from rich material, while the Denver clay with extra thin walls has proved better for low-grade silicious stuff. THE ACTUAL ASSAY. Fluxes. The cream of tartar must be 99.5 to 100 per cent. pure, and all the chemicals must be free from sulphur or sulphates. For soda, it has been found advisable to buy the "Natrona brand” sodium bicarbonate, which is fused in a large iron ladle, poured into a smooth iron dish, or ladle, to cool, and then pulverized through a sieve of 20 meshes to the linear inch. Refined crystallized borax is fused in a Dixon graphite crucible, poured into an iron dish and pulverized to pass a 20-mesh sieve. Powdered silica is occasionally employed for very ferruginous slags, such as blister slag, and powdered hematite (Fe2O1) for very silicious material, such as gravels, or carbonates and silicates of copper. Weighing. The moist material is dumped by the assayer on a sampling board about 15" x 20" in area. Operators formerly used a tray made of straight-grained wood, which has to be wet with water frequently and the superfluous moisture removed with a cloth. The best thing, in every way, that we have used is a vulcanite tray, 15′′ x 19′′ in size, such as is sold for photographic purposes. The long side of the tray next to the operator is cut out, and we have a sampling platform which does not absorb moisture, and does not break or wear out with months of use. The assayer quickly mixes the sample with a large steel spatula, and covers up the pile which is then pushed outward from the center and flattened to a layer about 1 inch thick and I foot square, or a little more. With a steel spatula (8-inch blade), vertical slices are quickly taken at regular distances apart (generally from 9 equi-distant points), dipping to the tray each time in such a manner as to make up 1000 grains on the pan of the pulp balance. This first sample is put in a dish or can, to be dried for the estimation of water, and a second portion is immediately taken for a copper assay and placed on oiled or smooth white paper. Mineral containing 70 to 95 per cent. copper makes so little slag of itself and needs so little flux, that there would not be enough from the gangue present, to cover the button and keep it deoxidized while cooling. On this account a large stock of old assay slag is accumulated, preferably from low grades, ground to pass a 20-mesh sieve, assayed for copper, and a certain quantity added to each crucible charge of rich material to act as a regulator and a cover for the button. If known grades, varying within narrow limits, are regularly assayed, a large quantity of complete fluxing mixture can be made up and added by measure instead of weight. Otherwise, weigh fluxes roughly, according to formulas determined by experience, mix with the spatula, transfer to crucibles and add a cover 1/16" to 1/8" deep of cream of tartar. Scrape away the hot coals from the surface of the rows of brick in the furnace, close each crucible when in position, with a clay. cover to keep out coal, poke the hot fuel around the crucibles, and fill in fresh coal (egg size) up to 1 inch below the top of the sand crucibles, or a little higher for Denver clay crucibles. The covers can be made in wooden molds from a mixture of fire-clay and ground fire-brick. Old crucible bottoms luted with clay could be employed. When using soft coal, the writer has the fire poked gently in from five to seven minutes from the start (according to the sharpness of the heat) coaled lightly in four minutes more, poked again five minutes later, and the assays taken out two to three minutes afterwards, the furnace having come to a very high heat -total time sixteen to eighteen minutes. Leave any charge not melted, until it just comes to quiet fusion. The melts should not be jarred while cooling. The success of the assay depends largely on the quick and sure regulation of the time and degree of heat. The above manipulation is for soft coal and a double-walled tight furnace. Coke, or gasoline fuel, would require different manipulation to get the same results in the time allowed (in general not over twenty minutes for complete fusion). If low grades are to be fired in Denver "E" crucibles, the furnace and stack should be thoroughly heated, but the grate cooled down to a bright red by taking the furnace cover off for a few minutes just before firing the assays. The heat can be brought up within the time allowed, but the slower start helps to prevent pellets from sticking to the crucible. Specimen charges are here given to show the average weights which each flux loses in melting down, and thus to indicate a method by which any one can approximately calculate what total slag will be produced in any case, and, in conjunction with the colorimetric assay of that slag, the total copper slagged away from the button. SMALL NO. 5 HESSIAN TRIANGULAR CRUCIBLES. Composition of charge. Losses in melting Rich mill concentrates, sample 1000 grains. (Cu + H2O)=88 70 to 50 grains sodium carbonate.... 70 to 50 grains borax glass.... 250 grains assay slag (0.3 per cent. copper).. 300 grains cream of tartar.... Calculated slag produced. Grains. 120 40 42 O 70 O 250 75 75 557 75 Actual weight of slag found to be 550 to 570 + grains. As the only appreciable impurity of copper assay buttons is metallic iron, if we desire an exact assay on any unknown sample, we have only to determine the iron in the button and the trace of copper lost in the slag, calculate both results to the 1000 grains of mineral tested, and apply the corrections to the weight of the button. This calculation will furnish also a guide as to the effects of experimental variation of fluxes and conditions of heat. Specimen charges employed in practice at lake refineries are presented in the following table. They are based on the original paper of Mr. M. B. Patch and on the later experience of the writer and assistants with a variety of material. It will be advisable to use the higher limiting quantity of soda on rich concentrates, when the old assay slag employed has been preserved from 1000 grain assays of low-grade material only, as such slag is proportionately lower in soda and oxide of iron than the usual slag produced years ago. Conglomerate mineral contains much more ferric and titanic oxides than the amygdaloid. In all cases there is some fine metallic iron in mill products, due to the wearing of the stamps. no sulphur. H or B 1000 350-400 300 350 50-150 17-20 1 Quick heat. Increase cream of tartar for black oxidized stuff. 2 Best assay for very coarse SIL It is evident, from the foregoing description, that the Lake Superior fire assay bears no relation to the empirical matting assay in use at Swansea, but is a practical method, easily controlled when once thoroughly learned, and sufficiently accurate for all technical purposes. NOTES ON THE ESTIMATION OF COPPER BY POTASSIUM PERMANGANATE. BY H. A. GUESS. ALTHOUGH at various times, during the past few years, different chemists have proposed the use of an alkaline thiocyanate as a precipitant in the estimation of copper,1 and although the results have been shown to be sufficiently accurate for technical purposes, the fact remains, that the method has made little headway, in competition with the electrolytic, cyanide, or iodide methods, in technical work, particularly in the metallurgical laboratories of the West, where the iodide method is greatly in favor. In the iodide method, as generally practiced, the copper is first precipitated by aluminum strips, and while this precipitation is fairly rapid and complete in concentrated solutions, it is not at all so in dilute solutions of 20 milligrams or less of copper in a bulk of 150 cc. of dilute sulphuric acid. Furthermore, other metals, particularly arsenic, interfere by being precipitated along with the copper, and have to be rendered inert by oxidation with potassium chlorate. On the other hand the precipitation of copper as thiocyanate is practically complete at the moment of adding the precipitant; no other metals interfere, silver being the only other metal precipitated by an alkaline thiocyanate from an acid solution, and its thiocyanate is unacted upon by the subsequent caustic alkali treatment. It has seemed to the writer that the main drawback to the method as so far proposed is the use of asbestos as the filtering medium, and the consequent necessity of filtering-pumps, which in a laboratory doing 100 or more copper assays in a day would render the process awkward and cumbersome. For the past six 1 This Journal, 20, 610; 22, 685, et seq; Am. J. Sci., 163, 20–26; Chem. Centrbl., I, (19), 1085 (1899). |