabric of the mantle. This treatment was probably the major outlet for beryllium until the end of the First World War. Attempts to produce beryllium by the electrolysis of a fused bath of beryllium chloride were first made in 1895 by Wilhelm Borchers. Goldschmidt in 1915 found that fused fluoride baths offered better production prospects. This approach, refined by Stock, Praetorius and Priess, yielded relatively pure beryllium for the first time in 1921. The properties of the metal obtained confirmed theoretical predictions that beryllium would be a light metal with a density around 1.8 grams per cm 3 and that 1ts direct modulus of elasticity would be significantly higher than that of steel. If its ductility could be improved, it was likely to be the light, stiff metal which had long been sought by the aircraft industry. The alloying behaviour of beryllium was studied by Regelsberger in 1924. Additions of up to 10 per cent by weight of beryllium improved the hardness and tensile strength of magnesium without making it brittle. Up to 10 per cent of beryllium could also be dissolved in copper to produce a pale yellow alloy. The colour of the alloy improved as the beryllium content decreased, and Regelsberger was the first to mention the beautiful golden colour of copper containing around 2 per cent of beryllium. Research on beryllium in the United Kingdom was initiated by the Imperial Mineral Resources Bureau at the Metallurgy Department of the NPL in 1923. By 1926 Dr A.C.Vivian had produced solid deposits of metal by the electrolysis of fused baths of beryllium and sodium fluorides similar to those used by Stock, Praetorius et al., whose detailed results had been published the previous year. The NPL beryllium was 99.5 per cent pure and, like the German product, was completely brittle when cold. By 1926 popular interest in beryllium had started to develop. With a density equivalent to that of magnesium, a stiffness higher than that of steel, and a melting point approaching 1300°C, this seemed destined to become the light metal of the future. Towards the end of 1926, when facilities for the production of relatively pure beryllium existed in Germany, the USA and the PART ONE: MATERIALS 124 United Kingdom, it was estimated that the metal could, if required, be produced for approximately 205 per lb. The electrolytic beryllium available in 1927, although reasonably pure, had a Brinell hardness of 140 and was still brittle when cold. In 1928 when the NPL improved the purity to 99.8 per cent, beryllium was still brittle and no improvement was noted when the purity level was gradually increased to the 99.95 per cent level. By 1932 attempts to produce ductile beryllium at the NPL were abandoned, and it was generally concluded, throughout Europe and the United States, that beryllium would in future find its major use as an alloying ingredient where the problems of purity and ductility were not so critically involved. Beryllium copper and beryllium nickel alloys were first produced commercially for Siemens and Halske by Dr W.Rohn at Heraeus Vacuumschmelze at Hanau, which began to operate just after 1923 at the height of the inflationary period in Germany. Also in the early 1920s, Michael Corson in the United States was pursuing similar lines of development. In 1926 he patented the composition of an age hardening beryllium-nickel-copper alloy, while Masing and Dahl of Siemens also protected the compositions of beryllium copper and beryllium cobalt copper alloys which had higher electrical conductivities than those of the American alloy and were therefore of greater industrial potential. Beryllium copper displayed a degree of age hardening far higher than that of any other copper-based alloy, and in spite of its cost beryllium copper soon found extensive industrial applications. It was particularly valuable as a spring material: the alloys were soft and ductile in the solution treated conditions and could then be fabricated into complex shapes. After heat treatment at temperatures between 300 and 350°C the best alloys developed tensile strengths approaching 100 tons per square inch, and as a result of the precipitation which had occurred the electrical conductivities improved to about 40 per cent that of pure copper. Even higher mechanical properties were obtained with beryllium nickel alloys, although these were not suited to electrical work because of their higher resistivity. DEVELOPMENT OF HIGH TEMPERATURE ALLOYS One effect of the rapid introduction of mains electricity into the domestic environment was an increasingly urgent requirement for improved alloys for electrical heating applications. The alloy required was one which combined a high electrical resistivity with extreme resistance to oxidation at high temperatures and mechanical properties high enough to ensure that it did not fail by creep after prolonged use at a good red heat. Prior to 1900 the only high resistivity alloys available were cupronickels such as Ferry and Constantan and iron alloys containing up to 20 per cent nickel. NON-FERROUS METALS 125 Nickel-chromium alloys The first satisfactory electrical heating alloys, introduced by A.L.Marsh in 1906, were based on the nickel-chromium and nickel-chromium-iron systems. Wires of these alloys had an electrical resistivity around no microhms per cm 3 , more than twice that of the best cupro-nickel alloys. They were, moreover, far more resistant to oxidation and stronger at high temperatures. Previous investigators had found that chromium additions tended to increase the oxidation rate of nickel. Marsh found that chromium additions in excess of 10 per cent rapidly decreased this rate and that alloys containing around 80 per cent nickel and 20 per cent chromium showed a good balance between oxidation resistance and resistivity. They were also ductile enough to be drawn into wire. It is now known that the oxidation resistance of alloys of this composition depends on the formation of a protective oxide layer. These alloys were initially induction melted in air and deoxidized with manganese. Air melted material was cheap to manufacture, although it was not always easy to draw into fine wire. At the end of the First World War, Siemens and Halske found that vacuum melted nickel-chromium alloys were easier to draw into wire and also had a longer high-temperature working life, compensating in some degree for the added expense of vacuum melting. Heraeus constructed at Hanau in 1921 a 300kg (660lb) capacity low frequency vacuum melting furnace of the Kjellin type for melting nickel-chromium based resistance alloys. By 1928 two 4000kg (8800lb) furnaces capable of casting two 2000kg (4400lb) nickel-chromium ingots were in regular operation. Siemens and Halske acquired Heraeus Vacuumschmelze in 1936, after which considerable quantities of nickel-chromium alloys, beryllium copper, and other specialized materials for the electrical industry were produced. The rare earth effect In the early 1930s it was found that some of the vacuum-melted alloys which had been processed from melts deoxidized with mischmetall (a rare-earth mixture) were remarkably resistant to high temperature oxidation. Rohn was able to show that the metal responsible for this effect was cerium, and in 1934 Heraeus applied for patents covering the manufacture and use of heating elements to which small quantities of cerium and other metals of the rare earth group had been added. This ‘rare earth effect’ was followed up in the Mond Nickel Laboratories at Birmingham, and led to the introduction of the well- known ‘Brightray C’ series of alloys. The mechanism which permits small quantities of the reactive metals, such as cerium, yttrium, zirconium, lanthanum and hafnium to improve the protective nature of the oxides which PART ONE: MATERIALS 126 form on the surface of nickel chromium and other high temperature alloys is still imperfectly understood, although the effect is widely employed. Aluminium containing nickel-chromium alloys It was found in 1929, by Professor P.Chevenard of the Imphy Steelworks, that small quantities of aluminium improved the oxidation resistance and high temperature strength of nickel-chromium alloys and made the alloys responsive to age-hardening. In 1935 he showed that the strengthening effect of aluminium could also be augmented by small quantities of titanium. When, around 1937, Britain and Germany both began to develop a workable gas turbine for aircraft propulsion, the main technical problem they encountered was that of producing turbine rotor blades which were strong enough at high temperatures to withstand the high, centrifugally imposed tensile stresses. The alloy selected was the 80/20 nickel-chromium solid solution alloy, strengthened in accordance with Chevenard’s findings by small quantities of titanium and aluminium. Work on this material started in 1939, and the first production batch was issued in 1941. Nimonic 80, as the alloy was called, was the first of the long series of nickel-based ‘superalloys’ produced by Mond. All of these depend for their high temperature strength upon substantial quantities of the precipitated ? (gamma prime) phase which, although based on the compound Ni 3 Al, also contains a good deal of titanium. The first Nimonic alloys were melted and cast in air and hot forged. As the alloys were strengthened and improved, however, very reactive alloying constituents were being added, and vacuum melting became mandatory. By 1960, most of the stronger alloys were being worked by extrusion rather than by forging. The use of molten glass lubrication, introduced by Sejournet in the 1950s, made it possible to extrude even the strongest Nimonic alloys down to rod as small as 20mm in diameter in one hot working operation. The alloys were progressively strengthened, initially by increasing the aluminium content and subsequently by additions of the more refractory metals such as tungsten and molybdenum. Nimonic 115, the last of the alloys to be introduced, marked the practical limit of workability. Stronger and more highly alloyed materials could not be worked and by 1963 a new generation of nickel-based superalloys was being developed. There were not worked, but were cast directly to the shapes required. Since the early years of the century it had been known that the grain boundaries of metals, although initially stronger than the body of the grains, tended to behave in a viscous manner as the temperature increased and lost their strength far more rapidly than single crystals. The concept of an equi- cohesive temperature, applicable to any alloy, above which the grains themselves were stronger than the boundaries and below which the boundaries were stronger than the grains, was first advanced in 1919 by Zay Jeffries. This NON-FERROUS METALS 127 suggested that the strongest high temperature alloys would have large rather than small grains. Difficulties were, however, caused on odd occasions by very large randomly orientated grains. After 1958 it became customary to apply a thin wash of cobalt-aluminate to the interior of the moulds in which the turbine blades were cast. This coating helped it to nucleate a uniform grain size in the castings. It also seemed logical to produce blades having as few grain boundaries as possible aligned at right angles to the axis of the tensile stress to which the blade would eventually be subjected. Methods of directionally solidifying superalloys in such a way that the structures obtained formed a bundle of longitudinally aligned columnar crystals were first described in 1966 by B.J.Piercey and F.L. Versnyder of the United Aircraft Corporation. This process, now known as directional solidification, resulted in an immediate improvement in high temperature blade performance even without change in alloy composition. Single crystal blades, which contained no boundaries, were soon produced as a logical development of the directional solidification concept. Because grain boundary strengthening additions such as hafnium were no longer required in single crystal blades, and tended to interfere with the perfection of their growth, single crystal turbine blades are now manufactured from alloys which have very much simpler compositions than the conventional casting alloys. Cobalt-base high temperature alloys While the British were using the 80/20 nickel-chromium resistance wire alloy as a base for their first turbine blade material American manufacturers were adopting a completely different approach to the problem of high temperature alloy strength. The blades and other components of gas turbines used for driving the superchargers of large piston engines had, for a considerable time, been very effectively and economically manufactured by casting them from cobalt chromium alloys. Alloys based on the cobalt chromium system had originally been developed by Edward Haynes. His original ‘Stellite’ alloy, which would ‘resist the oxidizing influence of the atmosphere, and take a good cutting edge’ was patented in 1909. Further improvements in hardness were obtained by tungsten and/or molybdenum additions. Although originally intended for dental and surgical instruments, the alloys soon found a considerable industrial market when it was found that they could be used for heavy turning operations. During the First World War they were used extensively by the Allies for shell turning, and in 1918 approximately four tonnes a day of the alloy were being cast. The Haynes Stellite Company was acquired by Union Carbide in 1920, and the alloy Vitallium was developed in the late 1920s. This PART ONE: MATERIALS 128 was successfully cast into turbine blades used to supercharge the engines of the Boeing Flying Fortress. The American engine manufacturers understood that no long-term future existed for workable turbine blade alloys: the high alloying content needed to achieve high temperature strength inevitably lowered the melting point of such alloys and simultaneously reduced their workability, and it was illogical that an alloy specifically designed to withstand creep failure at high temperature should also be expected to undergo severe plastic deformation during manufacture. From the very beginning of superalloy development, therefore, turbine blades were produced in the United States by vacuum melting and investment casting. The alloys employed were based on the cobalt chromium system, being strengthened, not by the ‘gamma prime’ phase on which the British nickel base alloys depended, but by the presence of substantial quantities of stable carbides. The way in which the high temperature capabilities of superalloys has increased with time is illustrated diagrammatically in Figure 1.11. It is significant that the safe operating temperature for the best single crystal alloys is only marginally higher than that of directionally solidified material. In view of the limitations imposed by the known melting points and oxidation resistance of existing superalloys it seems unlikely that new alloys having high temperature capabilities greatly superior to those currently available will be developed. Improved gas turbine performance will, most probably, result from the exercise of engineering rather than metallurgical ingenuity. POWDER METALLURGY Prehistoric iron must have been the first metal to have been consolidated from a spongelike mass (see Chapter 2). The legend of Wayland the Smith, which appears to date from the fifth century AD, obviously embodies a good deal of the folklore on which the whole fabric of powder metallurgy has since been erected. Various accounts describe him as making a steel sword by conventional blacksmithing techniques, then reducing it completely to filings, which were mixed with ground wheat, and fed to domestic geese or hens. The droppings of these birds were collected, reheated in his forge, consolidated, and then hammered into a sword. This again was reduced to filings and the process repeated, the end product being a weapon of unsurpassed strength, cutting power and ductility. Apparently even the early smiths were aware that metallurgical quality could be improved by subdivision, and that by repeating this process of subdivision a product would be eventually obtained which would combine hardness and ductility. The improvements which ferrous metallurgists have attributed to the passage of steel filings through the alimentary canal of a NON-FERROUS METALS 129 domestic fowl are more difficult to account for, however, as it is difficult to believe that this would reduce the phosphorous content of the metal to any significant effect. It is possible, however, that the high ammonia content of bird droppings might help to nitride the metal particles during the early stages of consolidation, although in the absence of reliable confirmation of this effect it seems logical to attribute the effectiveness of Wayland’s working approach to the way in which it was able to progressively refine an initially coarse metallurgical structure. The use of powder metallurgy by Wayland the Smith was obviously one approach to the problem which all swordsmiths have encountered: swords which were strong and hard had a tendency to break in service, while those which were free from this defect bent too easily and did not retain their cutting edge. Wayland’s technique must have produced a refined crystal structure which, like that which existed in Samurai blades, combined strength and ductility with the ability to maintain a fine cutting edge. It provides an effective demonstration of the ability of powder metallurgy to improve the quality of an existing product even when the metals involved could, if required, be melted together. Powder metallurgical techniques were used in pre-Columbian times by the Indians of Ecuador to consolidate the fine grains of alluvial platinum they were unable to melt. These platinum grains were mixed with a little gold dust, and heated with a blow pipe on a charcoal block. The platinum grains were, therefore, soldered together by a thin film of gold and it was possible by this method to build up solid blocks of metal which were malleable enough to withstand hot forging, and were fabricated into items such as nose rings and other articles of personal adornment. Although platinum was a metal unknown to the ancients, fragments of this metal and its congeners (elements belonging to the same group in the periodic table) were unwittingly incorporated into many of the earliest gold artefacts from the Old and New Worlds. The first important source of the metal was Colombia on the north-western corner of South America. Here, the grains of native platinum were regarded initially as the undesirable component of South American gold mining operations. The identity of platinum as a new element was recognized in 1750, by William Brownrigg. In 1754, William Lewis of Kingston upon Thames found that the grains of platinum could generally be dissolved in aqua regia, and that the ‘beautiful brilliant red powder’ now known to be amino-platino-chloride was obtained when sal ammoniac was added to the aqua regia solution. This precipitate, when calcined, decomposed to provide a dark metallic powder which appeared to be pure platinum. No satisfactory method of melting this platinum powder or sponge was found, however. The first truly ductile platinum was produced in 1773 by Rome Delisle, who found that if the platinum sponge, after calcination, was carefully washed and then reheated in a refractory crucible it sintered to form a dull PART ONE: MATERIALS 130 Figure 1.11 NON-FERROUS METALS 131 grey mass which could then be consolidated by careful forging at a good red heat. The density of the mass changed, as a result of this treatment, from 10.05 to 20.12g/cm 3 . This was the first time that the high theoretical density of platinum had been demonstrated. This sintering process was described by Macquer in 1778 in his Dictionnaire de Chimie, and in doing so he provided what is probably the first interpretation of the scientific principles upon which powder metallurgy depends: ‘a metallic mass quite compact and dense, but it completely lacks malleability when it has been exposed to a moderate heat, and only assumes it when subsequently subjected to a much greater degree of heat’. He goes on to report that: Particles of platina being infinitely divided in the precipitation it is not surprising that the heat penetrates such very small molecules more effectively than the ordinary grains of platina which, in comparison, are of enormous size; and their softening occurring in proportion, they should show the extraordinary effect on their agglutination in the proportion of their points of contact; moreover, these points being infinitely more numerous than can be those of much greater molecules, solid masses result which have all the appearance of quite dense metal, melted and solidified on cooling, but they are nothing but the result of a simple agglutination among an infinite number of infinitely small particles, and not that of a perfect fusion as with other metals. This powder metallurgy route was subsequently refined and applied by William Hyde Wollaston who in 1800 at the age of thirty-four retired from medical practice and decided to make his fortune by producing and selling ductile platinum on a commercial scale. Wollaston was a friend of Smithson Tennant, who between 1803 and 1804 had shown that the black residue which was left behind after native platinum had been dissolved in aqua regia contained the two hitherto unknown noble metals, osmium and iridium. The presence of Figure 1.11: Improvement in the high temperature capabilities of nickel based superalloys improved since their introduction in 1940. The strongest wrought alloy to be developed was Nimonic 115 which was introduced in 1960. Stronger alloys could not be worked, and cast alloys were then introduced. These were subsequently improved, firstly by directional solidification and finally by turning them into single crystals The upper curve on the diagram shows how the solidus of the workable alloys began to decrease rapidly as the alloys were progressively strengthened by alloying. By 1960 the gap between the melting point of these alloys and their working temperature capability had decreased to about 100°C. It was found, however that single crystal alloys could be based on very simple compositions, since no alloying additions were required to impart grain boundary strength. As these alloys developed, therefore, their melting points began to increase, thus providing a much needed margin of high temperature safety. . than those of the American alloy and were therefore of greater industrial potential. Beryllium copper displayed a degree of age hardening far higher than that of any other copper-based alloy, and. into many of the earliest gold artefacts from the Old and New Worlds. The first important source of the metal was Colombia on the north-western corner of South America. Here, the grains of native. stronger than the body of the grains, tended to behave in a viscous manner as the temperature increased and lost their strength far more rapidly than single crystals. The concept of an equi- cohesive