Nitre Refining

When the molten gold is obviously impure nitre is added, and as a rule its main action is to eat the pot away, for as fast as metals are oxidised they are again reduced by the carbon of the pots. Where fireclay crucibles are used the nitre does effect a small amount of oxidation, but the action is very imperfect, unless the bullion is very low grade. If the nitre is added before the retorted cakes have commenced to melt, so that each honeycombed lump of gold is penetrated by the fused salt, the action is much more perfect than when it is thrown on the molten surface of gold. In some cases where the alloy is brittle after melting, the slags from the fluxes are thickened with bone-ash or other inert substance, and are lifted off by means of a flat spiral of iron wire. This is lowered on the surface of the slag, some slag is cooled, and adheres. This is lifted out and pressed against a cold surface; the slag now forms a circular cake adhering to the spiral; this is again dipped down, raised and flattened, and the process repeated until the molten metal surface becomes apparent.

Refining With Mercuric Chloride

The temperature is raised until the metals are high above their melting points, when corrosive sublimate Hg Cl2 is thrown in, and a cover rapidly put on the pot. At a high temperature small quantities of base metals which render bullion brittle are removed by this treatment, and the alloy or metal is toughened. The process is only adapted for the removal of very small quantities of antimony, arsenic, or elements having a high atomic volume, and cannot be recommended on account of the poisonous fumes evolved in the process, for the bulk of the mercuric chloride is volatilised.

Refining with Ammonium Chloride

Ammonium chloride is also used for the same purpose in the same manner. When thrown on the molten metal it gradually sublimes and dissociates. The hydrochloric acid evolved attacks the base metals and removes part of them. The process, however, is incomplete and ineffective in the presence of a large quantity of base metal. If solid ammonium chloride is pressed below the surface of molten gold, much spitting is occasioned. Another method sometimes adopted for toughening brittle gold does so by the introduction of copper, which replaces equivalents of other metals. In this case oxide of copper is added with the gold. This at a high temperature reacts with lead and other metals; the resulting bullion in this case, however, contains copper instead of the metals removed, and the process is a toughening one more than a refining one.

Refining with Iron

Antimony and arsenic may be removed by stirring the molten metal with a bar of iron, a little nitre being added at the same time. These elements unite with the iron to form antimonides or arsenides, but at the same time it must be borne in mind that iron itself dissolves in molten gold, so that it is not possible to hit upon the exact point at which the whole of the base metals will be removed, and the iron left undissolved by the gold.

Removal of Tin

When tin is present carbonate of potassium is added, and heated with the molten metal.
Sometimes when nitre or other fusible corrosive refining agents or fluxes are used a layer of bone-ash is put over the surface of the molten bullion; small holes are made in this by scraping it to one side, and nitre or such fluxes added. When lead is present, sand and nitre are added.

Experimental Chlorinating Work

For some years the author has endeavoured to devise a method for refining retorted gold. It was recognised that once the metal becomes melted the surface exposed to the action of refining agents is so small, and the latter decompose so rapidly, that only a small degree of purification can be effected. With retorted gold, on the other hand, we have a porous material to deal with. The surface exposed is very large, and the refining fluxes are acting while the temperature is rising to melting point. The oxidation method with nitre and similar fluxes already described serves to cut out almost every metal save gold and silver. These cannot be used effectively in a plumbago pot, and clay pots are apt to become eaten through by the corroding action of the fluxes.

Refining with Sodium Bisulphate

It was considered that the silver as well as the base metals might be removed by some preliminary treatment. Nitric acid was ineffective on the retorted gold tried; sulphuric acid had little effect, but on testing with bisulphate of sodium it was found that base metals were removed or oxidised, and that silver sulphate formed The retorted gold was heated in porcelain or dense fireclay pots to a dull red heat for some time, generally as long as sulphur dioxide was evolved. The fused pyrosulphate was then poured out, and the cake drained as completely as possible. These were afterwards washed with hot water until the silver sulphate had been removed. The cake of melted pyrosulphate was also dissolved, and the silver present in the solution precipitated on iron, or by other well-known means. The gold could now be smelted, and invariably would be found to be much purer than when smelted by older methods. The silver and base metals were also removed at a fraction of the cost required by the chlorine method. The results obtained, however, were variable; one sample would be purified from about 80 to 96 per cent., yet another would only be increased by five or six per cent. The cause of this variation seemed to be that when coarse gold is amalgamated the mercury only affects the outer crust, so that on retorting there would be left pieces of gold porous on the outside, but unaltered within. In order to determine whether the gold and silver, if amalgamated in a fine state of division and then retorted would part, 80 grains of gold and 20 grains of silver were taken and amalgamated. The amalgam was retorted, and the resulting gold was of pale yellow colour. This was treated by boiling with strong sulphuric acid, which removed part of the silver; it was afterwards treated by fusion with sodium bisulphate, which removed more. The gold remaining was then washed with water, and after the silver sulphate had been removed the metal was weighed and parted. It was found to consist of: Gold, 97.2; silver, 2.8.

It is, therefore, a difficult matter to remove the whole of the silver, even when finely divided originally by this treatment: It might be stated here, although detailed reference will be made to it later on, that by using phosphoric acid with sulphate of sodium as a solvent that the retorted gold made in the same way was raised to:

99.2 gold.
0.8 silver.

Refining with Sulphate of Silver and Salt

These results were so promising that efforts were directed to shorten the process. The sulphate of silver takes a considerable time to wash out, and although the retorted gold is honeycombed, yet solution is retarded; and if any sulphate of silver remains it is decomposed on melting the bullion, and the silver becomes again alloyed with the gold.
If chloride of sodium is melted with sulphate of silver double decomposition takes place.

Ag2SO4 + 2Na2Cl=Na2S04 + 2AgCl.

The resulting silver chloride is not decomposed at a high temperature, and is not sensibly volatile when protected by a layer of fused salt. By melting the bisulphate with the retorted gold, and pouring any excess out, and then adding salt, and melting in a clay pot, the resulting bullion becomes almost as pure as if prolonged washing were given to remove the soluble silver salt.

A modification of this process consisted in first heating the retorted gold with molten common salt, so that the pores became filled with it. The crucible was then closed with a cover, and bisulphate of sodium dissolved in strong sulphuric acid allowed to drop in gradually. The sulphur trioxide and hydrochloric acid, both of which attack silver at a high temperature, had to escape through the retorted mass of gold. As soon as it was considered the salt was decomposed more salt was added to convert any sulphate of silver to chloride, and the whole melted down. The result of this treatment, so far as the refining was concerned, was satisfactory, but the bubbling in the pot, due to decomposition of salt, and evolution of hydrochloric acid are objectionable features. This modification gave purer gold than the bisulphate method.

Refining with Boric and Phosphoric Acid

It was then considered that if some relatively non-volatile acid were employed whose salts with silver were stable at a high temperature that these might be used in place of chloride of silver. The two which suggested themselves were boric and phosphoric acids. In order to try if the salts would decompose sulphate of silver, a mixture of sulphate of silver and boric acid were heated gradually, and then up to the melting point of gold. Sulphate of silver melted and decomposed, leaving metallic silver at a red heat, but both the borate and phosphate of silver which formed melted down and remained undecomposed at a temperature above the melting point of silver. Borax has commonly been used as a refining agent, but so far as I am aware, phosphoric oxide or acid has not. Indeed Percy goes so far as to say that “the phosphates of silver have no metallurgical interest.”

A strip of thin silver was then placed in boric acid, which was heated to redness; very slight action took place; when oxide of silver was substituted the oxide dissolved up and formed an opalescent mass, which was soluble in water, but from which a small quantity of black powder separated. The experiment was repeated with phosphoric acid. At a dull red heat the silver was attacked, and bubbles of gas were evolved. At a red heat no further action seemed to take place; the liquid remained clear when molten, but on cooling the mass solidified and turned black, when exposed to daylight. In this case the metaphosphoric acid probably decomposed to a slight extent into phosphoric oxide and water. The boric acid B2O3 3H20 at 100 deg. C. becomes B2 O3 2H20, and at 140 deg. C. B2 O3 H20, and at a high temperature B2O3. It is probable that while H20 is present the oxygen necessary for the formation of the borate is supplied, but the action ceases when this is expelled.

Phosphoric Acid and Oxidising Agents

Knowing that molten silver readily absorbs oxygen experiments were conducted to see if the addition of oxygen to alloys of gold and silver surrounded by phosphoric acid or boric acid would remove it. The first method thought of was by bubbling oxygen through the molten alloy, and having a cover of these acids, but it was considered that by heating weighed beads before the blowpipe in an oxidising flame that quicker results could be attained.

Boric Acid and Oxidising Agents

A boric acid bead was made on platinum wire, and an alloy of gold containing 72.14 per cent, of gold was used. The alloy taken weighed 0.192 grains. It was melted carefully, and the bead heated strongly in an oxidising blast. After about five minutes the bead, on cooling, became an opalescent grey colour, due to the oxide of silver dissolving in the boric acid. The bead weighed 0.188 grains. On re-heating for another five minutes the weight became 0.186 grains. Further heating for the same time did not reduce the weight appreciably. The weight of the silver was thus reduced by 11 per cent., but it does not seem possible to remove all the silver by this means.

Phosphoric Acid and Oxidising Agents

In the next experiment micro-cosmic salt H Na NH4 PO4 was used instead of phosphoric acid, since it has the same effect, and is not so fluid when melted as the latter, thus retaining the gold bead better. The same alloy was used, and 0.3 grains taken. On heating in the oxidising flame the bead became opalescent very quickly, and after five minutes the button weighed 0.290. Previous experiments had shown that it was somewhat difficult to keep the bead from running up and alloying with the platinum, so the bead was transferred to a small dish in which the flux was placed, and the bead heated as before in the oxidising flame. After the second fusion it weighed 0.280; after the third 0.270; after the fourth 0.263; after the fifth 0.258; the sixth 0.253; the seventh 0.250; the eighth 0.249. As the loss seemed to be approaching a limit a fine drop of sulphuric acid was added to the flux to determine the effect. The weight was reduced to 0.243; with another drop on re-melting the weight became 0.235, but after this the loss became very slight, only reaching 0.231. It is needless to state the button passed from a very pale to a bright yellow colour. The gold has been raised from 72.14 per cent, to 93.5 per cent., and 82 per cent, of the silver slagged off. This experiment shows that phosphoric acid in conjunction with oxygen, or an oxidising agent, has a powerful effect in tending to remove silver from gold. It was considered that the sulphuric acid formed hydrogen sodium sulphate, and liberated phosphoric acid in the cold, but on heating the non-volatility of the phosphoric acid would result in the expulsion of sulphur trioxide, and this would serve to attack the silver, but the second change would result in a relative stable phosphate of silver and sodium being formed.

In order to test the effect on a somewhat larger scale a piece of retorted gold was taken. The weight of this was 243 grains, and the amount of gold the button contained, as tested, by another sample, on melting was 86.4 per cent., the balance being nearly all silver. The retorted gold was placed in a porcelain dish, and 5 grammes of sodium bisulphate melted with it. After cooling a saturated solution of phosphoric acid was added, and the water driven off, and the whole placed in a muffle and left for some hours at a red heat. Bubbles of gas were freely evolved from the cake, and a clear, greenish, fusible slag remained. The gold was picked out, and the slag allowed to drain away from it. On being placed in water it gave a solution which reacted for silver. The slag, however, was not wholly dissolved out, but the cake was put in a French clay crucible, and melted down; a bright gold button weighing 215 grains was obtained. This assayed 977 fine gold; the balance being silver only.

Another experiment was made on a very base sample of retorted gold from Dargo (V.), which contained mainly gold, silver and copper. This was treated as in the previous case; the slag in this case, however, was turquoise blue. On melting down the button obtained an attempt was made to pour the small quantity of slag left in the pot, but the gold had not quite solidified, and ran down the side of the pot. When the small bar was taken out it was almost as white as silver. This coat proved to be merely superficial, and may have resulted from a fragment of reducing agent passing into the slag at the last moment, although a similar superficial white film can often be seen on pouring a high-grade gold bullion which has been melted under nitre. The weight of the retorted gold received, which was greenish-black in colour, was 322 grains.

When refined, 221 grains.
It assayed: gold, 98.15 per cent,
silver, 1.73
oxidisable metals, 0.12.

These examples serve to show that phosphoric acid might be made use of with great advantage in the purification of retorted base bullion. It is better than boracic acid or borax in not dissolving silica, and, therefore, it does not corrode the crucibles; it forms a readily fusible slag, capable of dissolving basic oxides, and if made use of after the manner indicated, it will remove a large quantity of silver from the retorted gold as well as base metals. There is no loss of gold by volatility through its action, and the silver can readily be reduced from the slags containing it by the action of iron in a slightly acid solution, or by the reducing action of suitable fluxes.

The small amount of silver remaining in the gold can be removed by the action of a current of chlorine being passed through the molten metal. It is also possible that some classes of retorted gold will yield a bullion over 995 fine.

Refining with Potassium Chlorate

In addition to these methods, chlorate of potash was tried, by melting it on retorted gold, so as to allow it to soak into the cake. The oxygen evolved served to oxidise base metals ; by adding some relatively fixed acid, such as is evolved from sodium pyrosulphate, chlorine is evolved. This serves to give a high-grade bullion, but the method was discarded when it was found that chloride of gold was carried away with the escaping vapours or gases evolved. Other oxygen carriers, such as chromates and manganates, were added, but the results were no better than those already described.