Search Documents

By G. H. F. Gardner

The velocity and attenuation of elastic waves in sandstones were measured as a function of both pressure and fluid saturation. A large change occurs in these quantities if water is added and the rock is not compressed, but the change is small if the rock is subjected to a large overburden pressure. Measurements were made by vibrating cylindrical samples in both the extensional and torsional modes at frequencies up to 30,000 cycles/sec. Formulas were derived which enable the attenuation of dilatational waves in dry rocks to be deduced from the data. Similar experimental methods were used to investigate the properties of unconsolidated sands. Velocities were found to vary with the 1/4 power of the overburden pressure and attenuations to decrease with the 1/6 power. The effects of grain size, amplitude and fluid saturation were studied. Formulas by which the effects produced by a jacket around the sample may be calculated were derived. The practical application of these results to formation valuation is discussed. INTRODUCTION The attenuation of elastic waves in the earth has been of interest to the seismologist and geophysicist for many years, but only recently to the petroleum engineer. Engineering interest has been brought about by the success of velocity logging devices, for it is possible by modification of these instruments to measure the attenuation of sound waves in addition to their velocity and, hence, deduce the mobility of formation fluids as well as the porosities of the rocks which contain them. The main problem is to decide whether field measurements can be made with sufficient accuracy to be of practical use. This problem can only be solved after we know the magnitude of the attenuations which are typical of the earth at various depths. The logarithmic decrement of a fluid-saturated rock is the sum of a "sloshing" decrement and a "jostling" decrement, the former caused by the mobility of the fluid contained within the rock and the latter by the granular framework of the rock. Sloshing decrements can be calculated' using Biot's theory, but the jostling losses are less well understood. The present paper reports an experimental investigation of jostling losses in consolidated and uncon- solidated sands, particularly with respect to the effect of overburden pressure and fluid saturation. Born' showed that the decrement of a sandstone may increase dramatically when only a few per cent by weight of distilled water is added, and that the additional loss is proportional to the frequency of vibration. His measurements were made with no compressive stress on the framework of the rock. M. Gondouin3 investigated similar phenomena for fluid-saturated plasters but also did not compress the samples. In the present paper it is shown that compression of the framework reduces this effect, so that at depth the jostling decrement of a sandstone may be expected to be almost independent of fluid saturation and frequency. Decrements for many sedimentary rocks have been given by Volarovich,4 but all for the state of zero overburden pressure. Anomalously low velocities have been logged in shallow unconsolidated gas sands. Results of the present investigation confirm that these velocities are not caused by correspondingly high attenuations, because the jostling decrement in a packing of sand grains is small and much less than in a consolidated sandstone at the same depth. Velocities in sands have been measured by Tsareva5 and by Hardin6 as a function of pressure, but the corresponding decrements do not appear to have been measured previously. The widely used "resonant bar method" of measuring velocities and decrements was employed. Comments on variations of this technique have recently been published by McSkimmin.7 The main novelty of the present technique was the application of pressure to the samples. It was found possible to do this by placing the apparatus inside a pressure vessel, provided the conditions leading to large additional losses were avoided. These conditions are discussed below. EXPERIMENTAL TECHNIQUE Cylindrical samples were caused to vibrate in both the extensional and torsional mode of vibration and the amplitude of vibration was measured as a function of frequency in the neighborhood of a resonant frequency. The resonant frequency, fr, is related to the corresponding elastic modulus by the formulas where E and N are Young's modulus and the modulus of rigidity, p is the density of the sample, and A the wavelength of the vibration.

Jan 1, 1965

By D. W. Rudd, D. W. Vose, S. Johnson

The permeability of Mo-0.5 pel Ti to hydrogen was investigated over a limited range of temperature and pressuire (709° to 1100°C, 1.i and 2.0 atm). The resulting permeability, p, is found to obey the The experimental data justifies the permeation mechanism as a diffusion contl-olled pnssage of Ilvdrogen atoms through the metal barrier. 1 HE permeability of metals to hydrogen has been investigated by a number of workers and their published results have been tabulated by Barrer' up to 1951. Since most of the work on the permeability has been accomplished prior to this date, the compilation is fairly complete. Mathematical discussion of the permeability process has been reported by Barrer, smithells, and more recently by zener. From these efforts several facts are observed. First, the permeability of metals to diatomic gases involves the passage through the metal of individual atoms of the permeating gas. This is evidenced by the fact that the rate of permeation is directly proportional to the square root of the gas pressure. Second, the gas permeates the lattice of the metal and not along grain boundaries. It was shown by Smithells and Ransley that the rate of permeation through single-crystal iron was the same after the iron had been recrystallized into several smaller crystals. Third, it has been observed that the rate of permeation is inversely proportional to the thickness of the metal membrane. Johnson and Larose5 verified these phenomena by measurirlg the permeation of oxygen through silver foils of various thicknesses. Similar findings were noted by Lombard6 for the system H-Ni and by Lewkonja and Baukloh7 for H-Fe. Finally, it has been determined that for a gas to permeate a metal, activated adsorption of the gas on the metal must take place. Rare gases are not adsorbed by metals, and attempts to measure permeabilities of these gases have proved futile. ~~der' found negative results on the permeability of iron to argon. Also, Baukloh and Kayser found nickel impervious to helium, neon, argon, and krypton. From what was stated above concerning the dependence of the rate on the reciprocal thickness of the metal barrier, it is seen that although adsorption is a very important process, at least in determining whether permeation will or will not ensue, it is not the rate determining process for the common metals. A case in which adsorption is of sufficient inlportance to cause abnormal behavior has been noted in the case of Inconel-hydrogen and various stainless steels.'' APPARATUS The apparatus used in this study is shown in Fig. 1. The membrane is a thin disc (A), but is an integral part of an entire membrane assembly. The entire unit is one piece, being machined from a solid ingot of metal stock. When finished, the membrane assembly is about 5 in. long. Two membrane assemblies were made; the dimensions of the membranes are given in Table I. The wall thickness is large compared to the thickness of the membrane, being on the average in the ratio of 13 to 1. There exists in this design the possibility that some gas may diffuse around the corner section of the membrane where it joins the walls of the membrane assembly, If such an effect is present, it is of a small order of magnitude, as evidenced by the agreement of the values of permeability between the two membranes under the same temperature and pressure. A thermocouple well (B) is drilled to the vicinity of the membrane. The entire membrane assembly is then encased in an Inconel jacket and mounted in a resistance furnace. The interior of the jacket is connected to an auxiliary vacuum pump and is always kept evacuated so that the membrane assembly will suffer no oxidation at the temperatures at which measurements are taken. The advantages of this configuration are: 1) there are no welds about the membrane itself, so that the chance of welding material diffusing into the membrane at elevated temperatures is remote. 2) It is possible to maintain the membrane at a constant temperature. Since the resulting permeation rate is very dependent upon temperature, it is advisable to be as free as possible from all temperature gradients. 3) It is possible to obtain reproducible results using different specimens. The only disadvantage to this configuration is the welds (at C) in the hot zone. The welding of molybdenum to the degree of per-

Jan 1, 1962

By J. J. Gilman

The data presented indicate that the critical shear stress and strain-hardening Thedatapresentedrate of a zinc monocrystal depend on the orientation of its slip direction with respect to its external boundaries. The tendency of a crystal to form deformation bands also depends on its shape. THE plastic behavior of pairs of zinc monocrystals in which both members of the respective pairs had the same orientation with respect to the longitudinal axis, but each had different orientations with respect to their rectangular external shapes, were compared in this investigation. The purpose of the investigation was to see what influence the shape or surface of a zinc crystal has on its mechanical properties. In a previous investigation of triangular zinc monocrystals,1 anomalous axial twisting was observed which seemed to be related to the triangular shape of the crystals. Wolff,' in 400°C tensile tests of rectangular rock-salt crystals bounded by cubic cleavage planes, found that, of the four equivalent slip systems, the two with the "shorter" slip directions yielded and produced slip lines at lower stresses than the other two. This observation and the work of Dommerich³ as formulated by Smekal4 as a "new slip condition" for rock-salt: "among two or more slip systems permitted by the shear stress law, with reference to the formation of visible slip lines by large individual glides, that slip system is preferred which has the shortest effective slip direction." More recently, Wu and Smoluchowski5 reported essentially the same effect for ribbon-like (20x2x0.2 mm) aluminum crystals at room temperature. Experimental Chemically pure zinc (99.999 pct Zn), purchased from the New Jersey Zinc Co., was the raw material. Glass envelopes, containing graphite molds and zinc, were evacuated while hot enough to outgas the graphite but not melt the zinc. At a vacuum of about 0.2 micron the envelopes were sealed off and then lowered through a furnace at 1 in. per hr so as to melt and resolidify the zinc and produce mono-crystals. One-half of one of the molds is shown in Fig. la. Each mold consisted of four pieces from a cylindrical graphite rod that was split longitudinally and transversely at its midpoints. Rectangular milled grooves 0.050 in. deep and % in. wide formed the mold cavity when the split halves were assembled with twisted wires. Fig. lb shows the specimen shape obtained when the top and bottom mold-halves were rotated 90" with respect to each other. Good fits prevented leakage and excess zinc was necessary to provide enough liquid head to fill the mold completely. In removing soft crystals from the molds it was impossible to avoid small amounts of bending. However, manipulations were carried out whenever possible with the crystals protected by grooved brass blocks. All specimens were annealed prior to testing. From the top and bottom sections of each crystal, X-ray specimens and tensile specimens 7 to 8 cm long were sawed. The tensile specimens were annealed inside evacuated tubes for 1 hr at 375°C. Next the crystals were cleaned and polished by 2-min dips in a solution of 22 pct chromic acid, 74 pct water, 2.5 pct sulphuric acid, and 1.5 pct glacial acetic acid.' Cleaning was followed by a 10-sec dip in a 10 pct caustic solution, then washed in water and alcohol, and dried. This treatment results in a bright surface covered by an invisible oxide film. The testing grips were a slotted type with set screws and were supported in a V-block during the mounting operations in order to avoid bending the crystals. A schematic diagram of the recording tensile-testing machine is shown in Fig. 2. The machine has been described elsewhere.' The head speed was 0.3 mm per sec for all tests. The crystal orientations were determined by the Greninger X-ray back-reflection method with an estimated accuracy of 1. Description of Crystal Geometry A schematic picture of a rectangular zinc mono-crystal is shown in Fig. 3. ABD designates the front edge of a basal plane (0001) of the crystal, the only active slip plane for zinc at room temperature. Of the three possible (2110) slip directions, the active one is indicated by an arrow. Cartesian coordinates are taken parallel to the specimen edges. The normal, n, to the basal plane (n is parallel to the hexagonal axis) has the direction cosines a, ß and ?. X0 = 90 — y is the angle between the longitudinal axis and

Jan 1, 1954

By James R. Holden, Frederick E. Wang

Through both single-crystal and powder X-ray diffraction methods, ten A6B23-type compounds have been confirmed to exist between lanthanides (A) (plus scandium and yttrium) and manganese (B); A = Y, Nd, Sm, Gd, Tb, Dy, Ho, Er, Tm, and Lu. The formation of a compound of this type is shown to he extremely atomic size-sensitive; hence it can be classified as a "size-factorH compound. The "enveloping effect", a geometrical consideration observed in its crystal structure, is proposed as the reason for the A6B23-type compound being size-sensitive. The approximate ideal geometrical ratio of the radii R/r is 1.31 while experimentally A6B23-type compounds have a radius ratio lying within the range 1.2 to 1.4. FLORIO t al.' characterized the structure of Th6MnZ3 as fcc, space group Fm3m, with 116 atoms in the unit cell. Since then, a number of isotypic binary compounds, and recently Gd n,,' have been confirmed to exist. The fact that strontium and barium form A6Bz3-type compounds with magnesium strongly suggested the possible existence of Ba6Liz3. However, investigation3 showed the compound Ba6LiZ3 to be absent. Since both strontium and barium are group 11-a elements and are therefore "open metals",6 the nonexistence of Ba6LiZ3 can hardly be explained satisfactorily by valence-electron considerations. On the other hand, the consistent atomic-radius ratio, (R/r),* observed for the known A6Bz3-type compounds,3 strongly suggests that the formation of compounds of this type is atomic size-sensitive. Therefore, one is tempted to explain the nonexistence of Ba6LiZ3 entirely on the basis of the atomic-size difference between strontium and barium. However, this approach is not entirely without objection. Atoms are not rigid spheres and are known to vary in size within certain limits.7 Since the atomic-radius difference between strontium and barium (0.07 to 0.09A) is within these limits, it is reasonable to assume that the size difference would have a negligible effect on the formation of Ba6LiZ3. This view is further supported by the fact that the radius ratio, R/r, in other known "size-factor" compounds is observed to range widely—for example, from 1.08 to 1.45 for ABz-type compounds (C15, MgCuz type)' and from 1.37 to 1.58 for AB5-type compounds (D2d, CaZn5 type).g The present investigation was undertaken in order to find a more satisfactory explanation for the non-existence of Ba6LiZ3 and, consequently, a better understanding of the nature of the A6Bz3-type compound. The primary objectives are to confirm the previous conclusion3 that the A6B23-type compound is indeed a "size-factor" compound and subsequently to determine the atomic-radius ratio range in which the A6Bz3-type compound can exist. In order to achieve these objectives, stoichiometric A6Bz3 alloys, where manganese (B) was alloyed with various lanthanide elements (A), were selected for investigation. The atomic-radius ratios of lanthanide elements with manganese range from 1.26 for Lu/~n to 1.46 for Eu/Mn. This radius ratio range includes and exceeds the range of all previously reported A6Bz3-type compounds—1.32 for Th/Mn' through 1.38 for Sr/Li. Furthermore, the atomic-size difference between successive elements of the lanthanide series in order of atomic number) is of the order of 0.01A (europium and ytterbium are exceptions). The series of lanthanon-manganese alloy systems is ideally suited to a precise determination of the limits of allowable atomic-radius ratio for A6Bz3-type compound formation. EXPERIMENTAL PROCEDURE The lanthanide metals, in ingot form, supplied by Michigan Chemical Corp. (St. Louis, Mich.) and Nuclear Corp. of America (Burbank, Calif.), were guaranteed by the suppliers to be at least 99.9 wt pct pure (traces of silicon, calcium, and other minor constituents present on occasion, not to be more than 0.05 wt pct) as shown by spectrographic analysis. Manganese metal, in polycrystalline form, was redistilled from the commercial, chemically pure grade and was analyzed to be at least 99.95 wt pct pure. In all cases, the atom ratio between the two elements in each charge was A (rare-earth meta1):B (manganese) = 6:23 and a constant weight, 3 g, of

Jan 1, 1965

By S. L. Craig, C. Orr, H. G. Blocker

WITHIN recent years gas adsorption methods have been developed for measuring the surface area of finely divided materials and have become extremely valuable in research on the corrosion and the catalytic activity of metals. Rather elaborate apparatus is required, and a single determination is so time-consuming that these methods have not been utilized to the fullest extent; the methods are un-suited for most routine control work such as that encountered in powder metallurgical operations and in processes employing metal catalysts. These difficulties are largely eliminated, and surface area is reduced to a routine determination if the liquid-phase adsorption of a surface-active agent such as a fatty acid can be used. When the affinity of the fatty acid carboxyl group for the solid surface is greater than its affinity for the solvent, a unimolec-ular layer of orientated fatty acid molecules will be formed at the solid-liquid interface in a manner similar to that of a compressed fatty acid film on a water surface. The measurement of surface area is then reduced to a measurement of fatty acid adsorption. This propitious circumstance, first investigated by Harkins and Gans,¹ has been employed with somewhat inconclusive results by a number of investigators in evaluating the surface properties of metals, metal catalysts, and metal oxides. The specific surface area values for nickel and platinum catalysts, determined from the adsorption of a number of fatty acids from various solvents, were found by Smith and Fuzek² to agree with values calculated by the gas adsorption technique of Brunauer, Emmett, and Teller," he so-called BET technique. And recently Orr and Bankston4 have also reported good agreement between nitrogen gas and stearic acid adsorption results in the measurement of the surface areas of clay materials. On the other hand, Ries, Johnson, and Melik5 found only order-of-magnitude agreement between these two methods in studying supported, cobalt catalysts having specific surface areas as great as 420 sq m per g; the reason is partially attributable to the very porous nature of the materials. Greenhill,6 investigating the adsorption of long-chain, polar compounds in organic solvents on a number of metal powders, concluded that a uni-molecular layer of stearic acid was formed on exposure of the solid to the acid solution and that the presence of an oxide or another film did not alter this result. Furthermore, the adsorption process appeared to be the same whether or not the sample was degassed prior to exposure to the solution. Greenhill estimated the surface area of one of the powders he investigated from microscopic diameter measurements, and obtained a rough check with surface area evaluation. Russell and Cochran7 found moderate agreement for alumina surface area results by fatty acid and gas adsorption methods. In addition, they also found that the prolonged heating and evacuating pretreatments previously used by investigators were unnecessary. The present work, however, considerably extends these previous investigations, shows that fatty acid adsorption can be used to determine the surface area of a variety of metals and metal compounds, offers further confirmation of the correctness of gas adsorption methods, and presents a simplified technique for the determination of the metal surface area which is suitable for routine work. Experimental Technique Basically, the fatty acid adsorption method is quite simple. It consists of exposing a sample of the material of which the surface area is desired to a fatty acid solution of known concentration. By analysis of an aliquot of the solution, the concentration after adsorption has occurred may be determined. The difference between the initial quantity of acid in solution and the final quantity is that quantity of acid adsorbed by the sample. The specific surface area of the adsorbent material may be calculated from the quantity adsorbed and the weight of the sample. In agreement with the findings of others as outlined above, it was found entirely unnecessary to degas or pretreat the nonporous materials employed other than by drying them thoroughly. However, precaution was necessary so that the dried sample entered the fatty acid solution with little exposure to moisture. The effect of moisture on the interaction of stearic acid with finely divided materials has been thoroughly investigated by Hirst and Lancaster." They found the presence of water merely reduced the amount of acid adsorbed by powders such as TiO2, SiO2, Tic, and Sic. With reactive materials such as Cu, Cu2O, CuO, Zn, and ZnO, however, water was found to initiate chemical reaction. Only with ZnO was reaction observed when the solid and the solu-

Jan 1, 1953

By J. B. Wagner, p. Hembree

The iron tracer diffusion coefficient of umstite has been measured at 110(fC across the phase field and at a single composition at 800°C. Assuming a simple cation vacancy model the tracer diffusion coefficient was found to be a linear function of the cation vacancy concentration at 1100°C. The equation is D = 3 x 20 29 where denotes the concentration of vacancies in numbers per cc. The tracer work at 800°C was carried out to investigate the reported "pinning" of tracer to the wustite surface at low temperatures. No evidence for the "pinning" of the tracer was found at 800°C in COz-CO gas mixtures. HIMMEL, Mehl, and Birchenall,&apos; Carter and Richardson,2 and Desmarescaux and La combe3 have measured the diffusion of iron tracer in wustite at several temperatures and compositions. The present work was undertaken to extend the measurements over a large composition range at 1100°C and to resolve certain apparent discrepancies in the data, expecially at lower temperatures. EXPERIMENTAL Wustite was prepared by oxidizing rectangular iron plates* in C02-CO mixtures. The samples were •The iron was supplied by the Battelle Memorial Institute courtesy of the American Iron and Steel Institute. The analysis is presented in Table I. quenched. Due to the inward flow of cation vacancies during oxidation, the center of the sample contained a thin void. The edges of the wustite slab were sanded until the sample could be split into two parts. Each part was then sanded on the front and back flat area until a smooth surface was obtained. The specimens were then replaced in the furnace and equilibrated at llOO°C in a predetermined COa-CO mixture by methods described elsewhere.4"6 The specimens were again quenched and the surfaces were lightly sanded to remove any roughness following the first equilibration. The specimens were then reequi lib rated in the same C02-CO mixture for thirty minutes in order to relieve any mechanical damage on the surface due to the polishing. The specimens were then quenched and the tracer was applied by an electroplating technique. The work of Carter and ~ichardson&apos; demonstrated that there was no systematic difference in the iron tracer diffusion coefficient in wustite if the tracer was plated, dried, or evaporated on the specimen. In the present study a piece of filter paper was saturated with an iron chloride solution of pH <* 3 that contained the tracer FeS5. The wustite was placed on the filter paper and made the cathode. A current density of 0.4 to 0.6 ma per sq cm was passed for about five to ten minutes. The thickness of the tracer layer was estimated to be about 7 x lom6 cm. This estimate was made by considering the area plated, the current flow, and time for plating and the activity of the iron in the plating solution. Different areas of the specimen were counted using a collimator to determine the uniformity of the tracer. Any specimen which exhibited a variation from the initial count rate (about 1500 cpm) by more than 15 pct was rejected. An estimate of the time necessary to convert the thin layer of iron tracer to wustite was made using the data of Pettit and wagner." he estimated time was 1 sec at 1100°C assuming linear oxidation kinetics. The shortest diffusion anneals were 1800 sec. The samples were suspended in the hot zone of a furnace by two platinum wires. Two separate specimens were run at the same time. Only the edges of each sample were in contact with the wires. The C02-CO gas of the same composition as that used in the pre-diffusion anneals flowed freely around the samples at a linear velocity of 0.9 cm per sec. To initiate a run, the specimens were lowered from the cold zone of a furnace to the hot zone by a magnetic lowering device." bout 60 sec were required for lowering. To terminate a run, the sample was withdrawn from the hot zone to the cold zone. Time zero for the beginning of the experiment was taken when the sample blended into the red glow of the furnace and conversely for the end of the experiment. The surface decrease method of measuring the tracer diffusion coefficient was used to collect the data. This method requires that counting geometry be reproducible because the specimen is counted before the diffusion anneal and after the anneal. A special jig was constructed for each specimen so the specimen could be removed from the jig and returned to the jig such that the well geometry was reproducible.

Jan 1, 1970

By G. M. Willis, F. L. Hennessy

The oxygen pressure in equilibrium with silver and Ag2O-B2O3 melts has been measured between 800&apos; and 900°C, to obtain the thermodynamic properties of the liquid. The compound Ag20. 4B20:1 appears to exist in the liquid, which shows marked heat content and entropy effects. A KNOWLEDGE of the thermodynamic properties of binary liquid silicates, borates, and phosphates would be of considerable assistance in the interpretation of the behavior of multi-component metallurgical slags. However, the literature contains comparatively few studies of the thermodynamics of binary slags. The system Ag20-B,O, attracted our attention as it was known to give a single liquid phase,&apos;,&apos; in which high contents of silver could be obtained (up to 61 pct Ag according to Foex2). Further, it would be expected that the partial pressure of oxygen over melts in equilibrium with metallic silver could be used to determine the activity of Ag2O in the Ag,O-B,O, system. In many respects, it may be expected that the reaction of a basic oxide with boric oxide would be analogous to its reaction with silica. Liquid immiscibility frequently occurs in both borate and silicate systems. With B2O3 and SiO reaction with a basic oxide presumably involves a breakdown of the three-dimensional network of the acid oxide by reaction with oxygen atoms common to more than one silicon or boron atom. Ag2O-B2O3 was therefore investigated as a model of a slag system in the hope that its thermodynamic properties would assist in understanding those of other systems. Several methods for determining the activity of a component in a slag have been described in the literature. Chang and Derge" used high temperature electromotive force measurements to obtain the activity of SiO2 in CaO-SiO2 and Ca0-Al203-Si02 slags, but the cell reaction in their work is not clear. low has used rate of volatilization and vapor pressure measurements combined with phase diagrams to obtain activities in the systems KO-SiO,, Na,O-SiO, and Li,O-SiO," and PbO-SiO26 Taylor and Chipman7 extrapolated their results for the distribution of FeO between liquid iron and CaO (+Mg0)-FeO-SiOl slags to obtain the activity of FeO in the binary FeO-SiO2 system. In principle, one of the most direct methods for obtaining the activity of a metallic oxide in a phase is by comparison of the equilibrium oxygen pressure for the system metal-pure oxide with that of metal oxide-containing phase. Schenck and othersa have studied the stabilization of Ag2O on combination with other oxides (MO,) in the solid state by measurements of the oxygen pressure in systems of the type Ag-Ag,O-xM0,-MOy-0, (gas). Schuhmann and Ensio" have determined the activity of FeO in iron silicate slags in equilibrium with solid iron, using CO/CO2 mixtures to establish known partial pressure of oxygen. Although the method gives the activity of FeO without ambiguity, the slag is not a binary system, and interpretation of the results in terms of the hypothetical binary system FeO-SiO, is not possible. If a metal is solid at temperatures at which the properties of the slag containing its oxide are to be studied, this method has the considerable experimental advantage that the metal can be used as the container for the slag, and contamination by contact with refractories is avoided. In this work, crucibles for Ag2-B,O, melts were made from silver, and the liquid brought to equilibrium with definite pressures of oxygen gas. The oxygen pressure PO, thus fixes the activity of Ag20 in the liquid silver borate. For the reaction at a given temperature. is substantially constant, is directly proportional to the square root of the equilibrium oxygen pressure. Varying the oxygen pressure changed the silver oxide content of the liquid and it was possible to obtain the activity of Ag2O over a range of composition. Experimental Procedure In principle, the method consisted of bringing melts in silver crucibles or boats to equilibrium at a fixed temperature under a definite pressure of oxygen and analyzing the glass after solidification. Materials: B2O3 glass was prepared from A.R. quality boric acid by fusion in platinum. The silver

Jan 1, 1954

By R. F. Shurtz

R. Duval (Mining Engineer, Ancien eleve de PEcole Polytechnique, Paris, France) I do not agree with the Eq. 3, reading: m =1/100- [(0.214x30.4) + (0.7B6 x0.00)] =6.5pct CaO If 0.214 and0.786 were proportions by weight, the equation would represent the well known mixtures law of the conventional arithmetics and 6. 5 pct CaO would be the correct average content. But it is not the case as the author states: "In samples consisting of single grains of mineral, those grains must, as already mentioned, be either of dolomite or magnesite. Since 78.6 pct of the deposit consists of magnesite and 21.4 pct of dolomite (excluding for present purpose the presence of other minerals), for any single grains picked at random the probability will be 0.214 that is it dolomite and0.786 that is it magnesite. In 1000 such samples the expected numbers of dolomite and mapesite grains will be 214 and 786 respectively." 0.214 and 0.786 would be proportions by weigbt under the necessary condition that all grains of dole mite and magnesite should have an identical weight. Obviously it is not the case, as the specific gravities are not the same for mapesite and dolomite and the volumes of the grains are different. Furthermore, because of these differences the conditions for a random sampling are not fulfilled and we are not authorized to state that the probabilities are, respectively, 0.214 and 0.786. The author however makes a simple application of Eq. 1: M = 1/n— ? fi x i . n Should we deduce that this relation is wrohg? Not at all, but when applying Eq. 1 you must not overlook what it actually. means. Eq. 1 gives a definition of the arithmetic mean of a total of n observed values Xi and nothing else. But the average conteht of a deposit has not the same significance. It is the ratio between the weight of concerned mineral in the deposit and the total weight of the deposit. As from 1000 particles the 214 of dolomite and the 786 of magnesite have not the same weight, the two definitions do not concur, and when applying Eq. 1 the result is an arithmetic mean of figures which has no connection with what is named average contentof a deposit. The situation is similar to the calculation of an average velocity. If a car travels a first mile over at 30 miles per hr and a second mile over at 60 miles per hr, when applying formula 1 you find as average velocity for the 2 miles: 30+60 ------- - 45 miles per hour. Many people calculate in this way and they do not realize that a mistake is involved. In fact the definition of he average velocity for the 2 miles is the quotient of the distance of 2 miles by the time (in hours) necessary for 2 miles travel, i.e.: 2 ---------- = 40 miles per hr. 1 + 1 30 60 In other words, the average volocity wanted is not the arithmetic but the harmonic average of the two velocities. The above mentioned bias in the calculation of the average contents of deposits is frequent, even in the assessments made by experienced engineers and is independant of what is named the sampling error. In order to supress the bias and to be able to use Eq. 1, you must apply a correction. An example on the subject can be found in an article by Duval et al. in the January 1955 issue of the &apos;&apos;Annales des Mines" (French), page 19. R. F. Schurtz (Author&apos;s Reply) Mr. Duval&apos;s position is quite correct. The proportions shown for dolomite and magnesite., respectively, of 0.214 and 0.786 are, in fact, proportions by weight uncorrected for specific gravity. In our day to day operation of producing magnesite from these mines at a very substantial rate, we do not normally make corrections for the difference between the specific gravity of dolomite and that of magnesite. If these corrections are made in Eq. 3 as shown in my article, then the numbers of grains turn out to be in proportions of 0.226 dolomite and 0.774 mapesite instead of the values actually shown in the equation. For the purposes of our work, and in view of the inherently low accuracy of the data, this correction was not deemed worthwhile making.

Jan 1, 1961

By P. R. Sperry, P. A. Beck, J. Towers

Dahl and Pawlek1 found that electrolytic copper develops extremely coarse grains at 1000°C after about 90 pct reduction by rolling. This coarsening occurs only under conditions of penultimate grain size, deformation, and alloying which lead to the "cube" recrystallization texture.l,2,3,5 The peculiar angular shapes and straight grain boundaries of the coarse grains were noted by several investigator.1,4,5 On the other hand, coarsening in Fe-containing aluminum or in AI-Mn alloys8 does not depend on a "cube" (or any well developed) recrystallization texture. It is true that increasing deformation by rolling, and, therefore, an increasingly well developed re-crystallization texture, are associated with decreasing incubation periods of coarsening.6-7-8 Nevertheless, coarsening readily develops in aluminum even after only 30 pct reduction by rolling, where the recrystallization texture is very weak.6,8 Also, coarsening was observed by Jeffries9 many years ago in sintered thoriated tungsten, which presumably has no preferred orientation. In all these cases coarsening is associated with grain growth inhibition by a dispersed second phase.8,9 The annealing temperature has to be suficiently high to overcome the inhibition at a few locations. But if it is too high, growth starts at many points, and the resulting grain size becomes much smaller.9 Normally, the coarse grains are more or less equiaxed, and the boundaries have a typical ragged appearance.6.8 Cook and Macquarie4 demonstrated that, in addition to the texture-dependent coarsening previously found at 1000°C,l electrolytic tough pitch copper may also coarsen at 800°C after 50 pct reduction by cross rolling. The coarse grains formed under such conditions have rounded shapes and ragged boundaries, like those in aluminum. When the annealing temperature is higher, the tendency for their formation decreases. All these observations suggest that the coarsening at 800°C is associated with inhibition by a second phase. Actually, coarsening at 800°C after 50 pct reduction by cross rolling was observed only in tough pitch copper,4 which contains Cu2O particles. On the other hand, the texture-dependent 1000°C coarsening occurs in both tough pitch and oxygen-free copper;4 it does not appear to depend on the presence of a dispersed second phase. However, the interpretation of the 800°C coarsening in Cu after 50 pct rolling as an inhibition-dependent process, similar to the coarsening in A1-Mn alloys, is somewhat weakened by the fact that this coarsening was reported4 to occur only after cross rolling, and not after straight rolling. It was, therefore, decided to re-examine this question. A 1 in. diam electrolytic tough pitch copper rod, No. 2 hard drawn, was annealed for 20 min at 700°C, rolled to 0.5 in., annealed 10 min at 700°C, and straight rolled to 0.064 in. It was then given a penultimate anneal of 20 min at 500°C and it was cut into four sections, which were given final reductions by straight rolling as follows: A 30 pct reduction of area B 50 pct reduction of area C 70 pct reduction of area D 90 pct reduction of area Specimens cut from the four sections were finally annealed at 800°C in an oxidizing atmosphere. Strip A remained fine grained up to 10 hr, but the specimen annealed 12 hr consisted of only 2 large grains. Strip B had a few scattered large (1/2 to 3/4 mm) grains after 1 min, although the balance of the specimen consisted of fine grains of about 0.02 mm. After 5 min there were several 10 to 15 mm grains present, and after 1 hr strip B was completely coarsened. The coarse grains had the same characteristics (see Fig 1) as those obtained by Cook and Macquarie at 800°C after cross rolling. Strip C had several grains of 0.05 to 1 mm after 1 min, but it was still largely fine grained after 12 hr. After 48 hr it consisted entirely of grains of about 0.5 to 4 mm, with an extraordinarily large number of twin bands. Strip D remained com- pletely fine grained after 4 hr at 800°C. These results indicate that, in the deformation range of 30 to 70 pct reduction, the incubation period for coarsening as well as the rate of growth and the final size of the coarse grains decreases with increasing deformation. Similar

Jan 1, 1950

By F. V. Lene, E. J. Westerman

The recrystallization behaviors of two extruded and cold-drawn experimental sintered aluminum powder alloys, containing 1.75 and 3.0 pct Al2O3 by weight, were compared with that of extruded and cold-drawn commercially pure alumirzum. The kinetics of recrystallization of the alloys are described semiquantitatively. For the alloy containing 1.75 pct A l203 the rates of nucleation and of growth were also semiquantitatively determined. THE most striking property of aluminum alloys strengthened by a dispersion of Al2O3, the so-called SAP alloys, is their stability at elevated temperatures. One of the manifestations of this stability is their resistance to recrystallization after they have been cold worked. Most of the commercial grades of either the Swiss SAP or of Alcoa&apos;s Aluminum Powder Metallurgy Products have not been recrys-tallized after cold working, even when they are heated for a long time at a temperature near the aluminum melting point. Lenel, however, observed that the dispersion strengthened aluminum alloys with a larger spacing between the oxide particles than that of most commercial grades would recrys-tallize.1 It appeared to be of interest to further investigate the mode and kinetics of recrystallization of these alloys, and to compare their recrystallization behavior with that of commercially pure aluminum. Because homogeneous deformation of these SAP alloys in tension did not provide sufficient cold work to induce recrystallization, they were cold worked by wire drawing; the nonuniformity of this deformation unavoidably complicated the interpretation of the recrystallization studies. EXPERIMENTAL DETAILS Extrusions—Two types of sintered aluminum powder extrusions were used in this study. One type, designated AT-400, was produced from Reynolds atomized aluminum powder consisting of spherical particles averaging 3µ in diam and containing 1.75 wt pct of Al2O3. This powder was very similar to the R3M powder from which extrusions were previously prepared with an average spacing of 0.9µ between oxide particles.2 The second type, designated MD 2100, was produced from Metals Disintegrating Co. flake powder containing 3.0 wt pct of Al2O3, with an average flake thickness of 0.8µ. The average spacing between oxide platelets in MD 2100 extru- sions was 0.45µ.2 Powder compacts of 3/4-in. diam were extruded at 1000°F into 0.097-in. diam (AT-400) and 0.093-in. diam (MD 2100) wires by methods previously described.3 In order to compare the recrystallization behavior of sintered aluminum powder extrusions with that of wrought commercially pure aluminum 3/4 in. rod stock of 1100 F aluminum was extruded at 1000°F into 0.102-in. diam wire. Wire Drawing—Tungsten carbide dies were used for the AT-400 and 1100 F alloys. They had an included angle of about 15 deg and reduced the wire area approximately 7 pct per pass. Steel dies with an included angle of 11 to 13 deg and an average reduction per pass of 10 pct were used for drawing the MD 2100 alloy, because drawing this alloy through the carbide dies produced overdrawing defects. Heat Treatment—The cold-drawn wires were cut into small samples, and the deformed ends were etched off. The samples were each wrapped tightly in a single layer of aluminum foil, and individually isothermally annealed in a lead bath. Metallography—The modes and kinetics of recrystallization were determined by metallography. Mounted and polished specimens were anodized in a solution of 1.8 pct HBF4;4 examination under polarized light clearly revealed their grain structures. The recrystallized grains were generally much larger than those of the unrecrystallized matrix, and could clearly be distinguished because they alternated between maximum and minimum light reflection when the microscope stage was rotated, while the unrecrystallized matrix had a comparatively homogeneous "salt and pepper" structure. The fractional recrystallized volumes of the dispersion hardened alloy wires were determined by cutting and weighing of recrystallized and total transverse areas on photomicrographs. The recrystallized grains in the 1100 F alloy were too small to be cut out individually; therefore a combination of cutting and lineal analysis was used in this case. RESULTS AND DISCUSSION Modes of Recrystallization—The modes of recrystallization of the three alloys varied widely. In the 1100 F alloy nucleation and growth started in the region midway between the center and the surface;

Jan 1, 1961

By E. R. Gilbert, D. E. Munson

Creep of copper under 75 to 1.50 kg per sq cm stresses at temperatures near the melting point was found to he a complex reaction controlled by three mechanisms acting in parallel. In order of appearance with decreasing temperature, the auerage activation energies, Qc , are 168, 79, and 24 kcal per mole. Stress dependence of the minimum creep rate was found to he an exponential for the two high-Qc processes and a power law for the low-Qc, process. Transition of control occurs from one mechanism to another. The relative transition temperature depends upon the applied stress, and the range oiler which the transition occurs depends upon the difference in the activation energies of the mechanisms. The creep behavior at high temperature is explained by the climb of dislocations through thermally or mechanically formed jogs, CREEP in pure fcc metals at temperatures in excess of one half the absolute melting point is normally controlled by dislocation c1imb.1,2 A climb model which seems applicable, according to an extensive analysis of data,3 was derived by Weert-man.4 This model assumes jog-saturated dislocations and predicts an activation energy, Qc, nearly equal to that for self-diffusion, USd. Although the requirement of jog saturation is restrictive, agreement between theory and experiment seems adequate. Many other theoretical treatments, including an early model by Mott,5 include detailed consideration of jog formation as an initial requirement for climb. These models predict activation energies for creep which differ from those of self-diffusion. Seeger6,7 postulates an observed activation energy related to the stacking-fault energy. Thus, Usd <Qc<5Usd + Uj where Uj is the jog-formation energy. Seeger incorporated qualitatively the influence of the relative numbers of thermal and athermal jogs. Expanding this concept, Shoeck8 explicitly states a function based on formation mode and relative numbers of vacancies: e r ci exp {-uf/k T} exp {-Um/k T} [ 1 ] where Uf and Um are energies of vacancy formation and migration, respectively. The concentration of jogs, Cj, depends upon the manner of jog production. For intersection jogs, Cj is not sensitive E. R. GILBERT, Junior Member AIME, formerly with De- to temperature; for thermal jogs, Cj is proportional to exp {—Uj /k T}. Schoeck regards each mode as a distinct mechanism; therefore, the mechanisms may act together.299 The diversity predicted by theory, surprisingly, has not been substantiated by experimental results. A significant investigation must include the extremes in stacking-fault energy. Extensive creep studies of aluminum10 and nickel,11,12 high stacking-fault energy metals, have been made. Comparable studies on a low stacking-fault energy metal, such as copper, have not. It is the purpose of this paper to report the results of an investigation of the creep of copper under conditions which favor thermal jogs. EXPERIMENTAL Cylindrical compression creep specimens (0.240 in. in diameter by 0.400 in. long) were machined from cold-drawn rods of electrolytic tough-pitch copper containing 0.0007 Mg, 0.002 Fe, 0.001 Ni, 0.0005 Ag, Cd < 0.005, and Pb < 0.005 wt pct impurities. Undetected spectrographically was a nominal 0.04 wt pct 0, which occurs as a Cu2O constituent distributed discontinuously at grain boundaries. Vacuum annealing at 900°C for 15 min produced a stable 0.03-mm average grain diameter. Testing was carried out using apparatus similar to that described by sherby,13 modified by enclosing platens and a push rod in a vacuum cylinder. Normally this arrangement resulted in pressures less than 10 Only a slight surface tarnish, less than 0.0005 in. in thickness, occurred during the test. The applied stress, corrected for atmospheric pressure, was maintained within 2 pct of the desired true stress by the addition of lead shot at fixed strain increments. Creep strain was measured with dial gages as a relative displacement of the upper and lower platens; accuracy of measurement was 0.0001 in. Two creep-test methods were used, the differential or cyclic temperature14 and the isothermal, to obtain creep data at stress levels of 150, 100, and 75 kg per sq cm over the temperature range of 620° to 1032°C. Minimum creep rates were used from both test methods; this was considered proper because comparable temperature tests or cycling to the same temperature gave the same creep rate, within experimental error. The cold vacuum test chamber, with the unstressed specimen in place, was heated to temperature by placing a preheated furnace over the chamber. Temperature equilibrium was attained within 30 min. For the cyclic tests, the stress was removed during the 5 to 10 min necessary to effect the temperature change and re-

Jan 1, 1965

By O. W. Simmons, L. W. Eastwood, C. M. Craighead

H. Schwartzbart and W. F. Brown, Jr.—The authors have divided the effects of recovery on the true stress-true strain curve into two types; metarecovery, which effects only the first part of the curve or the yield strength, and orthorecovery, which effects the flow stress at any strain. Both of these are said to be true recovery effects, involve no recrystallization, and are explained by the removal of two different types of imperfections caused by work hardening. However, there seems to be some question as to whether the data are sufficiently conclusive to exclude, as an explanation of the authors&apos; results, a mechanism based on the relief of residual stresses between the grains or slip bands and recrystallization. It appears that metarecovery could be interpreted in the same fashion as a customary interpretation of the Bausch-inger effect. The balanced system of internal stresses which exists between grains in a strained specimen due to varying orientation and, hence, yield strengths, of the different grains is responsible for a reduced yield strength in compression following pre-tension, and, similarly, for an elevated yield strength in tension following pre-tension. If the specimen is now heated so that the internal stresses are relieved by creep, then the yield strength in tension following tension will have been reduced and in compression following tension will have been raised. There seems to be a very strong case for the lack of recrystallization in the aluminum investigated by the authors, if one defines recrystallization as the presence of visually detectable new grains or accepts the X-ray evidence as conclusive. One must remember, however, that the appearance of spots on the back-reflection X-ray patterns cannot be taken as the time when recrystallization first started. The areas of recrystallized strain-free material must first have grown to a size large enough to give distinct spots on the patterns and this may take some time. Averbachl7 in an investigation of brass has shown that recrystallization can be detected by extinction measurements at temperatures lower than those based on hardness or X-ray line width determinations. It can be seen from fig. 10 that the rate of recrystallization is extremely low over a considerable time period at the onset of the process. Observations on the rate at which small amounts of recrystallization effect the flow stress would have given further insight as to whether undetectably small amounts of recrystallization might have been responsible for orthorecovery. Also, the question arises as to whether the effects observed in fig. 6 for various times and temperatures could not have been obtained if the time at 212°F were sufficiently long. In addition, the argument that the curve in fig. 10 is not sigmoidal seems weak in view of the scattering of the points. It is conceivable that an accurate determination of the curve for the first 100 hr would exhibit a relationship other than the one drawn. There is one point we would like to raise about the condition of the starting material. The authors annealed their material at 750°F for 15 min to remove the effects of any previous work hardening or machining strains. Reference to the work by Anderson and Mehl shows that this treatment may not have completely recrystallized the aluminum, so that the starting material may have had some strained areas. Higher temperatures or longer times may have been required to remove the effects of any small strains. We would like to mention some results of tests being conducted at the Lewis laboratory of the National Advisory Committee for Aeronautics in an investigation of the Bauschinger effect in relation to fatigue. Tests were performed on annealed electrolytic copper and several annealed brasses. Specimens were pre-strained 1 pct in tension and then tested in compression or tension with and without intermediate stress-relieving annealing treatments at 500°F for various times. Specimens heated at 500°F for 10 1/2 hr showed an elevation of the flow curve in compression and an approximately equal lowering of the flow curve in tension when compared with the curves for the un-heat treated specimens. After approximately 0.8 pct strain, all flow stresses coincided and were equal to the flow stress of the virgin material at this strain. This behavior is consistent with the metarecovery observed for aluminum by the authors and for which a residual stress model can be used. On the other hand, increas-

Jan 1, 1951

By Y. Nakada, U. F. Kocks, B. Chalrners

THE orientation dependence of work hardening has previously been studied over the entire range, i.e., including special orientations of high symmetry, in aluminum1-3 and silver.* The differences between various orientations are substantial, and the trend is the same in all fcc crystals. However, there are quantitative differences in behavior between aluminum and silver at room temperature, particularly in the (100) orientation. While many experiments on gold have been reported,5&apos;9 none included the special orientations. We therefore undertook tension tests on gold crystals of the axial orientations (100),(110), (Ill), (211), and, as a representative of single slip, one whose Schmid factor was 0.5 (hereafter referred to as m = 0.5). Single crystals of dimensions 1/4 by .1/4 by 6 in. were grown from the melt1&apos; at a rate of 4 in. per hr, using gold of 99.99 pct purity obtained from Handy and Harman. A growth rate of 1/8 in. per hr, or a purity of 99.999 pct,ll produced no difference in results. The crystals were annealed at 1000°C in air for 24 hr and furnace-cooled down to room temperature, after which they were electro-polished in a solution of potassium cyanide (40 g), potassium ferrocyanide (10 g), soda (20 g), and enough distilled water to make 1 liter of solution, at a current density of 0.02 amp per sq mm. After 2 or 3 hr, a very smooth surface was obtained by this method. Nine m = 0.5, two (loo), one (110), three (Ill), and one (211; crystals were tested at room temperature in a floor-model Instron machine at a tensile strain rate of 0.5 pct per min. The accuracy of the stress measurement was k10 g per sq mm up to 500 g per sq mm, k2 pct for higher stresses. The strain measurement was accurate to t2 pct. The scatter of stress at a given strain among the crystals of the same orientation was small, *5 pct being the largest. The representative stress-strain curves for various orientations are shown in Fig. 1. Table I summarizes the work-hardening parameters as used by seeger.12 Results of other investigators are also included in this table for comparison. There are no previous data for the corner orientations. Values for m = 0.5 crystals agree fairly well with those of Berner. Tm is defined as the stress at which the stress-strain curve begins to deviate from linearity of Stage 11. However, in practice this is a very difficult value to estimate because each investigator has a different idea of linearity. Therefore, the comparison with the values of other investigators may not be valid. In the present experiment, (100) crystals had the highest 111, followed by (111) crystals. The work-hardening rate in Stage I1 was highest for the (111) crystal followed by the (100) crystals. Our value for 0x1 of m = 0.5 orientation agrees very well with those of other investigators. 1) Single-Slip orientation (m = 0.5). These crystals were oriented so that the primary-slip vector was contained in one side face. The dimension perpendicular to this side should then not change if single slip indeed takes place. Within the accuracy of measurement, this dimension did not change during the deformation. Since the accuracy is k0.2 pct, the amount of secondary slip is less than 0.7 pct of the slip on the primary-slip system at 30 pct tensile strain. This is in conformity with the results obtained by Kocks" for aluminum crystals. The tensile axis moved toward (211) during the deformation. Slip bands, Fig. 2(a), were very fine and closely spaced. Some deformation bands were observed. There were no clear-cut cross-slip traces such as the ones observed on aluminum m = 0.5 crystals. 2) (111) Crystals. The orientation of the tensile axis was stable during the deformation. From this observation, one can deduce that at least three slip systems must have operated, and probably all six because the remaining three all have cross-slip relation to one of the first three.&apos; It was very difficult to observe the slip markings. Consequently, we could not confirm by this method that six systems were operative during the deformation. At high strains (above 50 pct shear strain), this orientation had the highest stress-strain curve. At 80

Jan 1, 1964

By J. H. Swisher

When an Fe-Mn alloy is internally oxidized, the inclusions formed are MnO which contains some dissolzled FeO. In the internal oxidation reaction, not all of the manganese is oxidized; some remains in solid solution as a result of the high Mn-0 solubility product in iron. Taking these factors into consideration, the rate of internal oxidation of an Fe-1.0 pct Mn alloy is computed as a function of temperature, using available thermodynanzic data and recently published data for the solubility and diffusivity of oxygen in iron. The predicted and experimentally determined rates for the temperature range from 950 to 1350°C are in good agreement. ThE rates of internal oxidation of austenitic Fe-A1 and Fe-Si alloys have been studied extensively.1"4 Schenck et al. report the results of a few experiments with Fe-Mn alloys at 854" and 956C, and Bradford5 has studied the rate of internal oxidation of commercial alloys containing manganese in the temperature range from 677" to 899°C. When Fe-Mn alloys are internally oxidized, the inclusions formed are solutions of FeO in MnO, the composition depending on the experimental conditions. Since the thermodynamics of the Fe-Mn and FeO-MnO systems have been investigated,6"9 and since the solubility and diffusion coefficient of oxygen in y iron have been determined recently,&apos; it is possible to predict the rate of internal oxidation from known data. The calculations used in predicting the rate of internal oxidation will first be outlined, then the results of the prediction will be compared with the experimental results of this investigation. PREDICTION OF PERMEABILITY FROM THERMODYNAMIC AND DIFFUSIVITY DATA Oxygen is provided for internal oxidation in these experiments by the dissociation of water vapor on the surface of the alloy. The dissociation reaction is: + H2(g) + [O] [1] where [0] denotes oxygen in solution. The equilibrium constant for this reaction is known as a function of temperature:&apos; log As oxygen diffuses into the alloy, oxide inclusions are formed which are MnO with some FeO in solid solution. The reactions occurring are: [Mn] + [0] = (MnO) [31 and [Fe] + [0] = (FeO) [41 where [ Mn] is manganese dissolved in iron and (FeO) is iron oxide dissolved in MnO. The overall reactions may be written as follows: [Mn] + HOte) = (MnO) + H2(£) [5] and [Fe] + H20(g) = (FeO) + Hz(R) [61 The standard free-energy changes and equilibrium constants for Reactions [5] and [6] are known.6 Therefore the equilibrium constants for Reactions [3] and [4] may be obtained by combining known thermodynamic data for Reactions [I], [5], and [6]. For Reactions [3] and [4]: K = and For the present purpose, both the Fe-Mn7,8 and FeO-~n0&apos; systems can be considered to be ideal, i.e., [amn] = [NM~] and (aFeO) = (NM~~) = 1 - (NFeO) where the Ns are mole fractions. These relations, together with Eqs. [I] and [8], permit us to compute both the oxide and metal compositions as a function of temperature and oxygen potential at any point in the specimen. For cases where the oxygen concentration gradient between the surface and the subscale-base metal interface is linear, the kinetics of internal oxidation is an application of Fick&apos;s first law: where dn/dt is the instantaneous flux of oxygen into the specimen, g-atom per sq cm sec; 6 is the instantaneous thickness of the subscale, cm; Do is the diffusion coefficient of oxygen in iron, sq cm per sec; p is density of iron, g per cu cm; h[%O] is the oxygen concentration difference between the surface and sub-scale-base metal interface, wt pct. B6hm and ~ahlweit" derived an exact solution to the diffusion equation for systems in which there is a stoichiometric oxide formed. They showed that the oxygen concentration gradient is given by a rather complex error function relation. For the Fe-Mn-0 system and for most other systems that have been studied, however, variations in oxide compositions are small and rates of internal oxidation are sufficiently slow that the deviation from linearity in the concentration gradient of oxygen is negligible. The mass of oxygen transported across a unit area of the specimen for the total time of the experiment is given by the mass balance equation:

Jan 1, 1969

By W. M. McKewan

Magnetite pellets were reduced in flowing hydrogen at pressures up to 40 atm over a temperature range of 350° to 500°C. The rate of weight loss of oxygen per unit area of the reaction surface was found to be constant with time at each temperature and pressure. The reaction rate was found to be directly proportional to hydrogen pressure up to 1 atm and to approach a maximum rate at high pressures. The results can be explained by considering the reaction surface to be sparsely occupied by adsorbed hydrogen at low pressures and saturated at high pressures. PREVIOUS investigation1,2 have shown that the reduction of iron oxides in hydrogen is controlled at the reaction interface. Under fixed conditions of temperature, hydrogen pressure, and gas composition, the reduction rate is constant with time, per unit surface area of residual oxide, and is directly proportional to the hydrogen pressure up to one atmosphere. The reduction rate of a sphere of iron oxide can be described3 by the following equation which takes into account the changing reaction surface area: where ro and do are the initial radius and density of the sphere; t is time; R is the fractional reduction; and R, is the reduction rate constant with units mass per area per time. The quantityis actually the fractional thickness of the reduced layer in terms of fractional reduction R. It was found in a previous investigation2 of the reduction of magnetite pellets in H2-H,O-N, mixtures, that the reaction rate was directly proportional to the hydrogen partial pressure up to 1 atm at a constant ratio of water vapor to hydrogen. Water vapor poisoned the oxide surface by an oxidizing reaction and markedly slowed the reduction. The enthalpy of activation was found to be + 13,600 cal per mole. It was also found that the magnetite reduced to meta-stable wüstite before proceeding to iron metal. The following equation was derived from absolute reaction-rate theory4,8 to expfain the experimental data: where Ro is the reduction rate in mg cm-2 min-&apos;; KO contains the conversion units; Ph2 and PH2O are the hydrogen and water vapor partial pressures in atmospheres; Ke is the equilibrium constant for the Fe,O,/FeO equilibrium; Kp is the equilibrium constant for the poisoning reaction of water vapor; L is the total number of active sites; k and h are Boltzmann&apos;s and Planck&apos;s constants; and AF is the free energy of activation. Tenenbaum zind Joseph5 studied the reduction of iron ore by hydrogen at pressures over 1 atm. They showed that increasing the hydrogen pressure materially increased the rate of reduction. This is in accordance with the work of Diepschlag,6 who found that the rate of reduction of iron ores by either carbon monoxide or hydrogen was much greater at higher pressures. He used pressures as high as 7 atm. In order to further understand the mechanism of the reduction of iron oxide by hydrogen it was decided to study the effect of increasing the hydrogen pressure on rebduction rates of magnetite pellets. EXPERIMENTAL PROCEDURE The dense magnetite pellets used in these experiments were made in the following manner. Reagent-grade ferric oxide was moistened with water and hand-rolled into spherical pellets. The pellets were heated slowly to 550°C in an atmosphere of 10 pct H2-90 pct CO, and held for 1 hr. They were then heated slowly to 1370°C in an atmosphere of 2 pct H2-98 pct CO, then cooled slowly in the same atmosphere. The sintered pellets were crystalline magnetite with an apparent density of about 4.9 gm per cm3. They were about 0.9 cm in diam. The porosity of the pellets, which was discontinuous in nature, was akrout 6 pct. The pellets were suspended from a quartz spring balance in a vertical tube furnace. The equipment is shown in Fig. 1. Essentially the furnace consists of a 12-in. OD stainless steel outer shell and a 3-in. ID inconel inner shell. The kanthal wound 22 in. long, 1 1/2, in. ID alumina reaction tube is inside the inconel inner shell. Prepurified hydrogen sweeps the reaction tube to remove the water vapor formed during the reaction. The hydrogen is static in the rest of the furnace. The sample is placed at the bottom of the furnace in a nickel wire mesh basket suspended by nickel wire from the quartz spring. The furnace is then sealed, evacuated, and refilled with argon several times to remove all traces of oxygen. It is then evacuated, filled with

Jan 1, 1962

By Robert E. Green, Wolfgang Sachse

Simullaneous ultrasonic attenuation measurements of both quasishear waves propagating in single cryslals of aluminum indicate that, in the undeformed annealed state, the dislocation density is generally not uniform on all slip systems. Change oof attenuation measurements made during plastic defortnation of crystals , which possessed specific orientations ideal for studying the orientation dependence of dislocation damping, indicate that, for low strain levels, dislocation motion occurs on additional slip systems besides the primary one, even for crystals oriented for plastic deformation by single slip. THE sensitivity of internal friction measurements permits such measurements to be used successfully in studying the deformation characteristics of metal crystals. On the basis of experimental observations, T. A. Read1 was the first to associate internal friction losses with various dislocation mechanisms. Since that time further work2-&apos; has been performed and a dislocation damping theory has been formulated by Granato and Lucke.6 In the amplitude independent region, this theory predicts the attenuation a to be dependent on an orientation factor O, a dislocation density A, and an average loop length L. if is a constant, independent of crystallographic orientation. For a given crystallographic orientation, changes in dislocation density and loop length give rise to the observed attenuation changes accompanying plastic deformation. The Granato-Liicke theory suggests the investigation of the orientation dependence of attenuation measurements in hopes of obtaining information to separate dislocation motion losses from other losses.7 An experimental study of the orientation dependence of attenuation in undeformed annealed single crystals should yield an insight into the uniformity of dislocation distribution throughout the entire specimen. A similar study on crystals plastically deformed in a prescribed fashion should give information about the alterations in the dislocation distribution on the slip systems activated during plastic deformation. The possible modes of elastic waves which can be propagated in aluminum,8 copper,9 zinc,10 and other hexagonal metals" have been calculated. Associated with each mode of wave propagation are dislocation damping orientation factors, which are based on the resolution of the stress field of that particular elastic wave onto the various operative slip systems in the material. These orientation factors have also been calculated as a function of crystallographic orientation in the papers cited above. Einspruch12 obtained agreement between predicted and observed attenuation values of longitudinal and shear waves in (100) and (110) directions of two undeformed aluminum crystal cubes. He ascribed the slight deviations between predicted and observed values to a nonuniform dislocation distribution, or to other loss mechanisms. In shear deformation of zinc crystals, Alers2 found that the attenuation of shear waves having their particle displacements in the slip plane was very sensitive to the deformation, while the longitudinal wave attenuation was affected only when the wave propagation direction was not normal to the slip plane. Using aluminum single crystals oriented for single slip, Hikata3 et al. found that during tensile deformation the change of attenuation of the shear wave (actually quasishear) having particle displacements nearly perpendicular to the primary slip direction exhibited the easy-glide phenomena, while longitudinal waves did not. Similar results were reported by Swanson and Green5 during compressive deformation of aluminum crystals. These results are in qualitative agreement with the calculated orientation factors for specimens of this orientation. In well-annealed, undeformed aluminum crystals, the damping is expected to be due to dislocations vibrating on all twelve slip systems. The orientation factors associated with this initial damping will be designated by O2 and O3, where a, represents the average orientation factor for the slow shear (or quasishear) wave and O3 represents the average orientation factor for the fast shear (or quasishear) wave. The calculation of these values for aluminum crystals by Hinton and Green8 shows that they vary very little as a function of crystallographic orientation—at most, by a factor of 2.47. If the dislocation density and loop length are uniform, then in the initial undeformed state, Here the subscript zero refers to the initial value of the attenuation. Also for aluminum, the calculations8 show that the orientation factors for primary slip only, associated with each shear wave, exhibit a sharp minimum for particular crystallographic orientations. A composite plot of the two shear wave orientation factors for primary slip only is shown in Fig. 1. Since these orientation factors are associated with dislocation motion occurring on the primary slip system only, the proper condition to check these factors might be attained by slightly deforming a single crystal oriented for primary slip. For dislocation motion on the primary slip system only,

Jan 1, 1969

By A. McLean, D. W. Gomersall, R. G. Ward

An experimental study has been made of the solubility of nitrogen in liquid iron and liquid Fe-C alloys using levitation melting and a rapid quenching device. Iron alloy droplets were equilibrated with nitrogen gas at 1 atm pressure, quenched, and analyzed. Previous techniques for studying the Fe-C-N system have produced data which me in marked disagreement. This disagreement is due largely to errors caused by reaction between the molten alloy and the crucible material. With the levitation procedure, errors from this source have been eliminated and precise solubility data obtained for temperatures between 1450° and 1750°C. C-N interactions in molten iron have been expressed in terms of first- and second-order free energy, enthalpy, and entropy parameters. ALTHOUGH the solubility of nitrogen in iron base alloys is in general small, the effects of nitrogen on the properties of steel may be quite profound. For most purposes nitrogen in finished steels is undesirable, particularly in the low-carbon grades, since on cooling to room temperature the solubility limit of nitrogen in the steel may be exceeded and this can lead to embrittlement and loss of ductility on aging. On the other hand, nitrogen can improve the work-hardening properties and machinability of steels while in certain stainless grades nitrogen is important in order to stabilize the austenite phase. It is, therefore, desirable that one should be able to predict the solubility of nitrogen in liquid iron alloys. To do this, information is required concerning the interactions between nitrogen and the various alloying elements which may be present in liquid iron. There have been several investigations of these effects in recent years1"7 and the interactions between nitrogen and many elements dissolved in liquid iron are now known to a high degree of precision at steel-making temperatures. Unfortunately, a number of iron alloy systems which are of interest in steelmaking have been difficult to deal with by the experimental techniques generally used for this type of investigation. Among the most important of these are the Fe-C alloys. In the past, two methods have been widely used for determining nitrogen solubilities: the Sieverts&apos; technique, in which the amount of nitrogen required to saturate a given mass of liquid metal at a particular temperature and pressure is measured volumetrically, and the sampled-bath technique in which liquid metal held in a crucible is equilibrated with a gas phase containing a known partial pressure of nitrogen, and samples drawn from the melt are quenched and analyzed. These two methods have been discussed in detail elsewhere.5,8 With the Sieverts&apos; technique, errors may be introduced from the following sources: i) Gas adsorption on metal films which have condensed on the cooler parts of the reaction chamber. ii) Uncertainty in the determination of "hot volume" calibrations. iii) Crucible-melt interaction, particularly if a gaseous reaction product is formed or if the melt becomes contaminated with material from the crucible walls. The sampled-bath method may also suffer from errors due to reaction between the melt and the crucible material. In addition, there is the possibility that gas may be lost from the sample during solidification and cooling. In the present investigation, the solubility of nitrogen in liquid iron alloys has been studied by means of a new technique based on the use of levitation melting equipment and a rapid quenching device. In addition to the fact that problems of the type outlined above are avoided, this particular approach has the following advantages: i) The high-frequency current induces vigorous stirring within the levitated droplet so that gas-metal equilibration is rapidly attained. ii) The gas phase surrounding the melt can be changed very quickly and is easily controlled. For example, a droplet may be levitated in helium, deoxidized in hydrogen, equilibrated with nitrogen, and quenched, within a period of 15 min. ii) Melts can be readily under cooled or superheated, thus extending the effective temperature range of an investigation and allowing temperature-dependent data to be determined with a high degree of precision. Excellent reviews of levitation melting techniques and their application to physical-chemistry studies at high temperature have been published recently by Jenkins et al.,9 Peifer,10 and Rostron.11 In the present investigation a levitation melting technique has been used to obtain data for the solubility of nitrogen in pure liquid iron and liquid Fe-C alloys at temperatures between 1450° and 1750°C. The solution of nitrogen in liquid iron can be described by the reaction:

Jan 1, 1969

By Charles M. Packer, Oleg D. Sherby, Roy H. Johnson

Originally round tensile specimens of a eutectic Zn-A1 alloy develop elliptical cross sections during superplastic deformation. This observation, coupled with a detailed study of the microstructure and preferred orieniation, suggests that crystallographic slip and continuous grain boundary migration or re-crystallization are important processes during super-plastic deformation. In spite of the extensive activity in superplasticity1-15 and the numerous explanations proposed, no single model has had universal acceptance. It has been established, however, that the general requirements for superplastic extension of two-phase alloys include an extremely fine, stabilized grain size of the order of a few microns, a temperature about equal to or greater than one-half the melting point, a critical range of strain rate, and a similarity in the mechanical strength of the major phases. The proposed models can perhaps best be characterized in terms of the important phenomena associated with them. These phenomena include: phase instability,1 diffusional creep by volume diffusion3 or grain boundary diffusion4,5 slip and continuous grain boundary migration or recrystalliza-tion,= grain boundary Sliding,7-9,13,14 and dislocation glide.&apos;5 In this paper, experimental observations will be reported which support a model involving slip and continuous grain boundary migration or recrystalliza-tion. Specifically, a correlation will be made between this model and the development of elliptical cross sections as originally round specimens are superplas-tically deformed. For the most part, superplasticity studies have been conducted with eutectic or eutectoid alloys. Probably the most thoroughly studied material has been the monotectoid Zn-A1 alloy.1,2,6,12,13,15 No attention to the eutectic Zn-A1 alloy has previously been reported, and the results discussed in this paper represent part of a general study of the superplastic properties of this alloy. MATERIALS The alloys used in this investigation were prepared by melting appropriate quantities of 99.99+ pct A1 and 99.999 pct Zn in air, mixing, and pouring into a water- cooled stainless-steel mold. Wet-chemical analysis was conducted with each heat of alloy prepared, using the procedure of Fish and smith.16 The composition of the eutectic alloy was 95.1 wt pct Zn. Ingots about 2 in. thick were rolled to 0.4-in. plate at about 300°C with a reduction of 5 to 10 pct per pass. Specimens were machined from the plate with the tensile axis parallel to the rolling direction. The specimens were round, with 0.150-in.-diam, 1.25-in.-long gage length, and 0.25-in.-diam threaded grip sections. EXPERIMENTAL PROCEDURE Specimens were mounted inside a uniform-temperature quartz tube which was surrounded by a double elliptical radiant furnace with a 12-in.-long uniform-temperature hot zone and a low thermal capacity. The tube extended through the top and bottom of the furnace and permitted rapid quenching of the loaded specimens when quickly filled with cold water at the conclusion of the test. The quench precluded any effects on specimen microstructure from a normal, slow cool. Constant stress was applied to test specimens by suspending a load on a constant stress cam of the type described by Hopkin.17 The design of this cam permitted application of a constant stress for elongations up to 200 pct. For greater elongation, approximately constant stress conditions were maintained by systematically reducing the load manually. RESULTS As part of an investigation of the superplastic properties of the eutectic Zn-A1 alloy, evidence was obtained for the development of elliptically shaped cross sections as originally round specimens were extended. For example, after an elongation of about 100 pct, a round specimen with an initial diameter of 0.150 in. became elliptical with major and minor axis of 0.128 and 0.88 in., respectively. Photographs are presented to illustrate the ellipticity developed during superplastic deformation, Fig. 1. The specimen shown was deformed at a stress of 500 psi, at a temperature of 285°C, and a strain rate of 2.28 x 10-2 min-1. The strain-rate sensitivity exponent* was measured at *The strain-rate sensitivity exponent, m, is defined as d In o/d In c where o is the steady-state flow stress and E is the strain rate. this temperature and in the strain rate range 10"3 to 10-1 min-1 was found to be about 0.5. This value is typical of those observed with superplastic materials. The material studied exhibited negligible strain hardening during superplastic deformation, the creep rate remaining constant under constant stress and temper-

Jan 1, 1969

By W. R. Hudspeth

AS an outgrowth of its spodumene recovery operation at Kings Mountain, N. C., Foote Mineral Co. has been recovering a heavy mineral by-product. Foote leased this idle plant in 1951, reactivated it, using a new spodumene recovery process, and purchased plant and properties in October 1951. While the operation at Kings Mountain is primarily concerned with the production of spodumene concentrate, pilot plant work determined that the pegmatites also contained heavy minerals including cassiterite. Plans were made to recover the heavy minerals as a by-product and the flowsheet incorporated these facilities when the mill was modified for the new spodumene recovery process. The orebodies consist of spodumene, feldspar, quartz and mica. Apatite, tourmaline, and beryl are present in small quantities. The wall rock is pre- dominantly hornblende shist. The heavy minerals, including cassiterite, columbite, pyrrhotite, monazite, pyrite, and rutile represent about 0.2 pct of the ore. The fine-grained heavy minerals are disseminated throughout the dikes, apparently unassociated with the spodumene. The pegmatites are quarried and secondary breakage is by mud-capping and block-holing. Power shovels load into trucks transporting the ore to a coarse ore bin. A Telesmith 10x36-in. apron feeder delivers the ore to an 18x36 in. Traylor Jaw crusher adjusted to discharge -3 in. product to a primary conveyor. The conveyor delivers to a 4x5-ft Tyrock single deck vibrating screen using 3/4 in. cloth. The screen undersize is elevated to the crushed ore bin. Screen oversize goes to an Allis-Chalmers Hydrocone Crusher fitted with 4 in. concave and set to deliver approximately 66 pct minus 3/4 in. The crusher discharge returns to the primary conveyor. The crushing and screening installation has a capacity of about 60 tons per hour. Spirals The crushed ore is delivered at a rate of 350 tons per day to two 6x8-ft Hardinge Pebble Mills, equipped with 20 mesh Ton-Cap trommel screens. The screen oversize is pumped to a 12-in. hydroclone for primary desliming. The hydroclone underfl spirals. There is no heavy mineral loss in the hydro-clone overflow. The spirals bank consists of eight 5-turn Model 24-A Humphreys Spirals. The top port and the last four ports of each spiral are blanked, the remaining nine port splitters are adjusted to remove about 5 pct feed weight. The heavy mineral rougher concentrates are upgraded on a Deister Overstrom table. The spiral concentrates contain approximately 4 pct heavy mineral, and the spiral reject, which goes to another section of the plant for spodumene recovery, contains about 0.03 pct heavy mineral. There is an interesting feature in the spirals installation. An adjustable splitter mounted on the discharge boxes splits out a mica fines product containing very little heavy mineral. The mica product is cleaned by spiralling and screening. Thus the spirals recover two products; mica, and a heavy mineral rougher concentrate. Table Treatment The rougher spiral concentrate goes to a Deister Plato table, modified to receive a Deister-Overstrom No. 6 rubber cover with sand riffles. The table is operated with a 5/8 in. stroke, 270 strokes per minute, and a slope of 1/2 in. per ft from feed to tailings side. There is no slope adjustment from motion to concentrate end. Wash water consumption is relatively high, since the large spodumene grains tend to report with the fine heavy minerals. A middling band about 4 in. wide is maintained in order to produce clean concentrate. The middling, representing about 10 pct of table feed, is recirculated by air-lift. A band of concentrate grade coarse spodumene occurs just below the middling. This is removed and delivered to concentrate storage. The table tailing, containing approximately 0.7 pct heavy minerals, is returned to the spodumene feed preparation circuit. The heavy mineral table concentrates are approximately 45 pct cassiterite, 33 pct columbite, 14 pct pyrrhotite and 8 pct monazite, together, with some rutile, pyrite, and copper from blasting wire. Concentrate is collected at 24 hour intervals. and dried. If the concentrate remains in wet storage appreciably longer surface oxidation takes place which seriously interferes with the subsequent magnetic separation process. About 150 lb of heavy mineral concentrate is produced per 24 hours and shipped to the company&apos;s plant at Exton, Pa. for final separation into tin and columbium concentrates.

Jan 1, 1952

By T. L. Chu, G. A. Gruber

Silica films hare been rleposited 011 silicon substmtes at 400° to 600°C by a chemical-transport technique using hydrogen fluoride as the transport agent ill a closed system. This transport takes place from a source materia1 1071: temperature to substrates at higher temperatures, as indicated by the thermochemistry of the transport reaction. The experimental variables of- the transport process, such as the substrate temperature, the pressure pi the transport agent, and so forth, have been studied. The rate -determining step of the transport process appears to he the ),ale of chemical reaction in the source region. The transported films are similar to thermally grown silica films in physical proper-ties with the exception of &apos;some what higher dissolrrtion rates. SILICA films deposited on suitable substrates serve many purposes in electronic devices. They are used for the fabrication of tunneling devices, the surface passivation of devices, and the shielding of devices from nuclear radiation: and as selective masks against the diffusion of specific impurities into semiconductors. Doped silica films can also be used as sources for the diffusion of impurities into semiconductors. Several oxidation and deposition techniques for the preparation of silica films have been developed to meet the requirements of these applications. The therma1 oxidation of silicon by oxygen or steam at temperatures above 900 C is commonly used in silicon technology. The deposition techniques are perhaps more advantageous since they usually require lower temperatures and are not limited to silicon substrates. Silica films have been deposited on silicon and other substrates by reactive sputtering and chemical reactions. The sputtering of silicon in an oxygen atmosphere is capable of depositing good-quality silica films on silicon&apos; and gallium arenide. Many chemical reactions are known to yield silica at room temperature or higher. These reactions may involve intermediate steps. However, the final step yielding silica should take place predominately on the substrate surface in order to produce adherent films. When silica is formed in the gas phase by volume reactions, no adherent deposit can be obtained. Generally, the experimental conditions of a reaction can be varied so that the surface reaction predominates over the volume reaction. The chemical reactions which have been used successfully for the deposition of silica films are briefly as follows. The pyrolysis of alkoxysilanes in an inert atmosphere or under reduced pressure has been employed to deposit silica films on germanium3 and silicon4 at 650" to 750°C in a flow system. The deposition of silica films from alkoxysilanes has also been achieved at nearly room temperature by a low-pressure plasma. Device quality silica films have been deposited on germanium and gallium arsenide by the deposition of an amorphous thin silicon film followed by oxidation at 600" to 700" . Silica films for high-temperature capacitors have been produced by the hydrolysis of silicon tetrabromide at 950°C in argon and hydrogen atmospheres.7 We have developed a chemical-transport technique for the deposition of silica films on semiconductor substrates at relatively low temperatures. The thermochemistry of the transport reaction, the experimental variables of the transport process, and the properties of the transported silica films are described in this paper. THERMOCHEMICAL CONSDERATIONS The transport of solid substances by chemical reactions in the presence of a temperature gradient has been used for the preparation of films and crystals of many electronic materials. In this technique, a gaseous reagent is chosen so that it reacts reversibly with the solid substance under consideration to form volatile products. Since the equilibrium constants of most reactions are temperature-dependent, the transport of these products to regions of suitable temperature in the reaction system would cause the reverse reaction to take place. depositing the original solid. When the equilibrium is shifted toward the formation of the solid as the temperature is decreased, the solid is transported from a high-temperature zone to a lower-temperature region, and vice versa. This chemical-transport technique can be carried out in a closed or gas-flow system. In a closed system, chemical equilibrium is presumably established in the different temperature regions of the system, and the transport agent regenerated in the deposition region repeats the transport process in a cyclic manner. The local chemical equilibrium may not be approached in a flow system: however, this system offers a greater degree of flexibility. Silica reacts reversibly with hydrogen fluoride and this reaction was chosen for the transport process. The over-all reaction between silica and hydrogen fluoride may be written as: SiO2(s) + 4HF(g-) = SiF4Ur) + 2H2O(^)

Jan 1, 1965