Platinum Metals Rev., 1958, 2, (2), 55
The History of the Melting of Platinum
Two Hundred Years of Progress
Those who came early into contact with native platinum, when the Spaniards were finding it in New Granada in the sixteenth century, were at once struck by its apparent infusibility, coupled with a metallic appearance. As it did not melt at any degree of heat available at the time, they decided that it could not be a metal and must therefore be classed as a “semi-metal” until someone melted it. This feat was achieved just two hundred years ago, in Paris, by the distinguished French chemist Pierre-Joseph Macquer and his assistant Antoine Baumé, who, by means of a large concave “burning-mirror”, managed to concentrate the heat of the sun on to a specimen sufficiently to produce partial fusion, yielding a brilliantly metallic silver-white substance that was fully malleable and could be hardened by working and softened by annealing just as other recognised metals.
This showed that the substance could be melted if a sufficiently high temperature could be achieved and maintained and, over the next generation, many attempts were made to boost up the available furnaces to produce higher and higher temperatures to that end. These efforts usually resulted in melting the furnaces and the crucibles without having much effect on the platinum, but there is no doubt that a few of the investigators, such as Guyton de Morveau, did succeed in bringing about melting. It is however noticeable that when these people were successful they always used considerable quantities of carbon, or of material that yielded carbon at a high temperature, and there can be little doubt that the fusion was assisted by the introduction into the metal of small quantities of impurity sufficient to lower its melting point so that the available heat could melt it. The effective elements were probably phosphorus or silicon or both, reduced by the carbon from phosphates or silicates in the crucible or the fluxes or in the ash of the carbon itself. This conclusion is reinforced by the fact that the metal was almost always brittle.
The first complete melting of platinum was a result of the discovery of oxygen, and was achieved by Lavoisier. He reported it in 1783 in the course of a series of attempts to melt a number of substances, hitherto accepted as infusible, by heating them on charcoal in a blast of the new gas. With platinum he obtained full fusion into rounded globules, but he found difficulty in keeping the piece of metal molten if it exceeded 12 to 15 grains in size.
Soon after this another method came into use when the development of electricity made possible currents heavy enough to raise the temperature of platinum wires to the melting point.
Hare’s Oxy-hydrogen Blowpipe
The next advance after Lavoisier’s in the technique of melting platinum came with the introduction of the oxy-hydrogen blowpipe. The blowpipe, as a mouth instrument, came down to us from prehistoric times and the fitting to it of bladders of air and simple air-pumps goes back a long way, but the use of oxygen and hydrogen was not simple because of the ease with which the mixture explodes. Nevertheless, by the second decade in the last century practical instruments were available, and two developments fall for description here. The earlier occurred in 1801 when Robert Hare, the son of a brewer of Philadelphia, read to the Chemical Society of that city a paper describing how he could feed a blowpipe with a continuous current of air or other gases by storing them in vessels from which they were slowly expelled by hydrostatic pressure, to be ignited at a common orifice. This system enabled oxygen and hydrogen to be used, and with them Hare was able to melt on cupels a series of extremely refractory substances, including of course platinum. From the winter of 1802 Hare collaborated with Benjamin Silliman and there were further communications on the blowpipe in 1803 and 1804. After that the two men seem to have gone on to other work and nothing more is heard of their melting of platinum until 1836, when Hare took the work up again and was able to produce pieces of metal of quite considerable size; one of 28 ounces is mentioned. He also managed to melt iridium and rhodium.
There is no evidence that Hare at any time produced melted platinum for sale, but by 1842 Joaquim Bishop, who had been instrument maker in the University of Pennsylvania under Hare, and had built his apparatus, put the blowpipe to commercial use in the business which he set up in Malvern, Pennsylvania. Afterwards he specialised in the making of platinum apparatus and was the first man to use the melted metal in this way. The business that he founded has continued to do so until the present day.
Sainte-Claire Deville and the Lime-Block Furnace
The next man to enter the story is Henri Sainte-Claire Deville of Paris. He had interested himself in the production of the then rare metal aluminium, and had achieved success by 1855. The work led him on to study refractory substances and to seek to melt those metals which were not easily fusible by the means then available. To do this he turned his attention to the burning of coal-gas and oxygen in a blowpipe, and by 1857 he and his assistant, H. Debray, had succeeded in melting platinum in considerable quantity in a simple apparatus consisting solely of two blocks of quicklime. Each was hollowed out in a saucer-shaped depression so that, when one block was placed on the other, the two saucers formed an enclosed space in which solid platinum could be melted by means of a coal-gas/oxygen flame introduced through a hole in the top. At the end of the operation the whole apparatus could be tilted and the molten metal poured out through a hole in the side.
From that time, melted metal became freely available in platinum refining, and thereafter all newly-formed fabricating firms used it alone for their work. The old established businesses did not, however, change over to it entirely for many years, since they found in the forged metal certain advantages which will be referred to later in this article. But something else happened to give the melting process its own firm niche in the industry, and that was the demonstration in 1862 by Deville and Debray of the important hardening of platinum brought about by alloying it with various amounts (normally ten per cent) of iridium. The production of this and other alloys could not readily be achieved by the forging process, although some producers did make a success of co-precipitation, and melting had to be employed to an increasing extent to this end alone.
But although the practice of melting came into use more and more as the years went on, at the same time shortcomings in connection with the lime-block method began to appear. In the first place it became difficult to locate suitable lime that would stand up to the high temperatures without cracking or even breaking up. Next, the precise control of the composition of the gas mixture was difficult and called for slight but definite differences at various stages of the process. If the composition erred on the reducing side, then the metal took up traces of impurities from the constituents of the lime, in the form of small quantities of calcium and magnesium.
In the early days these did not matter, but later the demand for higher and higher purity brought out the fact that these traces of impurity appreciably raised the hardness of the metal, and seriously altered some of its electrical properties, especially those used in its application to thermocouples. The third defect of the lime-block furnace was that if the composition of the gases was on the oxidising side, then the metal produced contained gas inclusions which vitiated its mechanical properties when required for very thin foil. The fourth fault was the impossibility of getting really efficient mixing of metal and alloy under the closed-up conditions of the operation. The final fault of the process was that the direct impingement of a high-pressure flame on the surface of the molten metal gave rise to losses in fume, which again were not important when the price of platinum was £1 per ounce, but became increasingly serious as it rose towards (and over) its present level. All this explains why the old fabricators were very loth to part with the forging process altogether and why, by the nineteen-twenties, there was a growing need for a major improvement in the method of platinum melting.
The High-frequency Induction Furnace
This came promptly in 1920 in the United States in the form of the high-frequency electric induction furnace invented by Dr E. F. Northrup and made by the Ajax Electrothermic Corporation. In this extremely simple but most ingenious piece of apparatus use was made of the property of electromagnetic induction, as in the ordinary electric power transformer. In this case the metal to be melted is placed in a crucible situated on the axis of a water-cooled copper helix, which carries the primary alternating current, while the secondary currents are generated inside the metal in the crucible. The electrical resistance of this metal to the rapidly alternating currents produces heat, and the key to the Ajax-Northrup achievement was to obtain sufficient heat to melt the most refractory metals. This was effected by using primary currents of very high frequency, and by careful design of the dimensions of the primary coil and its position relative to that of the crucible and its contents.
In 1921 a visit was made to the United States by S. S. Moore Ede and F. T. F. Toller of Johnson Matthey & Co. Limited, and while there they saw a very early example of the Ajax-Northrup furnace. At once they grasped its possibilities in connection with the melting of platinum and, having been assured that this could be done in crucibles of fused silica, they arranged for a furnace to be sent to them in London at the same time as one which had already been ordered by the National Physical Laboratory. These duly arrived and were the first to be brought to England, while the Johnson Matthey unit was the first to be applied to the industrial melting of platinum. In this task there was at the outset a great deal of trouble over the choice of a suitable refractory for the crucible. The softening point of fused silica proved to be too near the melting point of platinum to avoid serious distortion, and another solution of the problem had to be sought. This was not easy because the very high melting point of the metal caused the temperature gradient through the walls of the crucible to the water-cooled conditions outside to be excessively steep, and no ordinary refractory could stand up to it. Eventually, however, zircon was found to have the essential properties and the first crucible of this material for melting 100 ounces of platinum by the high-frequency method was tested in June 1922; since then it has been used generally for this work.
The Ajax-Northrup technique avoids most of the drawbacks of the lime-block furnace, with the sole exception of a slight danger of oxygen pick-up, but this is easily dealt with by suitable control of the melting conditions. The incorporation of necessary alloys or other additions is much facilitated by the fact that, when the currents have finished their melting work, they promote a vigorous “electromagnetic rotation” in the molten metal and so bring about a most efficient mixing of the charge.
Since the furnace itself is of such a simple design and the heat is generated entirely within the metal to be melted, the crucible, if suitably packed within a silica sheath, can be picked up with the gloved hand for the molten contents to be cast. In the early days the high-frequency currents were generated by means of a spark-gap operated in hydrogen in conjunction with a large bank of condensers. In the more modern equipment these currents are produced by valves somewhat resembling those used in radio transmission.
Nevertheless, it is quite evident that the high-frequency induction furnace is not necessarily the last word in the long story of the evolution of the melting of platinum and its alloys.
Recent developments in the electrical, electronic and chemical industries have shown that even the most minute traces of impurities may profoundly affect the physical and chemical properties of metals called into use today; modern research is therefore confronted with the problem of producing metals in a state of ultra-purity. This involves, in the case of platinum, rigid purification of the compounds from which the metal is prepared and careful control of the melting conditions to avoid contamination with impurities derived from the refractories and the melting atmosphere. There is little doubt, therefore, that sooner or later methods will be devised for melting platinum without using refractories and without allowing molten metal to come into contact with gases likely to lead to contamination.