Search Documents

ARTHUR F. BRAID,* New. York, N. Y. (written discussion ?). -I have read Mr. Cain's paper with a very great deal of interest. The subject is one that is worth a great deal of thought and research at this time. It is a well-known fact that the strong affinity of steel or, iron for oxygen is perhaps the main detriment to sound production, and, as the author states, it is-indeed remarkable that today there are no analytical methods in general use in steel works for determining the gaseous con-stituents of steel.

Jan 10, 1919

By Eckley B. Coxe

IN making surveys in the anthracite coal regions of Pennsylvania, the ordinary engineer's chain (50 or 100 feet long) is generally used, both above and below ground. Sometimes, where it is difficult to chain, as, for instance, across a chasm, a wire is stretched from one station to the other, the distance is marked on the wire and its length is then measured with the ordinary chain. Having had occasion lately to make some surveys where it was necessary to determine with great accuracy the position of the land or property line, not

Jan 1, 1874

By Eckley B. Coxe

In making surveys in the anthracite coal regions of Pennsylvania, the ordinary engineer's chain (50 or 100 feet long) is generally used, both above and below ground. Sometimes, where it is difficult to chain, as, for instance, across a chasm, a wire is stretched from one station to the other, the distance is marked on the wire and its length is then measured with the ordinary chain. Having had occasion lately to make some surveys where it was necessary to determine with great accuracy the position of the land or property line, not

By means of top-slicing methods, wide veins, masses, or thick beds of soft ore may be mined, where the caving of the overburden is of no consequence. The ore is mined in a series of horizontal floors or slices, taken in descending order, from the top of the deposit. Each floor is mined in small sections, the roof of each section being allowed or forced to cave before an adjacent section is attacked. The method as used in two mining districts—Marquette, iron, and Pachuca, silver-gold—is described in the two papers following:

Jan 1, 1925

By R. Hogg, R. Schaller, P. Luckie

OVERVIEW This paper is a review of two fine particle processing unit operations - classification and mixing. It may appear that these two processes have nothing in common. In fact, the simplified symbols used to represent them on flowsheets - a tipped equilateral triangle such that the base lies along the vertical - are totally reversed. For classification, a splitting action, the apex of the tipped triangle is to the left of the vertical base. A horizontal flow arrow indicating input is drawn to the apex while two horizontal flow arrows indicating outputs are drawn from the top and bottom of the vertical base, respectively. On the other hand, for mixing, a blending action, the apex of the tipped triangle is to the right of the vertical base. Two horizontal flow arrows indicating inputs are drawn to the top and bottom of the vertical base, respectively, while a single horizontal flow arrow indicating output is drawn from the apex. Classification is the process whereby a collection of particles is separated into two or more portions differing in some physical or chemical property. Mixing is the process whereby two or more portions of particles differing in some physical or chemical property are combined into a collection of particles. The common bond that links them together in the same review is the fact that they both require dispersion of the particles in order to achieve successfully their individual objectives.

Jan 1, 1980

By Bruce W. Gonser, Robert J. Nekervis

EACH metal has a unique combination of properties that distinguishes it from other metal;. Su& a combination may account for applications that cannot be met very well by anything else. This is particularly true of tin, one of the earliest metals used by man. Table i shows the consumption of tin in the United States by usage for the years 1949 and 1950, which may be considered to be generally typical of current usage. The major applications-tinplate, solders, bronze, and babbitt-were well established a century ago. Applications for the metal and (probably later) of tin alloys started more than 5000 years ago. Fortunately for the ancients, tin is one of the metals whose oxide is easily reduced by carbon. Nature further simplified the refining problem by providing only one ore of commercial importance, cassiterite (SnO2). Tin is easily reduced from cassiterite. All that is necessary is high heat in the presence of carbon. It is quite possible that back in the dawn of history, tin was discovered when black alluvial sand was thrown on a fire, either to bank it or to put it out. SOURCES OF TIN All the important tin-producing centers are in the southern hemisphere, far away from the' industrial consuming centers. Fig 1 shows their location and 1950 production. Reserves are adequate, and the industry is able to produce considerably more tin than is

Jan 1, 1953

By C. M. Loeb, A. M. Gaudin, J. H. Brown

THE object of most size reduction operations in the mineral industry is to liberate the grains of valuable minerals in the ore from those of the gangue. This is usually accomplished by crushing and grinding the entire mass of ore until there is only a small probability that any single particle contains more than one mineral. During this size reduction only limited control exists over size or composition of the particles exposed to the breaking action, and there is no control over the paths followed by cracks generated during the operation. This lack of control usually results in overgrinding and in production of large quantities of very fine material. The first detriment, overgrinding, is costly in itself, but when combined with the second factor it is doubly so. Not only is the fracture of a free particle unnecessary-the fracture of these particles may also make subsequent separation operations difficult, inefficient, and wasteful. It has been pointed out' that if the object of size reduction is to liberate the valuable mineral component of the ore then, ideally, fracture should follow intergranular paths to the exclusion of trans- granular ones. This would result in liberation of the valuable minerals with as little size reduction as possible. This ideal comminution operation is referred to as intergranular comminution, and it was the object of the investigation to determine the extent to which it could be developed by heat treatments.

Jan 4, 1958

By Desmond De Villiers Oxford, Peter H. Williams

Nchanga Consolidated Copper Mines, Ltd., is one of the four mines operated by the Anglo American Corp. in Northern Rhodesia, now the independent nation of Zambia. Today Nchanga is the second largest single producer of copper in the world, accounting for more than 200,000 long tons of copper annually at a monthly milling rate of 430,000 tons of ore. The average mill-head grade is about 5.20% total copper, of which approximately 60% is copper sulfide. The remainder, referred to as "acid soluble," is copper oxide. Nchanga mines three separate orebodies varying in grade and sulfide-to-oxide ratio. Ore is processed in two separate grinding circuits feeding into three separate flotation circuits. A fourth flotation circuit treats leach plant residue for additional recovery of sulfides.

Jan 1, 1965

By J. D. Rice, H. H. Rachford, G. H. Bruce, D. W. Peaceman

The problem of unsteady-state gas flow through porous media leads to a second-order non-linear partial differential equation for which no analytical solution has been found. In this paper a stable numerical procedure is developed for solving the equation for production of gas at constant rate from linear and radial systems. An electronic digital computer is used to perform the numerical integration using an implicit form of an approximating difference equation. Solutions are presented in graphical form for various values of dimension-less parameters. The solutions are compared with the laboratory study of gas depletion in a linear system. INTRODUCTION Production of fluids from porous rock reservoirs is essentially a transient process. Transient gradients develop as soon as production begins, and further withdrawals continue to cause disturbances which propagate throughout the reservoir, each adding in some way to the prior ones. A correct mathematical analysis of this behavior is complicated by the fact that the transient or unsteady-state flow of compressible fluids must be described by difficult second-order partial differential equations. As a practical matter, three distinctly different cases arise: 1. Flow of single-phase liquid 2. Flow of gases 3. Multiphase flow The first of these has been found to give a linear second-order equation similar to the well-known heat flow equation. where 7 is fluid density, 8 is time, and a a constant for the system, provided d 7 = r cdp, where the compressibility, c, is constant over the range of pressure, p, considered. Solutions of Equation (1) for both linear and radial flow are available in several forms.2'3' On the other hand. the second case, which is the flow of gas, gives a non-linear second-order equation, the solution of which is not known. Although a number of approximate solutions have been proposed, each is limited in value by the associated simplifying assumptions. Inasmuch as the analysis of transient flow is limited to liquid systems, a solution of the second case is necessary if further progress is to be made in studying underground fluid movement. For this reason a solution of Equation (2) was undertaken by means of numerical integration of approximating difference equations. BASIC DIFFERENTIAL EQUATION Equation (2; is derived by combining the equation of continuity, the perfect gas law, and Darcy's law to give the basic differential equation Strictly speaking, the fluids present in a gas reservoir are not perfect gases. Furthermore, there are indications that over certain velocity ranges, Darcy's law is not applicable. It would not be practicable. however, to attempt to obtain numerical solutions which would take into account all the possible variations from these laws which could occur in actual systems. The

Jan 1, 1953

By R. A. Perry

A speiss is an artificial arsenide, sometimes an antimonide, formed in lead smelting, smelting of oxide copper ores, and in some lead-refining operations. The production of speiss is closely allied with the amount of arsenic and metallic iron in the blast-furnace charge, while dross is produced mostly from the copper and sulphur, with varying amounts of antimony, arsenic and nickel. Most speiss is essentially an iron arsenide containing some sulphur, copper and nickel, if these elements are present in the furnace. The recent introduction of the soda treatment of lead drosses has produced a speiss high in copper. Speiss is not miscible to any extent with molten copper, lead, matte or slag. Miscibility increases with temperature. This property of the speiss is clarified in the latter part of this paper. Most speisses are produced in conjunction with matte and their sulphur content points to the solubility of matte in speiss. . Some plants have insufficient speiss to warrant separate treatment. When arsenic is high in the blast-furllace charge, there may be sufficient speiss to tap separately from the first settler. Arsenic From the foregoing, it is obvious that a study of the treatment of speiss would not ,be complete without consideration of arsenic. Inasmuch as perhaps go per cent of the blast-furnace feed may be sinter, a definite knowledge of the properties of arsenical ores and compounds that go in to the roaster and sinter feeds is of value. Lcad smelters obtain their arsenic from ores and concentrates, as well as from flue dusts shipped from other plants, particularly copper smelters. The percentage of arsenic in ores is usually low, although occasional lots of high arsenical products are smelted. These crude ores may be sulphides, as arsenopyrite; or oxides like scoiodite, which carry gold or lead minerals. The small amount of arsenic in the original ores received is greatly increased by a circulating load of arsenic in secondary products at are resmelted for their lead or precious metals. More important than the arsenic present directly in the concentrates treated is that in fumes and dusts that may originate in the blast furnace or sinter plant of the lead smelter, or may come from a copper smelter as lead-bearing converter fume or roaster dust. Thus the lead smelter not only must treat its own arsenic but is compelled to handie that from some copper and perhaps other lead plants. Where copper and lead plants exchange products, each plant may be forced to re-treat part of the arsenic it has shipped ,outside its own confines. Speiss from the lead blast furnace contains chiefly iron and arsenic with some lead, copper and sulphur and often appreciable amounts of gold and silver. If its copper content is sufficient this speiss may be sent to the copper smelter where usually it is treated in the converter and the copper is separated as blister. Much of the arsenic and most of the lead are recovered as converter fumes

Jan 1, 1944

Discussion of the paper of G. P. Hulst, presented at the Salt Lake meeting, August, 1914, and printed in Bulletin No. 92, August, 1914, pp. 1865 to 1871. L. S. AUSTIN, Salt Lake City, Utah.-One question that I would like to ask is in regard to the sampling kettle which stands beside the softening furnace. How is the sample taken or melted? G. P. HULST.-The sampling kettles in question are of 45 tons capacity and are fired with coal. The lead bullion comes in on a high line in car-load lots of 30 to 40 tons. The entire lot is trucked from car, weighed over scales and thrown into the sampling kettle, which has been previously cleaned. The kettle is fired and the bullion melted, then heated up until no frozen bullion' remains in the .bottom or adheres to the sides of the kettle. The kettle is scraped clean, sides and bottom, and the wet dross is skimmed Doff clean into molds. The kettle is stirred thoroughly during the period in which 24 to 30 samples of approximately ½ assay ton are taken. These samples are taken in a long-hand loci iron mold, providing for taking six to eight "gumdrops." The mold is inserted and heated to the same temperature as the molten metal, and gumdrops are clipped out and cooled in the mold by dipping 'the bottom of. the mold into Water. The samples should be free from fins. The gumdrops represent the contents of the kettle, and in assaying, they should not be clipped, but weighed up and the result computed. The dross bars are weighed and sawed according to old hammer and punch sample template. The sawdust represents a sample of the dross. In eases of high gold in the bullion there is a slight correction in the assay. This method of sampling was developed at the Omaha plant of the American Smelting & Refining Co., W. T.. Page, Manager, in 1903 or 1904, for adjusting differences in checking smelter sampling. L. S. AUSTIN.-The base bullion which you treat is very clean. It has been remelted, and skimmed before shipment. How rapidly can you get it through the softening furnace, and how does it compare with ordinary base bullion? G. P. HULST.-Twenty-four hours is the softening limit on these charges. In other words, it is charged into the furnace, heated up, and the first and second skimmings are taken off, and it goes into the kettle the next day; that is, 24 hr. elapse from the time the bullion is charged to the time it is in the kettle. L. S. AUSTIN.-It seems to be a long time to hold such a pure bullion. I suppose, though, it comes around conveniently so that you have a time for all the other operations of refining. Isn't that the case? MR. HULST.-Every refinery is based on those principles.

Jan 11, 1914

"The earth is the Lord's and the fullness thereof." Science knows no national boundaries, knows no country. These views might be taken as premises for a discussion of the development of the mineral resources of the world. Yet how far man is removed from such broadminded ideals! In this day of jealousy and dread, of political machinations and statesmen's maneuverings, of rantings and crass ravings against fellow-beings, man's attitudes towards God's gifts are but the measure of the smallness of man. That the exploitation of mineral wealth should be for the common good, none will deny, but what is meant by "common"? Is it for the common good of The Pennsylvania State College, Pennsylvania, the United States of America, or the world? If man is big enough and honest enough to state categorically that by "common," as used in the context above, is meant that the utilization of the mineral wealth should be for the mutual benefit of all mankind, and that all concerned strive to see that such a viewpoint is executed, then there is brighter hope for the future. It is to be hoped that the world may have at least a few years or possibly a generation of peace. What can be done in these years to work towards the ultimate goal of human improvement? The work of statesmen in the drafting of international agreements and the easing of small frictions which appear between nations, as they

Jan 1, 1950

By Fred C. Bond

AN improved method of measuring the surface area of a comminution product down to any desired particle size has been developed. The method is largely graphical, and requires relatively little calculation. It can be used with materials containing large amounts of clay, or excess fines of any kind. The surface area S is expressed in terms of square meters per 100 c.c. of solids. This expression is based upon volume rather than weight, and is considered to be more suitable for crushing and grinding problems, where the machines ordinarily are operated under constant volume conditions. It has the advantage that surface areas of different materials can be compared directly, even though their specific gravities may be unknown. The "specific surface" of a material, or the number of square centimeters of surface area per gram, is equal to 100S divided by the specific gravity. Particle diameters are expressed in microns, K being the diameter of the largest particle present and L the grind limit, or lower size limit of the particles formed directly by grinding. The value of K does not always coincide exactly with the diameters of the largest particles retained on the coarsest testing sieve, since the largest particles formed in crushing and grinding are often slabby, or have other irregular shapes. L has been chosen as 0.700 microns, or number 27.5 in the size number series, for reasons to be explained later. S is the total surface area of all particles larger than L microns in diameter. When a homogeneous comminution product is separated into fractions whose diameters bear a constant geometrical relationship p to each other, and the log per cent weight retained upon each size and passing the next larger size is plotted against the log diameter, a straight line results, which is called the distribution line, and which has a slope in called the distribution modulus. The distribution line can be extended to any desired limiting size, and the total surface area of all particles larger than that size can be calculated. The equation of the distributi0n line is [w = Cx" [I]] where w is the per cent weight retained on each successive sieve; x is the particle diameter; and C is the per cent weight retained on an opening of unit diameter.2 Theoretically, the distribution line can be extended to include infinitely small particles. When it is so extended, and the assumption is made that no fine particles are present other than those accounted for by the equation; the log per cent cumulative weight passing any series of sizes, plotted against the log diameter passed, will result in a straight line lying parallel to and above the plotted distribution line.

Jan 1, 1941

By J. K. Stanley

IT has been known for many years that the magnetic properties of a ferromagnetic material are very sensitive to internal strain. Any structure-sensitive property such as ferromagnetism, which is a-function of the regularity of the atomic arrangement, and discontinuities such as grain boundaries must show deviations from an equilibrium state when the regularity of the lattice is disturbed. Experience has shown that such distortions have a marked effect on the magnetic properties of a ferromagnetic material. The purpose of this paper is to show the feasibility of using magnetic methods in detecting internal strains. More specifically, it will be shown how such properties as permeability, remanence, and coercive force change on cold- working of aluminum iron and how these magnetic properties change during the annealing below the recrystallization temperatures. This work was conducted for two reasons. One was the desire to see in what manner strains are relieved in aluminum iron at low temperatures-a subject of practical importance in soft magnetic materials-and the other was the feeling that a study of cold-working and strain relief might shed some light honour understanding of what takes place in the deformation of metals DEFINITION IN TERMS It is necessary to define the terminology used in this paper because magnetic nomenclature is usually unfamiliar to the metallurgist, and certain other terms, such as recovery, have been used ambiguously by metallurgists and others. Magnetic terms are best explained by reference to the common hysteresis loop of ferromagnetic materials (Fig. I). In Fig. I the coordinates are magnetic induction, B, (ordinate) and magnetizing force, H, (abscissa). The c.g.s. unit for the magnetic induction is the gauss, the number of flux lines per square centimeter, and the unit for the magnetizing force is the oersted, the number of lines of force per centimeter in air or vacuum. If one starts with a completely de- magnetized specimen and magnetizes it, a typical virgin magnetization curve beginning at the origin is obtained. This curve is also the locus of tips of a family of smaller hysteresis loops. From this curve it is to be noted that the induction is not proportional to the magnetizing force, and that it appears to approach a limiting induction called the saturation value, and does so for values of H of 1000 oersteds. The ratio of the magnetic induction (gauss) to the magnetizing force (oersteds) is called permeability, p. Thus µ = B/H and if the magnetizing force H (oersteds) is specified as 10 or IOO oersteds, the permeability is written as µ10 or µ100. This is the accepted notation, and will be used in this paper. The values of permeability for H = 10 or 100 are chosen as indices for

Jan 1, 1945

By V. S. Tuman

In the first part of this paper, an estimate is made of the magnitude and extent of the thermal stresses which result from mud circulation. Our study is made for the period of relaxation, i.e., when the drilling operations are terminated and mud circulation is discontinued. In the second part, we have shown how the thermal stress can affect the wave propagation. An example is worked out for a Carbondale sandstone (Illinois basin). It is concluded that, in extreme cases when differential temperature is about 100°F, the travel time will be affected by about 2 microsec/ft. For more refined permeability studies, such corrections would be desirable. MAGNITUDE AND EXTENT OF THERMAL STRESSES RESULTING FROM MUD CIRCULATION When geological beds are penetrated by boreholes, it is observed that the formation temperatures increase with depth; however, the temperature profile varies considerably from place to place. During drilling operations, mud pumped from a pit at atmospheric temperature absorbs heat from the lower beds. As a result, the drilling fluid returning to the surface is at a higher temperature than the surrounding beds. The temperature gradient in a well depends on the rate of mud circulation and other factors; normally, this temperature gradient is smaller than the natural gradient in the formations. (See Fig. 1.) The temperature difference between the mud column and the surrounding beds gives rise to therma1 stresses (which are also important in the phenomenon of fracturing), and it is expected that they will affect the transit time of the sonic front in velocity logging.1 It is evident chat beds near the surface will undergo various cycles and will have a complicated dynamical and thermal-stress history. A limestone bed which is just penetrated will act as a source of heat, warming up the cold mud pumped down a few minutes earlier from the surface pit. The same limestone bed a week later will act as a sink, absorbing heat from the drilling mud which has already warmed up sufficiently to be at a higher temperature than the bed in question. No doubt a large section of the hole will undergo a heat treatment, which will make the calculation of the exact magnitude of the thermal stresses very difficult. In the solution of wellbore thermal equation, it is assumed that the rock around the wellbore is elastic and isotropic, and that the system is radially symmetrical. A complete dynamical solution of the thermal stresses set up at different horizons during the drilling operations would require a comprehensive history of the mud circulation, the surface temperature of the pit and the number of hours in which mud was not circulated. CONDUCTION OF HEAT AND THERMAL EQUATION OF STATE

By J. K. Stanley

It has been known for many years that the magnetic properties of a ferromagnetic material are very sensitive to internal strain. Any structure-sensitive property such as ferromagnetism, which is a function of the regularity of the atomic arrangement, and discontinuities such as grain boundaries must show deviations from an equilibrium state when the regularity of the lattice is disturbed. Experience has shown that such distortions have a marked effect on the magnetic properties of a ferromagnetic material. The purpose of this paper is to show the feasibility of using magnetic methods in detecting internal strains. More specifically, it will be shown how such properties as permeability, rema-nence, and coercive force change on cold-working of aluminum iron and how these magnetic properties change during the annealing below the recrystallization temperatures. This work was conducted for two reasons. One was the desire to see in what manner strains are relieved in aluminum iron at low temperatures—a subject of practical importance in soft magnetic materials—and the other was the feeling that a study of cold-working and strain relief might shed some light onour understandingof what takes place in the deformation of metals. Definition of Terms It is necessary to define the terminology used in this Paper because magnetic nomenclature is usually unfamiliar to the metallurgist, and certain other terms, such as recovery, have been used ambiguously by metallurgists and others. Magnetic terms are best explained by reference to the common hysteresis loop of ferromagnetic materials (Fig. I). In Fig. 1 the coordinates are magnetic induction, B, (ordinate) and magnetizing force, H. (abscissa). The c.g.s. unit for the magnetic induction is the gauss, the number of flux lines per square centimeter, and the unit for the magnetizing force is the oersted, the number of lines of force per centimeter in air or vacuum. If one starts with a completely demagnetized specimen and magnetizes it, a typical virgin magnetization curve beginning at the origin is obtained. This curve is also the locus of tips of a family of smaller hysteresis loops. From this curve it is to be noted that the induction is not proportional to the magnetizing force, and that it appears to approach a limiting induction called the saturation value, and does so for values of H of 1000 oersteds. The ratio of the magnetic induction (gauss) to the magnetizing force (oersteds) is called permeability, . Thus u= B/H and if the magnetizing force H (oersteds) is specified as ro or 100 oersteds, the permeability is written as or l00 This is the accepted notation, and will be used in this paper. The values of permeability for H = 10 or I00 are chosen as indices for 106

Jan 1, 1945

W: 0. HOTCHKISS, Madison, Wis. (written discussion*).-Mr. George's paper on. "The Wisconsin Zinc District" is a very complete and clear description, which it is a pleasure to read. His discussion of the origin of the ores is the least satisfactory part, and Mr. George correctly states present knowledge when he begins his paragraph on a origin with the words "the general supposition is, etc." There is still much to be learned regarding the details of the processes by which the ores have reached their present situation. In this the Wisconsin district differs not at all from other mining districts. It is my hope that Mr. George and others, equally clear-thinking, who are in close touch with these operations will be able at an early date to give us more information on this subject.

Jan 3, 1918

By H. Sato, I. Matsuoka, T. Kawai, S. Hasebe

In a limestone mine, a large amount of slime is washed off in order to remove impurities such as clay minerals. A fine sized calcite is, consequently, wasted and freight for carring the slime to reclaim land is expensive. The separation of calcite from the slime has been investigated by the flotation process, using sodium oleate as a collector. A satisfactory flotation of calcite from the slime requires great amounts of sodium oleate, because sodium oleate is consumed by clay minerals. But, the amounts of sodium oleate required are reduced to half by adding suitable amounts of sodium silicate. After rougher and cleaner flotation, the concentrate containing more than 97% calcite is obtained with 70% recovery; thereby the amount of slime to be reclaimed is reduced to 40%.

Jan 1, 1976

By Thomas M. Drown

At the meeting of the Istitote in New York, in February, 1880," I described a process of determining sulphur iu metallic allphides, with especial reference to the determination of pyrites in coal. The procem, as then described, is as follows: A solution of sodium hydrate, of 1.25 specific gravity, is saturated with bromine. If an excess of bromiue is used it must be neutralized by the addition of a little more sodium hydrate. The pulverized mineral or coal is moistened with 25 c.e. of this solution autl heated, then hydrochloric acid is added cautiously to just acid reaction. This operation is repeated with another 25 c.c. of the alkaline solution, and again rendered acid. The mixture should be kept hot. The contents of the beaker or dish are then

Jan 1, 1881

By R. L. Bleifuss

Most pellets swell only slightly during reduction, but some swell so enormusly that their increase in size is termed catastrophic. Since catastrophic swelling produces irregularities in blast furnace operation, an understanding of the phenomenon is of practical as well as theoretical interest. Observations have shown that catastrophic swelling is characterized by the growth of a network of fine iron filaments. Evidence is presented to show that the growth of these filaments is promoted by the presence of calcium in the magnetite lattice of the original ore. During reduction, the limited mutual solubility of CaO and FeO produces an unstable condition that inhibits the uniform nucleation of metallic iron. Other elements in the magnetite lattice whose oxides are continuously soluble in FeO in the critical temperature range have no effect. Magmatic segregation or pyrometasomatic replacements are the most likely deposits to contain calciferous magnetite. Over the past few years there has been an increasing concern throughout the iron and steel industry over the tendency of pellets to swell during reduction. Although the tendency for swelling is only one of several attributes that determine the behavior of pellets in the blast furnace, swelling-reduction tests are now part of the normal evaluation procedure of many pellet consumers. During a swelling-reduction test, pellets display either the moderate increase in size known as normal swelling, or the enormous expansion known as catastrophic swelling. This paper is concerned with the origin and control of catastrophic swelling and how it is enhanced or inhibited by the presence of minor elements. Swelling Behavior The swelling behavior of pellets during reduction is indicated in Fig. 1. The lower curve represents normal swelling and the upper curve catastrophic swelling.' During the initial stages of reduction, that is, up to about 20-25% reduction, the two curves are considered to be equivalent. Beyond this point normal pellets undergo a gradual decrease in size with increased reduction. By contrast, pellets which exhibit catastrophic swelling show a second dramatic expansion between 25 and 70% reduction. The initial normal swelling develops as the hematite of the pellet is successively reduced to magnetite, and then to wustite. Work by Edstrom' and others has shown that the reduction of hematite to magnetite is accompanied by a 25% volume increase, whereas the magnetite to wustite stage shows a volume decrease. The volume increase overlaps the wustite field because magnetite is formed simultaneously with the wustite during topochemical reduction of the pellet. The catastrophic expansion is clearly related to the wustite, to

Jan 1, 1971