Older processes explaining how to go about separating silver from gold by cementation were carried on practically in two ways, the first by heating gold for a long time with a mixture which evolved nitric acid on heating, and an absorbent into which any fusible salts ran; the mixture in the second method contained common salt as an essential ingredient, and some material which would assist in giving off hydrochloric acid and chlorine, also an absorbent for the fusible salts which liquidated out. Detailed historical accounts of these are supplied in Percy’s Metallurgy, and it will be sufficient to briefly outline the processes. The impure bullion was rolled into thin plates about 4½ inches square. Each plate was then coated with a layer of nitrate of potash and clay, and these were packed one above the other. They were then surrounded by clean, dry cow dung, which was set on fire. Cow dung burns slowly and perfectly, leaving a white ash. The heat generated was sufficient to decompose the nitrate, which, in the presence of siliceous matter becomes silicate of potash, and the water evolved from the dehydration of the clay, combining with the oxides of nitrogen, forms nitric acid, which, in its turn, attacks the silver on the surface of the alloy, forming nitrate of silver. The nitrate of silver is absorbed by the burnt clay and is also decomposed, thus supplying more nitric acid. It is somewhat remarkable that although nitric acid will not attack such alloys as are so treated when they are immersed in it, yet in both cementation processes the silver is removed for a considerable depth. The increased temperature and the expansion of the alloy, and the diffusion of the particles, probably accounts for the attack of such solvents.

Generally three firings with cow dung were given, the ashes being removed from the sides of the pile each time, and the silver which had liquated out in the form of nitrate was recovered from them. After the three firings the plates were taken and washed with water—the burnt clay and adhering silver being removed. The operation of applying the mixture and repeating the firings was again performed. Each dressing and heating caused the removal of some silver, but it was not until six dressings and eighteen firings had been applied that the gold was considered pure. The silver from the clays and ashes was recovered by smelting with litharge and charcoal, and the resulting lead bullion cupelled.

The second process of cementation, or the formation of a cementing mass, was by means of common salt and brick dust. This mixture was heated in contact with thin gold-silver plates for a long time. The action of the salt on the clay, together with the oxygen of the air and watery vapor, resulted in the formation of silicate of sodium, hydrochloric acid and chlorine. The result was that the chlorine compound attacked the silver, converting it into chloride, which was absorbed by the cementing medium.

Many modifications of this ancient method of refining were in use, such as the addition of ferrous and cupric sulphates to the chloride mixtures, and the addition of ammonium chloride, but the final action was invariably due to chlorine or hydrochloric acid. As a modification of the method of applying the mixture to gold in the form of plates, it was sometimes granulated, but the granules must have offered more difficulty in the separation of the gold from the adhering cementing mixture.

As in the case of nitric acid the alloys of gold and silver are acted upon by this cementing mixture, while if the plates themselves were immersed in hot hydrochloric acid they would be only superficially acted upon. There seems to be no doubt that diffusion plays a very important part in these cementation processes—not only of the gaseous products through or into the plates, which was looked upon as the sole reason of the formation of the silver compounds, but the actual transfusion of the metals themselves at a slightly elevated temperature. For instance, if an alloy of gold and silver, which is white on melting, be rendered gold-colored by having the silver superficially removed by a suitable solvent, be heated, it will become white. The superficial golden layer has been re-absorbed, although the temperature is far below the melting point of the alloy.

The diffusion of carbon in iron far below its melting point has been abundantly proved by Roberts Austin (Nature, Vol. XLI, p. 14). The chapter by the same author in his Introduction to Metallurgy, 4th edition, pp. 56, 66, shows that miscible solids diffuse through each other in a similar manner to liquids. Such diffusion is largely dependent on temperature. In the case of the gold-silver alloys, the explanation of the cementation process is that a superficial removal of silver occurs, the alloy is rendered poorer in this metal and more diffuses from the interior; this in turn is removed and the action goes on until very little remains. It naturally follows, the thinner the plate the more rapidly will the silver be removed. The term cementation has been transferred to another class of operations, in which solid elements are added to others by merely heating them in contact—for instance, the transference of carbon of metallic iron by packing the latter in charcoal dust and heating; the iron takes up carbon, and is thereby converted into steel; probably the method of packing the metals suggested the term, but in the case of steel the packing material does not cement, nor does it withdraw anything from the iron, yet in reality the action in both cases is due mainly to the diffusion of solids in solids.