Search Documents

By A. P. Miller, T. S. Washburn

Improved procedure in the manufacture of rail steel has come as the rail user demanded better wearing qualities combined with greater unit weight. With each weight increase per lineal yard has come greater difficulty in making a product that would successfully withstand all tests that modern rails must meet. With these facts in mind, and with a sincere desire to produce a rail whose service in track would be satisfactory in all respects, the steelmaker was quick to realize that all of the factors entering into his process would have to be more closely controlled. Control of rail-steel slags was recognized as one of the most important of many interdepending factors. The object of slag control is to hold the slag compositions within a range that gives the best results with respect to economy of production, uniformity of analysis, and quality of the finished rails. The most important component of the slag from the standpoint of these production and quality factors is the ferrous oxide content. This component is a good indicator of the basicity and oxidizing power of the slag, and these largely govern the effect of the slag on the steel. Consequently, the

Jan 1, 1935

By A. P. Miller, T. S. Washburn

Improved procedure in the manufacture of rail steel has come as the rail user demanded better wearing qualities combined with greater unit weight. With each weight increase per lineal yard has come greater difficulty in making a product that would successfully withstand all tests that modern rails must meet. With these facts in mind, and with a sincere desire to produce a rail whose service in track would be satisfactory in all respects, the steelmaker was quick to realize that all of the factors entering into his process would have to be more closely controlled. Control of rail-steel slags was recognized as one of the most important of many interdepending factors. The object of slag control is to hold the slag compositions within a range that gives the best results with respect to economy of production, uniformity of analysis, and quality of the finished rails. The most important component of the slag from the standpoint of these production and quality factors is the ferrous oxide content. This component is a good indicator of the basicity and oxidizing power of the slag, and these largely govern the effect of the slag on the steel. Consequently, the

Jan 1, 1935

By Francis B. Foley

Fundamentally, slag control in acid open-hearth steelmaking cannot be considered from any different viewpoint than can that in basic open-hearth melting. The same laws of mass action and distribution apply in both cases, so that while most, if not all, of the recent intensive study of slag-metal reactions have been carried on in basic furnaces the principles evolved are nevertheless applicable to acid open-hearth steelmaking and it is therefore not our intention to discuss the various possible chemical reactions in the slag and between slag and metal. The points we have to carry in mind, in considering acid melting practice, are the differences in the type of steels produced and the difference in the chemistry of both the steel and the slag. The acid furnace is used to produce, let us not say "quality" steel, for, after all, a "quality" steel for one purpose is not a "quality" steel for another, but steels that require acid melting practice to be of proper "quality." Steels that are required to have high physical properties in tests taken transversely from forgings are peculiarly the product of acid furnaces. In this class may be placed various ordnance materials, notably gun forgings and projectiles and large commercial forgings requiring high transverse ductility. Forgings requiring an extremely hard surface, polished and free from imperfections, are also typical acid-steel products. Of course, such forgings can be made of basic open-hearth steel, but not with the same assurance of success. At the start, the preparation of the slag in the acid and basic furnaces differs. Lime is charged in the beginning of a basic heat to form a slag, but in the acid furnace nothing is added for the special purpose of producing a slag when the heat is melted down. The result is that the acid furnace has much less volume of slag than the basic. The slag formed upon melting down is made up almost entirely of the oxides of elements contained in the metallic charge, notably oxides of iron, manganese and silicon, the amount of silica being augmented somewhat by erosion of the furnace lining. A typical acid open-hearth slag, when the bath is completely melted, will have an analysis about as follows: CaO, 3.40 per cent; SiO2, 51.9; Fe2O3, 1.25 ; FeO, 28.86; Al2o3, 1.82; MnO, 12.22. The high iron oxide content of the slag gives it a black appearance. The metal is high in carbon and some manganese and silicon will be present, depending on the amount in the original charge, the amount of

Jan 1, 1935

By Francis B. Foley

Fundamentally, slag control in acid open-hearth steelmaking cannot be considered from any different viewpoint than can that in basic open-hearth melting. The same laws of mass action and distribution apply in both cases, so that while most, if not all, of the recent intensive study of slag-metal reactions have been carried on in basic furnaces the principles evolved are nevertheless applicable to acid open-hearth steelmaking and it is therefore not our intention to discuss the various possible chemical reactions in the slag and between slag and metal. The points we have to carry in mind, in considering acid melting practice, are the differences in the type of steels produced and the difference in the chemistry of both the steel and the slag. The acid furnace is used to produce, let us not say "quality" steel, for, after all, a "quality" steel for one purpose is not a "quality" steel for another, but steels that require acid melting practice to be of proper "quality." Steels that are required to have high physical properties in tests taken transversely from forgings are peculiarly the product of acid furnaces. In this class may be placed various ordnance materials, notably gun forgings and projectiles and large commercial forgings requiring high transverse ductility. Forgings requiring an extremely hard surface, polished and free from imperfections, are also typical acid-steel products. Of course, such forgings can be made of basic open-hearth steel, but not with the same assurance of success. At the start, the preparation of the slag in the acid and basic furnaces differs. Lime is charged in the beginning of a basic heat to form a slag, but in the acid furnace nothing is added for the special purpose of producing a slag when the heat is melted down. The result is that the acid furnace has much less volume of slag than the basic. The slag formed upon melting down is made up almost entirely of the oxides of elements contained in the metallic charge, notably oxides of iron, manganese and silicon, the amount of silica being augmented somewhat by erosion of the furnace lining. A typical acid open-hearth slag, when the bath is completely melted, will have an analysis about as follows: CaO, 3.40 per cent; SiO2, 51.9; Fe2O3, 1.25 ; FeO, 28.86; Al2o3, 1.82; MnO, 12.22. The high iron oxide content of the slag gives it a black appearance. The metal is high in carbon and some manganese and silicon will be present, depending on the amount in the original charge, the amount of

Jan 1, 1935

By L. F. Reinartz

The furnace in which rimming steel is made has an important bearing on the quality of the steel produced. Furnace.—Particularly in the manufacture of low-carbon rimming steels, it is necessary to have an efficient, easily regulated furnace, because of the relatively high finishing temperature requirements. In recent years, the most important development in design of open-hearth furnaces has been the sealing and insulation of the greater part of the furnace to prevent air infiltration and loss of heat. Efficient, straight-line valves insure proper draft. Combustion controls guarantee correct gas-air ratios. Furnaces of all sizes ranging from 20 to 200 tons have been used successfully in the manufacture of rimming steels. Metal bath is varied from 20 to 40 in. in depth without appreciable variation in quality. Charging.—A normal open-hearth charge consists of pig iron (hot or cold), scrap of miscellaneous nature, and slag-making additions, such as limestone and burnt lime. Because of local conditions, some plants find it necessary to discard the use of pig iron and use some carbonaceolis substances with varying percentages of cast iron to replace the pig iron. It is questionable whether such practice will consistently produce the quality of rimming steel required for auto stampings and other severe drawing requirements. Pig Iron.—An open-hearth furnace under proper control should melt 150 tons of low-carbon rimming steel on about 35 to 38 per cent pig-iron charge basis. In some practices, best results are obtained when the total charge contains from 0.40 to' 0.50 per cent silicon. If reasonably low-

Jan 1, 1935

By L. F. Reinartz

The furnace in which rimming steel is made has an important bearing on the quality of the steel produced. Furnace.—Particularly in the manufacture of low-carbon rimming steels, it is necessary to have an efficient, easily regulated furnace, because of the relatively high finishing temperature requirements. In recent years, the most important development in design of open-hearth furnaces has been the sealing and insulation of the greater part of the furnace to prevent air infiltration and loss of heat. Efficient, straight-line valves insure proper draft. Combustion controls guarantee correct gas-air ratios. Furnaces of all sizes ranging from 20 to 200 tons have been used successfully in the manufacture of rimming steels. Metal bath is varied from 20 to 40 in. in depth without appreciable variation in quality. Charging.—A normal open-hearth charge consists of pig iron (hot or cold), scrap of miscellaneous nature, and slag-making additions, such as limestone and burnt lime. Because of local conditions, some plants find it necessary to discard the use of pig iron and use some carbonaceolis substances with varying percentages of cast iron to replace the pig iron. It is questionable whether such practice will consistently produce the quality of rimming steel required for auto stampings and other severe drawing requirements. Pig Iron.—An open-hearth furnace under proper control should melt 150 tons of low-carbon rimming steel on about 35 to 38 per cent pig-iron charge basis. In some practices, best results are obtained when the total charge contains from 0.40 to' 0.50 per cent silicon. If reasonably low-

Jan 1, 1935

To control the slag in an iron blast furnace is to control the quality of the pig iron produced (and to a certain extent the tonnage), and to control the whole operation of the furnace itself, including a large part of the cost of production. The character of the slag is the determining factor in the character of the pig iron as to the controllable limits of sulfur, silicon, and, to a certain extent, the manganese; it has an influence on the proportions of combined and graphitic carbons.

Jan 1, 1935

To control the slag in an iron blast furnace is to control the quality of the pig iron produced (and to a certain extent the tonnage), and to control the whole operation of the furnace itself, including a large part of the cost of production. The character of the slag is the determining factor in the character of the pig iron as to the controllable limits of sulfur, silicon, and, to a certain extent, the manganese; it has an influence on the proportions of combined and graphitic carbons.

Jan 1, 1935

By Karl L. Fetters

In the process of making steel in the open-hearth furnace the refining of the metal during the working period of the heat is largely accomplished through the agency of the slag. From the birth of the process until the present day steelmakers have recognized the importance of the slag in determining the composition of the metal and the quality of the product. The general relationships between slag composition and metal analysis have long been known, at least in a qualitative or approximate fashion. For examples, the iron oxide content of the slag is higher on lower carbon heats; the more basic slags are higher in iron oxide for a given carbon content; certain amounts of iron oxide and lime arc required for phosphorus removal and a high lime-silica ratio is needed if the sulphur content of the metal is to be diminished. In recent years a number of quantitative relationships involving these and other variables have been developed through plant and laboratory studies by numerous investigators in this country and abroad. The direct comparison of laboratory and open-hearth data involves the question of the degree to which each approaches the conditions of chemical equilibrium, a question which, from the open-hearth standpoint, we are scarcely yet in position to answer. And, although a considerable body of dependable equilibrium data has been obtained in the laboratory, this work has not yet been extended to cover the ranges of slag composition encountered in basic open-hearth practice. In the present study we shall not attempt to draw comparisons between data from the two sources; nor do we attempt to draw conclusions as to the approach to chemical equilibrium in the open-hearth furnace. Many of the operating variables that are likely to be erratic or unrecorded during the earlier stages of the heat become more consistent and arc more readily subject to observation and control during later stages. For example, undigested lime or ore gradually goes into solution, the rate of carbon drop levels off, and the temperature of the metal is brought under control. For these and other reasons a greater degree of self-consistency is likely to be found among data obtained during the

Jan 1, 1940

By Karl L. Fetters

In the process of making steel in the open-hearth furnace the refining of the metal during the working period of the heat is largely accomplished through the agency of the slag. From the birth of the process until the present day steelmakers have recognized the importance of the slag in determining the composition of the metal and the quality of the product. The general relationships between slag composition and metal analysis have long been known, at least in a qualitative or approximate fashion. For examples, the iron oxide content of the slag is higher on lower carbon heats; the more basic slags are higher in iron oxide for a given carbon content; certain amounts of iron oxide and lime arc required for phosphorus removal and a high lime-silica ratio is needed if the sulphur content of the metal is to be diminished. In recent years a number of quantitative relationships involving these and other variables have been developed through plant and laboratory studies by numerous investigators in this country and abroad. The direct comparison of laboratory and open-hearth data involves the question of the degree to which each approaches the conditions of chemical equilibrium, a question which, from the open-hearth standpoint, we are scarcely yet in position to answer. And, although a considerable body of dependable equilibrium data has been obtained in the laboratory, this work has not yet been extended to cover the ranges of slag composition encountered in basic open-hearth practice. In the present study we shall not attempt to draw comparisons between data from the two sources; nor do we attempt to draw conclusions as to the approach to chemical equilibrium in the open-hearth furnace. Many of the operating variables that are likely to be erratic or unrecorded during the earlier stages of the heat become more consistent and arc more readily subject to observation and control during later stages. For example, undigested lime or ore gradually goes into solution, the rate of carbon drop levels off, and the temperature of the metal is brought under control. For these and other reasons a greater degree of self-consistency is likely to be found among data obtained during the

Jan 1, 1940

By William Kelly, Frederick Laist

The slow, tedious and costly method of reducing copper matte to metallic copper in the reverberatory furnace, commonly known as the Welsh process, was displaced by the rapid and inexpensive converter process about fifty years ago. The art of refining copper was for many generations a carefully guarded secret among the Welsh furnacemen and was handed down among them from father to son. An attempt was made to apply it at Butte when the smelting of copper ore first started there but it proved too difficult and expensive under local conditions, and until the introduction of the converter process the entire output of copper was shipped to Swansea in the form of high-grade (60 per cent) matte. Before the construction of the Manhes converter, attempts had been made to reduce copper matte in the bessemer steel converter, but these failed, chiefly because the air was blown through the bottom of the vessel. The air holes became clogged with chilled copper as soon as the reduction started. Manhes partially overcame this difficulty by placing the tuyeres in the side of the vessel, some, distance above the bottom, thus allowing space beneath them for the copper to accumulate. Nevertheless the first attempts to operate such a converter at Butte were not successful, as no means had been provided for clearing away obstructions, which formed to some extent in front of the tuyeres. This difficulty was overcome by drilling holes in the outside wall of the air chamber directly opposite the holes in the shell and lining, so that these could be punched

Jan 1, 1934

By F. H. Parsons

At the time that converting or bessemerizing of copper matte first began to be practiced by the smelters in the copper industry, converters were lined by tamping the flux, usually siliceous ore, into the steel shell. About 1910, practice in several plants had demonstrated the superiority of a basic lining of nonfluxing material and the addition of flux through the mouth of the converter. Adoption of the new method by nearly the entire industry followed. Basic linings are made of refractory "magnesite" brick manufactured, usually, from imported dead-burned magnesite. Either the linings are laid up in the shell dry or the bricks arc bonded with some mortar dictated by the preference of the user. Foreign, usually Austrian, magnesite was the main source of supply for these bricks, although there were domestic sources, but it was not until costs began to rise a few years after the first use of basic linings that these domestic sources were investigated to any great extent. Exigencies created by the World War necessitated the use of the domestic magnesite during this period, but the brick manufacturers reverted to the use of the foreign product after the war, because the impurities therein made a brick with a better fire bond than the purer domestic magnesite. Therefore the domestic product became available at a favorable price for any metallurgical process that permitted its use. The United Verde converting practice using magnesite linings has resulted in a high consumption of magnesite because of the zinc content of the low-grade matte. It was realized that the cost of these linings represented a large portion of the cost of converting, and an investigation was started to find means by which these costs might be lowered. In the neighboring state of California rather pure magnesite was being mined and treated for the market. The product, however, was purposely made to simulate the dead-burned grain magnesite product from foreign sources by the addition of fluxing impurities before burning. It was learned that this dead-burned California grain magnesite could be laid down in Arizona at a price one-third that of the same tonnage of "magnesite" brick manufactured from foreign magnesite. It was also learned that at least one California producer was ready to supply the market with periclase, which was a grain magnesia produced by burning the

Jan 1, 1934

By B. L. Sackett

The converting of lead matte is not a general practice at lead smelters, therefore a description of the methods used and developments made during the past 20 years in converting both lead and copper matte at the Tooele plant of the International Smelting Co. in Utah may be of interest. The Tooele converting plant consists of five electrically operated stands in which are used 96 by 150-in. horizontal, barrel-type converters each of about 10 tons capacity. The main converter building is served by a 60-ton electric traveling crane and the casting shed by a 30-ton crane. Copper matte is tapped from near-by reverberatory furnaces into launders that deliver into 8-ton ladles on the converter floor. Lead matte is brought from the blast furnaces, about li mile away, in 5-ton ladles mounted on trucks, hauled by electric locomotives. As the Tooele copper plant has never been a producer of much copper, and has been Irregular in its operation, the original converting plant with its small converters has been adcquate, although decidedly out of step with large, modern plants. Tonnage figures in this paper are for comparison with each other only, as they are not at all in line with prevailing tonnages at larger plants today. The reasons for the changes made in converting practice at Tooele, as well as the improvements in the results obtained as these changes were perfected, are given in the following pages. Lead-matte Converting, 1914 to 1927 In 1914, Oscar M. Kuchs wrote a paper for the Transactions,' describing in detail the lead-matte converting methods at Tooele up to that time. The method perfected only a few months before Mr. Kuchs' paper was written consisted, briefly, of blowing the lead matte alone, without the addition of siliceous fluxing ores, to a point where the lead and zinc were largely eliminated by volatilization. During this period

Jan 1, 1934

By R. G. Knickerbocker

The copper smelter and refinery of The Messina (Transvaal) Development Co. Ltd., at Messina, South Africa, was erected in 1920 and 1921, but initial operations were deferred until late in 1922 on account of the collapse of the copper market in the interim. Because the anticipated production was to be only 20 tons of copper per day from a high-grade matte containing 60 per cent copper, and because the ore did not contain gold or silver in quantities sufficient to pay for separation from the copper bullion, it was decided to design the plant along lines quite different from the modern smelteries treating large quantities of sulfide ores carrying appreciable amounts of gold and silver. The plant was built, therefore, for the production of copper by the Nicholls-James process. By this method two-thirds of the copper matte produced was roasted in a Merton calciner and charged into a bath of the remaining one-third portion of unroasted matte in a reaction furnace. The original design was not only novel in providing for the Nicholls-James process, but also in comparison with new plants built elsewhere, because of factors inherent in the locality. Sources of supply for machinery were very remote; it takes three months after the placing of an order to accomplish delivery of machinery into Messina from the United States. Native labor was available at 30 to 40¢ per 8-hr. shift. These factors permitted plant design which provided no overhead cranes, no matte ladles, no slag pots or other usual mechanical means of transportation between departments. The design resulted in a compact simple plant with furnaces arranged close together and terraced so as to provide for gravity transfer of molten material from one furnace to another and for manual disposal of products. Fig. 1 shows the relative location of furnaces. The three furnaces are parallel on successively decreasing elevations. A superimposed feed floor was provided above the furnaces. All three furnaces were provided with coal-dust burners served from a Bonnot system of coal-dust transportation, and the waste gases were discharged into waste-heat boilers. A short trial of the Nicholls-James process developed several facts which brought about a decision to abandon the process. Owing to the

Jan 1, 1934

By George E. Beavers

Under the title used for this paper, the converter practice of the Tennessee Copper Co. is described in the TRansactions of the Institute.l That article contains the following statement: "The limit in capacity of the Peirce-Smith converter in terms of concentrate has not been reached but approximately as high as one ton of concentrate has been handled for each ton of liquid matte charged and good working conditions have been maintained." It is not the intention of the writer to review the conditions leading up to the introduction of a Peirce-Smith converter at Copperhill, but, instead, to offer later data to show how satisfactory the installation has been. Converter Campaign, June 15, 1930 to August 23, 1931 Blowing time (converter days)............... 349.90 Number of blows........................... 427.00 Tons matte charged........................ 38,335.00 Copper assay of matte, per cent.............. 13.67 Tons blister copper made.................... 9,246.00 Time to blow one ton of copper, minutes...... 54.50 Air used per minute, cu. ft................... 12,500.00 Blast pressure, lb........................... 12.80 Air per ton copper, cu. ft.................... 681,330.00 Air per ton iron, cu. ft....................... 199,040.00 Tons flux used............................. 11,845.00 Tons flux per ton matte..................... 0.31 Tons flux per ton copper.................... 1.28 Tons flux per ton iron....................... 0.37 Tons concentrate treated.................... 28,960.00 Copper assay of concentrate, per cent......... 17.13 Tons concentrate per ton matte.............. 0.76 Tons material charged per converter day...... 235.70

Jan 1, 1934

By George E. Beavers

It is well to understand at once that converter gas is always mixed with copper blast-furnace gas at Copperhill before entering the Glover tower, and, at times, roaster gas is also added along with the converter and blast-furnace gas. Fig. 1 is introduced to show the variations in the sulfur dioxide content of the gas and the ability of the acid-plant operator to control the variations, resulting in a very even grade at the Gay-Lussac tower entrance. The sulfur dioxide content of the gas entering the Gay-Lussac tower must be low and regular to assure low niter consumption. The blast-furnace gas alone is rather uniform as regards sulfur dioxide content and the variations, as shown in Fig. 1, are chiefly caused by converter changes—out for charging, slag pouring, copper pouring, and, of course, the "finish." The combined gas during the converter finish is, at times, almost free of oxygen and the sulfur dioxide content reaches 13.0 per cent. At such times air is added at the Glover tower.

Jan 1, 1934

By Charles R. Wraith

The property of the Roan Antelope Copper Mines, Ltd., is situated at Luanshya, north central part of Northern Rhodesia, and is connected with the main line of the Rhodesia Railways, Ltd., by a 24-mile branch line from N'dola. Most of the material and supplies that are not available in Africa are shipped into the country via Beira, on the East Coast of Africa, a distance of 1473 miles from Luanshya and the nearest ocean port. The output of copper is shipped by rail to Beira and thence by boat to the market. Passenger travel is mostly via Cape Town, a distance of 2162 miles from Luanshya. Excellent passenger trains operate on regular schedules between Cape Town and Northern Rhodesia. Luanshya is 13" south of the equator at an elevation of 4000 ft. above sea level, and has a mild healthful climate. During the rainy season, from October to March, the rainfall averages 53 in. The area is profusely covered with vegetation. An ample supply of water for industrial and domestic purposes is available throughout the year. The more important supplies for smelting purposcs, such as wood, coal, limestone, silica, quartzite, and clay for "dolly" mud, are conveniently available. Excellent grades of coal, coke and firebrick are obtained from the Wankie Colliery Co., Ltd., Southern Rhodesia. The Wankie colliery has just completed the erection of equipment for the manufacture of silica brick. An ample supply of quartzite of good grade is available for this purpose. Silica brick is also obtained from the Belgian Congo and from India. Limestone is quarried from a large deposit a few miles from Bwana Mkubwa, which is 30 miles by rail from the smelter. An abundant supply of clay for " dolly " mud and for lining launders is available within a short distance of the smelter. Metallurgy Copper is contained in the ore as chalcocite, chalcopyrite and bornite. Chalcocite predominates in the eastern end of the orebody almost to the exclusion of other copper minerals. The amount of chalcocite decreases with a corresponding increase of chalcopyrite and bornite toward the western extension, where the latter predominate.

Jan 1, 1934

By W. B. Boggs, J. N. Anderson

An account of the operation of the Noranda smelter covering the period from December, 1927, when the smelter started, until December, 1929, was given in a paper published by the Canadian Institute of Mining and Metallurgy early in 1930.' The object of the present paper is to review matters of interest in the original paper and to describe the present construction and operation of the smelter, with particular reference to the reverberatory-furnace practice. The Noranda smelter was built by Noranda Mines, Ltd., from the designs of A. E. Wheeler, consulting metallurgist of New York, for the reduction of ores and concentrates from the Horne mine and to treat, on a custom basis, ores and concentrates from other mines in the Rouyn district. The smelter is at Noranda, about 330 miles directly north of Toronto. Construction work was started in June, 1926, and copper was first produced in December, 1927. The smelter was built throughout as two duplicate units, each with a rated capacity of 500 tons of ore per day. Operations were started with one unit and after several months it was found that one unit alone could smelt 1000 tons of ore a day. At that time there was only one hoisting shaft, No. 3, and its hoisting capacity and the capacity of the original crushing plant were insufficient to supply ore for the two units at the increased tonnage. Accordingly No. 4 shaft was sunk and a new crushing plant provided to handle the ore hoisted there. Operating two units at capacity also required the installation of two more converters to treat the additional quantity of matte produced. This work was completed in the latter part of November, 1929, and since then both smelter units have been operating continuously. In 1928, a mill to treat 500 tons a day of concentrating ore was completed. Since then the concentrator capacity has been increased twice, and at present 2000 tons a day of ore are being concentrated. The layout of the surface plant of the mine, concentrator and smelter is shown in Fig. 1 and the general plan in Fig. 2. The smelter is shown in more detail in Fig. 3. The original hoisting shaft, No. 3, is now used only for handling men, supplies and waste rock, and the original crushing

Jan 1, 1934

By A. G. Suydam

Ancient as is the art of producing copper, so ancient is one of its eternal problems: refractories. Looking backward, in the light of present knowledge, clouded though it be, one cannot avoid a sense of profound respect for the patience, acumen and courage of those who have wrought so successfully despite difficulties, a more accurate understanding of which seems merely to accentuate. The problem of refractories arises from the necessity for constructing containers in which to bring natural ores into thermal solution in order to effect reduction and separation of the valuable metal content and free it of undesirable contamination while producing a waste slag. The very forces known to produce thermal solution of sufficient liquidity are necessarily destructive of exposed refractory linings in proportions varying inversely with thermal conductivity, directly with the porosity and ease of fluxing of the refractories used, and directly with the temperatures employed. Beyond this, mechanical forces are set up within the refractory lining incident to rapid heating or cooling, shrinkage at high temperatures, and unequal or excessive loading, that frequently pass the elastic limit, causing spalling or distortion. The effects of these forces are often accentuated by the necessity of providing at the same time as hot a bath and as cool a container as are consistent with fuel economy. Evolution in the Use of Refractories Clay Brick.—Fireclay and clay refractories had to yield wherever they were in contact with slags rich in basic oxides, first because of the combining power of those oxides with free silica and secondly because of the tendency of alumina to form either aluminates or alumina-silicates of varying complexity. Of no less importance, the incipient fusion point was low as compared with good clay brick of today, while the brick lacked strength at high temperatures. Expansion was slight, therefore it was difficult to make and maintain tight walls, roof, or bottom. However, the spalling tendency was low because of some of these properties and clay brick have retained their place in roasting furnaces, boiler furnace linings, and in the upper walls of some of the reverberatory types such as some refining, anode and wire-bar furnaces. They have served often and most efficiently for hot patching.

Jan 1, 1934

By Harold H. Goe, Louis V. Bender

The recovery of arsenic was started at the Washoe Reduction Works (later called the Anaconda Reduction Works) in 1904, when an arsenic plant was constructed adjacent to the lower end of the 60-ft. main flue.' This plant originally consisted of a Brunton roasting furnace for treatment of flue dust and a small reverberatory furnace for treating the crude arsenic produced in the roasting operations, together with the necessary kitchens for condensation of the fumes and grinding and barreling apparatus for preparing the refined white arsenic for market. The Brunton furnace was fired with wood. It was 14 ft. 6 in. in diameter with two outside fireboxes. The hearth revolved upon a central shaft. The charge was fed through a dust-tight hopper on to the center of the hearth and slowly moved to the outer edge of three rows of stationary rabbles fastened in the roof of the furnace. There was a 3-ft. space between the hearth and roof. A crude arsenic was condensed from this furnace. The hot flame from the fireboxes passed directly over the charge, vaporizing the As203 that was carried through brick flues to the condensing chamber. This chamber was 8 ft. high, 16 ft. wide and 240 ft. long, divided into small kitchens by partitions spaced every 7 ft. It was connected with the main flue system to provide draft and to receive any losses from the chamber. The small kitchens were connected by openings in such a way that the fumes took a zigzag course, causing them to come in contact with the cold walls and ceiling and condense the arsenic. There was an opening from the outside into each kitchen, for the removal of the deposited arsenic, which was closed by a wooden door. The refining furnace was a small coke-fired reverberatory having a hearth 10 by 20 ft. in dimensions. It had three hoppers through which the crude charge was introduced, and one firebox placed at one end. The hot gas from the burning coke passed over the charge, volatilizing the arsenic and carrying it to the kitchens.

Jan 1, 1934