The aim of instruction in a metallurgical laboratory is to make real the principles on which metallurgical processes and operations are based, and to foster the spirit of investigation. The materials with which experiments are carried out are ores, metals, and metallic compounds. The method varies with the end sought. A class may work as a whole, each member contributing his share to the solution of a problem, or the students may carry on investigations independently; the former exemplifies class-research; the latter, individual research. It is with a branch of the former, with special reference to ore-treatment, that the present paper deals.

In smelting an ore by a well-established process, the result is shown by analyzing the products to see whether their compositions correspond to those calculated in making up the charge; by taking account of stock to show the distribution of metal in the different products made and the losses from dust and volatilization; by casting a thermal balance to find the distribution and losses of heat; and by making a cost-sheet to ascertain, as far as possible, the necessary outlay of money.

In lixiviation and amalgamation, the mode of operating has to be varied to adapt a process to the individual ore. Here a number of tests become necessary. Each will consist of a series of experiments with one variable, in order to find the conditions under which the variable gives the best result; a summary of several tests will give the best method of operating. Working a number of charges with the best method will furnish the data desired for ascertaining the recovery of metal, the size of plant needed for a given capacity, and the cost of treatment.

The Washoe process, raw-amalgamation of a silver-ore in an iron pan, furnishes a satisfactory example of this class of ore-treatment. The object of the present paper is to show how pan-amalgamation is carried on as a class-exercise at the Massachusetts Institute of Technology.

Amalgamating Pan

The first amalgamating pan with settler was put in operation at the Massachusetts Institute of Technology in 1871. It was built on the Washoe pattern by Messrs. Booth & Co., San Francisco, Cal. In 1895 the laboratory bad three of these pans, respectively 30, 18, and 12 in. in diameter. They were, however, little used at that date for class-work on account of the time required to get them into good working-order, and of the difficulties met with in making a clean-up, as it was next to impossible to recover all the amalgam from a pan with detachable shoes and dies. While the percentage of extraction is usually based upon the assay of the tailings, it is of importance in a teaching-experiment to compare it with the actual yield from the amalgam. Instruction was given mainly with small pans, only 7 in. in diameter, three of which were of copper, hardened by a small percentage of silicon, and the rest of cast-iron. A drawing and a brief description of these have already been given in a paper read before the Institute. These pans were a great advance over the original Washoe model, but improvements in the details of construction suggested themselves, which led, in 1899, to the replacement of the pan of 1895 by the present form. This has met all the requirements of a pan that is to be used for class-work in the systematic testing of ores. These requirements are that it shall give a quantitative result which corresponds to working- conditions, and that it shall be small, easy to run, and easy to clean.

The battery of ten pans in the laboratory is represented in Fig. 1, and detailed drawings of the pan in Figs. 2 and 3. The pan, Fig. 2, is cast in one piece. It has a flat bottom, which forms the lower grinding-surface; its inside dimensions are : diameter at bottom, 7 in.; at top, 7.25 in.; height, 4 5/8 in. In the center is a hollow core, 2.75 in. in diameter and 3.75 in. high, to prevent the pulp from collecting. The pan has four legs, which stand on a wooden stool; the latter carries a flat evaporating gas-burner for heating (not shown in the illustration).

The muller, Figs. 2 and 3, is of special construction. It is cast in one piece, as is the pan. The upper part, the driver, is slipped over the rotating-shaft and fastened to it by a set-screw ; it has a vent, Fig. 2, to prevent hot pulp from being sucked into the core; the spider has two legs only; the form of the muller-plate and shoes is given in Fig. 3; details of construction are given in Fig. 2. The shoes have the usual form of an oblique sector of a circle. One peculiarity of the shoe, seen in half-section on D E and in section on L M of Fig. 2, and in Fig. 3, is that the part outside of the muller-plate tapers from 0.25 in. at the front to 7/8 in. at the rear end, and thus assists in the formation of a pulp-current by raising the pulp while the muller is being rotated through the driving- shaft.

This shaft, Fig. 2, is suspended from a bevel-gear with hub, fastened to it by a set-screw, and journaled in two boxes bolted to a wooden frame common to the ten pans, Fig. 1. The bevel-wheel is driven by an adjustable bevel-pinion, which is thrown in and out of gear by a forked lever (stopper), the arms of which end in a grooved hub; the lever is supplied near the bottom with a hook, to be lowered into an eye (not shown) when the pinion is working.

Operating an Amalgamation Plant

During the winter term the whole of Tuesday and the afternoon of Wednesday are given to class laboratory-work; and the following Saturday one hour is devoted to the discussion of results. Exercises in pan-amalgamation are so planned that a student starts his experiment Tuesday morning and finishes it Wednesday afternoon; the several results are handed to an instructor, who tabulates (Table I.) and plots them (Fig. 4), and prepares manifold sheets, which are distributed at the conference on the following Saturday, when the work as a whole is passed in review. Thus a series of as many as 10 tests, with one variable—for example, the time of grinding—is carried out by the class in one and one-half days, and the whole experiment finished up the same week.

Another section of students will make a test with the same ore, choosing a different variable—for example, the time of amalgamating—and carry it through in a similar way. In this manner the principal variables in pan-amalgamation are taken up by class-sections, each section benefiting by the work of the others, while at the same time the best- manner of treating the ore under consideration is being investigated.

The details of an operation may be given in connection with Table I. and Fig. 4, which represent a series of 10 tests, in which the time of grinding was the variable, ranging from 20 to 100 minutes.

The pan is first cleaned. For this purpose the muller is raised on the shaft and clamped, the wooden stool under the pan is withdrawn, the pan taken out and dusted, and the suspended muller freed from adhering particles of foreign matter. The pan is now put in place, the muller lowered, pressed down, and turned to and fro by hand and clamped. The necessary amount of water, 500 c.c., is charged, the muller set going, the lamp lighted, and the salt, 180 g., added. The ore, 1,800 g., of 40- mesh size material, is fed in slowly, and the time of grinding counted after all the ore has been charged—namely, after about 5 min. On account of heating the pulp with a lamp to about 80° C., there is considerable evaporation of water during the grinding- and amalgamating-periods, which has to be remedied by adding fresh water, in this case from a wash-bottle holding 500 c.c. The amount used is noted, as it gives an idea of the care with which the heating has been carried on. Allowing the pulp to become too thick requires an excess of water over the normal to thin it down in order that the desired current may be again established. Table I. shows that the water-additions ranged from 465 to 869 c.c. Water from the wash-bottle should be blown in small amounts against the side of the pan; it will

loosen parts of the top of the charge which have adhered to the warm pan and become hard; the pulp-current then will carry them towards the center and cause them to descend there. Scraping the sides with a spatula corrects the adhesion of parts of the charge, and has to be resorted to more or less during the larger part of a test, as repeated additions of water thin the pulp to such an extent as to spoil the current.
At the end of the grinding-period, the muller is raised 1/8 in. previous to adding the quicksilver. In order to fix the distance, a pencil is held against the rotating shaft and a line

marked off; the bevel-wheel is now thrown out of gear, the muller unclamped, raised, reclamped, and set rotating again. A weighed amount of quicksilver, 150 g., is then added during about 5 min. in a fine spray from a glass funnel on to the charge near the outer edge. The consistency of the pulp must be a little thicker than during grinding; the right degree will be determined by the manner in which the quicksilver is disseminated ; this will be found in globules if the pulp is too thick, in fine particles if right, and will collect on the bottom if too thin.

In making a clean-up, the first step is to remove the tailings and amalgam from the pan. The muller is stopped; a sheet- iron vessel, 18 in. square and 4 in. deep, having a thin coat of pitch and tar, is placed underneath the pan; the muller is again set going, and first 1 liter of water added to thin the pulp, then 2 liters more in about 5 min., which causes a large part of the tailings to overflow into the vessel. The amount of water desired and the time allowed for adding it had been settled by experiment before adopting this mode of operation. The rest of the pan-content is now transferred to a fiber pail holding about 2.5 gal., the pan and muller being scraped with a spatula and brushed with a dauber. The next step is the separation of the amalgam from the tailings and the recovery i of the latter. The tailings collected in the iron vessel are transferred to a filter, which is a simple wooden frame, 18 in. square and of 1-in. section, with heavy unbleached cotton cloth spread over it and nailed fast on the under side. The 10 filters are soaked for several hours in water before they are put to use, in order to close the pores. Nevertheless, small amounts of slime pass through, which are caught with the filtrates in buckets and allowed to settle over night, when the clear liquid is decanted and the slime collected from each filter, dried and weighed (“ through filter ” in Table I.). The tailings and amalgam, collected in a bucket, are separated by panning twice in a 16-in. gold-pan; the tailings go on to the filter-cloth and drain over night; the amalgam, collected in a porcelain dish, is dried and weighed. The discrepancy in weights of quicksilver and amalgam in the table requires explanation. The combined weights of quicksilver fed, 150 g., and silver contained in the charge, 5.499 g., give 155.499 g., while the weights of the amalgams recovered show a range of from 156 to 164 g. Part of this excess is due to a possible slight overweight in the quicksilver charged; part, however, to the presence of impurities in the amalgam. Thus, a partial analysis of retort-bullion gave: Ag, 51; Pb, 43.61; Fe, 5.12; Cu, tr.; total, 99.73 per cent. The 10 amalgams of a series of tests are placed separately in half-cylinder cast-iron vessels, transferred to a pair of retorts, and distilled in a two-muffle furnace, which is fired with soft coal; the muffles are 4 in. wide, 6 in. high, and 18.25 in. deep. Each vessel is coated with chalk, and

receives a layer of paper before the amalgam is charged, in order to prevent the retort-bullion from adhering to the iron.

The general arrangement of apparatus is clear from Fig. 5. It consists of two wrought-iron pipes, 3 in. in diameter and 24 in. long, each closed at one end by a disk, 0.75 in. thick, that has been welded in, and at the other by a reducing T, 3 by 3 by 1 in., and a reducing cross, 3 by 3 by 1 in., respectively, and joined by a 1-in. connecting-pipe; the T and the cross are closed by square-head screw-plugs. Into the lower retort is screwed the condenser, which reaches into a vessel filled with water.

The quicksilver is driven off in about 3 hr.; the water in the condenser has to be replaced at intervals; a continuous flow of water was found to be unnecessary. The retorts and the furnace are allowed to cool over night.

The morning after the run has been made, the drained filters are placed on steam-tables to become thoroughly dry, and the retort is opened. In the afternoon, each student receives the tailings and the retort-bullion from his pan; he passes the tailings through a 40-mesh sieve to break up the lumps, samples them down to 200 g., crushes the sample through a 100-mesh sieve, and makes a duplicate assay; at the same time, he scorifies his retort-bullion and cupels it. By using a large muflle, 8.5 by 5 in. and 18 in. deep, a number of crucible-fusions can be made at the same time, and thus the work expedited. The results are handed to the instructor, who records the weights and assays.

In Table I., the weights of tailings on the filters range from 1,775 to 1,797 g., equal to from 98.60 to 99.84 per cent.; those that passed through the filters weigh from 1 to 8 g., equal to from 0.06 to 0.44 per cent., which gives a loss in weight ranging from 1 to 20 g., or from 0.06 to 1.11 per cent. The weights of amalgam and retort-silver show some variations. The combined recovery in quicksilver from the 10 tests is high, 99.13 per cent. In making up the silver-account, the tailings show assay-values of from 13.92 to 16.68 oz. of silver per ton. A pan was charged with 5.499 g. of silver; this is the total to be found in the tailings and in the amalgam; the amount accounted for is seen to vary from 99.49 to 99.97 per cent. The last two columns give the extraction in silver based upon the tailings-assay and upon the recovery in the amalgam. The former figure is, of course, the only reliable one, but the other column is added to bring out any contrasts which may exist, as they form a valuable means for instruction. It is an accident that the figures of the two columns agree so closely; frequently considerable discrepancies occur, due to imperfect cleaning of the pan in a preceding test, or to hard amalgam adhering to the muller in one case or peeling off in another.

Fig. 4, finally, shows graphically the extraction of silver as influenced by the time of grinding. It is seen to increase with the time of grinding from 20 to 60 min., when it reaches a maximum of 84.76 per cent., and then to fall off. The probable reason for the diminished yield, after 60 min. of grinding, is the excessive sliming of the ore, which affects harmfully the pulp-current and flours the mercury, which increases the losses. Without a good current a satisfactory extraction is hardly ever obtained.

Summary of Amalgamation Tests

The following is a summary of a large number of tests made in extracting the silver from a single ore by raw-amalgamation in cast-iron pans of the construction given. They represent the first experiences of students in this kind of work, who, however, are familiar with assaying and panning. In the selection of samples only those tests have been omitted which in the class-conferences were decreed to be faulty for some well-ascertained reason.

The ore is a silver-ore from the Palmarito Mining Co., District of Mocorito, Sinaloa, Mexico. An examination, aided by the microscope, showed that it was composed mainly of quartz and kaolinite, and contained besides some hematite, galena, pyrite, native silver, and cerargyrite. In the pulp, crushed by means of rolls through a 40-mesh sieve, were found particles of metallic iron. The ultimate analysis gave : H2O, hygr., 0.07; SiO2, 86.10; Fe, 6.68; Al2O3, 2.66; S, 0.07; Pb, 0.28; Ag, 0.31 (89.1 oz. per ton); As, Sb, Cu, absent.

The rational analysis was determined by the following considerations : The Al2O3 was calculated as kaolinite (Al2O3, 39.8; SiO2, 46.3; H2O, 14.9); the remaining SiO2 was assumed to be quartz; the S not required by Pb to form galena was calculated as being bound to Fe as pyrite; metallic iron to the extent of 0.10 per cent, was extracted by a magnet from the pulp; the remaining Fe was figured as hematite; of the Ag present, 0.02 per cent, was extracted by means of sodium hyposulphite and calculated as cerargyrite; the rest was assumed to be present in the metallic state. This procedure was believed to be warranted by the facts that As and Sb were absent, that more than 70 per cent, of the total silver was amalgamated in 20 min., and that more than 80 per cent, was recovered in the pan in the absence of salt. The rational analysis of the pulp thus gave: H2O, hygr., 0.07; quartz, 83.01; kaolinite, 6.68; hematite, 9.36; galena, 0.38; pyrite, 0.06; metallic iron, 0.10; cerargyrite, 0.03; metallic silver, 0.29; total, 99.98 per cent.

In the tests there were examined the effects of varying the addition of salt, the time of grinding, the time of amalgamating, and, last, the influence of an addition of blue vitriol. Previous experiments had shown that, with a charge of 1,800 g. of ore, 500 c.c. of water at the start gave a satisfactory pulp, and 150 g. of quicksilver an amalgam of sufficient liquidity to reduce the loss in panning to a negligible quantity. These three items, therefore, were kept constant in all the work, as well as the temperature, which was held at about 80° C.

The results, recorded in Table II. and Fig. 6, show that, while the extraction in the absence of salt is high, it rises rapidly until 6 per cent, of salt has been added, falls with 10 per cent., and rises again to practically the former maximum when 15 per cent, of salt has been charged.

(The reason for the falling-off in extraction between 6 and 15 per cent, of salt is not clear, and will have to be looked into at a later date. The salt-series was the last one that was investigated; in the other tests the usual standard addition of 10 per cent, had been made; this explains the discrepancy between advocating 6 per cent, of salt and using 10 per cent.)

There is, therefore, no reason for going beyond 6 per cent. The high extraction without the presence of any salt whatever points to the supposition that a large part of the silver is present in the metallic state.

The extraction, Table III. and Fig. 7, is seen to resemble very closely that given in Table I. and Fig. 4.

The data, Table IV. and Fig. 8, show clearly the rapid rise in the extraction during the first hour of amalgamation, and the small increase during the next half-hour, when a maximum is reached. The slight falling-off later on is to be attributed to the inevitable flouring of quicksilver in every amalgamation-process, with a consequent loss in silver.

The addition of blue vitriol to the pan, as shown in Table V. and Fig. 9, has no beneficial effect whatever; on the contrary, the extraction decreases. The irregularities in the results are due to the amalgamation of copper, which causes losses in panning and in the subsequent assaying.

The inferences to be drawn as to the treatment of the ore are that, with a charge of 1,800 g.,with an addition of 500 c.c. of water at the start and smaller amounts later on to keep the consistency of the pulp constant, and with 150 g. of quicksilver, 6 per cent, of salt, 60 min. grinding and 90 min. amalgamating give the highest extraction.

Conclusion

The data and the curves drawn from them show that the results are satisfactory, and especially so when it is remembered that they represent the work of students making such tests for the first time. Considering pan-amalgamation as a laboratory-experiment, it teaches in a simple, quick, and effective way, with an apparatus that is inexpensive, the importance of series-work in making an investigation and the value of taking account of stock; it further gives results that can he used as the basis for work on a large scale.