This section is from the book "Leaching Gold and Silver Ores. The Plattner And Kiss Processes: A Practical Treatise", by Charles Howard Aaron. Also available from Amazon: Leaching Gold And Silver Ores.
155. Loss In Roasting. The only reliable method of determining the loss of gold or silver in the roast-ing of ore on the large scale is, to dry, weigh, sample, and roast a quantity, say from one to ten tons; then, after cooling, to weigh and sample again. It may weigh more or less than before, according to the character of the ore, the quantity of salt used, and the manner of roasting. The assay of the unroasted ore, coupled with the total weight taken, gives the quantity of precious metal in the lot before roasting, and the assay of the roasted ore, with its total weight, gives the quantity remaining after roasting. The difference is the loss, caused by dusting and volatilization. To make the test reliable, the furnace must be thoroughly cleaned, both before and after the roasting; if this is neglected, the weight obtained for the roasted ore will be incorrect.
As a part of the precious metal lost in the roasting may be recovered from the flues and dust chamber, while, besides the losses already discussed, others may occur in the leaching, collecting, and melting of the metal, it will be seen that a final result can only be reached by means of a general and complete clean-up; yet, as this cannot be had very frequently, the investigations described are indispensable as guides in working.
It is best, when any doubt is entertained as to the best way of roasting a given ore, to experiment in the laboratory on ounce or half ounce samples taken from crushed lots of considerable magnitude. The method employed by some metallurgists, of taking samples from the furnace at intervals during the progress of the roasting, and assaying them with the view of ascertaining the loss at successive stages, is open to the objection that it cannot be known exactly what correction must be made for the changes of weight sustained by the ore. It is, however, very proper to make such trials, because a heavy loss might be thus indicated in time to alter the treatment of the charge. The percentage of the silver which is soluble must be considered in connection with the loss by volatilization, in order to arrive at the most profitable manner of roasting; for it is better to make rich, rather than poor, tailings, if the difference should go up the smoke stack.
156. Solubility Assay. This is commonly, but incorrectly, called a "chlorination assay." It is made as follows: From a quantity of the roasted ore, two assays are weighed out, one of which is leached on a filter with hypo as long as any metal can be detected in the filtrate; it is then washed and dried. The leached and the unleached samples are then separately smelted, and the resulting lead buttons cupelled and parted. The result obtained from the unleached sample gives the assay value of the roasted ore; that from the leached sample the portion of the silver which is insoluble in hypo.
If the average change of weight which the ore sustains in roasting has been ascertained, a corresponding correction, applied to these assays, will show the loss in roasting, if an assay has been made of the unroasted ore; thus:
Raw ore contains 56 oz. of silver per ton. Loss of weight in roasting=5 per cent; therefore, Roasted ore should contain 56 + 2.9=58.9 oz. per ton, but actually contains 57.0 " " "
The loss in roasting is therefore 1.9 " " "
Unleached sample contains 57 oz. per ton Leached " " 1.5 " " "
Soluble silver 55.5 " " " = 97.3 per cent of the silver remaining in the ore after roasting. But a ton of roasted ore which has lost 5 per cent in weight corresponds to 1.052 ton of raw ore, containing 58.91 oz. of silver, of which we extract 55.5 oz. The total loss therefore is 3.41 oz., in 58.91 oz., or 5.79 per cent nearly, which, however, would be a better result than is often obtained.