LEARNING HOW TO REFINE AND CAST COPPERlibrary.aimehq.org/library/books/Choices of Methods in Mining and... · LEARNING HOW TO REFINE AND CAST COPPER A PAGE ... (NaCl or HCl) has been

LEARNING HOW TO REFINE AND CAST COPPER

A PAGE FROM THE HISTORY OF DEVEXBPMENT OF ELECTROLYTIC REFINING AND MECHANICAL CASTING

In June, 1893, while I was with the Old Dominion Copper Co. at Globe, Ariz., I received an offer to take charge, as manager, of the Baltimore Electrict Refining Co. which had recently built a plant at Baltimore, Md., having a capacity of about 1200 tons of electrolytic copper per month, using the Hayden series system. This was a very large output, for in those days the plants were small and the art was in its infancy.

At the beginning of 1893 there were 11 electrolytic refineries in the United States, most of them using the multiple system. The Hayden series system was used a t Baltimore and Bridgeport and the Smith series system was in operation a t the Electrolytic Copper Co. at Ansonia, Corn., and at the Pennsylvania Salt Co. at Philadelphia. In the latter system the plates were placed in a horizontal position, but the operating difliculties were so numerous it was soon discarded.

The total production of these 11 plants in 1892 was very small; I have not been able to discover reliable figures of production. There were about thirty electrolytic copper refineries in the world that year, according to Titus Ulke, and the entire production of these

Consulting Engineer, New York; Past Director A. I. M. E. t This should have been L'Electrolytio."

56

LEARNING TO REFINE AND CAST COPPER 57

refineries was 64,000,000 lb. annually, as much as could be produced in the United States at the present time, 1931, in 6>4 days. This shows conclusively what enormous strides have been made in the industry.

About this time the Nichols Chemical Co. at Laurel Hill, Long Island, was installing a plant to use the series system with cast anode plates instead of rolled plates, as in the Hayden system. These anodes are about 96 in. thick, 11 in. wide and 52 in. long. Five anodes are hung in a row and there are 130 rows in series in a tank. The operations a t this plant have been conducted most successfully under the able supervision of J. B. F. Herreshoff and his successors. It still competes with the multiple system in use a t the other large plants.

In the early days, from 1892 to 1898, when many of the large electrolytic refineries were erected, there was much rivalry between the users of the multiple system and those operating the series system. The former was used in the pioneer commercial plant installed by Edward Balbach and F. A. Thum in 1882, and the Balbach electrolytic copper refinery was the forerunner of this great industry in this country. Practically all the plants erected since 1893 have installed the multiple system. Still, the series system had its advantages.

In the multiple system the anodes were cast 2 by 3 ft., 1 to 1% in. thick. There were from 20 to 22 anodes in each tank* and the electric current passed through these anodes in multiple, through the electrolyte to the cathodes, out to the heavy copper bus bars and then on to the next tank. This system has often been described.

In later years the size of the anodes has been increased to 3 by 3 ft., in some plants to 4 by 4 ft., and 30 to 40 are sometimes placed in one tank.

58 CHOICE OF METHODS IN MINING AND METALLURGY

In the series system each tank held from 100 to 135 anodes, the current entered at the head end to the first anode and then through the electrolyte to each anode in series, depositing copper on the front of each of the anode plates, then to the cathode at .the back end of the tank, which was in contact with the conducting bar leading to the next tank.

The Hayden series system, which was installed in the new Baltimore Electric Refining plant, required careful preparation of the rolled anodes. The impure and argentiferous copper was cast into cakes 10% in. wide, from 2% to 3 in. thick and about 26 in. long. (The length has been increased considerably during the intervening years.) These cakes were heated in a furnace and then rolled into long sheets in grooved or closed rolls, so that the width would be uniform. The sheets were sheared into plates Ji in. thick, 11 in. wide and 24% in. long and assembled by putting two plates together in wooden-grooved sticks, % by % in., which held the two plates firmly together in a vertical position. The assembled anode plates were called "stands." The electrolytic tanks were made of slate slabs about two inches thick, grooved where the joints occurred and securely cemented with a paste of putty-like consistency, made of litharge and glycerin. The tanks were divided into six compartments, each approximately 26 in. wide, 27 in. deep and 10 ft. 6 in. long. Each compartment held 130 stands, the total weight of the copper being approximately 5400 pounds.

It can be readily understood from this brief description* that the multiple system was much simpler to

* For a detailed description see Keller: Mineral Zndzcetry, vol. VII, pp. 231-236, 1898.


operate than the Hayden system, but the latter had its advantages, as well as its disadvantages, and this subject will be discussed later. From 1892 to 1898 the Hayden system was conducted in a most successful manner at the Baltimore Electric Refining Company's plant and competed favorably with the multiple system used at other plants, both in cost and in the quality of copper produced. In later years the Hayden system at the Baltimore plant has gradually been replaced by the multiple system.

In the early nineties the electrolytic copper refiners found it practically impossible to produce copper of a standard quality regularly. We had a lot to learn. Sometimes the cathodes were tough, crystalline and pure; a t others the product was distinctly inferior. I have seen cathodes in the tank room of the Baltimore Electric Refining Co. covered with wide streaks of a brittle black deposit of copper containing impurities. It used to be said that when the cathodes were removed from the tank we might expect to take them out in sheets or with a shovel, which, of course, was an exaggeration. Still, at best the product was far from uniform, and on this account electrolytic copper did not command so high a price as the product from the Lake Superior mines, a standard brand of copper of excellent quality. The difference in price then was from $4 to cent per pound, in favor of the latter. Electrolytic copper producers had a serious problem to solve.

In 1893 Wesley Blair, foreman of our electrolytic department, made a discovery for which he deserved a great deal of credit and which, as far as I know, has never been recorded. On a certain occasion he noticed that the copper deposited on the cathode in one of the


sections in the tank room suddenly changed from a dull dark-red color to a bright crystalline deposit, and being an observing and thinking individual he noted this change and watched this section carefully. In a few days the copper turned dark again but in about a week he saw the bright crystalline deposit return. He started a careful investigation and, to make a long story short, found that one of the laborers, on the two occa- sions when the deposit changed, had run salt water into that section instead of fresh water. Blair reasoned that there must be something in the salt water that affected and improved the deposit. Naturally salt was his first thought, so he set up an experimental tank and added some of it to the electrolyte. Too much was put in a t first and the deposit became rough and covered with needles which developed into nodules. A second test with less salt added gave splendid results, the copper deposited was tough and pure and even at a current density of 25 amp. per sq. ft.-an extremely high current density in those days (usually 15 amp.)--excellent results were obtained. He then used hydrochloric acid, found i t equally efficacious and decided that chlorine was the element which produced the desired result. At that early stage we did not comprehend at once exactly what caused this change in the deposit but soon found that it was due to the formation of an oxychloride of antimony, which was insoluble in the electrolyte and dropped to the bottom of the tank with the anode slime, carrying arsenic down with it. Anode slime was the common name given to substances in the crude copper that were not dissolved by- the electrochemical action in the electrolytic refining process but formed a coating on the anode and finally fell to the bottom of the tank. It was composed largely

LEARNING TO REFINE AND CAST COPPER 6 1

of the impurities arsenic, antimony, bismuth, tellurium, selenium, etc., undissolved copper and the gold and silver in the anode copper. I n some cases these precious metals made up 50 per cent of the weight of the slime.,

This discovery by Blair was of paramount importance, and though other agents have been used later, chlorine in some form (NaCl or HCl) has been and is now added to the electrolyte a t all copper refineries. The average content varies from 0.002 to 0.006 per cent, depending mainly upon the amount of antimony in the anode.

The following notes taken from my record book, are interesting:

Nov. 7, 1893. Decided to put section No. 2 [one of the 12 sections in the tank room] on the "Blair experiment." Hung a lead test plate (about 4 in. wide and 8 in. long) from the top of a cathode to get a sample of the copper deposit before trying the experiment.

Nov. 8. Removed sample plate for analysis at 10:30 a. m. Then placed 1 lb. of salt in the gutter which conveyed the electrolyte back to the well or sump [each section had its own pump and circulating system and the weight of the electrolyte in each system waa about 116,000 lb.] and put in a second test strip, on which a reddish-colored deposit of copper soon appeared. About 11:30 a. m. the deposit on the top of the lead strip became lighter. At 1 :00 p. m. the color of the upper half of the deposit on the test strip was much brighter and by 3 :30 p. m. the entire deposit was bright and crystalline. Took it out and found copper tough. This complete change had taken place in 5 hr. Put in third test plate to obtain sample of new deposit.

Nov. 9. Third test plate taken out. Deposit tough, fine grained and of bright color.

The first and third sample deposits of copper were analyzed by Robert L. Whitehead, the well-known


chemist who was then in charge of the silver and gold parting plant, who found the following results:

Copper from test plates Arsenic Antimony

Deposit from test plate No. 1, per cent 0.050 p o s i t from t a t e NO. 3 . . . . . . . . . ;i~fl Trace

Salt was then added to this section at the rate of 1 lb. a week for some time and the cathodes produced were excellent. Its use was introduced gradually and in a few months all the sections in the tank room were on this system.

Before Blair's scheme for adding salt to the electrolyte was adopted the impurities in the latter were kept very low by withdrawing a number of cubic feet from the circulating system daily and sending the solution with- drawn to tanks where the copper was deposited on scrap iron and the acid liquor run to waste. Or if there hap- pened to be a market for sulphate of copper crystals, the electrolyte removed from the tank-room department was transferred to a blue vitriol plant where by a process consisting of evaporating and crystallizing, sulphate of copper was produced. By this means a large percentage of the acid in the removed electrolyte was usefully recovered instead of being run to waste.

It has always been necessary, of course, to remove some electrolyte from the circulating system, but after we adopted the practice of adding salt the number of cubic feet taken out daily was very much less, resulting in a large saving in the amount of sulphuric acid added to the electrolyte-about 50 per cent.

The use of salt did not entirely solve the problems of producing a uniformly high quality of electrolytic copper


but i t was an important link in the chain of innovation which finally enabled refiners to accomplish the desired result. In my opinion, i t was the most important contribution to this art. To the many others whose ideas, researches and suggestions were of great value, much credit is due.

It required years to convince the purchasers of copper that electrolytic copper of a high standard could always be regularly and uniformly produced, but the consumers finally were willing to concede that this class of copper was fully equal to the product coming from the Lake Superior district and, in fact, superior for electrical purposes on account of its higher electrical conductivity. From 1914 electrolytic copper has been used a4 the basis for official price quotations, Lake copper being sold at the same price to large consumers.

We soon discovered that in operating the Hayden system with a drop in potential of 22 to 23 volts per tank there was an enormous loss of effective current between the head and the tail ends on account of the current leakage through the solution, under and over the plates (the electrolyte alone having much less electrical resistance than the copper plates and electrolyte, owing, of course, to transfer resistance). Attempts were made to reduce this leakage to a minimum, but we soon reached the limit, as i t was necessary to have a t least 2% to 3 in. space under the plates, to provide for the accumulation of slime and prevent short-circuiting. The top of the plates must also be covered with solution to insure the complete dissolution of the anode. About the best ampere efficiency we could obtain was 70 per cent and sometimek it was even lower. The resistance between electrodes was very low, requiring only 0.16 to 0.17 volt


as against 0.30 to 0.33 volt in the multiple system. (In some plants the voltage now required is less.) The amount of copper produced per unit of power in the latter was much less than in the Hayden system.

We also conducted a careful research in the resistance of electrolytes of different composition and temperature and ascertained that we could reduce the resistance of these solutions greatly by reducing the copper content, increasing the acidity (HzS04) and the temperature of the electrolyte. These researches resulted in a large increase in the amount of cathodes produced per unit of power. In 1894 the production of the tank room was 154,585 lb. per working day (the power department was shut down and no copper deposited on Sunday) while in December, 1898, it was 230,746 lb. per day. This increase was due to changes and improvements made, based on the result of research work, as during this period there had been no increase in tank capacity.

We also investigated the influence of impurities, principally arsenic, antimony and cuprous oxide, on the electrical conductivity of copper, and it is remarkable how closely our results, which we thought were somewhat

. crude, agreed with determinations made much later and in a most efficient manner by others, notably Lawrence Addicks. *

In 1895 we made a study of the effect of cuprous oxide on the electrical conductivity of copper and after com- piling the results from 33 samples came to the following conclusions: Copper containing over 2.15 per cent CuzO was too brittle to be satisfactorily drawn into wire, and with over 1.20 per cent the conductivity was liable to be below standard. With electrolytic copper of the

Trans. A. I. M. E., vol. 36, p. 18.


standard grade the best results were obtained when the CuzO content was between 0.6 and 0.7 per cent, the oxygen being required to oxidize the impurities present. We also found that in two cases where the CuzO was very low, 0.18 and 0.24 per cent, the conductivity was high, 102 per cent, based on the old Mathiessen standard; this copper was almost pure. Subsequent research by others confirm these latter determinations.

HAND LADLING OF COPPER REPLACED BY MECHANICAL CASTING

The problem of improving the method used for taking molten copper from reverberatory furnaces and casting it into shapes received constant attention by metal- lurgists for many years. This was especially true during the latter part of the last century, when the increased scale of copper-smelting operations demanded a similar development in the refining and casting process.

From the establishment of the systematic s,melting of copper in Wales, about the middle of the sixteenth century, until 1895, copper shapes were cast by a very laborious hand method, the workman dipping the copper from the furnace in small ladles which were carried to the moulds into which it was poured.

As late as 1870, a normal charge was 10,000 Ib. and 15,000 lb. was a large one. About 1890 charges as large as 25,000 lb. were being melted, refined and cast, and even as late as 1895 the average furnace charge of copper at our plant was only 31,440 lb. It was then evident that we were nearing the maximum that could be attained by hand ladling in a Whr. cycle.

It is interesting to note that Prof. Edward D. Peters, in Mineral Industry for the year 1893, calls attention

66 CHOICE'OF METHODS I N MINING A N D METALLURGY

to the necessity for tapping fine copper into molds and mentions an invention of Egleston for doing this, which did not answer the purpose. He emphasized the need for doing away with the laborious work of hand ladling but appreciates the "great difficulties to be encountered before a method can be invented that will supersede the present system of ladling."

The first step taken to solve the problem was a device using ladles of 200 lb. capacity, running on a track or suspended from a trolley, dipping the copper direct from the furnace and pouring it into molds. This scheme was not a success because the copper splashed when poured into the molds, forming "cold sets"; it solidified on the ladle, forming "skulls," and owing to the increased exposure to the air absorbed oxygen and contained more cuprous oxide than when poured from small ladles. To avoid dipping copper out of the furnace by means of ladles, tapping was resorted to, and Edward Keller says:*

The first copper-refining furnaces with tapping arrangement were built by A. F. Schneider at the works of the Guggenheim Smelting Co. at Perth Amboy, N. J., but for reasons unknown to the writer these furnaces were never put into operation. Later Wm. H. Peirce introduced tapping (at the old plant of the Baltimore Copper Smelting & Rolling Co.) and the use of large ladles to convey the copper at his large furnaces, in which he refines converter copper with perfect success.

The tapping arrangement perfected by Peirce wm simple and ingenious. After a charge of copper had been removed, and while the furnace was being recharged, a clay dam was rammed into a vertical slot, about four inches wide, in the side of the furnace. This dam was reenforced on the outside by removable iron bars and

Mineral Industry, vol. VII, 1898.


the heat of the furnace soon burned the clay dam to a solid brick. After the charge was melted, refined and ready to cast, the iron bars were removed, the clay dam was gradually cut away with a sharp tool, and the charge run out of the furnace. The semi-mechanical device* designed by Peirce for casting anode cakes from impure converter copper was arranged so that the copper flowed from the furnace through the tap hole into a 750-lb. ladle suspended from a small turning crane. The vertical . motion of the ladle and the horizontal one along the line of the beam of the crane were effected by compressed air and hydraulic power, while the circular, horizontal swinging of the crane and ladle, and the tilting of the latter for pouring the copper into the molds, were performed manually. The molds were arranged in a straight row along the top of the water bosh.

Mr. Keller then says:?

Although these successful innovations by Mr. Peirce were somewhat of a revolution in the industry of copper (pyro-) refining, especially in the handling of the cruder grades of metals; the casting of electrolytic copper into wire bars, cakes, and ingots by the herein described mechanical appli- ances still remained unsuccessful, because splashing and cold sets, which makes these castings unacceptable, seemed unavoidable except by hand ladling.

After a long study and consideration of the difficulties inherent in casting fine copper shapes, I devised an apparatus which overcame all the difficulties that hitherto had proved insurmountable.

The schedule for ladling an average charge of 31,000 lb. of fine copper from the furnace and casting it into

* Keller: Mineral Zndusty, vol. VII, opposite p. 263. t Op. cil., p. 252.


ingots, cakes and wire bars called for the use of four ladlers up to 155-lb. bars; from that weight up to 250-lb. bars, five ladlers; and for 250-2b. bars and over, six ladlers. The number of ladlers to a furnace depended upon the size of the bars and not on the weight of the copper to be cast, the old Cornish tradition being that the larger the bar the more ladles should be poured at once to "float" the copper and prevent the occurrence of cold sets.

In February, 1895, I decided to try an experiment and told the foreman of the casting department to have a charge taken from the furnace by four ladlers and cast into bars of large size. The ladlers said it could not be done; we asked them to try it, but they replied that it would be useless to attempt any such radical change. When we insisted, the ladlers on all the furnaces struck; they were not dissatisfied with anything but their reputation was at stake. During the next two weeks, fortunately, we were able to ship cathodes on foreign orders; the strikers were friendly and there was no unpleasantness. At the end of that time a committee of the ladlers came to me and said that after careful consideration they had decided to try to cast large bars with four ladlers. They succeeded at once in doing what they thought was impossible; the fact that they put forth their best efforts to accomplish this result was greatly to their credit.

In thinking over the question of casting copper, it seemed to me unwise to be dependent on a particular and independent class of skilled labor for the final stage in production and that in spite of the many unsuccessful attempts made in the past some mechanical method of casting fine copper should be devised.


We first tried large dipping ladles suspended from a trolley but had no more success with them than others had had in the past. Careful experiments showed that in order to cast fine copper successfully it must be poured into the mold with a minimum drop or fall and be exposed to the air as little as possible.

Working along this line for more than a year, I designed a machine which I felt confident would cast fine copper in the usual shapes successfully and economically. When I placed my plans before William Keyser, * he was doubtful and did not wish to appropriate the money necewary to build the machine and accessories. I offered to furnish the funds myself, under certain conditions, and then he told me to go ahead. The machine was to serve a new reverberatory furnace we had already decided to build. This was in February, 1897, according to my notes, in which appear the following items: "Commenced work Mar. 15. Machinery ordered Apr. 30. Machine installed May 31. Furnace ready and started drying brickwork June 22. Took out first charge July 8. Minor troubles till July 21 when machine was pronounced a complete success."

In December, 1897, after the machine had been thoroughly tried and tested, it was decided to apply this system of mechanical casting to all the furnaces in the Baltimore Electric Refining Co. plant. The attitude of the furnace ladlers during this period was admirable. While they knew that a successful casting machine would throw them out of their much specialized class of work, they realized that the time had come when something

* In the last half of 1896 the Baltimore Electric Refining Co. was can- solidated with the Baltimore Copper Smelting & Rolling Co., with Mr. Keyser as president, but the management of the two plants was separate and independent.


in this line must be accomplished and rendered all the assistance that lay in their power. I have always felt most grateful for their co-operation, as the process of adopting a new idea to exacting conditions is always difficult at best.

The machine is described in detail elsewhere* and only a brief note is necessary here. The copper is drawn from the furnace by using the tapping scheme installed by Peirce at the plant of the Baltimore Copper Smelting & Rolling Co. and flows into a specially designed receiving ladle placed underneath the furnace spout. The ladle is pivoted on trunnions, placed as near the pouring mouth as possible, which rest in curved supports fastened to a rocking shaft, so that by rotating this shaft the mouth of the ladle can be moved forward while the copper is being poured into the mold and then backward. The ladle is supported at the back by a hook engaging in a yoke fastened to a flexible cable, which in turn is con- nected with the piston of a hydraulic cylinder. By this means the ladle can be raised at the back for pouring and lowered when the mold is Ned. While pouring the copper into the molds, which pass directly under the mouth of the receiving ladle, the level of the mouth is not changed, the height of the fall of the stream of copper into the mold being reduced to a minimum; as a matter of fact, the mouth of the ladle is directly over the molds, as in the case of hand ladling.

The molds are supported on the outer end of arms which are suspended from the rim of a central wheel or turn-

* Edward D. Peters: Practice of Copper Smelting, pp. 552-566, 1911. Minmal Indust~y, vol. VII, pp. 252-257, 1898. Hofman and Hayward: Metallurgy of Copper, pp. 276-279. Also other works on the metallurgy of copper.


table. The latter revolves on a set of rollers running on a circular track and is rotated by an electric motor (at first a hydraulic cylinder was used but the movement was not even). I n the center of the wheel is a stationary platform around which the wheel revolves and upon which are placed the three levers that regulate the entire operation.

In casting by this machine, after the copper is poured into the mold the wheel is rotated, and when the mold has described an arc of 180 deg. the copper shape in the mold has solidified and is dumped into a large water bosh by automatically turning the mold upside down. The copper casting is picked up by a slow-moving conveyor, is cool by the time it reaches the surface of the water and is delivered to the inspection platform, from which it is loaded on to cars.

With these machines a casting speed of 75,000 lb. of copper per hour can be attained and in some cases this rate has been exceeded.

The size of the furnaces was rapidly increased, bearing in mind the desirability of completing a furnace cycle in an even 24 hr. and it was found possible to charge, melt, refine and cast 300,000 to 400,000 lb. of copper per day, 10 times as much as when the practice of hand ladling was in vogue. With the introduction and use of these machines the cost of casting has been reduced about one-half; the item of labor alone to less than one-third.

This result could not have been accomplished, however, without the use of the charging crane designed by Ladd and Prosser and used first, I believe, at the Chrome, N. J., plant of the U. S. Metals Refining Co. This crane replaced the laborious and time-consuming process of charging the furnace by hand.

72 CIIOICE OF METHODS IN MINING AND METALLURGY

A straight-line casting machine was used by J. C. McCoy at the Raritan Copper Works when that plant was built in 1899, and later the same type of apparatus was installed by the Nichols Copper Co. in its plant at Laurel Hill. The straight-line machine at the Raritan plant was abandoned about 1907 and replaced by the Clark machine, the chief reason for doing so being that in the former there was lost motion in jarring, and the cost of maintenance was high.* The Clark machine is practically the same as the Walker casting apparatus except that on the former the molds are carried parallel to the radial arms while on the latter they are placed circumferentially.

It is safe to say that in the past 30 years Walker machines have cast approximately one-half of the crude copper and refined copper produced in the world. They are used in practically every copper-producing country.

CHOICE OF ELECTROLYTIC SYSTEMS

In 1899 1. was called from Baltimore by the Guggen- heim Brothers to take charge of their plant at Maurer, near Perth Amboy, N. J. In 1901 it became evident that the electrolytic copper refinery at this plant, which had an output of about 33 tons per day, was inadequate to treat the crude copper we expected to receive and that an increase in capacity was necessary. The plant had been running six years, the tanks were smaller than those in the newer installations and the design, though up to date in 1895, was antiquated. An entirely new plant was therefore in order, and the question arose, Should we use the multiple or series system?

Lawrence Addicks: Met. and Chem. Eng., p. 585, Nov. 15, 1917.


Comparing one with the other, very briefly, it can be stated that the advantages in the series system are:*

1. For a given amount of power more copper can be deposited. While the leakage of current around the electrodes in this system is large, the voltage between plates is much less. The production in the Hayden system is about 140 per cent and in the Nichols cast anode system about 170 per cent of that in the multiple system per unit of power.

2. Less carry of metals in process; electrodes are much thinner.

3. Less scrap produced; about 6.7 per cent in the Hayden system, 10 to 15 per cent in the Nichols and 14 to 18 per cent in the multiple system.

4. Less tank room space required for a given output. Tanks can be placed closer to each other, and there are many more electrodes in each.

5. Much less copper is required for bus bars and conductors.

The disadvantages of the series system can be con- sidered as advantages of the multiple system. They are:

6. Ability to treat copper of any quality, no matter how impure or how rich in precious metal.

7. Less loss in precious metals in the cathodes produced-about 0.35 per cent of the gold and silver in the anodes of the multiple system; 1 per cent in the Hayden system and 2 per cent in the Nichols process, due to the very long cathodes.

8. Ability to handle electrodes and scrap in larger units and with less cost for labor.

For more complete discussion see A. L. Walker: Mineral Industry, vol. XVII, p. 327, 1908.


9. Requires less care in maintaining the purity of the electrolyte, as it is possible to effect a much better circulation.

10. Cost of casting and preparing anodes much less, especially when compared with the Hayden system, where the plates must be rolled.

The crude copper we expected to receive was very impure and rich in silver and gold; therefore, considering items 6, 7 and 9, i t was imperative to use the multiple system. In view of my experience in Baltimore, I thought it worth while to see whether some of the advantages of the series system could not also be attained. At that time the multiple-system plants used wood tanks lined with lead and each was separate from the one next in series. They were usually accessible to a working aisle on both sides, but in the Balbach and Guggenheim plants the tanks were arranged in pairs, with a space between them, so that each tank was accessible to a working aisle on one side only. When the tanks were entirely separate two heavy copper bus bars were required for each, but when placed in pairs only one was necessary.

In all the multiple plants, the tank room area per ton of cathodes produced per day was several times that necessary in the series plants, and the cost of the copper conductors and bus bars was a heavy item. In the Guggenheim plant the tanks were 9 f t . 10 in. long, 3 f t . deep and 2 ft . 6 in. wide. Each contained 22 anodes, 2 ft . 6 in. long, 2 f t . wide and about 1.25 in. thick, and 23 cathodes. There were 360 tanks arranged in 12 rows in pairs, forming so-called double tanks. These double tanks were set in line on terraces with a difference in level between each of about two inches.


All of the tanks were in series and the current, 3000 amp., was carried from the switchboard by a copper conducting bar, about 1.5 by 3 in., to a bus bar of the same size supported along the outer edge of the first tank.

In each pair of tanks the anode lugs in the f ist rested on the bus bar and the current was delivered to the anodes in multiple. I t then passed through the electrolyte to the cathodes, hung from the cathode crossbars, and through the latter to a strip of copper placed on the board which covered the adjacent sides of the pair of tanks. This strip of copper, 4/4 by 1 in., distributed the current to the lugs of the anodes of the second tank, which also rested on it. The current passed through the second tank in the same manner as through the first and out to the bus bar supported along its outer edge; then on to the next pair of tanks in series.

It occurred to me that if we could extend this idea of two tanks with a distributing bar between them, to a number of tanks (using improved methods of construc- tion), we should be able to combine the advantages, in part at any rate, credited to the series system in items 1, 4 and 5 with those of the multiple system, and a dis- tinct gain would be made in the practice of electrolytic copper refining.

After a careful study I decided that it was quite possible to arrange the tanks in such a manner that any number desired could be placed in series, in close prox- imity to one another, with a triangular distributing bar of small cross-section between tanks. An arrangement like this would reduce the power necessary for a 2ven production due to knifeedge contacts between the triangular distributing bar and the anode lugs and cathode bars of the adjacent tank. The tank room area


for a given output would be very much less and the amount of copper required for bus bars would be enor- mously reduced. We had already proved, in November, 1900, the advantage of line or knifeedge contacts in a test made on two blocks of tanks in which the conditions were exactly the same.

With the regular flat contacts, the drop in voltage from the cathode crossbar of one tank to the anode of the next involving two contacts was from 12.5 to 12.8 per cent of the total tank voltage. In the block equipped with edge contacts this loss was only from 6.7 to 7.7 per cent, a saving of about 44 per cent in the contact losses at this point. There were, of course, other contact losses, which were the same in each case.

The idea of placing the tanks in nests met with con- siderable criticism from many, the principal objections being that (1) there would be no working aisle from which the contacts, electrode spacing, etc., could be easily taken care of, except at the two end tanks in the nest, and men could not attend the tanks properly by working on top of them; (2) leaks in the lead lining of the tanks undoubtedly would occur, the wood partitions between two adjacent tanks would become damp and the electric-current leak from one to the next would cause a loss in current efficiency and rapid deterioration of the tanks.

Their objections were serious, but we were sure that they could be overcome. Before making plans for the new plant, however, we decided to test the scheme on a small scale. We built a pilot plant consisting of two nests of 15 tanks each. They were the same size, inside, as those in the old tank room. The partition between two adjacent tanks in the nest was 5 in. thick and, of


course, the lead lining of each tank was completely insulated from the lining in the next one.

The pilot plant was completed and put in operation early in 1901 and, after a few minor troubles had been corrected, no operating difficulties were experienced and the result exceeded our expectations. The men found it easy to work on top of the tanks, test the voltage, take care of the contacts, etc. We also felt sure that any danger from current leakage between the tanks could be made negligible by the use of a ventilated partition, which we designed.

After the pilot plant had operated successfully for a number of weeks, we decided to install a new electrolytic refinery of 100 tons' daily capacity, using this system. We also deviated in another manner from the practice of the old refinery. There the long axis of the anodes and cathodes was vertical and as a result the loss of silver and gold due to adhering slime was large. Carefully conducted tests proved conclusively that the lower 6 in. of the cathode plate held about as much silver and gold as the upper 24 in. This was an important matter, as the anodes were very rich, often containing as much as 300 to 400 oz. of silver and 3 to 4 oz. of gold per ton of 2000 lb. In one lot the silver ran as high as 1000 oz. per ton. We wished to use somewhat larger anodes than in the old plant, 3 by 2 ft., and decided to suspend them in the tank with the short axis vertical, in order to reduce the danger of slime falling on a deep cathode. This was fortunate in another respect, as a number of years later, when an increase in capacity was required, the end walls and partitions of the tanks were raised 12 in.-a simple matter-the size of the anodes increased to 3 by 3 ft. and the current (amperes) increased in proportion.


The production was increased 50 per cent by this means without the addition of more tanks or increase in tank room area.

The new tank house was erected in 1901 and 1902. Provision was made for 816 tanks about 11 ft. long, 3 ft. 6 in. wide and 2 ft. 655 in. deep. They were placed in 24 batteries containing two nests, each nest being composed of 17 tanks, arranged according to the "Walker systemM* with a ventilated partition 5 in. wide between each tank.

There were about 26 anodes and 27 cathodes in each tank. The bus bars at the end of each nest were designed to carry a current of 5000 amp. The current density at the anodes was about 15 amp. per sq. ft. On the partition between the tanks a cover board, M by 5 in., was placed and the cqpper triangular distributing bar (x in. on a side, in. in set,.ion) rested on the cover board. The cathode crossbars of one tank and the anode lugs of the next were in contact with the distributing bars in the manner already described. We had no difficulty at all in operating the plant from the time it was put in commission.

The De Lamar electrolytic copper rehery built in 1902 and 1903 at Chrome, N. J. (later acquired by the U. S. Metals Refining Co. and still later by the American Metal Co.), installed the Walker system at once, and the Raritan plant (Anaconda Copper Co.) used it in its new tank room in 1906 and also in the old tank room, built in 1899, in 1911. A large majority of the electro-

Hofman and Hayward: Metallurgy of Copper, pp. 376-377, 394-396. Titus Ulke: Electrolytic Copper Refining, pp. 88-89, 1903. Lawrence Addicks: Met. and Chem. Eng., p. 667, Nov. 16, 1916. A. L. Walker: Mineral Zndzlstry, vol. XIX, p. 218, 1910.


lytic refineries constructed since 1902 have been equipped with and are now using this system.

When we made a comparison between the new Guggen- heim* tank room and the old one we found that there was a decided increase in the amount of copper produced per unit of power, due to lower resistance in the contacts and also, naturally, to improved insulation and con- struction. The tank room area required for 1 ton of cathodes produced per day was reduced from more than 1000 to 650 sq. ft., and afterward when the tanks were made 12 in. deeper and the anodes 50 per cent larger the latter figure was automatically reduced to 433 sq. ft. In the U. S. Metals Refining Co. plant at Chrome, when especial attention is paid to the casting of smooth anodes they can be placed in the tanks very close together, and one ton of cathodes is produced daily per 330 sq. ft. of tank room area, a figure that compares favorably with the series system, when the space required for scrapping series cathodes is taken into account.

The amount of copper for bus bars in the new system was greatly reduced. Had we adopted the same arrangement for tanks in the new 100-ton electrolytic refinery as was used in the old one we would have required 8.5 times as much copper for bus ban-a large saving. This is especially true when a very heavy current is used. The weight of the cathode crossbars, leading in and leading out conducting bars, would of course have been the same.

In 1916 Robert L. Whitehead devised a scheme by which the distributing bars on top of the tanks in the Walker system can be dispensed with. He placed the

* All the plants of the Messrs. Guggenheim Brothers were consolidated with the American Smelting & Refining Co. about 1802.

80 CHOICE OF METHODS I N MINING AND METALLURGY

cathode crossbars of one tank directly on the anode lugs of the anodes in the adjoining tank, thus reducing by one the humber of contacts for each tank. ,A notch in the cathode bar fits over a projection on the anode lug and gives a good connection. This method has found favor in a number of electrolytic reheries.

After many years of experience I am convinced that success in attainment in this world rarely depends on the efforts of anyone alone; i t is usually,due to gradual development, assistance from others and, last but not least, as Charles M. Schwab once said, "a lot of hard work."

Documents

LEARNING HOW TO REFINE AND CAST COPPERlibrary.aimehq.org/library/books/Choices of Methods in Mining and... · LEARNING HOW TO REFINE AND CAST COPPER A PAGE ... (NaCl or HCl) has been