Chapter IX (1)

SILVER, GOLD, CYANIDES, PLATINUM, MERCURY.

SILVER.

Silver is widely diffused, and has been found in most mining districts. It occurs native in sufficient quantity to constitute one of the chief ores of the metal. It also occurs combined with sulphur (as in argentite), with sulphur and antimony (as in stephanite or brittle silver ore, and in pyrargyrite or ruby silver), and with copper, sulphur, antimony, and arsenic, as in polybasite. Chloride of silver occurs native as horn silver or kerargyrite. Silver is found in the ores of other metals, such as fahlerz, which sometimes contains from two to ten per cent. of the metal, and galena, which is an important source of it; in fact, galena is never found entirely free from silver. It is present also in greater or less quantity in the ores of copper and zinc.

Silver dissolves readily in nitric acid, forming silver nitrate. It only forms one family of salts, and of these the chloride and nitrate are of chief importance to the assayer. The formation of the chloride of silver on the addition of hydrochloric acid or a soluble chloride to the nitric acid solution, serves for the recognition and separation of silver. The precipitated chloride is white (becoming violet on exposure to light), insoluble in nitric acid, soluble in ammonia, hyposulphite of soda, or concentrated solutions of chlorides. The best confirmatory test is made by wrapping the precipitate in a little sheet lead, and cupelling, when the silver will be left in the metallic state, and is easily recognized.

~Dry Assay.~--This assay is made up of two parts: (1) the concentration of the silver in a button of lead; and (2) the cupellation of the resulting alloy. The concentration of the button of lead may be effected either by scorification or by fusion in a crucible.

The scorification assay is performed in a scorifier, which is a shallow open-mouthed dish about 2-1/2 inches across, with a very thick bottom to enable it to withstand the corrosive action of the slag. A charge of more than 3 or 5 grams of the ore cannot be worked in one, and with such small charges the unavoidable variations have a serious effect on the figures reported. A difference of one milligram on the weight of the button of silver got represents a difference of 6 or 10 ounces per ton. With rich ores such variation is unavoidable under any conditions, and the only safe plan is to take the mean of several assays. But with poorer ores the accuracy of the assay, as well as convenience in working, is much increased by working in a crucible with larger charges.

In scorification the proportion of lead required for scorifying 1 gram of ore is in average cases from 10 to 15 grams, sinking in the case of galena to 2 grams, and rising with earthy and refractory substances to from 30 to 40 grams. But by fusing in a crucible with well-selected fluxes, a proportion of 4 of flux to 1 of ore is generally sufficient; and not only is the proportion of added matter less, but it is also easier to manipulate large quantities in crucibles, so that, although in some cases the crucible assay is more troublesome and less satisfactory, yet with poor and earthy ores it is the best method of dealing with them; while when properly worked it yields results as accurate as scorification does. As a general rule, if more than 5 grams of ore must be taken, the crucible assay should be adopted.

~Scorification Assay.~--The charge of ore is usually 3 grams, sometimes 5; the lead varies from 30 to 70 grams, and the quantity of soda, borax, or powdered glass added varies from 0.3 to 3 or 4 grams. It is generally recommended to have the lead granulated,[9] and to mix the ore with about half of it in the scorifier; then to put on the rest of the lead; and finally to sprinkle the borax or glass on the top. It answers just as well, however, to use the lead in the shape of foil, and wrap the ore up in it; and if the ore contains much sulphur, the borax may with advantage be added (wrapped in a little tissue paper) some five or ten minutes after the operation has started.

The process of scorification is as follows:--A scorifier (fig. 38) of convenient size having been selected (one 2-1/2 inches across is most generally useful), it is dried at a gentle heat for about ten minutes. The charge is then put into it, and it is introduced, with the help of a scorifier tongs (fig. 39), into a muffle heated considerably above redness. The muffle is then closed, and when the metal has melted down, it is opened, but the temperature is kept up. A ring of slag will, after a time, form around the metal, and when this appearance (known as the eye) presents itself, the temperature may be lowered. When the eye has disappeared--that is, when the layer of slag has quite closed in--a pinch of powdered culm wrapped in tissue paper is added. As soon as the slag has again become tranquil, the scorifier is taken out, and its contents are poured into a mould (fig. 40), the slag is detached, and saved. If the button of metal weighs more than 30 grams, its size is reduced by another scorification in the same scorifier, which should have been replaced in the muffle immediately after the contents had been poured out. If the ore is not a very rich one, the button of lead will carry practically all the silver; but with rich ores it is more satisfactory to save the slag, and subsequently to melt it down with the cupel on which the lead has been treated, so as to recover the silver lost in the slag, together with that absorbed in the cupel, at one operation. Or, if the cupellation loss is neglected or calculated in some other manner, the slag or slags from the scorifier may be powdered and mixed with 20 grams of oxide of lead, 5 grams of borax, and 1 gram of charcoal. This should be melted down in a small crucible, and the resulting button of lead cupelled.

If the scorification has been unsatisfactory, the quantity of silver obtained from the slag will be by no means inconsiderable. The usual explanation is that with sulphury ores compounds of metallic oxides and sulphides (oxysulphides) are formed, which remain in the slag, retaining considerable quantities of the precious metal. It is said that under certain conditions such a slag may contain as much as 10 per cent. of silver. An excess of lead and a high temperature prevents the formation of these oxysulphides. But if much silver is present in the ore, the slag cannot be safely thrown away, even if sulphur is absent, and the process has been satisfactorily performed.

If the crust which appears on the surface of the lead does not clear, add a small lump of borax and 20 grams more lead; then close the muffle, and keep the temperature as high as possible. If the slag forms properly, but shows unfused or only half-fused lumps, even when the scorification has proceeded for some time, add more borax, and stir with an iron rod. The slag adhering to the rod must be detached by hammering, and replaced in the scorifier.

If the ore consists largely of quartz, soda should be added instead of borax; or, if it contains much copper, powdered quartz may be used. If the scorifier at the end of an operation is more than usually corroded, the borax should be replaced in subsequent assays on similar ores by powdered glass or quartz.

If a fairly fluid slag is formed which does not clear from the metal and show the eye, more lead and a higher temperature is wanted.

As a general rule, it may be stated that when a scorification is unsatisfactory, what is wanted is more heat, more lead, or more borax.

It is a safe plan when work has to be done on a strange ore, to make three or four assays with varying quantities of lead. The proportion of lead is right when a further addition does not yield a higher result. The proper proportion having been found, a note of it should be made for future use.

POT ASSAYS.

The object of the fusion in a crucible, like that of scorification, is to concentrate the silver in a button of lead which is to be subsequently cupelled; and to retain the earthy and waste matters in the slag. It is necessary to consider the quality of the slag and the weight and quality of the lead. The slag when fused should be liquid and homogeneous, and not too corrosive on the crucible. The button of lead should be soft, malleable, and free from a coating of regulus.[10] In weight it should not differ much from the ore taken. With 20 grams of ore, for example, a button of lead weighing from 18 to 25 grams will be satisfactory: less than this would leave an undue proportion of silver in the slag; and more would be unnecessarily large for cupelling, and would increase the loss in that operation.

With average ores, take 20 grams of the powdered ore and mix with 30 grams of "soda," 40 grams of red-lead or litharge, 5 grams of borax, and from 2 to 2.5 grams of flour, and place in an E crucible (Battersea round). Put these in the furnace at a red heat, cover the crucible, and gradually raise the temperature until the whole charge has melted down and is in a state of tranquil fusion. Pour into a mould, and replace the crucible in the furnace. As soon as the lead is solid, detach the slag and put it back into the crucible; and when it is again fluid, charge on to it with a copper scoop a mixture of 20 grams of oxide of lead, and 1 gram of charcoal: when fusion has again become tranquil, pour and detach the button of lead. The lead buttons should be hammered into discs with rounded edges, and be freed from slag; if too big for a cupel they may be scorified together in a small scorifier, but it is better to cupel them separately.

~Ores containing Metallic Oxides.~--Peroxides of iron, manganese, and copper interfere by counteracting the effect of the charcoal or flour, and thus reducing the size of the lead button. Peroxide of iron will reduce the weight of lead by a little more than its own weight; and peroxide of manganese has about twice this effect. When these oxides are present an additional quantity of flour must be used, and precautions must be taken to prevent reoxidation of the slag by the furnace gases. This may best be prevented by using a layer of common salt as a cover to the charge. When the ores contain a good deal of quartz or stony matter, the fluxes just given (for average ores) will do; but the proportion of soda should be diminished, and that of the borax, oxide of lead, and flour increased as the quantity of metallic oxides become greater. If the ore contains practically no quartz, the soda may be altogether omitted, and some glass or powdered quartz added. The following charge may be taken as an example: weigh up 20 grams of the powdered ore, 15 grams each of "soda" and borax, 60 grams of oxide of lead, and 5 grams of flour. Mix and place them in an E crucible, and cover with a layer of from a quarter to half an inch of common salt. Place in the furnace as before. The salt will give off a considerable amount of fume, which will, to a certain extent, conceal the state of the charge: when the crucible has been in the furnace for about 25 minutes remove it and pour out the contents immediately. With ores that produce a thick slag the addition of 5 grams of fluor spar will be an advantage. It may happen that with an unknown ore the first assay will be more or less unsatisfactory: but from it the necessity for adding more or less flour will be learnt, and a second assay, with the necessary modification of the charge, should give a good result.

~Ores containing much Sulphides.~--Ores of this class may be easily recognized, either by the appearance of the minerals they contain or by the odour of sulphurous oxide (SO_{2}) which they evolve when roasted on a spatula. The sulphides most commonly present, in addition to the sulphurized minerals of silver, are pyrites, galena, blende, and mispickel. When they are present in only a moderate amount, their effect is simply to increase the weight of the button of lead; and this is easily counteracted by reducing the amount of flour, or by omitting it. When in larger amounts, they not only yield large buttons, but also render the metal sulphury, sometimes even giving a button of regulus instead of lead. This last evil may be remedied (1) by putting in a rod of iron as soon as the charge has fused, or (2) it may be counteracted by a proper addition of nitre, or (3) when the sulphides present are only those of iron or copper the sulphur may be removed by calcining, and the ore converted into one of the class containing metallic oxides. The calcination is effected as follows:--Weigh up 20 grams of the powdered ore and place it in a wide-mouthed crucible sufficiently large to perform the subsequent melting down in. The roasting must be done at a gentle heat at first, so as to avoid clotting: the mouth of the crucible should project considerably above the coke, and should slope forward towards the worker. The charge must be occasionally stirred with the stirrer (fig. 10) so as to expose fresh surfaces to the action of the air, and to prevent adhesion to the sides of the crucible. The stirrer should not be removed till the calcination is finished. The temperature should be raised at the end to a good red heat; and (to ensure the decomposition of any sulphate that may be formed) the roasted ore should be rubbed up in a mortar with a pinch of anthracite, and again calcined. It is then mixed with fluxes as described, and fused in the same crucible.

The calcination of an ore is a work occupying a good deal of time, and, in most cases, it is better to take advantage of the desulphurizing power of red lead or nitre. Red lead by itself will do, but a large quantity of it will be required; 1 part of a metallic sulphide needs from 20 to 50 parts of red lead to yield a button free from sulphur; whereas at most from 2 to 2-1/2 parts of nitre are sufficient. There is sometimes an advantage in having a considerable excess of oxide of lead in the slag, but where there is no such reason, 2 parts of red lead to 1 of ore is enough. A charge which will do for most sulphides is the following: 20 grams of ore, 40 to 100 grams of red lead, 20 grams of "soda," 5 of borax, and sufficient nitre (or perhaps flour) to give a button of about 25 grams of lead. How much this must be (if not already known) may be approximately determined by fusing 3 grams of the ore and 3 grams of "soda" in a small crucible (C) with 50 grams of litharge (not red lead) under a cover of salt, and weighing the resulting button of lead. Subtract 3 from the weight of lead obtained, and the difference multiplied by 1.3 will give the quantity in grams of nitre required. If the button of lead weighs less than 3 grams flour must be added. If this is not satisfactory repeat the assay, adding an extra gram of nitre for each 4 grams of lead in excess of that required, or 1 gram of flour for a 12-gram deficiency.

In the method in which iron is used as a de-sulphurising agent, only as much oxide of lead should be added as will give a button of lead of the required size. Rather a large button of lead should be got, and the slag should be strongly alkaline; if the ore does not already carry a large amount of sulphur some should be added. The fusion should be performed at a low temperature (similar to that for a galena assay), and should be continued for some time after it has become tranquil. Take 20 grams of the ore, 40 grams of "soda," 40 grams of oxide of lead, and 5 or 10 grams of borax; place this mixture in a crucible (with a rod of iron, as in the galena assay), cover, and fuse for about half an hour. Take out the rod, washing it in the slag, and, in a minute or two, pour. Clean and cupel the button of lead.

~General Remarks on the Fusion.~--Other things being equal, the smaller the quantity of the slag the better, provided there is sufficient to cover the metal. The presence of peroxides of the heavy metals is prejudicial, since they tend to increase the quantity of silver retained in the slag. It may be given as a general rule that when iron, copper, manganese, &c., are present, there is a more than ordinary need for cleaning the slags, and care must be taken to keep these metals in the state of lower oxide.

In selecting the fluxes, it should be remembered that soda is the best for quartz, and borax for lime and metallic oxides. And that with ores almost free from gangue some quartz or glass should be added to protect the crucible. Two parts of soda are enough to flux 1 part of quartz; whilst of borax, or oxide of lead, 4 parts are barely sufficient. Oxide of lead has the advantage of being heavy and so does not occupy much space in the crucible; on the other hand, if the melting down be performed too quickly, or if oxide of lead only is used, this high specific gravity is a disadvantage, for the lighter earthy matter floats as a pasty mass on the more fluid oxide of lead, and thus escapes its action.

When metallic sulphides are present in the ore, an excess of oxide of lead helps to keep the sulphur out of the button of metal. In addition to the oxide of lead required as a flux, some will be required to provide the lead in which the silver is to be collected. Oxide of lead, mixed with charcoal or flour, yields, when heated, a multitude of minute buttons of metal uniformly distributed through the mass of the charge; as the charge melts down these run together and fall to the bottom; this shower of lead collects the silver more easily than a single button at the bottom of the crucible could do. Only that portion of the oxide of lead which remains in the slag can be considered as a flux; very often the first indication of an excessive reduction of lead is the pastiness of the slag rendered thick by the withdrawal of the oxide of lead which would have kept it fluid. If, in an assay, it is found that 5 parts of flux are not sufficient for 1 part of ore, the remedy lies in using a different flux rather than in taking a larger quantity.

_On the Reducing Effect of Charcoal, Flour, and Tartar._--The weight to be got from a given charge will depend (provided sufficient oxide of lead is present) upon the proportion of the reducing agents in it. We have thought it well to illustrate this part of the subject by a series of experiments which the learner will do well to practise for himself before proceeding to the assay of actual ores. Take 80 grams of litharge and 20 grams of a mixture of borax and soda. Fuse three lots (1) with 1.5 gram of charcoal, (2) with 3 grams of flour, and (3) with 7.5 grams of tartar. Weigh the buttons of lead obtained, and divide each by the weight of reducing agent used. The results will differ somewhat with the dryness and quality of the flour, etc., used; in one series of experiments they were as follows:--

Gram. Grams. Gram. Grams.
1.5 charcoal gave 34.0 lead .'. 1 charcoal = 22.6 lead.
3.0 flour " 33.5 " .'. 1 flour = 11.2 "
7.5 tartar " 38.0 " .'. 1 tartar = 5.0 "

The use of flour as a reducing agent has many advantages, and it is well to remember that _1 gram of flour reduces about 11 grams of lead_; and that charcoal has twice, and tartar one-half, this reducing effect.

_On the Reducing Effect of Charcoal, &c., on Red Lead._--It is often easier to obtain red lead of good quality than it is litharge, and by a large number of assayers red lead is the form of oxide of lead always used. Red lead, however, contains an excess of oxygen which will use up some of the reducing agent before lead separates out. On making a series of experiments (similar to the last, but using 80 grams of red lead instead of the litharge) the results were, with the same quantities of the reducing agents:--

With charcoal, 18 grams of lead.
" flour, 18 " "
" tartar, 22 " "

Comparing these with the results with litharge, in the previous table it will be seen that the same quantity of reducing agent has in each case brought down 16 grams less of lead, so that a larger amount of the reducing agent must be added to get a button of the same weight as that obtained with litharge. To get a button of a desired weight, say 22 grams, we must add reducing agent sufficient to throw down 22 + 16 or 38 grams of lead, which would require 3.4 grams of flour. If this amount of flour is fused with 80 grams of red lead, a button of lead weighing 22 grams will be formed, the other 16 grams being kept up by the oxygen of the red lead.

If the quantity of red lead differs from 80 grams, this rule must be modified. With 40 grams of red lead, for example, we should add an excess of reducing agent sufficient to throw down 8 grams of lead instead of 16. Similarly, with 160 grams of red lead, we should add enough to throw down 32 grams.

The following rule will enable one to calculate the weight of flour required to produce a button of lead of any desired weight from any given quantity of red lead. Each 5 grams of red lead present diminishes the weight of the lead by 1 gram. If then we _divide the weight of red lead in a charge by 5, and add this to the weight of lead required, the sum divided by 11 will give the weight of flour which must be added_. Using 80 grams of red lead and wanting a button of 20 grams, we should add 3.3 grams of flour.

80/5 = 16; 16+20 = 36; 36/11 = 3.3 nearly.

The following are some results obtained which will illustrate the rule:--

Red Lead used. Flour used. Lead got.
40 grams 3 grams 25.0 grams
100 " 3 " 13.5 "
80 " 4 " 30.0 "
80 " 5 " 40.0 "

_On the Reducing Effect of Metallic Sulphides, and the Counteracting Effect of Nitre._--The sulphides found in ores will reduce a button of lead from oxide of lead just as flour does; and, as charcoal, flour and tartar differ in their reducing power, so equal weights of the different mineral sulphides throw down different weights of lead.

One gram of iron pyrites yields about 11 grams of lead. One gram of copper pyrites, blende, fahlerz, or mispickel, yields 7 or 8 grams of lead, whilst 1 gram of antimonite will give 6, and 1 gram of galena only a little over 3 grams. It is evident that if an ore carries much of these sulphides, the quantity of lead reduced will be very much larger than that required for an assay. To counteract this effect nitre is added; _1 gram is added for each 4 grams of lead in excess of that required_. For example: with a 20-gram charge of an ore containing 50 per cent. of pyrites, if no nitre were added, 110 grams of lead would be got; or, if there was not sufficient oxide of lead to yield this quantity of metal, the button would be sulphury. To reduce the weight of the button by 80 grammes, we should add 20 grams of nitre, if litharge were used; or if red lead were used, we should add 16 grams of nitre, _since the oxidizing effect of 20 grams of red lead is equivalent to that of 1 of nitre_, and since 80 grams of red lead are generally used in a charge. Two assays of an ore of this kind with these quantities of nitre gave 26.0 grams of lead with litharge, and 22.5 grams with red lead.

It is best to use in these assays 80 grams of red lead, 20 of soda, and 5 of borax, with 20 grams of the ore. If the lead got by the preliminary fusion in a small crucible with litharge (described under "_ores containing much sulphides_") is known, the following table will indicate the quantity of nitre, or flour, to be added with this charge:--

-------------------------------+-------------------+------------------
Lead got in Preliminary Fusion | Flour to be added | Nitre to be added
with 3 grams of Ore. | to the Assay. | to the Assay.
-------------------------------+-------------------+------------------
0.0 gram | 3.3 grams | none
3.0 grams | 1.3 gram | --
6.0 " | none | 4.0 grams
9.0 " | -- | 9.0 "
12.0 " | -- | 14.0 "
15.0 " | -- | 19.0 "
18.0 " | -- | 24.0 "
21.0 " | -- | 29.0 "
-------------------------------+-------------------+------------------

If litharge is used in the assay instead of red lead 4 grams more nitre, or 1.5 gram less flour must be used. When more than a few grams of nitre are added to a charge the proportion of "soda" and borax should be increased, because one of the products of the reaction of nitre upon sulphides in the presence of soda is sulphate of soda, and because the "soda" thus used up no longer serves as a flux; more borax should be added, as it is the best flux for the metallic oxides which are formed in the process. If in an assay too large a button of lead is got, even after this calculation has been made, and the assay is repeated, add 1 gram more nitre for each 4 grams of lead in excess. Sometimes the assay appears tranquil before the nitre has produced its full effect; in such cases it is well to seize the crucible with the tongs and mix its fused contents by rotating them; if this causes an effervescence, the crucible should be replaced in the fire and the fusion continued. The following experiments will illustrate the extent to which the above rules may be relied on. In all of them the standard flux was used, viz.:--80 grams of red lead, 20 of soda, and 5 of borax.

_Pyrites_ 5 5 5 5 2.5 5 10 15 20
Quartz -- 20 -- 20 17.5 15 10 5
Nitre -- -- 5 5 -- 4 16 28.5 41
Lead got 42.5 36.0 16.0 19.0 11.5 22.5 22.5 26.5 27.5

_Copper Pyrites_ 8 8 8 8
Quartz -- 12 -- 12
Nitre -- -- 4 4
Lead got 47.5 34.0 33.0 26.0

_Antimonite_ 8 8 8 8
Quartz -- 12 -- 12
Nitre -- -- 4 4
Lead got 29.0 26.0 13.0 13.0

_Galena_ 10 10 10 10 15 20
Quartz. -- 15 -- 15 5 --
Nitre -- -- 3 3 3.5 7
Lead got 17.0 19.0 8.0 8.0 18.5 18.5

A similar set of experiments, with 80 grams of litharge instead of 80 grams of red lead, gave:--

_Pyrites_ 4 4 4 4 7 10
Quartz -- 15 -- 15 13 10
Nitre -- -- 5 5 12.5 20
Lead got 46.5 40.5 25.5 24.5 27.0 26.5

_Copper Pyrites_ 5 5 5 5
Quartz -- 15 -- 15
Nitre -- -- 5 5
Lead got 44.5 32.5 23.0 25.0

_Blende_ 5 5 5 5 10
Quartz -- 15 -- 15 10
Nitre -- -- 5 5 15
Lead got 41.5 38.5 21.5 22.5 21.6

_Antimonite_ 5 5 5 5 10
Quartz -- 15 -- 15 10
Nitre -- -- 5 5 10
Lead got 31.0 32.5 11.5 12.5 18.7

_Galena_ 10 10 10 10 15 20
Quartz -- 15 -- 15 5 --
Nitre -- -- 5 5 7.5 11
Lead got 33.5 33.5 13.0 14.0 19.5 22.7

The variation in some of these experiments, in which we might have expected similar results, is due to the fact that the sulphur, and in some cases the metals, are capable of two degrees of oxidation. For example: theoretically 1 gram of iron pyrites (FeS_{2}) would yield 8.6 grams of lead if the sulphur were oxidised to sulphurous oxide (SO_{2}), and the iron to ferrous oxide (FeO); whilst if the sulphur were oxidised to sulphate (SO_{3}), and the iron to ferric oxide, 12.9 grams of lead will be thrown down. Similarly the yield with copper pyrites would be 7.5 or 11.6; with blende, 6.4 or 8.5; with antimonite, 5.5 or 8; and with galena, 2.6 or 3.4. As regards the metals, the lower oxide will always be formed if the assay is carried out properly (fused under a cover, and with a sufficiency of reducing agent). But the proportion of sulphur oxidised completely will vary with the conditions of the assay. With a slag containing much soda the tendency will be to form sulphate, and, in consequence, a big reduction of lead; whilst with an acid slag containing much quartz the tendency will be for the sulphur to go off as sulphurous oxide (SO_{2}). In a fusion with litharge alone all the sulphur will be liberated as the lower oxide, whilst with much soda it will be wholly converted into sulphate. For example: 3 grams of an ore containing a good deal of pyrites and a little galena, gave, when fused with litharge, 16.5 grams of lead. A similar charge, containing in addition 20.0 grams of soda, gave 22.5 grams of lead.

It will be noted from the experiments that 1 gram of nitre kept up on the average 4 grams of lead; the range being from 3.2 with acid slags to 5.3 with very basic ones. These facts serve to explain some apparently irregular results got in practice.

CUPELLATION.

The process is as follows:--The cupels, which should have been made some time before and stored in a dry place, are first cleaned by gentle rubbing with the finger and blowing off the loose dust; and then placed in a hot muffle and heated to redness for from 5 to 10 minutes before the alloy to be cupelled is placed on them. The reasons for this are sufficiently obvious: the sudden evolution of much steam will blow a cupel to pieces; and, if the whole of the water has not been removed before the cupel is filled with molten lead, the escaping steam will bubble through, and scatter about particles of the metal. If some particles of unburnt carbon remain in the bone ash, a similar result will be produced by the escape of bubbles of carbonic acid as soon as the fused litharge comes in contact with them. The cupels having been prepared are arranged in a definite order in the muffle, and the assay buttons are arranged in a corresponding order on some suitable tray (cupel tray, fig. 41); the heat of the muffle being at bright redness. Then with the help of the tongs (fig. 42) the assay buttons should be placed each in its proper cupel; a note having been previously made of the position it is to occupy, and the door of the muffle closed.

This part of the work should be done promptly, so as not to unduly cool the muffle: the start requires a fairly high temperature, and is a critical part of the process. A black crust forms at once on the surface of the lead; but this ought soon to fuse and flow in greasy drops from off the face of the metal, so as to leave the latter fluid with a well-defined outline, and much brighter than the cupel. If this clearing does not take place, the buttons are said to be frozen; in which case the temperature must be raised, some pieces of charcoal put in the muffle, and the door closed. If they still do not clear, the heat must have been much too low, and it is best to reject them and repeat the assays.

When the buttons have cleared it is well to check the draught of the furnace, and to partly open the door of the muffle, so as to work at as low a temperature as is compatible with the continuation of the process.[11] Too low a temperature is indicated by the freezing of the buttons and the consequent spoiling of the assays. Experience soon enables one to judge when the heat is getting too low. A commoner error is to have the heat too high: it should be remembered that that which was high enough to clear the buttons at starting is more than sufficient to keep the process going. At the finish a higher temperature is again required: therefore the door of the muffle should be closed and the furnace urged. The finish is easily recognised. The drops of litharge which in the earlier stages flow steadily from the surface of the alloy, thin off later to a luminous film. At the end this film appears in commotion, then presents a brilliant play of colours, and, with a sudden extinction, the operation is finished. The metal again glows for an instant whilst becoming solid.

If the button is a small one the cupel is withdrawn at once and placed on that square of the cupel tray which corresponds to the position it occupied in the muffle. If, however, it is fairly large precautions must be taken to prevent spirting.

Molten silver dissolves oxygen from the air and gives it off on solidifying; the escape of the gas on sudden cooling is violent and, by throwing off particles of the metal, may cause loss. This is called "vegetation" or "spirting." The silver is apparently solid when spirting takes place; the crust breaks suddenly and some of the metal is forced out. The evil is best guarded against by slow cooling and avoiding draughts. With large buttons of silver precautions should never be omitted. One plan is to allow the cupels to cool in the muffle itself, the mouth being closed with hot charcoal. Another is to cover the cupel with another cupel previously heated to redness; in this case the silver cools between two hot cupels, and, of course, cools slowly. A third plan is to withdraw the cupel to the door of the muffle, holding it until it begins to get solid and then immediately to put it back into the hotter part of the muffle.

Silver remains after cupellation in flattened elliptical buttons, adhering but only slightly to the cupel. Its upper surface should show faint markings as if it were crystalline. The presence of platinum renders it still more crystalline, but removes the characteristic lustre and renders the metal dull and grey. Copper, if not completely removed, has a very marked effect on the appearance of the button: the metal is spread out, damping, as it were, and firmly adhering to the cupel, which latter in the neighbourhood of the metal is almost black with oxide of copper. Sometimes the silver button is globular, or even more sharply rounded on its under than on its upper surface; it is said that this is due to the presence of lead. Gold may be present even to the extent of 50 per cent. without showing any yellow colour.

The appearance of the cupel affords some useful information. The presence of cracks evidently due to shrinkage indicates a badly made cupel. If, however, they are accompanied by a peculiar unfolding of the cupel, the margin losing its distinctness, it is because of the presence of antimony. When lead is the only easily oxidisable metal present, the stained portion of cupel is yellow when cold. A greenish tint may be due to small quantities of copper or, perhaps, nickel, cobalt, or platinum. Larger quantities of copper give a greenish grey or almost black colour. A dark green and corroded cupel may be due to iron. Rings of pale-coloured scoria may be due to tin, zinc, antimony, or arsenic. When the cupel shows signs of the presence of these metals in objectionable quantity, it is well to repeat the assay and scorify so as to remove them before cupellation.

The button should be detached from the cold cupel by seizing with a pair of pliers: the under surface should be distorted by squeezing or hammering the button so as to loosen the adhering bone ash. The cleaning is easily completed by rubbing with a clean hard brush. After cleaning the buttons are best put on a tray of marked watch-glasses, and then taken to the balance and weighed. The weight of silver got needs a small correction; (1) by deducting for the amount of silver introduced by the lead or oxide of lead used in the assay;[12] and (2) by adding for the cupellation loss.

~Loss in Cupellation.~--During the whole process of cupelling a silver lead alloy a more or less abundant fume may be observed rising from the cupel. This furnishes an evident loss of lead and a possible loss of silver; for although silver at the temperature of cupellation gives off no appreciable vapour, it is known that such fume formed on a large scale contains silver. It is, however, difficult to believe that the small amount of lead vapourised carries with it a weighable amount of silver. That it does not do so in the ordinary way of working is shown by the fact that a button of silver equal in weight to the silver lost in cupelling may be got by smelting the cupel and cupelling the resulting button of lead. The loss of silver by volatilisation is altogether inconsiderable, unless the temperature at which the operation is performed is much too high.

Another possible source of loss is the infiltration of small particles of alloy into the cupel. The cupel is necessarily porous, and particles of metal may perhaps drain into it, more especially if the bone ash is not in fine powder; but if this is the main source of loss it is hard to see why, in cupelling equal weights of silver and gold, the loss is not equal in each case. It is not easy to believe that the mere filtration of the fused alloy will effect such a change in the proportion of the metals as that which actually occurs. For example: a cupel on which an alloy consisting of 0.80 gram of silver, 0.47 gram of gold, and 25 grams of lead had been cupelled, was found to contain 7-1/2 milligrams of silver, and rather less than half a milligram of gold. Assuming, for the sake of argument, that the gold present had filtered into the cupel in the form of small drops of alloy, it would have been accompanied by less than a milligram of silver, and the presence of the extra 6 or 7 milligrams of silver must have been due to a different cause. There can, thus, be little doubt that the cause of the greater part of the "cupellation loss" is a chemical one and cannot be counteracted by a mechanical contrivance.[13] In cupellation, then, there is a loss, apart from imperfect working, inherent in the process itself; and as the amount of this loss varies under different conditions, it is necessary to study it somewhat in detail.

The following experiments are taken without selection from the work of one student. Three experiments were made for each determination, and the mean result is given. By "range" is meant the difference between the highest and lowest result and the percentage loss is calculated on the silver present. The silver added in the lead used has been deducted.

~Effect of Varying Lead.~--In each experiment 0.4 gram of silver was taken and cupelled with the lead. The silver loss and "range" are expressed in milligrams.

------------+--------------+--------+------------------
Lead Used. | Silver Lost. | Range. | Percentage Loss.
------------+--------------+--------+------------------
Grams. | | |
10 | 6.5 | 1.0 | 1.62
20 | 7.0 | 1.0 | 1.75
40 | 12.0 | 1.5 | 3.00
60 | 12.7 | 0.5 | 3.17
------------+--------------+--------+------------------

The loss increases with the lead used.

~Effect of Varying Temperature.~--0.4 gram of silver was cupelled with 20 grams of lead.

Temperature. Silver Lost. Range. Percentage Loss.

Bright red 7.0 1.0 1.75
Clear yellow 17.3 1.7 4.32

The difference in temperature in these experiments was much greater than would occur even with careless work.

~Effect of Varying Silver.~--20 grams of lead were used in each cupellation.

---------------+--------------+--------+------------------
Silver Taken. | Silver Lost. | Range. | Percentage Loss.
----------------------------------------------------------
Milligrams. | | |
12.5 | 0.7 | 0.2 | 5.6
25.0 | 1.4 | 0.1 | 5.6
50.0 | 1.6 | 0.4 | 3.2
100.0 | 2.9 | 0.3 | 2.9
200.0 | 5.6 | 0.5 | 2.8
400.0 | 7.0 | 1.0 | 1.7
800.0 | 9.7 | 1.0 | 1.2
---------------+--------------+--------+------------------

It will be seen that, although the quantity of silver lost increases with the silver present, the percentage loss is greater on the smaller buttons.

The following results are often quoted:--Cupelling 1 grain of silver with 10 grains of lead, the loss was 1.22 per cent.; 10 grains of silver with 100 grains of lead, loss 1.13 per cent.; 25 grains of silver cupelled with 250 grains of lead, lost 1.07 per cent. The proportion of silver to lead was the same in the three experiments, and the largest button gave the best result. Evidently, if the quantities of lead had been the same in the three experiments (say, 250 grains in each case), the loss on the smaller quantities of silver would appear worse in the comparison.

In judging these results, it must be borne in mind that it is difficult to regulate the temperature, &c., in consecutive experiments so as to get exactly similar results, so that the range in consecutive cupellations is greater than that in a batch cupelled side by side.

~Effect of Copper and Antimony.~--0.1 gram of silver was cupelled with 20 grams of lead, and to one batch 0.5 gram of antimony, and to another 0.5 gram of copper was added.

Loss in
Silver Lost. Range. Percentage.

Without addition 2.9 0.3 2.9
With antimony 3.2 0.2 3.2
With copper 4.9 1.7 4.9

Perhaps the antimony has so small an effect because it is eliminated in the earlier part of the process, while the silver is still alloyed with, and protected by, a large proportion of lead; whilst the copper on the other hand makes its fiercest attack towards the close, when the silver is least capable of resisting it. The ill effects of copper are most strongly felt when the quantity of lead present is not sufficient to remove it: the coppery button of silver got under these conditions is very considerably less than the weight of silver originally taken.

Although the above is a fair statement of the loss attending average work, it will not do in very important and exact work to place too much reliance on the figures given, or, indeed, on any other set of figures, with the object of correcting the result of an assay. Each man must rely on his own work.

It is easy to determine what allowance must be made for the loss in cupellation by cupelling side by side with the assay piece an alloy of similar and _known_ composition. For, if the two pieces are very nearly alike, we may justly conclude that the loss on each will be the same; and if, further, we take the average of three or four such determinations we shall get results accurate within 0.1 per cent. The method of getting such results may be best explained by one or two illustrations. This method of working is termed "assaying by checks."

Suppose we have an alloy of silver and lead in unknown proportions and that by cupelling two lots of 10 grams each there is got from I. 0.1226 gram of silver, and from II. 0.1229 gram. We should know from general experience that the actual quantity of silver present was from 2 to 4 milligrams more than this. To determine more exactly what the loss is, the following plan is recommended:--The two silver buttons are wrapped up each in 10 grams of lead, and cupelled side by side with two other lots of 10 grams of the original alloy. If now the two buttons I. and II. weigh 0.1202 and 0.1203, they will have suffered in this second cupellation an average loss of 2.5 milligrams. Suppose the two fresh lots of alloy gave 0.1233 and 0.1235 of silver, the average loss on these would also be 2.5 milligrams. Add this loss to each result, and take the mean; which is in this case 0.1259.

If copper is present in the alloy as well as silver, it is necessary to add about the same quantity of copper to the checks as is supposed, or known, to be present in the assays. If the substance to be assayed is an alloy of silver and copper, first cupel 0.5 gram of it, with, say, 10 grams of lead, and weigh the resulting button of silver, in order to get an approximate knowledge of its composition. Suppose the button weighs 0.3935 gram. We know that this is below the truth: for the sake of round numbers take it as 0.4, and assume that the rest of the alloy (0.1 gram) was copper. Two check pieces are then weighed out, each containing 0.4 gram silver and 0.1 gram of copper wrapped in 5 grams of lead. Of course the silver must be pure. And there is also weighed out two (or better, four) assay pieces each containing half a gram of the alloy wrapped in 5 grams of lead. The whole lot are then cupelled as nearly as possible under the same conditions. With four assay pieces, the cupels should be placed close together in two rows of three across the muffle; the two check pieces are put in the middle cupels. Suppose the buttons of silver got weighed as follows:--

Check pieces I. 0.3940 II. 0.3945
Assay pieces I. 0.3905 II. 0.3912
III. 0.3910 IV. 0.3909

The average loss on the two check pieces is 5.7 milligrams, and the average result of the four assay pieces is 0.3909. Add the average loss to the average result, and there is got the corrected result, 0.3966. And if 0.5 gram of alloy contain 0.3966 of silver, 1000 will contain 793.2 of silver, and this is the degree of fineness.

A correction for the loss in cupellation is always made in this way when rich alloys are being assayed; and in the case of rich ores it may be done after the manner of the first of the above illustrations. There is another method of working which relies more on experiment. This is to smelt the cupel as described further on (p. 114), and to again cupel the resulting button of lead. The button of silver got in this second cupellation is added to that first obtained. It will sometimes, but not often, happen that the two buttons together will slightly exceed in weight the silver which was actually present. This is because of the retention in the buttons of a small quantity of lead. It has been stated that the proportion of lead thus retained may be as much as 1% of the silver present; this, however, can only be under exceptional conditions. A determination of the actual silver in the buttons got in the series of cupellations quoted on pages 102, 103, gave an average percentage of 99.85, so that even with the larger buttons the effect of the retained lead would be only to increase the weight by about 1 milligram. In the method of working with checks, the retained lead has no disturbing influence.

~The proportion of lead required~ for the cupellation of any particular alloy requires consideration. With too much lead the time occupied in the process is increased, and so is the loss of silver; on the other hand, too little lead is of greater disadvantage than too much. From 8 to 16 parts of lead are required for each part of silver alloy, or, if gold is present, about twice as much as this must be used. For the cupellation of 1 gram of a silver copper alloy containing different percentages of copper, the following quantities of lead should be used:--

Percentage of Copper
in Alloy. Lead Required.

5 6 grams
10 8 "
20 10 "
30 12 "
40 14 "
50-100 16-18 "

The alloy, in not too large pieces, is wrapped in the required weight of lead foil and charged into the cupel at once; or the lead may be put in first, and, when the cupellation has fairly started, the alloy may be added wrapped in tissue paper; or a portion of the lead may be first started and the alloy wrapped in the remaining lead and subsequently added. The cupellation of large quantities of alloy or of alloys which contain tin, antimony, iron, or any substance which produces a scoria, or corrodes the cupel, must be preceded by a scorification. The advantages of this are that the slag is poorer in precious metal than that found on a cupel and is more easily collected and cleaned; that larger quantities of metal can be treated, and that, even if the substance is in part infusible, or produces at the start a clinkery mass or scoria, the oxide of lead gradually accumulates, fluxes the solid matters, and produces a good final result; but if the oxide of lead by itself is not sufficient for the purpose, borax or some other flux can be easily added.

If the button of silver got is very small its weight may be estimated from its size; but it must be remembered that the weight varies as the cube of the diameter. If one button has twice the diameter of another it is eight times as heavy and so on. Scales specially constructed for measuring silver and gold buttons may be purchased; but it is much better to make the measurement with the help of a microscope provided with an eyepiece micrometer.

If the length of the long diameter of a silver button be taken the following table will give the corresponding weight in milligrams:--

------------+-----------++------------+---------
Diameter. | Weight. || Diameter. | Weight.
------------------------------------------------
0.04 inch | 3.6 || 0.015 inch | 0.19
0.035 " | 2.4 || 0.014 " | 0.15
0.03 " | 1.5 || 0.013 " | 0.12
0.025 " | 0.9 || 0.012 " | 0.097
0.02 " | 0.45 || 0.011 " | 0.075
0.019 " | 0.4 || 0.010 " | 0.056
0.018 " | 0.33 || 0.008 " | 0.028
0.017 " | 0.27 || 0.006 " | 0.012
0.016 " | 0.23 || 0.004 " | 0.004
------------+-----------++------------+---------

The weight of a corresponding button of gold is got by multiplying by 2.25. These figures are based on those given by Plattner, and apply only to buttons of such shape as those left after cupellation. A sphere of silver 0.01 inch in diameter would weigh 0.09 milligram, and a similar sphere of gold weighs 0.167 milligram.

It is safer, however, to compare with a micrometer the diameter of the button whose weight has to be determined with that of a standard button of nearly equal size whose weight is known. The weights of the two buttons are proportional to the cubes of their diameters. This plan of working is described more fully in Appendix B., page 440.

~Calculation of the Results.~--After deducting for the silver added, and correcting for the cupellation loss, the calculation is made in the usual way; reporting as so many parts per thousand in the case of rich alloys and as so many ounces and pennyweights, or better as ounces and decimals of an ounce, in the case of poor alloys and ores.

In this last case, however, it is less fatiguing to refer to a set of tables which give, either directly or by means of simple addition, the produce corresponding to any weight obtained from certain given weights of the substance. The following table gives the produce in ounces and decimals of an ounce per ton of 2240 pounds:--