Prof. V. Ya. Mostovich.

Cyanidation of Gold Tellurides

(From the Metallurgical Laboratory of the Tomsk Technological Institute).

Among gold-bearing ores, after pyritic ores, telluride ores are of very great importance in the metallurgy of gold; in these ores gold is chemically combined with tellurium in the form of the so-called tellurides, such as calaverite, sylvanite, krennerite, and others. If we accept the established fact that in almost all gold-bearing minerals and rocks known to us gold occurs in the free, metallic state, then gold tellurides are, for the present, the only widespread chemical compounds of gold in nature.

Given the great chemical similarity between selenium and tellurium, the possibility is not excluded that, analogous to tellurides, selenides of gold also exist, but the latter have so far not been isolated in a free and pure state; moreover, seleniferous deposits are very rare. Although selenium is found in noticeable quantities in the telluride ores of Goldfield and in some silver- and gold-bearing ores in Sumatra, it is very probable that in these ores selenium is associated with gold tellurides or with other components¹).

Gold tellurides are found by us in placer and vein deposits, where they are associated with both basic and acidic rocks, and especially often with sulphides, forming deposits of gold sulphotellurides. In his review of gold-bearing telluride ores, Kemp²) gives a rich bibliographic index of the deposits of the latter and says that tellurium is far more widespread in nature and is a more frequent companion of gold than is usually supposed. Gold and silver tellurides are so often associated with bismuth and its sulphotelluride, tetradymite, \(\mathrm{Bi_2(STe)_3}\), that the latter serves as an indicator of the presence of tellurides of the noble metals in a given deposit. The very frequent association of the three elements—gold, bismuth, and tellurium—in ores is also attested by R. Pearce³) and Sharwood⁴). The latter specially studied deposits of gold-bearing telluride ores and supplements Kemp’s list with a number of new deposits. Among the latter, “several deposits” in the Urals deserve attention, though Sharwood unfortunately does not indicate their exact location. In the article “On Some Natural Companions of Gold,” Lincoln⁵) says that gold in veins is associated with more than 173 minerals; if from this number one excludes the nonmetallic minerals and two common sulphides, pyrite and galena, it turns out that all the remaining minerals—companions of gold—contain the following elements: arsenic, bismuth, tellurium, and selenium, elements very close to one another in the periodic system, but very remote from gold.

Deposits of gold-bearing telluride ores are known in many states of North America and in Australia.

Of especially great industrial importance are the deposits at Cripple-Creek in the state of Colorado in the United States of North America, and Kalgoorlie and Coolgardie in Western Australia. How significant the latter localities are is evident from the fact that the gold industry of Australia, which yields \(12\%\) of the world’s gold, is based chiefly on the treatment of telluride ores, from which half of all Australian gold is obtained⁶).

The lack of more detailed information on the degree of distribution of tellurium in gold-bearing ores is explained, first, by the very small content of tellurium in such ores and, second, by the absence of sufficiently sensitive and reliable methods for detecting this element. And inter—

because, owing to the fact that gold tellurides, and the spongy brown gold produced upon their destruction, are extremely difficult to extract by ordinary metallurgical processes, the refractory character of certain ores may be due to the presence of tellurium in them, and, when choosing a method for treating such refractory ores, investigation of the question of the presence of tellurium in them may prove very important.

Let us take, for example, a rich telluride ore containing one ounce of gold per ton, or 13.17 zolotniks per 100 poods. Assuming that all the gold is combined with tellurium in the form, at least, of calaverite, $\mathrm{AuTe_2}$, the amounts of gold and tellurium in such a rich ore will be expressed as only 0.00343% for gold and 0.0044% for tellurium, although in reality the amount of tellurium in telluride ores is usually considerably greater than that required to bind all the gold in the form of one telluride or another, since part of the tellurium is combined with other elements, e.g., with silver, bismuth, lead, iron; nevertheless, the tellurium content of ores is usually measured in hundredths or tenths of a percent. Thus, for example, typical telluride ores of Kalgoorlie contain 0.03 to 0.1% Te, the ores of Crepple-Creek[^7] contain 0.02–0.2%, and the ores of South Dakota[^8] contain 0.012–0.024% Te with 0.3–0.5 ounce of gold per ton.

If, with respect to gold, fire-assay methods make it possible to determine its content in ores with an accuracy of up to 0.00003% (11 parts in 100 million), then hundredths of a percent of tellurium, in ordinary ore analysis, completely elude quantitative determination. The situation is no better with qualitative analysis. Thanks to the extraordinary sensitivity of the reaction for gold, the presence of the latter can still be clearly established at a content of 1 part by weight of Au in 5 to 100 million parts by weight of solution[^9]. For tellurium, however, we have no especially sensitive reactions, so that hundredths and tenths of a percent of tellurium may fail to give clear indications of the presence of this element in ores.

For testing minerals for tellurium, a test with sulfuric acid is used. When finely ground telluride is heated with concentrated sulfuric acid until white fumes appear, tellurium dissolves with the formation of a carmine-red compound, $\mathrm{TeSO_3}$, which serves as the reaction for tellurium. But the reliability of this, the only sensitive reaction, is limited by a number of conditions[^10]. The reaction is suitable only for pure, unoxidized tellurides; it is not applicable to samples containing oxidized tellurium compounds, many foreign sulfides—especially pyrite—or manganese compounds. The latter, under analogous treatment, give the same reaction owing to the formation of manganic acid. The compound $\mathrm{TeSO_3}$ forms only within a limited temperature range. All these circumstances greatly weaken the reliability of the method. Better results are given by the following modification of the method. A 50–100 g charge of ore with an addition of silver is subjected to crucible smelting for gold with litharge, ordinary fluxes, and a reducing agent. In this way, owing to the high solubility of tellurium in lead and silver, almost all the tellurium is collected in the lead button, and during cupellation of the latter it reveals its presence by special, sharply expressed phenomena. Tellurium[^10-a] lowers the surface tension of lead, as a result of which, toward the end of cupellation, the latter does not brighten, but spreads out drop by drop and breaks up into the smallest globules with a matte surface; the larger of these globules have the appearance of frozen beads and are surrounded by a violet ring. Holloway[^11] proposed the following method. The lead button from the crucible smelt is dissolved in nitric acid; the solution is diluted with water and, after precipitating tellurium on lead foil, the resulting precipitate is tested for tellurium by the sulfuric-acid reaction. In comparison with the accuracy and sensitivity of the qualitative and quantitative gold assay, the analytical—

methods with respect to tellurium and other analogous rare elements occurring in gold-bearing ores in the form of so-called traces are extremely imperfect. Owing to this circumstance, the indicated elements are often not determined at all, or escape determination; meanwhile these “trace” elements may be quite sufficient to bind all the gold in a given ore, whose quantity, when the gold is disseminated, is usually expressed by the ratio of 1 part by weight of gold to 100 or 500 thousand parts by weight of rock. The traces of rare elements may very often account both for the refractory character of ores and for the richness of the tailings from their treatment.

In the metallurgical treatment of ores, the rare elements contained in them in the form of traces gradually become concentrated and reveal themselves in appreciable quantities in the final products.

R. Pearce¹²) reports the following facts. The sulphide ores of Leadville were distinguished by great purity, representing almost pure pyrite; in them there were no traces of bismuth or tellurium; however, these elements were discovered during the metallurgical treatment of the indicated ores. Nearly pure gold-bearing quartz from one California deposit yielded its gold poorly on amalgamation. Only when the ore was concentrated were bismuth and tellurium found in the concentrates; in Pearce’s opinion these accounted for the refractory character of this ore.

V. O’Brien¹³), who investigated the ores of South Dakota, gives the following interesting data on the composition both of the ores themselves and of the zinc precipitates from extracts obtained by cyaniding these ores.

Material	% Te	% Se
Golden Reward ores	0.006	traces
In the zinc precipitates from these ores	0.407	0.107
Dakota ore	traces	traces
In the zinc precipitates from the ore	0.239	0.061

The selenium content in our Uralian cupriferous pyrites is so small that scarcely more than traces of selenium can be found in several kilograms of pyrite. In smelting these pyrites, the selenium is to a considerable extent concentrated in the copper, and during electrolytic refining of the latter it passes into the anode slime. At the present time the anode slime from electrolytic copper refining at many plants is the starting material for obtaining selenium and tellurium. Another interesting example of the occurrence of rare elements in ores is the presence of gallium, indium, and germanium in the form of traces in the zinc ores of the Joplin district in the state of Missouri. The rich zinc blendes of this district are distinguished by great purity. Only by spectroscopic examination of 64 samples of blende from various deposits of the district did Urbain¹⁴) find the presence of Ga in 59 samples, In in 41 samples, and Ge in 38 samples. In refining metallic zinc smelted from these ores, an alloy of gallium with indium was found in the lead residues from the secondary distillation.

The most important of the natural tellurides of gold are the following:

calaverite, $\mathrm{AuTe_2}$
sylvanite, $\mathrm{AuAgTe_4}$
krennerite, $\mathrm{AuTe_2}$, part of the Au replaced by Ag
petzite, $(\mathrm{AgAu})_2\mathrm{Te}$

Nagyagite—sulphotelluride of lead, gold, and silver of variable composition.

The following table gives analyses of gold-bearing tellurides[^15].

Analyses of gold-bearing tellurides.

%	A	B	C	D	E	F
Au	38.95	34.77	29.35	25.16	23.69	28
Ag	3.21	5.87	11.74	41.87	41.06	16
Cu	—	—	0.34	—	—	—
Fe	—	—	0.59	—	—	—
Te	57.27	58.60	58.91	33.21	32.0	56
Mo	—	—	—	0.08	—	—
Sb	—	0.65	—	—	—	—
F₂O₃	0.12	—	—	—	—	—
Insoluble	0.33	—	—	—	—	—
Total	99.88	99.89	100.93	100.32	96.75	100

A — Calaverite from Crepple Creek, Colorado
B — Krennerite from Nagyag, Hungary
C — Sylvanite from Boulder County, Colorado
D — Petzite from Calaveras County, California
E — Petzite from Ontario, Canada, according to Smith[^16]
F — Sylvanite from Crepple Creek, according to T. Rickard[^17]

A characteristic feature of natural tellurides of gold is the complexity and inconstancy of their composition, which varies greatly even in pure specimens of one and the same mineral. Thus, for example, calaverite, most closely corresponding to the formula $\mathrm{AuTe_2}$, contains from 40.6 to 42.7% gold and from 0.4 to 3% silver; the theoretical composition of $\mathrm{AuTe_2}$ corresponds to 43.68% gold. In sylvanite and petzite we encounter a strongly varying ratio of Au:Au, etc.

Ores containing tellurides of gold belong to the category of refractory ores, since the most commonly used methods of extracting gold are either not directly applicable to them or give a very low extraction. Thus, owing to the extreme brittleness of gold tellurides, their separation from the gangue by ore concentration is often accompanied by large losses in the slimes, into which part of the tellurides is carried. Direct amalgamation is likewise unsuitable, since gold tellurides are not wetted by mercury, cause “flouring” of the latter, and pass into the amalgamation tailings. All methods of ordinary cyanidation of these ores, according to general opinion, give a very low extraction of gold, since the latter is chemically combined with tellurium, which does not dissolve in potassium cyanide[^18]. In view of these circumstances, in practice a variety of processes have been used for extracting gold from telluride ores; of these, the most economical and those that have found wide application in centers for treating telluride ores have up to the present time been only two processes: the dry and the wet. The dry process, also called the Marriner process[^19], consists in preliminary roasting of the crushed ore and subsequent extraction of gold from the roasted ore by amalgamation or cyanidation, or by both processes in succession. The roasting of telluride ores, as in the case of sulphide ores, has the aim of destroying the gold tellurides.

remove the tellurium in the form of volatile tellurous anhydride, free the gold from its association with tellurium, and make it accessible to the action of mercury and potassium cyanide. The dry process has found application chiefly in America, in the Cripple Creek district, where, owing to cheap fuel and the introduction of large mechanical roasting furnaces, the cost of roasting has been brought down to an insignificant amount with a high percentage of gold extraction[^20]. The same process is encountered at some plants in Australia.

The wet process consists in the direct treatment of crushed ore with a mixture of solutions of potassium cyanide and cyanogen bromide, as an oxidizing reagent that accelerates the dissolution of gold. There are several variants of the wet process. Thus, for example, amalgamation, concentration, and classification are introduced into the sequence of operations. The variant most widespread in Western Australia is Diehl’s process[^21][^22]. The typical Diehl process includes the following operations. The ore is subjected to wet grinding and amalgamation. The pulp is classified and concentrated. The concentrates, as the richest and most refractory material, are dead-roasted, reground, and cyanided. The ore freed from concentrates is either converted into slime and the latter is agitated with a mixture of potassium cyanide and cyanogen bromide, or, after concentration, the pulp is classified into sand and slime and each of the products is bromocyanided separately: the sand goes to percolation, while the slime is agitated and filtered by means of filter presses or vacuum filters.

Both the dry and the wet process have their own advantages and disadvantages, and the choice of one or the other process is determined both by a number of technical conditions and, chiefly, by economic considerations.

In the dry method the most expensive operation is the roasting of the ore, which must be carried out to dead roast. Therefore this method is advantageous only with cheap fuel and for large-scale enterprises, where the construction of mechanical roasting furnaces of high capacity is possible.

The very composition of the ores also plays a very important role. Ores rich in carbonates, after roasting, acquire the property of setting, like cement, which is a great difficulty in subsequent treatment. Such ores, especially characteristic of the deposits at Kalgoorlie, have the following composition:

According to Knutsen[^22]	Value	According to James[^19]	Value	Notes
SiO₂	50–76%	SiO₂	50%
FeCO₃	10–30%	CaCO₃	6–17%
CaCO₃	9–38%	Fe	10	in the form of carbonates
Au	1–3 oz. per ton	Mg	1–5%	in the form of carbonates
		Al₂O₃	5–20%
		S	3–7%

To prevent setting, the roasted ore is subjected to wet grinding and amalgamation in pans, and only the tailings from amalgamation are cyanided. Regrinding and preliminary amalgamation are, in general, necessary, since gold tellurides are very fusible; as a result, during roasting they easily liquate and give large aggregates of gold in the form of lumps or beads, in which form the gold amalgamates with difficulty and dissolves very slowly in potassium cyanide. Thus, according to Rose[^23], the melting point of gold tellurides containing up to 69% Au lies below 688°; for example, calaverite with 43.59% Au melts at 452°. There are also indications that TeO₂, formed during the roasting of telluride ores, gives with iron oxides the corresponding salts [FeTeO₄ or Fe₂(TeO₄)₃], remaining partly in the ore in the form of films on the gold, which is a serious obstacle to the extraction of gold by the percolation method[^24].

When introducing the wet process in Western Australia for ores rich in carbonates, Diehl aimed to make the process cheaper, to avoid roasting and all the inconveniences and difficulties associated with it. For the Crepple-Creek ores, however, in which SiO₂ and Al₂O₃ predominate, while calcite and other carbonates are almost absent, the dry method proved more advantageous.

The wet process also has its advantages. First, the dissolution of gold is accelerated and the participation of oxygen is not required, so that aeration in any excessive form is unnecessary; second, no free alkali is formed during the process, which readily reacts with sulfides present in the ores and gives compounds capable of decomposing potassium cyanide; third, preliminary roasting of the ores is not required, which considerably cheapens the process and in some cases—for example, with ores rich in carbonates—has, as indicated above, a special advantage.

But bromocyanidation also has negative aspects. First, bromine cyanide is expensive, and therefore, like any auxiliary reagent, it can be used only when the additional expenditure on the reagent is compensated by a reduction in the consumption of potassium cyanide, a shortening of the ore-treatment time, or an increase in the percentage extraction of gold. Further, bromine cyanide is extremely sensitive to many substances found in the ore or in solutions. Thus, bromine cyanide is decomposed by alkalis, forming potassium cyanate, which is not a solvent of gold.

$$ \mathrm{BrCN} + 2\,\mathrm{KOH} = \mathrm{KBr} + \mathrm{KCNO} + \mathrm{H_2O} $$

Therefore, in bromocyanidation only a very insignificant content of free alkali in the solution is permissible, approximately 0.01% and even less. The process is carried out in two stages. First the ore is agitated with a solution of potassium cyanide in the presence of the required amount of free alkali until that portion of the gold which is soluble in potassium cyanide has passed into solution; then the free alkali is neutralized with sulfuric acid, the required amount of bromine cyanide is added, and agitation is continued until the maximum extraction attainable under the given conditions is obtained; or the ore is first agitated with a potassium cyanide solution, then agitation is continued with bromine cyanide, and the alkali is added toward the end of the agitation, not long before filtration of the pulp²⁵). Further, bromine cyanide reacts rapidly with pyrite and metallic iron, which leads to useless consumption of an expensive reagent. Pyrite and, especially, metallic iron are rarely absent from ores that have been subjected to grinding.

In the dry method one product is obtained—the roasted ore—which is cyanided in the usual manner; in the wet method two or three products are obtained (concentrates, sand, slimes), which makes it necessary to apply two or three separate processes.

All these circumstances complicate bromocyanidation, and conducting this process requires much care and experience, which, understandably, makes the process itself more expensive. Therefore at present, at most plants treating telluride ores, the conclusion has been reached that the dry method is ultimately cheaper than the wet one.

Where the greater part of the gold is combined with tellurium, dry grinding and complete roasting are employed. Where the wet method has already been established, the ore is nevertheless subjected to careful concentration before bromocyanidation; selected concentrates are roasted and subjected to ordinary cyanidation, while the tailings are reground into slimes and cyanided with the addition of bromine cyanide.

It was indicated above that the chief reason for seeking new methods of extracting gold from telluride ores was the refractory nature of gold tellurides with respect to potassium cyanide. On this we encounter

in the literature the most contradictory data. According to the assertion of the majority of metallurgists, gold tellurides do not dissolve at all in potassium cyanide, or this solubility is very slight.^26) F. Smith^27) asserts that potassium cyanide dissolves both gold and tellurium in gold tellurides. In one of his recent articles on cyanidation at the Nipissing plants in Canada, G. Clevenger^28) says that the data of experiments and of plant practice show that the established opinion regarding the insolubility of gold and silver tellurides in potassium cyanide is not entirely correct.

The question of the solubility of pure compounds of gold, in particular tellurides, in potassium cyanide was studied only by Christy^29). Thus, according to his observations, freshly precipitated aurous sulfide, $\mathrm{Au_2S_2}$, dissolves extremely readily in potassium cyanide even in the absence of oxygen, according to the reaction:

\[ \mathrm{Au_2S_2 + 4KCN = 2KAu(CN)_2 + K_2S + S,} \]

since gold has a great affinity for cyanogen, and potassium for sulfur. At the same time, gold tellurides obtained by fusing tellurium with gold dissolve with extreme difficulty in potassium cyanide even in the presence of oxygen. Christy explains this phenomenon by the fact that the affinity of the ion $\mathrm{K}$ from $\mathrm{KCN}$ for tellurium in gold telluride is much weaker than for the sulfur ion in aurous sulfide. In addition, in its tendency to be oxidized to $\mathrm{TeO_2}$, tellurium absorbs oxygen, thereby slowing the oxidation of potassium and, consequently, the dissolution of gold; for the transfer of gold in the presence of potassium cyanide into a soluble complex it is necessary to oxidize the potassium in one molecule of potassium cyanide according to the scheme:

\[ \begin{array}{c} \mathrm{KCN}\\ \mathrm{KCN} \end{array} \quad \mathrm{Au} \quad +\quad \begin{array}{c} \\ \\[-0.6em] \mathrm{KCN}\\ \mathrm{KCN} \end{array} \quad \mathrm{Au} \quad +\mathrm{O} = \mathrm{2KAu(CN)_2 + K_2O} \]

On the other hand, certain phenomena in the process of formation of tellurides and the properties of the latter raise doubts as to whether gold tellurides are in fact stable chemical compounds of tellurium with gold, or whether the latter is not chemically bound but is present in a mixture or alloy with tellurium^15).

1) Thus, in natural gold tellurides we observe neither constancy of composition nor correspondence of the latter to chemical formulas, as we have in the analogous sulfides of most metals, for example copper, lead, zinc, etc. The fluctuations in the composition of natural tellurides are very great, and deviations from the composition of normal compounds are very considerable even in the purest mineral specimens.

2) According to the investigations of V. Lenher,^30) both elemental tellurium and natural and artificial gold tellurides quantitatively precipitate metallic gold from a solution of gold chloride according to the reaction:

\[ \mathrm{4AuCl_3 + 3Te = 4Au + 3TeCl_4} \]

In Lenher’s opinion, the question arises whether a true chemical compound can precipitate one of its components from a solution of the same compound? Sharwood^31) does not see in this reaction proof that gold tellurides are not true compounds, since, although cases of precipitation expressed by the equation: $\mathrm{AB + AC = 2A + BC}$ are not especially frequent, we know quite a few examples of such reactions. For example, when moist $\mathrm{SO_2}$ and $\mathrm{H_2S}$ are mixed, sulfur is liberated:

\[ \mathrm{2H_2S + SO_2 = 2H_2O + 3S} \]

In metallurgical processes we know many examples of reduction of a metal through the interaction of two of its compounds. Such, for example, are all reactions between the sulfides and oxides of copper and lead that take place in reaction smelting:

\[ 2\mathrm{PbO} + \mathrm{PbS} = 3\mathrm{Pb} + \mathrm{SO}_2 \]

\[ 2\mathrm{Cu}_2\mathrm{O} + \mathrm{Cu}_2\mathrm{S} = 6\mathrm{Cu} + \mathrm{SO}_2 \]

These objections of Sharwood are valid, since they concern reactions proceeding at high temperatures; at ordinary temperature and in aqueous solutions, reactions analogous to the precipitation of gold from a solution of gold chloride by tellurides are nevertheless an exception. In any case, it would be more correct to suppose that in this reaction the precipitant of metallic gold from its solution is not the gold tellurides, but tellurium, and that in the present case we have cementation of gold solutions by tellurium, similar to the cementation of copper and silver solutions by metallic iron and copper. With such cementation, and with an excess of the precipitant (Te), the precipitated gold will contain free tellurium in greater or lesser quantity. It is possible that in nature, with respect to gold-bearing solutions, tellurium and tellurides play the role of mineralizers, and that this may explain the varying composition of natural tellurides, which are, in a sense, cement gold deposited in the mass of the tellurides or on their surface. And indeed, Lenher often observed specimens of native gold together with tellurides, and a gold incrustation in the form of a film on the surface of crystals of natural tellurides.

1. Gold tellurides are extremely unstable compounds. Even in the cold, weak nitric acid readily dissolves the tellurium in tellurides, while pure gold remains in the residue. Such a weak solvent as \(\mathrm{S}_2\mathrm{Cl}_2\) quantitatively extracts tellurium from gold tellurides:

\[ \mathrm{AuTe}_2 + 4\mathrm{S}_2\mathrm{Cl}_2 = \mathrm{Au} + 2\mathrm{TeCl}_4 + 4\mathrm{S}_2, \]

the sulfur dissolving in the excess \(\mathrm{S}_2\mathrm{Cl}_2\).

1. Lenher attempted to prepare the gold telluride \(\mathrm{AuTe}_2\), as most nearly corresponding to a chemical compound, by both the dry and the wet method. On heating a mixture of gold with an excess of tellurium in a stream of carbon dioxide, the tellurium was gradually distilled off in an amount depending on the temperature and duration of heating. The residue obtained was a mixture of metallic gold and an alloy of gold with tellurium, the composition of which depended on the conditions of the experiment. Under the action of hydrogen telluride on gold chloride, only metallic gold is precipitated. Likewise, it proved impossible to obtain gold tellurides by the simultaneous precipitation of both elements from a mixture of solutions of gold chloride and tellurium tetrachloride by means of sulfurous gas. First pure gold was precipitated, then tellurium. All these facts are difficult to reconcile with the conception of gold tellurides as chemical compounds.

In view of the foregoing, it seemed of interest to clarify the question of the influence of tellurium on the solubility of gold in potassium cyanide, especially since the quantitative aspect of the question has so far not been touched upon by anyone. We have abundant material from plant practice, where telluride ores, i.e., complex systems in which the solubility of gold is affected both by many components of the ore and by the working conditions, are subjected to the action of potassium cyanide solution. It is possible that these circumstances explain the extreme contradictions that we encounter in the conclusions of metallurgists concerning the solubility of gold tellurides in potassium cyanide. Since I did not have at my disposal specimens of natural gold tellurides in sufficient quantity for experiments, I had to turn to the investigation of alloys of gold with tellurium.

The system \(\mathrm{Au}:\mathrm{Te}\) was studied by Rose\(^{23}\) thermally and microscopically. Rose fused the pure elements in various proportions in a fireclay crucible under a layer of charcoal, took freezing curves of the alloys, and studied their structure under ...

microscope. In Fig. 1 the constitution diagram of the Au : Te system is reproduced for the concentration range from 0 to 100% of each of the components. From the diagram it is evident that gold forms with tellurium only one chemical compound, AuTe₂, containing 43.59% Au, with a melting point of 452°, and two eutectics with 20 and 60% Au and melting points of 397° and 432°; in composition the first eutectic corresponds to the formula AuTe₆, the second—to AuTe. In alloys containing up to 43.53% Au, tellurium is the free structural component; in alloys richer in gold, metallic gold is the free component. The melting point of pure tellurium was determined by Rose as 440°. All alloys containing up to 60% Au melt within the temperature range 397–452°. The presence of a eutectic was traced in alloys with a content down to 0.45% Te. Thus, by means of thermal analysis Rose was able to establish in the Au : Te system only one definite chemical compound, AuTe₂, to which in composition correspond the minerals sylvanite or calaverite; in the latter, part of the Au is often replaced by silver. Thermal analysis could not confirm the existence in the system of other compounds that would correspond to the minerals petzite, Ag₂Te, krennerite, Me₂Te₃, or calaverite, Me₅Te₃, Me₃Te₂.

The composition of natural gold tellurides lies in the left-hand part of the diagram up to and including the compound AuTe₂. I studied the solubility in potassium cyanide of two series of alloys containing from 0 to 1% and from 30 to 90% Te. In the alloys of the first series, according to Rose’s diagram, the tellurium-containing component is the eutectic AuTe₂ : Au; in the alloys of the second series, with a content up to 42.59% Au, there is no free gold; the free component in them is tellurium; in alloys from 43.59% Au and higher, gold is the free component.

The need to include in the research program the study of the solubility of alloys containing up to 1% Te was dictated by the circumstance that usually even minimal impurities of certain elements have a pronounced effect on the physical and mechanical properties of pure metals. Here the influence of traces on the masses is manifested, and this influence of impurities is the stronger, the greater their atomic volume. Numerous examples of this are found in copper, zinc, lead, silver, gold, and other metals. Thus, 0.05% Pb or 0.03% Bi causes hot shortness in deoxidized copper (deprived of oxygen); 0.05% Bi makes copper cold-short; 0.0013% As or 0.0071% Sb lowers the electrical conductivity of copper by 1%³²). In the case of gold, Bi, Te, and Pb strongly reduce the strength and ductility of gold, increasing its hardness and brittleness. Thus, 0.01% and less of lead makes standard gold brittle and unsuitable for coining. The influence of Bi and Te is especially strong; 0.05% Bi and 0.05–0.1% Te cause noticeable brittleness of gold³³); at 0.5% Te, gold can be rolled only with difficulty. This strong and identical influence of Te and Bi is partly understandable if one compares the atomic volumes of gold, tellurium, and bismuth. These volumes are related as 10.2 : 20.48 : 21.32; the last two elements, Te and Bi, possessing similar atomic volumes, must exert an analogous influence on gold.

In view of such a sharp influence of a small quantity of impurities of foreign elements on gold, it did not seem excluded that these impurities might have the same influence also on the chemical properties of gold, in particular on its solubility in potassium cyanide.

Experimental Part.

1. Solubility of gold alloys with 0–1% tellurium in potassium cyanide.

The starting materials for the preparation of these alloys were chemically pure gold and tellurium from Kahlbaum. The gold was obtained by precipitating a pure solution of AuCl₃ with oxalic acid and by remelting the precipitate in a graphite crucible in an electric kryptol furnace. The purity of the tellurium was verified--

verified by analysis. For this purpose, weighed portions of 0.3–0.5 g of tellurium were dissolved in nitric acid (sp. gr. 1.2); the solution was first evaporated to dryness, and then twice with concentrated hydrochloric acid with the addition of NaCl until the \(HNO_3\) was completely removed. The residue was dissolved in HCl of sp. gr. 1.175, and the hot solution was precipitated with sulfurous gas.

Under these conditions selenium precipitates, but none was found. The solution was then diluted with an equal volume of hot water and, while boiling, saturated with \(SO_2\) until the solution above the flocculent precipitate of tellurium became completely clear. The latter was washed in a Gooch crucible with water until the HCl was removed, then with alcohol, and dried at \(105^\circ\) to constant weight. The preparation proved to be 100% pure.

The introduction of small quantities, usually 100–200 mg, of powdered tellurium into superheated molten gold in order to obtain alloys containing up to 1% Te would, because of the low melting point, easy oxidizability, and volatility of tellurium, have involved a considerable loss of the latter. Therefore an alloy of gold with 40% Te was first prepared, and calculated quantities of this alloy, in powder form, were introduced into molten gold, which after the addition of tellurium was stirred with a carbon rod. The method described gave very satisfactory results, as shown by analysis of the prepared alloys. The alloys were rolled without annealing into plates 0.1 mm thick. Even 0.3% Te made the alloy hard, and cracks appeared at the edges of the plate; at 0.5% Te it was necessary to roll with extreme care, only very slowly reducing the thickness of the plate, since the latter tore and showed great brittleness. The tellurium content in the alloys was determined by fire assay. A 150 mg weighed portion of alloy turnings, together with 600 mg of pure silver, was immersed in 40 g of assay lead melted on a scorifier. Scorification was carried out at a bright-red heat and very slowly, in order, as far as possible, to oxidize all the tellurium. The lead button was cupelled, and the bead of silver-bearing gold was parted in the usual way. In parallel with the tests for determining the losses of gold, two weighed portions of pure gold were subjected to all the operations described above. The Te content was determined from the loss in weight of the alloy and with correction for gold. For 150 mg of pure gold, the average loss was 0.045 mg.

In the alloys prepared, analysis showed the following quantities of tellurium: 0.11, 0.57, and 1.14%.

From the middle part of the plate of each alloy, disks 16.1 mm in diameter were punched out by means of a steel tube with a sharpened edge. The edges of the disks were turned on emery paper; in each disk, near the edge, a hole was pierced with a needle, and the disks were suspended on glass hooks fused to glass rods. The systems were placed in Erlenmeyer flasks with a KCN solution. For each alloy two tests were set up. In view of the fact that the dissolution of gold was influenced, in addition to the area of the disks, by the amount and strength of the solution, the relative position of the disks in the solution, and the temperature, all these conditions were maintained, as far as possible, identical in each series of experiments. Before immersion in the solutions, the disks were boiled in concentrated HCl, washed with water and alcohol, dried at \(110^\circ\), and weighed on gold balances with an accuracy of 0.025 mg. After a definite time had elapsed, the disks were removed from the solutions, washed with water, alcohol, and ether, dried at \(110^\circ\), and weighed.

The amount of gold that had passed into solution was determined from the loss in weight of the disks, and in some experiments by direct precipitation of gold from the solutions. Usually these results agreed well. Under the action of the potassium cyanide solution the alloy disks darkened, and the degree of darkening increased with the increase in the tellurium content of the alloy. Black spots appeared on the surface of the disks. On the disks of alloys with 0.58 and 1.14% Te, noticeable

was the thinnest mesh of elemental tellurium, which separated from the disk and was washed off with water. For these alloys, the solubility of gold was determined by the quantity of the latter in the solutions.

In parallel, for comparison, two samples of pure gold were included in each series of tests. All the experiments were carried out with the same set of disks. The results of the experiments are given in the following tables; in them the Nos. of the alloys indicate the tellurium content in the corresponding alloys; thus, Au₁ and Au₂ are pure gold; 0.37a is the first, 0.37b the second sample with 0.37% Te, etc.

Table 1.

I Series. Amount of solution 200 g. Strength of solution 0.24% KCN. Duration of experiment 14 h.	I Series. Amount of solution 200 g. Strength of solution 0.24% KCN. Duration of experiment 14 h.	I Series. Amount of solution 200 g. Strength of solution 0.24% KCN. Duration of experiment 14 h.	II Series. Amount of solution 200 g. Strength of solution 0.11% KCN. Duration of experiment 14 h.	II Series. Amount of solution 200 g. Strength of solution 0.11% KCN. Duration of experiment 14 h.
No. of alloy	Weight of disk, mg	mg Au in solution	Weight of disk, mg	mg Au in solution
Au₁	177.6	15.05	151.4	12.95
Au₂	160.3	16.85	135.1	12.50
0.11a	165.0	14.25	138.35	13.30
0.11b	172.0	14.35	143.9	13.50
0.37a	196.9	14.0	169.25	13.10
0.37b	252.9	12.9	219.8	12.85
0.57a	181.1	12.81	153.3	12.02
0.57b	175.64	13.67	148.65	12.97
1.14a	197.2	12.45	175.2	12.01
1.14b	172.4	15.20	143.7	12.80

III Series. Amount of solution 200 g. Strength of solution 0.11% KCN. Duration of experiment 22 h.	III Series. Amount of solution 200 g. Strength of solution 0.11% KCN. Duration of experiment 22 h.	III Series. Amount of solution 200 g. Strength of solution 0.11% KCN. Duration of experiment 22 h.	IV Series. Amount of solution 200 g. Strength of solution 0.24% KCN. Duration of experiment 20 h.	IV Series. Amount of solution 200 g. Strength of solution 0.24% KCN. Duration of experiment 20 h.
No. of alloy	Weight of disk, mg	mg Au in solution	Weight of disk, mg	mg Au in solution
Au₁	138.15	20.70	117.45	17.90
Au₂	122.05	20.25	102.0	15.95
0.11a	—	—	125.05	15.95
0.11b	—	—	133.70	14.70
0.37a	—	—	157.60	15.0
0.37b	—	—	211.40	14.90
0.57a	143.15	19.70	125.05	15.70
0.57b	137.80	20.20	117.60	15.60
1.14a	161.6	19.45	144.20	15.60
1.14b	129.4	19.55	133.0	15.10

From the table we see that tellurium in an amount up to 1.14% does not affect the solubility of gold in potassium cyanide. The latter, under otherwise identical conditions, increases only with an increase in the duration of the experiments.

Since in these experiments the amount of gold capable of dissolving is determined by only two factors—the size of the surface of the disks and the tellurium content in the alloys—and since the surface of the disks immersed in the solution is almost the same for all alloys, the influence of tellurium on the solubility of gold will appear more clearly if we compare the figures expressing the amount of gold dissolved from a unit surface, for example, from one square centimeter of disks with different Te contents over the same inter-

interval of time, for example, in 10 hours. If the very small lateral surface of the disks is neglected, in view of their insignificant thickness (about 0.1 mm), then the immersed surface of each disk, with a diameter of 1.61 cm, is 4.0716 sq. cm. Dividing by this value the total amount of gold that passed into solution from each disk, we obtain the relative solubility per 1 sq. cm. These data are given in Table 2.

Table 2.

No. of alloy	Initial weight of disk, mg	Final weight of disk, mg	Total quantity of dissolved gold, mg	Duration of experiments, hours	Gold dissolved from 1 sq. cm, mg	Gold dissolved from 1 sq. cm in 10 hours, mg
Au₁	177.6	99.55	66.60	70	16.36	2.34
Au₂	160.3	86.05	65.55	70	16.10	2.30
0.11a	165.0	109.10	43.50	48	10.68	2.23
0.11b	172.90	119.0	42.55	48	10.45	2.18
0.37a	196.90	142.60	42.10	48	10.34	2.15
0.37b	252.90	190.50	40.65	48	9.98	2.08
0.57a	181.10	109.35	60.22	70	14.79	2.11
0.57b	175.64	102.0	62.44	70	15.34	2.19
1.14a	197.20	128.60	59.51	70	14.62	2.09
1.14b	172.40	97.90	62.65	70	15.38	2.19

A comparison of the figures in the last column of Table 2 shows quite clearly that, in alloys with gold, tellurium in amounts up to 1.14% does not affect the solubility of gold in potassium cyanide. It is evident that in these experiments the gold was under conditions very unfavorable for dissolution. Around the disks of alloy there formed a layer with a high concentration of gold, and the latter could pass into the surrounding medium by diffusion. The absolute amount of gold that could dissolve depended on the rate of diffusion, since there was always a large excess of potassium cyanide in the solutions. But since dissolution proceeded under identical conditions, the circumstance indicated above could affect only the absolute, and not the relative, amounts of dissolved gold, on the basis of which we could draw conclusions about the influence of tellurium on the solubility of gold in potassium cyanide.

2. Extraction of gold from tellurides containing from 10 to 70% gold.

The tellurides for these experiments were prepared by melting calculated quantities of powdered tellurium and shavings of pure gold in borosilicate-glass test tubes. First the mixture was heated on a small flame of a Bunsen burner until melting began; then the alloy was covered with a layer of sugar charcoal, overheated to approximately 600°, and thoroughly stirred with a thin carbon rod. Under these conditions it was possible to obtain alloys of nearly the calculated composition with minimal loss of tellurium. All tellurides up to the compound AuTe₂ (43.59% Au) were very brittle and were easily ground in an agate mortar to fine powders. The brittleness of the alloys increased with increasing tellurium content. As the gold content increased, the dark-gray color of the alloys became lighter; the alloys became harder, showing noticeable malleability. On examination under the microscope, in alloys with a content up to 43.59% Au no separation of the latter was observed, and the ground alloys passed entirely through a Tyler 150-mesh sieve (size of the opening—

of aperture 0.104 mm). In alloys containing more than 43.59% Au, the separation of small globules of gold was clearly visible.

When the alloy corresponding to eutectic B (60% Au) was ground and screened through a 150 sieve, the following results were obtained: 489.9 mg of alloy yielded 38.8 mg of lumpy product +150, which could not be further ground. This product contained 82.99% Au, while the −150 material contained 59.33% Au.

In view of the extremely limited quantity of materials, especially tellurium, that I had at my disposal, it was necessary to work with the smallest possible amounts of reagents, to prepare alloys each time weighing 1–2 g, and therefore the possibility of checking the solidification curve of the Au : Te system given by Rose was precluded.

But microscopic study of the fracture of the alloys and their behavior on grinding partly confirm Rose’s data.

The analysis of the alloys was carried out by the dry and wet methods. In the dry method, a weighed portion of alloy powder of 25–50 mg was wrapped in lead foil and lowered into 40 g of assay lead melted on a scorifier. After closing the muffle, the lead button, with an addition of 200 mg of silver, was cupelled and the gold was determined in the usual way. In the wet method, a weighed portion of alloy of 25–50 mg was moistened in a porcelain crucible with water and decomposed with dilute nitric acid. Dissolution was begun cold; the assay was heated until the tellurium had completely dissolved; the gold was thoroughly washed with hot water, dried, and weighed.

Both methods of analysis give, as Rose also indicates, identical results. Thus, for example:

	dry method		wet method
alloy D	20.6 20.7	} 20.65% Au	20.6 20.6	} 20.6% Au
B	60.3 60.0	} 60.15% »	60.0 60.0	} 60.0% »
Au Te₂	41.90 41.70	} 41.8% »	41.8 41.8	} 41.8% »

With careful work, the wet method gives, for several assays, almost always coincident results, owing to the fact that losses of gold in it, because of the limited number of manipulations, are smaller and less possible than in the dry method. Since, moreover, the performance of the method is distinguished by great simplicity and rapidity, subsequently, in testing all alloys, only the wet method was used.

To determine the influence of tellurium on the solubility of gold in potassium cyanide, tellurides were selected with gold contents of 10, 20 (eutectic D), 30, 43 (chemical compound AuTe₂), 50, 60 (eutectic B), and 70%. Only material that had passed through a 150 sieve was studied. Since it was important to establish the relative extraction of gold from the various tellurides by means of potassium cyanide, the experiments were carried out under identical conditions as regards the grain size of the tellurides, the quantity and strength of the solutions, the duration of treatment, the temperature, and the total gold content in the separate assays, for which purpose weighed portions of alloys containing the same quantity of gold were taken.

In parallel, assays of pure powdered gold were set up; this gold had been obtained by rubbing gold precipitated by oxalic acid through a 150 sieve with the aid of a soft brush.

The criterion for determining the degree of extraction of gold from tellurides is the rate of dissolution of gold, and therefore the only variable factor in the experiments was the duration of dissolution.

For the experiments, a weighed portion of the alloy was placed in an Erlenmeyer flask with a potassium cyanide solution. The flask was closed with a stopper through which a glass stirrer with a four-bladed propeller was passed. The stirrer was supported in a glass bearing and was set in rotation by means of a mounted pulley driven by a caloric motor.

After a definite interval of time had elapsed, the precipitate was filtered off and washed through a previously dried and weighed filter. The precipitate with the filter was dried and weighed. Gold was determined both in the solution and in the precipitate, and from these data the extraction of gold and tellurium was calculated. To determine gold in the undissolved residue, the latter was incinerated together with the filter on a bedding of fine quartz on a scorifier in the muffle furnace mouth. The precipitate was then roasted until the vapors of $\mathrm{TeO_2}$ had been completely removed, scorified with 40 g of assay lead and 0.5 g of borax. The resulting button was cupelled with the addition of 200 mg of silver, and the bead of argentiferous gold was subjected to parting. From the cyanide solution the gold was precipitated with zinc dust. Since tellurium was precipitated together with the gold, the precipitate $\mathrm{Au + Te + Pb}$ was first scorified with the addition of assay lead. For each alloy, 3–4 extractions were performed.

In view of the contradictory data in the literature on the solubility of tellurium in potassium cyanide$^{13,25}$, this solubility was determined in special experiments. Tellurium ground to 150 mesh was agitated with a KCN solution of the same strength as that used in the experiments with tellurides. From the weight of the undissolved residue the amount of dissolved tellurium was determined. The following results were obtained:

250 mg of tellurium were agitated with 100 cc of KCN solution containing 0.24% KCN.

After agitation for 6 h, 9.8 mg of Te dissolved, or 3.9% of the charge.

” ” ” 24 ” 15.2 ” ” ” 6.1% ” ”

The experiments show that the solubility of free tellurium in potassium cyanide is very small.

From solutions, zinc dust precipitates tellurium in the form of a reddish-brown turbidity. It is quite possible that, on dissolving in KCN, tellurium gives with the latter the compound $\mathrm{KCNTe}$, analogous to $\mathrm{KCNS}$.

The solubility of tellurium in potassium cyanide explains the considerable accumulation of tellurium in working cyanide solutions. When these solutions are precipitated in extractors, tellurium is precipitated together with the gold and is gradually concentrated in the zinc precipitates, which causes brittleness of the gold pressed out from them, and sometimes the formation over the gold of a matte of tellurides of silver and gold. This phenomenon is observed in the Diehl process, and Knutsen$^{22}$ gives an example of such a matte containing 30% Ag, 30% Au, and 40% Te.

The results and details of the experiments are given in Tables 3 and 4.

The following alloys were subjected to investigation:

Alloy No.	% Au	% Te
10	9.91	90.09
D	21	79
30	28.8	71.2
$\mathrm{AuTe_2}$	43.5	56.5
50	49.7	50.3
B	60.5	39.5
70	67.8	32.2

Table 3.

Amount of solution — 100 cu. cm. Strength of solution 0.11% KCN.
Temp. 15° C. Duration of agitation 5.5 hours.

Alloy No.	Weight, mg	Content Au, mg	Content Te, mg	Passed into solution Au, mg	Total residue Au + Te, mg	Content in residue Au, mg	Content in residue Te, mg	Sum of Au in solution and residue, mg	Passed into solution Te, mg	% extraction Au	% extraction Te
10	504,6	50	454,6	23,55	474,6	26,2	448,4	49,75	6,20	47,1	1,19
10	»	»	»	30,0	463,2	20,1	443,1	50,10	11,50	60,0	2,63
D	250	52,5	197,5	20,6	223,8	32,3	191,5	52,90	6,0	39,24	3,04
D	»	»	»	27,4	216,2	25,7	190,5	53,10	7,0	52,2	3,55
D	238,1	50,0	188,1	32,8	194,4	17,6	176,8	50,4	11,3	65,6	6,07
D	»	»	»	27,0	203,2	24,55	178,65	51,55	9,45	54,0	5,02
30	173,6	50	123,6	17,0	153,6	33,8	119,8	50,8	3,8	34,0	3,07
30	»	»	»	30,5	133,0	20,3	112,7	50,5	10,9	61,0	8,87
43,5	124,4	54,12	70,28	24,0	92,4	30,3	62,1	54,3	8,18	44,35	11,64
43,5	»	»	»	28,3	87,6	26,4	61,2	54,7	9,08	52,3	12,92
43,5	115,0	50,0	65	24,2	83,4	25,9	57,5	50,1	7,5	48,4	11,54
43,5	»	»	»	19,1	90,6	30,9	59,7	50,0	5,3	38,2	8,15
50	101	50	51	35,7	56,2	13,6	42,6	49,3	8,4	71,4	16,47
50	»	»	»	38,7	51,8	11,3	40,5	50,0	10,5	77,1	20,59
B	86,4	51,18	35,22	39,6	38,2	10,8	27,4	50,4	7,82	77,35	22,2
B	»	»	»	43,5	35,8	8,2	27,6	51,7	7,62	84,96	21,64
B	82,65	50	32,65	36,15	40,4	12,85	27,55	49,0	5,1	72,3	15,62
B	»	»	»	41,0	33,8	9,15	24,65	50,15	8,0	82,0	24,5
70	73,5	50	23,5	37,2	31.8	12,3	19,5	49,5	4,25	74,4	17,9
70	»	»	»	29,3	41,8	20,7	21,1	50,0	2,65	58,6	11,16
Au	53	53	—	52,8	--	—	—	52,8	--	100	—
Au	50	50	—	50,0	—	—	—	50	—	100	—

Under the conditions of the experiments, 50–53 mg of powdered gold dissolved completely in the course of approximately one hour.

Table 4.

Agitation duration 24 hours; the remaining conditions are the same (see Table 3).

Alloy No.	Weighed amount, mg	Content: Au, mg	Content: Te, mg	Passed into gold solution, mg	Total residue Au + Te, mg	Content in ° atm.: Au, mg	Content in ° atm.: Te, mg	Sum of Au in solution and residue, mg	Passed into tellurium solution, mg	Extraction, %: Au	Extraction, %: Te	Average extraction, %: Au	Average extraction, %: Te
9,53	524,6	50	474,6	46,4	434,6	—	—	—	43,6	92,8	9,19	95,1	7,71
9,70	516,1	50	466,1	48,7	438,4	1,3	437,1	50	29,0	97,4	6,22	95,1	7,71
19,9	251,3	50	201,3	42,3	179,4	7,0	172,4	49,3	28,9	84,6	14,36	87,2	55,4
19,9	»	»	»	44,9	173,2	5,0	168,2	49,9	33,1	89,8	16,44	87,2	55,4
29,5	169,5	50	119,5	47,1	92,8	2,6	90,2	49,7	29,3	94,2	24,52	94,7	25,23
29,5	»	»	»	47,55	90,8	2,25	88,55	49,8	31,0	95,1	25,94	94,7	25,23
43,5	115,0	50	65	44,75	51,4	5,9	45,5	50,65	19,50	89,5	30,0	88,4	31,58
43,5	»	»	»	43,65	50,2	6,75	43,45	50,40	21,55	87,3	33,16	88,4	31,58
49,75	100,5	50	50,5	46,4	38,0	3,65	34,35	50,05	16,15	92,8	31,98	93,5	35,20
49,75	»	»	»	47,1	34,0	2,90	31,1	50,0	19,4	94,2	38,42	93,5	35,20
59,33	84,28	50	34,28	48,1	11,6	2,3	9,3	50,4	25,0	96,2	72,93	93,4	67,87
»	»	»	»	46,75	17,0	4,65	12,35	51,4	21,53	93,5	62,81	93,4	67,87
»	»	»	»	45,2	18,8	4,7	14,1	49,9	20,18	90,4	58,87	93,4	67,87

Turning to the analysis of the data obtained in the experiments, we can draw the following conclusions:

1. A potassium cyanide solution of 0.11% strength extracts both gold and tellurium from gold tellurides.

2. The quantity of tellurium passing into solution from the tellurides during 5.5-hour agitation does not exceed the solubility of free, elemental tellurium under the same conditions.

3. If one compares the highest and mutually closest figures for the percentage extraction of gold for alloys of different concentration, one may note an increase in the extraction of gold on both sides of the alloy \(AuTe_2\), as is evident from the following comparison:

Gold content in the alloy, %	Gold extraction, %
10	60,0
20 (D)	58,0
30	61
\(AuTe_2\)	50
50	74
60 (B)	83,5
70	74,4

1. In general, with 5.5-hour agitation, strongly differing results are obtained; when the duration of agitation is increased to 24 hours, under otherwise identical conditions, these results even out and agree quite satisfactorily for a series of experiments. At the same time the extraction of gold reaches 95%, and the same tendency toward an increase in the extraction of gold on both sides of the alloy \(AuTe_2\) is observed, as is shown by the following figures:

Gold content in the alloy, %	Gold extraction, %	Tellurium extraction, %
10	95.1	7.71
20	87.2	15.4
30	94.7	25.23
43.5 (\(AuTe_2\))	88.4	31.58
50	93.5	35.2
60	93.4	67.87

As is evident from the third column of figures, the percentage extraction of tellurium with 24-hour agitation increases with an increase in the gold content of the alloys.

The almost complete agreement (within the limits of the errors of determination) of the amounts of gold found in the solutions and in the undissolved residues with the amount of gold taken in the charges of the corresponding alloys, for individual experiments, shows that the large differences in the extraction of gold for one and the same alloy are explained not by losses of gold, but by other causes.

On examining the tests, it was not difficult to notice that during agitation by means of a stirrer the following phenomena occurred. With poor alloys, for example with 10–20% Au, part of the alloy floated up in the form of a film on the surface of the solution; with richer alloys, the formation of hollow balls of alloy was observed, some floating up, and some accumulating at the bottom of the vessel beneath the stirrer when the speed of the latter was low. When the number of revolutions of the stirrer was increased, the alloy particles rose as a result of centrifugal force and collected in the form of a ring on the walls of the flask outside the solution. At a certain speed of the stirrer, the balls and film began to move at almost the same speed as the medium surrounding them, and mixing of the liquid and the solid did not occur. The phenomena described above were not eliminated even after thorough washing of the flask and stirrer with hot chromic mixture; consequently, the accidental presence of organic contaminants (oil, grease) did not cause these phenomena.

Thus, during agitation in the flask, three facts were observed: finely divided tellurides are poorly wetted by water, exhibit a floating capacity, and, at a certain number of revolutions of the stirrer, the solid particles and the liquid begin to rotate at the same speed. All this, understandably, made the extraction of gold difficult and slowed it down, caused large differences in the results obtained, and did not permit definite conclusions to be drawn regarding the dependence between the extraction of gold and the composition of the alloys.

Similar phenomena were noted by V. Fischer^34) in his work “Investigation of Supersaturated Solutions of Salts.”

Fischer says that these phenomena are always observed when a liquid is mixed with a solid in cylindrical vessels by means of a central stirrer, and do not depend on the shape of the stirrer. In order to achieve more intensive mixing of the solution and the solid particles, for example salt crystals, Fischer used a two-bladed glass propeller for mixing, the diameter of whose blades was approximately equal to half the diameter of the vessel. Into the latter, near the wall, a glass tube was lowered. The solution and crystals, carried along by the stirrer, encounter an obstacle, as a result of

in which intensive mixing takes place. A similar construction of vessels is used for flotation experiments in the laboratory. But in the case of tellurides, which in general are poorly wetted by water, this method did not promise any noticeable improvement. It is of interest to note that analogous phenomena are encountered in work on the plant scale, when devices are used for agitation in which circulation of the slime pulp is effected. The pulp circulates in the apparatus in a horizontal or vertical direction, as, for example, in mechanical agitation in a Pachuca tank.

With both methods of agitation the solution and the solid particles move, first, in one and the same direction and, secondly, with a very small difference in their velocities of motion, owing to which the position of the ore particles relative to the medium surrounding them changes very slowly. The consequence of this is that the length of time required for a given degree of extraction of the precious metal is much greater than it would be if the ore particles frequently changed the velocity and direction of their motion, or, in other words, if the ore particles and the solution did not coincide in their motion either in velocity or in direction[^35]. Incidentally, these considerations formed the basis for the design of more improved agitators for the slime process, for example, Parral’s pneumatic agitator, or the still more recent Dorr agitator for pneumo-mechanical agitation. In the latter apparatus the principles of proper agitation are realized: frequent change of the direction and velocities of motion of the solution and of the ore particles. As is known, the Dorr agitator has recently almost displaced all other analogous apparatus, since it permits a higher extraction to be obtained.

In view of the foregoing, in order to obtain a clearer dependence between the extraction of gold from tellurides and their composition, another method of agitation was used. The charge of alloy and the solution were placed in an ordinary reagent bottle, which was tightly stoppered with a rubber stopper. Four such bottles were fastened in a horizontal position on a platform, which, by means of a motor and the corresponding transmission, received a rapid rocking motion in the horizontal direction. In this way a very vigorous shaking of the substance with the solution was attained.

In view of the use of material for the experiments, a new series of alloys was prepared, the composition of which is given below.

Alloy No.	% Au
10	9.63
D	19.50
30	29.50
Au Te₂	43.08
50	49.25
B	60.25

All the alloys were ground to —150 mesh, and charges containing 50 mg of gold each were taken for the experiments.

Since numerous determinations in the preceding experiments showed that the total amount of gold found in the solutions and the corresponding precipitates agrees well with the gold content in the alloy charge, the extraction of gold was determined from the solutions carefully separated from the precipitates. To make up for the loss of oxygen, part of which was consumed in dissolving the gold, agitation was stopped every 2 hours and the bottles were opened for 1–2 minutes.

The results and details of the experiments are given in Table 5.

Table 5.

Amount of solution: 100 cubic cm. Temperature: 15°. Strength of solution: 0.12% KCN. Duration of agitation: 6 hours.

Alloy No.	Charge, mg	Au content, mg	Passed into solution Au, mg	% extraction of gold	Average % extraction of gold
10	519.2	50	36.8	73.6	70.7
10	»	»	33.9	67.2	»
D	256.4	50	33.5	67.8	64.6
D	»	»	30.7	61.4
30	169.5	50	30.02	60.04	57.3
30	»	»	27.3	54.6
Au Te₂	110.8	50	29.0	58.0	60.3
»	»	»	31.3	62.6
50	101.5	50	37.2	74.4	74.4
B	83	50	41.1	82.2	82.2

Duration of agitation: 24 hours; the remaining conditions are the same.

Alloy No.	Charge, mg	Au content, mg	Passed into solution Au, mg	% extraction of gold	Average % extraction of gold
10	519.2	50	48.4	96.8	97.4
»	»	»	49.0	98
D	256.4	50	46.25	92.5	92.6
D	»	»	46.35	92.7
30	169.5	50	44.70	89.4	90.5
30	»	»	45.8	91.6
Au Te₂	110.8	50	45.1	90.2	89.9
Au Te₂	»	»	44.8	89.6
50	101.5	50	48.6	97.2	97.6
50	»	»	48.95	97.9
B	83	50	49.0	98.0	98.1
B	»	»	49.1	98.2

Turning to consideration of the figures for the extraction of gold from tellurides presented in Table 5, we find in them confirmation of the conclusions drawn on the basis of the experiments of the first series. But in view of the more perfect agitation and the elimination of all the shortcomings inherent in agitation in flasks by means of stirrers, in these experiments we obtained a higher extraction, both with six-hour and with twenty-four-hour agitation. Moreover, the results of parallel experiments agree with one another much better than in the first series.

A solution of potassium cyanide extracts gold from tellurides of the latter, and the percentage extraction from the various alloys increases on both sides of the compound or alloy $\mathrm{AuTe_2}$, at which this extraction is at a minimum. This dependence is seen quite clearly from the following diagram in Fig. 2, where the fusibility curve and the curve of gold extraction for the $\mathrm{Au}:\mathrm{Te}$ system are compared.

The causes producing an increase in the extraction of gold are different for alloys lying, in composition, to the left and to the right of $\mathrm{AuTe_2}$. For alloys lying between 100% Te and $\mathrm{AuTe_2}$, the increase in extraction is explained by the following reasons. In these alloys the components are tellurium and $\mathrm{AuTe_2}$; in an alloy corresponding to the formula $\mathrm{AuTe_2}$ we have the least favorable conditions for dissolving gold, for the alloy is too concentrated with respect to gold and consists of a single gold-bearing component, $\mathrm{AuTe_2}$, in which all the gold is held by tellurium in chemical or mechanical combination. With an increase in the tellurium content, the concentration of the alloy with respect to gold decreases. At the same time tellurium is a substance that mechanically dilutes the alloy and promotes greater dispersion of the gold, whereby the surface of the latter, exposed to the action of potassium cyanide and oxygen, is, as it were, increased. Here an increase in the amount of tellurium plays the same role as fine grinding of an ore, by which, as is known, a greater degree of isolation or exposure of the gold particles is achieved, and consequently an increase in the extraction of the latter, both in amalgamation and in cyanidation. Another reason for the increased extraction of gold with increasing tellurium content in the alloy is the circumstance that the degree of brittleness of the alloys is proportional to their tellurium content, which is observed when the alloys are ground. Thus, whereas the $\mathrm{AuTe_2}$ alloy is still fairly hard, alloys rich in tellurium approach elemental tellurium in their brittleness. Owing to this, at one and the same degree of grinding, determined by the size of the sieve opening, the quantity of particles finer than size −150 in the ground alloy is the greater, the more tellurium it contains. In connection with this circumstance, the extraction of gold from these alloys naturally increases.

In alloys lying between $\mathrm{AuTe_2}$ and 100% Au, the gold-bearing components are both $\mathrm{AuTe_2}$ and free fine gold, the presence of which accounts for the increase in the percentage extraction of gold.

The experiments carried out showed that direct cyanidation of finely ground synthetic gold tellurides gives a very complete extraction of gold, close to that obtained in the cyanidation of gold-bearing sulfides, which, as is known, contain gold in the free state. And with sulfides the practical maximum of extraction rarely exceeds 95–97%. The extraction of gold from natural sulfides, or so-called concentrates, increases as the grain size decreases, and modern methods of cyaniding sulfides amount to applying to them a complete slime process, i.e., grinding to −200 mesh and prolonged agitation with a solution of potassium cyanide.

It is quite possible that, with still finer grinding of tellurides, it would have been possible to attain the practical maximum of extraction, i.e., 95–97%, even with a shorter duration of treatment; but the extreme shortage of starting materials did not permit a new series of experiments to be undertaken to clarify this question as well.

Of greatest interest, understandably, is the extraction of gold from alloys lying between 100% Te and $\mathrm{AuTe_2}$, since this region of alloys includes all the natural gold tellurides known in nature. It would also be very important to investigate the question with respect to natural gold tellurides; and it is possible that the latter would prove more refractory or would show other relationships, but, unfortunately, it was impossible to obtain natural tellurides in the quantity needed for the experiments.

Conclusions.

1. Weak solutions of potassium cyanide (0.12–0.14%) extract gold from tellurides of the latter.

2. The percentage extraction of gold increases with the duration of treatment of the tellurides and depends on their composition.

3. In the Au:Te system the alloy \( \mathrm{AuTe_2} \) gives the lowest extraction; on both sides of it the extraction increases. In the concentration range 0–43.59% Au (\(\mathrm{AuTe_2}\)), the increase in extraction is explained by the greater dilution of the alloys with respect to gold and by their greater brittleness. Both circumstances condition a greater degree of exposure of the gold particles.

In the concentration range 43.59–100% Au, the increase in the extraction of gold is explained by the presence in these alloys of fine metallic gold as a free component.

1. In pure form, tellurium is only slightly soluble in potassium cyanide. From alloys of tellurium with gold, potassium cyanide extracts both elements. The percentage extraction of tellurium increases with an increase in the gold content of the alloy. Zinc precipitates gold and tellurium from the solutions.

2. The experiments show that gold tellurides of a composition corresponding to natural tellurides cannot, with respect to cyanidation, be regarded as substances more refractory than gold-bearing sulfides. The extraction of gold from these tellurides is no lower than from gold-bearing sulfides under identical treatment conditions.

III. Bromination and bromocyanidation of gold tellurides.

Turning to the methods of treatment of telluride ores in practice, a survey of which was given at the beginning of the article, we see that, of these methods, industrial application has been found by the cyanidation of preliminarily roasted ores, bromocyanidation, and in some cases bromination of raw ores. The latter method was successfully used as early as 1896 by Engelhardt[^39], who extracted gold from the telluride and sulphotelluride ores of Colorado with an aqueous solution of bromine.

The use of cyanogen bromide lies at the basis of Diehl’s process. This process was patented in 1894 by Sulman and Teed[^37] and bears the name of its authors. But Prof. Christy denies them the right of priority, since long before them cyanogen bromide had been used for extracting gold from telluride ores. Instead of expensive cyanogen bromide, Christy[^38] prefers to add bromine to a solution of potassium cyanide, which makes the operation more convenient. According to Mulholland’s experiments[^40], bromination in connection with cyanidation renders the majority of refractory ores amenable, and the extraction of gold thereby reaches 97%. The consumption of KCN decreases, and the extraction of gold is greatly accelerated. The introduction of bromine into the solutions is carried out in various ways. The required quantity of bromine is added to the solutions with agitation, or a mixture of air with bromine vapors is blown through the pulp.

As for the theory of bromocyanidation, in Christy’s opinion the role of cyanogen bromide, as well as of the other halide compounds of cyanogen, consists in the fact that these substances liberate from potassium cyanide a free CN group, which promotes the energetic dissolution of gold in potassium cyanide. The category of such accelerators in the process of dissolving gold also includes free halides and some oxidizing agents. According to Christy, and also Sulman and Teed, the following reactions take place in bromocyanidation:

\[ 1,\ \mathrm{KCN} + \mathrm{BrCN} = \mathrm{KBr} + (\mathrm{CN})_2 \]

\[ 2,\ 2\mathrm{Au} + (\mathrm{CN})_2 + 2\mathrm{KCN} = 2\mathrm{KAu}(\mathrm{CN})_2, \]

and since both reactions proceed simultaneously, the overall reaction has the following form:

\[ 3,\ 2\mathrm{Au} + 3\mathrm{KCN} + \mathrm{BrCN} = 2\mathrm{KAu}(\mathrm{CN})_2 + \mathrm{KBr}. \]

Analogous reactions take place when bromine water or chlorine water is used:

\[ \begin{aligned} &1,\quad 2\mathrm{KCN} + \mathrm{Br} = \mathrm{KBr} + \mathrm{CN} + \mathrm{KCN} \\ &2,\quad \mathrm{Au} + \mathrm{CN} + \mathrm{KCN} = \mathrm{KAu}(\mathrm{CN})_2 . \end{aligned} \]

As is evident from these reactions, cyanogen bromide or bromine plays the same role in the process of dissolving gold in potassium cyanide as oxygen does, i.e., the role of oxidizers. From comparison of Elsner’s equation:

\[ 2\mathrm{Au} + 4\mathrm{KCN} + \mathrm{H}_2\mathrm{O} + \mathrm{O} = 2\mathrm{KAu}(\mathrm{CN})_2 + 2\mathrm{KOH} \]

with the preceding equations, it is not difficult to calculate that:

1 g of oxygen causes the dissolution of 24.5 g Au
1 » bromine » » 2.46 » »
1 » cyanogen bromide causes the dissolution of 3.72 g Au.

Thus oxygen is much more active than the other oxidizers, but the greater convenience of using bromine, cyanogen bromide, and others lies in their greater solubility in cyanide solutions and, especially, in pulp, than the oxygen of the air.

For typical telluride ores from the Kalgoorlie and Coolgardie deposits, the use of cyanogen bromide as an accelerator in the cyanidation of raw ores was accompanied by considerable technical and economic success. For some ores, recoveries of 95% were achieved. Nevertheless, owing to the delicacy of the process, which for its successful execution required much attention and experience on the part of the metallurgist, and also to its somewhat greater cost, the bromocyanidation process gave way to the dry process, i.e., to the cyanidation of previously roasted ores, as we see from the practice of Australian plants.

On the other hand, some experimental data appear not to confirm the accelerating action of cyanogen bromide on the extraction of gold from telluride ores^41).

It therefore seemed of interest to determine the behavior of tellurium, gold, and synthetic gold tellurides toward solutions of bromine, cyanogen bromide, and a mixture of the latter with potassium cyanide. For this purpose a series of corresponding experiments was carried out.

Bromine water was prepared by shaking bromine with water, and the bromine content was checked by titrating the solution by iodometric methods. Cyanogen bromide was prepared by the following method described by Sholl^42).

A solution of 65 g KCN in 120 cc of water is cooled to 0° and, with frequent shaking, added dropwise to 150 cc of bromine covered with a small amount of water and cooled with ice. Each drop of potassium cyanide causes effervescence and heating of the solution, the temperature of which is maintained between 0° and 30°. After the free bromine has disappeared, several more drops of potassium cyanide are added, until the thick dough-like mass in the flask becomes light yellow. This point is extremely important, since further addition of potassium cyanide causes browning of the mass and decreases the yield of cyanogen bromide.

It is preferable to take pure potassium cyanide (99–100%), in which case the end of the reaction coincides almost with the complete utilization of KCN. The reaction product is distilled from a retort heated on a water bath to 65°–70°. Cyanogen bromide sublimes at 61° in the form of white needle-like crystals. Cyanogen bromide is readily soluble in water, and its dilute solutions are fairly stable. Thus, for example, in a solution that had stood for three months, the initial cyanogen bromide content of 1.88% had fallen to 1.62%.

The content of cyanogen bromide in solution was determined as follows^43). To 10–20 cc of solution were added 5–7 cc of concentrated hydrochloric acid, and then 2 g of potassium iodide; the iodine liberated according to the reaction:

\[ \mathrm{BrCN} + \mathrm{HCl} + 2\mathrm{KJ} = \mathrm{HCN} + \mathrm{KBr} + \mathrm{KCl} + \mathrm{J}_2 \]

iodine was titrated with thiosulfate; the amount of cyanogen bromide is calculated from the iodine on the basis of the reaction indicated above.

\[ \begin{array}{ccc} 2J & — & \mathrm{BrCN}\\ 2 \times 126{,}92 & — & 105{,}93 \end{array} \]

Definite weighed portions of substances ground to 150 mesh were shaken with aqueous solutions of the corresponding reagents.

The solubility of gold and tellurium was determined from the content of these elements both in the solutions and in the undissolved precipitates.

The results of the experiments are given in Table 6:

Table 6.

No. of experiment	Substance	Weight, mg	Solvent	Amount of solution, cc	Concentration of solution, %	Duration of agitation, hours	Gold passed into solution, mg	Au in residue, mg	% extraction of Au	Weight of residue at 105°, mg	Te in residue, mg	Te passed into solution, mg	% extraction of Te
1	Te	150	Br	100	0,975	1	—	—	—	—	0	150	100
2	»	»	»	»	0,48	1	—	—	—	—	Solution decolorized; part of Te did not dissolve.
3	»	200	Br CN	100	1,8	4,5	—	—	—	208	200	0	0
4	»	»	»	»	»	»	—	—	—	211	200	0	0
5	»	»	»	»	0,55	7	—	—	—	207,8	200	0	0
6	»	»	»	»	»	»	—	—	—	206,6	200	0	0
7	Au	100	Br	100	0,975	1	100	0	100
8	»	»	»	»	0,48	3	100	0	100
9	»	50	»	»	0,90	1,5	46,7	—	93,4
10	»	»	Br CN	100	1,88	1,5	0,25	—	0,50
11	»	»	»	»	»	»	0,2	—	0,40
12	»	»	»	100	0,55	3	0,02	—	0,04
13	»	»	»	»	»	»	0,02	—	0,04
14	»	»	»	»	»	6	0,02	50,2	0,04
15	»	»	»	»	»	»	0,06	50,1	0,12
16	»	»	»	»	»	15	0,10	—	0,2
17	»	»	»	»	»	15	0,08	—	0,16
18	»	»	»	»	»	24	0,1	49,9	0,2
19	»	»	»	»	»	24	0,15	49,9	0,3
20	»	»	»	»	0,52	24	0,05	—	0,10
21	»	50	KCN	100	0,25	1,5	48,4	—	96,8

On the basis of the experiments and the observations made, we can draw the following conclusions concerning the behavior of gold and tellurium toward solutions of bromine and bromocyanide.

Tellurium dissolves very readily in dilute bromine water, the latter being decolorized owing to the formation of tellurium bromide. The composition of the latter may be determined from the following data.

In experiment No. 1, the solution before agitation contained 975 mg of bromine.
» » after agitation, 322.4 mg of bromine remained.

Consequently, 150 mg of Te combined with 652.6 mg of Br, which corresponds to a bromide of composition \(\mathrm{TeBr_7}\), and not to the known bromide \(\mathrm{TeBr_4}\), since theoretically 150 mg of tellurium require 376 mg of Br for the formation of \(\mathrm{TeBr_4}\). In experiment No. 2 the solution became completely decolorized, contained no free bromine, and part of the Te did not dissolve, as was to be expected in view of the deficiency of bromine in the solution, of which there was only 480 mg instead of the 652.6 mg required on the basis of the results of the first experiment.

Tellurium is completely insoluble in an aqueous solution of bromocyanide. Judging from the weight of the precipitate dried at \(105^\circ\), which exceeded the original weight of the tellurium by 3–5%, the latter, under the action of \(\mathrm{BrCN}\), is either oxidized or forms some compound with \(\mathrm{BrCN}\). In any case, the product of the interaction of Te and \(\mathrm{BrCN}\) in bromocyanide does not dissolve.

Metallic gold, like tellurium, dissolves very readily in bromine water and is practically insoluble in an aqueous solution of bromocyanide. The results obtained for the relative solubility of gold in bromine water and potassium cyanide do not agree with the data of Rose \(^{44}\), who determined the relative solubility of gold in various reagents. Rose used in his experiments a gold parting-leaf weighing 500 mg and placed it in quietly standing solutions of various reagents, taken in excess. I, on the other hand, agitated gold in the form of a fine powder (—150 mesh) with the solutions; this, of course, explains the difference in the figures obtained by us for the solubility of gold in the same reagents. For comparison these figures are given below.

Author	Solvent	Concentration %	Duration of experiment, hours	Temp., °C	Au dissolved, mg
Rose	Br	1	\(1\frac{1}{4}\)	60	25.85
»	»	0.2	5	15	5.8
Mostovich	»	0.975	1	15	100
»	»	0.48	3	15	100
»	»	0.90	\(1\frac{1}{2}\)	15	46.7
Rose	KCN	1	\(1\frac{1}{4}\)	15	0.75
Mostovich	»	0.25	\(1\frac{1}{2}\)	15	48.4
»	»	0.12	\(1\frac{1}{2}\)	17	40.6

From the behavior of gold and tellurium toward bromine water it was to be expected that gold tellurides would readily dissolve in the latter, which was fully confirmed by specially conducted experiments. All gold tellurides in the finely ground state dissolved very rapidly and without residue in bromine water. This fact may explain the high extraction of gold obtained by Engelhardt in treating telluride ores with bromine water. Since an aqueous solution of bromocyanide dissolves neither gold nor tellurium separately, it seemed highly probable that the same solution would not...

extract gold from the tellurides of the latter. To verify this assumption, two slags with 43.5 and 41.5% Au were agitated with a solution of cyanogen bromide and, in parallel, with a solution of potassium cyanide.

The results of these experiments are given in Table 7.

Table 7.

Experiment No.	Slag composition, % Au	Charge, mg	Au content, mg	Solvent	Amount of solution, cubic cm	Concentration of solution, %	Duration of agitation, hours	Passed into solution, mg	% extraction of gold
1	43.5	110.8	50	BrCN	100	1.80	4.5	26.6	53.2
2	43.5	—	—	—	—	—	—	32.8	65.6
3	41.5	120.5	50	—	—	1.88	3	23.2	46.4
4	—	—	—	—	—	—	—	23.5	47.0
5	—	—	—	KCN	100	0.16	3	15.7	31.4
6	—	—	—	—	—	—	—	16.0	32.0

Contrary to the assumption made, it turned out that cyanogen bromide extracts gold from gold tellurides, but does not extract tellurium, since the latter could not be detected in the solution. Such unexpected results led us to set up a new series of experiments on the agitation of gold tellurides with cyanogen bromide. In these experiments the amounts of gold and tellurium were determined both in the solutions and in the undissolved residues, dried at 105° to constant weight. After KCN was added to decompose BrCN, the solutions were evaporated, and the gold was precipitated from them with zinc dust. From the residues, separated and washed on the filter, tellurium was extracted with hot dilute nitric acid.

The solution of Te(NO₃)₂ was evaporated three times with conc. HCl and NaCl to dryness; the dry residue was dissolved in hot hydrochloric acid (1:10), and tellurium was precipitated from the hot solution with sulfurous gas for 1½–2 hours. The precipitated tellurium was filtered after 12 hours with the aid of a Gooch filter, washed with water and alcohol, and dried at 105°. The results of the experiments are presented in Table 8.

Table 8.

Experiment No.	Slag composition, % Au	Charge, mg	Au content, mg	Te content, mg	Solvent	Amount of solution, cubic cm	Strength of solution, %	Duration of agitation, hours	Passed into solution, Au, mg	Au in residue, mg	Total residue at 105° Au + Te, mg	Te in residue, mg	% extraction Au
1	41.7	120.5	50.2	70.3	KCN	100	0.38	6	34.3	—	—	—	68.6
2	—	—	—	—	—	—	—	—	33.9	—	—	—	67.8
3	—	—	—	—	BrCN	—	1.88	—	28.0	—	—	—	56
4	—	—	—	—	—	—	—	—	27.85	—	—	—	55.7
5	—	—	—	—	—	—	0.55	3	15.85	—	—	—	31.7
6	—	—	—	—	—	—	—	3	14.85	—	—	—	29.7
7	—	—	—	—	—	—	—	6	24.9	26.25	103.8	by difference: 77.55	49.8
8	—	—	—	—	—	—	—	6	25.0	24.70	101.0	76.3	50.0
9	—	—	—	—	—	—	—	24	33.3	19.3	100.8	81.5	66.6
10	—	—	—	—	—	—	—	24	34.0	18.55	98.0	79.45	68.0
11	—	—	—	—	—	—	—	7	25.0	27.0	100.8	73.8	50
12	—	—	—	—	—	—	—	7	22.6	27.25	101.4	74.15	45.2
13	38.8	129	50	79	BrCN	1.00	0.52	24	31.1	—	—	by analysis: 74.0	62.2
14	—	—	—	—	—	—	—	—	32.45	—	—	75.8	64.9
15	—	—	—	—	—	—	—	—	31.8	—	—	74.0	63.6
16	—	—	—	—	—	—	—	—	33.25	—	—	78	66.5

In experiments 7–12, “tellurium in the residue” was calculated from the difference: weight of residue at 105° minus gold in the residue; in experiments 13–16, the tellurium in the residue was determined by analysis of the residue.

As is evident from the experimental results, an aqueous solution of bromocyanide actually extracts only the gold from gold tellurides; the tellurium does not dissolve, and the difference of 0.95–5 mg of tellurium, compared with the amount of the latter present in the original alloy charges, represents only the natural losses associated with the numerous operations for determining tellurium in the residues from agitation. It is also necessary to note the following fact, following from experiments 1–4 of Table 8. In these experiments the solutions contained the same amount of CN, both in the form of KCN and of BrCN. It turned out that the extraction of gold from tellurides by means of potassium cyanide is only slightly higher than the extraction of gold from the same tellurides by means of bromocyanide. From this it may be concluded that, when both reagents, KCN and BrCN, act jointly upon tellurides, the percentage extraction of gold from them will differ little from the extraction obtained when only one of these reagents is used. On the other hand, this conclusion contradicts many facts known from the practice of cyaniding tellurous ores. Thus, in the Diehl process, bromocyanide is used in cyaniding tellurous ores as a reagent that especially strongly accelerates and increases the extraction of gold. Knutsen²²) gives the following examples. Metallic gold, he says, does not dissolve either in bromocyanide or in potassium cyanide free from air. A plate of metallic gold immersed in a solution of potassium cyanide dissolves extremely slowly, as is illustrated, among other things, by Rose’s figures cited above for the solubility of Au in KCN.

If bromocyanide is added to the solution and the mixture is shaken, the gold dissolves almost instantaneously. Further, in agitating one and the same ore with a pure solution of KCN and with a mixture of the latter with bromocyanide under identical conditions, Knutsen obtained the following results.

The ratio of liquid to solid was 1:1; duration of agitation, 24 h.

Composition of solution	% extraction of gold
0.15% KCN	44.64
0.20 » »	53.24
0.15 » KCN + 0.0375% BrCN	85.71
0.20 » » + 0.05 » »	95.87

Consequently, the addition of bromocyanide to a solution of potassium cyanide in an amount not exceeding 25% of the latter increased the extraction of gold almost to 100%.

In view of the obvious contradiction between the data of practice and the conclusion I made on the basis of the experiments on bromocyanidation of gold tellurides, it appeared necessary to determine experimentally the influence of bromocyanide on the extraction of gold from synthetic tellurides by means of potassium cyanide. The corresponding experiments were carried out, the results of which are summarized in Table 9.

In these experiments the amounts of KCN and BrCN were taken in considerable excess over theory, on the assumption that dissolution of gold in a mixture of these reagents proceeds according to the equation:

\[ 2\mathrm{Au} + 3\mathrm{KCN} + \mathrm{BrCN} = 2\mathrm{KAu}(\mathrm{CN})_2 + \mathrm{KBr} \]

Theoretically, 50 mg of Au require 24 mg of KCN + 13 mg of BrCN; for the experiments, in the form of the corresponding solutions, 120 mg of KCN and 50 mg of BrCN were taken.

In parallel, under the same conditions, experiments were set up on extracting gold from telluride with potassium cyanide alone.

Experiment No.	Composition of alloy, % Au	Weight of alloy, mg	Au content in alloy, mg	Solvent	Amount of solution, cc	Duration of agitation, hours	Passed Au into solution, mg	Gold in residue, mg	Sum of gold in solution and residue, mg	% extraction of gold
1	38.8	129	50	KCN 0.12%	100	3	19.0	—	—	38.0
2	—	—	—	—	—	—	19.35	—	—	38.70
3	—	—	—	—	—	—	20.45	30.3	50.75	40.9
4	—	—	—	KCN + BrCN	111	3	20.45	29.5	49.95	40.9
5	—	—	—	—	—	—	20.9	30.2	51.1	41.8
6	—	—	—	—	—	—	22.6	—	—	45.2
7	—	—	—	—	—	—	22.3	—	—	44.6
8	—	—	—	KCN 0.12%	100	4	24.5	25.6	50.1	49.0
9	—	—	—	—	—	—	21.2	30.0	51.2	42.4
10	—	—	—	KCN + BrCN	111	4	25.5	24.9	50.4	51
11	—	—	—	—	—	—	24.0	25.7	49.7	48
12	100% Au	50	50	KCN 0.12%	100	0.5	40.6	—	—	81.2
13	—	—	—	KCN + BrCN	111	0.5	49.7	—	—	99.4

Comparing the quantities of gold extracted in the experiments, we see that, although bromine cyanide considerably increases the dissolution of pure gold in potassium cyanide, it has only a very slight effect on the extraction of gold from synthetic tellurides, as was also to be expected from the behavior of gold tellurides toward solutions of potassium cyanide and bromine cyanide separately.

In the extraction experiments only one alloy with 38.8% Au was subjected to treatment, but there is no doubt that all the other tellurides, up to and including AuTe₂, would give analogous results. Thus, the opinion of Knutsen and other authors concerning the role of bromine cyanide as a reagent that especially vigorously increases the extraction of gold from tellurides is not confirmed with respect to synthetic gold tellurides.

Above we have seen that bromine cyanide, completely inactive with respect to gold and tellurium taken separately, extracts only gold from alloys of both elements. In chemistry I find no analogy to this strange phenomenon. It is known, for example, that if a given substance is a solvent for only one of the components of an alloy, then under certain conditions either the entire alloy dissolves in this solvent, or part of the insoluble component does. Thus, for example, nitric acid dissolves silver and does not dissolve platinum. But from alloys of both elements nitric acid, together with silver, dissolves part of the platinum, and the greater part the larger the Ag : Pt ratio; thus, with an eightfold amount of Ag relative to Pt, the entire alloy dissolves completely in nitric acid^45). In view of this, in…

in assaying alloys of Au, Ag, Pt, nitric acid cannot be used for the separation of Ag and Pt. This is explained by the fact that an acid solution of silver nitrate dissolves platinum.

Further, from the practice of assaying gold-bearing copper alloys, e.g., anode copper, blister copper, etc., it is known that a dry assay for gold usually gives higher and better mutually agreeing results than a combined assay, in which all or part of the copper is first removed from the alloy by dilute nitric acid. The difference for gold between the two methods reaches 3%. Attempts were made to explain the losses of gold in the combined assay by the solvent action on gold of copper nitrate or of the products of reduction of nitric acid formed when the alloy is dissolved in nitric acid. But Keller[^16], by numerous and carefully conducted experiments, established that gold contained in an alloy with copper dissolves in nitric acid. This solubility of gold depends not at all on impurities in the alloy, but on the manner in which the alloy is cooled. It reaches a minimum for alloys rapidly cooled by granulation. From his experiments Keller draws the following hypothetical conclusion concerning the causes of the solubility of gold in nitric acid. The solubility of gold in its alloys with copper in nitric acid may be explained by the chemical nature of the gold, which is different in slowly and rapidly cooled alloys. It is possible that, upon slow cooling of liquid alloys, part of the gold forms with the components of the alloy certain chemical compounds soluble in nitric acid. Rapid cooling, e.g., granulation of the alloy, contributes to the gold remaining in an uncombined, free form, in which it is insoluble in the dilute nitric acid used for dissolving the alloy.

In the examples cited above, the solvent, nitric acid, dissolves one of the components of the alloy (Ag, Cu). In the case of gold tellurides, however, the solvent, bromocyanide, dissolves neither gold nor tellurium separately. The supposition that, upon alloying gold with tellurium, compounds of both elements are formed that are soluble in nitric acid is excluded in view of the fact that in solutions obtained by agitating tellurides with bromocyanide only gold was found, while tellurium, which should likewise have passed into solution, was entirely absent.

On the basis of the fact that, under the action of bromocyanide on gold tellurides, the tellurium, without dissolving, is merely oxidized, while the gold passes into solution, the mechanism of this process could be explained as follows. The process of extracting gold from its tellurides by means of bromocyanide is electrochemical in nature. In an aqueous solution of bromocyanide, gold and tellurium form a couple, in which the tellurium acquires a positive charge and the gold a negative charge. Under the action of this couple, bromocyanide dissociates into the anion Br′ and the cation (CN)′. At the cathode, i.e., at the gold, aurous cyanide is formed, which with BrCN gives the soluble complex AuCN·BrCN(?). This is all the more probable since some gold compounds readily form analogous complexes and adsorption products, such as AuCN·Br, KCN·AuBr₃, HBr, etc. Lindbom[^49] prepared compounds in which two radicals (CN) in the complex (AuCN)₄ are replaced by halides. These compounds are obtained by the action of halides on a solution of KAu(CN)₂; the latter compound takes up halides and forms, for example, the ion AuCN′Br₂CN′, the halides entering into the composition of the neutral part of the complex.

In order to confirm the correctness of the above explanation of the mechanism of extraction of gold from its tellurides by bromocyanide, there were determined, first, the potentials of both components of the alloy, gold and tellurium, with respect to an aqueous solution of bromocyanide, and, second, the character of the dissociation of bromocyanide under the action of an electric current.

To determine the potentials, the electromotive force of the following pairs was measured:

\[ \begin{aligned} 1\quad &\mathrm{Hg}\mid \mathrm{HgCl} + N/1\,\mathrm{KCl} — \mathrm{Au}\mid \mathrm{BrCN}\ (0.52\%)\\ 2\quad &\mathrm{Hg}\mid \mathrm{HgCl} + N/1\,\mathrm{KCl} — \mathrm{Te}\mid \mathrm{BrCN}\ (0.52\%), \end{aligned} \]

in other words, Ostwald’s normal mercury electrode was combined with the electrode under investigation. The potential difference between mercury and the solution $\mathrm{HgCl} + N/1\,\mathrm{KCl}$ was taken as equal to $+0.56$.

The arrangement of the apparatus for determining the potentials is shown schematically in Fig. 3¹.

Explanations for the diagram:

NO—Ostwald normal electrode.
C—vessel with $\mathrm{BrCN}$ solution (0.52%).
M—gold or tellurium, clamped in the platinum tips of tweezers.
R—rheostat of 300,000 ohms.
G—mirror galvanometer of W. G. Pye.
T—tube with a scale divided into 100 cm.

The calibration of the scale was carried out with the aid of a normal Weston cell, whose electromotive force at 10° is 1.0189 volts. For the pair: Ostwald electrode—gold in bromine cyanide, an electromotive force of $+0.1286$ was obtained, i.e.,

\[ \pi_{\mathrm{Au}}-\pi_{\mathrm{Hg}}=0.1286, \]

whence

\[ \pi_{\mathrm{Au}}-\pi_{\mathrm{Hg}} = \pi_{\mathrm{Au}}-\mathrm{BrCN}+\pi_{\mathrm{KCl}}-\mathrm{Hg} \]

\[ 0.1286=x+(-0.56) \]

\[ x=0.6886 \]

For the pair: Ostwald electrode—tellurium in bromine cyanide, an electromotive force of $+0.2507$ was obtained, i.e.,

\[ \pi_{\mathrm{Te}}-\pi_{\mathrm{Hg}}=+0.2507 \]

\[ \pi_{\mathrm{Te}}-\pi_{\mathrm{Hg}} = \pi_{\mathrm{Te}}-\mathrm{BrCN}+\pi_{\mathrm{KCl}}-\mathrm{Hg} \]

\[ 0.2507=y+(-0.56) \]

\[ y=0.8107 \]

Consequently, for the pair $\mathrm{Te}:\mathrm{Au}$ in an aqueous solution of bromine cyanide (0.52%), the potential difference is

\[ 0.8107-0.6886=+0.1221\ \text{volt} \]

Direct measurement of the potential difference for this pair gave 0.11 volt. The potential decreased throughout, and the maximum deflection of the galvanometer was taken as correct.

These measurements show that both tellurium and gold are electropositive with respect to the bromine cyanide solution, i.e., neither of them dissolves in bromine cyanide; this was confirmed by numerous agitation experiments with finely divided elements and bromine cyanide. In order to clarify how bromine cyanide dissociates in aqueous solution under the action of current, several experiments were carried out in which the bromine cyanide solution was subjected to electrolysis between two silver electrodes in the form of thin plates. The negative electrode apparently did not change; on the positive electrode a pale-yellow precipitate formed, which upon examination proved to be silver bromide. This shows that bromine cyanide dissociates into the anion $\mathrm{Br}'$ and the cation $(\mathrm{CN})'$. From this it could be concluded that, in the case of gold electrodes, both electrodes under the action

of the current will dissolve in bromocyanide, since at the anode there will occur the formation of readily soluble bromide of gold, \(AuBr_3\), and at the cathode—cyanide gold, which, as is known, readily gives complexes and adsorption products with haloid compounds of cyanogen.

The experiments carried out fully confirmed this supposition. Two gold plates, accurately weighed, were immersed in a solution of \(BrCN\) (0.52%). When current from the lighting mains was passed through a lamp rheostat, owing to the extraordinarily high resistance of the solution, the latter became strongly heated.

After 3 minutes the plates were removed from the solution, washed, dried, and weighed. Below are given the weights of the plates before and after the experiments.

	Anode (Au)	Cathode (Au)
Initial weight	130.5 mg	172.6 mg
After the first experiment	129.6 mg	171.6 mg
After the second experiment	128.7 mg	170.7 mg
Passed into solution	1.8 mg	1.9 mg

In another series of experiments, two accumulators connected in series were taken as the source of current. Under these conditions the solution did not heat up. At the anode the liquid turned yellow and sank to the bottom of the vessel. The anodic liquid showed a strongly acid reaction, and the cathodic liquid a strongly alkaline reaction. The following amounts of gold passed into solution from the electrodes:

	At the anode	At the cathode
After 1 hour	0.1 mg Au	0.5 mg Au
After 3 hours	0.2 mg Au	0.8 mg Au
After 3 hours	0.15 mg Au	2.0 mg Au
After 4 hours	0.3 mg Au	1.85 mg Au

The electrolyte was periodically stirred.

The experiments show that much more gold dissolves at the cathode than at the anode. In order to establish more precisely the composition of the products formed during electrolysis, experiments were made on the electrolysis of a solution of bromocyanide (0.52%) between gold electrodes in a U-shaped vessel consisting of two test tubes communicating by means of a connecting tube. The current was taken from the lighting mains through a lamp rheostat. In this way it was possible to separate the anodic liquid from the cathodic liquid with sufficient accuracy and to investigate each separately.

At the anode oxygen was evolved; the solution became yellow-brown, showed a strongly acid reaction, and contained free bromine and gold bromide (according to reactions with \(CS_2\) and \(SnCl_2\)). At the cathode hydrogen was evolved; the solution had a strongly alkaline reaction and, on slight heating, evolved ammonia.

The following amounts of gold passed from the electrodes into solution:

	At the anode	At the cathode
After 40 min.	0.85 mg	1.70 mg
After 60 min.	1.35 mg	1.70 mg

The phenomena that took place during the electrolysis of bromocyanide may be explained as follows.

The bromine liberated at the anode partly reacts with water, forming hydrobromic and hypobromous acids, which is analogous to the action of chlorine on water:

\[ HOH + BrBr \rightleftarrows HBr + HOBr \]

This accounts both for the acid reaction of the anodic liquid and, probably, for the decrease in the solubility of gold at the anode, since bromine water by itself dissolves gold very energetically.

At the cathode a complex process takes place, the scheme of which is probably as follows. The cyanogen that is liberated binds gold into $\mathrm{AuCN}$, giving, with cyanogen bromide, a water-soluble complex. Part of the cyanogen passes at the cathode into hydrocyanic acid, which is reduced by the action of the hydrogen liberated at the cathode into methylamine[^48].

\[ 2\mathrm{CN} + \mathrm{H}_2 \longrightarrow 2\mathrm{HCN} \]

\[ \mathrm{HCN} + 2\mathrm{H}_2 = \mathrm{CH}_3\mathrm{NH}_2. \]

This can explain the strongly alkaline reaction of the cathode liquid and the evolution of ammonia from it upon heating.

The very slight solubility of gold at the electrodes is due both to the reactions indicated above and to the fact that, at the electrodes, owing to the absence of stirring of the electrolyte, layers of high concentration are formed which impede the passage of gold into solution.

The experiments performed make it possible to explain in the following way the action of an aqueous solution of cyanogen bromide on gold and tellurium and on their alloys.

Both elements in the pure form and in the free state do not dissolve in cyanogen bromide. Under the action of the latter on gold tellurides, an electromotive force arises between tellurium and gold, with tellurium becoming the positive electrode and gold the negative electrode. Under the action of this local current, cyanogen bromide dissociates into the anion $\mathrm{Br}'$ and the cation $(\mathrm{CN})'$. Bromine is directed toward the tellurium and oxidizes it; the product thereby formed does not dissolve in the aqueous solution of cyanogen bromide. The cation $(\mathrm{CN})'$ is directed toward the gold, converting it into aurous cyanide; the latter forms with cyanogen bromide a soluble complex, as a result of which the gold is extracted from the tellurides under the action of cyanogen bromide upon them.

All these processes account both for the extraction of gold from its tellurides and for the phenomena that were observed in experiments on agitating gold tellurides with an aqueous solution of cyanogen bromide.

The data of these experiments confirm the correctness of the explanation given above of the mechanism of extraction of gold from its tellurides.

Summary. In the present work, the extraction of gold from synthetic gold tellurides by dilute solutions of potassium cyanide, cyanogen bromide, a mixture of these reagents, and bromine water was investigated.

The investigation gave the following results.

1. In alloys with gold, tellurium in an amount up to 1.14% does not affect the solubility of gold in potassium cyanide.

2. From tellurides containing from 10 to 60% gold, potassium cyanide extracts both elements. The extraction of gold reaches 98%; it increases with the duration of the action of potassium cyanide and depends on the composition of the alloy. The alloy $\mathrm{AuTe}_2$ (43.59% Au) gives the lowest extraction; on both sides of this alloy the extraction of gold increases. The experiments show that gold tellurides of composition corresponding to natural tellurides cannot, with respect to cyanidation, be considered substances more refractory than gold-bearing sulfides.

3. Both elemental tellurium and finely ground gold tellurides dissolve very readily in bromine water.

4. An aqueous solution of cyanogen bromide is completely inactive with respect to gold and tellurium taken separately. From alloys of both elements, cyanogen bromide extracts only gold. Cyanogen bromide has very little effect on the extraction of gold from tellurides by means of potassium cyanide.

The extraction of gold from tellurides by cyanogen bromide is an electrochemical process. In an aqueous solution of cyanogen bromide, gold and tellurium form a couple, with tellurium becoming positive and gold—nega-

To the article by V. Ya. Mostovich
Cyanidation of Gold Tellurides

Fig. 1. Freezing curve of Au–Te alloys (Robos)

• Vertical axis: °C
• Horizontal axis: Weight %% Au
• Curve labels: Tellurium; Compound of Au and Te₂; Gold
• Point labels: A, B, C, D, E
• Scale markings visible: 10, 20, 30, 40, 50, 60, 70, 80, 90; 400, 500, 600, 700, 800, 900, 1000

Fig. 3.

Visible labels in the apparatus diagram: NO; KCN; HgCl; Hg; C; M; G; R; T.

Fig. 2

Vertical axis: % extraction
Horizontal axis: Weight % Au

Curve labels: freezing curve; composition AuTe₂; tellurium; gold.

electrode. Under the action of this pair, bromocyanogen dissociates into the anion Br′ and the cation (CN)′. Bromine is directed toward the tellurium and oxidizes it; the cation (CN)′ converts the gold into auric cyanide, which forms a soluble complex with bromocyanogen. This process explains the extraction of gold from its tellurides by means of bromocyanogen.

Literature

1) Economic Geology, VI, 1911, 275
2) Mineral Industry, VI, 1898, 295
3) Trans. Am. Inst. Min. Eng. XVIII, 1890, 447
4) Economic Geology VI, 1911, 22
5) do „ „ 302
6) Metall u. Erz X, 1913, 690
7) Trans. Inst. Min. Met. VIII, 1899, 484
8) Trans. Amer Inst. Min. Eng. XXVI, 1897, 485
9) T. K. Rose. Metallurgy of Gold, (1906), 27
10a) Economic Geology, VI, 1911, 35
11) Trans. Inst. Min. Met. VIII, 1908, 175
12) Trans. Amer. Inst. Min. Eng VIII, 1890, 447
13) Mines and Minerals XXIX, 1909, 427
14) Bull. Amer. Inst. Min. Eng., September, 1917, 1257
15) Bull. U. S. Geolog. Survey № 491. F. Clarke. Data of Geochemistry, 616
16) Trans. Amer. Inst Min. Eng XVIII, 1890, 439
17) Trans. Inst. Min. Met. VIII, 1899, 76
18) Mineral Industry VI, 1898, 368
    Mc. Farren. Text--Book of cyanide practice (1912), 25, 204, 206
19) Trans. Inst. Min. Met. VIII, 1899, 484
    do XIII, 1906, 22
20) Trans. Am. Inst. Min. Eng. XLIX, 1915, 166
21) Engin. Magazin XXXVI, 1908--9, 600
22) Trans. Inst. Min. Met. XII, 1902, 2, 23
23) do „ XVII, 1908, 285
24) do „ VIII, 1899, 101 (Discussion)
25) Min. Sci. Pr. XCVII, 1907, 562
    H. Bain. More recent cyanide practice (1910), 226
26) Mineral Industry VI, 1898, Ph. Argall p. 368
27) Trans. Am. Inst. Min. Eng. XXVI, 1896, 485
28) do XLIX, 1915, 166
29) do XXVI, 1897, 735
30) Journ. Am. Chem Soc. XXIV, 1902, 355, 918
31) Economic Geology, VI, 1911, 34
32) H. O. Hofman. Metallurgy of Copper (1914), 16
33) T. K. Rose. Metallurgy of Gold (1906), 11, 18
34) B. M. Fisher. Investigation of supersaturated solutions of salts. (Dissertation) Riga, 1913, 60, 616.
35) Trans. Am. Inst. Min. Eng. XLI, 1910, 368
36) Metall u. Erz XIII, 1916, 53
37) Trans. Inst. Min. Met. III, 1894, 203
38) Trans. Am. Inst. Min. Eng. XXVI, 1896, 735
39) Eng. Min. Journ. LXII, 1896, 295
40) Berg. u. Hüttenmän. Ztg. L, 1896, 66
41) Trans. Am. Inst. Min. Eng XXVII, 1897, 846
42) Ber. d. Deutsch Chem. Gesell. 29, 2, 1896, 1822
43) J. E. Clennell. The Chemistry of Cyanide solutions (1910), 100
44) T. K. Rose. Metallurgy of Gold (1906), 11, 12
45) C. Fulton. Manual of fire assaying (1907), 149
    C. Schiffner. Einführung in d. Probierkunde (1912). III
46) Trans. Am. Inst. Min. Eng. XLIII 1912, 582
47) do XXX, 1900, 864
48) J. Clennell. Cyanide Handbook (1910), 54
    A. Bernthsen. Lehrb. d. organisch. Chemie (1902), 269.

1. Reference number visible in the source. ↩