When Charles Butters began to take up the work of flotation in our Oakland laboratory, one of the first points brought to our attention was the treatment of the concentrate produced by flotation; J. E. Clennell was accordingly instructed to undertake the researches detailed in the present paper.
The whole value of the process hinges on two points, namely, the grade of tailing produced and the net realisation of the value contained in the concentrate, these two considerations being of equal importance. This last point is complicated by questions of geographical situation, for if the concentrate cannot be treated locally the cost of realisation may be so heavy that flotation would be entirely precluded.
The results obtained in our laboratory by the combination of flotation and cyanide have been so remarkable that a serious study of the disposal of concentrates has been forced upon us.
The difficulties attending the treatment of concentrate by cyanide are well known. The process of concentration collects in a small bulk not only the valuable constituents of the ore but also those substances that act as cyanicides, or which are readily converted by oxidation or otherwise into cyanicides, so that their influence, per ton of material treated, is greater than would be the case with the unconcentrated ore. Heavy minerals such as the sulphides of iron, copper, lead, arsenic, antimony, zinc, and double sulphides such as mispickel, proustite, pyrargyrite, and bornite, naturally tend to accumulate in the concentrate. If some interval elapses between the formation of this concentrate and its treatment, oxidation may take place, with formation of sulphates, arsenates, and antimonates. which are still more detrimental to cyanide treatment than the original minerals. These difficulties have been wholly or partly overcome by the adoption of modifications in the treatment, such as preliminary water, acid, or alkali washing, roasting, fine grinding, the use of special solvents, such as bromo-cyanide, and prolonged contact of the material with cyanide, extending in some cases to over a month.
In the case of concentrate produced by flotation, the minerals composing the product are substantially the same as those obtained by gravity concentration, consisting of the sulphides and double sulphides of the heavy metals, and it is to be expected that the same difficulties will be encountered in their treatment. But as the concentrate also contains a considerable part of the oil, tar, or other flotation agent, the presence of this foreign matter must be taken into account. In some cases, this circumstance introduces an additional difficulty. A part of this organic matter is soluble in the cyanide or alkali used in the process, and the solution so formed may be capable of absorbing oxygen. The effect produced by carbonaceous matter in precipitating gold and silver previously dissolved by cyanide is well known and has been a source of much trouble in many localities. Some of the constituents of this matter are not easily eliminated and appear to resist oxidation even at a high temperature; roasting under ordinary conditions does not completely remove the carbon; it is probable that a portion derived from tar remains in the graphitic form, capable of acting as a precipitant for gold or silver.
The experiments detailed below were made on concentrates produced from typical gold and silver ores by a modified type of the Minerals Separation flotation machine. Most of the tests were made in neutral or alkaline media. The frothing agents employed were those in general use, consisting of tar, creosote, carbolic acid, pine-oil, and fuel-oil. It is not proposed to discuss these in detail in the present paper; it will be sufficient to state that the concentrate was collected and drained on a vacuum-filter and in some cases dried at a moderate temperature before treatment.
As an example of an ore in which the value consists essentially of gold we may take the product of the San Sebastian mine, in Salvador, operated by the Butters Salvador Mines, Ltd. For preliminary work a composite sample was made from 21 lots taken from different parts of the mine, and concentrate produced by treating the finely crushed ore in a 10-lb. flotation machine. The sample taken for this test assayed originally 1.54 oz. gold and 0.28 oz. silver. The concentrate obtained by flotation assayed 4.92 oz. gold and 1.14 oz. silver. As this constituted 25% of the weight of ore taken, the gold recovered in the concentrate amounted to 79.9% of the total. An analysis of the concentrate showed:
together with small quantities of molybdenum, tellurium, and other elements. The tailing carried 0.04 oz. gold per ton.
The first tests were made with the object of determining whether this material could be treated advantageously raw by agitation with cyanide. In addition to direct cyanide treatment various modifications were tried, as shown in Table I, including addition of lead acetate, preliminary alkali treatment, desulphurizing with alkali and aluminum, and bromo-cyanide. The last procedure showed a marked improvement over every other method of raw treatment, but still failed to yield a satisfactory extraction. The extraction was increased by increasing cyanide strength, but with strong solution the consumption of cyanide became prohibitive, and alkaline sulphides were formed. This effect can be prevented and cyanide consumption much reduced by addition of lead acetate, some improvement in extraction being obtained. Preliminary alkali treatment with or without aluminum showed no benefit whatever. The fact that bromo-cyanide has a marked effect on the extraction suggests that a portion of the gold may be present as a telluride. This conclusion is supported by experiments made by direct treatment of the original ore, without concentration; these tests showed that a certain proportion of the gold is inaccessible to cyanide even after very fine grinding and prolonged contact. (See Table XI.)
As these results did not appear encouraging for any system of raw treatment, attention was next turned to roasting. It was soon found that roasting within certain limits of temperature converted a considerable part of the copper into sulphate, which could be leached with water, together with some sulphate of iron, leaving the residue in a favorable condition for cyanide treatment. Preliminary acid wash of the roasted material was also tried; this would have the advantage of dissolving a further quantity of copper that might have become converted into oxide in the roasting, but the results show that the benefit obtained would not warrant the additional cost. Another test was made in which the concentrate was cyanided raw before roasting and acid-washing, and re-cyanided after the washing: this also showed no advantage either in extraction or cyanide consumption over direct roasting. water-wash, and cyanide. In all cases the roasted material was agitated with cyanide, using a dilution of 3:1. The results obtained by these three methods are shown in the following table. (No. II.)
In Test No. 2 the acid-washing was made with 1% H2S04, using approximately 5 tons of wash per ton of concentrate treated. Before agitation with cyanide, the pulp was re-ground in a model tube-mill with glass marbles.
In calculating ‘recovery,’ allowance is made for losses of gold and silver in roasting.
Test No. 1. Oxidizing roast, water-wash, and cyanide.
Test No. 2. Oxidizing roast, acid-wash, and cyanide.
Test No. 3. Cyaniding raw, roasting, acid-wash, and re-cyaniding.
In Test No. 3 the preliminary raw treatment was made with 0. 1% KCN using a dilution of 2:1, for two days; the extraction of gold was 12%. Acid treatment was made with 1% H2S04, dilution 1:1, agitated 18 hours, and then leached with water before cyaniding. Roasted and washed concentrate agitated with cyanide for 4 days.
Test No. 1 on Table II indicates that the flotation concentrate from the San Sebastian ore may be successfully treated by a simple process of roasting, water-washing, and cyaniding. This conclusion was confirmed by numerous experiments on a large scale in which the material was roasted in a hand-reverberatory furnace, and the roasted product treated by agitation in tanks with mechanical stirrers, adding water, settling, and decanting until the bulk of the copper and iron salts was removed, finally collecting the material on a vacuum-filter and washing on the filter to remove the last traces of soluble salts. The washed concentrate was then re-pulped with lime and cyanide solution in an agitation-tank, and treatment continued in the ordinary way. The results of the bottle tests were fully confirmed.
Attempts to treat the material by percolation were not successful. Owing to the fine grinding of the ore previous to flotation, the roasted material showed a tendency to slime; percolation took place slowly and irregularly, through channels formed in the mass, so that the extraction by this means was always imperfect.
In the tests made in the large muffle the oxidation was somewhat more effective, but a rather longer time was required to reach the temperature at which roasting began. Temperature was approximately determined by Seger cones.
On examining the details of Table III, it will be apparent that the most favorable results were obtained when roasting was carried out at a low temperature; under these conditions a maximum amount of copper was extracted by water-washing, and the highest extraction of gold obtained with a minimum cyanide consumption.
In this ore the silver is negligible, but it is significant that the silver extraction on the roasted material is poor in all cases. This condition will be noted in most cases where attempts have been made to treat silver ores by cyanide after an oxidizing roast.
With these results as a guide, tests were made on a larger scale on the same material, roasted by hand in an oil-fired reverberatory furnace. A charge of about 400 lb. was dried slowly in a sample drier, and charged into the furnace; the temperature was gradually raised till it approximated that obtained in the muffle-roasts, probably about 550° C. After 3½ hours, the flame was turned off and the charge allowed to cool in the furnace over-night. The concentrate roasted in this way, showed little or n0 tendency to sinter 0r form lumps, but in subsequent tests when the material was charged without previous drying, portions of the concentrate agglomerated into comparatively hard lumps, which contained a core of unroasted material, and which it was necessary to sift out and re-roast after grinding. Possibly in practice it would be advisable to pass the material, after drying and before roasting, through a ball-mill or similar pulverizer.
A bottle-test made on a scale of 100 gm. on a sample of roasted concentrate from the above reverberatory charge showed the following results:
The loss shown in this test seems to be mostly mechanical, due to dust carried off while stirring the charge; it could probably be much reduced by using a suitable roaster with revolving rabbles and a dust-chamber.
AGITATION Leach TESTS
The following tests were made in small tanks fitted with wooden paddles.
No. 1. Agitated with cold water, washed by settlement and decantation, then drained by vacuum on a horizontal filter-tray; re-pulped with lime and cyanide solution.
No. 2. Agitated with hot water, washed by settlement and decantation, neutralized with lime and agitated with cyanide without previous filtration. Cyanide treatment by decantation, finally washing without water.
Portions of the same roasted charge as were used for the previous tests were leached in tanks with a canvas filter, using vacuum to aid filtration. After washing out soluble salts as far as possible in this way, the residue was mixed with lime and treated by percolation with cyanide solution in the same manner.
In view of the unsatisfactory results obtained by percolation and the fact that further extraction was made by increased water-washing, the residue of charge No. 1 was mixed with the residue of charge No. 1 treated by agitation (see Table V) and the united charge agitated further with water, then with weak cyanide solution and finally with water again.
The result of these tests indicates that thorough washing is essential for a high extraction. Filtration without vacuum was found to be practically impossible.
The following tests were made on another portion of roasted concentrate, to determine the influence of cyanide strength on extraction. Eight tests were made; in the first four a preliminary wash was given with hot salt solution, 10% NaCl using 2 tons of material treated. In the remaining four tests the preliminary wash was given with water alone, using 4 tons per ton of material. The salt wash showed some extraction of silver, but it does not appear that any advantage derived from this would warrant the additional cost.
The cyanide treatment is detailed in the accompanying tables:
From these results it appears that the extraction is scarcely affected by variation of cyanide strength within the limits and under the conditions of the tests. The cyanide consumption, however, increases with increasing strength. Apparently the best results are obtained with a strength of 0.125% KCN.
Two tests were also made on another portion of roasted concentrate from the same lot of ore to determine the influence of time on the cyanide treatment.
From this test it is evident that the gold in the roasted concentrate is rapidly soluble in cyanide. The small insoluble portion seems to be quite inaccessible to prolonged treatment or to stronger solutions.
The foregoing tests sufficiently indicate that the San Sebastian concentrate, produced by flotation, may be treated successfully on a commercial basis by the method of roasting, water-washing, and cyaniding. Some tests were, however, made by alternative methods for the purpose of comparison.
The processes thus tried were:
- Chlorination by saturating the roasted charge with chlorine gas and leaching with water, as in the old Plattner process.
- Direct cyanide treatment of the raw ore after fine grinding in a tube-mill with steel balls.
A charge of roasted concentrate was moistened with about 15% of water, and placed loosely, without any paper or other filter, in a porcelain funnel with flat perforated diaphragm. Chlorine gas was introduced from below through the neck of the funnel, and after saturation, the charge was allowed to stand covered for 24 hours. It was then leached out with water and the residue dried and assayed. The extraction was found by difference of head and tail assays; it was also checked by precipitating the filtrate with ferrous sulphate, allowing to settle and collecting the deposited gold on a filter. This was dried and cupelled.
In one case the residue after chlorination was further treated by agitation with cyanide. The results obtained by chlorination are detailed in the accompanying table.
The residue from test No. 3 (Table X) was further treated by cyanide, by agitation for 4 days with a solution originally at 0.2% KCN, and increased toward the end of the treatment to 0.5% KCN, using a dilution of 3 :1. This treatment yielded the following results:
From these figures it would seem that the results to be expected from chlorination, or from chlorination followed by cyanide, are in no way superior to those obtainable by water-washing and cyanide. Either method will give a satisfactory extraction and the choice would depend on relative cost under local conditions.
Direct Cyanidation of Raw Ore
It is interesting to compare the results obtained on this ore by direct cyaniding without any form of concentration, with those given by the combination of flotation and cyanide.
The following tests were made on portions of the same lot of ore as was used for tests detailed in the preceding tables. Three charges were treated; the first two were taken from a portion crushed in a small tube-mill with manganese-steel balls, using the following quantities:
Ore, 25 lb.; lime, 0.25 lb.; water, 17 lb.
Time of grinding, 6 hours.
The pulp was drained on vacuum-filter to 26.4% moisture.
The third test was made on part of a larger portion of ore crushed in the same manner, but in a larger mill, for use with a 200-lb. flotation machine.
Cyanide Treatment of Flotation Tailing
A flotation test was made on the ore used in Test No. 3 (Table XI), resulting as follows:
The tailing thus produced was agitated with cyanide, with results as shown below:
Comparison of Methods
For the sake of comparison, we may assume in view of previous results that an extraction of 98% could be obtained from the concentrate yielded by the above flotation test, by the method detailed, namely, roasting, water-washing, and cyaniding. The values shown in the middling may be eliminated on the assumption that in practice the middling would be constantly returned to the head of the machine, and that finally only two products would be obtained, concentrate and tailing, having the same assay-values as in the test. The result of the flotation test would then appear as follows:
Taking the figure of Test No. 3 (Table XI) as indicating the possible recovery by direct cyanide treatment we have:
In addition to this there is a saving in cyanide consumption as follows, per ton of raw ore:
showing a saving in cyanide of 2.99 lb.; taking cyanide at 16c. per lb. of KCN equivalent, this would amount to 48c. per ton of raw ore treated, bringing the total saving to $2.74 or about $9600 on a monthly output of 3500 tons.
The tests given in Table No. III show that about 0.4% of copper, or 8 lb. per ton of raw concentrate (in this case 2 lb. per ton of raw ore) can be extracted in a soluble form, and might be recovered as an additional source of revenue.