Search Documents

By L. J. Buck

Lead smelting in the ore hearth, first practiced in Europe, whence the traditional name "Scotch hearth, and later developed in America to its present efficiency, is based upon the roast reduction process and is applicable to relatively high-grade galena concentrates. The roast reduction process embraces not only the roast reactions: 2PbS + 70 = PbO + PbSO4, + SO2 PbS + 2PbO = 3Pb + SO2 PbS + 2PbSO4 = Pb + 2PbO + 3SO2 but also, by reason of the addition of extraneous fuel, the reduction reaction: 2PbO + C = 2Pb + CO2 Thus in the ore hearth the roasting of lead sulphide, forming lead oxide and lead sulphate, the reaction of lead oxide and lead sulphate with lead sulphide producing lead, and the reduction of lead oxide by carbon producing lead, go on simultaneously. The Ore Hearth Function of the Ore Hearth.—The ore hearth, through the roast reduction process, provides for the extraction of lead and the immediate disposal of the smelting products—fume, hearth slag and lead, all of which are subsequently treated as follows: the fume is recovered in the flue system and baghouses, mixed with concentrates, and returned to the ore hearth in the form of ore mix; the hearth slag is smelted in a blast furnacc for the further recovery of lead; and the lead is refined and cast into a marketable product. Concentrates Amenable to Ore-hearth Smelting.—Only concentrates that are high in lead content, 70 per cent or more, are ordinarily considered amenable to ore-hearth smelting. Silica should not exceed 2 per cent and iron is allowable only to 4 per cent. These slag-forming constituents occlude both sulphide and mctallic lead, prevent the elimination of sulphur and decrease the lead extraction. Lime and magnesia are beneficial in that both favor a dry slag, promote elimination of sulphur, and prevent occlusions of metallic lead.

Jan 1, 1937

By H. J. Stehli

Probably the earliest sinter from lead ore was made in the product sinter boxes attached to some hand reverberatory roasters. The ore was drawn from the roaster bed when dead roasted, and then by laborious effort fritted together by rabbling at a high temperature in the sinter box. This was not a true sinter, but was a great improvement over the dusty dead-roasted fine ores. The Huntington-Heberlein pots were the first real advance in the art, and produced some real sinter. They required that the sulphides be preroasted to 5 or 6 per cent sulphur, and the long cycle of operating and deep bed of material used in the pots fostered the formation of densely fused material and fines. This necessitated crushing of the densely fused cake and returning to the next charge large amounts of fines screened out. The Huntington-Heberlein pots made comparatively high lead losses because the deep bed and high internal temperature together favored the volatilization of lead. Lime (CaO) was added to the charges, and this addition constituted the patented feature of the Huntington-Heberlein process. Its sole function we now know was to act as an isolator of the sulphide particles, and any other substance—for instance, iron oxide—of equal fineness performed the same function. The Savelsberg process used limestone instead of burned lime, and in this there was a definite advantage. The converting of CaCO3, to CaO is a highly endothermic reaction, so the lime acted not only as an isolator of sulphide particles but absorbed excess heat and to some extent prevented premature fusion of the sulphides. The Carmichael-Bradford process used gypsum instead of limestone, and had the disadvantage of adding more sulphur to a charge from which it was desired to eliminate sulphur. Here again the gypsum served only as an infusible isolator of sulphide particles. The first Dwight-Lloyd apparatus to sinter lead ores commercially was installed at Perth Amboy in 1907. The fact that lime in any form was necessary to sintering was quickly disproved by making excellent sinter from Bunker Hill and Sullivan table concentrates, silica sand and iron oxide only.

Jan 1, 1937

By G. L. Oldright, Virgil Miller

It is well known, in metallurgical circles, though less recognized in the technical press, that there have been remarkable increases in the capacity of the blast furnace in the last five .years. In places this capacity has been almost doubled. This paper will brief the experimental work carried on during more than five years with the direct assistance of the technical staffs of three lead smelters treating three distinctive types of feed. The first is the custom smelter of the International Smelting and Refining Co. at Tooele, Utah1, with a multiplicity of ores; the second, the Kellogg, Idaho, smelter of the Bunker Hill and Sullivan Mining and Concentrating CO.2, treating an exceedingly rich lead charge, such as is now being made more and more as the practice of concentration is improved; and the third, the smelter at Trail, B. C., owned by the Consolidated Mining and Smelting Company of Canada, Limited3, which treats highly zinciferous charges and later recovers zinc from the slag. In comparing operations at these three smelters it must be clearly recognized that the objectives in smelting differ through differences in economics. Many operations are not, then, strictly comparable. There have also been decided improvements in operation since all of the smelters were examined, though this was but a relatively short time ago. The general factors that govern speed in smelting and the loss of metals in the smelters examined will be stressed. (For details of operation the reader is referred to the original papers.) This paper will then be an examination of the fundamentals of operating units in a search for improvements, and not a description of three great lead smelters. Capacity of Blast Furnace Assuming a well prepared and self-fluxing charge to start with, enough coke must be added: (1) to melt the charge, particularly the slag-forming

Jan 1, 1937

By Arthur E. Hall

Tax operation of "softening" in lead refining is designed, as the word implies, to separate from the bulk of the lead the elements that tend to make lead hard. These elements, which invariably are present in argentiferous lead, interfere with the desilverization, and must be almost entirely absent in refined desilverized lead. They are mainly copper, antimony, arsenic and tin. The properties utilized in accomplishing this separation are the low solubility of copper or copper sulphide in lead at its freezing-point temperature, and the greater and selective affinity for oxygen of the other elements at temperatures of the order of 700° to 900' C., or, in the prcsence of sufficient, alkali, at a temperature of about 500' C. The major problems of softening lead bullion lie in the necessary subsequent treatment of the by-products of the primary operation, which contain, either chemically or mechanically, lead or reagents that must be isolated for market or re-use. Therefore an important objective of the primary operation must be the highest practicable concentration of these elements in the by-products. Within the past three decades, during which the writer has been engaged in lead refining, the technique of softening lead bullion has had as many variations as there were lead refineries. This was the result of the variation in composition and character of lead bullion recived. The writer has received for refining shipments of lead bullion varying as widely as indicated by the following table: Sb, Per CEnt As. Per Cent Sn, Per Cent Cu. Per Cent Ag. Oz. per Ton 1.00 0.05 nil 0.05 70 0.50 2.50 2.00 0.50 200 3.50 0.50 nil 0.30 350 At times these have been received uniformly and at other times irregularly with a variation from week to week as wide as those shown. Under such conditions, establishment of a universal softening technique presented many problems. Such variations existed because refinery tolls too often were based on a ton of bullion, whatever its composition, and

Jan 1, 1937

By O. P. Chisholm

Dross emerges from the blast furnace either with the lead through a lead well or by tapping from a forehearth or settler, but until a dozen years or so ago few dross reverberatories were used in western lead smelters because it was the usual procedure to circulate the dross from the drossing lead kettles back to the blast furnace. It was quite customary for lead smelters to produce matte in their blast furnaces, as this was the form in which copper could be collected, concentrated if necessary, and shipped to the copper smelters for further treatment. The main prerequisites for good matte formation arc., of course, a more or less definite amount of copper on the charge and sufficient sulphur. Such troublesome impurities as arsenic and antimony were ignored, in so far as possible, although they contaminated the bullion and thereby increased refinery costs. At the same time stocks of blastfurnace speiss containing gold, silver, lead and copper in addition to iron, arsenic and antimony, were being accumulated for want of a market to absorb them. As the dross was being fed back to the furnaces as rapidly as it was being produced, arsenic and antimony tended to build up until the arsenic could effect a combination with copper and iron and an alliance with lead, and the antimony a combination with copper and an alliancc with lead. Sulphur, being the most active agent in taking up the copper, left little of that element for the arsenic and antimony, so that on the whole a somewhat unsatisfactory metallurgy resulted. During this same period, certain lead plants were beginning to encounter trouble with zinc due to the replacement of crude ores of lot zinc content by concentrates containing high percentages of lead and zinc in sulphide form, the blast-furnace slags resulting from the smelting of such concentrates being characterized by a decidedly increased zinc content. It was soon realized that zinc plus sulphur was "poison" to a lead blast furnace and steps were taken to reduce the sulphur content,. At this stage some plants found that any substantial matte fall was out.

Jan 1, 1937

By J. L. Bray

Lead has often been suggested as a protective coating for iron and steel. Such a protective coating should possess: (1) good adhesion, (2) durability, (3) ease of application, (4) freedom from pinholes, (5) good appearance and (6) low cost, but, unfortunately, lead coatings so far produced do not have all of these characteristics. Because of the necessity of carrying out the coating process cheaply, and because lead solutions have poor plating characteristics, most of the efforts have been directed toward the hot-dipping process rather than electroplating. The hot-dipping processes using lead have not been successful chiefly because of the lack of adherence between iron and lead. In any coating process involving iron a bond can be obtained only by the formation of an intermetallic compound, the formation of a solid solution, or mechanical means. Lead forms with iron two immiscible liquid solutions, and in the solid state no intermetallic compounds or solutions. It is necessary, therefore, to employ some other element or elements as a binding agent. Tin may be used as in terneplate, or copper in the form of a thin film of electrolytically deposited metal to which the lead will adhere. Because of its physical and chemical properties, lead as a coakkg medium would undoubtedly be limited in its use to sheet and wire and then only when the article is not subject to abrasion or wear. A very large tonnage of these products is being used under such conditions and a successful lead-coating process should be a valuable contribution to the science. Two other difficulties stand in the way of successful lead coating—that of poor appearance of the weathered surface and the tendency of the lead to form pinholes during solidification. Since lead is commonly accepted as cathodic to iron, these exposed areas are ones of accelerated corrosion. Use of Zinc.—The method under consideration involves zinc as a binding agent on both the steel sheet and in the molten bath. This

Jan 1, 1937

By J. B. Austin, H. H. Pierce

In developing a consistent picture of the thermal expansion of the commercial chromium steels, which cover a range of carbon contents, it is desirable to know the expansion in the limiting case of the binary, carbon-free, iron-chromium alloys. We have, therefore, determined the thermal expansion of a series of five special alloys, ranging in chromium content from 3 to 10.5 per cent, with not more than 0.03 per cent carbon, over a temperature range extending beyond the transformation region; this low carbon content enables us to draw reasonably reliable conclusions as to the behavior of the carbon-free system. Moreover, the measurements serve to determine the beginning and the end of the alpha-gamma transformation; and they emphasize once more the fact that a transformation temperature determined by any method in which the specimen is cooled at an appreciable rate is apt to be seriously in error, particularly if the metal contains a transformation-retarding element such as chromium or managanese. The lag, or undercooling, observed at the ? — a transformation is explained in terms of the influence of temperature on transformation velocity; the rather unusual shape of the ? loop is attributed to a change with temperature and with composition of the distribution of chromium between the alpha-iron and gamma-iron solid solutions. The exact composition of the alloys, which were obtained through the courtesy of Dr. W. Rohn, of the Heraeus Vakuumschmelze A-G., is given in Table 1. Determinations of the linear thermal expansion and of transformation temperature were carried out in the vacuum dila-tometer previously described,' using a rate of heating and of cooling of 3.0" to 3.5" per minute unless otherwise stated. Magnetic susceptibility, which was measured in studying the rate of transformation, was determined with a magnetic balance, which also has been described.2 Results and Discussion Thermal Expansion Coeficients.—Numerical values for the linear expansion coefficient, derived from a smoothed curve based on at least

Jan 1, 1935

By H. H. Pierce, J. B. Austin

In developing a consistent picture of the thermal expansion of the commercial chromium steels, which cover a range of carbon contents, it is desirable to know the expansion in the limiting case of the binary, carbon-free, iron-chromium alloys. We have, therefore, determined the thermal expansion of a series of five special alloys, ranging in chromium content from 3 to 10.5 per cent, with not more than 0.03 per cent carbon, over a temperature range extending beyond the transformation region; this low carbon content enables us to draw reasonably reliable conclusions as to the behavior of the carbon-free system. Moreover, the measurements serve to determine the beginning and the end of the alpha-gamma transformation; and they emphasize once more the fact that a transformation temperature determined by any method in which the specimen is cooled at an appreciable rate is apt to be seriously in error, particularly if the metal contains a transformation-retarding element such as chromium or managanese. The lag, or undercooling, observed at the ? — a transformation is explained in terms of the influence of temperature on transformation velocity; the rather unusual shape of the ? loop is attributed to a change with temperature and with composition of the distribution of chromium between the alpha-iron and gamma-iron solid solutions. The exact composition of the alloys, which were obtained through the courtesy of Dr. W. Rohn, of the Heraeus Vakuumschmelze A-G., is given in Table 1. Determinations of the linear thermal expansion and of transformation temperature were carried out in the vacuum dila-tometer previously described,' using a rate of heating and of cooling of 3.0" to 3.5" per minute unless otherwise stated. Magnetic susceptibility, which was measured in studying the rate of transformation, was determined with a magnetic balance, which also has been described.2 Results and Discussion Thermal Expansion Coeficients.—Numerical values for the linear expansion coefficient, derived from a smoothed curve based on at least

Jan 1, 1935

By A. T. English

X-ray dif action techniques were used to investigate ordering and domain coalescence for the B2 (CsC1) superlattice in this alloy. Essentially complete disorder is obtainable by drastic quenching. isothertmal ordering occurs very rapidly at temperatures between 565° and 680°C, only a few seconds being required for development of substantial long-range order. A very fine antiphase dormain structure, found in samples annealed for short times, undergoes coalescence with further annealing. Domain "size" is found proportional to (t)1/2 . This Yate law can be derived by adapting the formalism of Hillert's analysis 01- metallurgical grain growth. For this Purpose the material is descvibed in terms of a "Swiss-cheese " structure, consisting of two multiply connected, mutually intertwined domain volumes. The observed activation enevgy for coalescence, 70 kcal per g-atom, agrees with an estit~lated value ,for volume diffusion in disordered rrzaterial. It is suggesled that the experimental value vejlects the substantially disordered environment of mobile vacancies located in the moving antiphase domain boundary. The alloy Fe-Co-2V is bcc below about 900'C and orders into a B2 (CsC1) superlattice below about 730c It has important applications in magnetic components, including many requiring extensive cold rolling.* example, Marcinkowski and Brown6 observed a re1atively small as-quenched domain size of about 250Å in B2 Fe3A1. Since there appear to be no published data on domain coalescence in B2 alloys, this process was studied in detail. MATERIALS AND EXPERIMENTAL TECHNIQUES A commercial high-purity alloy, nominally 49 Fe-49Co-2V, was obtained from the Arnold Engineering Co. in the form of cold-rolled sheet 0.004 in. thick. Samples were annealed for 1 hr at 800°C in dry H2 to produce disorder and quenched into iced brine. They were then annealed in a nitrate-nitrite salt bath for various times at temperatures between 525° and 625°C. Estimates of the LRO parameter S and the APD size D were made using established X-ray diffraction techniques. Superlattice peaks are best observed using CoKa radiation, for which the value of (fFe-fCo) is maximized. Peak-to-background intensity ratio was improved by filtering CoKß from the incident beam to eliminate sample fluorescence, and by using a pulse-height analyzer. The LRO parameter S was computed from the ratio of integrated intensities I 100/I 200- Owing presumably to systematic errors associated with the presence of crystallographic texture and X-ray extinction, only rough agreement was obtained between experimental and theoretical (absolute) intensity ratios. For this reason certain experimental refinements (e.g., use of balanced filters) were considered superfluous and were omitted. Emphasis was placed on maximizing intensity in order to reduce the time taken to accurately determine the contour of the very small (loo), peak. This was done by point counting 105 counts at appropriate intervals. Background intensity averaged about 1200 cps.

Jan 1, 1967

By L. I. Cothern

The introduction of mechanized mining has created a demand for long working faces. It has also prompted mining men to contest the old theory that longwall methods can be used only where roof conditions are ideal and where pack-wall material is plentiful. "Longwall" as used in this paper means "modified longwall," "semi-longwall," or any of the various ''wall" or "long-face" methods of mining. Many mines are now using long conveyor faces that differ only in their methods of development, types of roof support and method of loading coal onto the conveyor. Some types of longwall operations are being carried on successfully under conditions heretofore considered impossible by men experienced in longwall mining. With possibly one or two exceptions, long-face methods in Alabama are a direct result of the need for lower mining costs. The advantages over room-and-pillar systems were clearly expressed in an article by Messrs. Fies and Lacy of the DeBardeleben Coal Corporation1 as follows: (1) increased tonnage per miner; (2) increased percentage of lump coal; (3) reduction in cost of development; (4) elimination of the cost of room yardage; (5) reduction in the cost of haulage, from ability to use larger cars; (6) reduction in the cost of machine cutting; (7) reduction in cost for tracks; (8) reduction in the cost of deadwork; (9) reduction in the cost of supervision. An attempt on the part of the operator to gain these advantages by changing from room-and-pillar to longwall requires considerable courage, and credit should be given to those men who have done their bit to meet the competition of rival fuels. The initiatory problems to be solved are in reference to: (1) retreating or advancing system, (2) length of wall, (3) method of roof support, (4) method of entry development, (5) type of conveyor or loader, (6) hand vs. machine loading (onto conveyors), (7) hand vs. machine loading in entries, (8) control of roof pressure,

Jan 1, 1940

By L. I. Cothern

The introduction of mechanized mining has created a demand for long working faces. It has also prompted mining men to contest the old theory that longwall methods can be used only where roof conditions are ideal and where pack-wall material is plentiful. "Longwall" as used in this paper means "modified longwall," "semi-longwall," or any of the various ''wall" or "long-face" methods of mining. Many mines are now using long conveyor faces that differ only in their methods of development, types of roof support and method of loading coal onto the conveyor. Some types of longwall operations are being carried on successfully under conditions heretofore considered impossible by men experienced in longwall mining. With possibly one or two exceptions, long-face methods in Alabama are a direct result of the need for lower mining costs. The advantages over room-and-pillar systems were clearly expressed in an article by Messrs. Fies and Lacy of the DeBardeleben Coal Corporation1 as follows: (1) increased tonnage per miner; (2) increased percentage of lump coal; (3) reduction in cost of development; (4) elimination of the cost of room yardage; (5) reduction in the cost of haulage, from ability to use larger cars; (6) reduction in the cost of machine cutting; (7) reduction in cost for tracks; (8) reduction in the cost of deadwork; (9) reduction in the cost of supervision. An attempt on the part of the operator to gain these advantages by changing from room-and-pillar to longwall requires considerable courage, and credit should be given to those men who have done their bit to meet the competition of rival fuels. The initiatory problems to be solved are in reference to: (1) retreating or advancing system, (2) length of wall, (3) method of roof support, (4) method of entry development, (5) type of conveyor or loader, (6) hand vs. machine loading (onto conveyors), (7) hand vs. machine loading in entries, (8) control of roof pressure,

Jan 1, 1940

By H. B. Pulsifer

One of the most common basic open-hearth furnace products is a simple carbon steel with a carbon range from 0.05 to 0.15 per cent. The material is widely used for sheets, tubes, bars, wire and the innumerable special objects of secondary fabrication. The properties of the material vary over wide ranges, depending on the more exact composition and details of manufacture or treatment that affect its structure. The present paper is a review of those properties and structures that have been found by studying this steel as used in the cap screw industry. At first glance this might be thought a narrow application of the steel, but, since the industry uses the material in the cold-worked, annealed and heat-treated conditions, the title of the paper may be excused for its comprehensiveness. Yet, in studying the wide ranges of properties and structures that preclude a more specific title, there can be no suggestion of completeness or finality. Even a summary of a few years' work is no more than a preliminary survey of a field that will be fertile in scientific data and industrial products for many years to come. An outline of the structures will be given first, then a list of the previous contributions that are at hand and finally some details of the work and related problems, especially concerning this low-carbon steel in the surface-hardened condition. Wire Used in Manufacture of Cap Screws The wire that is used for making cold-upset and roll-threaded cap screws has been cold-drawn from hot-rolled rods that were pickled and limed. The common sizes used in the industry range from 0.185 to 0.750 in. dia., the exact diameters depending on the plant practice and having the usual tolerances of 0.002 in. for the smaller sizes and 0.003 in. for the larger sizes. The carbon range, by pit analysis, is usually between 0.07 and 0.12 per cent. Manganese should be within the usual 0.30 to 0.50 per cent. range. Sulfur rarely exceeds 0.045 per cent. and phosphorus seldom exceeds 0.030 per cent. Silicon is usually below 0.03 per cent. but in the silicon-killed steels may be in the customary 0.10 to 0.25 per cent. range. Rimmed steel is not considered most suitable for cap screws that are cold-headed or for parts that are to be surface-hardened.

Jan 1, 1931

By Benjamin Lustman, Robert F. Mehl

The study of the high-temperature oxidation of pure metals, intensively pursued experimentally since the pioneer work of Pilling and Bedworth1 and supplemented by the recent theoretical work of Wagner2 is now well advanced. The important field of low-temperature oxidation—i.e., oxidation extending only through the temper color range of thicknesses with the formation of thin oxide layers in the range from 0 to 1000-2000 ?ngstroms—has not, however, reached so satisfactory a state, owing largely to the experimental difficulties that attend the formation and study of thin films; most of the work has been performed on abraded specimens whose surface irregularities approximated the thickness of the films formed. It has been pointed out that the rates of oxidation of different crystal faces are different and that the difierence may be related to the orientation relationships that obtain between the metal crystal and the superimposed oxide crystal; it has also been suggested that this orientation dependence should manifest itself in rates of oxidation and perhaps also of corrosion in aggregates bearing preferred orientations.3'4 An attempt to measure the rates of oxidation of different crystal faces of alpha iron demonstrated this difference in rate,4 but the dependence of rate on orientation was not simple; the method of measuring thickness, consisting of a color comparison with barium stearate multi-films of known thickness, was not wholly satisfactory. For these reasons it was felt that a study of the rate of oxidation of single crystals prepared with surfaces as smooth as possible, using a sensitive method giving true film thickness, should be profitable. Copper was chosen because of the ease of reduction of the primary oxide film and because the orientation relationships between metal and oxide and the composition of the oxide films formed at low temperatures are known; for thin films formed at low temperatures, Cu2O alone occurs.22-24 Experimental Procedure Various methods of surface preparation were attempted in order to obtain a surface that would oxidize uniformly and repro-ducibly. Electrolytic polishing by the method of Jacquet5 and electrolytic and mechanical polishing followed by annealing in vacuum and in purified hydrogen at various temperatures were tested. The procedure finally adopted consisted in a fine metallographic polish followed by an anneal in purified hydrogen at temperatures of 900° to 1000°C. for 12 hr. or more. The surface so produced was very bright and smooth, with no trace of specular etching (though some grain-boundary etching was observed); indeed, fine scratches produced by the final polish were entirely removed by the high-temperature anneal. After this treatment, the specimens were cooled in hydrogen to room temperature and placed in the oxidation furnace, where

Jan 1, 1941

By Benjamin Lustman, Robert F. Mehl

The study of the high-temperature oxidation of pure metals, intensively pursued experimentally since the pioneer work of Pilling and Bedworth1 and supplemented by the recent theoretical work of Wagner2 is now well advanced. The important field of low-temperature oxidation—i.e., oxidation extending only through the temper color range of thicknesses with the formation of thin oxide layers in the range from 0 to 1000-2000 ?ngstroms—has not, however, reached so satisfactory a state, owing largely to the experimental difficulties that attend the formation and study of thin films; most of the work has been performed on abraded specimens whose surface irregularities approximated the thickness of the films formed. It has been pointed out that the rates of oxidation of different crystal faces are different and that the difierence may be related to the orientation relationships that obtain between the metal crystal and the superimposed oxide crystal; it has also been suggested that this orientation dependence should manifest itself in rates of oxidation and perhaps also of corrosion in aggregates bearing preferred orientations.3'4 An attempt to measure the rates of oxidation of different crystal faces of alpha iron demonstrated this difference in rate,4 but the dependence of rate on orientation was not simple; the method of measuring thickness, consisting of a color comparison with barium stearate multi-films of known thickness, was not wholly satisfactory. For these reasons it was felt that a study of the rate of oxidation of single crystals prepared with surfaces as smooth as possible, using a sensitive method giving true film thickness, should be profitable. Copper was chosen because of the ease of reduction of the primary oxide film and because the orientation relationships between metal and oxide and the composition of the oxide films formed at low temperatures are known; for thin films formed at low temperatures, Cu2O alone occurs.22-24 Experimental Procedure Various methods of surface preparation were attempted in order to obtain a surface that would oxidize uniformly and repro-ducibly. Electrolytic polishing by the method of Jacquet5 and electrolytic and mechanical polishing followed by annealing in vacuum and in purified hydrogen at various temperatures were tested. The procedure finally adopted consisted in a fine metallographic polish followed by an anneal in purified hydrogen at temperatures of 900° to 1000°C. for 12 hr. or more. The surface so produced was very bright and smooth, with no trace of specular etching (though some grain-boundary etching was observed); indeed, fine scratches produced by the final polish were entirely removed by the high-temperature anneal. After this treatment, the specimens were cooled in hydrogen to room temperature and placed in the oxidation furnace, where

Jan 1, 1941

By L.L. Wyman

The exact nature of the changes that take place in the iron-nickel alloys, giving rise to the interesting and useful expansion alloys in the Invar range, has yet to be fully understood. Similarly, the ternary iron-nickel-cobalt alloys, which also possess such commercially useful expansion characteristics,1,2 are concerned with transformations that appear to be identical with those found in the parent iron-nickel alloys. The investigations on iron-nickel alloys in the range 26 to 29 per cent Ni have been reported by several authors3-9 wherein it is shown that there are factors other than temperature, such as stress and grain size, which play an important part in determining the temperature at which the low-temperature transformation takes place. It will be shown in this paper that the ternary iron-nickcl-cobalt alloys having nickel contents comparable to the alloys mentioned above behave in a parallel manner. The alloys in the range of 26 to 31 per cent Ni and 16 to 21 per cent Co have a coefficient of expansion of about 4.0 X 10-6 per degree Centigrade, and the alloys in this range of composition find considerable application in making glass-to-metal seals, because these expansion properties so well match commercial glasses. The behavior of these alloys is not a simple matter, for it has been observed in alloys of 28.5 per cent Ni and 18.5 Co that severe working ran cause the appearance of a second phase having a needlelike structure similar to the martensite structure in carhon steels. Fig. 1 shows the change brought about in an alloy of this composition when a strip of this material was cut with tin shears. This second phase has been identified as the body-centered cubic alpha phase, and it is the purpose of this paper to describe the results of some of the preliminary investigations in the study of the changes involved in these alloys. Materials The past work on these alloys has concerned itself primarily with materials made by the usual casting procedure and involving the use

Jan 1, 1939

By L. L. Wyman

The exact nature of the changes that take place in the iron-nickel alloys, giving rise to the interesting and useful expansion alloys in the Invar range, has yet to be fully understood. Similarly, the ternary iron-nickel-cobalt alloys, which also possess such commercially useful expansion characteristics,1,2 are concerned with transformations that appear to be identical with those found in the parent iron-nickel alloys. The investigations on iron-nickel alloys in the range 26 to 29 per cent Ni have been reported by several authors3-9 wherein it is shown that there are factors other than temperature, such as stress and grain size, which play an important part in determining the temperature at which the low-temperature transformation takes place. It will be shown in this paper that the ternary iron-nickcl-cobalt alloys having nickel contents comparable to the alloys mentioned above behave in a parallel manner. The alloys in the range of 26 to 31 per cent Ni and 16 to 21 per cent Co have a coefficient of expansion of about 4.0 X 10-6 per degree Centigrade, and the alloys in this range of composition find considerable application in making glass-to-metal seals, because these expansion properties so well match commercial glasses. The behavior of these alloys is not a simple matter, for it has been observed in alloys of 28.5 per cent Ni and 18.5 Co that severe working ran cause the appearance of a second phase having a needlelike structure similar to the martensite structure in carhon steels. Fig. 1 shows the change brought about in an alloy of this composition when a strip of this material was cut with tin shears. This second phase has been identified as the body-centered cubic alpha phase, and it is the purpose of this paper to describe the results of some of the preliminary investigations in the study of the changes involved in these alloys. Materials The past work on these alloys has concerned itself primarily with materials made by the usual casting procedure and involving the use

Jan 1, 1939

By N. A. Zeigler, T. D. Yensen

Aging is a term that connotes a slow change in properties under ordinary operating conditions. It can be accelerated by increasing the temperature and by mechanical straining. The magnetic properties of iron and some of its alloys have for many years been known to be subject to such a change. But the cause of the change was not recognized until a few years ago after mechanical age-hardening in non-ferrous alloys of the "duralumin" type had been definitely shown, by means of solid solubility curves, changes in electrical resistance and microanalysis, to be caused by slow precipitation of one or more constituents from the solid solution. The similarity between the solid solubility curves of such alloys (Fig. 1) and those for iron-nitrogen (Fig. 2) and iron-carbon alloys (Fig. 3) below the A1 transformation point suggested at once that magnetic aging of iron might be caused by a similar process; namely, by the slow precipitation of carbon and nitrogen in the form of Fe4N and Fe3C at ordinary temperatures because of a cooling rate that had been too rapid to permit complete precipitation according to equilibrium conditions. This suggestion has since been amply confirmed, particularly by the comprehensive work of Köster1 on the effect of nitrogen and carbon. Typical curves are shown in Fig. 4 for the relationship between aging temperature and coercive force. Previous to the establishment of the precipitation-hardening theory it was thought that the effect of nonmetallic elements like C, N, O, etc., on the magnetic properties of iron was greatest when these elements were retained in solid solution, atomically dispersed throughout the interstitial spaces of the iron lattice2. Based on the theory that magnetic properties are intimately connected with the regularity of the lattice, it now follows quite logically that a greater effect on this regularity may be obtained when the above impurities are precipitated as very fine particles of Fe3C, Fe4N, FeO, etc., more or less uniformly dispersed within the grains. But if this is the case, it follows also that there must exist a relationship between the size

Jan 1, 1935

By N. A. Zeigler, T. D. Yensen

Aging is a term that connotes a slow change in properties under ordinary operating conditions. It can be accelerated by increasing the temperature and by mechanical straining. The magnetic properties of iron and some of its alloys have for many years been known to be subject to such a change. But the cause of the change was not recognized until a few years ago after mechanical age-hardening in non-ferrous alloys of the "duralumin" type had been definitely shown, by means of solid solubility curves, changes in electrical resistance and microanalysis, to be caused by slow precipitation of one or more constituents from the solid solution. The similarity between the solid solubility curves of such alloys (Fig. 1) and those for iron-nitrogen (Fig. 2) and iron-carbon alloys (Fig. 3) below the A1 transformation point suggested at once that magnetic aging of iron might be caused by a similar process; namely, by the slow precipitation of carbon and nitrogen in the form of Fe4N and Fe3C at ordinary temperatures because of a cooling rate that had been too rapid to permit complete precipitation according to equilibrium conditions. This suggestion has since been amply confirmed, particularly by the comprehensive work of Köster1 on the effect of nitrogen and carbon. Typical curves are shown in Fig. 4 for the relationship between aging temperature and coercive force. Previous to the establishment of the precipitation-hardening theory it was thought that the effect of nonmetallic elements like C, N, O, etc., on the magnetic properties of iron was greatest when these elements were retained in solid solution, atomically dispersed throughout the interstitial spaces of the iron lattice2. Based on the theory that magnetic properties are intimately connected with the regularity of the lattice, it now follows quite logically that a greater effect on this regularity may be obtained when the above impurities are precipitated as very fine particles of Fe3C, Fe4N, FeO, etc., more or less uniformly dispersed within the grains. But if this is the case, it follows also that there must exist a relationship between the size

Jan 1, 1935

By John Wulff, R. Buehl, H. Hollomon

Although the main features of the transformations occurring in 18-8 have been published already, certain conclusions merit questioning and discussion. The questions may be summarized as follows : 1. If ferrite is formed by cold-working austenite, what happens to the carbon with which the ferrite and austenitc are both presumably supersaturated? 2. What is the nature of the alpha-ferrite formed at grain boundarices by chromium-carbide precipitation at these boundaries?

Jan 1, 1940

By R. Buehl, John Wulff, H. Hollomon

Although the main features of the transformations occurring in 18-8 have been published already, certain conclusions merit questioning and discussion. The questions may be summarized as follows : 1. If ferrite is formed by cold-working austenite, what happens to the carbon with which the ferrite and austenitc are both presumably supersaturated? 2. What is the nature of the alpha-ferrite formed at grain boundarices by chromium-carbide precipitation at these boundaries?

Jan 1, 1940