The boiling point of pure gold has not been determined; calculated according to Wiebe’s formula it would be about 2,240°, or nearly 500° above the melting point of platinum. However, contrary to the belief of the older experimenters, it is sensibly volatile in air at far lower temperatures. Robert Boyle was unaware of this fact, but Homburg gilded a silver plate in 1709 by holding it over gold strongly heated in the focus of a burning mirror, and St. Claire Deville volatilised and again condensed gold when melting it with platinum. The rapid volatilisation of gold, when heated by an ordinary blowpipe, was first proved in by Dr. Robert Hare, a purple stain being thus produced on bone-ash in a few seconds. Lastly, a discharge of high-tension electricity from gold points causes its volatilisation, and if the discharge is sent through a fine gold wire stretched on paper, it converts it into a purple streak of finely divided condensed particles of the metal. The rapid distillation of gold caused by heating it in a current of air of considerable velocity, such as that furnished by a blow-pipe, by which the liquid is thrown into waves, may be shown at any time by heating a fragment of the precious metal of the size of a pin’s head on a bone-ash cupel in the oxidising flame of a good mouth blowpipe. Almost immediately after the fusion is complete, a purple stain of condensed gold begins to form on the outer margin of the cupel. The author has found that a piece of gold weighing 0.5 gramme loses half its weight in an hour, if heated on a cupel by a foot-blowpipe (the temperature attained being probably less than 1,300°), and only a few minute beads are observable, detached from the main button. Alloys of copper and gold disappear much more rapidly. No doubt most of the gold passes off as spray, but perhaps part of the loss may be due to rapid volatilisation, and could not be correctly described as mechanical loss.

The volatility of gold, both when pure and when alloyed with silver and copper, has been investigated by Napier, who found that an alloy of 100 parts gold to 12 parts of copper, if kept for six hours at a temperature just high enough to keep it melted, lost 0.234 per cent, of its gold contents, and at the highest temperature attainable in an assay muffle, it lost 0.8 per cent, in six hours. An increase in the amount of copper present caused an increase in the loss of gold. In the simple operation of pouring about 30 lbs. of a gold-copper alloy from a graphite crucible into moulds, fumes were given off, of which the part condensed in a wet glass beaker held above the crucible contained 4.5 grains of gold. Napier also found that gold does not appear to volatilise so readily when alloyed with silver only, as when copper is also present.

Makins found that gold volatilises sensibly along with silver and lead, when melted with these metals in a muffle in an ordinary bullion assay. The loss of gold by volatilisation on melting its copper alloy is the common experience of mints. At the Sydney Mint it was estimated to be 0.017 per cent., or £170 per million sterling melted, and is probably seldom less than 0.01 per cent., or one part in ten thousand. At the same Mint, Leibius found that the sweepings from the top coping-stone of a chimney 70 feet high contained 1.46 per cent, of gold and 6.06 per cent, of silver. Similar results have been obtained at some other mints.

A number of experiments were recently made at the Royal Mint, with the view of determining the effect of variations in the temperature and other conditions on the volatility of gold and some of its alloys. The test pieces were heated in a muffle furnace, the temperature of which was determined by the optical pyrometer, and by the Le Chatelier thermocouple, which consists of platinum and rhodio-platinum wires. The results of some of the experiments on fine gold, and certain alloys of gold and copper, are given in the table on the next page.

The following conclusions may be drawn from the table:

1. The loss of gold on heating the pure metal rises with the temperature, being four times as great at 1,250° as at 1,100°, whilst it is insignificant at 1,075° and probably nil at 1,045°, the melting point (Violle).
2. An atmosphere consisting largely of carbonic oxide is apparently favourable to the volatilisation of gold, the rate being six times greater than in an atmosphere of coal gas at the same temperature. The increased loss when the graphite crucible was substituted for a cupel may be noticed in this connection; clay crucibles of similar shape to the graphite one have an opposite effect, the loss being less than on cupels.
3. The increase of loss of gold alloyed with copper, when the percentage of the latter metal is increased, which is observable in one of Napier’s experiments, is confirmed.

The volatilisation of gold-copper alloys merits further investigation, as it is of great industrial importance. The high percentage of the losses shown in the table is, no doubt, due to the smallness of the masses treated, and to the comparatively large surfaces exposed in consequence of this.

A number of experiments were also made at the same time on various gold alloys to determine the relative effect of the presence of other metals on the volatilisation of gold. The temperatures and conditions used were similar to those described above. It was found that the volatility of gold was increased by the presence of any metallic impurity, even by the non-volatile metals, such as platinum. The tellurium alloys suffered the heaviest losses, gold containing five per cent, of tellurium losing from 16.5 to 39.5 parts per 1,000 by volatilisation per hour at a temperature of 1,245° (i.e., 200° above the melting point of pure gold), a point of much importance in connection with the metallurgy of telluric gold. Lead and platinum had a very slight effect in increasing the volatility of gold; copper and zinc a more marked effect, while five per cent, of antimony or mercury caused losses amounting to about two parts per 1,000 of gold per hour at 1,245°. Further conclusions which were drawn were as follows:—

1. Although the temperatures employed were higher than those at which pure zinc, cadmium and tellurium, and probably antimony and bismuth, are distilled, yet these elements were never completely volatilised. The “temperature of dissociation of the alloys,” as in the case of the bismuth-arsenic alloy investigated by Edward Matthey in a paper read before the Royal Society, January 26, 1893, is much higher than the ordinary distillation point of the more volatile constituent. Thus zinc boils at 950°, but its gold alloy loses little or no zinc at 1,120°, and still retains part at 1,250°, whilst antimony was not driven off to any extent by the highest temperature attained. Copper appears to pass off more easily than some of the so-called volatile metals, perhaps because the dissociation of its alloy with gold may not form an initial stage of the operation.
2. The amount of gold lost depends partly on the volatility of the alloying metal, but perhaps too much stress has been laid on this factor in the past, the results of heating alloys of mercury, zinc, antimony and copper pointing to that conclusion. A metal with a strong attraction for gold, such as copper, may carry it off, perhaps as a vaporised alloy, more easily than one which mixes with it less intimately.
3. It is observable that those impurities which reduce the surface tension of a button of liquid gold appear to increase, the vapour pressure of the metal as indicated by the loss on heating. This was to be expected.
4. A current of air or coal gas insufficient to disturb the surface of the liquid metal does not appear to increase the volatilisation.

It may be added that Hellot stated that if an alloy of one part of gold and seven parts of zinc is heated in air, the whole of the gold comes off in the fumes.