THE ANALYSIS OF COPPER McGraw-Hill BookGompany Electrical World The Engineering and Mining Journal Engineering Record Engineering News Railway Age Gazette American Machinist Signal E,ngin-; ** >*" f ANALYSIS OF COPPER SPECIAL EQUIPMENT 2. Office Records. Large works keep a systematic record of all tests except for those of temporary character. A serial number is given to every sample that comes to the labora- tory, and printed cards or tickets are provided upon which are entered all the data of weights, sampling, and analysis. These records are subsequently bound, catalogued, and filed away for ready reference. A few plants have adopted a system of 5 x 8 inch tabbed cards which are filed in cases, each drawer holding 1000 cards. The U. S. Metals Refining Co. use a triplicate series of tickets, on which the descriptions are entered by a clerk to relieve the chemists. The assay slips are printed in three colors, white, red, and yellow, with carbon papers between them. The white slip is filed eventually in the clerk's office, the red slip is sent to the laboratory with the sample, and the yellow one to the assay department. After the chemist in charge has checked the work and signed the assay certificate filled in by the clerk, the yellow and red slips are filed in the laboratory. The work of the Mansfeld Smelter in Germany is of a peculiar nature. The ore deposits are partly surface slates and partly true ores which contain a refractory mineral "typolite." In addition to daily reports, the completeness of German system requires that the chief chemist shall furnish a monthly thirty- page report of all the work of the month, including the ore and metal returns of mines, electrolytic refinery, and brass works. 3. Assay Furnaces of portable type are a necessity in remote places. The best of coal or gas fired muffle and pot furnaces are the rule in modern plants. A plan of a good pot furnace, adapted to either native copper or gold ores, is inserted. The depth of the furnace from cover to grate is 61 cm. and the grate area 32 x 46 cm. The ash-pit door is made with an extra weight, so that a blast of air may be admitted under the grate through a pipe 6.3 cm. in diameter. The air may be furnished by the smallest Roots' blower, belted to a shaft in the sampling room and driven by a small five-horse power electric motor. 4. The Furnace Accessories, devised by Dr. E. Keller, are of assistance in rapid work, especially on bullion. 1 A cupel charger 1 Trans. A.I.M. E. 36 (1906), 3. INTRODUCTION permits 48 cupels to be placed, or withdrawn, as a unit. The front row of assays are blanks, .the other 40 being charged with the lead buttons. Other devices are parting baths with holders, multiple scorifier tongs, a stirring machine to be used with Cross Section ErF Cross Section C~B Common Brick Cross Section GrH Fig. 1. Assay Furnace. beakers, and a mechanical filtration apparatus in which 20 silver assays, contained in 750 c.c. beakers, may be filtered simultane- ously. Although riffles of correct design are easily purchased, they are quickly and cheaply made. Haultain has described their manufacture. 1 1 Eng. and Min. Jour. 73 (1907), 232. ANALYSIS OF COPPER APPARATUS STANDARD FLASKS AND ALIQUOTING PIPETTES 5. Copper Flasks, as they are called, and automatic " wash- out" pipettes are used in the electrolytic assay of crude copper. Water-jacketed " deli very pipettes" have been described by W. C. Ferguson of the Nichols Co. 1 A. M. Smoot gives the following directions for his simple " wash-out" pipette, to which the liter "copper flasks" are adjusted in mutual agree- ment. "In assaying inhomogeneous materials, such as blister or converter borings, which consist of roughly pulverized coarse and fine parts differing from each other in composition, it is necessary to weigh out a charge large enough to cover the variations. In some Fig. 2. Liter Flask for Copper Assay. cases, a preliminary separation of the coarse and fine parts of the whole sample is essential. The charge for assay is then formed by taking proportionate weights of each part. In any case, the sample to be weighed for assay will be larger than can be conveniently elec- trolyzed directly. "Aliquoting apparatus is therefore necessary, but the ordinary flask and hand pipette are not sufficiently accurate. Measurements within an error of one or two hundredths of a per cent are essential in dividing strong solutions of material rich in copper. A convenient liter flask is shown in Fig. 2 and the pipette in Fig. 3. " The flask has an enlarged neck which not only facilitates the introduction of the sample and reagents, but also renders the mixing of the solutions much easier when the volume is made up to the mark. The necks of these flasks below the enlargement should be narrow, so that the meniscus may be clearly defined against the mark. "Mechanical pipettes 2 with an automatic zero point may also be made to deliver the required volume, but in order to retain the required accuracy, such pipettes must be rigidly mounted to i J. Ind. and Eng. Chem. 2 (1910), 187. 2 Ibid. Fig. 3. Pipette. INTRODUCTION 5 prevent vibration, and must be kept scrupulously clean, since a trace of adhering dirt or grease will make a difference in the volume actually delivered. Such delivery pipettes are usually provided with a lower stop cock^land the slightest variation from the perpendicular in opening it will retard the flow of liquid, caus- ing a slight error. Each form of pipette has its advocates. "Hand Pipettes may be used if carefully compared and con- structed to hold instead of deliver an aliquot part of the contents of the flasks, and if the pipette is subsequently washed out." The Smoot pipette (Fig. 3) has a stop cock which itself serves as an automatic zero point. When the pipette is full, it holds an exact fraction of the contents of the liter flask from the bottom of the stop cock to the end of delivery tube. On turning the valve through an angle of 45, air is admitted through the funnel, allowing the contents to discharge; the pipette is then washed with three successive changes of water of 10 c.c. each, introduced through the funnel. The wash water spreads over the inner surface, removing every trace of reagent. The pipette is filled from the bottom and is worked automatically by a filter pump. The mounting and connection are shown in Chapter X under the title, " Determination of copper by electrolysis." ELECTRICAL DATA 6. Normal Current Density, in the electrolytic deposition of metals, is stated as "amperes per square decimeter" (100 sq. cm.) of immersed cathode surface, counting both sides of the plate upon which the deposit is made. 7. Electrical power for the determinations by electrolysis is usually delivered as a direct dynamo current of 100-115 volts potential. If the current is only available by day, or is alternat- ing, it is better to install a small motor-generator set or charge storage cells to deliver a current at 6 to 10 volts tension than to bother with inferior primary batteries. Accumulators with a capacity of 10 amperes and a voltage of 2.2 to 2.5 per cell cost about $10 each. The current may be reduced by small ventilated rheostats, one of which is connected to each independent group of six pairs of electrode clamps. The wires leading to each set should be provided with a special switch connection to an ammeter and voltmeter. The first instrument may read up to 10, the latter to 5 units. ANALYSIS OF COPPER 8. Electrolytic Cabinets are varied in design according to the work. The Tennessee and Cananea Companies use the Guess-Haultain cabinet, which is specially adapted to ore-testing, because the electrodes are made of very thin corrugated foil and sand-blasted. A heavier equipment is required for the assay of metal. There are two systems in use for connecting the individual assays, or clamps, to the main wires from the generators. In the first sys- tem, represented by the rack of A. M. Smoot (Fig. 4), or the T5 .WTfr E JW42L1L, BED Vf<)Vf ^)| (%) 6 ii iT/.i " y M =Q I fl -) j: ife a a A == y . a ^^ a y a a a a k iij /* Mahogany fase .- Aluminum Ang/e 0- Aluminum Rod H- Cut-out Plugs, Brass, Tops. Fig. 4. Rack for 10 Electrolyses in Series. cabinet of E. Keller, the assays are arranged in series, and a plug must be inserted as each assay is withdrawn, to allow the current to pass. In the second system, also in general use, the assays are ar- ranged in small sets of five or six assays, in parallel connection. The withdrawal of any assay does not break the circuit, but causes the remaining assays to receive more current, temporarily. (a) Series System. Mr. Smoot has a special device to per- mit the rapid withdrawal of electrodes without breaking the circuit, or permitting the contacts, at any time, to mar the wires and alter their weight. There should be no exposed screws or brass connections liable to corrosion above the assay beakers. The simple form of rack, illustrated in Figs. 4, 5, 6, is arranged INTRODUCTION A ' Mahogany Base B s Alumirwm Posf C - 5 fee I Spring D * Hard 'Rubber Cap E = Aluminum Angle F- Brass Machine Screw to hold ten assays in series, the maximum number which may be placed on one wooden bar of this type, although any number of racks ,may be joined, according to the voltage of the available current. The details of the special rubber binding post are shown in Fig. 5. The wires of the electrodes pass through the holes in the rubber caps and are held firmly in the V-shaped slots in the aluminum rods by the pres- sure of the springs, which tend always to push the caps to one side. An immediate release is effected simply by pushing the sides of the rubber caps against the springs. There are no wires, the current being carried through aluminum or brass rods of larger cross-section. All metal over the beakers is protected, the con- nections being shown in illustrations No. 5 and 6. The current of two cells is em- ployed, although current from a dynamo, through two rheostats, may be used instead. Fig. 5. Plan of Binding Post. The wires from Fig. 6. Arrangement of Circuit. battery I pass through all assays. At point P, a plug is inserted connecting bar X with bar Y, and the variable resistance VR of battery 2 is regulated so that the ammeter A registers .40 amperes. 8 ANALYSIS OF COPPER When it is desired to remove the assays, remove plug P, thus permitting the current from battery I only to pass through all the tests. Insert a plug at point S which will permit both batteries to supply assays 2, 3, 4, etc., but allow only the current from bat- --*--. IP.M!! J-LLJ _JK_. <& s\ ; -*$ mm H.LLLLU-U, ILUJAILLV. r -Side Thread R.H. All fo be Threads Standard Fig. 7. Brass Electrode Clamp. tery I to pass through test No. 1. This device permits the re- moval of the first assay without cutting off the current from the others. By a similar process the other assays are withdrawn, one by one. If it is desired to take off one (say No. 4) without disturbing the others, it may be done by inserting the plug at INTRODUCTION 9 point "T," thus short-circuiting test No. 4, which may be with- drawn so quickly that no coppep ^an redissolve. (6) The Parallel System, with tests arranged in small groups, as adopted by the Calumet & Hecla Laboratories, and by the American Brass Co., seems to permit as rapid manipulation as any rack yet designed. In the C. & H. apparatus, brass terminals rest upon a continuous shelf of hard vulcanite perforated with small holes just large enough for the electrode wires, so there is no chance for injurious corrosion of the terminals, which are soldered to the insulated leads. The electrode clamps are arranged in pairs on the long strip of vulcanite, or hard rubber, which is Fig. 8. Rack for Electrolysis Parallel System. 6.3 mm. thick and 10 cm. wide and is supported on brackets, with the under side of the strip 30 cm. above the main shelf. About six pairs of brass terminals are connected in parallel with two large copper wires which lead directly to an individual switch, rheostat, and ammeter switch. Assays may be taken off, one by one, without inserting any plugs, the current being gradually reduced after each removal, by means of the variable resistance. This system consumes a little more current than the first or series arrangement. The electrode clamps are made from square brass rods, 1 cm. diameter and 5 cm. (2 inches) in length. A V-shaped slot is made on the rear end of each, and the conducting wire, with the insulation removed at this point only, is laid in the slot and soldered. It may be more convenient, in small installations, to 10 ANALYSIS OF COPPER employ as variable resistances, small banks of incandescent lamp sockets, in which may be inserted lamps of varied candle-power, or amperage. Storage accumulators are conveniently charged through lamps as resistances. Figs. 7 and 8 show the design. The beakers rest upon wooden spools. The manipulation of the assays is described in Chapter XL (c) The European Equipment of the Mansfeld Works merits a short description. According to Hermann Koch, a current at 3000 volts is reduced by a transformer and motor-generator to a 100 ampere current at 8 volts tension, which is used to charge if Fig. 9. Mansfeld Electrolytic Table. three large accumulators, furnishing ordinarily about 25 amperes for electrolysis at the present time. Fig. 9 illustrates a special table for electrolysis. The resistances, which reduce the tension to 2-2.8 volts for each assay, are iron frames with 12 bobbins. Special connec- tions give equal tension at all electrodes, independent of changes in total load. The ampere meters read up to 5 amperes for each group of 3 to 16 assays. For the final division of the current (.1-.5 ampere per assay), small marble slabs are fastened to the back board, bearing two rows of brass connectors which carry the current to the positive and negative clamps and tripods. The American Brass Co. use a double system of conducting INTRODUCTION 11 metal posts, of similar design, but connected below the table, so that the wiring is concealed. - A rheostat is provided for each group of five assays. DEVICES FOR THE CIRCULATION OF ELECTROLYTES 9. Systems in General Use. 1st. Rotation of the anode, or cathode, by attachment to a spindle driven at high speed by a small electric motor. 2d. Rotation of the solution in the beaker by electro-magnetic force, as in the " Rotary Device" of Professor F. C. Frary. 1 (W. B. Price has obtained good circulation in brass analysis, by blowing compressed air upward through a small tube near the anode.) For the first system, a good anode may be made from a stout platinum wire to the bottom of which is clamped a round bladed pro- peller, 2.5 cm. in diameter. According to the inventor, 2 the Frary Solenoid, or "Rotary Device/' has been patented only in Germany, Fig. 10a. Frary Rotary Device. but can be obtained from the Vereinigten Fabriken fur Laboratoriums-Bedarf , Berlin, or their agents, the Standard Scientific Co., New York. Fig. 10a illustrates this model, and 106 shows a modification which the author was obliged to devise some years ago (before Professor Frary's paper had been noted) , in order to accommodate a beaker 12.5 cm. high and 5.7 cm. diameter. .A large size was designed to accommodate a No. 5 (750 c.c.) beaker, and provided with small tubes to conduct cold water between the beaker and the copper cylinder surrounding it. The copper cylinder is brazed water-tight to an upper and lower plate of soft steel, Z. Electrochem., 13 (1907), 308. 2 Letter. 12 ANALYSIS OF COPPER doff forming a reel, which is coated with mica and wound with about 500 turns of No. 13 (B. & S. gauge) insulated copper wire. 1 The electric current is usually passed through the electrodes and coil in series. A magnetic field is produced in the solu- tion which constantly diverts the lines of force passing radially from anode to cathode. With the largest size of appa- ratus, the author uses a 5 >Sfeel V A i \ i 'cm. ? (Copper Cylinder i *\Coil 00 Turns- 1 ( No._ 13 B&S. Gage i n-L \ !| IT H Fig. 106. Sectional View of Water-cooled Solenoid. cathode cylinder of plati- num sheet, 7.5 diameter, 11 cm. high, and perforated with holes 3 mm. diameter, spaced 1 cm. apart. EQUIPMENT FOR DRILLING OF CAST COPPER 10. Strong power drills must be used generally, and it is convenient to have one with a gauge on the vertical spindle. Small direct-connected electric drills are sometimes employed for gold and silver bullion. For ingots and the thin square plates recommended by Edward Keller for argentiferous metal (Chapter II), the author uses boxes 15 cm. in depth, made from T V-mch sheet steel, to preserve all the chips from the drill. The drill should be free from all traces of oil. The ingots, etc., are easily held in place by large flat wedges, and the square plates are supported by shoulder pieces brazed to the bottom sheet, inside the box. WEIGHTS 11. Standard Weights. In regard to this matter, the au- thor cannot too strongly recommend that all large laboratories preserve as a reference standard, a special gold-plated set of double-checked weights from 100 g. to 1 mg., to which the weights of all other balances shall be periodically adjusted by the method of substitution. 1 J. Ind. and Eng. C. 3 (1911). INTRODUCTION 13 For the daily weighing of the five-gram samples of refined copper, F. D. Greenwood employs a special five-gram brass weight, which is again placed 6n the balance in taking the weight of the electrode and deposited Copper. This brings the total de- ficiency, or difference from the original weight of copper, on the milligram weights and rider (see Chapter XI). CHAPTER II SAMPLING AND CRUSHING Treatment of the Subject. A number of the best mining experts have recorded their testimony to the fact that faulty sampling has been the cause of greater disagreement and losses than any inherent defect in the assay methods usually employed or any personal equation in their operation. For the mathe- matical theory of the subject, any one may consult the exhaus- tive papers of D. W. Brunton (1) and S. A. Reed (2) on ores, and E. F. Keller for metallic copper (26). Only the regular practice of mines and metallurgical works will be considered in this chap- ter (3). Whatever plan is locally adopted, it should furnish good agreement between contracting parties and, where samples are taken throughout the various mining and smelting operations of any company, the sampling should be done with such care that a full exhibit may be made of the absolute and relative gains, or losses, of the successive operations of ore and metal treatment. Divisions of the Subject. 1. Mines. 2. Western sampling mills. 3. Stamp mill control. 4. Carload lots of ores at eastern smelters. 5. Slags and copper products within the smelting works. DIVISION 1 MINE SAMPLING 1. In the valuation of prospects and control of mines by sampling and assaying there are three general systems. For a full description several papers are listed at the end of the chapter where all the references are tabulated. The general opinion of the best authorities is made the basis of the directions outlined. System A. The selection of a large number of small samples, cut quickly across the vein, at points evenly placed in parallel lines across the whole area, as exposed. The first, and least expensive, system gives correct results only on soft ores of uni- form grade and is used for testing new stopes, etc., in producing mines of uniform character. The grooves are made about 3 SAMPLING AND CRUSHING 15 inches (7.5 cm.) in width and 6 inches deep in soft ores or about 1 inch deep in quartz, and finally trimmed to uniform size. System B. A smaller number of heavy samples, at equal distances but farther apart, aod large enough, if blasting of grooves is absolutely necessary, to represent all the variations along the lines of cutting. Such a sample, in operating proper- ties, may weigh from 25 to 50 tons to represent a stope, although with new properties, the usual amount is 50 to 400 pounds (Kir by). Philip Argall assumes 5 pounds per foot across 10 feet of vein, or stope. This system is advisable for all American hard ores, where the values are very uneven and is the most frequently employed for accurate valuation. System C consists of actual "mill runs" of lots of 50 tons, and more, of regularly mined ore from each stope, or chute. The sample may be cut out by mechanical sampling mill after the ore has left the mine. Such a large works sample is necessary for native copper (see description by G. H. Benedict [7] ). Even such tests may be deceptive with regard to the quality of ore that can be regularly produced. 2. Practical Notes. These are based on papers of E. P. Mathewson, D. W. Brunton, and others. Ordinary duck, etc., will catch wire gold, or copper metal. The best cloth to use for mine rock, where the pieces are small, is soft flexible enameled oil cloth. Samples of over one ton, and made up of pieces at least three inches in diameter, may be reduced at the mine to one ton by throwing aside every tenth shovel, if the work must be done by hand. Such fractions are then reduced to one-half inch size, and five to fifty pounds weight, by crushing and dividing. A hammer tends to break off hard projections, while a geolo- gist's hammer seeks the soft spots; hence the best tool, in the opinion of many experts, is a moil struck by a four-pound hammer. The sampling stations should be tied to the survey stations on the mine maps to permit re-sampling. A conservative defi- nition of the phrase "ore developed" is that it means ore exposed on four sides. METHODS AT THE MANSFELD MINES 3. At Mines. The largest mines of Germany still find, after eighty years' experience, that the first system, A is best 16 ANALYSIS OF COPPER suited to their conditions. A description of the system, by H. Koch, has been necessarily condensed in translation. The sampling of the hoisted ore (including the mineral typolite and shales from the quarries) is carried out at both the shafts and the smelter, so that the same smelting stock is examined twice. The necessity for this procedure arises from the exceedingly variable nature of the slates. Only an average result of the metallic contents of the doubled samples can be considered accurate, although the metal can be closely determined in each one as received. The samples at the shafts are taken by small scoops or troughs out of the mine cars or cable-buckets, while they are being filled, and in such proportion that the total weight of sample amounts to nearly ^iir of the ore hoisted. The material, keeping that from the quarries and mines separate, is united to form daily catch samples from each shaft, stamped or crushed in machines, pulverized, sifted, well mixed, reduced to about one hundred grams, and then examined by color test to assign its grade as to copper (assay weight, 2 g.) . From the same samples, one gram is taken (for each metric ton loaded), to form a weekly catch sample, which again is re- duced, mixed and graded as to copper contents by the colori- metric test; then delivered to the central laboratory, where the accurate determination of the copper and silver is completed by electrolysis and cupellation. 4. At the Smelting Works, samples are taken during the dis- charge of the ore cars from the mines, and are collected sepa- rately from each shaft or quarry to constitute daily quarry and mine samples, which are ground down as at the mines and tested colorimetrically. For each ton of ore received one gram is re- served for a monthly sample, and this material, after proper mix- ing and reduction, is tested by color test at the mines, and by gravimetric analysis at the smelter. The values obtained from the monthly average samples taken at the smelter are compared with the monthly averages com- puted from the weekly average test samples taken at the shafts. The arithmetical mean of the mine and smelter samples is. taken as the basis of the office reports. To judge of the (probable) values, frequent tests are also made of gossan, and surface strata, and assayed partly by color at the mine, partly for copper and silver at the main laboratory. SAMPLING AND CRUSHING 17 Raw Ore from Smelting Ore-beds. When a stream of ore is cut out of the bedding-pile, 9 to J.O spoon samples are taken (5 to 6 from each cut of 7 metric ifons), and collected as one sample until 25 tons are smelted. This Average unit sample (reduced to about 200 g.) goes to the laboratory for the valuation of the copper and silver contents. From the finished samples, a monthly average sample is made up for complete analysis. DIVISION 2 MECHANICAL SAMPLING 5. The daily control of the large western mines is effected by diverting a fixed proportion of the daily product of the mine (perhaps one-tenth of the cars if the ore is uniform) to a sampling mill, located at the smelter, or at the mill if the ore requires concentration. , At the Anaconda Mines, Montana, the sample lots of ore are elevated to bins at the top of the sampling building, and then passed down through an alternating succession of rolls, crushers, and four Vezin samplers until the ore is cut to a portion of 3.2 pounds for each ton of the original sample. ixixx = BTS and sib X 2000 = 3.2. The new five-story mill of the Calumet & Arizona will reduce a ton by four Snyder machines to 1.6 or 3.2 pounds, according to the nature of the ore sampled. The product of the mechanical samplers is further reduced on an iron floor by Brunton shovels. The Garfield smelter is said to reduce the ore by causing the sample to be heaped into cones which are flattened and divided by iron crosses against which the material is shoveled. These crosses are made, in several sizes, from lOx-rV inch iron plates of suitable lengths, set vertically. (A short discussion of the requirements of a good sampler, and of Eastern practice, is given in 13, Division 4.) When the sample is reduced to 100 or even 25 pounds (vary- ing with the locality) the ore is dried on a steam drier to deter- mine the moisture, and the dry material put through an Engelhardt sample grinder, divided by riffles, bucked to pass a 100-mesh sieve (100 to the linear inch), and put in four packages of 6 to 12 ounces each. One goes to the laboratory, one to the owner of the ore, and two are filed away in case of dispute. In some works the product of the last machine, reduced to 100 or even 25 pounds, 18 ANALYSIS OF COPPER goes direct to the laboratory for all the further reduction and mois- ture test. There it is ground, quartered, and riffled down to about 50 ounces by a fixed rule, which must be the result of experiment with different classes of ores. The rule actually depends on the principle, that during the reduction of the ore by grinding, sift- ing, and quartering, there is in each successive product a fixed ratio of the largest particles of valuable metal, or sulphides, to the whole weight of each size which must not be exceeded or the final error will be too great, i.e., quartering must not be carried too far, before grinding finer. The separation of the fine and coarse particles in transit through the modern mechanical sam- pling mill is taken care of satisfactorily by reducing the length of the spouts between cutting devices, and using a shaking feed for all crushers and rolls. Any metallics on the sieves are weighed along with the ore portion from which they were taken and sep- arately assayed, unless they can be ground in with the ore by passing through the grinder a second time. IRON FROM GRINDING PLATES 6. Iron is introduced into all hard ores, when finely pulver- ized by steel plates. In furnace slags, the iron derived from wear may amount to 0.3 per cent. To make a correction factor, pass some of each class of material through crusher, and reduce in large porcelain mortar, and agate, or Abbe ball mill. Compare the iron in this sample with the other. DIVISION 3 SAMPLING IN STAMP MILLS 7. Concentrates in Western mills are passed through me- chanical samplers. The following description by C. H. Benedict, of the sampling in the Calumet & Hecla Stamp Mills, illustrates a correct principle, adapted to the study and control of milling operations in any district. There is no custom milling, as generally accepted, on Lake Superior. A mine without a mill leases the exclusive use of one or more stamp heads and owns the product. In this way, there is no necessity of sampling mills, if, in fact, the nature of the rock made such sampling possible. The quality of the rock is found by adding the assay value of the concentrates to that of the tailings. SAMPLING AND CRUSHING 19 In order to test out a given milling process, however, tests are frequently made by taking samples of all products for a week on amygdaloid (or if on conglomerate rock, about four days), Discharge Fig. 11. Sampler for Mill Tailings. in order to get the variations in stamping. During such tests, samples are ordinarily taken of each product for one second every hour. For the overflow of jigs, the sampler makes use of a long shallow iron trough, " U " shaped, and a little longer than the width of the jig, holding the trough cross- wise under the whole stream 20 ANALYSIS OF COPPER for the time specified. "From these time samples, the calculated tonnage must, of course, check closely with the car weights of the rock, originally stamped." Concentrates are weighed in bulk and tested in one of two ways by means of a slotted sampling pipe or by the "spoon and shovel." In the first case, a f-inch (1.9 cm.) pipe (with a slot cut out in the side and one edge faced out to make a projecting lip) is driven into the car of wet mineral, the pipe turned, and the resulting sample shaken out into long pans. In the second method, a portion is taken by a spoon, or similar device, from a shovelful, at regular intervals. Mid- dlings are sampled in the same way as the tailings. 8. Tailings of the Calumet & Hecla Mining Co. are all ele- vated by sand wheels and leave the mill in a launder 54 inches (1.37 m.) wide, and with a stream about 8 inches deep. A sample from this launder is taken every hour and a half, day and night, by means of the slotted sampling machine shown in Fig. 11. This sampler consists essentially of a fan-shaped receptacle with a slotted end, which is carried through the stream by being suspended on a track over the stream (and is afterward hoisted by the spindle, in order to discharge it over the edge of the launder into a metal barrel). The samples from each shift are then com- bined, and should be dried in large tarred pans over steam coils, then screened through a succession of sieves (of 10, 20, 40, and 60 meshes to the linear inch). The resulting fine products are weighed, sampled, and the calculated assay from the sizings checked up against the original assay. Laboratory reduction of these samples of tailings, or mineral, is accomplished ordinarily by permitting the dried sample, if large, to flow through a hopper, and cutting the resulting stream of dry pulp with a rectangular pan, moving in time with a pen- dulum. If the sample is small it is poured through a funnel, held in the hand over cut-out riffles. The size of the sample for the assayer is regulated by the coarseness of the material, and is usually about 40 grams. This is crushed through a 60-mesh sieve, the resulting metallics picked out to constitute a separate sample, and the fines reduced to portions of 5 grams for electrolytic assay. The combined assay values of tailings from the different machines must check with the assay of the general tailings from the main waste launder. SAMPLING AND CRUSHING 21 DIVISION 4 SAMPLING AT REFINERIES * 9. Mill concentrates of native copper are sampled in the refinery from hand carts; or from chutes, when weighed out to the furnaces, taking say two small cups for each thousand pounds of the wet product. The daily samples from each grade, and from each mine, are kept in water-tight copper cans, inclosed in a water-tight wooden box, until the percentage of water and copper can be determined. At Custom Sampling Works, New York (By A. M. Smoot). There should be no invariable rules for the sampling of copper ores and mattes, because, while the grade and physical condition are the principal factors in determining the procedure, varying conditions of sale and delivery must also be considered. 10. Hand Sampling. Low-grade copper ores, containing only small amounts of gold and silver, may be conveniently tested in 100-ton lots. If the ore shows no pieces larger than 4 inches (10 cm.) diameter, take every tenth barrowful as it is discharged from cars, or every tenth bucketful as it is taken from vessels, and set it aside for a sample. Crush the 10 per cent sample to 2-inch size, or 5 cm. diameter, and by half shoveling, reduce it to five tons (or 4536 kg.). Crush this to 1 inch (2.5 cm.) diameter, and reduce it again by half shov- eling to one and a quarter tons. Crush to | inch and reduce by shovel to 625 pounds (284 kilos). Crush to \ inch (or 4.2 mm.) diameter, and reduce by half shoveling, or preferably by riffling, to about 80 pounds (36 kilos), crush to pass a sieve of 16 meshes to the linear inch and reduce by riffling to about five pounds (2.3 kilos). Reduce this to pass an 80- or 100-mesh sieve, mix well, and divide by riffling into the required number of sample packages. Quartering should be avoided, since it may easily lead to errors which will tend constantly in one direction. It is more laborious and costly than half shoveling. 11. Rich Ores, containing appreciable gold and silver values, as well as high copper, should be divided into fifty or even twenty- five ton lots (22.7 metric tons) for sampling, unless they are very uniform. If possible, all rich ores should be crushed to 2 inches (5 cm.) diameter before any division is attempted. Take every fifth shovelful as a sample, crush to f inch (1.9 cm.), and reduce by half shoveling to 2J tons, crush to \ inch (6.3 mm.), and reduce to 22 ANALYSIS OF COPPER five hundred pounds, crush to pass eight meshes to the linear inch and reduce to fifty pounds; then crush to pass 20 mesh and reduce to five pounds (2.27 kilos). Grind this through a 100-mesh sieve (39 meshes per cm.), mix well, and divide into the required number of assay samples. 12. Mattes. These are sampled like rich ores, but some material with high gold and silver values will require finer crush- ing throughout and more intimate mixing between the reductions. See note on western methods. Duplicate samples of such rich mattes should be taken from the very beginning of the operations. Duplicates, which are made by dividing a single sample at some point toward the end of the coarse crushing, only serve to check the final work. Dupli- cates are especially desirable with very variable material; for instance, where two or three small lots of widely different char- acter are to be sampled as one lot. In such cases as this all of the material should be crushed to at least f inch (1.9 cm.), and mixed by coning before any subdivision is attempted. Fine crushing throughout. It is the practice at some eastern works to crush the whole ten or twenty per cent sample from the original lot to J inch (6.3 mm.), and then by repeated mixing and half-shoveling to reduce it at once, without further crushing, to 250 or 300 pounds. This is done on the ground that it is cheaper to crush the whole sample than to introduce several intermediate samples of coarser materials. The crushing of the whole lot is insisted upon by one or two western plants, when a car is loaded with a mixture of reverbera- tory and cupola furnace mattes. Although a satisfactory sample is obtained, the production of a large amount of fines is often objectionable to the smelter. If care is used throughout in hand sampling, the results are accurate, but treatment of large lots is too slow and costly. NOTE. Western mechanical practice has been outlined under Division 2. 13. Machine Sampling. For large lots of uniform material, machines are more economical and practicable, as stated in the ac- count of the systems of the western reduction works. The obser- vation of several important points in construction and operation (which were not mentioned under mill practice) is essential to good SAMPLING AND CRUSHING 23 work. The sampler should be constructed so that the scoops or diverters cut the whole ore stream at frequent intervals. The scoops should permit a free fall to. the stream after it is cut out, so that it shall not be retarded and so no pieces shall bounce outside. The openings in the scoops or diverters should be at right angles to the ore stream, and they should move across it with uniform speed, and at uniform intervals. The openings should be at least four times the diameter of the largest pieces which pass through them. In machines such as the Vezin, in which the sample scoops revolve around a shaft, the shape of the scoops should be sectors of a circle whose center is the center of the shaft. The feed to the machine and the flow of the ore should be constant and uni- form. The whole apparatus should be housed to prevent the escape of dust, and at the same time all parts should be easily accessible. As in western mills, the final product of the machine sampler passes to a closed box, or car, from which it is taken to a sampling floor and finished by hand. Ordinarily 50 tons would be cut by three machines to 800 pounds. The last machine sample would then be cut with a Jones riffle to 400 pounds, crushed to 12 mesh, and riffled to 50 pounds. Finally, crush to 20 mesh and riffle to 5 pounds, grind this to 100-mesh sieve, mix and divide it into the required number of samples. 14. Moisture Samples. This question is very important, especially in ores. Since all ores and mattes are assayed on the dry basis and settlement is based on the dry weight, moisture samples should be taken at the time of weighing. Weighing is generally done immediately before sampling. For a moisture sample, at least 10 kilograms of ore should be taken after crush- ing to at least one-half inch, or preferably finer, but too much crushing should be avoided. The reject from the third mechani- cal crushing of the general sample usually affords a good moisture sample. Small moisture samples are useless. DIVISION 5 FURNACE PRODUCTS 15. Reverberatory Slags. Waste ore slags from matting furnaces can be taken at the door, or sluice, about once in 2 hours, or when skimming. Mineral slags from native copper contain 15-20 per cent of copper as silicate and shot metal and the last re- fining slags are full of carbon, so that the best average is ob- 24 ANALYSIS OF COPPER tained by breaking all, or a half at least, of the cooled cakes from the slag buggies and taking pieces of equal size from the top, middle, and bottom of each block broken. 16. Matte in Western Works. Small hand-ladles can be filled at the settler or hot metal ladle while tapping, or pieces can be taken by shoveling from the cool material, after crushing, if it is to be bagged for direct shipment. As already noted, some experts advise a preliminary fine crushing of the whole, in the case of a car of mixed reverberatory and cupola matte. 17. Cupola Waste Slags. The method, sometimes adopted, of dipping an iron rod in the slag stream, and chilling it in water, does not show the average amount of shots of metal, or prills of matte. To secure an average sample, a small hand-ladle should be held under the full slag stream, and the contents quickly granulated with water in an iron pail, kept in a safe place. At Lake Superior, samples are taken once an hour and combined into one sample for each working shift. 18. Anode Slags (H. D. Greenwood). The small slag pots are dumped and the shells cooled with water, which will help to disintegrate the large pieces. The large pieces are then broken up with hammers and sent through the crushers, where they are crushed to one-half, or one inch, pieces, all me tallies being returned to the furnaces. The sample is coned and worked down by hand to about 200 pounds (or 90.7 kilos) ; then crushed finer and worked down to a laboratory sample, recrushing finer with each division. The metallics of each quartering are kept separate (as in the case of Lake Superior slags), and sampled, assayed, and proportioned accordingly. 19. Sampling of Molten Copper. Edward Keller and others have proved that the difficulty of sampling pigs and other irregular deep castings at anode furnaces is eliminated, if thin square plates are cast with a full hot ladle directly from the molten charge, after it is well mixed and ready to pour into molds. Keller l recom- mended that a plate, 15 inches square by one inch in thickness, should be taken from the middle of each third of the charge as it was tapped from the furnace. With refined native copper, it has been proved that plates 6 x 6 x \ inch are large enough. A uniform number of holes should be drilled through each plate, but the sampler should keep away from the edges for a distance 1 Trans. A. I. M. E. 27 (1897), 106. SAMPLING AND CRUSHING 25 equal to twice the thickness of the plate, at least. A granulated shot sample, or one of the thin plates, may be correct for gold and silver if all skulling is avoided by pouring from a hot clean ladle. Keller insists that large snot samples should be re-melted, if variable, and a sample cast or granulated. The copper per- centage, when determined upon the shot, cannot be as good an average of a lot as a result obtained from the plates, owing to the absence of some surf ace oxides which should be included in this case. Superintendent William Wraith, of the Washoe Smelter, con- cluded, after many tests, that the pouring of shot from a hand- ladle tends to give high results for silver. His statement follows : l "The only method of furnace sampling which uniformly checks results obtained by drilling every fourth anode by a 99-hole template system con- sists in batting out samples from the molten stream of metal, while pouring, and allowing the copper to fall into water. "The first sample is batted from the stream with a wooden paddle, 30 minutes after starting to pour, three other samples being taken at one- hour intervals, each portion weighing from 4 to 6 ounces. These samples are dried, examined for burnt wood, screened on a 10-mesh screen of No. 8 wire, to remove fines, the oversize then screened on a 4-mesh screen of No. 20 wire, to remove the coarse, the undersize of this screen being taken as the sample. The four portions are thoroughly mixed, and split in half by passing over a 16-slot splitting device, slots being 0.5 inch wide, one-half being kept as a reserve sample, and one-half assayed." SLABS AND ANODES 20. (From A. M. Smoot, F. Andrews, and H. D. Greenwood.) The method of A. M. Smoot, for the custom sampling of pig copper and anodes, is presented first, as it gives a full explanation of the reasons for every step taken by the large refineries in this rather difficult problem. Nearly all custom refineries adopt the same principle. The template system of sampling takes into account the segrega- tion of the constituents which occur when any molten alloy is cast into a mold and chilled. The word "bar," as used here, includes any commercial shapes into which crude copper is cast for shipment. By this system, the top and bottom surfaces of any given bar, cake, or slab are divided into a number of squares, each of which is a bounding surface either of a parallelopipedon or a wedge. By drilling holes in the centers of these squares, 1 Trans. A. I. M. E. 41 (1910), 318. 26 ANALYSIS OF COPPER clear through the bars, borings representing the contents of the solid figures are obtained. By making the squares small and drilling a large number of holes in rotation, an accurate sample representing the metallic contents of any given bar may be ob- tained. An accurate sample of the lot may be obtained by tak- ing a large number of bars of the same shape and character, and drilling successive holes, one in each bar, each corresponding in position to one of a series of squares marked off on the top of a single bar, that is, by advancing the drill one hole in the template with each successive bar. The number of bars taken to represent a lot should depend on the number of bars in the lot and on the character and tenor of the crude copper. In ship- ments made up of several converter charges of varying gold and silver contents, all of the bars should be drilled. When the material is of nearly even composition, such as the product of a smelter handling uniform ore from a single mine or district, a smaller proportion (for instance one-fourth or even one-tenth) of the bars will represent the lot. In sampling copper which is high in gold and silver, even when it is of fairly uniform grade, every bar should be drilled. When a fraction of the num- ber of bars in a lot is taken, the number should correspond with the number of holes in the template or to some even multiple of this number, so that one or more complete cycles shall be used and every hole in the template be represented in the sample. This is not important where a number of successive lots of the same material are sampled. In such cases, the rotation of drill holes may overlap from one lot to the next. (Refer to Andre ws's account.) Since segregation on cooling takes place quite uniformly from or towards every cooling surface, a quarter or half a bar may be assumed to contain all the elements of segregation and a quarter or half section template may be used. This is advanta- geous because a larger number of small squares may be laid out on the smaller templates, and still keep the same number of holes within the number of bars to be drilled. In drilling bars with a flange or wedge, the template should be laid out so that the weight of borings from the wedge-shaped part of the bar is in the same ratio to that taken from the rectangular part as the weight of the wedge portion is to the weight of the rectangular portion. All holes are usually drilled with a half-inch (1.27 cm.) drill, and must SAMPLING AND CRUSHING 27 extend clear through the bar. Too much emphasis cannot be laid on this point. The drill should be driven at high speed with a light feed, so that the drillings may be thin and easily ground. Forcing the drill produces thick drillings. The top surface of crude copper is usually very rough from the escape of gases, especially in the central portion. The top surfaces at the edges are usually fairly smooth. The rough top surface frequently contains undecomposed matte and sometimes small pieces of slag, as also an excess of cuprous oxide. As cuprous oxide is a solvent for silver, the top skin of crude copper bars is frequently higher in silver than the underlying metal. It is, of course, poorer in copper. In drilling bars with the top surface up or towards the drill, some of the smaller particles of the borings become lost in the interstices of the rough surface; they cannot be completely collected. The smaller particles consist in part of the brittle top surface; thus drillings from rough bars made with the top surface up are not an average sample. When bars are drilled with the bottom surface up, the pressure of the drill is apt to break off rather large pieces, forming craters in the brittle crust when the drill is thrust through the bar. It is, therefore, customary to drill half of the sample bars with the tops of the bars up and half in the reverse position. The' drill holes in the wedge must neces- sarily be taken with the top of the bar uppermost. Shipment of Samples. Containers for powdered ore or drill- ings of crude copper usually consist of sacks, stout manila paper bags, or printed envelopes with patent fasteners. Drillings of refined copper should, however, always be kept in tight glass- stoppered bottles and shipped in capped bottles in special mailing cases to provide absolute protection from oxidation in transit. 21. Reduction to Assay Sample. The weight of drillings from the usual shipping lots may be twenty to thirty pounds (9 to 14 kilos). All drillings must be ground and mixed before division. The grinding is done in a mill with slightly corrugated or toothed plates. The movable plate revolves horizontally against the fixed plate. They may be of good cast iron, but chrome steel is better, although it is difficult to get good castings of this material and inequalities cannot be adjusted by machining. All the drillings should be ground to pass a screen with 8 meshes to the linear inch, and then thoroughly mixed and divided 28 ANALYSIS OF COPPER on a riffle, or split sampler, to obtain a sample of seven or eight pounds (3.2 to 3.7 kilos). This sample should be ground re- peatedly, until all particles pass at least a 16-mesh screen. A fineness of 20 mesh is preferable and drillings may easily be ground to this fineness if they are thin enough originally. The finely ground drillings should be thoroughly mixed and divided with a split sampler or riffle into the required number of packages. No attempt should be made to dip with a spatula, or to quarter the ground borings, but the greatest care must be used to include the proper proportion of fine and coarse in each package. If much gold and silver are present and there is a large differ- ence between the assay values of the fine and coarse parts of the ground sample, the whole of the 16- or 20-mesh sample should be weighed and separated into fines and coarse on a screen of 40 meshes to the linear inch. The fines should be weighed and the finished samples should include separate parcels of the coarse and the fine parts together with a memorandum of weights and their mutual ratio. 22. Method of F. Andrews. The principle given is the same as the one first quoted. This system provides for material of variable composition. From 20 to 100 per cent of the lot is sampled by drilling a J- inch hole through each sample piece, and in rotation as already directed. The templates are of the exact size of the surface to be covered, and each brand of material has, of course, its own inde- pendent template. The drillings are ground twice and then cut down by a divider to four or five pounds. This amount is then further ground until it all passes through a 16-mesh sieve. The whole sample is then put through a 40-mesh sieve, the coarse and fines weighed, and the required number of samples put up with proportionate parts of coarse and fines in separate bags. Each sample then comes to the laboratory composed of two parts, one (coarse) which has remained on the 40-mesh sieve and the other (fine) which has passed through. In weighing up for the gold-silver assay, the assay ton is made up of proportionate parts of coarse and fines, but the copper assay is made on each part separately, 5-gram portions being used and the correct assay figured from the proper weights. 23. Method of F. D. Greenwood (for uniform material). It is recommended that the ground borings should be directly cut SAMPLING AND CRUSHING 29 down on a split sampler in such a way as to obtain a sample of about 1 assay ton (29.167 grams) for the gold or silver assay. For the copper assay of crude oullion, 80 grams are taken, and in each case the final sample must Include the proper proportion of the finer and coarser parts of the borings. This sample must be very carefully obtained for reasons already stated. Portions "dipped" from the sample bottle or from the sample spread out on paper, are likely to contain undue amounts of coarse or fine. 24. Metallic Iron in Drillings (A. M. Smoot). There is a very small amount of iron introduced in the drilling and grinding of converter and blister copper, but it is wrong to attempt to remove it from the ground borings by a magnet, because crude copper always contains magnetic particles, due to matte, etc., which properly belong to the sample. Any attempt to remove such iron with a magnet will introduce a larger error than it will cure. Practically, the amount of iron derived from the tools is negligible, since ground turnings seldom contain more than .03 per cent, and a large part of this is present in the original copper. Iron due to the tools may be separated in part by treating the original turnings with a magnet and carefully saving the magnetic particles. After grinding in the mill and dividing by riffle to the amount required for the sample packages, go over the drillings again with a magnet and discard any magnetic particles obtained. Restore to the sample a proportionate part of the magnetic par- ticles originally found in the whole sample and mix thoroughly. 25. Moisture in Converter and Blister Copper. The rough surface of such metal contains many cavities which may retain appreciable moisture if the copper has been exposed to the weather. This is apt to occur in winter when bars have been stored in yards exposed to snow, or shipped without protective covering. The per cent of superficial moisture is, of course, very small, and the average amount may be easily ascertained by moderate drying. Bosh-cooled crude copper nearly always contains " occluded" water held in large cavities under the rough surfaces which could not be affected by weather. The " occluded moisture" may in some cases amount to 0.1 per cent or even 0.2 per cent by weight. This moisture is difficult to remove. Long continued drying at a temperature of about 350 F. is necessary. Of course, such 30 ANALYSIS OF ^COPPER moisture is unevenly distributed, and a large number of bars must be dried to secure even reasonable accuracy. This necessitates a special drying chamber. SAMPLING OF REFINED WIRE-BARS, CAKES, AND INGOTS 26. The method specified by the American Society for Testing Materials in the year 1913, for drilling cast metal, involves the driving of several holes clear through the casting in three direc- tions, from top, side, and bottom, after removing and rejecting the surface oxide from the space drilled. F. D. Greenwood, and others, drill the holes J inch to \ inch deep, rejecting the drillings, then drill in the same holes until the drill is within J inch from the bottom. The drillings are cut up with snips and only clean bright drillings are accepted. The author drills two holes, at least, in the top, side, and bottom, with the precautions specified, but only halfway through, which gives practically the same sample but causes less heating and danger of oxidation. All borings must be tested with a magnet and freed from dirt by sifting through a 40-mesh sieve, then placed in clean dry bottles, as they oxidize rapidly if stored, or shipped, in envelopes. Flat anodes or sample plates are drilled through, including the oxide, as indicated under "The sampling of molten copper at furnaces." SAMPLING OF COAL AT SMELTING WORKS A special system is recommended for the sampling of cars and cargoes. The methods of subsequent reduction are those proposed by a joint committee of the American Society for Test- ing Materials and the American Chemical Society, and published jointly during 1914. 27. Shipment of Samples. If samples are shipped from a distance, much moisture will be lost unless the containers are sealed. When the moisture content is important, the sample should be broken to half -inch size as accumulated, the mixing done quickly, and the sample transferred to a glass fruit jar or tin with screw cap, which may be made air-tight by sealing with adhesive rubber tape and gaskets. Three pounds is a usual sample for long-distance shipment. 28. Car Sampling. The 1899 Coal Committee l recom- 1 /. Am. Chcm. Soc. (1899), 1116. Sec also "Coal," by Sommermeier. SAMPLING AND CRUSHING 31 mended that the sampler should take a sample of six scoop shovels at regular intervals on each side of the car. ' The shovel should be brought out full. Spread ou a tight floor and break all lumps larger than an orange. Shovel, quarter, break finer (in a power crusher if possible), ancf quarter or riffle until the sample is re- duced to 3 pounds. Manipulate quickly and transfer to a large fruit jar. A carload of such material as lignite may lose several hundred pounds in transit, and the U. S. Geological Survey has proved that during a 150-mile haul, there is a decided tendency for slate to settle. In such a case, sample the whole face of the load at the middle and ends of the car, while unloading. According to A. D. Little, the ratio of the largest pieces of coal to the total weight of sample at each stage of the reduction process should be less than 0.01 per cent, if errors of 1 per cent in the ash determination are not to be exceeded in a single sampling of a fuel containing 5 per cent of ash. N. W. Lord insists that this ratio should be that of the maximum slate sizes to the total sample, which is a good proviso for very low-grade fuel. 1 29. Vessel Cargoes. To sample coal on a very large scale, as unloaded by several power hoists, the system adopted by the Calumet & Hecla Mining Co. is recommended. A large covered barrel of corrugated galvanized iron is placed beside the scales on the elevated platform of each hoist. The buckets of coal are dumped on a grizzly, or coarse screen, from which the clean lump coal flows to the gravity distributing car. The weigher grabs a piece of coal (without selection) from one car out of five, or in such proportion that the total sample shall be about one part in five thousand. The fine coal slides beneath the hoist down a closed chute, from which it is drawn into railroad cars and sent to boilers for immediate consumption. The fines are separately weighed, sampled, and analyzed. The analysis of the whole cargo is then calculated from the tests of the lump coal, A, and the fines, B, according to their actual relative weights. By storing only clean lump coal, the danger from spontaneous combustion is minimized, and a better knowledge of the composition is obtained than by any attempt to estimate the proportions of coarse and fines. If such an elaborate sj^stem is out of the question, a small propor- tion of the total load may be dumped separately and shoveled 1 Bailey and Brady, /. Ind. and Eng. Chem. 1, 161, 263, 316. Ibid. 6, 517. 32 ANALYSIS OF COPPER ' over an inclined screen, and the parts separately weighed, to obtain a practical estimate. 30. Reduction of Lump Coal for Assay. The car, or cargo, sample is run through a large jaw crusher to 1.5-inch size, and rapidly reduced by coning and half -shoveling to 50 or 75 pounds. This is now to be reduced to a 2- to 5-pound sample for assay. If the laboratory is near at hand, the whole sample is transferred in iron pails, and, if very moist, is broken to J-inch size and dried at a low heat in iron pans (moisture 1). If fairly dry, the sample is broken down at once to pass a sieve of 4 meshes to the linear inch and quartered to 3 or 5 pounds before drying. Samples are preserved in tight fruit jars with screw caps, as already described, if the preliminary reduction must be made at some distance from laboratories. If a sample of 3 to 5 pounds is dried at 15 C. above room temperature, Appalachian bituminous coal and anthracite will be air-dry if placed in the drier in circulating air over night. Illinois coals may require 48 hours and lignites 72 hours for the preliminary air-drying. 31. Reduction to Assay Sample. Immediately after the last weighing, the entire dried portion should be rapidly pulverized to 10-mesh size; mixed and reduced to 450-500 grams with an inclosed riffle sampler having J-inch divisions. If reduction is carried directly to 60-mesh fineness, there will be loss of moisture. With ordinary open grinders, the author finds it most accurate to grind quickly through a 20-mesh sieve, then dry over steam plate, or in an oven below 100 C., for 1| hours (giving moisture per cent No. 2) . 100 grams is afterward reduced to pass a sieve of 60 meshes to the linear inch (25 per cm.). The Official Committees (35) recommend that the 500-gram sample be ground directly to 60 mesh in a closed porcelain Abbe ball mill at 60 revolutions per minute. Bituminous coals require 1 hour's time and anthracites 2 hours'. The jar should contain about one-third its volume of 1-inch well-rounded pebbles. The best machine for the grinding of 3- to 5-pound samples to 10 or 20 mesh is an inclosed coffee mill or a Hance Bros. & White drug mill with corrugated plates, and fitted with pulleys. In the analysis of the fines another moisture test is often made by drying one hour at 100-110 C. Neither the grinding, or drying, of coals should be carried too far, and uniform conditions must be maintained to obtain satisfactory results. The latest modifica- SAMPLING AND CRUSHING 33 tions of standard methods of analysis are found in the reports of coal committees; for which see reference (35) at the close of this chapter. > GOLD AND SILVER BULLION 32. Method of U. S. Mint. The following description is based on a paper of F. P. Dewey and the practice of refinery chemists, who follow, quite closely, the Mint method. In the purchase of bullion by the Mint the size of the deposit has an important bearing on the sampling. When the weight of a deposit reaches 300 ounces Troy, the samples become important, and with bars weighing 700-1200 ounces, correct work is essential. High-grade bullion and coin gold do not segregate, but when we come to consider bullion of more complex composition, the matter assumes greater importance. It is safe to assume that a " brittle" bar of gold bullion will not be homogeneous. According to Dewey, there is only one satisfactory method of sampling of general gold bullions. (a) "Dip" samples are taken by pouring a small portion of the well-mixed molten metal into water so as to produce globules or granulations. Granulations are sometimes made by pouring directly out of the black-lead crucible into water, the operation of casting being interrupted for the purpose. A good sample of silver bullion is obtained from a silver refining furnace by taking a small "dip sample" directly from the metal bath after every second (1000-ounce) bar cast. The granulations may then be mixed and reduced in size by shears, if necessary, or remelted. (6) "Chips." There are various cases where a bar of solid bullion can be satisfactorily tested without melting by cutting a chip with a chisel, preferably of special design. Power-driven punches may be used and machines are also in use which take out a triangular piece of metal by means of a projection on a lever operated by a cam. The chips must be taken from a corner or along the edge of a bar. In systematic sampling of large bars, two chips are generally cut, one from the top and one from the bottom of the bar, and are properly identified. (c) Drill samples are taken according to fixed system, but, in large bars, there is a wide choice in the location of holes. A common practice in the Mint service is to drill halfway through at diagonally opposite corners of the top and unite the drillings for the top sample. The remaining corners are drilled halfway 34 ANALYSIS OF COPPER through from the bottom, and the drillings mixed for the bottom assay. Occasionally, with large bars, the four drillings are kept separate, and sometimes holes are drilled near the center of the bars, also. Drill samples are often better than chips, especially where large, fairly uniform bars are sampled by a well-designed plan. Drill samples of brittle bars are, however, liable to be inaccurate because of difference between the coarse and fine parts, although on high-grade gold the difference may be as low as .0001 between the top and bottom. F. P. Dewey concludes that, in sampling bars of gold weighing over 300 ounces, when the assayer is acquainted with the bullion, he may accept either a chip or drill sample. On an unknown bul- lion, it is unsafe to accept anything but a properly prepared dip sample. Some metal, such as " cyanide" bullion from the cyanide process, must be refined before the gold can be accurately determined. 33. Special Methods of Refineries. An easy practical way of sampling silver bullion at furnaces, while casting, has been described under the title of "dip samples." Dr. E. Keller, in discussion of the Mint system, suggested that the thin plate method used successfully at anode furnaces for copper, might also be applied to silver bullion. H. D. Greenwood uses such a method, casting Dore bullion into (18 x 7 x f inch) plates, which are drilled with sVirich holes on the checker-board plan. The drillings are then ground to pass 30 mesh, or a sieve of 12 holes per cm. Any iron is removed from samples with an electro-magnet. STANDARD METHOD FOR ZINC SPELTER 34. Virgin Spelter, that is, spelter made from ore or raw concentrates by process of reduction and distillation and not produced from reworked metal, is considered in four grades by the American Society for Testing Materials (36), and by the sub- committee on alloys of the American Chemical Society (37). For the methods of analysis and the composition of each grade, refer to methods 31-35, Chapter XIV. With the exception of two or three minor details, the same procedure is followed by each Committee in the sampling of zinc spelter. SAMPLING AND CRUSHING 35 Ten slabs are taken at random from each carload, or lot, received. The first authority permits a smaller sample to be taken for smaller lots, but in rio case less than three slabs. In case of dispute, half of the sample is to be taken by the maker and half by the purchaser; and the whole shall then be mixed. The slabs taken as a sample are to be sawed completely across and the sawdust used as a sample. In case no saw is available, the slabs should be drilled completely through and the drillings cut up into short lengths. The Committee of the American Chemical Society recommend that three 9-mm. holes should be drilled along one diagonal of each slab. One hole should be drilled as nearly as possible at the middle and the others halfway from the middle toward each end of the diagonal line. No lubricant is permitted in either sawing or drilling. The saw or drill must be thoroughly cleaned, and the sawdust or drilling must be carefully treated with a magnet to remove any particles of iron derived from the tools. 36 ANALYSIS OF COPPER LITERATURE OF SAMPLING Name Subject Reference 1. D. W. Brunton .... Mine sampling Trans. A. I. M. E. 25, 826. 2. (a) S. A. Reed . . Theory of sampling Ibid. 40, 567. School of M Quarterly 1885 (6) E. Keller Pig copper sampling Trans A I M E 27, 106 (c) " " Mattes. . Mineral Industry 10, 242 (d) " " Mathematics of Eng. and Min J 93, 703 (e) D. M. Liddell . . (/) A. M. Smoot. . . (I U (( Ibid. 90, 897. Ibid. 93, 1213. 3. Edmund Kirby .... Mine sampling Ibid. 69, 196-247. 4. J. A. Church Machines Ibid 86, 291 338 5. T. A. Rickard Mines Ibid. 76, 213-305. 6. Philip Argall Mines Ibid 76, 888 7. G. D. Bancroft .... Mines. . . Ibid 75, 323 8. E. E. White Drill holes Proc. Lake Sup Min Inst 16, 9. W.J.Adams.. 10. Wm. Wraith Ore shipments JVlolten copper anodes 100. MiningSci. Press, 1904, Aug. 6. Trans A I M E 41 318 11. E.Keller 12. E. A. Hersam (i it u Sizes of Screens Ibid. 42, 905. Ibid 37, 265 13. R. H. Richards 14. Whitehead & Ulke . 15. F. P. Dewey. . . Screen standards Gold bullion Gold bullion at Mints Ore Dressing, 3, 1103. Eng. & Min. J. 66k, 189. Trans. A. I M E 43, 870 16. John Jewett 17. Bailey & Brady 18. G. C. Stone Smelter deductions . . . Accuracy with coal . . . 1075. Eng. & Min. J. 73, 217. J. Ind. & Eng. Chem. 1, 161. Ibid. 1, 262. 19. E. P. Mathewson. . . 20 Sampling Mills Anaconda Eng. & Min. J. 86, 338. Ibid. 82, 624. Ibid. 82, 1165. Ibid. 81, 509. Mineral Industry, 16, 300. Ibid. 14, 164. Ibid. 10, 206 Eng. & Min. J., 90, 1059. Ibid. 82, 257. Ibid. 82, 261. Eng. & Min. J. 83, 232. Ore Dressing, 2, 844, 1570. Eng. & Min. J. 70, 549. Ibid. 71, 534. Ibid. 76, 729. Proc. Am. Soc. Test. Materi- als, XIII. Proc. Am. Soc. Test. Mat. XIV, 409. J. Ind. & Eng. Chem. 6, 517. Am. Soc. Test. Materials Year Book, 1914, 284. J. Ind. & Eng. Chem. 7, 547 (1915). Greene Cananea Cobalt, Ont. 21. F. F. Colcord 22. . Garfield, Utah U.S. Metals Co Michigan Smelter. . . . B. C. Copper Co Tooele Utah 23. .. 24. .. 25. Paul Johnson 26. Repath & McGregor 27 Copper Queen Britannia Co., B. C.. . Laboratory Machines Design of riffles Riffles & samplers .... Lab. sampler 28 29. Haultain 30. R. H. Richards 31. Snyder 32. Calkins 33. Jones " riffle sampler. . . . Lab. screens, ratio. . . . Sampling of Coal Report, 1914 Of Amer. Chem. Soc. . " Coal Analysis" Zinc Spelter Specification 34. G. A. Disbro 35. Committee, E-4, 1914 .. Joint Committee .... E. Sommermeier. . . . 36. Committee Report . 37. Sub-committee Approved Report CHAPTER III REAGENTS AND STANDARD SOLUTIONS Introduction. The following paragraphs treat of the impuri- ties in chemical reagents which may injuriously affect the accuracy of the analysis of copper. The fact that less than .0050 per cent of the usual impurities is required to affect the physical proper- ties of refined copper, together with the demand of consumers for metal of very high purity and of strictly uniform quality, make it necessary to employ for such tests the purest chemicals and to make blank analyses of such reagents. The chemist of the Wallaroo Company has found it necessary to redistil all imported English acids. All acids mentioned in succeeding chapters should be understood to be of the purest commercial grade obtainable and of the maximum strengths given below, unless dilute reagents are locally specified. Definition of Density. "Specific Gravity" is a longer term than the one of " Density" lately recommended by the U. S. Bureau of Standards. The author prefers to adopt the latter term for the sake of brevity. Density is defined as the weight in grams of a cubic centimeter, or as the ratio between the weight of a substance at a fixed temperature (15 or 20 Centigrade) and the weight of an equal volume of water measured at 4 Centigrade, its point of greatest density. COMMERCIALLY PURE REAGENTS The formulas for the most important solutions have been arranged in a regular alphabetical order and by number. Such a system is easily added to or interpolated and permits immediate reference to the formula if the numbers are also placed on the corresponding stock bottles. Hydrochloric acid, HC1 (density, 1.18 to 1.20), usually contains a trace of iron and arsenic. The product of the electrolytic process is to be preferred for all arsenic and antimony distillations as it is strictly pure. 38 ANALYSIS OF COPPER Nitric acid, HNO 3 (d., 1.42), should be free from every trace of chlorine and nitrous acid when used for electrolysis or gold assay- ing. The red-fuming acid for oxidation of sulphides should have a density of 1.60 at least. Sulphuric acid, H 2 SO 4 (d., 1.83-1.84), often contains a trace of arsenic. Some of the high-grade product shows a trace of man- ganese, which becomes noticeable in the determination of iron and nickel in refined copper. Ammonium Hydroxide, NH 4 OH (d., .90), unless of extreme purity, will show injurious amounts of iron in addition to the usual traces of pyridine. Not more than .0001 per cent of the latter should be, present in ammonia used for the separation of arsenic or antimony from copper by precipitation with excess of ferric salts. PREPARATION OF PURE GASES CARBON MONOXIDE Carbon Monoxide, CO, may be prepared by dropping strong sulphuric acid into 98 per cent formic acid. 1 The acid is placed in a small flask, and the gas must be purified from a little carbon dioxide by passing it through potassium hydroxide solution, then through alkaline pyrogallate of potassium to remove traces' of oxygen, and finally through drying tubes. To secure an even flow of gas, raise the temperature of the flask. CARBON DIOXIDE Carbon dioxide, COz, may be obtained in a very pure state by dropping strong sulphuric acid into a flask half filled with a thick paste of sodium bicarbonate and water, as recommended by Bradley and Hale. 2 This method is inconvenient for analytical purposes. Gas of sufficient purity for use in the "Determination of oxygen in copper" (Chapter XIII) is generated by the action of dilute hydrochloric acid on lumps of white marble. The traces of oxygen present are absorbed by passing the gas through a 15-cm. roll of red-hot copper gauze, contained in a 25-cm. ignition tube. To avoid rapid deterioration of the copper, the burner should not be lighted until the air has all been driven out of the bottle and tubes. The coil is easily regenerated by reversing the valves, opening a stopper in the rear of the tube, and passing hydrogen 1 E. Rupp, Chem. Ztg. 32, 983. 2 J. Am. Chem. Soc. 30, 1090. REAGENTS AND STANDARD SOLUTIONS 39 gas from a second generator for 15 minutes. The purification train of tubes may be arranged- in order thus: (a) Casamajor generator, or a small Kipp with, pressure tube above; (b) Bulb of saturated neutral ^potassium permanganate, or mercuric chloride, for absorption of hydrocarbons; (c) U-tube of silver sulphate; (d) Bo wen's potash bulb with strong sulphuric acid; (e) Tube with the roll of copper gauze; (/) Two drying tubes of calcium chloride, sulphuric acid, or phosphorous pentoxide, as preferred. HYDROGEN Hydrogen gas may be more easily purified if it is generated by pure zinc and dilute sulphuric acid (not hydrochloric acid). A. M. Smoot recommends the addition of one drop of platinic chloride solution to two liters of (1:4) acid. The purifying train consists of: (a) The generator, or gas cylinder; (6) Allihn washing bottle of 10 per cent potassium hydroxide, saturated with potassium permanganate; (c) Tubes for the removal of traces of oxygen. Here there is a choice of at least three alternatives, the first being preferred, (1) Allihn 250 c.c. bottle of potassium hydroxide solution (d., 1.27) in which is dissolved 5 grams of pyrogallic acid; (2) a heated tube of 5 per cent palladium asbestos; (3) a tube containing a long roll of reduced copper gauze heated by a short furnace; (d) Finally, two tubes of drying agents, as described under " Oxygen," Chapter XIII. SULPHUR DIOXIDE Sulphur dioxide, S0 2 , has been obtained by heating turnings of copper with sulphuric acid. Whenever a cylinder of the compressed gas is not at hand, the easiest method of generation consists in dropping a saturated solution of sodium sulphite from a separatory funnel, through a tube of soft glass, which dips under the surface of 400 to 500 c.c. of sulphuric acid, con- tained in a liter flask. NOTE. Oxygen, O 2 , simply requires a purification from traces of carbon dioxide, with the addition of drying tubes, un- less it is to be used for the combustion of carbon. In that case, a heated catalyzer tube should precede the potassium hydroxide. 40 ANALYSIS OF COPPER STANDARD SOLUTIONS The standard solutions are numbered as a means of ready reference, and are those employed in the methods described in succeeding chapters. Unless otherwise specified, they are made with the com- mercially pure reagents, sold as " chemically pure." 1. Acid Mixture, for the electrolytic assay of copper (method 2, Chapter XI). This mixture, which has been adopted by several works, permits direct electrolysis without any evapora- tion, as soon as the copper drillings are dissolved upon the steam plate and the solution diluted. It is made up in large quantities in the following proportions by volume: 7 parts of nitric acid (d., 1.42), 10 parts of sulphuric acid (d., 1.84), and 25 parts of distilled water, strictly free from chlorides. 2. Ammonium Oxalate, (NH 4 ) 2 C 2 O 4 .H 2 O. A saturated solution is used for the precipitation of calcium, or, if preferred, a 10 per cent solution of the crystals. 10 c.c. of the latter pre- cipitates about .4 gram CaO. 3. Ammonium Molybdate, (NH 4 )6Mo70 2 4.4H 2 0, is required in the titration of lead in the " Western method" for ores or slags (Chapter V). Dissolve 4.3 grams in 200 c.c. of water, add a few drops of ammonia, and make up to one liter. 1 c.c. is nearly equivalent to .005 gram of lead (Pb). To standardize, dissolve .200 gram of chemically pure lead foil in dilute nitric acid. Make the liquid alkaline with ammonia, boil for a minute, then render the solution acid with acetic acid and titrate. Use a solution of 1 gram of tannin in 300 c.c. of water as an indicator. Deduct the amount of molybdate required to affect the indicator from the total amount discharged from the burette. 4. Ammonium Molybdate is also employed as an indicator in the titration of zinc. Prepare a 1 per cent solution with distilled water. 5. Ammonium Acetate, (NH 4 C 2 H 3 O2), is used as a saturated solution in the extraction of lead from various precipitates, such as silica, or barium sulphate (Chapter V). Glacial acetic acid may be very slowly added to concentrated ammonium hydroxide until the liquid is slightly acid. For ordinary work, the solution may then be made slightly alkaline. 6. Ammonium Phosphate, Secondary, (NH 4 ) 2 HPO 4 , may be REAGENTS AND STANDARD SOLUTIONS 41 made up as a 10 per cent solution for the precipitation of mag- nesium. 1 c.c. will precipitate about .0305 gram of the oxide, MgO. The same reagent is employed in the " Phosphate method" for alumina, described in Chapter V. 7. Ammonium Sulphydrate, NH 4 HS, must be frequently pre- pared, as it does not keep unaltered many days. Saturate strong ammonia, free from pyridine, with pure hydrogen sulphide until the iron is precipitated as black sulphide and the supernatant liquid is yellow. After settling, the solution should be filtered into a brown glass-stoppered bottle. 8. Ammonium Thiocyanate, NH 4 SCN, is required as a pre- cipitant, and also as a standard volumetric solution. In method 3, Chapter IV, a solution of 40 grams per liter is adopted as a precipitant of copper in the assay of copper in ores. 9. Ammonium Thiocyanate for the titration of silver arsenate, in method 12, Chapter V, is made of such a strength that 1 c.c. equals .001 or .005 gram of arsenic. It is standardized against a weighed quantity of pure silver dissolved in nitric acid. 107.88 parts of silver = 25 parts of arsenic. 10. For method 7, Chapter IX, the thiocyanate, or sulpho- cyanate, is adopted as a finishing solution in the refinery method for titration of silver bullion, following a standard solution of hydrochloric acid (19). 1 c.c. equals .001 gram of metallic silver. 11. Ammonium Thiocyanate, as a precipitant of copper in complete analysis of the metal, is prepared by dissolving 1 pound, or 453.6 grams, of the crystals in 2 liters of distilled water. 1 c.c. will precipitate about 0.12 gram of copper in the presence of an excess of sulphur dioxide, sufficient to saturate the solution and complete the reduction of the precipitate to the white cuprous thiocyanate. 12. Barium Hydroxide, Ba(OH) 2 , is employed in the form of a saturated solution for the absorption of carbon dioxide in steel analysis, or for carbon, etc., in ores. Stir 20 grams of the hy- droxide in 1 liter of hot, freshly boiled distilled water, and after settling, filter quickly through a covered funnel into a closed flask, provided with a rubber stopper perforated for two tubes. Through one hole is passed the stem of a 100 c.c. pipette, and through the other, a short bent tube of soda lime to protect the liquid from the carbon dioxide of the air. For each analysis, 90 to 100 c.c. are pipetted into a straight 10-bulb absorption tube. 42 ANALYSIS OF COPPER 13. Barium Chloride, BaCl 2 .2H 2 O. A solution of nearly 10 per cent strength is used as a precipitant only. 14. Bismuth Sulphate. For the color method of F. B. Stone, in the analysis of refined copper (7, Chapter XIII) . Dissolve the chemically pure metal, or oxide, evaporate with a slight excess of sulphuric acid until any other acid is removed, then dilute until 1 c.c. contains .0001 gram of bismuth. In making color tests, a little sodium sulphite solution is added to decolorize any trace of ferric salt. 15. Cadmium Chloride, CdCl 2 , in an ammoniacal solution, absorbs sulphur, when evolved as hydrogen sulphide from metals. Dissolve 12.5 grams of the choride (according to J. M. Camp), in 125 c.c. of water, add 100. c.c. of ammonia (d., .90), and filter. Then dilute to 1 liter. 10 c.c. are taken for the absorption of the sulphur from 5 grams of steel, and about 5 c.c. for sulphur evolved during the reduction of copper by hydrogen (method 15, Chapter XIII). 16. Copper-Potassium Chloride is used as a solvent of steel, which permits the separation of carbon. It is prepared by dis- solving 1 pound, or 453.6 grams, of the green crystals in 1 liter of distilled water, filtering through a plug of ignited asbestos, and adding 50 c.c. of strong hydrochloric acid. For carbon in nickel alloys, a saturated neutral solution is adopted, adding acid to each analysis (16, Chapter XIV). The tops of the bottles should be protected from dust. 17. Cupric Chloride, CuCl 2 . 2H 2 O, as a solution of 300 grams of crystals in 1 liter of hydrochloric acid (d., 1.20), is combined with No. 58 to make a distilling solution for arsenic (method 10, Chapter V). 18. Ferric Ammonium Sulphate, Fe 2 (NH 4 ) 2 (S0 4 ) 4 .12 H 2 O, as a 10 per cent solution, is added to copper solutions as a precipi- tant of the arsenic group of metals. Ferrous ammonium sul- phate is much used to standardize permanganate, but the results do not always agree with metallic iron. A special solution, No. 40, is used with No. 39 for manganese. 19. Hydrochloric Acid, HC1, diluted to the " normal" solution (36.468 grams of HC1 per liter) is run against soda in alkalimetry. For method 7, Chapter IX, a solution is prepared of such a strength that 100 c.c. precipitates a few milligrams less than 2000 milligrams of silver. REAGENTS AND STANDARD SOLUTIONS 43 20. Standard Iodine. As a titrating solution for arsenic in ores (Chapter V), dissolve 12.828 grams of pure iodine and 19 grams of potassium iodide in '200 c.c. of water and dilute to 1 liter. 1 c.c. nearly equals .005^gram of arsenious oxide, As 4 O 6 . To standardize, dissolve .2 gram of pure arsenious oxide in a very little sodium hydroxide. Dilute to 300 c.c., make acid with sulphuric acid, using only one or two drops of dilute acid in excess, as determined by a floating piece of litmus paper. Add an excess (4 to 10 grams) of sodium bicarbonate, then 3 c.c. of No. 56 starch solution, and titrate cold to a purplish blue. 21. Iodine Solution, for the titration of arsenic from refined copper, is made by dissolving 3.5 grams of "reagent" iodine with 7 grams of the potassium iodide in a little water and diluting to 1 liter. Allow to stand a day before use. 1 c.c. titrates about .001 gram of arsenic, and about .04 c.c. is required to produce the end-point. Standardize with .06 gram of the purest (99.9 per cent) arsenious oxide, just as for solution 20, but add five drops of a 10 per cent solution of potassium iodide to the arsenious solution immediately after the starch. The iodide is added to obtain a definite end-point and an immediate formation of iodide of starch in presence of very small amounts of arsenic. See method 6d, Chapter XII. 22. Iodine Solution for the titration of hydrogen sulphide (when liberated from cadmium sulphide in the determination of sulphur in steel, or of oxygen in copper), may be made of such strength that 1 c.c. equals .0025 gram of sulphur. To obtain this strength, dissolve 2 grams of purest " reagent" iodine in 100 c.c. of water with 5 grams of potassium iodide and dilute to 1 liter. For very exact work, the iodine solution may be made by Payne's formula. See " Methods of Iron Analysis," by F. C. Phillips, page 22. 23. Magnesium Sulphate, MgSO 4 .7 H 2 O, is used in the form of the well-known " magnesia mixture" for the precipitation of arsenic and phosphoric acids. Magnesium sulphate is better than the chloride because it contains only one-fifth as much calcium salt, which is present as an impurity. Fresenius' formula gives good results : Dissolve 1 part of magnesium sulphate crystals, 4 parts of ammonium chloride, and 4 parts by volume of am- monia water (d., 0.9) in 8 parts of water. Allow impurities to 44 ANALYSIS OF COPPER settle and then filter. An old solution, which has perceptibly attacked the glass, should be rejected unless any precipitated silica is deducted. 24. Magnesium Chloride, MgCl2.6H 2 0, for the separation of phosphoric acid from ores, slags, or limestone (Chapter V), may be made up by the formula of A. A. Blair. Dissolve 110 grams of the crystals in water and filter. Dissolve 280 grams of ammonium chloride in water, add a little bromine water, and a slight excess of ammonium hydroxide, heat nearly to boiling, and then filter. Mix the two liquids, render the mixed liquid faintly alkaline, dilute to 2 liters, shake occasionally, allow to stand two or three days, and filter a portion into a small bottle as required. 10 c.c. precipitates .15 gram of magnesium oxide, MgO. 25. Manganous Sulphate, MnSO4.4 H 2 O, is combined with phosphoric acid to make a titrating solution known as the Zimmerman-Reinhardt mixture. It is added in their method for the titration of reduced iron by potassium permanganate, in order to obtain a clear end-point. Dissolve 160 grams of the manganese salt and dilute to 1750 c.c. Add 330 c.c. of strong phosphoric acid and 320 c.c. of concentrated sulphuric acid. 10 to 20 c.c. of the mixture are added to each iron solution after reduction of the iron by stannous chloride and mercuric chloride. The preceding is the improved formula of Mixer and Dubois. 26. Mercuric Chloride, HgCl 2 , is used as a saturated solu- tion. ,5 c.c. are added to a solution of ferrous iron after reduction by stannous chloride, in order to destroy the slight excess of the latter, which should not be more than one or two drops. 27. Mercuric Nitrate is used for the amalgamation of copper borings in the " assay of gold in copper" (method 3, Chapter X). Twenty-five grams are dissolved in one liter. A saturated solu- tion of the sulphate in dilute sulphuric acid is even better. Take 10 c.c. of the nitrate, or 25 c.c. of the sulphate for each "assay ton" sample of copper. 28. Dimethyl Glyoxime, for the precipitation of nickel in ammoniacal solution, is dissolved in the proportion of 1 gram to 100 c.c. of ethyl alcohol. Used in method 3, Chapter XIII. 29. Nitroso-/3-Naphthol is employed as a precipitant for cobalt and for the separation of cobalt from nickel (method 2, Chapter XIII). REAGENTS AND STANDARD SOLUTIONS 45 Dissolve the salt in acetic acid of 50 per cent strength until the acid is saturated. 30. Potassium Cyanide, itCN, for the titration of copper in ores and mattes, according to' " western methods." Dissolve 22.5 grams in water and make up to one liter. Use the chemi- cally pure salt. 1 c.c. equals .005 gram of copper. Standardize it with matte, or ore, in which the copper has been determined electrolytically. 31. Potassium Cyanide, for the titration of nickel in ores, or mattes, is made of a strength of 24.5 grams per liter and stand- ardized against pure nickel, and diluted so that 1 c.c. equals .005 gram nickel (method 1, Chapter VII). 32. Potassium Ferrocyanide, K4Fe(CN) 6 .3H 2 O. For a standard solution, to be used in the titration of zinc in ores and slags, dissolve 21.63 grams of the salt with 7 grams of sodium sulphite crystals in water and make up to 1 liter. 1 c.c. is equivalent to nearly .005 gram of zinc. To standardize, dissolve .200 gram of chemically pure zinc, or freshly ignited chemically pure zinc oxide, in 15 c.c. of hydrochloric acid. Add 7 grams of ammonium chloride. Dilute to 200 c.c. with boiling water and titrate, using a solution of one gram of ammonium molybdate in 100 c.c. of water for an indicator. Deduct amount of ferro- cyanide required to affect the indicator from the total amount used (methods 16 and 17, Chapter VII). 33. Potassium Ferrocyanide, as a standard solution for the titration of zinc in brass and German silver (Chapter XIV), is prepared as follows: Dissolve 80 grams of the salt in 2500 c.c. of distilled water, and allow to stand at least six weeks before use, to obtain a permanent solution. To standardize, weigh 2 grams of zinc, dissolve in nitric acid, and make up to one liter. Take 100 c.c. (.2 gram) zinc, add 5 c.c. nitric acid, 3 c.c. strong ferric chloride solution, 20 c.c. of saturated citric acid solution, dilute, and make distinctly alkaline with ammonia. The final volume should be 250 c.c. Boil the liquid and titrate the boiling solution. Fill pits in a porcelain test plate with 50 per cent acetic acid, and add 2 drops of zinc solution to determine the end-point which changes from greenish to clear blue. The correction for the amount necessary to reach the end-point in a solution free from zinc, must be made by each operator, and is about c.c. 46 ANALYSIS OF COPPER 34. Potassium Bichromate, K 2 Cr 2 O 7 , is used to make up a standard titrating solution for iron in the " Western methods" for ores (Chapter V). Dissolve 3.408 grams in water' and make up to one liter. 1 c.c. is equal to approximately .005 gram of ferrous oxide, FeO. With 4.39 grams of fused salt per liter, the iron value is about .005 gram per c.c. To standardize in terms of iron, or ferrous oxide, dissolve .2 gram of analyzed iron wire in 10 c.c. of (1:1) hydrochloric acid with the addition of a few crystals of potassium chlorate to destroy any trace of organic matter. Or 1 gram of pure ferrous ammonium .sulphate may be dissolved in 100 c.c. of hot water and 10 c.c. of the hydrochloric acid. In either case, add one or two drops of stannous chloride in excess of the amount required to decolorize the liquid. Remove this excess of stannous chloride by adding 5 c.c. of a saturated solution of mercuric chloride, which should pro- duce a white (not dark) precipitate. Cool by diluting to 300 c.c. or more and titrate. Test drops of the liquid on a por- celain tile with a drop of potassium ferricyanide. 35. Potassium Hydroxide, KOH. A normal solution is 56.1 grams per liter. Potash (purified by alcohol) should not be used for gas analysis. 36. Potassium Chromate, as adopted in the Guess method for lead in ores, or mattes, is made up with 100 grams of the salt, K 2 CrO4, per liter. 1 c.c. precipitates about 106 milligrams of lead. 37. Potassium Iodide, KI, in the form of a 10 per cent solu- tion, is added in the titration of arsenious oxide by standard iodine (No. 20) in order to have the ions required to produce an immediate formation of the iodide of starch with the slightest excess of iodine, even when the solution titrated contains but a trace of arsenic. Add five or six drops after the solution has been made alkaline by the large excess of sodium bicarbonate. The titration is described in detail under the "Determination of arsenic in refined copper" (method 6d, Chapter XII). 38. Potassium Iodide, KI, is also employed as an active reagent for the titration of nickel in matte (method 14, Chapter VI). 40 grams are dissolved in 1 liter. 39. Potassium Permanganate, KMnO4, for the titration of moderate amounts of iron in solution, is made up of such a REAGENTS AND STANDARD SOLUTIONS 47 strength that 1 c.c. will oxidize about .005 gram of iron. To this end, dissolve 2.86 grams in water and dilute to 1 liter. This strength may be doubled for iron ores or any rich material. The iron (Fe) value of^ 1 c.c. multiplied by .2951 gives the value in terms of manganese (Mn.), as determined by Volhard's method. The solution is standardized by various methods in different laboratories. The author prefers soft iron wire of known composition, for the determination of the iron value. In the titration of small amounts of manganese in copper ores, the value in terms of manganese may be found by multiplying the value in calcium oxide, CaO, by .5878. 40. Potassium Permanganate, for the titration of oxalate of calcium, obtained from copper ores and slags, may be best standardized by perfectly dry, pure sodium oxalate. The strength of permanganate employed in " western methods" is 5.643 grams per liter. 1 c.c. is nearly equal to .005 gram calcium oxide, CaO. 56.07 parts of calcium oxide, CaO, require 134 parts of very pure standard oxalate. The salt should be thoroughly dried in an air oven at 100 C. and preserved in a small glass-stoppered bottle. Sodium oxalate X .41843 equals CaO, and 1 part of iron (Fe) equals .50206 part of calcium oxide (CaO). 41. Potassium Permanganate, for the estimation of manganese by the "bismuthate method/' is made up very dilute and run against a special ferrous-ammonium-sulphate solution (42) or with sodium arsenite (47). A .03 normal solution (1 gram per liter) will titrate convienently .02 gram of manganese (Mn). For rich ores, use a .1 normal solution (or 3.1 grams per liter). 42. Special Ferrous Ammonium Sulphate, for the " bismuth- ate method," is dissolved to form a solution of 12.4 grams of the salt. The concentrated liquid is then treated with 50 c.c. of a mixture of equal volumes of strong sulphuric acid and strong phosphoric acid, and diluted to exactly 1 liter. This will give a standard solution almost equal to the weaker .03 normal permanganate (41). For .1 normal permanganate, use 39.2 grams of the iron salt and double the amount of the two acids em- ployed. See volumetric method for chromium (15, Chapter VI). 43. Potassium Thiocyanate, KSCN, is dissolved and diluted to such a degree that 1 c.c. will precipitate about .12 gram of 48 ANALYSIS OF COPPER copper. Refer to solutions 7 and 8, and method 1, Chapter XII. For ores, a more dilute solution of 40 grams of the crystals per liter is sufficient, as described in method 3, Chapter IV. 44. Silver Nitrate, AgNO 3 , is employed in method 5, Chapter V, as a precipitant of arsenic acid. Make a 10 per cent solution by weight and use 7 c.c. for .1 gram of arsenic. 45. Silver Nitrate, in method 1, Chapter VII, is used as an indicator in the titration of nickel. One gram is dissolved in water, and the solution brought to a volume of 1 liter. 46. Silver Nitrate, for the precipitation of chlorine in copper electrolyte, is made up as a normal solution of 169.89 grams per liter. 1 c.c. will combine with .03546 gram of chlorine. For very exact standardization, the solution would, of course, be run against standard salt solution, or standardized ammonium thiocyanate (No. 9). 47. Sodium Arsenite. A stock solution is made by heating in a flask on a water bath 15 grams of arsenious oxide (As 2 O3), 45 g. of sodium carbonate and 150 c.c. of water. Cool and make up to 1000 c.c. with water. A standard solution of the proper strength for titration of chromium or manganese in ores or metals is made by diluting 300 c.c. to 1 liter and standard- izing it with permanganate. The value of the permanganate itself is best obtained by comparison with pure sodium oxalate (see 25, chapter VI). 47 a. Sodium Citrate, 200 grams per liter, is a reagent for 1, Chapter VII. 48. Sodium Chloride, NaCl, in the assay of copper bullion, Chapter X, is a precipitant of silver. If 19 grams are dissolved in 1 liter, 1 c.c. will precipitate 350 milligrams of silver. Refer also to No. 17. 49. Sodium Hydroxide, NaOH, is made up as a normal solution for the determination of the free acid in copper elec- trolyte, as described in Chapter IX. Dissolve 40 grams of the purified sticks in 1 liter, and standardize against normal sul- phuric acid, using a .1 per cent aqueous solution of methyl orange as an indicator. The acid is valued by titrating it against .8-gram portions of chemically pure sodium carbonate, which has been ignited just below a red heat. 50. Sodium Thiosulphate (or hyposulphite), Na2S 2 O3.5H 2 O, is the titrating solution in the " iodide method" for copper in REAGENTS AND STANDARD SOLUTIONS 49 ores or mattes, as given in method 2, Chapter IV, also in method 22, Chapter VI. If 19.59 grams are dissolved in 1 liter, 1 c.c. will react with about .005 gra*m of copper. 51. Sodium Phosphate, secondary, Na 2 HPO4.12H 2 O, is dis- solved to form a 10 per cent solution of the crystals, for the precipitation of either aluminum, magnesium, manganese, or zinc. The ammonium salt (5), when it is permissible, is prefer- able to the sodium compound, as the ammonium salt is easily volatilized, if traces are inclosed in a washed precipitate. 52. Sodium Sulphide, Na 2 S, as a filtered solution of the pure commercial salt, extracts or separates the sulphides of the " arsenic group" of metals from copper, lead, or silver, but not so well from bismuth, unless a little potassium hydroxide is added. A filtered solution with a density of 1.08 is diluted with about two parts of water for use in copper analysis. For elec- trolysis of antimony sulphide, it is necessary to provide a strictly colorless monosulphide. The yellow color may be removed by heating with sodium peroxide, hydrogen peroxide, etc., or the sodium sulphide may be prepared by a method described by A. Classen and H. Koch. Dissolve about 60 grams of the purest commercial sodium hydroxide in water, dilute until the liquid has a density of about 1.2, then saturate one half of the liquid with hydrogen sulphide. Add the other half of the solution, and if the monosulphide is required, filter the solution directly into stoppered bottles. Dilute for use, according to conditions. 53. Stannous Chloride, SnCl 2 .2H 2 O, for reduction of iron. Dissolve metallic tin foil in hydrochloric acid, or use the formula of J. M. Camp: Dissolve 300 grams of chloride crystals in 500 c.c. of strong hydrochloric acid and 500 c.c. of water. Boil with a few scraps of tin until clear and then bottle. 1 c.c. reduces about 1 gram of iron. 54. Normal Sulphuric Acid, H 2 SC>4, is produced by mixing with water an amount of the acid equivalent to 49.04 grams of the 100 per cent acid, and then diluting to 1 liter. It is stand- ardized against sodium carbonate ignited, and may be compared with normal sodium, or potassium hydroxides (Nos. 35 and 49). Use an aqueous solution of methyl orange as an indicator, except in cases where bicarbonates are to be estimated. In very accurate work, the temperature should be noted with each titration, to apply a correction, if necessary. 50 ANALYSIS OF COPPER 55. Standard Titanium Solution, for method 13, Chapter VII, is prepared by dissolving the pure oxide by fusing it, or by treating with 10 c.c. of hot strong sulphuric acid and 5 grams of potassium bisulphate. The solution is mixed with dilute sulphuric acid at first, to prevent precipitation. Then add water until 1 c.c. contains 1 milligram of titanic oxide, Ti0 2 . 56. Starch Indicator (for titrations with iodine) . According to the formula of J. M. Camp, add to .5 gallon (2 liters) of boiling water about 25 grams of pure wheat starch, previously stirred up into a thin paste with cold water. This is boiled for ten minutes, and, when cold, about 25 grams of pure zinc chloride dissolved in water are added and the solution diluted to 1 gallon, or 3800 c.c. Mix thoroughly and allow to settle over night; siphon the clear solution into a glass-stoppered bottle. It will keep indefinitely. This preparation is suitable for the titration of the acidified solution of cadmium sulphide obtained in the evolution method for sulphur in steel. 57. Special Starch is prepared by hydrolyzing the ground material with 1 per cent hydrochloric acid, allowing it to stand over night in the cold; then filtering, and washing with cold water. Dry the washed starch and heat for three hours in the air oven at 100 C. Dissolve 2 grams of this starch in 500 c.c. of boiling water, and add 15 drops of oil of cassia to prevent fermentation (method 6, Chapter XII). 58. Zinc Chloride, ZnCl 2 , finds its principal use in a copper laboratory as one ingredient of a concentrated distilling solution used by chemists in the western states for the distillation of arsenious chloride from the sulphide in the analysis of ores. Distilling Solution. Dissolve one pound (453.6 grams) of chemically pure zinc by adding to it gradually a mixture of 1250 c.c. of hydrochloric acid (d., 1.2) and 500 c.c. of water. When the zinc has dissolved, evaporate the solution to 1100 c.c. and mix the whole amount with 1 liter of the concentrated solution of cupric chloride (No. 17). CHAPTER IV THE ASSAY OF COPPER IN ORES AND FURNACE PRODUCTS THIS chapter is restricted to the assay of copper in its solu- tions. The following methods are in regular use for the estima- tion of copper in ores, native copper, furnace by-products, or mill tailings: (1) titration with potassium iodide; 1 (2) tit ration with potassium cyanide; 2 (3) precipitation with potassium thio- cyanate and titration of the resulting compound with potassium permanganate in presence of excess of caustic alkali; 3 (4) the colorimetric assay; 4 (5-9) electro-analysis. 5 The selection of method varies with the character of the material and the preference of the chemist, but the electrolytic assay is the most accurate. 1. Titration by Iodide. 1 According to standard western practice, the weighed sample of 1 gram (.5 for ores over 25 per cent copper), is decomposed in a 3-inch (7.5 cm.) casserole by treatment with nitric and hydrochloric acid in excess and 10 c.c. of sulphuric acid. Evaporate until dense white fumes of the last acid are evolved. After cooling, dilute with water and boil until all soluble sulphates are in solution. Filter off the insoluble residue and wash well with hot water, allowing the filtrate and washings to run into a 350 c.c. beaker. To this filtrate add 10 c.c. of a saturated solution of sodium thiosulphate and boil until the precipitated sulphides settle readily. Filter and wash with cold water until free from iron salts (usually 4 to 7 times). Carefully wrap the paper, fold it into a dry filter, and place it in a porce- lain crucible. The crucibles are placed in a muffle that is kept 1 Method 1. Eng. and Min. Jour. 89, 498; /. Am. Chem. Soc. 24 (1902), 1082 and 580; Eng. and Min. Jour. Nov. 17, 1904; (1910) 1221; 74 (1902), 846. 2 Low's Technical Analysis. 3 J. Am. Chem. Soc. 20 (1898), 610. 4 J. Am. Chem. Soc. 19 (1897), 24; Trans. A. I. M. E. 30, 851. 5 Eng. and Min. Jour. 84, 773; also 77 (1909), 159; 94, 315. J. Am. Chem. Soc. Nov. 1907. El. Chem. and Met. Ind. 6, 19 and 58. 52 ANALYSIS OF COPPER barely red hot, to roast and dry the precipitate, thus removing the sulphur and much of the arsenic. Too much heat must be carefully avoided at this stage, or the mass will spit before it is dry, and later on will fuse into the porcelain. It is to prevent " spitting" that the extra filter paper is used, and when properly ignited, the residue will be found enveloped in ash, and . easily transferable to a flask. Now transfer the residue to a small " copper flask," decompose it with 5 c.c. of nitric acid-potassium chlorate mixture, and evaporate almost to dryness to oxidize any remaining arsenic, etc. Add dilute ammonia to alkaline reaction and boil the liquid well for two minutes. Acidify slightly with acetic acid without boil- ing, cool, and add about 3 grams of solid potassium iodide. Determine the copper by titration with sodium thiosulphate (Na 2 S 2 O 3 ), using starch as an indicator. (Solutions 48 and 54, Chapter III.) The Western chemists, just quoted, did not find the reduction of copper by aluminum to be satisfactory. It not only took more time, but with special ores and conditions, the precipitation was not complete. Besides, particles of the fine copper were easily lost on filtering, or were oxidized and dis- solved by the wash water. Reduction by Aluminum. A. M. Fairlie, 1 A. H. Low, 2 and others prevent oxidation of the copper by the addition of 15 c.c. of hydrogen sulphide water as soon as the reduc- tion by aluminum is completed. This scheme prevents the formation of a bulky mass of sulphides from rich ores. If such reduction is preferred, it may be effected after all silver chloride has been filtered off, by boiling the sulphuric solution in a wide lip- less beaker with three pieces of sheet aluminum, each 1^ inches (or 3.8 cm.) square. Continue the boiling until the copper is out of solution and the aluminum appears bright and clean on shaking. Pour through a 9 cm. filter directly after the addition of the hydrogen sulphide water, but hold back most of the metal and wash it two or three times with diluted hydrogen sulphide water. Drain thoroughly and redissolve through the funnel by 10 c.c. of dilute (1:1) nitric acid. Finally, wash the aluminum 1 Eng. and Min. Jour. 84, 773; also 77 (1909), 159; 94, 315. J. Am. Chem. Soc. Nov. 1907. El. Chem. and Met. Ind.. 6, 19 and 58. 2 J. Am. Chem. Soc. 30 (1908), 760. COPPER IN ORES AND FURNACE PRODUCTS 53 well, cleanse any sulphur on the filter by treatment with 5 c.c. of saturated bromine water, and. finish the washing with hot water. Boil out all the bromine 1 and titrate by iodine as before. Kendall 1 recommends, the addition of sodium hypochlorite and phenol in order to secure more even results in the titration, and A. H. Low 2 advises a slight boiling after the addition of the acetic acid to prevent a return of color after titration. Refer also to the special " iodide" titration of nickel in nickeliferous copper matte (14, Chapter VI). 2. Titration with Potassium Cyanide. 2 The action of this reagent on copper solution is a function of several variables. The conditions observed in standardization must be exactly observed in the work on the ore samples. The main points are uniform temperature, regularity in titration, the same excess of alkali, and the same final volume, in each case. Standard Solution. 22.5 grams of chemically pure potassium cyanide per liter make a solution, 1 c.c. of which will have a value of about .005 gram of copper. Refer to formula 28 of Chapter III. Unless the ores are very impure, western operators usually prefer to standardize against matte, or ores, in which the copper contents have been determined electrolytically. Others prefer, instead, to take .2 gram to .3 gram of chemically pure copper foil, dissolve in 5 c.c. of strong nitric acid, dilute with 5 c.c. of saturated bromine water plus 25 c.c. of distilled water, and boil out the bromine. Add 10 c.c. of ammonia (d., .90) with 50 c.c. of water and cool quickly to room temperature, then titrate exactly as for ores. The standard cyanide is subject to change, and it is necessary to protect the liquid from the sun, or any strong light, and retest it frequently. (a) Assay of Pure Ores. Such ores, or standard mattes, are dissolved in nitric acid. Add 10 c.c. of strong acid to 1 gram of sample in a 500 c.c. flask. Take only .5 gram when over 25 per cent copper. Heat on a steam bath until the brown fumes are driven off, then add 200 c.c. of cold water and 20 c.c. of ammonium hydroxide (d., .90). Titrate slowly, adding po- tassium cyanide in small amounts only and allowing the deposit 1 J. Am. Chem. Soc., 33 (1911), 1947; 34, 347. 2 Personal communications. 54 ANALYSIS OF COPPER to settle after each addition. When the solution in the flask is a pale violet color, filter into another 500 c.c. flask and add the cyanide until colorless. The burette reading is multiplied by the factor for its copper value, which is determined by run- ning a half-gram sample of "standard" matte as a standard with each set of assays. (6) Samples containing organic matter, or much arsenic or manganese, are decomposed with nitric acid and a small amount of potassium chlorate. Large amounts of organic matter may keep some iron in solution, producing greenish tints. In this case, roast sample gently in a scorifier; then treat the residue in a platinum dish with 5 c.c. of nitric acid, 5 c.c. hydrofluoric acid, and 2 c.c. sulphuric acid, and evaporate to fumes of sul- phuric anhydride. Add 5 to 10 c.c. of nitric acid, wash into the flask, and proceed as before. Good results are often obtained by adding small portions of potassium chlorate to the boiling nitric acid solution. Some oxidized ores and slags require a preliminary treatment with hydrochloric acid or the addition of hydrofluoric acid for complete solution of the copper. In such a case, add 5 c.c. of each of the acids, heat gently for ten minutes, add 10 c.c. of nitric acid, and take to dryness. Take up with water and 5 to 10 c.c. of nitric acid, wash into the flask, and proceed as before. If much silver is present, remove it with a few 'drops of hydrochloric acid before adding the ammonia to any of the samples under treatment. (c) Impure Ores, or Mattes, etc. In presence of much zinc, cobalt, or nickel, a preliminary separation of the copper must be made, preferably by sodium thiosulphate, as in method 1. The roasted sulphides should be dissolved in nitric acid and titrated as before, except that chemically pure copper foil is used for the standard, because the titration is made in the absence of iron. 3. Thiocyanate Method (By F. G. Hawley). 1 Weigh .5 to 1 gram of ore into a tall 300 c.c. beaker, add 12.5 c.c. of "acid mixture" (1 part sulphuric acid, 2 parts nitric acid and 1 of a satu- rated solution of potassium chlorate in nitric acid). Then, when nearly decomposed, add 10 drops of hydrofluoric acid, evaporate to strong white fumes of sulphuric acid, and cool the residue. Add 60 c.c. of water and just neutralize with ammonia. Add 5 c.c. of hydrochloric acid, then 10 to 12 c.c. (according to cop- 1 Eng. and Min. Jour. 90 (1910), 647. COPPER IN ORES AND FURNACE PRODUCTS 55 per content) of potassium thiocyanate (40 grams per liter). Boil for two minutes and remove ^from the plate. Let the beaker stand for 5 minutes covered, then for 5 minutes uncovered, and filter through a 12.5 cm. jilter, No* 597 S. and S. Wash four times with hot water (60 to 70), then place the original beaker under the funnel, and, with a wash bottle, treat the copper salt with a boiling 5 per cent solution of sodium hydroxide. Use a medium fine jet and stir the precipitate thoroughly.- Wash four times with hot water, cool the filtrate somewhat, make acid with slightly diluted sulphuric acid, and immediately titrate with standard potassium permanganate, of which 1 c.c. equals .01 gram of metallic iron, Fe. High coppers should be titrated cold and in a volume of not less than 200 c.c. A conversion table is employed. The thiocyanate may also be dissolved as in 5c and the solution electrolyzed. 3a. Method of D. J. Demorest. 1 This modification is said to permit an accurate titration by permanganate without the use of any empirical conversion table. Weigh out enough of the ore to have present .05 to .30 gram of copper. Transfer the sam- ple to a 200 c.c. beaker, add 5 c.c. of strong hydrochloric acid, and heat for several minutes; then add 10 c.c. of nitric acid and digest on a hot plate until the 'ore is completely decomposed. Then add 10 c.c. of (1:1) sulphuric acid and boil down until white fumes appear. Cool, add 50 c.c. of water containing 3 grams of tartaric acid, and heat until all soluble salts are in solution. Cool and add ammonia until the solution turns a deep blue, then add sulphuric acid until the liquid is just acid, then 1 c.c. more. Now add 1 gram of sodium sulphite dissolved in 20 c.c. of water, heat nearly to boiling, and add slowly, with vigorous stirring, one gram of potassium thiocyanate dissolved in 20 c.c. of water. Heat at a nearly boiling temperature for several minutes to coagulate the precipitate and dissolve out all the tartaric acid. Cool somewhat and filter hot, preferably through an asbestos mat on a Gooch filter. Wash well with water and rinse out the suction flask. Then pour through the felt 30 c.c. of a hot 10 per cent sodium hydroxide solution and wash well with water. The assay should be finished hot. Warm the filtrate to about 50 and titrate, running in the 1 J. Ind. and Eng. Chem. 5 (1913), 215. 56 ANALYSIS OF COPPER permanganate slowly and shaking the flask vigorously. The solution turns green. After about 10 c.c. have been run in, take out a drop of liquid, place it in a drop of hydrochloric acid on a white paraffined plate, and add a drop of 10 per cent ferric chloride solution. If a red color appears, continue to add the permanganate, testing after each 5 c.c. until the red becomes weaker, then test often until the red tint is faint. Add 30 c.c. of (1:1) sulphuric * acid, shake until dissolved, and finish the titration. The copper value is .1897 times the iron value. Refer to methods 5c and 8 for electrolytic modifications of the thiocyanate method. For safety, titrate under a hood. 4a. The Colorimetric Assay. Western Method (after Thorn Smith) . Standards for the assay of blast furnace slags are prepared as follows: Take 3 grams of sample of an ore, tested electrolytically, cover it with water, add 10 c.c. of nitric acid (d., 1. 4) and 1 c.c. of hydrochloric (d., 1. 2); heat for a few minutes on a steam bath; dilute with 100 c.c. of water, and then add a slight excess of dilute ammonia. Filter into bottles of uniform size, for compari- son of color, and wash until the bottle is filled to the 200 c.c. mark. If the true copper in this sample was .20 per cent, this standard is called 'B. 2. To prepare B. 3, add .003 gram of copper to another 3 grams of the same sample. B. 4 is prepared by adding .006 gram copper and B. 5 by add- ing .009 gram of copper, the copper always being added before the ammonia and the samples each treated as with B. 2. Tailings. Heat 1 gram of sample on the steam bath with 5 c.c. of nitric acid and a pinch of potassium chlorate; dilute, filter, and wash as in the assay of slags. If the copper per cent by electrolysis was .50, this standard is called T. 5. To pre- pare T. 6, add .001 gram of copper to anothe^ 1-gram sample; .002 gram for T. 7; .003 gram for T.8; .004 gram for T. 9; .005 gram for T. 10. Reverberatory Slags. Take 2 grams of any of the Mon- tana, or Arizona, products, add 10 c.c. of hydrochloric acid and 2 c.c. of nitric acid. Heat, dilute, filter after the addition of ammonia, and wash as in tests of slags from cupolas. If the true copper on this sample was .30 per cent, this standard is called R. 3. To prepare R. 4, add .002 gram of copper to another 2 grams of sample. Add .004 grams for R. 5, and COPPER IN ORES AND FURNACE PRODUCTS 57 treat each as in the preparation of R. 3. If samples low enough to furnish the lowest standard -are not at hand, remove the copper first by electrolysis from a sample and then add the re- quired amount of pure^standarcl copper solution for the first number in the set. A set of standards having been prepared for each product as above, another set of bottles is filled almost to the mark with water and 10 c.c. of ammonium hydroxide (.90). A standard solution, containing .001 gram of copper to the c.c., is then run into each from a burette until the color matches the above standards, and the burette reading in each case is carefully noted. The following results were obtained by the chemists. From this table prepare the standards. BURETTE READINGS FOR STANDARDS B.2 required 4.4 cc. B.3 " 6.7 " R.2 required R.3 3.2 cc. 4.6 " T.4 required T.5 3.8 cc. 4.7 " B.4 " 9.1 " R.4 6.1 " T.6 5.7 " B.5 " 11.1 " R.5 7.9 " T.7 6.6 " R.6 9.7 " T.8 7.6 " R.7 11.4 " T.9 8.5 " R.8 13.2 " T.10 " 9.5 " Determination of Copper (in unknown samples) . Treat the blast, reverberatory slags, or tailings, exactly as described above and match the filtrates with the standards, which were prepared by separating the copper in presence of nearly the same interfering elements found in the regular works samples reported to the office for analysis. 4b. Lake Superior Method. The color test has been super- seded by rapid electrolysis in testing slags or tailings high in soluble iron, but is still found useful for the classification of lots of sludges from the diamond drill, and for the rapid assay of Lake mill tailings, and assay slags from the fire assays of native copper products. The permanent standards are diluted from a strong solution to the uniform volume of 200 c.c. by means of dilute ammonia, 1 volume of ammonia (d., .9) to 6 volumes of water. If preserved in thin-walled, cylindrical bottles with tight glass stoppers, the solutions will remain unchanged for about a year. Oil sample bottles will answer but are inferior. 2.5 grams of sample are taken as a standard sample for analysis and also for the set of standards. Dissolve .3 gram 58 ANALYSIS OF COPPER of pure copper in 5 c.c. of nitric acid in a 500 c.c. flask, treat with 5 c.c. of sulphuric acid, boil out the nitric acid, and make the solution up to 1500 c.c. with dilute ammonia as already described. Then 1 c.c. ' contains .0002 gram of copper. Fill a burette with the well-mixed solution and run into each standard bottle the amount required to make a set ranging from .1 to 1 per cent of 2.5 grams. The diameter of the bottles is 4.4 cm. and the height to the 200 c.c. mark, 15.2 cm. Cupola Slags. Treat 2.5 grams of powdered slag in a No. 4 porcelain casserole with 7 c.c. of nitric, 7 c.c. of sulphuric acid, and 7 c.c. of water. Warm, and finally boil down to white fumes over a Bunsen burner, stirring to break up any clots and form a paste. Stir in 50 c.c. of water while still warm and, when the soluble salts are dissolved, add sufficient ammonium hydroxide to pre- cipitate the iron and alumina. Pour the solution through a No. 3 Munktell 15 cm. filter into a bottle of the same volume as the standard set. Wash the mass with dilute ammonia until the washings appear colorless, and then make the solution up to the 200 c.c. mark with the same dilute ammonia prescribed for the standards. When the bottle is nearly filled, if the tint on shaking appears greener than the standard bottle of about the same depth of color, then complete the dilution with pure water instead of ammonia. If the ferrous oxide is over 10 per cent, the first precipitate of ferric hydroxide should be washed only once, drained, and then washed back into the casserole with a jet of water, using as little as possible. Redissolve the iron in a very little dilute sulphuric acid, precipitate again with a little ammonia, run through the filter into the bottle, and wash until the filtrate is clear. 4c. Color Test of Mansfeld Shales. H. Koch weighs 2 grams of substance into a small porcelain crucible and roasts the contents on a sand bath, afterwards decomposing the sample exactly as in the electrolytic assay of ore and shales (method 5). From the residue on evaporation, the sulphate of copper is dis- solved as usual, then the solution (50 c.c.) with the residue is transferred to a thick-walled, cylindrical glass, 6 to 7 cm. wide and 14 cm. high, which is marked at 250 cm. After the addi- COPPER IN ORES AND FURNACE PRODUCTS 59 tion of 30 c.c. ammonium hydroxide (d., .91), it is filled to the mark with water and stirred. After settling, the fluid is passed through a dry 14 cm. filter intb a square-cornered bottle with ground sides. The area.of the battle adopted is 60 by 40 mm., the height to shoulder about 70 mm., and the whole contents about 110 cm.; the thickness of the walls must be uniform. The copper values are ascertained by comparison with standard solutions which are prepared in the central labora- tory from shales of known copper content. The copper is usually calculated in kilos per metric ton. The minus error increases with the copper, just as in tests on American ores or slags, hence the standard solutions are prepared according to an empirical scale of perhaps 20 bottles, to read colors from ma- terial varying in copper contents from 2 kg. to 140 kg. per ton. The copper in shales of the usual contents of 25 to 35 kg. (2.5 to 3.5 per cent) is estimated with an accuracy of 1 kg. per metric ton (.1 per cent). 5. Copper by Electrolysis. In routine work, as observed by A. M. Smoot, the electrolytic assay has advantages over all others, because it is applicable to any sample from refined copper to the lowest tailings; and further, it admits the use of. large charges, thus dividing the errors inherent in all such processes, which is an especial advantage with high grade material. This fact is so well recognized that practically no other method is used by large firms for control and umpire assays. It is wrong to assume, however, that the copper is thus per- fectly separated from other associated elements, and even with the purest refined copper, high results may be obtained by oxida- tion of the cathode or by occlusion of gases by the deposit. A preliminary separation of copper from associated impurities is sometimes necessary, thus making the electrolysis a finishing process to obtain a weighable deposit. Chemists are referred to Chapter I for a description of the most convenient apparatus. Perforated cathodes are described in 1, Chapter XI. Acid Electrolyte. In Western reduction works, nitric acid is often used alone, particularly with leady gres, if a deposit of lead peroxide is also required, but there is a tendency to oxidation of copper unless the per cent of copper is small and the time of deposition rather short. In accurate work on metallic copper 60 ANALYSIS OF COPPER (Chapter XI), sulphuric acid is therefore added in such amount that when the electrolysis is completed, sufficient free sulphuric acid shall be present to retain arsenic, iron, and other impurities in solution after the greater part of the nitric acid has been re- duced to ammonia by the electric current. 5a. Western Assay of Sulphide Ores. Weigh 1 gram of ore into a beaker (3.5 inches high and 2.25 inches diameter); add 8 c.c. of nitric acid and a little potassium chlorate. Take to complete dryness on the steam bath. Take up with water and 6 to 10 c.c. of nitric acid, fill the beaker with water, allow to settle, and place on the battery. It is often better to use a little sulphuric acid, 2 to 5 c.c., and drive off nearly all the nitric acid, finally adding an exact amount (4 c.c.) of nitric before electrolysis. 5b. Oxidized Ores. Take a one-gram sample to dryness with 8 c.c. of nitric acid; add 10 c,c. of hydrochloric acid and 2 c.c. of sulphuric acid and take down to fumes of sulphuric anhydride. Dissolve the salts with 8 c.c. of nitric acid and water, allow to settle, and place on the battery. 5c. Mattes. Moisten a one-gram sample with a few drops of water, add 8 c.c. of nitric acid and 1 c.c. of sulphuric acid, and take to dryness on a steam bath. Dissolve with water and 8 c.c. of nitric acid, filter, and electrolyze. The percentage of silver, as determined by fire assay, is deducted from the per- centage of copper plus silver found by electrolysis. 291.66 ounces per ton equals 1 per cent. If the silver is considerable, it may be precipitated with just sufficient dilute hydrochloric acid, avoiding an excess. The silver chloride is filtered off, the traces of hydrochloric acid removed by evaporation and the copper alone deposited by the current. 5d. Western Slags. Decompose 2 grams in a platinum dish with 6 c.c. of nitric acid, 8 c.c. of hydrofluoric acid, 1 to 2 c.c. of sulphuric acid, and evaporate to white fumes. Take up with water and 10 c.c. of sulphuric acid and place on the battery. Allow about .11 ampere per assay (with a 110-volt current); this gives a potential of 1.4 to 2.3 volts across the elec- trode terminals of each assay. Finally, test a little of the liquid with a drop of hydrogen sulphide water on a spot plate to prove that the copper has all been deposited. Wash the platinum cathodes first by dropping them very rapidly into a beaker of COPPER IN ORES AND FURNACE PRODUCTS 61 water, then by a jet from a bottle. Remove the water by means of 94 per cent denatured alcohol, carefully burn off the slight excess of alcohol (keeping the plate in rapid motion) and then cool and weigh the cathodes. If the deposits are either slightly grayish, or show dark spots due to arsenic, they may be dissolved in 8 c.c. of nitric acid, or in an acid mixture, and the copper again deposited, removing the electrode just as soon as the process is completed. Copper in ores and furnace products may be separated from bismuth, antimony, and arsenic by precipitation as thiocyanate (3). The white salt is washed thoroughly, carefully ignited in a porcelain crucible, dissolved in 7 c.c. of nitric acid, and the copper esti- mated by electrolysis as before. Instead of igniting the precipi- tate, the excess of ammonium thiocyanate may be destroyed by evaporation with nitric acid. Compare also ''Eastern methods.' 7 6. Lake Superior Method for Chilled Slags. If a sample of slag, or other furnace product, is evaporated to fumes with acids, without any addition of hydrofluoric acid, enough copper is fre- quently retained in the silica to cause a serious error. One of the author's assistants noted that chilled slags, granulated in water, are almost completely soluble in dilute boiling sulphuric acid without separation of silica. The copper is thus taken entirely into solution. Grind the slag to pass through a sieve of 100 meshes to the linear inch (40 per cm.). Weigh 2 grams of the powder, from which any shot has been sifted out and separately weighed. Transfer from the balanced watch glass to a tall 300 c.c. lipless beaker, 12.5 cm. (5 inches) in height, and 5.7 cm. in diameter. Add 80 to 100 c.c of distilled water, stirring rapidly until the sample is entirely in suspension. Continue the rapid stirring and add from a graduate 12 c.c. of sulphuric acid (d., 1.84), place over a lamp, and bring to boiling, while stirring to preserve the suspended condition. Boil for about three minutes, and break up any particles of slag which may have settled. The solution should be clear or slightly milky and no more than a trace of silica should remain on the bottom. Now add carefully three or four drops of nitric acid. A violent effervescence will ensue. As this subsides, and the iron rapidly changes to the ferric state, add 7 c.c. of nitric acid if the copper is to be deposited over night, or 8 to 9 c.c. if the cop- per assay is to be placed in the Frary rotary device, or in con- 62 ANALYSIS OF COPPER nection with a rotating anode spindle. Dilute to 150 c.c., and electrolyze over night with a current of .8 to 1 ampere per assay, or in the rotary device with 3 to 4.5 amperes. When the copper is nearly deposited, wash down the split cover glasses, and, a few minutes later, test with hydrogen teulphide water on a porce- lain spot plate in the usual manner. In the case of yellow solutions, the color of the test should be compared with some untreated solution. Wash the plates rapidly with water and alcohol, ignite, and weigh as in the preceding method. In the Frary solenoid, the deposition of the copper may be completed in 45 to 90 minutes, but the current should not be allowed to act long after the deposition is complete. The solutions must be kept cold, or the copper may commence to redissolve. A solenoid may be cooled by the circulation of cold water be- tween the beaker and copper cylinder, if it is to be kept in con- tinuous operation. The operator should then be able to increase the current to 4.5 amperes per square decimeter, and complete the electrolysis of a sample of waste blast furnace slag in half an hour. Such a method is fairly rapid and far more accurate than any color test. The platinum cathode is made from a sheet 5 cm. high and 10 cm. wide, giving a total immersed depositing surface of 100 sq. cm., or 1 sq. decimeter, which will be regarded as a normal area in the measurement of current density. UMPIRE AND CONTROL ASSAYS AT THE PORT OF NEW YORK 7. Siliceous Ores. 1 Such material is generally tested, when high-grade, by the same method to be prescribed for high-grade matte. As much as 3 to 5 grams are taken if the ore is low- grade. A small amount of copper may be deposited in 6 to 8 hours with a current of .25 ampere. The large samples taken for electrolysis render these methods more suitable for umpire work than the more rapid titrations generally used in the western part of the United States. 8. Heavy Sulphides ; Thiosulphate Modification. Pyritous ores usually contain impurities which would contaminate the copper deposits, and the presence of much sulphate of iron in the electrolyte causes interference, just as it does in the assay of 1 A. M. Smoot, personal communication. COPPER IN ORES AND FURNACE PRODUCTS 63 ferruginous reverberatory or blast furnace slags from native copper. In the case of rich ores, the nitric acid should nearly all be driven out by evaporation; or with impure material, a preliminary separation may bei< effected by precipitating the copper as cuprous sulphide in the following manner. Moisten 5 grams of ore with 10 c.c. of water, add 10 c.c. of nitric acid (d., 1.42), and when the action has subsided, 10 c.c. more. Digest the mixture on a steam bath until the ore is de- composed and* the sulphur is clean, then boil for a few minutes, cool a little, and add 10 c.c. of sulphuric acid (d., 1.84). Evapo- rate on a hot plate until the first acid is expelled and white fumes are evolved. Cool, add 150 c.c. of water and 10 c.c. of sulphuric acid. If silver is present, precipitate all of it with sodium chloride, boil, filter, wash the residue with hot water, dilute to 300 c.c., and heat to boiling. To the boiling liquid, add, drop by drop, a saturated solution of sodium thiosulphate. The ferric iron will be rapidly reduced, and when colorless, the copper will begin to separate as cuprous sulphide. The point at which this begins is easily seen. Add about 2 c.c. of thiosulphate after the copper sulphide begins to form, and boil until the sulphide agglomerates. Filter the hot solution and wash with hot water, or better, hydrogen sulphide water. Dry and ignite in a No. 2 round-bottomed porcelain crucible. Moisten the residue with 3 c.c. of water, add 6 c.c. of nitric acid, and digest for a few minutes; finally, boil, transfer the liquid to a tall beaker and electrolyze for 8 hours or more, according to the per cent of copper, using a current of .25 ampere for cylinders 5.5 cm. wide by 3.8 cm. in height. A better deposit is produced by adding 10 c.c. of sulphuric acid also. The Thiocyanate Separation of Copper. This modification is similar to the titration method (3), or the western method for slags, and where antimony is present, is preferable to the thio- sulphate, which causes the deposition of at least a part of the antimony. According to A. M. Smoot, treat 5 grams of the ore as above, until the filtered solution from the insoluble residue has been obtained. Add ammonia until the liquid is alkaline and then make acid with hydrochloric acid, leaving an excess of about 1 c.c. more than is necessary to dissolve the iron hydroxide. Dilute to 300 c.c., add 50 c.c. of a saturated solution of sulphur 64 ANALYSIS OF COPPER dioxide, and heat on a steam bath until the iron is reduced. Add 3 to 5 c.c. of a solution of potassium thiocyanate (150 grams to 1000 c.c.) and heat again until the precipitate settles. Filter through double papers, wash with warm 2 per cent solution of ammonium nitrate, and proceed as in the thiosulphate method. Compare Western method for slags (5). 9. Roasted Ores, Iron Oxides, and Cinders usually contain copper which is insoluble in acids, even after prolonged digestion. Digest 3 grams of such ores with 25 c.c. hydrochloric acid until the iron oxide is partly dissolved, then add 10 c.c. nitric and boil until only about 5 c.c. remains. Cool, add 5 c.c. sulphuric acid, and evaporate until SO 3 fumes are given off. Cool, add 100 c.c. of water or more, and boil until all soluble salts are dissolved. Add sodium chloride, if necessary; filter and wash the residue with hot water. Transfer the residue to a porcelain crucible; dry, ignite, and fuse it with potassium pyrosulphate. Dissolve the fusion in hot water and filter it into the main solu- tion; or, if the insoluble copper is small, as is usually the case, estimate it separately, colorimetrically. Filter the ammoniacal solution through asbestos, and read the color in 100 c.c. tubes, according to method 4b. 9a. D. J. Demorest claims that a precipitation as sulpho- cyanate in ammonium tartrate gives a more complete separa- tion from antimony, arsenic, or bismuth, than is possible in sulphuric acid alone. If electrolysis is preferred to titration, the precipitate as obtained is dissolved by repeated treatment on the filter, in a covered funnel, with 18 c.c. of nitric acid (1:2), after which the liquid is boiled 5 minutes under a hood to de- stroy all the cyanogen compounds before electrolysis. (See 3a.) 10. Eastern Methods for Mattes are the same as those for Western mattes already described. One hundred per cent copper foil is, however, adopted for standards. An operator is justified in using any one of methods 5, 6, 7, 8, 9, that is best adapted to the material treated and the commercial requirements. 11. Copper by Electrolysis in Mansfeld Ores. For the assay of about 800 shale samples per month, 1 this method is said to offer the advantage that it permits the electrolysis of the un- filtered liquid for copper, the liquid only clarifying by long set- tling. The procedure is as follows: A sample of 2 grams of 1 H. Koch, personal communication. COPPER IN ORES AND FURNACE PRODUCTS 65 the fine powdered shale is burned at a dark red heat, in a small porcelain crucible in a Plattner's assay muffle, in order to drive off the bitumens. The roasted 'mass is transferred to a 150 c.c. beaker, 20 c.c. of a mixture of equ&l parts of nitric acid (d., 1.42) and sulphuric (d., 1.2) are added and the mixture evaporated to dry ness on the sand bath. The cooled mass is taken up in about 70 c.c. of a diluted nitric acid (1:7) with the addition of a few drops of sulphuric acid. The electrolytic precipitation of the copper follows without filtration, employing a small platinum cylinder of 66 sq. cm. total area, .and a small platinum wire spiral as anode. As finally placed, the flat spiral wire stands horizontally without touching the glass;- the cylinder is fastened to a special tripod stand. The electrolyses of shales are ordinarily arranged as a set of 16 assays to one working stand, or rack, the strength of current being .1 ampere per assay. When connected at evening, the extraction of the copper proceeds without attention. In the morning, the assay glass is washed with water and the assayer notes whether any copper deposits on the newly moistened por- tion of the cathode. If this is not the case, the cylinder is quickly withdrawn from the solution, then washed off with water and alcohol, dried in an air bath at 90 C., and weighed on the balance. (A color test by hydrogen sulphide is a more certain indication of the end-point than the one above described.) Raw Mansfeld Ores. With ore samples from the deep mines 2 grams are evaporated in a 14 cm. casserole on a sand bath, with 30 c.c. nitric acid (d., 1.2) and 20 c.c. sulphuric acid (d., 1.2) the heat being raised so that the separated sulphur burns. The residue is taken up with about 100 c.c. water and 10 c.c. of sulphuric acid, and then 4 c.c. of a normal solution of hydro- chloric acid is added for the separation of silver. After standing 12 hours, the liquid is passed through an ordinary 8 cm. filter into a beaker 15 cm. high and 8 cm. wide, washed with water, and a few drops of sulphuric acid added to render the lead in- soluble. The electrolysis of the filtrate (after the addition of 20 c.c. nitric acid and dilution to 400 c.c.) is carried out, gener- ally at night, with the aid of the platinum electrodes already described, and with a current of .3 ampere for each assay. For the estimation of Zn, Ni, and Co, in ^his material, refer to a later method (3, Chapter VII). 66 ANALYSIS OF COPPER Typolite Ore. This is treated like ordinary ore (q. v.) with the exception of a larger addition (8 c.c.) of yi^ normal hydro- chloric acid, on account of the higher silver contents. An excess of the acid in this case must be carefully avoided, or the copper will deposit in a spongy form. Current, .3 ampere per assay. 12. Desilverized Residues. In this oxidized product, con- taining about 70 per cent copper, it is not necessary to treat in a casserole, to burn off sulphur, etc. Two grams are digested with 60 c.c. nitric acid and 10 c.c. concentrated sulphuric acid until a white anhydrous deposit of copper sulphate is formed. After dilution with water to about 400 c.c., 10 c.c. ammonia is added to lessen the acidity, and the acid liquid electrolyzed with .4 ampere. The assays are usually combined in sets of 3 to 4, but the samples must be similar to insure an equal division of current. 13. Mansfeld Ore Slags. Five grams of powdered slag are mixed in a casserole with 30 c.c. of nitric acid and 15 c.c. of sulphuric acid and evaporated to dry ness on a sand bath. It has been proved that chilled slag is quickly decomposed, while stiff-tempered slag (i.e., gradually cooled) requires an alkali melt. The dry residue is treated hot with 100 c.c. water and 10 c.c. sulphuric acid, then filtered through an ordinary 12 cm. paper into a beaker (15 by 8 cm). The filtrate, after adding 20 c.c. nitric acid, is electrolyzed with platinum electrodes and a current of 0.11 ampere per assay. 14. Typolite Slags. Two grams are evaporated to dryness with 30 c.c. aqua-regia and 15 c.c. sulphuric acid. The residue is extracted with 100 c.c. of water and 10 c.c. sulphuric acid. The solution, after warming, is filtered into a beaker. Twenty c.c. nitric acid are added to the filtrate which is then treated with a few drops of cold saturated oxalic acid in order to overcome the harmful effect of the heavy iron contents of this material (40 per cent Fe) . Tests are arranged in sets with a current of .3 ampere per assay. The routine electrolyses are carried on over night. It is only in urgent cases that the process is finished in the daytime at the sacrifice of accuracy, in which case the liquids are warmed to 40 to 45 C. The author con- siders that as the local conditions require a simple assay, the use of mechanical stirrers is prohibited. Magnetic Separation. The original ore is peculiar to the COPPER IN ORES AND FURNACE PRODUCTS 67 district. A description of the slag produced has been necessarily condensed in the translation of the original paper. Most of the copper contents exist in such felag in metallic form. Some sul- phide and silicate are also present, hence it is erroneous to report all the copper as metal. Magnetic separation has accordingly been taken up for the successive mechanical division of the copper, as metal and sulphide, from the silicates. Magnetic treatment is very efficient in consequence of the considerable proportion of magnetic oxide of iron. By a small electro- magnet, connected with six Meidinger elements, pieces of nut size are attracted with ease. The separation can be only approximate as the slag grains will inclose, or be covered with, non-magnetic particles. Nevertheless, four treatments of typolite slag (containing 7.75 per cent copper and .161 per cent sulphur) yielded 15.9 per cent of non-magnetic residue which assayed nearly 33.8 per cent metallic copper and 3.48 per cent copper sulphide. Method of Assay. The total copper is determined by the usual methods. For the valuation of the metallic copper only, the most suitable method is the extraction of powdered slag with standard solution of silver nitrate, followed by the dry assay of the silver precipitated by the copper; and as a control, the titration of the silver remaining in the solution by potassium thiocyanate after the Volhard method. The copper sulphide and the metallic iron present (perhaps 1.25 per cent) also take full part in the reaction of the silver nitrate with the metallic copper and only the silicate remains, so that the relative proportions of copper metal, sulphide, and silicate are easily calculated. For a total of 7.75 per cent as noted above, about 6.85 per cent would exist as metal, .58 per cent as sulphide, and .32 per cent as silicate. NOTE 1 . The magnetically separated copper is itself argen- tiferous. It carries the silver excess of the original typolite slag and a notable amount of gold even after skimming off some half- melted ore which floats on the slag in the descending series of skimming pots at the furnace. The slag copper is enriched with the whole nickel content of . the ore, and copper bottoms are formed at the same time in the pots. When this same slag is, however, remelted in a furnace with regulus, the magnetic sepa- 68 ANALYSIS OF COPPER ration of the ground slag is hardly profitable in spite of the high iron content (nearly 38 per cent), for the reason that the iron is then present in the final slag as non-magnetic ferrous silicate. NOTE 2. When typolite slag is cooled in the series of pots, it forms zones or layers which increase in magnetic iron oxide towards the bottom and consequently increase in electrical con- ductivity. If the upper zone contains 1 to 3 per cent copper and to .003 per cent silver, the middle will be twice as rich while the lowest zone will show 4.5 to 5.4 per cent copper and .015 to .017 per cent silver. It was proved by a series of tests that it is impossible to obtain a proper average by ladle samples from the settling pots. CHAPTER V ANALYSIS OF ORES, SLAGS, MATTE, AND FLUE DUST CONTROL OF SMELTING FURNACES THE methods described in this chapter include only the determinations ordinarily required to furnish the data to metallurgists for regular daily control of furnace operations. Occasional tests for arsenic, bismuth, nickel, sodium, and rarer elements, are reserved for Chapters VI and VII. The methods of this chapter are arranged in the sequence adopted in practical work, rather than in alphabetical order. Classification. Furnace material, for the purposes of analy- sis, may be divided into two classes : First, material decomposed by acids, such as chilled blast furnace slags; second, refractory material, leaving an insoluble residue after acid treatment; for example, reverberatory slags, calcined ores, and flue dust. Such a residue evidently requires fusion. The methods are those of the largest western reduction works, except in cases where other methods are specified. DETERMINATION OF INSOLUBLE MATTER 1. In Raw Ores. For a special test on Butte ores, refer to the method for sulphur (16). As a general method, heat a half- gram sample in a small beaker, or casserole, for ten minutes with 10 c.c. of hydrochloric acid, add 1 to 5 c.c. of nitric acid (depending on the sulphides present), cover with a watch glass until violent action ceases, then uncover and evaporate to dry- ness on the steam bath. Sulphides may be treated with 5 c.c. of nitric acid and a little potassium chlorate, 10 c.c. of hydro- chloric acid added when the action becomes quiet, and the liquid evaporated to dry ness as before. When dry, remove from the bath, add 15 c.c. of hydrochloric acid and 50 c.c. of hot water. Bring to boiling and filter. Wash the insoluble matter with dilute (1:9) hydrochloric acid, then with boiling water. 70 ANALYSIS OF COPPER Fold the filter about the " insoluble," ignite, and weigh. This assay may be combined with the iron titration. Another operator 1 recommends a first treatment with 7 to 10 c.c. of nitric acid, followed by evaporation to dryness, and cooling. Then add 30 c.c. of (1:1) hydrochloric acid and heat until the solution is as complete as possible. In presence of carbonaceous matter, it may be necessary to bake for a long time, as such residues hold acid very tenaciously. Roasted Ores. Digest with hydrochloric acid, without boil- ing, until the oxidized part is dissolved, add about 3 c.c. of nitric acid to decompose sulphides, and dry as before. Barium Sulphate Ores. Treat with 10 c.c. of hydrochloric acid (1: 1), boil a few minutes, add 4 to 5 c.c. of nitric acid, and after action has ceased, dry and bake, then proceed as already indicated. After the " total insoluble matter" is weighed, fuse it with sodium carbonate, or mixed carbonates, 8 parts to 1 of residue, digest the fusion with water until disintegrated, filter, and wash. Wash out the crucible* with 5 c.c. of hydro- chloric acid (1: 1) and with this acid dissolve the residue upon the filter, being careful that it is all dissolved and the filter washed out clean. Precipitate the barium as sulphate from a boiling solution, and deduct this barium sulphate from the "total insoluble" to obtain a result defined by custom as "the insoluble residue." SILICA 2. In Chilled Blast Furnace Slags. To .5 gram of slag in a small porcelain casserole, add six drops of water and about 3 c.c. of strong hydrochloric acid. Stir well with a glass rod until all lumps are broken up and a smooth jelly results. Now add a few drops of nitric acid and work the jelly up and around the sides of the casserole to a height of about 1.2 cm. This will permit of very rapid dehydration, and will reduce the loss by "spitting." Place the casserole, uncovered, upon the plate and evaporate off all traces of acids, care being taken, however, not to prolong the baking at too high a temperature, or some alumina will unite with the silica and give too high a result when the direct weight of silica is taken; the residue, after ignition, being of a gray instead of a white color. 1 Western Chem. and Met. 3, 120. ORES, SLAGS, MATTE, AND FLUE DUST 71 Slightly cool the casserole, add about 20 c.c. of hydrochloric acid (d., 1.20), and boil for a few- minutes. Dilute with a little hot water and filter while hot, Cashing well with hot water until free from chlorides:' Ignite and weigh as silica. If exact results, or the percentage of true silica is required, additional precautions are taken as stated in the paragraphs on ores. In a statement of a "co-operative analysis," Thorn Smith 1 has indicated a necessary precaution. At least two evaporations to hard dryness with an intermediate filtration are required to render precipitated silica insoluble in dilute acids. When ac- curate work is attempted, the silica should finally be ignited ten minutes with the blast lamp. It is also necessary to correct, in such work, for the small amounts of oxides of iron or alumina, and sulphates of alkaline earths, which remain after treatment of the silica with sulphuric acid and excess of hydrofluoric acid. The manner of calculation of this correction to the original weight of siliceous residue depends on the forms in which the bases may be assumed to exist in the ignited siliceous residue. Platinum dishes are best suited to accurate work in complete analysis of a slag or ore. Solubility of Jena Glass. Thorn Smith and A. M. Smoot have proved that Jena glass, although heat resistant, is easily attacked by alkalies. This undesirable property is due to the presence of oxides or silicates of zinc as a constituent. Special Solvents. For complete solution, Smith decomposes Tennessee cupola slags by digestion in a platinum dish with a mixture of hydrofluoric acid and nitric acid. To dissolve the slag for an iron titration only, hydrochloric is substituted for nitric acid. Another solvent used for ores and slags is the " chlorate mixture" (1 part sulphuric acid, 1 part nitric, and 1 part of a saturated solution of potassium chlorate in nitric acid). 3. In Lake Superior Slags. Waste cupola slags are chilled by granulation in water at the furnace, the sample dried, then crushed and divided to obtain a 25- to 50-gram sample for assay which will pass a sieve of 100 meshes to the linear inch (39 per cm.). The copper prills, remaining on the sieve, are separately 1 Eng. and Min. Jour. 75, 295. W. F. Hillebrand, J. Am. Chem. Soc. 24, 362. Bull. 422, U. S. Geol. Survey. 72 ANALYSIS OF COPPER weighed. In accurate complete analysis, the free iron is esti- mated by magnetic separation according to the principle outlined in Chapter II, under " Correction for iron from grinder." In some slags this iron contains copper. The total copper is determined most rapidly by method 6, Chapter IV, but may also be estimated in the filtrate from the silica. Silica. Dissolve 2 grams of slag in a 300 c.c. casserole by boiling with a mixture of 15 c.c. distilled water, 15 c.c. strong nitric acid, and 10 c.c. of sulphuric acid (d., 1.84) until white fumes appear and the residue, when gently rubbed, has become a smooth paste. Partially cool the residue, then add 90 c.c. of distilled water, wash the glass cover and replace it. Boil for a few minutes until the soluble matter is dis- solved, transfer to an electrolytic beaker, and place on the battery stand for the estimation of copper unless silica is the only constituent to be determined, in which case the solution is filtered at once. For direct estimation of silica, boil the residue with 75 c.c. of (1:2) hydrochloric acid, decant through an 11 cm. filter placed in a platinum cone over a suction flask. Then boil the residue with 50 c.c. of (1:1) hydrochloric acid to extract calcium sul- phate, dilute to 75 c.c., filter, wash, ignite, and weigh. Treat the residue with hydrofluoric acid and one drop of sulphuric after moistening with water. Ignite and re weigh. Fuse with potas- sium pyrosulphate, dissolve, and add to main solution. If a test for barium is advisable, separate it from the original residue by fusion with sodium carbonate amd proceed as in method 1, under the head of "Barium Sulphate Ores." SILICA IN REFRACTORY MATERIALS 4. In Reverberatory Slags (also Calcines, Briquettes, and Flue Dust). Such material yields an insoluble residue requir- ing fusion. Reverberatory furnace slag, from smelting of ores, contains such a small amount of copper that it can be fused without any preliminary treatment. With other refractory products, weigh .5 gram, place in a small casserole, and add 6 c.c. of hydrochloric acid and 3 c.c. of nitric acid. Cover the casserole, heat for a few minutes on the hot plate, bring to boiling, dilute with boiling water, and filter ORES, SLAGS, MATTE, AND FLUE DUST 73 through a fine paper into a larger casserole. Wash the residue on the paper three times. Fold the filter about the residue and ignite it in a cup or crucible. Place the filtrate on a hot plate while the fusion is being made. To fuse the slag or ignited oxides, mix them well with 6 to 8 grams of anhydrous so- dium carbonate in a platinum crucible (or a dish if preferred), and then cover the mixture with 1 or 2 grams of sodium car- bonate. Place on a scorifier, preferably, and fuse in a muffle. Raise the heat gradually from dull to bright red, and keep at a red heat for 15 minutes. Dip the crucible in water to cool. Partly fill with water and warm for a few minutes on a hot plate. The cake, as a rule, can be loosened from the crucible by this treatment. Wash it into a casserole containing the main solution which has already been evaporated to dryness. Cover, then add 10 to 30 c.c. of hydrochloric and 1 c.c. of nitric acid. If the fusion was made in a dish, cool, add water, then add 10 c.c. of hydrochloric acid and warm until the contents can be easily transferred to the original casserole. If the color of the fusion indicates the presence of much manganese, it must be removed from the dish without the use of acid, as there is danger of liberation of chlorine. Evaporate the contents of the casserole to complete dryness again, but do not bake long above 115 C., as some silica is liable to recombine with the alumina. 1 When dry, remove from plate, add 10 c.c. hydrochloric acid and 30 c.c. of water, and boil for a few minutes. Filter, wash ten times, ignite, and weigh. Test the residue for purity, by the usual, treatment with hydrofluoric acid, and ignition. For very accurate results increase the sample weight. ALUMINA 5. Alumina in Ores and Slags. Decompose the sample, dehydrate, and filter off the silica as in the silica determinations. Purify the silica by evaporation with a drop of sulphuric acid and excess of hydrofluoric acid and fuse the residue with a little potassium pyrosulphate, if necessary, to bring it into solution. In the filtrates, determine the alumina by the Western phosphate method, to be described. 1 Methods of Rock Analysis, by W. F. Hillebrand; also J. Am. Chem. Soc. 24, 362, and Bull. 422, U. S. GeoL Survey. 74 ANALYSIS OF COPPER Instead of treating with hydrogen sulphide, ferric hydroxide and alumina may be precipitated with ammonia. The mass is dissolved in hydrochloric acid, the solution diluted to 400 c.c., and the alumina precipitated as before. This serves to separate the alumina from the small amount of copper contained in such waste ore slags. The other metals affected by hydrogen sulphide are usually present in such small quantities that they may be neglected. In making the ammonia separation, a large excess of ammonium chloride must be present and the solution boiled for fifteen minutes to break up any aluminates which may have been formed. Precipitation. After the excess copper has been removed, dilute the acid filtrate from the silica with cold water to about 400 c.c., add 30 c.c. of a 10 per cent solution of ammonium phosphate, and then dilute ammonia until a slight permanent deposit forms. Now add 1.5 c.c. concentrated hydrochloric acid and 40 c.c. of a 20 per cent solution of sodium thiosulphate and heat to boiling. When boiling has continued a couple of minutes, add 15 c.c. of a 20 per cent solution of ammonium acetate and 6 c.c. of strong acetic acid, and boil about 15 minutes longer. This addition of ammonium acetate and acetic acid after boiling gives a much more granular precipitate, which, after allowing it to settle for about 20 minutes and decanting the clear superna- tant fluid, filters very rapidly. Wash with hot water ten times, dry, ignite gently at first, and weigh as A1 2 03, P2O 5 41.85 per cent of which is alumina (A^Oa). Instead of the previous method, the following plan may be adopted for the removal of the copper and similar metals. Should metals affected by hydrogen sulphide be present, add ammonia to the filtrate from the silica until the liquid is nearly neutral, then add 2 or 3 c.c. of hydrochloric acid in excess, reduce the solution with sodium sulphite, and boil off the excess of sulphur dioxide. Add 15 c.c. of hydrochloric acid, pass hydrogen sulphide through the solution, filter off and wash the sulphides with cold water. Boil off all hydrogen sulphide from the filtrate, dilute to 400 c.c. with cold water, and proceed as before. IRON OXIDES 6. Direct Titration of Ores. Oxidized iron ores (carrying copper) often dissolve more rapidly if a little stannous chloride ORES, SLAGS, MATTE, AND FLUE DUST 75 is added with the hydrochloric acid. If the tin solution is added to aid the solution of the iron ? then the reduction of iron before titration must be made with stajmous chloride (as tin interferes with a solution reduced with lead), or an excess of mercuric chloride should be added to the solution after decanting from lead. To the filtrate, add 5 to 10 grams of test lead and boil until colorless. Decant from the lead (filtration often being neces- sary), and wash the lead very thoroughly. Cool, add 10 c.c. of hydrochloric acid, and titrate with potassium bichromate, if preferred. If the solution appears yellowish after decanting from lead, add 2 drops of stannous chloride solution and mer- curic chloride in excess before titration. When copper, arsenic, and antimony are absent, heat the filtrate from "insoluble" nearly to boiling, reduce as before, cool, and titrate. Separation of iron from copper by ammonia is not complete in one treat- ment, however. The ferric and aluminum hydroxides, with the filter, are placed in a beaker, 50 c.c. of boiling water added, and the solution titrated as already described. If care is taken not to add the stannous chloride in large excess, the mercuric chloride may be added to the hot liquid. It is then diluted with cold water and titrated. If a large excess of stannous chloride is added by accident, run in potassium permanganate solution from a burette until the iron solution is a pale yellow; add one or two drops stannous chloride in excess, then mercuric chloride, and proceed as before. 7. Lake Superior Method. Dissolve 2 grams of sample in acids, evaporate to fumes with 10 c.c. sulphuric acid and dilute. (See method 5, Chapter IV.) Two precipitations of iron by ammonia do not effect a complete separation. The copper may be very rapidly and exactly removed by placing the solution in an electrolytic beaker (see electrolytic assay of copper), and passing 3 to 4.5 amperes of current through the solution after placing the beaker in a Frary solenoid, or rotary device. To get a bright deposit, add 5 to 8 c.c. nitric acid and keep the solution cold. When testing for the end-point of electrolytic deposition, preserve the test portion, and return it to the main solution after the plate has been withdrawn. The sulphur should first be destroyed by heating the test portion with a little potassium permanganate. The time of electrolysis for a 2-gram sample is 30 to 60 minutes. 76 ANALYSIS OF COPPER For the exact estimation of iron it is necessary to filter the original solution (or electrolyte) and decompose the washed insoluble matter by fusion. Combine the two solutions, oxidize the iron, precipitate twice with ammonia, redissolve the iron in 10 c.c. hydrochloric acid (1:1), and wash the paper. Reduce as above, destroying the slight excess of stannous chloride by the addition of 5 c.c. of a saturated solution of mercuric chloride. Dilute to 400 to 500 c.c. and titrate with potassium perman- ganate (39, Chapter III). Add 10 to 20 c.c. of titrating solution 25 to prevent the formation of a yellow iron coloration. One c.c. of permanganate oxidizes about .01 gram of iron, Fe. If a test of the electrolyte is to be made for iron only, it may be evaporated to fumes to remove nitric acid. Add 30 c.c. of water and 10 of hydrochloric acid, heat to boiling, then add sufficient potassium permanganate to produce a yellow color and destroy a trace of sulphur. Reduce as before and titrate, after dilution with cold water to 400 c.c. Calculate to the compound in which the iron exists in the ore. 8. Iron in Slags. The percentage of iron oxide, FeO, is found by titration of a 0.5 to 2 gram sample. Chilled waste slags from Lake Superior furnaces may be dissolved as directed in method 3 for silica, and the copper removed by rapid electrolysis. Boil the solution down to white fumes after electrolysis to re- move nitric acid. Add 30 c.c. of water and 10 c.c. of hydrochloric acid, heat to boiling, and add a little potassium permanganate, if necessary, to color the solution yellow and destroy any sulphur compound. Reduce and titrate as with ores. For ordinary furnace tests it is not necessary to remove the copper. Proceed as in the next paragraph. Ore slags, from the blast furnace, are dissolved in a similar manner. To .5 or 1 gram of slag in a beaker, add about 50 c.c. of boiling water, and then while stirring to keep the slag in suspen- sion, pour in 10 to 15 c.c. of hydrochloric acid (d., 1.2). Boil for a few minutes until the solution clears (any slight residue of coke dust being neglected). Oxidize to yellow color with per- manganate. Reduce with stannous chloride in very slight excess, followed by mercuric chloride in excess as already mentioned in 6, and titrate with either potassium bichromate, or permanga- nate, as preferred. ORES, SLAGS, MATTE, AND FLUE DUST 77 With the dichromate, use potassium ferricyanide as an indi- cator, placing the test portions in contact on a white spot plate. t CALCIUM ANQ< MAGNESIUM 9. Direct Method for Lime. Precipitation of calcium oxalate in oxalic 'acid solution, in presence of the other bases, is used in the western States as a special test for lime only. To the filtrate from the silica, add ammonia in excess and then oxalic acid, little by little, until the hydroxides just dis- solve. Make the solution slightly alkaline once more, and then redissolve the iron by adding oxalic acid in slight excess. The solution should now be of a light apple-green color. Boil well for a few minutes, filter, and wash well with hot water until free from oxalic acid, six or seven times usually being sufficient. Drop the filter and calcium oxalate into a beaker containing 150 c.c. of boiling water. Add 10 c.c. of dilute (1:1) sulphuric acid and titrate at once with potassium permanganate (solution 40, Chapter III). The permanganate may be standardized with chemically pure sodium oxalate which should be dried at 100 C. before use and preserved in a small glass-stoppered bottle. Na 2 C 2 O 4 x .41843 = CaO. Correct sodium oxalate may be ob- tained from the U. S. Bureau of Standards. 10. Calcium and Magnesium (in chilled cupola slags) . To the filtrate from a silica determination add an excess of ammonia and 5 grams of ammonium chloride (prepared from chemically pure reagents), then 10 c.c. of bromine water and about .1 gram of ammonium persulphate, as in the "Estimation of zinc," 15, Chapter VII. Bring to a boil, filter, dissolve the precipitate in dilute hydrochloric acid, add bromine water and ammonium persulphate as before, heat to boiling and filter again. After washing the hydroxides, precipitate the calcium (in the combined filtrates from iron, alumina, and manganese), with ammonium oxalate, using .5 to 1 gram of the dry salt or 20 c.c. of the saturated solution. Filter, wash well, and in the filtrate, separate the magnesia by stirring five minutes with an excess of ammonium phosphate (6 and 51, Chapter III). Have the solution strongly ammoniacal with about one-fourth its volume of ammonium hydroxide (d., .90) to prevent the pre- cipitation of zinc. Dissolve the precipitate in a very small amount of hydrochloric acid, add a small crystal of the phos- 78 ANALYSIS OF COPPER phate, then ammonia drop by drop, until alkaline, stirring vigorously. Wash with . dilute ammonium hydroxide (1 :10) containing 10 per cent of ammonium nitrate, ignite gently at first, and weigh when cooled. Magnesium pyrophosphate, Mg 2 P2C>7, x 0.36207 equals magnesium oxide, MgO. In the absence of manganese, omit the bromine water and persulphate. In some Lake Superior slags which contain much lime, the calcium oxalate should be dissolved and reprecipitated to obtain a perfect separation from magnesium salts. About 75 c.c. of water and 5 c.c. of sulphuric acid are finally used to dissolve the calcium oxalate for titration. Place the filter and contents in the original beaker, cover with the dilute acid, and heat nearly to boiling. A few drops of manganous sulphate in the solution causes the permanganate to act more quickly. 11. Calcium and Magnesium (in reverberatory slags and ores). Precipitate hydroxides with ammonium hydroxide and wash six times, then dissolve in dilute hydrochloric acid and titrate the iron. The calcium oxide in these refractory samples is usually below 6 per cent, and only .2 to .4 per cent of calcium oxide would be recovered by redissolving and reprecipitating the hydroxides. With higher percentages, a second separation with ammonia must be made even in routine technical work. Rich oxidized slags from Lake Superior furnaces treating concentrates of native copper, often contain as much as 5 per cent of free shot copper which must be separated and weighed during the grinding and sizing of the 25- to 50-pound sample. The insoluble portion of the slag is fused, preparatory to analysis for bases. After the slag has been totally decomposed with the exception of silica, the bases are determined as customary with blast furnace slags. Magnesium is determined by precipitation with phosphates, as in (10). 12. Sodium and Potassium Oxides. The determination of these elements is seldom necessary in furnace control. Refer to Chapter VII (6, 7, 9). SULPHUR IN CRUDE PRODUCTS Principle. Sulphur, or sulphate, is usually estimated by oxidizing the sulphur to sulphuric acid, then precipitating with barium chloride in a pure hydrochloric acid solution. ORES, SLAGS, MATTE, AND FLUE DUST 79 Lake Superior copper deposits contain but a trace of sulphur, excepting in a few small cross- veins. The Western methods, to be described, are the same as , those used in the eastern States and Europe for all furnace produces. 13. Sulphur in Roasted Ores. Treat .5 gram, or f the sulphur weight (.687 gram), with 10 c.c. of a mixture of nitric acid and potassium chlorate. Evaporate to dry ness on a steam bath. Take up with 5 c.c. of hydrochloric acid and 15 c.c. of water. Boil for a few minutes and filter off the insoluble mat- ter. Bring the filtrate to a boil, add barium chloride solution, BaCl2.2H 2 O, using an amount equivalent to 1 gram of the crystals for calcines and 2 grams for ores. Allow the barium sulphate to settle, filter, and wash 6 to 10 times with boiling water. Dry, ignite at a dull red heat, and weigh the cooled crucible and barium sulphate. Barium sulphate (BaSO^ x 0.13735 = sul- phur (S). If accurate work is required, the solution must be repeatedly evaporated with excess of pure hydrochloric acid in order that the solution may be free from all nitrates and chloric acid before the sulphuric acid is precipitated. To avoid errors which are introduced by direct precipitation in presence of iron, proceed as follows: After repeated evapora- tions with hydrochloric acid, dissolve in the dilute acid and filter. Heat the filtrate nearly to boiling, make ammoniacal, and then add an excess of barium chloride. Now make the liquid slightly acid with dilute hydrochloric acid, allow to settle in a moderately warm place, but do not boil. Filter, wash very thoroughly, dry, and ignite. Good results in presence of iron may also be obtained by adding the barium salt in a cold solu- tion and allowing to stand, cold, for 12 hours before filtra- tion. Compare with the semi-electrolytic method, 20. 14. Sulphide Ores and Matte. Add 10 to 15 c.c. of nitric acid (d., 1.36) to a .5 gram sample. Place on the front of the hot plate and add small quantities of potassium chlorate at frequent intervals. When the sulphur is all in solution, remove the watch glass and evaporate to dryness. Then proceed as in the previous method. The solution of sulphur in some mattes is difficult and requires prolonged treatment with nitric acid and potassium chlorate. The following scheme may hasten matters in such 80 ANALYSIS OF COPPER cases: Sprinkle 1 gram of potassium chlorate on .5 gram of sample, add 10 to 15 c.c. of water, and bring to a boil. Add 10 to 15 c.c. of nitric-chlorate mixture and heat to boiling, keeping the solution well covered until all the sulphur is dis- solved. Remove the watch glass, evaporate to dryness on the steam bath, and proceed as before. One gram of barium chloride crystals will precipitate 23.4 per cent of sulphur on a half-gram sample, and may be added in solution. The purest barium sulphate is obtained by having the solution of sulphates very dilute, with only a few c.c. of hydro- chloric acid in excess, and by adding the barium chloride drop by drop, while stirring. 15. With Sulphates. For the determination of total sulphur in ores containing barium sulphate or large quantities of calcium sulphate, one may employ a fusion method (sodium carbonate and potassium nitrate) with sodium chloride. 1 16. Sulphur in Lead Ores. Proceed as before, after evap- oration to dryness. Take up the residue with 5 c.c. of hydro- chloric acid and 5 grams of ammonium chloride and 15 c.c. of boiling water. These reagents, on boiling, will dissolve any lead sulphate which may have formed during evaporation. 17. Sulphur in Ores Carrying Zinc. Beckman's method consists in fusing .5 gram of ore with 25 grams of a (6:1) mix- ture of sodium carbonate and potassium chlorate. The sintering process of Waring for heavy zinc ores is of a similar nature. 18. The Sintering Process of F. G. Hawley is more simple than that of Waring. For decomposition, mix .5 gram of ore thoroughly with six to eight times its weight of a mixture of zinc oxide and sodium carbonate (4:1); sinter at a low red heat for 15 minutes in a porcelain crucible, leach with warm water, and filter. Acidulate the filtrate with hydrochloric acid, add 5 c.c. in excess, and add by degrees to the boiling solution a sufficiency (about 20 c.c.) of a semi-saturated solution of barium chloride. Slags may be similarly treated. 19. Sulphur hi Heavy Pyrites. The next two modifications are two new devices of New York chemists for the solution of rich copper-bearing pyrites and for the elimination of the injuri- ous effect of ferric chloride. (Compare method 13.) 1 Fresenius, Quantitative Analysis. ORES, SLAGS, MATTE, AND FLUE DUST 81 Allen and Bishop 1 recommend solution in a mixture of bromine and carbon tetrachloride, and reduction by powdered aluminum as follows: 1.3735 grams of ore are placed^in a dry 300 c.c. Jena beaker, 10 c.c. of a mixture of 2 parts of liquid bromine and 3 parts of carbon tetrachloride by volume are added, and the covered beaker shaken gently for 15 minutes at room temperature. Then 15 c.c. of nitric acid (1:4) are added and the mixture allowed to stand 15 minutes longer, after which the beaker is heated up gradually on the steam bath and the solution taken to dryness. Ten c.c. of hydrochloric are next added, the solution evaporated again, and the silica thereby dehydrated. Filter and reduce the ferric chloride by gradual addition of powdered aluminum, free from sulphur. Precipitate the barium sulphate in the cold by adding a 5 per cent solution of barium chloride at the rate of 5 c.c. per minute, the total volume of the solution being 1600 c.c. 20. Semi-electrolytic Method for Sulphur. The next method quoted is more accurate than 19, because the volume of the solution is kept very small, thus favoring higher results, and the interfering metals are nearly all eliminated from the solution. 2 The ore should be ground just to pass a sieve of 80 meshes to the linear inch. .5 gram of dried ore (or .687 gram = J factor weight) is placed in a 250 c.c. beaker with a mixture of 3 parts nitric acid (d., 1.42) and 1 part of hydrochloric acid to which 4 or 5 drops of bromine have been added. Cover the beaker tightly, and allow it to stand at room temperature for one half-hour. Transfer to a steam bath, heat gently until action ceases, then raise the cover, and evaporate to dryness. Add 5 c.c. of hydro- chloric acid to the residue, heat until action ceases, wash the cover, and evaporate to dryness. Treat with hot water until residue is disintegrated, and wash the solution into an electro- lytic beaker, containing a mercury cathode with insulated wire connection. Dilute to 75 c.c., pass a current through a platinum spiral anode to the mercury for 5 to 6 hours at .8 to 1 ampere, or at a lower rate overnight. All metals pass into the mercury. Siphon, or pour off, the colorless liquid, wash by decantation four times with 25 c.c. of water, then pour off mercury into 1 Eighth Inter. Congress of Appl. Chem. 1, 38. 2 A. M. Smoot, Eng. and Min. Jour. 94, 412; J. Am. Chem. Soc., 25, 911. 82 ANALYSIS OF COPPER small beaker, and wash the beaker once (especially the lip) with water. Filter into an 800 c.c. beaker and wash 4 to 5 times with hot water. Dilute to 450 to 600 c.c., heat to boiling, and add to boiling liquid 25 c.c. (or 34 c.c. for .687 gram ore) of a 10 per cent solution of barium chloride in a thin stream. The barium sulphate contains impurity amounting to .06 per cent to .09 per cent on .5 gram of pyrites, but the correction for solu- bility nearly balances this, leaving a final plus error of .01 to .04 per cent of sulphur, as calculated. The objection to this method for a large amount of work is the necessity for electrolysis and the treatment and purification of all the mercury. It seems well adapted to umpire assaying. Compare method 13. CHAPTER VI SPECIAL ELEMENTS IN ORES, SLAGS, AND MATTE ARSENIC 1. Distillation (Method of Skinner & Hawley). A distilling solution is required, (a) Dissolve 300 grams pure cupric chloride crystals in one liter hydrochloric acid (d., 1.20) (solution 17, Chapter III), (b) Dissolve one pound (453.6 grams) zinc (free from arsenic) by adding to it gradually a mixture of 1250 c.c. hydrochloric acid (1.20) and 500 c.c. of water. When the zinc has dissolved, evaporate the solution to 1100 c.c. (solution 58, Chapter III). Mix (a) and (6). The principle on which the assay depends is the evolution of arsenious chloride directly from the sulphide by distillation in a saturated solution of chlorides of copper and zinc. Analysis. Add 5 c.c. nitric acid to a half-gram sample in a small beaker. (A little potassium chlorate is added with flue dust samples. If desired, Low's method of decomposition may be used as described in method 7.) When the action becomes quiet, add 6 to 10 c.c. hydrochloric and evaporate to complete dryness on steam bath. As there is danger of loss, the tempera- ture should be low. Take up with 5 c.c. hydrochloric acid and 25 c.c. of water. Bring to a boil and filter off the insoluble mat- ter. Dilute the filtrate to 200 c.c. with boiling water. Add sufficient sodium sulphite to render the solution colorless. Boil off the excess of sulphur dioxide and add 15 c.c. hydrochloric acid. Pass a current of hydrogen sulphide through the solution until it is saturated. Filter off the precipitated sulphides, wash out the iron salts, and place the sulphides in a flask connected to an 8-inch (20 cm.) Allihn condenser, with large straight bulb tube, set vertically. Never allow the lower end of condenser to be more than sealed by the water in beaker (200 c.c.). Add 50 c.c. of " distilling solution" and distil carefully until the thermometer reads 115 C. Remove the flask from the heat and 84 ANALYSIS OF COPPER add 25 c.c. of hydrochloric acid. Distil again until the ther- mometer reads 115 C. Pour distillate into a No. 3 beaker, make alkaline with ammonia (about 25 c.c. being required), just acidify with dilute hydrochloric acid. Cool, add 2 to 4 grams of bicarbonate of soda and a little starch solution, and titrate with standard iodine solution. (Use 20, Chapter III.) On samples low in arsenic, 1 to 5 grams of sample may be taken. Rapid and complete precipitation of arsenic, in samples containing only small quantities of other hydrogen sulphide metals may be affected by adding 100 mg. of pure copper to the sample. Antimony and tin may be estimated in the residue from distillation. 2. In Slags (Skinner & Hawley). Treat 5 to 10 grams with 10 c.c. nitric acid, 10 c.c. hydrofluoric, and 2 c.c. sulphuric acid. Evaporate to sulphuric fumes. Take up with 10 c.c. hydro- chloric acid and 25 c.c. water, then boil and filter. Dilute fil- trate to 500 c.c. Almost neutralize with ammonia. Reduce with sodium sulphite and proceed as before. Hydrofluoric acid is not necessary in the case of chilled slags. 3. Sintering Method (F. G. Hawley). Mix .5 to 1 gram of ore with six to ten parts of an equal mixture of zinc oxide and sodium carbonate. Sinter in a porcelain crucible for 15 to 20 minutes. Start at a low red heat and increase to full redness. Leach with hot water and filter. Boil solution carefully, neutral- ize with nitric acid, and add just 4 drops excess, using litmus paper as an indicator. See that any alumina or zinc oxide, that may have run through the filter, is dissolved. Boil the solu- tion to remove carbonic acid gas, remove from the hot plate, and add a solution of silver nitrate. Seven-tenths gram of silver nitrate is sufficient for .1 gram of arsenic. No red pre- cipitate should be visible. If any appears, add a few drops of nitric acid until dissolved. Now add about 1 gram of sodium acetate, and stir rapidly. Let stand for 20 minutes, filter, and wash. Dissolve the silver arsenate through the filter with dilute nitric acid, dilute the solution, and titrate with a standard solution of ammonium thiocyanate, using ferric sulphate as an indicator. (Solution 8, Chapter III.) 4. Evolution as Arsine l (Modification of Cobeldick) . For 1 F. W. Schmidt, J. Anal and Appl Chem. (1892), 408; Heath, Eng. and Min. Jour. 63, 663. ELEMENTS IN ORES, SLAGS, AND MATTE 85 determination of traces of arsenic, 2 grams of ore are weighed into a 150 c.c. beaker, 10 c.c. of strong nitric acid added, covered with a watch glass, and allowed tg digest all night in a warm place. The liquid is diluted to 100 c.c. and excess ammonia added, boiled, cooled, filtered, and washed in cold water. The hydroxides are dissolved in dilute sulphuric acid (1:10), evapo- rated dry, and fumed on a silica plate. The mass is taken up with water, and a little sulphuric acid if necessary, the bulk of solution being about 50 c.c., warmed to dissolve all soluble mat- ter and transferred to a Marsh apparatus which is generating hydrogen from pure zinc and sulphuric acid or hydrochloric acid. (A better reduction is secured by adding, also, .5 c.c. of 80 per cent solution of stannous chloride.) The arsenic mirror is concentrated in the constricted portion of the tube and is com- pared with standard mirrors made with known amounts of arsenic brought into the solution (in presence of a trace of iron salt). If the material contains more than .1 per cent of arsenic, the portion of the tube containing the mirror is cut off, weighed, the tube cleaned, and re weighed. (Ericsson proposes to con- duct the gas through .1 normal silver nitrate, add excess hydro- chloric, filter, neutralize, and titrate with .002 normal iodine solution.) Antimony reacts also, if present. 1 Gutzeit Test. Allen & Palmer have recently presented a modified Gutzeit test which may be useful for continuous work in the estimation of traces of arsenic. 2 5. Nitrate Fusion. 3 Fuse 1 to 5 grams of the ore with six to 'ten times its weight of an equal mixture of sodium and potassium nitrates, and precipitate the arsenic from the aqueous solution with silver nitrate. The crucible is, however, con- siderably attacked. 6. Semi-fusion with Potassium Bisulphate. The following method of A. H. Low 4 is remarkably accurate for ores which can be thus decomposed. The first, or distillation, method is so much quicker that it seems to be preferred in the largest works, although it may be combined with the method of solu- tion described in the next paragraph. 1 Svensk Farm. Tidskrift, 18, 473 (1914). 2 Eighth Inter. Congress Appl. Chem. 1, 9. 3 Method of Dr. R. Pearce. 4 J. Am. Chem. Soc. 28, 1715. 86 ANALYSIS OF COPPER Digest .5 gram of ore with 7 grams potassium bisulphate, .5 gram tartaric acid, and 10 c.c. strong sulphuric acid, to a clear melt, sulphur-free. After cooling, the melt is dissolved in 60 c,c. of (1: 5) hydrochloric acid with 2 to 3 grams tartaric acid, diluted and treated with hydrogen sulphide. Sulphides are dissolved in potassium sulphide and filtered into a 300 c.c. flask. Three grams bisulphate and 5 c.c. sulphuric are added, the mixture boiled down again, cooled, and taken up with 75 c.c. of (2:1) hydrochloric acid and hydrogen sulphide passed again. Moisten the filter with acid of the same strength, and wash with the mixture. The arsenic is now on the filter, and antimony and tin should be in the filtrate. From this point the three are treated separately. First, the arsenic is determined by dis- solving the sulphide in a little ammonium sulphide, washing into a 300-c.c. Jena flask, and reducing with complete expulsion of sulphur by boiling vigorously with 2.5 grams of potassium bi- sulphate and 5 to 10 c.c. of sulphuric acid (d., 1.84). Titrate the cold aqueous solution of the paste by iodine as directed for "arsenic in copper," Chapter XII. Use iodine solution 20, Chapter III. It is necessary to boil the solution in the Jena flask very hard, after the water is expelled, in order that the acid fumes may carry all sulphur out of the neck of the flask. The solu- tion should be titrated within a few hours after reduction. One cubic centimeter of iodine = .005 gram arsenic. ANTIMONY 7. Antimony in Ores is finally estimated, according to Low, by diluting the filtrate from arsenic (6) with four parts of water, separating the antimony as sulphide by treatment with hydrogen sulphide, and reducing by digestion in a Jena flask, in the same way as for arsenious sulphide, but with double the quantity of potassium bisulphate and sulphuric acid. Expel all the sulphur and most of the free acid, cool, add 50 c.c. of water and 10 c.c. of hydrochloric acid. Dilute to 200 c.c. with cold water and titrate with standard potassium permanganate. Multiply the iron value by 1.0751, or the oxalic acid value by .9532 to obtain the titer for antimony. 8. Antimony may also be estimated by dissolving the ores, or slags, etc., in a mixture of acids, filtering and fusing any ELEMENTS IN ORES, SLAGS, AND MATTE 87 residue with carbonates and nitrate of sodium, then separating the arsenic and antimony frojn" the precipitated sulphides by treatment with dilute sodium sulphide solution. Evaporate the alkaline solution to dry ness on the water-bath or steam plate. Then oxidize the sulphur by digestion with red, fuming nitric acid, evaporate to dryness again, and dissolve the salts in 25 c.c. of water, or more. Add an amount of hydrochloric acid (d., 1.2), equal to twice the volume of the solution of the arsenic group, and precipitate the arsenic alone by hydrogen sulphide. Filter, testing the filtrate again. Wash with acid of the same strength. Then dilute the filtrate with 4 parts of water and precipitate the antimony, and tin if present, observing the pre- cautions of method 3, Chapter XII. Antimony may likewise be dissolved out of the mixed sulphides as in 10. Tin may also be separated from antimony according to method 11, the arsenic having been previously removed as just described. 9. Antimony by Electrolysis. Simple ores are decomposed by digestion with acids or by Low's process. Refractory ores may be decomposed, after Hawley's formula, by mixing 1 to 2 grams of ore with 8 to 10 parts of a (1:1) mixture of sodium carbonate and flowers of sulphur and heating slowly in a covered porcelain crucible for 15 minutes, finishing at a moderate red heat. Cool with the cover on, leach with hot water, and boil for five minutes. If the precipitate does not settle readily, or the solution appears green, add 2 to 4 grams of sodium sulphite, and boil again. Make up to 200 c.c. and pass through a dry filter into a dry beaker. Remove 100 c.c. to a 300 c.c. beaker, acidify with acetic acid, using 10 c.c. in excess, and boil for one minute. For accurate work, the filtered sulphides should be dissolved in ammonium sulphide (or sodium and potassium sulphides in presence of copper), and precipitated a second time with acetic acid. The antimony may be dissolved out by boiling (1 : 1) hydro- chloric acid and titrated, ignited with nitric acid in a porcelain crucible and weighed, according to Fresenius, as tetroxide, or the compound may be electrolyzed, if the antimony is present in quantity. Electrolysis. Dissolve the sulphides in 15 c.c. of sodium monosulphide (solution 52). W. B. Price recommends a solu- tion of 1.18 specific gravity, or density. Dilute to 70 c.c., add 88 ANALYSIS OF COPPER 3 grams of potassium cyanide, and electrolyze in a Frary rotary apparatus (or solenoid), with a current of 6 amperes and an electrode tension of 4 volts. TIN (WITH ARSENIC OR ANTIMONY) 10. If Tin is present in quantity, evaporate the filtrate from arsenic to dryness on the steam plate after adding enough potassium chloride to form a double salt with the tin and anti- mony. Redissolve by boiling for one hour, if necessary, with a mixture of 5 grams of ammonium oxalate and 5 grams of oxalic acid dissolved in 100 c.c. of water (a solution proposed by G. W. Thompson) . Precipitate the antimony from the hot liquid by gaseous hydrogen sulphide and filter the hot solution. Dissolve the sulphide in a very little hot dilute sodium sulphide, pour into a boiling solution of half the former quantity of the oxalic mix- ture, and repeat the precipitation and filtration. Boil out the hydrogen sulphide, and electrolyze the combined filtrates for tin. The solution should be kept in circulation by a slow stream of air, or by a revolving anode. Time 2.5 hours, and current .8 ampere per sq. decimeter. This procedure gives a much better separation of tin and antimony than is possible with the original method of Clarke. Refer to the special volumetric method, 12, Chapter VII and to " Antimony," Chapter XII. BISMUTH 11. Combination Assay. An approximate test of ores and slags may be made by fire assay as for lead, adding enough pure lead to obtain an average button. This button is then flattened, dissolved in dilute nitric acid, the solution made up to a definite volume with dilute sulphuric acid, shaken, and a known portion quickly filtered through a dry filter into a calibrated flask, and sub- sequently evaporated to fumes of sulphur trioxide. The "den- sity" of the lead sulphate is taken as 4-f nearly, or the volume of the sulphate from 75 grams of lead = 16.875 c.c. (L. G. Eakins) . The volume of the lead precipitate is deducted from the total volume. After the residue from evaporation has been taken up with water, any lead sulphate is filtered out, hydrogen sulphide is passed in for 10 to 15 minutes, the sulphides filtered off, and ELEMENTS IN -ORES, SLAGS, AND MATTE 89 extracted with yellow potassium sulphide. Dissolve the insolu- ble sulphides in fuming nitric acid, and evaporate again with 3 to 5 c.c. of sulphuric acid. Filter and wash again, make the liquid slightly alkaline with sodium carbonate, and add a few drops of potassium cyanide. Boil, allow to settle, filter on a fine paper, wash with warm water, and redissolve in a little nitric acid. Finally, separate the bismuth by precipitation with an excess of ammonia and ammonium carbonate, ignite very care- fully in a porcelain crucible at a low red heat, avoiding undue reducing conditions, and weigh as bismuth sesquioxide, Bi 2 3 . 12. Bismuth (Accurate Analysis). Decompose the ore with nitric acid and evaporate nearly to dry ness. The sample of .5 to 1 gram is then treated with about 5 c.c. of hydrochloric acid and heated until the solution clears. Then add 10 to 15 c.c. of sulphuric acid, and evaporate to fumes of sulphur trioxide. Fil- ter out the lead sulphate, treat the solution with hydrogen sul- phide gas to saturation, filter and extract the mixed sulphides with yellow potassium sulphide, or ammonium sulphide. Bis- muth is rather soluble in sodium sulphide, alone. As copper sulphide is generally present, wash the mass back into the beaker, and warm for some time with 3 to 4 grams of potassium cyanide. Cadmium sulphide and a trace of lead will remain with the bis- muth. Dissolve in nitric acid (1 : 2), filter, wash well, dilute to 250 to 300 c.c., neutralize the boiling liquid with ammonia (1 : 2), until the last drops make the solution faintly cloudy. Then add 1 c.c. of dilute hydrochloric acid (1 : 3), and keep hot for about an hour. Filter on a weighed filter, or asbestos felt, dry at 100 C., and weigh as the basic chloride, BiOCl, con- taining 80.17 per cent of bismuth. The basic salt may be also redissolved, if preferred, then precipitated as carbonate, and finally ignited to oxide. BARIUM 13. Barium, as obtained in the " determination of insoluble residue," or silica, and ferrous oxide (Chapter V), remains in- soluble after the silica is driven off with excess of hydrofluoric and sulphuric acids. When an insoluble residue is obtained from a leady ore, the first insoluble residues should be boiled with water containing 5 grams of ammonium chloride. Then fuse the weighed residue as described in the method for " insoluble matter 90 ANALYSIS OF COPPER in barium sulphate ores" (1, Chapter V), and obtain the barium in the form of pure sulphate for weighing. CADMIUM 14. Cadmium is a rare constituent of regular copper ores, and its determination would only be required in heavy zincy ores. One gram of such material may be decomposed with acids until the metals are dissolved. Evaporate the liquid with 10 c.c. of (1 : 1) sulphuric acid, until strong fumes of sulphur triox- ide are evolved. Cool, dilute to 50 c.c., filter off the lead sul- phate, if much is present and wash with a very little dilute sulphuric acid (1 : 20). Filtration is not, however, necessary at this stage, if the heavy metals are to be removed from the solution by reduction with aluminum. Such a reduction may carry down a part of the cadmium, which must be recovered by dissolving the precipitated metals and repeating the reduc- tion. To separate the iron, zinc, and cadmium, proceed accord- ing to "BreyerV method for zinc." (18 a, Chapter VII.) The heavy metals may also be precipitated as sulphides from hot (1:5) sulphuric acid, or the copper may be removed as thiocyanate (3, Chapter IV). If preferred, the purified cadmium sulphide may finally be dissolved and titrated like zinc, standardizing the potassium ferrocyanide with pure cadmium. The cadmium may also be precipitated as phosphate according to the conditions prescribed in 21, Chapter XIV for "zinc in standard brass." Factor: Cd 2 P 2 7 X .56358= weight of cadmium. 15. Chromium in chromite or furnace refractories may be accurately estimated by the volumetric method of A. G. McKenna. 1 It is necessary to grind the material in an agate mortar (after coarse crushing) to pass a sieve of 100 meshes to the linear inch (or 40 per cm). Fuse .5 gram of the powder with sodium peroxide in a nickel crucible for one minute, or until decomposed. The mass is extracted with water, filtered into a 500 c.c. flask and the filtrate boiled ten minutes to destroy the peroxide. Acidify the cool solution with a large excess of dilute (l :4) sulphuric acid, transfer to a liter beaker and dilute to 800 c.c. with cold 1 Proc. Eng. Soc. W. Pa., 16, 119 Methods of Iron Analysis, Phillips, 156. ELEMENTS IN ORES, SLAGS, AND MATTE 91 water. To this solution add 100 c.c. of ammonium ferrous sulphate solution (equivalent ^tcr 7 grams of metallic iron per liter). This will reduce chroma^e equivalent to .3176 gram of chromium sesquioxide. '" The excess of ferrous sulphate may then be titrated with potassium permanganate (3.692 grams per liter), although solution 41 and 42 of Chapter III may be used equally well. For continuous work the two are made of equal value in chromium 3 Fe=Cr, or 167.52 grams Fe reduce 52.0 grams Cr. The iron in the insoluble residue may be determined by titration, as in the analysis of refractories, 23, Chapter VII. For the complete analysis of chrome ores refer also to method 23 of the next chapter. 16. Cobalt is included with nickel, methods 1-3, Chapter VII, in order to avoid duplication. FLUORINE 17. Fluorine is known to exist in ores chiefly in the form of calcium fluoride, and some metallurgists contend that it is of no value as an active base. Accordingly a reliable method for its estimation is included. It can be roughly determined by the process recommended by A. H. Low for such material. In car- bonates (fluxes), this is a simple problem. The carbonates are boiled in strong acetic acid, after fine grinding. Then add some dilute acid and the calcium fluoride with silicates will remain insoluble. For an accurate estimation, however, proceed accord- ing to Kneeland's method in the following paragraphs. 1 Fuse .5 to 1 gram of ore, or slag (according to the percentage of fluorine), in a porcelain crucible, with 10 times its weight of a mixture of equal parts of sodium and potassium carbonates. When the whole mass has come to a quiet fusion, raise the heat to a bright red, and pour into an iron mold, saving the crucible. Cool, break up the crucible into small pieces, and transfer with the fused mass to a 15 cm. casserole (agate ware preferred to avoid bumping). Add 200 c.c. of water, and digest for one hour at a temperature near boiling, breaking up the fused lumps with a thick glass rod. If any lumps are still noticed, remove them with pincers, grind them in an agate mortar, and wash the 1 Modification of Berzelius' Method Berichte 21 (1888) 2843. Notes /. Amer. Chem. Soc. 37 (1915) 258. 92 ANALYSIS OF COPPER mass back into the casserole with hot water. Now boil for 10 minutes, and filter through a loose paper into a liter beaker. Wash first with hot water, then with a hot solution of ammo- nium carbonate, discarding the residue. Add to the filtrate 10 grams of ammonium carbonate, boil five minutes, and allow to stand in the cold for two hpurs. Filter through a loose filter into an agate-ware casserole, decanting as much as possible of the fluid. Wash once, or twice, with cold water. To eliminate final traces of silica, add 20 c.c. of an emulsion of zinc oxide in ammonium hydroxide, and boil, with the dish uncovered, until no more odor of ammonia is detected. Filter into a 750 c.c. beaker and wash with hot water. To the filtrate, add a solution of calcium chloride, stirring with a rubber-tipped rod until no more precipitate is formed. Allow to settle and filter, washing with hot water. Test the filtrate for carbonates and fluorine with a few drops of calcium chloride solution. Now transfer the filter and precipitate to a platinum dish of suitable size. Dry first, then ignite at a red heat for 20 minutes. Cool and disintegrate the mass with hot water. Add acetic acid until the solution is clear and evaporate to dryness, being care- ful not to decompose the residue. Moisten again with acetic acid and evaporate until there is no more odor of the acid. Wash the mass into a 400 c.c. beaker with hot water, add more water and warm until the calcium acetate is all dissolved, and add, finally, 150 c.c. more of hot water, while stirring. Digest for a few minutes in a warm place and filter, washing first with hot water, then with hot ammonium chloride, and again with hot water. Next, transfer filter and contents to a platinum dish, dry, and ignite. Cool, moisten with cold water, add 6 c.c. of sulphuric acid (d., 1.84), and heat for a few minutes, cool again, add 3 c.c. of hydrochloric acid, and heat for a few minutes more. Cool, dilute, and transfer the contents to a 250-c.c. beaker. Add 5 grams of ammonium chloride, boil for a few minutes, cool, and add an excess of strong ammonia. Add 2 to 3 c.c., of strongest hydrogen peroxide, boil, and filter. The lime is then all pre- cipitated from the filtrate with ammonium oxalate and deter- mined as calcium oxide in the usual manner by titration with permanganate of potash (1 c.c. = .005 g. CaO). CaO x 1.392 gives calcium fluoride (CaF 2 ), and CaF 2 X 0.4782 = fluorine. ELEMENTS IN ORES, SLAGS, AND MATTE 93 In the analysis of fluor-spar, add 4 parts of silica before the first fusion. LEAD IN ORES^AND MATTES 18. Western Assay, Ores. Decompose .5 gram in a cas- serole with 10 c.c. of nitric and 10 c.c. of sulphuric acid (diluted with 1 part of water) . Heat until fumes of sulphur trioxide have been escaping for at least five minutes. It is extremely impor- tant that all nitric acid be expelled. Cool, add 50 c.c. of cold water and boil until all soluble sulphates are dissolved. Filter, wash several times with hot dilute sulphuric acid (1 : 10 of water), then once with hot water. All the iron must be removed from the precipitate. Spread the filter on a watch glass and wash into a beaker with hot water, followed by hot solution of ammonium acetate in sufficient quantity to dissolve all the lead sulphate. Finally, heat to boiling and titrate with standard ammonium molybdate (solution 3, Chapter III). With some ores, it may be necessary to make a preliminary digestion in hydrochloric acid before the nitric and sulphuric acids are added. 19. Lead in Tailings. Take 5 grams and proceed as before, except that the solution of the lead sulphate in ammonium acetate should be filtered from the insoluble matter in order to give a clear solution for titration. This scheme will not give a very exact assay of small amounts of lead, but will always show the presence of lead, which may be estimated with fair accuracy, by the cloudiness produced. The method above described is reliable in presence of the elements usually present, except barium and strontium. In this case, proceed as before until the lead sulphate and insolu- ble matter is on the filter; then dissolve it by boiling in a mix- ture of 50 c.c. of water, .5 c.c. of hydrochloric acid, and 10 grams of crystallized ammonium chloride. From the solution, pre- cipitate the lead with a strip of aluminum; wash the deposit thoroughly, dissolve it in dilute nitric acid, and proceed as in the standardization of the molybdate (solution 3), by lead foil. (See Chapter III.) 20. Lead by Electrolysis. The original description of F. G. Hawley 1 is modified as follows : Digest .8643 gram of ore in a tall 300 c.c. beaker with 15 c.c. of chlorate mixture and evapo- 1 Eng. & Min. Jour. (1910) 648. 94 ANALYSIS OF COPPER rate to fumes of sulphur trioxide. The acid mixture is composed of 1 part sulphuric acid, 2 parts of nitric acid, and 1 part of a saturated solution of potassium chlorate in nitric acid. Cool, add 25 c.c. of water, and bring to a boil to insure per- fect solution of soluble matter. Now set the beaker in an in- clined position in a funnel so that the lead sulphate may collect in one place. Cool again and decant with care through an S. & S. 597 filter, keeping the lead, as far as possible, in the beaker. Wash once with a very little water, allow to settle, decant again, and wash the filter once with a little cold water. Place the beaker under the funnel, and wash the filter with 40 c.c. of a boiling mixture of the following formula (20 c.c. nitric acid, 15 c.c. of saturated ammonium nitrate, and 5 c.c. of water). Boil to ensure complete solution of the lead, rinse into a small 90- c.c. electrolytic beaker, and electrolyze the hot solution for two hours with a current of 1.5 to 2 amperes (at a temperature of about 70 C.). Wash the anode deposit with hot water, then with alcohol, dry over a hot plate, and weigh. According to Hawley, the factor .855 then gives the percentage of lead in the deposit, and the weight of lead peroxide gives by inspection the percentage of lead. Prof. E. F. Smith has proved, however, that the factor .8643 is more uniformly correct for the lead in the peroxide, provided that the anode deposit is dried 20 minutes, or more, in a hot-air oven at 210 to 230 C. The author recommends this procedure. Some operators prefer to dissolve the lead sulphate with warm saturated ammonium carbonate and excess of ammonia, then mix suddenly with the proper amount of nitric acid. Dr. Toisten uses ammonium tartrate, in which case the lead may be pre- cipitated as metal (10, Chapter XIII). 21. Rapid Electrolysis. In the absence of antimony, bis- muth, molybdenum, or tellurium, the assays, already described, may be shortened, and may also become an accurate process for the estimation of traces of lead which cannot be accurately determined as the sulphate. Treat the factor weight of ore (.8643 g.) in a tall 90 c.c. beaker with 10 c.c. of nitric acid. When decomposed, add 15 c.c. of nitric acid, fill with hot water, and electrolyze the nearly boiling solution, as before; employing the Frary solenoid, if desired, and increasing the ELEMENTS IN ORES, SLAGS, AND MATTE 95 current to 3 or 5 amperes. Time of electrolysis, 15 to 20 minutes. Copper in solution has a good influence on the deposition of lead. If copper or similar metalfc are absent, it is necessary to increase the volume of the nitric acid in the hot solution to about 30 per cent. It is sometimes an advantage to interrupt the current for a minute, during electrolysis, to allow any trace of lead to dissolve from the cathode. Bismuth and tin tend to come down with the lead, if appreciable amounts are present, but may be removed after the deposit is weighed. 22. Chromate Method for Calcareous Ores. H. A. Guess l devised a process for low-grade, limy material which is not readily tested by the molybdate method. The standard solutions are: potassium chromate, 100 grams per liter (solution 36, Chapter III); sodium thiosulphate solution containing either 18 or 36 grams per liter. Solution 50, Chapter III, adopted for the iodide titration of copper, may also be used for lead, if stand- ardized against pure lead. Analysis. To an ore-charge of 1 to 5 grams in a broad 250 c.c. beaker, add 3 to 5 c.c. of strong nitric acid, 15 c.c. of hydrochloric acid, and digest until all soluble matter is dis- solved and the excess of acid is reduced to about 8 c.c. The time required is 15 minutes. Remove the flask and add a slight excess of dilute ammonia. Eighty per cent acetic acid is then added slowly with vigorous shaking until the smell indicates a decided excess. Follow with 5 c.c. of concentrated ammonium acetate to insure the solution of any other lead salts. In ab- sence of antimony or gelatinous silica, add to the hot undiluted and unfiltered solution, an excess of about 10 c.c. of a 10 per cent solution of potassium chromate, the total volume being less than 50 c.c. After shaking and settling about five minutes, the contents are to be filtered through a close filter. If these directions are followed, the result will be a granular precipitate. Wash this lead salt several times with hot water containing .5 per cent acetic acid, omitting the latter acid, how- ever, if the ores are low in iron, or manganese, and if the main solution was strongly acidified. Set the funnel over the beaker and pass hot hydrochloric acid (1 : 1) through it until the lead is dissolved, then wash the filter until free from chromate. Add 1 Trans. A. I. M. E. 35, 367. 96 ANALYSIS OF COPPER .5 to 2 grams of potassium iodide, and titrate directly with the standard thiosulphate; of which 1 c.c. equals 5 mg. of lead. By having about 50 c.c. of (1 : 1) hydrochloric acid in 200 c.c. of the warm solution, any tendency to the formation of lead iodide is completely checked, and the end reaction is just as sharp as in a second method devised for richer ores. MANGANESE 23. In Slags. Manganese is usually determined, in the western States, by titrating the neutral solution of the sulphate (according to Volhard), with potassium permanganate. Dis- solve .5 gram of slag exactly as for iron (6 and 7, Chapter V), but use less hydrochloric acid. Add about 5 c.c. of nitric acid and boil until most of the chlorine is expelled. All Lake Superior cupola slags are dissolved as in 7 just quoted, using a beaker 5.5 cm. in diameter and 12.5 cm. in height, which is filled with boiling water before neutralization, within one-half inch of the top. Add a weighed excess (5 to 25 grams) of fine dry zinc oxide with rapid stirring, as in the analysis of steel. If a blank test of 25 grams of reagent requires more than .50 c.c. of the per- manganate, the zinc oxide is not sufficiently pure for the purpose. For Western slags, dilute with hot water to 100 c.c. and add the emulsion of zinc oxide, or the powder, until the iron is com- pletely precipitated. Boil for two minutes and titrate while hot, in the presence of the precipitate, with potassium perman- ganate. On standing a few seconds after stirring, the mass quickly settles, and the pink color at the end-point can be easily detected in the supernatant liquid. The standard solution, used for lime, may also be taken for this titration; 0.5878 x its calcium oxide value giving the value of the solution in manga- nese, Mn. The titer in metallic iron may be multiplied by .2951 if desired. To make a test of the zinc oxide, fill a beaker with hot water, add 1 c.c. of sulphuric acid, stir in the requi- site weight of zinc oxide used in regular work, and titrate as usual. 24. Manganese in Rich Ores. According to A. A. Blair, we may take enough rich material to contain .05 gram of manganese, when using .1 normal permanganate. Fuse 1 gram of such ore in a large platinum crucible with 10 grams of potassium bisulphate, 1 gram of sodium sulphite, and .5 gram of sodium ELEMENTS IN ORES, SLAGS, AND MATTE 97 fluoride. The heating should be slow until effervescence ceases. After complete fusion, cool the product, heat carefully with 10 c.c. of sulphuric acid (1.84), Ihen cool, dissolve in water and make up to a definite volume, from which, after mixing in a dry beaker, an aliquot portion may be titrated. The slight amount of barium sulphate has no influence. NOTE. For the determination of manganese as phosphate, refer to the analysis of brass, 16, Chapter XIV. 25. Manganese by " Sodium Bismuthate." This simple titration process is capable of great accuracy, particularly with iron, steel, and rich ferruginous ores. Originally devised by Schneider, 1 it has been developed by Reddrop & Ramage, Bre- arley and Ibbotsen in England and improved by Blair, 2 Brinton, 3 Blum, 4 and the Society for Testing Materials in the United States. Titration with Permanganate. .015 to .02 gram of manga- nese may be conveniently titrated with .03 normal reagent. Rich ores are titrated with .1 normal permanganate (3.1 grams of salt per liter). The reagents employed in standardization of permanganate are: First, nitric acid diluted by mixing 500 c.c. of acid (d., 1.42) and 1500 c.c. distilled water. Acid for washing a 3 per cent solution by volume. Second, potassium permanganate for titration, 41, Chapter III. Third, a solution of 12.4 grams of crystallized ferrous ammonium sulphate and 50 c.c. (of a mixture of equal volumes of concentrated sulphuric and phosphoric acids) diluted to one liter. For .1 normal permanganate, 39.2 grams of the double salt, or 27.8 grams of crystallized ferrous sulphate, and 50 c.c. each of sulphuric and phosphoric acids are diluted to a liter. Test the iron solution every day against the permanga- nate. If the reagent is slightly aged at first, it will remain practically unaltered for months. The same conditions should prevail in the standardization and analysis. Standardization. Measure into a 200 c.c. flask 50 c.c. of the (1:3) nitric acid, cool in ice water to less than 5 C., add a very 1 Dingier Polytech. J. 269, 224. 2 /. Am. Chem. Soc. 26, 793. 3 J. Ind. and Eng. Chem. 3, 237. 4 Eighth Inter. Congress of Chem., 1, 62. 98 ANALYSIS OF COPPER little sodium bismuthate, dilute with 50 c.c. of the 3 per cent acid, filter through an asbestos felt into a 300 c.c. flask, wash with 50 c.c. of the 3 per cent acid, and titrate at once, after the addition of 25 c.c. ferrous sulphate solution. If the felt is well coated with bismuthate, it is unnecessary to add any to flask. This gives the value of the ferrous sulphate. The permanganate may be valued: First, by multiplying the iron value by .4919. Second, by the use of a solution of pure manganous sulphate, in which the manganese is determined by evaporating a portion and heating for 4 hours at 450 to 500 C., then multiplying the weight by .3638. Oxidize and titrate 1 to 3 grams of this solution, which contains 5.749 grams of anhydrous sulphate in 1000 grams. Or 4.124 grams may be made up to 500 grams with distilled water and 1 gram will contain .003 gram manganese. A color is given to 50 c.c. of liquid by .00005 gram of manganese. Every trace of nitrous acid must be previously boiled out and the titration completed within ten to fifteen minutes after filtration. The solution changes in fifteen minutes at 40 C., but will remain unaltered several hours at 5 C. A large excess of bismuthate, over .5 to 1 gram, is inadvisable. Third method for valuation of permanganate, comparison with manganese in a standard steel or ore. Fourth, by specially purified (reagent) sodium exalate, which has been well dried in an air oven at 100 to 105 C. before use and preserved in a small glass stoppered bottle. Dilute the oxalate solution to 75 c.c. for .03 normal permanganate and 250 c.c. for the .1 normal reagent. The initial temperature should be 80 to 90 C. and the final 60 or higher. This is the best method if strictly correct oxalate is at hand. Such a correct salt may be obtained from the U. S. Bureau of Standards. The value of the permanganate in terms of the oxalate is then to be multiplied by .16397 to obtain the equivalent in manganese. The stronger permanganate may also be used to test the sodium arsenite used in the second alternative method of titration of steels and ores. 1 Titration by Sodium Arsenite. Make up a stock, solution by heating in a flask on the water bath 15 grams of arsenious oxide (As 2 3 ) 45 grams of sodium carbonate and 150 c.c. dis- 1 American Society for Testing Materials. Official Method for Steel in Year Book, 1914, 177. ELEMENTS IN ORES, SLAGS, AND MATTE 99 tilled water. Cool the solution and make up to 1000 c.c. with distilled water. A standard solution is made by diluting 300 c.c. of the " stock solution" to one liter and titrating against the 1. normal permanganate which has Seen standardized by the fourth method above given. The solution may be adjusted so that 1 c.c. is equivalent to .1 per cent of manganese when a one-gram sample is taken; formula 47, Chapter III. Analysis of Ores and Slags. If the ore dissolves rather easily, treat 1 gram with 12 c.c. of hydrochloric acid (d., 1.2) in a 120 c.c. Erlenmeyer flask, evaporate almost to pastiness, and then add 4 c.c. of sulphuric acid (d., 1.84). By boiling down to fumes over a free flame with the flask in a holder, the hydro- chloric acid is so completely expelled that no cloud test can be obtained with silver nitrate. This is very essential. Take up the residue with 50 c.c. of the (1:3) nitric acid, cool at least to 15 C., but to 5 C. for accurate work. Then add an excess .5 to 1 gram of sodium bismuthate and agitate for three to four minutes. Pour in 50 c.c. of the cold 3 per cent nitric acid (as in standardization) and filter through an alundum crucible or asbestos felt. Wash with about 50 c.c. of 3 per cent acid. Ti- trate at once by potassium permanganate after addition of 25 c.c. ferrous sulphate solution, or directly with the sodium arsenite if preferred. A few ores will not yield all the manganese to this treatment. If the residue is dark, it should be filtered off and fused with a very small amount of potassium bisulphate. Metzger and McCrackan claim good results by treating the sample with as much as 10 to 15 c.c. of sulphuric acid. They add subsequently 1 to 2 grams of the sodium bismuthate after the hydrochloric acid has been removed, and boil the solution twenty minutes to oxidize the manganese. Filter and wash with about 50 c.c. of the 3 per cent nitric acid as before. MOLYBDENUM AND OTHER RARE METALS 26. In copper refineries, tests are seldom required for any of the rarer metals except selenium, tellurium, gold or platinum in copper or ores, and occasionally platinum or palladium in electrolytic slimes. For the assay of the precious metals, consult 9, 10, Chapter VIII. The determination of selenium and tellurium is taken up in 100 ANALYSIS OF COPPER 10, Chapter VII and again in the analysis of metallic copper, 17, Chapter XIII. The latest methods for the identification of tungsten, vana- dium, and other rare metals are summarized in the " Report on Rare Metals to the Committee on Analysis," 8th Inter- national Congress of Applied Chemistry (1913), pp. 1 to 24. CHAPTER VII SPECIAL DETERMINATIONS IN ORES, SLAGS, AND MATTE CONCLUDED FURNACE REFRACTORIES NICKEL AND COBALT 1. Titration of Nickel in Matte. For nickeliferous matte, one may use a rapid method for the titration of nickel (and copper) which involves the prior separation of the copper as sulphide, and its subsequent titration in the usual way by potassium iodide, as in Chapter IV. The nickel is titrated in alkaline sodium citrate solution with potassium cyanide. This rapid assay is practiced by the Canadian Copper Company. To prepare for titration, weigh .5 gram of matte into a 200 to 300 c.c. lipless beaker. Add 15 c.c. of strong hydrochloric acid, cover, and set the beaker on the hot plate. The solution will be complete in 3 to 5 minutes. When the matte is all dissolved, add a single drop of nitric acid, and boil for a few moments. Dilute the solution to about 100 c.c. with hot water and pass hydrogen sulphide gas until the precipitation of the copper group is complete. Filter rapidly through a 597 S. & S. filter, receiving the filtrate into a 500 c.c. griffin-form, Jena beaker. Wash five to seven times with hot water, and set the filtrate on a hot plate to boil. Spread out the filter of copper sulphide against the side of the beaker, keeping that part of the paper holding most of the precipitate, below the rim of the beaker. Wash the paper as free as possible from the sulphide, using as little hot water as possible. Now add 3 to 4 c.c. of strong nitric acid and place the beaker to heat. Remove any adhering copper from the paper, or rod, by saturated bromine water. Evaporate the copper solution over a free flame, almost, but not quite, to dryness. The yellow globule of sulphur should not show any dark color. Wash down the sides of the beaker, using, however, no more than 5 c.c. of water. Add, drop by 102 ANALYSIS OF COPPER drop, a strong solution of sodium carbonate just in sufficient amount to form a slight permanent precipitate. Redissolve this in dilute acetic acid, using a slight excess, keeping the volume less than 10 c.c. Titrate as in method 1, Chapter IV, with sodium thiosulphate (50, Chapter III) adding six times as much potassium iodide (formula 38) as there is copper present. Nickel. The nickel solution is treated while the copper is being prepared for titration. The nickel solution should be boiled for 2 to 3 minutes to expel the hydrogen sulphide remain- ing from the copper precipitation. While still boiling, and after expulsion of the hydrogen sulphide, add about 1 gram of potas- sium chlorate and boil for one half minute. If the chlorate has been added before all the gas is expelled, sulphur will separate and it is very difficult to obtain a clear solution. However, prolonged boiling with the addition of more potassium chlorate will usually clear the solution. Cool slightly and add 20 c.c. of sodium citrate solution (47, 200 grams per liter). Neutralize with ammonia, using a distinct excess. Cool in running water. When the solution is quite cold, adjust the alkalinity by adding cold dilute hydrochloric acid, or ammonia, as required. The correct degree of alkalinity is a distinct but small excess of ammonia. The liquid is now titrated for nickel by a modification of the method of T. Moore. Add 5 c.c. of potassium iodide solution and 5 c.c. of a solution of silver nitrate (1 gram per liter). If a permanent cloudiness forms after the addition of the potassium iodide and before the addition of silver nitrate, more ammonia must be added until the solution becomes quite clear again. Then add the silver nitrate, which will cause a permanent cloudiness. Run in potassium cyanide with constant stirring till the solution is almost clear again. Now allow to stand while titrating the copper, as already described. Before that operation is completed, the nickel will have become cloudy again. Care- fully add more potassium cyanide till the nickel solution becomes quite clear. These standard solutions are standardized against matte which has been analyzed by exact electrolytic methods. NOTE. The potassium cyanide solution contains 24.5 grams of 98 per cent salt per liter (formula 31, Chapter III). For very accurate work, the amount of potassium cyanide required DETERMINATION IN ORES, SLAGS, AND MATTE 103 to dissolve the precipitate of silver iodide should be deducted from the total number of cubic } centimeters required to clear the solution. (This usually amounts to about .15 c.c.) For high grade and" " Bessemer" mattes, the procedure is the same except that 40 c.c. of hydrochloric acid are required in dissolving the matte. Five c.c. of the sodium citrate is sufficient. With the Bessemer mattes, the potassium cyanide must be run in very rapidly until near the end-point, otherwise nickel cyanide may form. This is difficult to dissolve and the end-point may be passed. (Zinc uses up cyanide. Small amounts of cobalt do not interfere, but larger amounts spoil the titration.) 2. Nickel with Cobalt by Electrolysis. If copper is not to be determined, an ore may be treated in a lipped beaker with a mixture of acids and evaporated to strong fumes of sulphuric anhydride. Some ores, or mattes, may be decomposed very easily by digestion with sulphuric acid and potassium bisulphate, in the same manner as for arsenic. The copper, etc., may then be precipitated from the diluted solution by hydrogen sulphide and the remaining gas boiled out of the filtered solution. A more rapid and satisfactory separation, however, is that by rapid electrolysis, placing the beaker in a Frary rotary apparatus. The solutions are subjected to 4 to 4.5 amperes at 2.75 volts electrode tension when a five gram sample is tested, or 3 amperes for a one-gram sample. Conducted through a rheostat, or lamp resistance, the current from a 110- volt circuit will deposit 5 grams of copper in 2.5 hours. When the iron percentage is low, the best electrolyte contains about 6 to 7 c.c. of nitric acid and 10 c.c. of sulphuric acid. When the iron is over 10 per cent, an excess of nitric acid may retard deposition. If any manganese is present, it may be re- moved with the iron, or after it, by boiling the faintly ammonia- cal solution with an excess of bromine or ammonium persulphate. If a separate assay for zinc is required, that element may be first removed by rendering the solution faintly acid with acetic acid, adding a quantity of glacial acetic equal to one-fifth of the volume of the solution, and saturating the cold solution thoroughly with hydrogen sulphide. If preferred, the Waring method of precipitation in formic acid solution may be conducted 104 ANALYSIS OF COPPER as described under "zinc." The white sulphide may be colored with a trace of lead or copper sulphide. The zinc is filtered out and washed with hydrogen sulphide water containing a little ammonium acetate, after which the hydrogen sulphide is boiled out of the filtrate. The solution can now be electrolyzed, although a little better result may be obtained by adding 5 c.c,. sulphuric acid and boiling down to fumes to remove most of the acetic and any nitric acid present. Neutralize the solution with ammonia (d., .9) and add 30 c.c. excess. The total volume may be 100 to 200 c.c., according to the amount of cobalt and nickel present. If the Frary solenoid is used (Chapter I), the strength of current may be 3 to 4 amperes, with an electrode tension of 2.5 volts, and the time about one and a half hours, or less. The solution must not become acid during the deposition. The end-point is tested by withdrawing 1 c.c. and testing with hydrogen sulphide. The current for deposition by the slow method is .5 amperes per sq. dm. at 2.5 volts potential. Two accurate methods for preliminary separation of cobalt arid nickel are given later. The cobalt and nickel are then electro- lyzed separately. The cathode deposits must always be tested for a trace of iron, and occasionally platinum or copper. 3. Ether Method for Cobalt, Nickel, and Zinc. This separa- tion was devised by H. Koch primarily for Mansfeld ores. The separation of the large quantity (about 25 per cent) of iron was formerly made by the " basic acetate." Of late, the well-known "ether method" of Rothe is used to advantage. The shaking apparatus of the inventor is not used. The same purpose is accomplished by a simple shaking cylinder and a removable flask adjoining, as shown in Fig. 12. A 5-gram sample of powdered crude ore in a 16 cm. evaporat- ing dish is moistened with water to avoid adhesion of the fines to the dish, which would occur if the dry substance were heated with acid. After decomposing with acid, the liquid is evaporated to dryness several times on the water bath with hydrochloric acid to convert the nitrates to chlorides, and the separated sul- phur is burnt out by heating on a small sand bath. The cooled residue is treated again with hydrochloric acid, evaporated, and taken up with 12 to 15 cm. of water. L This concentration must be adhered to, since with greater DETERMINATION IN ORES, SLAGS, AND MATTE 105 dilution the later separation of ferric chloride is incomplete. The solution is washed into the shaking cylinder by means of 40 c.c. of hydrochloric acid (l:2)^with 50 c.c. ether added, and strongly shaken. After ^he subsidence of the copper chloride solution, the etherial extract is drawn off by the bottle syphon, a second equal quantity of ether added, and the shaking repeated. A third similar extraction is sufficient to remove the last trace of iron. (This operation may be hastened by making one ammonia precipitation of the iron, followed by one ether separa- tion only.) The liquid remaining in the shaker is transferred to a 750 c.c. lipped beaker, to pre- cipitate the copper and arsenic with hydrogen sulphide. For safety, it is recommended to shake out the ether extract twice with some hydro- chloric acid to recover a trace of nickel as chloride, which will then be added to the main portion in the beaker. After filtration, the sulphides are evaporated to dryness with nitric and sulphuric acids, taken up with water, the separated Fig. 12. Cylinders for Ether Method. lead sulphate filtered off, and washed with dilute sulphuric acid. In the filtrate, manganese and the small trace of remaining iron are precipitated with ammonia (and bromine water). Repeated precipitation and solution of the latter in hot sulphuric acid with the addition of a few drops of solution of sulphur dioxide is found necessary, as otherwise cobalt remains in the man- ganese. The liquid is then neutralized with dilute sulphuric acid until the last drop renders it slightly acid (to methyl orange indicator), and the zinc is precipitated with hydrogen sulphide. After standing twelve hours, the sulphide is filtered off with the aid of the hydrogen sulphide water, to which some ammonium 106 ANALYSIS OF COPPER sulphate has been added, and the filtrate is concentrated on the water bath to destroy the hydrogen sulphide. The solution must first be acidified with sulphuric acid, since nickel sulphide might separate from it if nearly neutral. The concentrated liquid is then poured into a 1.5-liter beaker, saturated with ammonia, and electrolyzed for the precipitation of cobalt and nickel. 4. Separation of Cobalt from Nickel. Precipitation with potassium nitrite is rather uncertain, and two more delicate separations are now preferred. The first scheme is adapted to the separation of a little cobalt from much nickel. This is accom- plished by the use of nitroso-/3-naphthol, CioH 6 (NOH) ; a method due to Knorre and Illinski. 1 Remove the heavy metals from solution, either by hydrogen sulphide or by electrolysis, then take out the iron and alumina and follow with a second separation by ammonia. A very large amount of iron, with little alumina, may be separated by Koch's ether method, but if the cobalt, nickel, and zinc are only traces, as in pure ores and high grade metal, two precipitations with ammonia are sufficient. (See Waring's method for zinc.) If manganese is present, this is removed by bromine or am- monium persulphate. To the nearly neutral solution of the sulphates of cobalt, nickel, and zinc (or the solution of the cathode deposit of combined elements), add 4 to 5 c.c. strong hydrochloric acid, warm, and add a hot saturated solution of the naphthol reagent in 50 per cent acetic acid, as long as precipitation continues. Allow to settle for a few hours at room temperature, wash first with warm 12 per cent hydrochloric acid until all the nickel is removed, then wash with water to remove the acid. The precipitate can be ignited with oxalic acid in a Rose cruci- ble, finally reducing in hydrogen. The best way, however, is to dissolve in a little nitric acid and 3 to 5 c.c. sulphuric acid, and evaporate to fumes. Then dilute, neutralize with ammonia, add 30 c.c. in excess, and electrolyze with a platinum cathode. The data for the method are : volume 30 to 150 c.c., accord- ing to the amount of cobalt; area of cathode surface 50 or 100 sq. cm.; current .1 to .5 ampere, at about 2.5 volts potential. With revolving anode or solenoid, pass 4 amperes through coil and sheet cathode; time about 1 to 1.5 hours. The final test 1 Ber. Deutsch. Chem. Gesell (1885), 699. DETERMINATION IN ORES, SLAGS, AND MATTE 107 is made with hydrogen sulphide water. Test the deposits for purity. , 5. Separation of a little NickeJ from much Cobalt or Zinc. - The beautiful reaction of nickel with dimethyl glyoxime, dis- covered by Tschugaeff and Brunck, 1 is made the basis of the method. This may even be applied in presence of the iron and alumina by holding them in solution with an excess of sodium citrate, or tartaric acid (2 to 3 grams), as in the titration method already described. It is better, in this case, to add also 2 to 3 grams ammonium chloride. Prepare the solution, ordinarily, by removing the heavier metals, including iron, alumina, and man- ganese, by the separations already described (4). Then acidify, after boiling out all hydrogen sulphide, until the solution contains about 5 c.c. of free hydrochloric acid. The solution should be diluted so that 100 c.c. does not contain more than .1 gram of cobalt ; and a standard 1 per cent or 2 per cent alcoholic solution of the glyoxime reagent is then added to the boiling solution until the amount is about five times that of the nickel and cobalt. Ammonia is then added until the solution is slightly alkaline, and the precipitate allowed to settle some time on the steam plate. The liquid is filtered hot through a pure asbestos felt which has been dried and weighed. The washed salt is dried to constant weight at 110 to 120 C. The compound has the symbol CgHnN^Ni, which corresponds to 20.325 per cent of nickel. A method for the recovery of the oxime is to be found in the papers quoted. The salt is almost insoluble in water and only very slightly soluble in alcohol or acetic acid. If accurate results are desired with zinc ore and other rich products, it is necessary to dissolve the pink precipitate in hydrochloric acid (1 : 1), wash the filter, and repeat the separa- tion with a little more of the glyoxime. The filtrates from the nickel may then be treated for the separation of zinc from the cobalt by precipitation with hydrogen sulphide in a formic, or acetic acid solution (as in 2), unless the cobalt was removed before the nickel by method 4. Instead of weighing the nickel glyoxime, it may be dissolved in diluted nitric or hydrochloric acid, evaporated to fumes with 5 c.c. of sulphuric acid, and the sulphate neutralized with excess 1 J. Soc. Chem. Ind. 24, 941; Zeit. Angew.Chem. 20, 38, 44; also abstract J. Am. Chem. Soc. 2, 240. 108 ANALYSIS OF COPPER of ammonia and electrolyzed as in 2. The accuracy required and the amounts or relative proportions of cobalt, nickel, and zinc, should influence the choice of the method of separation of these elements. NOTE (on Zinc). The direct determination of zinc in slags, or ores, is recorded in 17, near the end of this chapter. POTASSIUM AND SODIUM 6. Potassium and Sodium (with Lithium) are estimated in silicates, such as clay, slags, or ores, by removing all other elements from the solution successively, or by sintering in a platinum crucible with an excess of calcium carbonate and ammonium chloride, extracting with hot water and precipitating any lime from the extract. Solution Method. Treat two grams of dried clay, or furnace product, with an excess of hydrofluoric acid and sulphuric acid in a platinum dish. In the case of slags, it is sometimes advisable to add nitric acid. Heat gently without baking until all free sulphuric acid is removed. A. H. Low recommends the direct extraction of this residue with hot ammonium hydroxide. Such treatment should be proved correct before acceptance with un- known material. Usually, the residue from slags, or ores, should be dissolved in dilute 1 hydrochloric acid, and the ferric hydroxide and alumina removed by two successive precipitations with ammonia in a platinum dish. Filter and wash, then remove any heavy metals by hydrogen sulphide, and filter again. Evapo- rate to dryness in a platinum dish and volatilize nearly all the ammonium chloride. Dissolve and precipitate calcium and magnesium by boiling with a slight excess of ammonia and ammonium carbonate. Filter, evaporate filtrate, and remove salts by gentle ignition as directed in the next paragraph. Determination as Sulphates. After the metals and alkaline earths are separated, evaporate the filtrate or extract to dryness in a platinum dish, and heat carefully at the lowest visible red heat, or below that point, until the ammonium salts are removed. Dissolve the residue in 5 to 10 c.c. of hot water. If no separa- tion of potassium from sodium is desired, the hot solution may be tested with a few drops of ammonium carbonate, filtered into a weighed platinum crucible, and the solution and washings evapo- DETERMINATION IN ORES, SLAGS, AND MATTE 109 rated again to dry ness on the steam plate or water bath. Ignite very gently as before and aft^r "cooling ten minutes in a desic- cator, weigh the sulphates. The,,sulphur trioxide in combination may then be found by dissolving the residue in a very little water, adding a drop of hydrochloric acid, and precipitating with barium chloride. Allow to settle, filter, wash, ignite, weigh the barium sulphate, and calculate the sulphur trioxide. Deduct this amount from the total weight of sulphates to obtain the total weight of oxides of potassium, sodium (and lithium if present). Determination as Chlorides. To remove the small amount of sulphates from the purified solutions of potassium and sodium, either add barium acetate, or a slight excess of barium chloride. In the former case, boil the solution, filter, wash out soluble matter, evaporate to dryness, and then heat to the lowest visible red to destroy the acetates. Treat the residue with a little water, filter out any insoluble barium or magnesium carbonates, add 2 to 3 drops of barium hydroxide solution, and take again to dryness. Dissolve in 5 c.c. of water in all cases, filter into a weighed crucible, concentrate to a low point, and if the liquid does not remain clear, filter into another clean, weighed crucible. Add two drops of hydrochloric acid, heat very carefully to the lowest perceptible red, cool in a dessicator for ten minutes, and weigh as chlorides of sodium, potassium, and lithium, if present. Test the purity of the residue by dissolving in 5 c.c. of water, evaporating with a few drops of ammonium carbonate, and redissolving in 5 c.c. of water. If the liquid is not clear, filter on a very small paper, wash with a few drops of water, evaporate in a weighed crucible, ignite gently, and weigh again. SEPARATIONS POTASSIUM FROM SODIUM, (AND LITHIUM) 7. Potassium is separated from the sodium (or lithium) by dissolving the weighed chlorides in the least amount of warm water, and treating the clear solution in the crucible, or a No. 3a casserole, with sufficient 10 per cent solution of platinic chloride " to convert all the sodium and potassium into double salts. Any unchanged sodium chloride would not dissolve in absolute alcohol. Assuming the mixture to be mostly sodium chloride, 116.92 parts of sodium chloride will require 195.2 parts of platinum, or its equivalent as chloride. A slight excess of .5 to 1 c.c. should be 110 ANALYSIS OF COPPER present in addition to the calculated amount. Concentrate below the boiling temperature until the mass solidifies upon cooling. Add a little absolute alcohol (according to A. H. Low, methyl is the best); stir well with a bent rod or wire, then filter through a small paper or a weighed asbestos felt. Repeat this extraction and washing with the alcohol until a pure golden yellow residue remains. Lithium, if present, is extracted with the sodium. William Crookes recommends the addition of J part of ether to the alcohol. If a dry filter is used, dry the washed contents in an oven at 80 to 90 C., then transfer as much as possible to a watch glass, and wash the adhering salt with hot water into a weighed platinum crucible. Evaporate the solution at a low heat, add the rest of the precipitate, dry the whole at 160 C., and weigh the potassium platinic chloride. The percentage of potassium oxide, corresponding to the K 2 PtCl 6 , may be found by multiplying by the factor .19376 (or .30673 for KC1). A. H. Low considers that the older factor of .3056 is the more correct one as slight changes occur during evaporation. The sodium chloride is taken by difference and multiplied by .53028 to obtain sodium oxide, Na 2 O. Lithium may be separ- ated as phosphate. 1 8. Indirect Separation. This is only applicable when the sodium and potassium are present in nearly equal amounts. First, determine the total percentage of chlorine in the weighed chlorides of potassium and sodium by precipitation, or by titra- tion, with silver salt. Finally, apply the rule of A. H. Low, which is the most simple form published. Subtract 47.56 from the percentage of chlorine in the weighed chlorides of potassium and sodium. Divide the remainder by 13.098 and the result will be the per cent of sodium chloride in the mixture of the two chlorides. This assumes that the lithium is negligible. The chloride of ammonium, which so often causes trouble by creeping during evaporation, may be decomposed very easily by adopting a suggestion of J. L. Smith. Evaporate the solution (after removal of the lime and magnesium salts) to a low point, transfer to a flask or tall beaker with a very little water, add 3 to 4 c.c. of hydrochloric acid for every gram of ammonium chloride present, and heat at a temperature a little below 100 C. until action ceases. Transfer to a small porcelain casserole, 1 Hillebrand, Bull. 422, U. S. Gcol. Survey. (Fresenius, Quant. Anal.) DETERMINATION IN ORES, SLAGS, AND MATTE 111 evaporate with hydrochloric acid to dryness, treat again with a little ammonium carbonate, preceded by barium acetate, as already described, and then filtdr back into the weighed platinum dish, or crucible. 9. Potassium and Sodium (by the sintering method of J. Law- rence Smith l ). The alkaline metals in nearly all silicates, except possibly spinel, or some of the silicates of heavy metals, may be converted into soluble chlorides by slow sintering in a platinum crucible with an intimate mixture of pure ammonium chloride and calcium carbonate. The alkalies are then leached out with hot water. Triturate 1 gram of the sample (in the form of impalpable powder) with 1 gram of pure ammonium chloride in an agate mortar, then mix with 8 grams of calcium carbonate, transfer to a large platinum crucible, and cover tightly. Heat in a slightly inclined position with a very small flame at a heat which will gradually drive off ammonia gas but not ammonium chloride. When the ammonia is removed (which should require fifteen to twenty minutes), raise the temperature so that the lower half or three-fourths of the crucible are maintained for one hour at a dull red heat. Transfer the cooled mass to a large platinum dish with 100 c.c. of hot water, boil, and break up any large particles with a small pestle. Filter and wash with hot distilled water, boiling the residue with more hot water. Test this washed residue with excess of hydrochloric acid to prove that the clay, or furnace product, was all decomposed. To the nitrate, add 1.5 grams of dry chemically pure am- monium carbonate, evaporate carefully to 40 c.c., add a little more carbonate and a few cubic centimeters of ammonia, and filter through a small paper. Wash the filter with a little water, and then proceed to remove the last traces of calcium and magnesium by evaporation, gentle ignition, and repeated treatment with ammonium hydroxide and carbonate. No ad- dition of barium salt is necessary, if no sulphates are present. The chlorides are separated as in the former method (7). SELENIUM AND TELLURIUM IN SLAGS OR ORES i 10. Selenium and Tellurium are separated from large quan- tities of iron salts, and from some heavy metals, by precipi- 1 Am. J. of Set. and Art, 3d series, 1, 269. 112 ANALYSIS OF COPPER tation with sulphur dioxide in acid solution. The metals are redissolved, when a separation of the two is required, and sepa- rated from each other by fractional precipitation from a hydro- chloric acid solution by means of sulphur dioxide. The amount of sample to be taken (5 to 25 grams) depends on the amount of the rare elements judged to be present in the material tested. Ores and Slags. Digest in a mixture of strong nitric and sulphuric acids treating the insoluble residue with hydrofluoric acid, if necessary to complete decomposition. Evaporate on the water bath with a slight excess of sulphuric acid until the other acids are practically removed. In the absence of copper or similar metals a known amount of ammonium nitrate, potassium nitrate, or zinc nitrate should be added before evaporation, sufficient to form double salts with the selenium and tellurium and prevent any loss of those elements. Redis- solve the salts in water and filter out any lead sulphate, etc. At this point there should be no hydrochloric acid, whatever, in the solution. Selenium and tellurium may now be separated com- pletely from iron and copper, or other metals, with the exception of gold, by heating to 85 to 90 C. and saturating the liquid with sulphur dioxide (prepared according to Chapter III). Keep hot for one hour, or long enough to remove any yellow- ish color in the solution; finally cool the liquid while the gas current is passing and allow to settle over night. Filter on a pure asbestos felt, made from acid- washed, ignited material, wash with dilute acid, then with water, dry at 100 to 105 C., and weigh. If accurate results are desired, ignite the felt before use, cool, and weigh; then moisten with water, dry in the oven at 100 to 105 C., and weigh again. This weight may be .5 to 1 mg. heavier than the first. After drying and weighing the dried precipitate, ignite to redness, and obtain the elements by loss as a check, using the first ignited weight of the felts for comparison. In this, or the following method of separation, the filtrates from the precipitation of either selenium or tellurium should be charged a second time in the same way with sulphur dioxide and allowed to settle, then filtered through a second weighed filter, if a trace of precipitate appears. Separation. If selenium and tellurium are to be separated, the first felt need not be weighed. The reduced metals are DETERMINATION IN ORES, SLAGS, AND MATTE 113 filtered off, then redissolved in a very little strong nitric acid, the asbestos filtered out, and ^the solution evaporated to dry- ness on the water bath with tjhe addition of an amount of ammonium or potassium nitrate sufficient to combine with the metals. If five drops of sulphuric acid are also added to convert the salts to sulphates, the subsequent reduction takes place quickly. Dissolve the salts in 90 c.c. of hydrochloric (d., 1.2) and 10 c.c. of water, heat nearly to boiling for one minute to reduce any nitrates and convert the selenium to a lower chloride. Saturate the solution with sulphur dioxide, passing the gas until the solution is cold. Settle as before, filter, wash with 90 per cent hydrochloric acid, dry, and weigh as directed for the estima- tion of the two elements together. The washing is completed with water to dissolve any sodium chloride, but this wash water should not be allowed to dilute the filtrate until .a second test has been made with sulphur dioxide to insure the complete separation of the selenium. To determine the tellurium in the total filtrate, dilute it with four volumes of water, charge again with sulphur dioxide, and proceed exactly as directed . for se- lenium, finally igniting the dried felt as a check on the weight by drying. The main solution of the original sample may be taken for the estimation of arsenic and antimony by usual methods. See also Chapter XIII. 11. The second method for selenium and tellurium (that of Edward Keller) depends on the preliminary separation of the rare elements from the copper by repeated precipitation with ammonia and excess of ferric salts. The iron oxide should be at least twenty times the weight of the selenium and tellurium. Ammonium ferric sulphate is added if necessary, to obtain the requisite amount. The copper should be completely removed by repeated precipitation and filtration, or copper selenide will be formed in the subsequent treatment. This compound is insoluble in sodium sulphide. Finally dissolve the iron hydrox- ides in (1 : 3) hydrochloric acid, dilute to 400 c.c., and precipitate the selenium, etc., in the cold by hydrogen sulphide. Filter, extract by repeated treatment with cold, then hot, dilute sodium sulphide solution, and evaporate the extract to dryness on the water bath. Decompose the sulphur with fuming nitric acid, or with hydrochloric acid and potassium chlorate. 114 ANALYSIS OF COPPER Evaporate to dryness again on the water bath, then proceed as in the previous method for the precipitation of the selenium in strong hydrochloric acid solution by means of sulphur dioxide. TIN IN ORES BY TITRATION 12. Tin, in regular analysis, is obtained with the sulphides of arsenic and antimony (method 10, Chapter VI). A. H. Low has devised a good volumetric method, the titration of stannous chloride by iodine. The liquid remaining in the still after distillation of arsenic may be taken. Or tin and antimony may be obtained from the filtrate remaining after the precipitation of arsenic by hydrogen sulphide in (2 : 1) hydrochloric acid. Dilute this filtrate with four parts of water and precipitate the tin and antimony, then filter and wash. Dissolve the sulphides in a little pure yellow ammonium sulphide, and transfer the solution to a 300 c.c. Kjehldahl flask with a long neck. Add 15 c.c. of sulphuric acid and 3 grams of potassium sulphate and boil down to fumes of sulphur trioxide. Now add .25 gram of solid tartaric acid and boil until all sulphur is expelled and most of the free acid. The antimony will then be reduced to the "ous" condition, and is ready for titration with tenth .normal permanganate which has been standardized against pure antimony. Dissolve the cooled residue in 50 c.c. of water and 10 c.c. of strong hydrochloric acid, boil out any sulphur dioxide, add 10 c.c. more of the acid, dilute to 200 c.c., cool to room tem- perature and titrate directly for antimony. The tin is determined by washing the titrated solution into a 500 c.c. round-bottomed flask with 50 c.c. of strong hydro- chloric acid and adding a reagent which will reduce the stannic compound to stannous chloride, ready for titration with iodine. Dilute to 200 c.c., add 1 gram of fine antimony powder (chemi- cally pure), and replace on the steam bath, shaking occasionally. A. H. Low uses a coil of sheet nickel, made from a 7 x 1 inch strip of the pure metal, boiling twenty minutes to reduce the tin completely to the stannous condition. When the tin is completely reduced by one of these pure metals, add a 1 cm. cube of marble to the flask, cool rapidly in running water, keeping the flask covered, then titrate care- fully and rapidly with tenth normal iodine, which has been DETERMINATION IN ORES, SLAGS, AND MATTE 115 standardized under the same conditions against chemically pure tin foil. NOTE. The foregoing directions apply to samples in which the tin is rather small in amount. When a large amount of tin is present, it is most conveniently separated from antimony, etc., by Thompson's oxalic acid-ammonium oxalate method and electrolyzed (10, Chapter VI). ^ TITANIUM 13. Titanium is determined in ores and slags by separating it from the bulk of the iron and alumina, converting it with a soda-fusion to insoluble titanate. The titanate is finally dis- solved and titanic hydroxide precipitated from an acetic acid solution in presence of sodium acetate. The gravimetric method of A. A. Blair (" Analysis of Iron"), may be adapted to copper products, if the sample is dissolved in acids as for electrolytic assaying, and the copper removed by the electric current. In- soluble residues are fused, as directed in Chapter V. The copper from blast furnace slags may be removed by rapid electrolysis in thirty to sixty minutes. Treat the silica in a platinum cruci- ble with a few drops of sulphuric acid and an excess of hydro- fluoric acid, evaporate the latter acid, and preserve. Transfer the electrolyte and solution of residue to a 500 c.c. beaker, evaporate to fumes, and redissolve in 150 c.c. of water. Neutralize the liquid with ammonium hydroxide and add 50 c.c. of strong solution of sulphur dioxide, which should redissolve any slight precipitate. Now pour in a clear filtered solution of 20 grams of sodium acetate and acetic acid (d., 1.04) equal to one-sixth of the total volume of solution. Boil for a few minutes, allow to settle, filter, and wash with 17 per cent acetic acid. Ignite the filter and contents; then fuse this insoluble residue with 5 grams of sodium carbonate for about half an hour. Run the fusion well up on the side; cool, dissolve in water, and filter to extract any soluble alumina. Wash the insoluble sodium titanate, re-fuse it as before, and cool the crucible. Then pour in very gradually strong sulphuric acid, finally warming the crucible slightly until the fusion is dissolved and fumes of sulphur trioxide are copiously evolved. Pour the fluid contents into 250 c.c. of cold water and wash out 116 ANALYSIS OF COPPER the crucible. Add 50 c.c. of saturated aqueous solution of sul- phur dioxide (or 3 c.c. of saturated solution of acid ammonium sulphite), filter if necessary, neutralize with ammonia, treat the clear and almost colorless liquid with the same amount of so- dium acetate and acetic acid as before, and precipitate the titanium hydroxide by boiling. If the ignited oxide is still dis- colored, repeat the fusion and precipitation. Titanium oxide (TiOi) X .6005 = titanium. The manner of decomposition and number of purifications for the titanic hydroxide must depend on the material treated. If the original "insoluble matter" contains lead sulphate, this should be extracted with slightly alkaline ammonium acetate before fusion of the residue is attempted. 14. Titanium by Colorimetric Assay. The color test de- vised by Weller, 1 and improved by H. L. Wells and W. A. Noyes, 2 is as follows : Mix .1 gram of ore with .2 gram of finely powdered sodium fluoride in a platinum crucible, adding 3 grams of sodium bi- sulphate without mixing. Fuse gently for two or three min- utes until copious fumes are evolved. Dissolve the cooled mass from the crucible with 15 to 20 c.c. of cold water, and filter and wash with about '10 c.c. of water. Treat any residue over again in the same way, although the amount of titanium usually found by a second fusion is very small. To the solution add 1 c.c. of the strongest hydrogen peroxide and a few cubic centimeters of dilute sulphuric acid, when the solution is ready for comparison in a Nessler tube with a standard. To prepare a standard solution, dissolve pure titanic oxide in hot strong sulphuric acid, add dilute sulphuric acid at first to prevent precipitation, then water until 1 c.c. of the solution con- tains 1 mg. of titanic oxide, TiO2- A new volumetric method has recently been devised by P. W. Shimer. 3 ZINC 15. Western Method for Ores and Mattes. This separa- tion is designed for rapid routine work. The principle involved is the evaporation of the acid solution of the sample and the 1 Berichte, 1882, 2592. 2 Trans. A. I. M. E. 14, 763. 3 Report to Eighth Inter. Congress of Appl. Chem. DETERMINATION IN ORES, SLAGS, AND MATTE 117 direct extraction of the zinc from the residues by boiling with a large excess of ammonia and apimonium chloride. Add to a .5-gram sample in %e started, as already directed, and allowed to discharge into the air at the three-way valve. A quick turn of two glass valves will then substitute carbon dioxide for hydrogen, and twenty minutes additional heating in the carbon dioxide will expel the hydrogen (about .01 per cent) which would otherwise have been retained by the copper. Cool in the current of carbon dioxide for ten minutes, using a small air blast, and then replace the gas by dry air from a pipe attached temporarily to the three-way valve. Place in the balance and weigh in ten minutes. To determine the correction for the traces of sulphur evolved, place 70 c.c. of water in a tall No. 1 beaker, and transfer the cadmium solution and sulphide as rapidly as possible with the . aid of 30 c.c. of (1 : 1) hydrochloric acid. Titrate at once with standard iodine. If 2 grams of iodine are dissolved in one liter of water, 1 c.c. will equal .00025 gram of sulphur. Deduct the weight of sulphur from the apparent loss of weight of the copper, and the result will express the weight of oxygen in the metal. A check, or experimental proof of the completeness of the reduction, may be obtained by a careful electrolytic assay of the reduced drillings from the tube, correcting the assay for the trace of copper in electrolyte. Rapid Method. In routine work, results may be obtained in quicker time by heating several samples in one tube in por- celain boats for two hours. If cooled in hydrogen (after Cobel- dick), the method is more rapid, but approximate. If finally heated and cooled in carbon dioxide, as recommended by the author, the only appreciable error is that due to the loss of a trace of sulphur. 15. Photo-micrographic Method. Following Heyn, 1 and the work of Hofman, Green & Yerxa 2 in the measurement of a microscopic field with a planimeter, E. S. Bardwell 3 projects the area of the microscopic field directly upon a piece of duplex paper in a 16-inch circle, and then traces and cuts out the copper areas, leaving a net-work of paper, representing the eu- 1 Mittheil. aus den Konigl. Versuchtsanstalten zu Berlin 18, 315. 2 Trans. A. I. M. E. 34, 671. 3 Ibid. Bull. 79 (1913), 1429. 232 ANALYSIS OF COPPER tectic. The two lots of paper are carefully weighed and the weights are proportional to the areas of copper and eutectic (Cu + Cii20 containing 3.45 per cent Cu 2 O). The polished copper is etched by heating it for three or four minutes in a current of hydrogen gas at about 300 C., or low red heat, first passing the gas for ten minutes to drive out the air. This scheme can only be approximate in the case of cast copper because there is so much variation in the crystalline aggregates, and it requires about as much time per sample as the rapid method already described in 14. TIN IN COPPER 16. Special Method. The separation of tin from antimony or arsenic has been described (10, Chapter XII). Copper refined from foundry scrap, or from metal which may have been elec- trolyzed with lead anodes, will occasionally contain a trace of tin. The following is developed from the method of W. H. Bassett, devised for this material. Dissolve two portions of 75 to 100 grams each in 500 c.c. of distilled water and 250 to 300 c.c. of pure, strong nitric acid. The distilled water employed should not be distilled from a tin- lined condenser. After the acid has been added gradually and the copper has dissolved, boil the solutions down to a sirup. When a skin begins to form on the surface, cool slightly, dilute with a few cubic centimeters of nitric acid and 600 c.c. of water, and heat until all is dissolved. If a little basic copper remains, add a very little acid until it clears up, and allow to stand on the steam plate for at least two hours but better overnight. The solution should not be allowed to cool until it is filtered. Pass both solutions through the same small doubled filter, reserving all insoluble residue until the last. Remove the large beaker and replace by a small clean beaker, then trans- fer the residue to the filter. The precipitate may contain a trace of iron and phosphate of tin, also a part of the antimony present. It may be purified by the method used in the analysis of bronze, Chapter XIV. Dissolve the tin oxide, etc., in a little yellow ammonium sulphide containing about 3 per cent of pure ammonium chloride. Filter, precipitate the tin with a slight excess of acetic acid, and filter again. Wash with water acidified with acetic acid, and ignite carefully to METHODS FOR FOREIGN METALS IN COPPER 233 oxide with the precautions noted in Fresenius' Quantitative Analysis. SELENIUM AND ^TELLURIUM 17. Special Method. The rapid combination method of the author for the estimation of the two elements, in connection with the arsenic and antimony, has already been described (7, Chapter XII), and is recommended. The papers of E. Keller 1 and C. Whitehead 2 present two original schemes which are slower, but are capable of equally good results. The elements are separated by excess of ferric hydroxide. The copper must be completely removed by repeated treatment, or a loss will result when the ferric hydroxide is finally dissolved in acid and the solution treated with hydrogen sulphide. If any silver, or copper, is present, some silver or copper selenide may be formed, which is afterwards insoluble in alkaline sulphides. Proceed as in the next paragraph. Dissolve two portions of 50 grams of refined copper in sepa- rate beakers, using in each case 200 c.c. of nitric acid (d., 1.42), of tested purity. Add to each 30 c.c. of a 10 per cent solution of ferric ammonium sulphate, or 2 grams of ferric nitrate, render sufficiently ammoniacal to redissolve all the copper hydroxide formed, then heat to boiling and allow to settle. Filter, wash with dilute ammonia, wash most of the ferric hydroxide into the original beaker, and dissolve in (1 : 20) sulphuric acid with a little hydrochloric. Precipitate again and filter on the same filter. Repeat the operations until the filtrate shows no blue color of copper. Finally, dissolve the hydroxide in the least possible amount of hydrochloric acid diluted with 10 volumes of water. From this point, there are two or three separations which are a matter of individual preference. (a) Modification of E. Keller. Saturate the cold hydro- chloric acid solution of the ferric hydroxide, etc., with hydrogen sulphide gas. Copper must be absent and the solution must be cold to render the selenium sulphide soluble in sodium sulphide. Filter, wash, and digest with a little sodium sulphide solution. Filter off insoluble matter, wash the filter, acidify the solution with nitric acid, and evaporate to dryness. The latter operation must be performed with care on the water bath or steam plate. If five drops of sulphuric acid are added during concentration, no 1 J. Am. Chem. Soc. 22, 242. 2 /^ 17> 2 80. 234 ANALYSIS OF COPPER nitric acid will remain in the residue. To this residue add 180 c.c. of hydrochloric acid (d., 1.2) and 20 c.c. of water. Boil just long enough "to destroy any nitrous compounds and reduce selenium and tellurium to the lower chlorides. If the salts are not excessive, 90 c.c. of acid with 10 c.c. of water are sufficient, and, if the sulphuric acid was added exactly to replace the nitric, it is better to heat just to boiling and then remove to prevent loss of selenium. Cool the solution and filter out any sulphur or insoluble halide through an asbestos felt on a Gooch crucible. Wash the residue several times with 90 per cent hydrochloric acid. The filtrate is now ready for saturation with sulphur dioxide gas, which may be supplied from a steel cylinder, or generated by copper borings and acid. The author prepares the sulphur dioxide more easily from a solution of sodium sulphite which is allowed to drop into strong sulphuric acid (Chapter III). Charge the hot solution with the gas, allow to cool while charging, and set the beakers away until the precipitate has settled. Filter the selenium on a weighed Gooch felt and wash three times with 90 per cent hydrochloric acid. Set the filtrate aside and wash the selenium free from salts by successive treatment with dilute hydrochloric acid, water, and strong alcohol. Dry at 100 to 105 C. for one hour, cool, and weigh immediately. Saturate the solution with the sulphur dioxide a second time to be sure of complete precipitation. The filtrate from the second precipitation contains the tel- lurium. Add water to double the volume, then boil the solution for several minutes, while sulphur dioxide is again conducted through the liquid. Filter when cooled, wash, dry, and weigh as before. The felts containing both selenium and tellurium may be ignited to drive off the two. elements, and the felts reweighed as a check. With some asbestos, there is a notable difference between the ignited weight and the weight after moistening the ignited felt and drying in an oven at 105 C. The latter weight corresponds to the original dried weight of the felt. (6) Modification of C. Whitehead, The preceding method is exactly followed until the solution of ferric hydroxide has been obtained, strictly free from copper salts. Add 1 gram of solid tartaric acid, make alkaline with an excess of potassium hydrox- ide, and pass hydrogen sulphide gas into the solution for thirty METHODS FOR FOREIGN METALS IN COPPER 235 minutes. Filter, decompose the sulphides of selenium and tellurium with dilute hydrochloric acid and allow the liquid to stand in a warm place until tfe hydrogen sulphide is removed. Filter again, dissolve the sulphides in aqua-regia (nitric and hydrochloric acids), add .2 gram of potassium chloride, and evaporate the liquid to dryness on the steam plate. Take up with 90 per cent hydrochloric acid, heat to boiling, and precipitate with sulphur dioxide as in the first modification. (c) Precipitation by Stannous Chloride. When the selenium is largely in excess, a fairly good separation is most easily obtained by adding an excess of stannous chloride to the hot ferric chloride solution until the iron is decolorized, and allowing the covered beaker to heat until the liquid boils. Allow to settle overnight, if possible. Prepare an asbestos felt by extracting with hydrochloric acid, igniting it, and then taking the "ignited" weight. Moisten the felt, dry in the oven, cool, and take a " dried" weight. After weighing the washed and dried precipitates of selenium or tel- lurium, or both, deduct the tare weight obtained by drying. Finally ignite the felt as a check. A felt will occasionally suffer a slight loss in washing, in spite of the usual care. Selenium and tellurium are separated, if desired, by fractional precipitation with sulphur dioxide as in (a). Refer to the shorter combination method of the author, in connection with the arsenic and antimony determination (7, Chapter XII). Zinc. The separation of zinc is included and described with the method for cobalt (6) and nickel (7). PART IV CHAPTER XIV ANALYSIS OF THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER Introduction. This chapter presents the standard technical methods of one of the largest American brass companies. A description forwarded by Dr. Toisten of the Mansfeld Brass Works shows that the European methods for straight brass or bronze are practically identical with those to be described. Tests for nickel and spelter are also included, the latter being the one proposed as a standard by the Committee on non-ferrous alloys of the American Chemical Society. The analysis of antifriction metals and other special complex alloys is well described in another recent work. 1 COPPER IN ORDINARY BRASS 1. In the Absence of Tin. One gram of drillings are weighed into a tall 200 c.c. beaker and dissolved in a mixture of 5 c.c. of sulphuric acid (d., 1.84), 2 c.c. of nitric acid (d., 1.42), and 18 c.c. of water. (As in previous chapters, all acids are understood to be of full strength and of the best grade, unless otherwise specified.) Dissolve the sample by heating, dilute the solution to about 120 c.c., and introduce a split platinum cylinder, having a total surface of 100 sq. cm. Electrolyze overnight with a current density of .5 ampere. The process may be hastened by the use of rotating electrodes and gauze cathodes, or preferably by the use of the Frary selenoid 2 described in Chapter I. If the assays are finished by the slow process, wash off the electrodes and split watch-glass covers at the beginning of work the next day, and continue the current for about one hour longer. Test the electrolyte for the end-point by transferring about 1 c.c. of the liquid to a porcelain test plate and adding a few drops of fresh hydrogen sulphide water. Continue the elec- 1 Price & Meade, The Technical Analysis of Brass. 2 J. Am. Chem. Soc. 29, 1592. THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 237 trolysis until there is no visible discoloration in the test. The cathode is removed quickly, plunged into a large beaker of water, dipped twice into alconol, the excess shaken off, and the remainder burned, moving the cathode continually. (Colorless denatured "pyro" alcohol of about 94 per cent strength is as safe to use as the best grade, at a fraction of the expense.) The use of dilute mixed acid avoids the necessity of evaporating to fumes, as the nitric acid remaining in solution secures uniformly good results. LEADED BRASS CARRYING TIN 2. Technical Estimation of Tin. (In this connection, refer also to alloys with iron and phosphorus.) To 1 gram of drill- ings or clippings add 10 c.c. of strong nitric acid. When the action has ceased, bring to a boil, add 50 c.c. of boiling water, and let stand until the metastannic acid has settled (about one hour), keeping the temperature just below the boiling point. It is important to keep this solution hot and filter hot, for if the liquid cools, the metastannic acid becomes partly soluble. Filter off the tin on a double 7 cm. paper, keeping the solution hot. It may be necessary to return the first portion of solution to the filter in order to have it run clear. J. T. Baker's ashless filters hold the tin better than any other tried. Some of the especially close grades of washed papers hold the moist metastannic acid properly but work so slowly that nothing is gained by using them. Wash the tin residue with boiling water and ignite, while moist, in a porcelain or platinum crucible, slowly at first and finally to the full heat of a Tirrell burner. The tin oxide must be ignited to a constant weight, and if more than 20 mg. in weight, the blast lamp is necessary for the final heating. Weigh as tin dioxide, SnC>2 factor, .7881. Limitations. The metastannic acid, obtained in this way, is free from copper, lead, zinc, and nickel, but will contain iron if it is present, also phosphorus. Also, the preceding method for tin is inaccurate in the analysis of an alloy containing iron in any considerable amount. Metastannic acid precipitated by nitric acid usually carries a part of the iron present. If the quantity of iron approximates that of the tin, nitric acid will render only a portion of the tin insoluble even with evaporation to hard dryness, but it is very unusual to find such conditions 238 ANALYSIS OF COPPER in wrought brasses and bronzes. Ordinarily, iron is present in such small amounts as to be negligible in its effect. If much iron or phosphorus is present, proceed as described in the methods to follow : " Alloys containing tin with iron," and " Alloys containing phosphorus." 3. Technical Method for Lead. To the filtrate from which the metastannic acid has been removed, or to the nitric acid solution of the alloy if no tin is present, add 40 c.c. of "lead acid/' the preparation of which is given in the following para- graph. Evaporate to fumes and allow to cool. Take up with 35 c.c. of water, heat to boiling, then allow to cool and settle for five hours better overnight. Filter off the lead sulphate on a Gooch crucible, wash with "lead acid," and remove the filtrate. Wash out the "lead acid" with a solution of equal parts of water and alcohol, finally with alcohol alone. Ignite and weigh the lead sulphate with the precautions noted in Fresenius's "Quanti- tative Analysis." (The factor for lead is .6831.) Composition of Lead Acid. This is a solution of one volume of sulphuric acid (d., 1.84) in seven volumes of water, saturated with lead sulphate. The solution is prepared as follows : 300 c.c. of sulphuric acid are poured into 1800 c.c. of water; 1 gram of lead acetate is dissolved in 300 c.c. of water and added to the hot liquid with stirring. The solution is allowed to settle for three or four days and pouted off through a thick asbestos filter for use. In dealing with small amounts of lead, it has been found in the precipitation of lead in (1 : 20) sulphuric acid (the best dilution) that the volume of solution is often so large in propor- tion to the lead present that a serious loss results. When lead acid is used, it is unnecessary to consider the solubility of the lead sulphate (that is, in technical foundry work), since the solution is always brought back to the same volume as the volume of lead acid originally added. Consequently, when the lead sulphate is filtered, no more lead remains in the filtrate than was originally added in the dilute sulphuric acid. NOTE. In a mutual investigation of a sample of complex "rolling-mill brass" for the U. S. Bureau of Standards, it has recently been determined that in an exact analysis of such material, the electrolytic deposition of lead from nitric acid THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 239 solution as the peroxide is 111019 accurate than precipitation as sulphate. Copper and zinc salts have a slight solvent action on sulphate of lead, so that the assumption on which leaded acid is prepared and used is not strictly true, if the solution contains much copper and zinc. After the lead peroxide has been cor- rected for a trace of tin oxide and iron oxide which it may contain, the result is still a little higher by electrolysis. 4. Lead by Electrolysis. Lead in leaded brass may also be determined by dissolving 1 gram in 10 c.c. of nitric acid and proceeding according to the exact electrolytic method 8, described under the title " Impurities in Brass." One gram is the quan- tity used in works tests for the rapid determination of foundry mixtures. 5. Copper in Leaded Brass. To the filtrate from the lead sulphate in 3, add 3 c.c. of (1 : 1) nitric acid, bring to the proper volume, and electrolyze as in the first method for copper (1). 6. Zinc in Leaded Brass. This element is usually taken by difference but may be precipitated as phosphate. The electro- lyte from 5 may also be titrated as described in the " Analysis of German silver" (23). EXACT DETERMINATION OF IMPURITIES IN BRASS 7. Lead as Sulphate, in Absence of Tin. When the amount of lead is small (less than .5 per cent), a 5-gram sample is dis- solved in 25 c.c. of nitric acid, 120 c.c. of lead acid is added, and the liquid, if tin is absent, evaporated to fumes. Take up with 105 c.c. of water, boil, settle, and filter as described under " leaded brass." If tin is present, as is shown by a milkiness in the nitric acid solution after expulsion of nitrous fumes, add 125 c.c. of boiling water, allow the metastannic acid to settle, and filter off exactly as for the 1-gram sample. Add 120 c.c. of lead acid to the filtrate and proceed as above directed. 8. Exact Electrolytic Method for Lead. For small amounts of lead the previous method has been largely superseded by "the electrolytic, which is more exact (as noted in 3), if corrections are made for traces of oxides of tin and iron contained in the anode deposit, when these elements are in solution with the lead and copper. If manganese is present in the alloy, it will also partially deposit with the lead. 240 ANALYSIS OF COPPER Weigh a 5-gram sample into a 250 c.c. wide beaker. Dissolve in 25 c.c. of nitric acid and dilute to 150 c.c. with water. Elec- trolyze with current reversed, i.e., with the sheet electrode for the anode, first, for one-half hour at .1 ampere per solution and then three hours at .5 ampere. Test by. exposing a clean surface to the liquid. When this test shows the deposition to be complete, remove the electrode, wash with water, and then with alcohol. Dry in an oven at least half an hour, and finally cool and weigh as lead dioxide. For lead the factor .8643 is used, following the results of Dr. E. F. Smith and others. If the lead is excessive, it is best to employ a gauze electrode to prevent flaking of the deposit. In Presence of Tin. Dissolve 5 grams in 25 c.c. of nitric acid, dilute with 125 c.c. of boiling water, and filter off the metastannic acid exactly as in 2. In very exact work, dissolve the tin compound in hot ammonium sulphide containing am- monium chloride, filter, dissolve the washed black sulphides from the paper, and return them to the copper solution. In the solu- tion determine the lead as already directed. The deposit may be removed after weighing by the solvents described under " Electrolysis of Lead in Copper" (page 222). To obtain the correct weight of the platinum, it should be ignited and weighed again after cleaning, as it loses weight perceptibly during electrolysis. 9. Iron, Exact Method. Dissolve 5 grams of brass in 25 c.c. of nitric acid, boil off nitrous fumes, dilute, and add ammo- nium chloride and ammonia in excess. Boil, filter on an S. & S. black ribbon paper, wash with dilute ammonia and then with hot water. Dissolve the iron precipitate in hot (1:1) hydro- chloric acid and reprecipitate with ammonia. Dissolve the ferric hydroxide after the second precipitation in hot (1 : 4) sul- phuric acid on the filter, washing out thoroughly with dilute acid and hot water. Add 40 c.c. of (1 : 1) sulphuric acid, pass through a Jones' reductor (or add hydrochloric acid and decolorize with one drop of stannous chloride in excess). Then titrate with potassium permanganate. For the use of the reductor, consult "The Chemical Analysis of Iron" by Blair. If the reduction is made by stannous chloride, add 5 c.c. of a saturated solution of mercuric chloride and 10 c.c. of titrating solution to secure a good end-point. (See 7, Chapter V.) For brass, German silver, THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 241 and spelter make the permanganate solution with .3 gram of the crystals per liter, and standardize against .015 to .018 gram of pure iron wire. One c.c. (ft solution then equals about .0005 gram of iron. In Presence of Traces of Tin. In such a case, the insoluble metastannic acid must be removed as in 8, and the filter and contents extracted with about 30 c.c. of warm yellow ammonium sulphide, containing 1 gram of ammonium chloride. The iron sulphide, thus obtained, is dissolved in dilute nitric acid and returned to the original solution of the alloy. ALLOYS CONTAINING TIN WITH IRON (Examples: Manganese Bronze, Phospho-Bronze) 10. Copper Assay, for Control of Mixtures. Dissolve a 1-gram sample in 10 c.c. of nitric acid, boil, wash down, and bake thoroughly. Moisten with nitric acid, dilute, boil, allow to settle, and filter off all possible tin oxides, exactly as for brass (2). The filtrate contains some of the tin, notwithstanding the baking. Add to the filtrate 10 c.c. of (1 : 1) sulphuric acid and evaporate to strong fumes. Dilute to about 120 c.c. and elec- trolyze as usual for copper. When the deposition is complete, the solution may show a yellow color on the spot plate with hydrogen sulphide water, due to the tin present in solution, but there should be no darkening of the cathode copper. If darken- ing occurs, the cathode deposit may be dissolved in nitric acid and the tin removed easily (since iron is no longer present). The copper may then be reprecipitated. This purification is seldom necessary in routine work. 11. Exact Electrolysis for Copper. The next modification was developed by the author in analyzing the " Rolling Mill Brass" for the U. S. Bureau of Standards. Dissolve 5 grams of drillings in 25 c.c. of nitric acid and 75 to 125 c.c. of water, then filter off and purify the metastannic acid with warm yellow ammonium sulphide exactly as in 8. Dissolve the trace of copper sulphide, etc., in 5 c.c. of nitric acid, diluted with 2 parts of water, and return it to the main solution which has been already treated with 10 c.c. of sulphuric acid and evaporated to fumes. The diluted solution, which contains traces of dissolved tin, should then be filtered from the lead sulphate. Five cubic centimeters of nitric acid should be present in the 242 ANALYSIS OF COPPER solution. It is possible to deposit the copper and lead simultane- ously from a pure nitric acid solution, if preferred, the one on the cathode, and the other on the anode, which should also be a cylinder, or dish. The trace of tin in the first deposit of copper is easily re- moved by standing the cathode in a tall beaker and adding 40 c.c. of the "acid mixture" used for the electrolytic assay of re- fined copper (2, Chapter XI). Dilute with sufficient water to cover the plate, cover the beaker with a glass perforated with a small hole for the stem of the electrode, and allow to stand on the steam plate until dissolved. Repeat the electrolysis. As the second electrolyte is free from iron, the last deposit will be free from tin, and the second electrolyte may be subsequently com- bined with the first to obtain a complete recovery of the tin. Rapid Assay. Use a rotating anode, or better, the Frary solenoid, observing the precautions indicated in the rapid assay of refined copper (Chapter XI) : Current requirement, 4 to 4.5 amperes per square decimeter of immersed cathode surface. 12. Tin and Iron in Bronze (Special Method) . This modi- fication was reported by Bassett and Merrill in the assay of U. S. standard " Rolling Mill Brass." Five-gram samples should be treated, if- results are to be reported to .01 per cent. Otherwise, dissolve a separate 1-gram sample in 10 c.c. of hydrochloric and 5 c.c. of nitric acid. Dilute, and make a double precipitation with ammonia to separate tin and iron from the copper. Dis- solve the oxides the first time in hot (1:1) hydrochloric acid and the next time in very hot (1 : 2) sulphuric acid. Wash very thoroughly with this acid. Dilute, filter off any trace of lead sulphate, and saturate the solution with hydrogen sulphide. Filter off the tin sulphide, wash, and ignite to stannic oxide, Sn(>2, as described under "phospho-bronze." Heat the filtrate from the tin sulphide to expel the hydrogen sulphide, oxidize with nitric acid, and add ammonium chloride and ammonia in excess. Boil, filter, wash the ferric hydroxide, dissolve in dilute sulphuric acid, and titrate as in method 9 for iron in brass. T. J. Demorest has recently devised a new scheme for complex alloys which deserves a trial (see reference 1 )- 13. Tin in Phospho-Bronze. In alloys of this kind, copper, zinc, and tin are determined in one sample. The metastannic 1 J. Ind. and Eng. Chem. 6, 842. THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 243 acid (containing phosphorus) js? filtered off exactly as in the analysis of brass (2), and the jcopper and zinc estimated as described. The paper with the tin precipitate is extracted in a small beaker with ammonium polysulphide and ammonium chloride, then filtered and the residue washed thoroughly with dilute yellow ammonium sulphide. The lead and iron sulphides are filtered from the dissolved tin, and the polysulphide solution is made acid with acetic acid (instead of sulphuric), heated to boiling, and filtered. The precipitated tin sulphide is ignited directly with the moist paper to stannic oxide, SnO 2 , taking care to ignite to constant weight. This ignition must proceed slowly and carefully until the sulphur is roasted off. The sulphur should not be allowed to burn, for some tin sulphide might be volatilized. The ignition is best started in a porcelain crucible, placed upright in a hole in a piece of asbestos board in which it fits tightly near the top. When the sulphur is expelled, the cru- cible is ignited in a slanting position upon a triangle. If the tin oxide weighs over 20 mg., finish with the blast lamp. Impure metastannic acid may also be purified by igniting, mixing with three parts of sodium carbonate and three parts sulphur, and fusing in a porcelain crucible. Some chemists prefer to place the small crucible inside a larger one. 1 The fusion is taken up with hot water, boiled to insure complete precipitation of foreign sulphides, filtered, and the tin finally separated with acetic acid as above described. PHOSPHORUS 14. By Titration. Dissolve a separate 1-gram sample in a casserole by an acid mixture containing two parts of hydro- chloric acid and one of nitric acid. Evaporate to dryness. Moisten with hydrochloric acid, evaporate to dryness again, and heat to dull redness. (This removes most of any arsenic present.) Moisten with hydrochloric acid, add water and about 3 c.c. of strong ferric chloride solution, make alkaline with am- monia, boil and filter, wash with dilute ammonia, and then with hot water. Dissolve the iron precipitate (which carries the phosphorus) in hot (1 : 1) hydrochloric acid, dilute, and satu- 1 Blair, The Chemical Analysis of Iron, 7th Edition, p. 95; Year Book Amer. Soc. Testing Materials (1915). 244 ANALYSIS OF COPPER rate the liquid with hydrogen sulphide. Filter off the tin and copper sulphides, oxidize the filtrate with nitric acid, and repre- cipitate with ammonia. Dissolve the ferric hydroxide with hot (1:4) sulphuric acid into an Erlenmeyer flask. Add ammonia to the solution in the flask until it is alkaline, make slightly acid with nitric acid, add ammonium molybdate solution, and pro- ceed as in the determination of "phosphorus in steel." The contributing chemists employ the ferric-alum modifica- tion of the permanganate titration, adopting the same strength of solution as for iron or manganese, 3 grams of crystals per liter. The " yellow precipitate" may be dissolved in ammonia, if preferred, the solution made acid, then alkaline, and treated with excess of magnesia mixture. MANGANESE 15. Manganese in Alloys. If the sample contains tin, dis- solve 5 grams in 25 c.c. of nitric acid, dilute with 125 c.c. of boiling water, and filter off the metastannic acid as described under " Brass" (2 to 8). If tin is absent, dissolve as before with- out filtration. In either case, evaporate the solution to a sirup in a tall 500 c.c. beaker, add 100 c.c. of nitric acid and 5 grams of potassium chlorate, and proceed as in William's modification of Ford's method. 1 In this process, the first precipitate of manganese dioxide may contain a little impurity, which would require re-solution and reprecipitation with nitric acid and potassium chlorate. The ferrous ammonium sulphate solution should be composed of 100 grams of the crystals, 50 c.c. of sulphuric acid (d., 1.84), and 2000 c.c. of water. Usually, 10 c.c. of solution are sufficient to dissolve the precipitated manganese dioxide. The potassium permanganate contains 3 grams of the crystals per liter, or ten times the strength used for iron in brass (9). It is standardized against .10 to .12 gram of pure iron wire. To estimate the manganese, dissolve the dioxide in 10 c.c. of the ferrous sulphate solution, protecting it from the air. Immediately titrate back the unused ferrous salt with the permanganate. Fe X .4919 = Mn. 16. Gravimetric Method. Manganese may also be deter- mined by Ford's method, 1 dissolving the sample and oxidizing the manganese as in the previous method. The first precipitate 1 Blair, Chemical Analysis of Iron, 7th Edition, 114-118. THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 245 should be dissolved and reprecipitated for purification. After dissolving the purified dioxide/ the manganese is finally thrown down by a large excess* of amrifonium phosphate in a boiling, slightly ammoniacal solution. Add the precipitant slowly to the hot liquid, stirring rapidly. Continue heating and stirring until the crystals assume the true color and silky appearance. Filter, wash, and dry the precipitate; then ignite carefully to avoid any reduction. Use the conversion table of Chapter XVI. The sodium bismuthate method is also very exact. (See method 25, Chapter VI.) ALLOYS CONTAINING A LITTLE NICKEL 17. Nickel. In alloys where there is a relatively low ratio of nickel to zinc (not over 1 to 15), a method is desirable which will permit a quick and accurate separation of the small amount of nickel from the large amount of zinc, with the final deter- mination of the nickel by electrolysis. In the complete analysis of any complex alloy, the filtrate from tin and iron is taken for the separation of nickel and zinc. Precipitate the nickel by dimethyl glyoxime from the hot, slightly ammoniacal filtrate according to the special method below, and follow with the estimation of the zinc in the filtrate from the nickel. The sodium potassium tartrate is omitted, as no iron is present. 18. Special Method for Nickel. Dissolve a special 1-gram sample exactly as in the determination of copper in brass (1, 5), and remove the copper by electrolysis. A little copper in the electrolyte does no harm. Concentrate the solution to about 100 c.c., add 1 gram of ammonium chloride, then 10 c.c. of a 20 per cent solution of sodium potassium tartrate. Neutralize with ammonia, adding about 2 c.c. in excess, and then stir in 5 c.c. of a 1 per cent alcoholic solution of dimethyl glyoxime for each per cent of nickel. This method is due to O. Brunck and Tschu- gaeff. 1 Allow the solution to stand for one-half hour, filter, and wash with very dilute ammonia. Dissolve the red precipitate in hot (1 : 1) hydrochloric acid, add 10 c.c. of (1 : 1) sulphuric acid, and evaporate to strong fumes of sulphur trioxide. Dilute, neutralize with ammonia, and add 20 c.c. in excess; electrolyze 1 Zeit. Angew. Chem. 20, 834 and 3844. 246 ANALYSIS OF COPPER with a current of .3 ampere per solution. Test for the end-point with fresh hydrogen sulphide water, as in the " determination of copper in brass" (1 and 5). NOTE. The Frary solenoid may be used to advantage with a current of 4 amperes per solution. With this rapid circulation and heavy current the time is reduced to one hour. SULPHUR IN BRASS OR NICKEL ALLOYS 19. Precipitation as Barium Sulphate. Weigh duplicate 10- gram samples of drillings into a 600 c.c. tall-lipped beaker. Add about .5 gram of pure sodium carbonate. Cover and then add 50 c.c. of nitric acid. When the action has ceased, boil off red fumes, evaporate off the bulk of the solution, and allow to bake overnight at tlje temperature of the steam bath. Fill to the lip with warm water. There should be present a layer about i inch (6 mm.) thick, composed of basic salts of copper. Add 3 c.c. of (1 : 1) nitric acid and electrolyze, placing the beaker in a large Frary solenoid, or rotary device. The lead may be removed at the anode as copper is taken out at the cathode. When the copper is gone, evaporate to small volume, cover the beaker, add 75 c.c. of hydrochloric acid, and boil down to a small volume. Add 75 c.c. more hydrochloric acid and evap- orate to dryness. Take up with water, make alkaline with ammonia, and dilute to the lip of the beaker. Then proceed to remove the nickel and zinc by electrolysis, using a large iron cathode and the Frary rotary device. When the solution is colorless, wash down the electrodes; remove them, and heat the solution to boiling. Let settle and filter through an 11 cm., black ribbon paper into a 600 c.c. tall beaker. Make the solution acid with hydrochloric acid and evaporate until the ammonium chloride starts to crystallize. Cover and add 50 c.c. of nitric acid. Boil down to small bulk; add 25 c.c. of hydrochloric acid and boil until no more chlorine is evolved. Wash down the cover and evaporate to dryness to dehydrate any silica present. Moisten with hydrochloric acid and take up with 20 c.c. of hot water. Filter into a small beaker. Heat the filtrate to boiling, add 5 c.c. of 5 per cent barium chloride solution, drop by drop, with stirring. Allow to settle at least five hours, filter on* a small ashless paper, and wash with hot THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 247 water. Ignite cautiously in .%a small porcelain, or platinum, crucible, and then heat to bright redness for twenty minutes. Cool and weigh the- barium sulphate. BaSC>4 X .1374 = sul- phur (S). NOTE. The utmost precautions should be taken to prevent contamination with sulphur in any form. Avoid gas-heating, except by gasolene gas. Run a blank analysis; that is, take a 600 c.c. beaker, add to it all the reagents, and subject the con- tents to the same operations as the samples of drillings, and at the same time. Deduct any barium sulphate obtained from the result of the actual assay. If this blank amounts to .0020 gram, it is a sign of poor work and the analysis should be repeated. Wash out all beakers, funnels, etc., with distilled water before use. ZINC IN BRASS AND BRONZE 20. Volumetric Method. The filtrate from the iron pre- cipitation, or from the nickel separation when that element is present, is taken for the determination of zinc. In brass works, the custom is to employ rapid titration. Add 20 c.c. of nitric acid (d., 1.42) to the filtrate from the nickel precipitation. Boil, add some ferric chloride, citric acid, and ammonia; then titrate with potassium ferrocyanide exactly as directed in the analysis of nickel alloys (23). 21. Gravimetric Method. For an occasional analysis, or in umpire work, some chemists prefer a precipitation and igni- tion of zinc phosphate. The filtrate from the nickel separation (16) may be used after boiling out the alcohol, or a new sample may be treated as follows : Precipitate lead and copper from the acid solution of the sample on sheet aluminum according to (25) Chapter VI. To the final solution (volume about 50 c.c.) add 50 c.c. of a 10 per cent solution of ammonium phosphate, make just alkaline to litmus, then acid with 1 c.c. of acetic acid in excess. Heat just below boiling for about an hour. This treat- ment should cause the zinc to become granular. Allow to settle, filter, wash with hot water, and dry the substance. Char the filter separately, ignite the phosphate with care, cool the crucible, and weigh the precipitate. (To obtain the weight of the zinc, multiply the weight of the ignited pyrophosphate, Zn 2 P2O7 , by .4289. 248 ANALYSIS OF COPPER ANALYSIS OF NICKEL ALLOYS (Example, German Silver, Monel Metal.) 22. Copper. The copper is determined as in brass. In arranging the cathode, allow it to project about J inch above the solution. When the copper is apparently all precipitated, wash down the cover glasses and sides of the beaker, which will cause the solution to cover the cathode. If no sign of copper is seen on the clean platinum, it indicates that the deposition is complete. If copper does appear, the electrolysis must be continued several hours, or until the solution polarizes and gas bubbles come off freely from the cathode. 23. Zinc. After removing the copper as above, concen- trate the solution to about 75 c.c. and transfer to a tall 500 c.c. beaker, make just alkaline with ammonia, then acid by adding 2 c.c. of formic acid (d., 1.2), then add enough ammonia to barely restore the blue color. Finally, add formic acid of the same density in the proportion of 38 c.c. for every 200 c.c. of final volume of solution. Heat to boiling and pass a rapid stream of hydrogen sulphide gas into the liquid for fifteen minutes, using a fine jet tube. Start the gas before putting the tube into the liquid, for if the interior of the tube becomes wet with the solu- tion, nickel sulphide will deposit on the inside of the tube. Filter off the pure white zinc sulphide and wash with hot water; dis- solve in hot dilute hydrochloric acid; filter off the sulphur and wash the paper. Any slight black residue may be dissolved in a few drops of nitric acid and returned to the nickel solution. The nickel solution should then be heated up and more hydrogen sulphide passed to make sure that the zinc is all out. To the hydro- chloric acid solution of the zinc sulphide, add nitric acid and boil to oxidize the hydrogen sulphide. Cool and add 3 c.c. of a strong solution of ferric chloride, 20 c.c. of a saturated citric acid solu- tion, and ammonia to make distinctly alkaline. A very large excess of ammonia should be avoided. Heat this solution, which should have a volume of from 250 to 300 c.c., to boiling, and titrate with potassium ferrocyanide solution. Titration. To determine the end-point with ferrocyanide, the pits in the porcelain test plate are filled with a 50 per cent acetic acid solution. When the titration is nearly completed (which may be judged from the conditions), two drops are taken THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 249 out and added to the acetic acid in one of the pits. This gives a slight greenish tint, and a few drops more of the ferrocyanide are added and another portion taken out. This is continued until the end-point, which is a distinct blue, is reached. On making acid with the two drops of acetic acid, a Prussian blue is not formed until all the zinc has been precipitated as ferro- cyanide. It requires about J c.c. of the ferrocyanide to give a distinct end-point in 250 c.c. of solution containing no zinc. The ferric chloride is added as an indicator, but if the solution con- tains sufficient iron, as it does in the case of certain zinc ores, the addition of this iron is not necessary. Small amounts of aluminum, lime, and magnesium make no difference in this titration. The correction for the blank in getting the end-point must be made by the operator. This titration is difficult for an inex- perienced operator, but for one used to the method it is one of the most accurate methods for the determination of zinc and is both rapid and easy to handle. It is accurate within about .2 per cent and the experienced operator can check his own results within about .1 per cent. This titration must be carried out in every case in exactly the same manner as the standardiza- tion. The potassium ferrocyanide solution is made up by dissolving 80 grams of the salt in 2500 c.c. of water and standardizing against a solution of pure metallic zinc. The potassium ferro- cyanide should be allowed to stand at least six weeks before use; in that time changes which occur in such solutions will have completed themselves, and the standardization value will remain permanent. This value is ascertained by weighing out 2 grams of zinc, dissolving the metal in nitric acid, and making up to 1000 c.c. Take 100 c.c. of the solution, add 5 c.c. of nitric acid, 3 c.c. of ferric chloride, and 20 c.c. of saturated citric acid solution, dilute, and make distinctly alkaline with ammonia. The final volume should be 250 c.c. Boil and titrate while boiling hot with the ferrocyanide solution and figure the zinc equivalent. 24. Alternative Titration for Zinc. This alternative method is based on the paper entitled "A Proposed Standard Method," written by Frank G. Breyer. 1 1 8th Int. Cong, of Appl Chem. 1, 162. 250 ANALYSIS OF COPPER Proceed as in 23 until the pure white zinc sulphide has been filtered off and washed with hot water. Return the paper and precipitate to the original 500 c.c. beaker and add 20 c.c. of (1:1) hydrochloric acid. When the zinc sulphide is dissolved, filter off the paper together with some sulphur and possibly a little nickel sulphide, which may be treated as directed in 23. Boil the solution, containing the zinc, slowly for about fifteen minutes to expel hydrogen sulphide; cool and wash down. Add 13 c.c. of ammonia (d., .90), and if the solution is not then alka- line, make it so by cautious addition of ammonia. There must be enough hydrochloric acid present to neutralize at least the above amount of ammonia. Make the solution barely acid again with hydrochloric and add 3 c.c. of concentrated acid in excess. Add 1 c.c. of a solution of ferrous sulphate containing 3 grams of the salt per liter, dilute nearly to 200 c.c., heat almost to boiling, and titrate with potassium ferrocyanide. The ferro- cyanide employed for this titration contains 44 grams of the potassium ferrocyanide and .3 gram of potassium ferricyanide per liter and is prepared at least six weeks before use. The end-point is a sharp change in the color of the solution from a turquoise blue to a "pea green," and with several more drops to a " creamy yellow." This end-point occurs a little sooner than the one with uranium nitrate, and is easier to use, as the change is seen directly in the solution. If the sample does not contain at least 3 per cent of zinc, which can be estimated by the appearance of the zinc sulphide, 20 c.c. of a standard solution, containing 6 grams of pure zinc in 2000 c.c., are added from a pipette just after the solution to be titrated has been boiled to expel the hydrogen sulphide, and the solution is then made ready for titration as described above. This addition of zinc is necessary in order to obtain a good end-point. The ferrocyanide solution is standardized in the following manner : Six grams of zinc are dissolved in 40 c.c. of hydro- chloric acid in a 2-liter flask, the solution made up to the mark with the usual precautions, and 100 c.c. taken for titration. Add 10 c.c. of hydrochloric acid (d., 1.2) and 13 c.c. of ammonia (d., .90) If the solution is not then alkaline, it is made so with further addition of ammonia. Acidify again with hydro- chloric acid, adding the acid drop by drop, and an excess of 3 c.c. of the strong acid. After addition of 1 c.c. of the ferrous THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 251 sulphate solution previously > described, the solution is diluted nearly to 200 c.c., is heated almost to boiling, and is treated as described above. NOTE, on Gravimetric Method. Any laboratory making an occasional analysis only, will find it less work to employ a modification of the Waring method described under "Zinc in ores" (17, Chapter VII), and in the "Analysis of Brass" (21, this chapter) . 25. Nickel in German Silver. Nickel is usually taken by difference after the iron has been determined. If the estimation of nickel is required, after filtering out the sulphide of zinc, evaporate the nickel solution to dryness, then cover the beaker, add 25 c.c. each of hydrochloric and nitric acids, and heat to destroy the formates. This oxidation with aqua-regia is repeated until all the formates are destroyed. The small amount of nickel sulphide, which is brought down by the zinc sulphide and which is filtered off from the hydrochloric acid solution of the zinc sulphide, must also be added. After destroying the formates, add 3 c.c. of sulphuric acid and evaporate to strong fumes. Dilute, neutralize with strong ammonia, and add 20 c.c. of strong ammonia in excess, making about 125 c.c. in all. Electro- lyze with a current of .3 ampere per solution. Test the solution for the end-point by fresh hydrogen sulphide water as in the case of copper. Rapid Electrolysis. The author has found the Frary solenoid to be well adapted to the deposition of nickel. The nickel ammonium sulphate is a better conductor of the current than a neutral solution, hence the liquid should contain the ammonium sulphate and be kept strongly alkaline. Current 4 amperes per solution. In very accurate work, the deposit should be dissolved in nitric acid, the cold dilute solution treated with 10 c.c. of hydrogen sulphide water, filtered, and the residue washed, ig- nited, and weighed. If no copper is found, the residue may be assumed to be platinum carried over from the anode by the strong current. 26. Iron in German Silver. Dissolve 2 grams of the alloy in 15 c.c. of nitric acid and proceed exactly as in the determina- tion of iron in brass. 27. Carbon by Combustion. Place 5 grams of drillings in a 252 ANALYSIS OF COPPER 500 c.c. beaker and add 300 c.c. of a saturated solution of copper- potassium chloride and 22.5 c.c. of hydrochloric acid (d., 1.19). Stir and allow to stand in a warm place overnight, or, if neces- sary, hasten the operation by a mechanical stirrer. Filter on a platinum crucible (made up with ignited asbestos), wash thor- oughly with dilute, hydrochloric acid and fintilly with hot water to remove all chloride. Place the Gooch crucible inside a Shinier crucible, ignite in a combustion train, and determine the carbon in exactly the same way as " carbon in steel." It is absolutely essential that the beakers and solutions shall be protected from dust by double coverings until the carbon has been transferred to the combustion apparatus. REFINED NICKEL 28. Electrolytic Assay. Weigh duplicate 2-gram samples into 250 c.c. beakers, cover, and dissolve in 30 c.c. of (1 : 1) nitric acid. Dilute, add about 5 grams of ammonium chloride, and make alkaline with ammonia. Heat carefully to boiling and filter off the ferric hydroxide on a small "black ribbon" S. & S. paper, receiving the filtrate in a tall 500' c.c. beaker. Wash with dilute ammonia and with hot water. Dissolve the precipitate into the original 250 c.c. beaker with a little hot (1:1) hydrochloric acid. Reprecipitate with ammonia, filter, and combine this filtrate and washings with those from the first precipitation. Stand this solution on a hot plate to concentrate and evaporate off the excess of ammonia. Cover and add 50 c.c. of a mixture of equal parts of nitric and hydrochloric acids, and boil to small volume. Cool, add 20 c.c. of (1 : 1) sulphuric acid, and repeat the treatment with nitric and hydrochloric acids, until the ammonium salts are entirely expelled. Wash down, remove cover glass, and allow the solution to evaporate to strong fumes. Dissolve the nickel sulphate in water, neutralize with ammonia (d., .90) and add 20 c.c. in excess. Electrolyze overnight with a current of .25 ampere per solution. Test with hydrogen sulphide as in preceding methods. The cathode, after burning off the alcohol, is baked in the oven for one hour at about 105 C. The deposit is nickel plus cobalt plus copper. No attempt is ordi- narily made to determine the nickel and cobalt separately. Copper is estimated on a separate sample, and the amount found is subtracted from the total amount of the copper + nickel + cobalt. THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 253 Iron, if present, causes slightly high results, probably being brought over mechanically 3$ hydroxide and weighed as oxide. Electrolyze with a gauze cathode and a current of 5 amperes per solution for 1J hours if preferred, using some rotary device. This deposit may be purified as outlined in the last paragraph of 25. 29. Special Assay for Copper. Dissolve a 5-gram sample in a casserole, using 50 c.c. of (1 : 1) nitric acid; add 20 c.c. of (1:1) sulphuric acid and evaporate to strong fumes. Add about 20 c.c. of water, heat until nickel sulphate is dissolved, add about 10 grams of ammonium chloride, and saturate with hydrogen sulphide gas. Filter off the copper sulphide (which may contain a little nickel), place the filter and contents in a tall 200 c.c. beaker, add 10 c.c. of (1 : 1) sulphuric and 25 c.c. of strong nitric acid, then boil down to strong fumes. If the filter is not de- stroyed leaving a clear solution, more nitric acid may be added, and the solution boiled down to fumes again. Dilute to 120 c.c., add 3 c.c. of (1 : 1) nitric acid, and electrolyze as usual for copper. 30. Iron in Refined Nickel. Dissolve a 2-gram sample in 30 c.c. of (1 : 1) nitric acid, dilute, add ammonium chloride and ammonia in excess, and proceed exactly as in "the estimation of iron in brass." STANDARD ANALYSIS OF ZINC SPELTER 31. Classification. The sub-committee on alloys of the American Chemical Society recommends a method of testing spelter, which has been elaborated from those originally pro- posed by Elliot and Storer l and Price, 2 and which closely agrees with the procedure of the American Society for Testing Materials. 3 The sampling of spelter is described in 36, Chapter II. Spelter ordinarily used for brass and similar alloys is usually considered in three grades: A, "High Grade"; B, "Inter- mediate 1 '; C, "Brass Special," according to the amount of lead and other impurities present. A fourth grade, D, "Prime Western," contains more impurities than the three grades preced- 1 Mem. Acad. Arts and Sciences, 8, Pt. 1, May 20, 1850. 2 Chem. Eng. 9, 4. 3 Year Book, 1914, pags 384; also personal communications. 254 ANALYSIS OF COPPER ing. In 1916, a new "Selected Grade," with .80% lead, has been proposed. HIGH LIMITS. Grade A. High B. Interm. C. Special D. Western Lead Per cent .07 Per cent 20 Per cent 75 Per cent 1 50 Iron. 03 03 04 08 C admiu.ni .05 .50 75 ? The sum of lead, iron and cadmium shall not exceed Aluminum .10% .00 .50% .00 1.20% .00 ? ? ANALYSIS 32. Lead, by "Lead Acid" Method. Weigh into a 350 c.c. beaker 25 grams, 15 grams, 10 grams, or 5 grams of sa wings or drillings, according to whether the spelter is of grade A, B, C, or D, respectively, and add, according to weight of sample, 300 c.c., 180 c.c.. 120 c.c., or 60 c.c. of "lead acid." For preparation of lead acid, see note. When all but one gram of zinc is dissolved, filter on a close paper and wash out the beaker twice with "lead acid" from a wash bottle. Wash the met allies back from the filter to the original beaker with water and dissolve in a little hot (1 : 1) nitric acid. Add 40 c.c. of lead acid, and evaporate to strong fumes of sulphuric acid. When cool, add 35 c.c. of water (which is the quantity of water evaporated from the lead acid) and heat to boiling. Add the first filtrate (containing most of the zinc and possibly a little lead), stir well, and settle at least five hours, filter, wash with lead acid, then with a solu- tion of equal parts of alcohol and water and finally with alcohol alone. Ignite, and weigh the lead sulphate as usual. NOTE. "Lead acid" is a solution of one volume of sulphuric acid (d., 1.84) in seven volumes of water, saturated with lead sulphate. It is prepared as directed in the first part of this chapter (3), under the method for "lead in brass." 33. Lead, by Electrolysis. Weigh out 8.643 grams of the sample into a 400 c.c. beaker and add sufficient water to cover the sample. Then 30 c.c. of concentrated nitric acid (d., 1.42) are added gradually and cautiously until solution is complete, THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 255 and the solution is boiled for a few minutes to dispel the nitrous fumes. Wash off the cover* glass and sides of the beaker and transfer the liquid to^a 200 0*3. electrolytic beaker. Dilute to 125 c.c. and electrolyze with a current of five (5) amperes. The time required for the electrolysis is from J to f of an hour, depending on the amount of lead present in the sample. Solu- tions are tested for complete precipitation of lead by washing down the cover glasses and sides of the beaker, so that the depth of the solution is increased about \ inch (1.2mm.). The current is then continued for fifteen minutes and if the newly exposed surface is still bright, the lead is completely deposited. The anode is then washed three or four times with distilled water, once with alcohol; then dried in an oven at 210 C. for J hour and weighed. The weight of the lead dioxide, Pb0 2 , in milligrams, divided by 100 will give the percentage of lead. The dioxide deposit can be readily removed by covering the anode with dilute nitric acid and inserting a rod of copper. The electrodes recommended by the Committee are cylinders of platinum gauze, 48 meshes to the linear inch (or 18 per cm.). The anodes are 1J inches (316 mm.) in diameter and of equal height, with a stem 4J inches (108 mm.) long of 16-gauge wire, making the total height 5J inches (or 14 cm.). The cathodes are \ inch (12.7 mm.) in diameter, but having the same height and length of wire stem as the anodes. 34. Iron. Weigh 25 grams of zinc into a tall 700 c.c. beaker and dissolve cautiously in 125 c.c. of nitric acid. Boil, dilute, add considerable ammonium chloride, and then ammonia until the white zinc salts have dissolved. Boil, let settle, and filter on a 11 cm. S. & S. "black ribbon" filter paper. Wash with dilute ammonia and with hot water. Dissolve the precipi- tated ferric hydroxide with hot (1 : 4) sulphuric acid, add 40 c.c. of (1:1) sulphuric acid, pass through a Jones reductor, wash first with 150 c.c. of water, then with 100 c.c. of water, and titrate with permanganate. The potassium permanganate solution contains .2 gram of the crystals per liter. One c.c. of the permanganate will equal .000334 grams of iron. The solution is standardized against sodium oxalate, whose correctness has been guaranteed by the maker. Weigh out duplicate samples of sodium oxalate, weigh- 256 ANALYSIS OF COPPER ing .0200 gram. This will take between 49 and 50 c.c. of the permanganate solution. To convert sodium oxalate to iron, use the factor .8334. Run through a blank using the same amounts of water and acid and deduct from the tit ration. NOTE. If before passing the solution through the reductor, a large amount of lead sulphate is present, it is well to filter it off to prevent it from clogging the reductor. i 35. Cadmium. Weigh 25 grams of drillings into a tall beaker; add 250 c.c. of water and 55 c.c. of concentrated hydro- chloric acid and stir. When the action has almost ceased, add more acid with stirring, using about 2 c.c. at a time, and allowing to stand after each addition of acid, until finally all but about 2 grams of the zinc have been dissolved. About 60 c.c. of acid in all are usually required; it is best to allow the first 55 c.c. of acid to act overnight. Filter, first transferring one of the undissolved pieces of zinc to the filter, and wash a couple of times with water. Discard the filtrate. Wash the metallics on the filter paper back into the 500 c.c. beaker, cover, and dissolve in nitric acid. Transfer to a casserole, add 20 c.c. of (1 : 1) sulphuric acid and evaporate to fumes, take up with about 100 c.c. of water, boil, cool, and let settle for several hours (best over- night). Filter off the lead sulphate on paper, wash with water, retain the filtrate, and discard the lead sulphate. Dilute the filtrate to 400 c.c., add about 10 grams of ammonium chloride, and pass hydrogen sulphide into the liquid for one hour. It is occasionally necessary to start the precipitation of the cadmium sulphide by the addition of one or two drops of ammonia to the dilute solution. Allow to stand until settled, filter off the impure cadmium sulphide in a loose-bottomed Gooch crucible; remove the cadmium sulphide by carefully punching out the bottom into a 200 c.c. tall beaker. Wipe off any adhering cadmium sulphide, using a little asbestos pulp, add 60 c.c. of (1 : 5) sulphuric acid, and boil for one-half hour. The dilute acid dissolves cadmium and zinc sulphides, but not copper or lead sulphides. Filter and dilute to 300 c.c., add 2 grams of am- monium chloride, and precipitate with hydrogen sulphide in order to get rid of traces of zinc. Let settle, filter, and transfer to a platinum dish, cover, and dissolve in (1 : 3) hydrochloric acid. THE PRINCIPAL COMMERCIAL ALLOYS OF COPPER 257 Then the sulphide remaining; on the filter paper is also dissolved by hot (1:1) hydrochloric aci