Search Documents

By A. Karim, T. G. Langdon, C. Y. Cheng, J. E. Dorn

The effect of changes in temperature and stress on the creep behavior of Fe-4 pct Si was investigated over the temperature range from 650° to 1175K Over the temperature range of 799° to 110l°K, the shear strain rate, ?', obeys the relationship: where K is a constant, 7 the applied shear stress, G the shear modulus, g the free activation energy for diflusion, R the gas constant, and T the absolute temperature. The results strongly suggest that the creep in this alloy is controlled by a dislocation climb mechanism. In a recent investigation, Ishida, Cheng, and Dorn1 observed that the apparent activation energy for creep of a! Fe from 480" to 775°K corresponded to that for self-diffusion. Over the magnetic transformation range (775" to 1045"K), the apparent activation energy for creep increased steeply with increasing temperature in agreement with the known decrease in the free activation energy for self-diffusion. Passing through the Curie temperature, the apparent activation energy decreased abruptly following which it then increased mildly as the temperature further increased. Inasmuch as the apparent activation energy for creep of a! Fe below the Curie transformation range of temperatures appeared to decrease with increasing values of the stress, the data were analyzed in terms of a model based on the nonconservative motion of jogged screw dislocations. But since the observed creep rate was much higher than that which could be accounted for on the basis of usual models, the authors found it necessary to assume that pipe diffusion became significant over the lower temperature range. Furthermore, the abnormally high activation energy for creep above the Curie temperature was attributed to grain boundary migration which became prevalent in this temperature range. It was thought that a number of the interesting issues that arose in the creep of a Fe might be resolved by investigating the creep of Fe-4 pct Si alloy: This alloy has about the same Curie range and therefore should give creep results that closely parallel those for a Fe and in this way provide comparable data for analysis of creep mechanisms. Furthermore, it can be annealed at higher temperatures which could provide some stabilization against grain boundary migration. In addition, it might serve to illustrate the possible effects of short-range ordering on creep.

Jan 1, 1969

By Matti J. Saarivirta

The relative machinabilities of the precipitation-hardenable Cu-0.9 pct Cr, Cu-0.25 pct Zr, and Cu-0.9 pct Cr-0.25 pct Zr alloys are improved from 18 to 80 pct by 0.75 to 1.25 pct Pb. The Lead also improves the tensile strength, hardness, ad resistance to softening of the Cu-Cr and Cu-Cr-Zr alloys. When the alloy containing 0.7 to 2.2 pct Cr, 0.10 to 0.25 pct Zr, 0.75 to 1.25 pct Pb, and balance copper is solution-annealed at 1000°C (1830°F) and then aged at 525°C (Y75°F), tensile strength of 65,000 psi, 16 pct elongation, 62Rb hardness, and 84 to 87pct IACS electrical conductivity are obtained. The copper base Cu-Cr, Cu-Zr, and Cu-Cr-Zr alloys have high electrical and thermal conductivities, high strength and resistance to softening at elevated temperatures, but their machinability is poor, i.e., close to that of the pure copper. In the manufacture of various components such as welding electrodes or electrical and thermal conductors it is ofter desirable that the alloy have good machinability for high production rates and long tool life. Thus, alloys that combine these properties may be used not only for existing applications, but for new ones where good machinability is a requirement. This investigation was undertaken to improve machinability without impairing other desirable characteristics of the Cu-Cr, Cu-Zr, and Cu-Cr-Zr alloys. These commercial alloys contain about 0.4 to 1.2 pct Cr, 0.10 to 0.25 pct Zr, and balance copper. Lead was selected as an element that should improve the machinability and possibly other characteristics of the alloy. WORK PROCEDURE This investigation began with a study of the effect of lead on the hardness, tensile strength, electrical conductivity, and machinability of the binary alloy containing 0.10 to 1.30 pct Cr and balance copper. This was followed by an investigation of the effect of lead on the Cu-0.1 to 0.4 pct Zr and Cu-Cr-Zr alloys. The alloys were made from initially oxygen-free copper and for comparison some castings were made from phosphorus-deoxidized Lake Copper. Commercially pure chromium, zirconium, and lead were used for alloying. In melting the chromium was added to molten copper at 1400°C (2550°F), the heat held for 15 min at temperature, and the temperature then lowered to 1350°C (2460°F) when the other elements were added. The alloys were held for 15 min at this temperature and cast into a cast-iron mold. One-kilogram, 3-in.-diam castings were made. An argon cover was maintained during melting and casting to prevent oxidation. The chemical compositions of the alloys are shown in Table I. The micro structure and properties of the alloys were studied in the cast and wrought conditions prior to and after various solution-annealing and heat-treating cycles. The relative machinability was determined as a function of the lead content and compared with the free-machining Cu-lpct Pb alloy. To determine the machinability, 13/16-in.-diam cast specimens were cut by a machine hack saw using a constant cutting speed of 50 fpm, a fixed feed force, and a new blade for each alloy specimen. Molybdenum highspeed steel saw blades, 0.032 in. thick with 14 teeth per in., were used in all tests. No cutting fluid was used. Five tests were made on each specimen and the average elapsed time per test was taken as a measure of the relative machinability for that alloy. The variation of cutting time for each alloy was less than 110 pct.

Jan 1, 1969

By R. D. Stover, F. O. Shuck

The variation of binary-diffusion coefficients with composition in the liquid Bi-Sn system at 300°C was measured using the capillary-reservoir technique. The experimental coefficients did not exhibit a simple composition dependence. A number of tests were also made to determine the effect of variation of density with composition on convection within the capillaries. These tests indicate that the density changes arising from composition duferences may give rise to high diffusion coefficients unless care is taken to maintain a stable configuration. In recent years there has been considerable interest in diffusion in liquid metals. However, the bulk of the work done to date has involved measurements of self-diffusion or diffusion of dilute concentrations of "solute" metals. Because there are possible applications to liquid phase processing of metal alloys and since additional data may help to clarify the nature of diffusion in liquid metals, investigations of the concentration variation of diffusion coefficients in binary alloys have been undertaken. Of past studies only two have covered large ranges of binary composition. Niwa et al.' investigated the Pb-Sn, Bi-Pb, and Bi-Sn systems and Grace and Derge2 investigated the Bi-Pb system. Because of the discrepancy between the data of these investigators, it was deemed desirable to check the possibility of convection arising from the variation of density with composition. The capillary-reservoir technique of Anderson and saddington3 was chosen for the present investigation since it has been widely used for measurements involving dilute solute concentrations and it was the technique used by Grace and Derge.2 It was necessary, however, to modify the normal procedure of using a pure metal bath to avoid measurement of integral diffusion coefficients over large concentration ranges. The alloy system chosen for study was the Bi-Sn system for which there are some data available.' EXPERIMENTAL PROCEDURE In order to utilize the capillary-reservoir technique for measurements of diffusion coefficients as a function of composition, it was necessary to modify the normal procedure slightly. This modification consisted of using a reservoir composition which was only slightly different from that in the set of capillaries, in order that the integral coefficients measured might be taken over a small composition range and might thereby be assumed to approximate the differential coefficient at some average composition between that of the capillary and that of the bath. It was found that a composition difference of approximately 0.10 mass fraction provided a satisfactory compromise between a small increment in composition and reasonable accuracy in measuring changes of composition. The alloys were prepared from tin and bismuth of 99.999 pct purity. The materials were first cleaned of surface oxides, then weighed and melted to form a homogeneous alloy in a Pyrex filling apparatus. The capillaries used were of precision bore Pyrex tubing 1.000 *0.005 mm bore. Capillaries of various lengths (from 2.85 to 5.85 cm) were cut and sealed at one end by fusing a 3-mm Pyrex rod to the end. The capillaries were filled by inserting them into the filling chamber, evacuating the chamber, submerging the capillaries, and pressurizing the chamber with argon. Two sets of capillaries were prepared for each run. One set contained an alloy more dense than the bath alloy and one set a less dense alloy. This procedure was followed so that the effect of density difference on the measurement could be determined. After filling one set of capillaries, these were placed open end up in a capillary holder and transferred to the diffusion chamber which consisted of a Pyrex tube inserted into a graphite block. The second set of capillaries was then filled and placed into the diffusion chamber. The reservoir alloy, which had been melted in a separate crucible, then was poured carefully into the diffusion chamber. Sufficient reservoir alloy was used to cover all capillaries to a depth of several centimeters. All of these operations: capillary filling, melting of alloys, and diffusion anneal, were carried out in a fused-salt bath which maintained a temperature of 300"*1°C. All capillaries were permitted to diffuse for approximately 48 hr. At the end of the diffusion period the capillary holder was removed from the bath and

Jan 1, 1969

By K. A. Gschneidner, O. D. McMasters, T. J. O’Keefe

X-ray diffraction, differential thermal, ad rnetallo-graphic methods were used to establish the Dy-Pb Phase diagram. Lead additions lower the 1377°C transformation temperature of dysprosium to 1360°C leading to an inverted peritectic reaction. The 327°C melting point of lead is lowered by dysprosium additions to about 326°C yielding a eutectic reaction. A second eutectic reaction occurs at 13.3 at. pct Pb and 1200°C. The dysprosium-richest intermetallic compound DysPb3 melts congruently at 1695°C and crystallizes in the hexagonal Mn5Si3 (D8,) type structure. The peritectic decomposition temperatures for the remaining compounds are Dy5Pb, at 1555C, DyPb2 at 955C, and DyPb3 at 880°C. A fifth compound near the DyPb stoichiometry exists over a 310°C temperature range decomposing at 1130°C by means of an inverted peritectic reaction and melting incongruently at 1440°C. The crystal structures of the compounds are discussed. A systematic study of the rare earth-lead alloy systems is underway in an effort to supply information concerning the alloying behavior of the rare earth metals. The Dy-Pb phase diagram is the fourth system to be investigated in this study. The Yb-Pb,1 Y-Pb,2 and Eu-Pb 3 diagrams have been published recently. Utilization of the rare earth series of metals as a research tool in this manner should yield a better understanding of alloy formation. EXPERIMENTAL PROCEDURE Materials. The lead used in this investigation was obtained from Cominco Products, Inc., and was specified to be 99.99 pct pure. The dysprosium was prepared in this Laboratory by the calcium reduction of the fluoride followed by distillation of the dysprosium. The major impurities in the dysprosium in ppm are: A1 (<40), Ca (400), Er (<50), Gd (<200), Ho (<200), Mg (<50), Si (30), Ta (400), Tb (<100), Y (<50), 0 (651, H (15), N (not detected), F (430), C (35). Alloy Preparation. Most of the alloys were prepared by melting weighed amounts of dysprosium and lead in sealed tantalum crucibles. The tantalum crucibles were sealed by are-welding in a He-Ar atmosphere welding chamber. Thus the alloys are in contact with He-Ar at about 1 atm pressure. Homogenization was achieved by holding them in the liquid state for about 1 hr, cooling, inverting the crucibles, remelting, and repeating the process at least twice. Since these alloys were prepared in sealed tantalum crucibles, chemical analysis for final composition was thought to be unnecessary. No detectable reaction of these alloys with the tantalum crucible was observed by metallographic examination. Metallographic evidence was also used to confirm the homogeneity of some of the alloys prepared in this manner. The compositions of a few alloys, which were prepared by nonconsum-able are-melting, were corrected for the small weight losses involved by assuming that the weight loss is due to vaporization of lead. The specimens obtained from the alloy samples were prepared under a dry-argon atmosphere because they were rapidly attacked by air and moisture. Thermal Analysis. Differential thermal analysis methods were used to determine the liquidus curves and reaction horizontals of the system. Both Pt vs Pt + 13 pct Rh and W + 5 pct Re vs W + 26 pct Re thermocouples were used to measure the temperature. An X- Y recorder was used to record the specimen temperature and differential electromotive force between the specimen and molybdenum standard. The arrest temperatures were measured potentiometri-cally. The accuracy limits (* values) associated with the reaction temperatures obtained by this method were estimated on the basis of both the reproducibility of the particular temperature value and the accuracy of the thermocouple at a given temperature. Liquidus temperatures were obtained from cooling arrest data while both heating and cooling arrest data were used to establish the horizontals of the diagram. Heat treatments during the thermal analyses of the alloys between 40 and 70 at. pct Pb were necessary in order to approach equilibrium conditions. The samples were held at temperatures between the various peritectic horizontals for l to 2 hr before the thermal analyses were continued. The entire range of compositions was investigated at the expense of a minimum amount of materials by adding appropriate amounts of lead to master alloys. More than sixty alloys were analyzed by this differential thermal method and for each alloy the results given herein are taken from two or three heating and cooling cycles. X-Ray and Metallographic Methods. Slice specimens for metallography and powder specimens for X-ray diffraction were prepared from rod-shaped samples which had been melted in sealed 0.62 5-cm-diam tantalum crucibles. The specimens were heat-treated in sealed tantalum crucibles which were protected by sealing them in argon-filled quartz ampules. Quenching was accomplished by breaking the ampules in ice water after heat treatment. X-ray powder specimens were sealed in 0.3-mm-diam glass capillaries under a dry-argon atmosphere. Copper, iron, and chromium radiation were used to obtain the powder patterns for these alloys. More than 150 powder patterns were obtained for specimens of various compositions and heat treatments. Included in these were several patterns for specimens which had purposely been oxidized. Patterns from specimens which had been accidentally exposed

Jan 1, 1969

By R. L. Skaggs, D. T. Peterson

The effect of carbon in solid solution on the plastic behavior of thorium was studied by measuring the flow stress of Th-C alloys from 4.2" to 573°K and at several strain rates. Carbon was found to strengthen thorium primarily by increasing the thermally activated component of the flow stress. The strengthening due to carbon was directly proportional to the carbon content and decreased rapidly with increasing temperature up to 423" K. The flow stress also increased with increasing strain rate. The strengthening appears to be due to a strong short-range interaction between carbon atoms and dislocations. A yield point was observed in the Th-C alloys which increased with increasing carbon content. JTREVIOUS study of the mechanical properties of thorium has been confined largely to the measurement of the engineering properties. Work prior to 1956 has been summarized by Milko et al.1 who reported that additions of carbon to thorium sharply increased the room-temperature strength. In addition, the yield strength was observed to decrease rapidly over the temperature range from 25" to 500°C. In 1960, Klieven-eit2 measured the flow stress of thorium containing 400 ppm C. He found that over the temperature range from 78" to 470°K the flow stress was strongly dependent on temperature and rate of deformation. A drop in the load-elongation curve, or a yield point, was observed over most of the above temperature range. Above 470°K, the flow stress was nearly independent of temperature and strain rate. This strong temperature and strain rate dependence of flow stress is not generally observed in fcc metals. It is, in fact, more typical of the behavior reported for bcc metals. Bechtold,3 Wessel,4 and conrad5 have pointed out the striking difference between the commonly studied bcc metals and fcc metals in regard to the effect of temperature and strain rate on the flow stress. Zerwekh and scott6 studied the plastic deformation of thorium reported to contain 12 ppm C. They found that this material did not obey the Cottrell-Stokes law as expected for fcc metals. In addition, they found values of the activation volume smaller by an order of magnitude than expected for an fcc metal. They concluded that thorium was strengthened by a randomly dispersed solute. Thorium differs from many other fcc metals that have been studied extensively in that it shows a relatively high carbon solubility at room temperature. Mickleson and peterson7 report the solubility limit at room temperature to be 3500 ppm C. The lowest value reported is that of Smith and Honeycombe8 who report the limit to be 2000 ppm C at 350°C. The pres- ent investigation was a systematic study of the flow stress and yield point phenomenon of thorium over a broad range of carbon content, temperature, and strain rate. EXPERIMENTAL PROCEDURE The thorium used in this investigation was produced by the reduction of thorium tetrachloride with magnesium as described by Peterson et a1.&apos; Chemical analysis of the original ingot after arc melting and electron beam melting is shown in Table I. Alloys were prepared by arc melting this thorium with high-purity spectrographic graphite. Threaded specimens with a gage length 0.252 in. diam by 1.6 in. long were used for the constant stress or creep measurements. These specimens were machined from rod which had been cold-rolled and swaged to % in. diam. Tensile specimens were prepared by swaging annealed 3/8 -in.-diam rod to 0.102 *0.001 in. The as-swaged wire was cut to lengths of 2 in., annealed, and the center 1-in. gage length elec-tropolished to 0.100 ±0.001 in. The specimens were gripped for a length of 3 in. at each end by a serrated four-jaw collet which was tightened by a tapered compression nut. No slipping occurred in the grips and negligible deformation was observed outside the 1-in. gage length. Both the creep and tensile specimens were annealed at 730°C under a vacuum of 1 x X Torr. The resulting structures consisted of equiaxed recrystallized grains with a grain size of 3200 grains per sq mm for the tensile specimens and 2200 grains per sq mm for the creep specimens. After the specimens were prepared, samples were analyzed for nitrogen, oxygen, and hydrogen. The results of these analyses are given in Table 11.

Jan 1, 1969

By R. K. Buhr, H. Thresh, M. Bergeron, F. Weinberg

The microsegregation of nickel and chromium in directionally solidified AISI 4340 steel castings has been measured using electron probe microanalysis. Minimum concentrations were observed to occur at the center of the dendrite stalks, as reported previously. However, the position of the maximum concentrations could not be determined from the dendrite configuration. An etching technique has been developed which shows the position of maximum concentration. Measurements were made of the distribution of nickel and chromium in the dendritic structure. The ratio of maximum to minimum concentration was determined as a function of primary dendrite arm spacing for both as-cast and partially homogenized materials. The results are compared with observations reported in the literatuve. A S part of an investigation of solute segregation in directionally solidified AISI 4340 steel casting, electron probe microanalysis was used to measure the distribution of nickel and chromium in the dendritically segregated material. This complemented the observations on the segregation of radioactive phosphorus obtained by autoradiographic techniques in the same castings.&apos; Kattamis and I?lemings2 have reported similar measurements, using electron probe microanalysis, on dendritically segregated AISI 4340 steels. They found minimum solute concentrations occurred at the center of dendrites, maximum concentrations occurred in the center of interdendritic regions, for symmetrically oriented dendrites, and a regular variation of concentration within dendrite arms. The ratio of maximum to minimum concentration was found to be essentially independent of position in the castings, except near the chill surface where it was somewhat higher. Assuming complete diffusion in the liquid and no diffusion in the solid during freezing, they considered that isoconcentration curves in the steel were homothetic. As a result, isoconcentration curves of the various elements, contour surfaces of dendrites produced by isothermal-transformation studies, and the contour of the solid-liquid interface would all be similar, and could be superimposed by a uniform expansion or contraction of the contours. Preliminary results of the present investigation suggested that isoconcentration curves were not homothetic. It was also evident that points of maximum concentration could not be predicted from the dendrite configuration. Since the ratio of maximum to minimum concentration appeared to be an important parameter in defining the overall segregation in a casting, other techniques for determining the point of maximum concentration were investigated. In addition the overall solute distribution associated with the dendritic structure was examined for both nickel and chromium, and the consistency of the distribution for neighboring dendrites was determined. METHOD Fifty-pound heats of AISI 4340 steel were cast in a manner similar to that used by Flemings et al3 with a water-cooled copper base plate and exothermic sides and top. The nominal composition of the material was C, 0.41 pct; Mn, 0.66; Si, 0.35; S, 0.012; P, 0.010; Ni, 1.88; Cr, 0.95; Mo, 0.28. The steel solidified directionally from the chill surface, with a dendritic structure which increased in dendrite arm spacing with increased distance from the chill. Results of the general investigation of the cast structure and segregation of radioactive phosphorus in these castings have been reported elsewhere.&apos; For the present measurements, sections were cut from the casting at several distances from the chill surface. The sections were subsequently heat-treated, mounted, and then polished on a plane perpendicular to the general freezing direction of the casting. Electron probe microanalysis was carried out with the JEOL-JXA3 microprobe using a voltage of 25 kev and beam current of 0.15 µa. The take-off angle is 22 deg. Quartz crystals were used in the spectrometers and measurements made for Ka radiation from both nickel and chromium simultaneously. A thin layer of carbon was deposited on all specimen surfaces prior to examination in the microprobe. Early measurements were made on polished surfaces, with indentation marks produced by a hardness tester to indicate the area of interest. It was found desirable to use etched surfaces, as opposed to polished surfaces, to reduce sample preparation time and more readily relate the microprobe scan to the structure. To determine if this could be done, without introducing significant errors, a series of measurements were made on polished and etched surfaces of the same specimen. Both polished and etched surfaces produced similar results within normal scatter as shown in the observations. Subsequent measurements were then made on etched surfaces. Measurements with the microprobe analyzer were made by point counting, for periods of 100 sec, at 10-µ intervals. Ten-second counts were made on pure nickel and chromium standards, incorporated in the specimen mount before and after each set of measurements. The results were adjusted for background, then corrected for dead time using the expression:

Jan 1, 1969

By M. A. Rigdon, R. E. Grace

The dissociation of cupric oxide to cuprous oxide and oxygen was studied with a microbalmce technique at 700" to 750°C. In this temperature range the dissociation pressure of the reaction 2CuO= Cu2O + 1/2 O2 was: The theory of near-equilibriwm reaction rates predicts a linear relationship between reaction rate and the Gibbs free-energy change for small values of AF/RT. This correlation is found to be valid for AF/RT = -0.1 or approximately 200 cal per mole away from equilibrium. THE theory of reaction rates near equilibrium as developed from classical thermodynamics suggests a small range of linear behavior between the overall reaction rate and the free-energy change accompanying the process. The extent of this linear correlation is generally unknown, even to the order of magnitude. The idea of linearity is also found in the formal mathematics of coupled transport phenomena, e.g., chemical diffusion, thermal diffusion, thermoelectricity, and so forth. The linear array of forces and fluxes which leads to various main and cross coefficients is predicated upon linearity between reaction rates and the bulk thermodynamics which govern the process. The general case in reaction kinetics is well-known: the magnitude of the free-energy change does not cor- relate with observed reaction kinetics, irrespective of whether the process occurs by diffusion or chemical reaction or both. The real need is to investigate specific reaction paths near equilibrium in order to learn of the limitations in attempting to correlate reaction rates and thermodynamics. It is the main purpose of this paper to examine the dissociation kinetics of cupric oxide under conditions of known driving force where the dissociation process is likely to occur by chemical reaction. A secondary objective is to obtain equilibrium data for the CuO-Cu20-O2 system. In a prior paper Edmiston and race&apos; investigated reduction kinetics of wustite in H2-H2O gas mixtures. The present study was carried out in oxygen according to the following reaction:

Jan 1, 1969

By J. Szekely, R. G. Carr

A method is presented which allows the extension of the Ergun equation for the description of gas flow through packed beds in a nonisothermal regime. The pressure profile in the packed bed is calculated by the establishment of a differential energy balance and the iterative integration of the integral equation thus obtained. The predicted values of the pressure distribution were found to be in excellent agreement with those measured experimentally, under carefully controlled conditions. MANY metallurgical processes are carried out in systems in which a gas stream and particular solids are contacted in a packed bed arrangement under strongly nonisothermal conditions. Ore sintering, the firing of pellets, the operation of regenerators, blast furnace technology may be quoted as examples. Although many aspects of these systems have been extensively studied, an important variable, the pressure distribution and the relationship between pressure head and flow rate, received only little attention to date. The purpose of the paper is to present a brief survey of relevant previous work, followed by a more rigorously formulated model, together with experimental results that afford ready comparison of theoretical prediction with measurement. PREVIOUS WORK A great deal of information is available in the chemical engineering literature on flow phenomena in porous media.&apos; The experimental results pertaining to gas flow were generally interpreted in terms of models based on the analogy with the flow through a pipe of an incompressible, isothermal fluid. Of the numerous, semiempirical correlations that have been proposed to relate pressure drop to the volumetric flow rate, the Ergun equation2 is generally regarded as the most satisfactory: The strong nonisothermality of the metallurgical systems previously mentioned would render the direct applicability of Eq. [I] rather doubtful as both density and viscosity are temperature-dependent. Ergun has used Eq. [I] for estimating the pressure drop in the stack region in the blast furnace by neglecting the viscous term and evaluating the gas density at the integral mean temperature.3 His calculations showed a rather variable agreement with measurements and it was uncertain whether the discrepancy found was attributable to the model or to the inaccuracies often inherent in measurements on an industrial scale. Other work on nonisothermal systems includes studies of the pressure distribution in sinter beds, notably by Voice et a1.4 and Mitchell.5 Mitchell developed a method involving the subdivision of the bed into several zones and assigning average property values (temperature, gas density, and so forth) to each zone. The pressure drop was then calculated for each zone by using an expression analogous to Eq. [I]. This method was an improvement over that described by Voice et al., but even here prediction and measurement differed by as much as 30 pct in certain cases. It would appear that no really satisfactory methods are available for the prediction of pressure drop in packed beds with a strongly nonisothermal regime. The method involving subdivision of the bed into small enough sections, that may be regarded as isothermal, would show promise; however its validity has to be tested by both a rigorous formulation and comparison with experimental data obtained under controlled conditions. In the following a formulation is presented together with experimental data allowing ready practical evaluation of the theory. FORMULATION The mechanical energy balance over an infinitesimal section of a packed bed may be written as follows:

Jan 1, 1969

By R. W. Gould, B. G. LeFevre, A. G. Guy

The resistivity anomaly known as the K state was studied in Ni-Mo alloys containing 10.5 and 14.0 at. pct Mo. Both these alloys exhibit a large K effect which depends on the mechanical and thermal treatment. On the basis of X-ray diffuse scattering studies which were correlated with resistivity measurements, it appears that the K state in dilute Ni-Mo alloys can be associated with changes in the degree of short-range order within the a phase. An interesting phenomenon that has received much attention in recent years is the K state. The K state is marked by anomalous changes in some of the physical properties of certain alloys without the occurrence of observable microscopic structural changes. One of the early pieces of work in this area was by Thomas&apos; who studied alloys of Ni-Cr, Ni-Cu, Ni-Cu-Zn, Fe-A1, Fe-Si, and Ni-A1. Upon annealing specimens which had been previously cold-worked or quenched from an elevated temperature he found an anomalous increase in resistivity over a certain temperature range. He also found that specimens which had been appropriately annealed to develop the K state showed a decrease in resistivity upon subsequent cold working. These effects are opposite to those found in normal alloys. Although the resistivity anomaly has been rather arbitrarily taken as the "definition" of the K state, there are several other interesting effects which accompany the resistivity increase. In Ni-Cr alloys,2, 3 for example, it was found that the hardness increases with increasing resistivity. It was also found that specimens which have been treated to develop the K state can be cold-worked for as much as an 80 pct reduction of area without an increase in the hardness. In Fe-A1 and Fe-Si alloys4 the K state is accompanied by an increase in flow stress and by a lattice contraction. In Ni-A1 alloys,5 specimens which have been treated to develop the K state also show an increase in elastic modulus. In Ag-Pd alloys6 the increased resistivity observed on annealing a cold-worked specimen is accompanied by an increase in the thermoelectric power and an increase in the Hall coefficient. The explanations of the K state phenomena are varied and depend upon the particular alloy in question. Several theories have been advanced to explain the increased conductivity with cold work on the basis of changes in the electronic configuration of the alloy as a result of local lattice distortions.7"9 Most investigators, however, believe that some type of local order in the solid solution, either short-range order (SRO) or clustering, is responsible for this effect. Theories concerning the relationship between ordering and the K state have for the most part been speculative, since there is little direct X-ray evidence that can be correlated with the above property changes. Much of the previous work on the K state was done in the Ni-Cr system where the small difference in the X-ray atomic scattering factors of the components nickel and chromium makes it very difficult to use X-ray diffuse-scattering measurements to determine the role of local order. In the Ni-A1 system, however, Starke et al.10 succeeded in detecting a connection between local order and the K state. It was found that a small but measurable K effect correlated with increasing SRO in the nickel-rich a phase. The manner in which local order might increase the resistivity of K state alloys is not completely clear. Since most of the known K state alloys contain at least one transition element, significance has frequently been attached to the presence of an unfilled d shell. It has been suggested that during the formation of the K state the number of conduction electrons decreases as a result of the transfer of s electrons to the d shell where they are more tightly bound.1&apos;11&apos;12 Koster and Rocholl13 have proposed that SRO can cause an increase in resistivity for alloy systems in which the number and mobility of charge carriers are reduced when the percent solute is increased. According to this hypothesis, the local environment of a given solvent atom changes in the same manner with increasing percent solute as it does with an increasing degree of SRO; hence the change in physical properties should tend in the same direction. In this hypothesis, SRO is considered only in a statistical sense, and the increased resistivity of the K state is attributed to a change in the mean distribution of electrons and holes in the s and d states as a result of SRO. From the work of Chen and Nicholson on Ag-Pd alloys,6 it appears that the K state can occur in systems for which the d shell is completely filled. These investigators explained the increased resistivity by picturing the SRO as small domains of some form of long-range order (LRO). According to ~ibson,&apos;~ the number of effective electrons can be reduced by the creation of a new Brillouin zone boundary near the Fermi surface of an alloy as a result of the changing crystallographic symmetry that accompanies the formation of a superlattice. This idea may be expressed in terms of the superzone concept.15 In the present work the role of local order in the formation of the K state in Ni-Mo alloys was investigated. The principal tools used in this study were X-ray diffuse scattering and electrical resistivity measurements; however, these data were supplemented by electron microscopic and field-ion microscopic data. The purpose of the work was to determine whether or not the K state in Ni-Mo alloys can indeed be attributed to the formation of SRO as has been proposed by previous investigators.

Jan 1, 1969

By Claude H. P. Lupis, John F. Elliott

The classic Siez~erts&apos; technique has been employed to determine the solubility of oxygen in liquid pure silver and the effect on the oxygen solubility of solute metal additions to liquid silver. The results indicate that oxygen deviates slightly from Sieverts&apos; law in pure silver as = 3.9. Gold, platinum, md palladium decrease the solubility of oxygen in silver. First-order free energy, enthalpy, md entropy interaction coefficients are determined from the data. An important aspect in the establishment of the formalisms to be applied to represent the thermody-namic behavior of components of solutions is the accuracy of the experimentally determined thermody-namic properties. Wagner developed the first-order free energy interaction coefficient.&apos; Recent papers2&apos;3 extend the treatment formally to other thermodynamic properties and define the second- and higher-order interaction coefficients. A current communication4 considers in some depth the requirements and limitations imposed because of experimental error on the use of several formalisms now under active consideration in the field. This communication reports the results of an experimental study of the behavior of oxygen in liquid silver alloys that was carried out concurrently with the development of Refs. 2 and 3 with the hope of providing sufficiently precise information so that enthalpy and entropy interaction coefficients could be determined. THE EXPERIMENTS The Sieverts technique was employed using an improved design of the reaction bulb over that developed originally by Humbert5 and subsequently improved by pehlke6 and Weinstein.7 The Vycor reaction bulb as shown in Fig. 1 was air-cooled and contained a port for the thermocouple angling off to one side and a vertical port for connections to the gas supply and

Jan 1, 1969

By V. Koump, T. F. Perzak, R. G. Olsson

The rate of dissolution of recrystallized alumina in molten iron oxide in equilibrium with iron at 1450°C was studied by rotating alumina disks in the melt. The samples were rotated from 1 to 12 min at speeds from 54 to 3270 rpm. It is concluded from the experimental results that the process is limited by diffusion in the liquid boundary layer. The interdiffusion coefficient for the dissolution of alumina disks in pure molten iron oxide was estimated to be about 3 xl0-5= sq cm per sec. THE erosion of refractory materials is of considerable importance in iron- and steelmaking processes. Since many refractories contain alumina, this investigation was undertaken to develop a further understanding of the erosion mechanism. Shurygin et al.1-3 have investigated the rate of dissolution of alumina in molten silicates and molten fluorides by rotating alumina disks in various melts. The rates were reported to be limited by diffusion in the liquid boundary layer. In the present investigation the rate of dissolution of alumina in molten iron oxide in equilibrium with iron was studied in a similar manner; alumina disks were rotated at different speeds in molten iron oxide at 1450°C. The relationship between the rate of dissolution of the flat surface of the disk and the speed of rotation was used to determine whether the process is limited by diffusion in the liquid boundary layer or by some chemical reaction mechanism. APPARATUS AND EXPERIMENTAL PROCEDURE A schematic diagram of the experimental apparatus is shown in Fig. 1. The iron oxide melts were contained in 2-in.-diam high-purity iron crucibles that were centered in a cylindrical graphite susceptor. Prior to each sequence of experiments, sufficient reagent-grade ferric oxide was added to a crucible to form approximately 350 g of molten iron oxide; the depth of the melt was about 11/2 in. To prevent excessive erosion of the crucible, about 70 pct of the required iron was added to the crucible in the form of iron powder. The mixture was melted in an atmosphere of purified argon by induction heating with a 250-kc generator. After the oxide was molten, a Pt/Pt-10 pct Rh thermocouple at the exterior of the crucible was calibrated against an iron-sheathed thermocouple immersed in the melt and was thereafter used to regulate the melt temperature. All experiments were conducted at 1450°C with a probable error of 10°C. The alumina disks were constructed from pure re-crystallized alumina having a density of 3.7 g per cu cm (porosity approximately 6 pct). For each disk sample a 1-cm-diam alumina cylinder of known length (0.9 to 1.0 cm) was cemented into a ground seat at the end of a 1.27-cm-OD centerless ground alumina tube. The ends of the tube and cylinder were flush and served as the disk surface. When mounted in the apparatus the alumina tube was guided by tungsten carbide bearings and rotated by a variable-speed motor. The bearing and drive assembly could be raised and lowered in order to allow the sample to be quickly moved in and out of the melt. At the start of an experiment, a sample was preheated for several minutes directly above the melt, and then immersed in the middle of the melt to a depth of a in. and rotated for 1 to 12 min at speeds of 64 to 3270 rpm. The speed of rotation was measured by a hand tachometer. At the start and completion of each experiment, a melt sample was withdrawn with a cold copper rod and the alumina content was determined by chemical analysis. During individual experiments the alumina concentration of the melt changed by the order of 0.25 wt pct. Average melt compositions for individual experiments were used in the subsequent computations and were in the range between 0.12 and 4.3 wt pct alumina. The extent of dissolution was determined by sectioning the end of the sample and measuring the final length of the alumina cylinder with a microscope. Since the exposed end of the cylinder, i.e., the disk surface, was etched in a manner characteristic of this geometry, a number of measurements were made on

Jan 1, 1969

By C. A. Stickels, C. M. Yen

Secondary recrystallization was investigated in vacuum-melted electrolytic iron to which 70 pm N was vacuum-meltedadded. The secondary texture is "near {554}<225>" for material cold-rolled 75 to 90 pct, the sharpness of the texture increasing with increased rolling reduction and with decreased annealing temperature. At reductions of 95 and 97.5 pct the secondary texture is &apos;"near {322)(296)". Both secondary orientations also exist as major components of the primary re-crystallization texture. Development of a strong "near {554) (225)" secondary texture appears to depend on the evolution of the Primary texture to a transition texture depleted in orientations near the secondary orientation before the onset of secondary growth. A variety of qualitative experinzents have been used to show that nitrogen is important in limiling primary grain growth. The presence of nitrogen does not seem essential for the establishment of a transition texture, but a loss of nitrogen during annealing may facilitate growth of grains in the secondary orientation. Secondary grains we shown to form initially at the specimen surface. This is not thought to indicate that surface energies are important in the growth process. It is proposed that the quasi-two-dimensional character of surface grains permits discontinuous growth parallel to the surface before secondary growth of interior grains is possible. An earlier study of recrystallization textures in 90 pct cold-rolled electrolytic iron showed that secondary recrystallization occurred after annealing for several days at 700C1 This type of secondary recrystallization, which had not been reported previously, results in the formation of a strong texture, best described by the indices "near {554}(225)". The purpose of the present work was to investigate the effect of various processing variables on secondary recrystallization in this material and determine the mechanism of secondary grain growth. LITERATURE REVIEW An understanding of the mechanism of a secondary recrystallization process depends on knowing: 1) how grains in the secondary orientation come to be in the primary recrystallization texture; 2) why normal grain growth does not occur; and 3) what factors determine the strength of the secondary texture. For secondary growth of grains of a particular orientation, a certain minimum fraction of the grains must be in that orientation after primary recrystallization. This requirement is apparently satisfied "naturally" in certain systems, i.e., when the primary texture obtained by rolling and recrystallizing material initially randomly oriented contains a sufficient fraction of primaries in the secondary orientation. However, in other cases, e.g., {110}<001> and {100}<001> secondary growth in silicon iron,2 it is necessary to enhance the fraction of primary grains in the secondary orientation by rolling and recrystallizing textured material. In the present case, the "near {554}<225>" orientation is contained within the spread of orientations found in the primary recrystallization texture of iron or bbc iron-base alloys. In systems where the main driving force for secondary growth is the reduction in total grain boundary energy, secondary growth is observed only when normal grain growth is minimized. Four ways in which normal grain growth can be limited are: 1) Limitation by a strong primary texture. When a very strong primary texture consisting of a single component or twin-related components develop, most primary grains are separated from one another by relatively immobile small-angle grain boundaries. The classic instance of this is secondary growth into the primary cube texture in some fcc metals. 2) Limitation by precipitates. Precipitates present in the proper volume fraction with a suitable dispersion will limit primary grain growth. The role of MnS inclusions in impeding normal grain growth in Si-Fe is well-documented.5 3) Limitation by sheet thickness. Normal grain growth slows drastically when the mean grain diameter is of the same order as the sheet thickness. This effect has been used to obtain secondary recrystallization in thin sheets of high-purity silicon iron.&apos; 4) Limitation by solute impurities. It is well-established that certain impurity elements in solution can have a large effect on grain boundary mobility.&apos; However, there does not seem to be any secondary recrystallization process in which primary grain size stabilization has been shown to be due to the drag exerted on grain boundaries by dissolved impurities. In certain systems, e.g., secondary recrystallization in silver,&apos; the means by which normal grain growth is limited has not been identified, and solute-impurity limitation might be suspected. In order to understand the factors which determine secondary texture strength in three-dimensional growth, it is necessary to examine in more detail the current picture of general secondary recrystallization processes. Following Cahn,9 it is assumed that the primary grains have a range of sizes and that secondary growth of one of the large grains in this distribution is possible when it exceeds a critical size with respect to its neighboring grains. The critical size depends on the ratio ?S/?p, where ?s is some sort of average grain boundary energy between the potential secondary and the primary grains and ?p is some sort of average grain boundary energy between primary grains. For a constant primary grain size, the critical size for secondary growth increases as ?$/?p increases. May and Turnbull5 have incorporated the

Jan 1, 1969

By D. E. Coates, G. R. Purdy, S. V. Subramanian

The morphological stability of the planar solid-liquid interface in dilute ternary alloys, undergoing steady-state unidirectional solidification, is analyzed in terms of both the constitutional supercooling principle and the perturbation methods recently developed by Mullins and Sekerka. First, various steady-state solutions for the two solute distributions ahead of a planar interface are examined. The nature of the solutions depends on the size and concentration dependence of the off-diagonal diffusion coefficients. W~thin the framework of the constitutional supercooling principle, a cumulative contribution to instability frorn the two solutes is found to exist in the absence of diffusional interaction. It is shown that the latter can produce a further enhancement of instability or can have a stabilizing influence, depending on the form of the liquidus surface and on the sign of the solute-solute interaction. A perturbation analysis, which ignores diffusional interaction, verifies the cumulative influence of lhe solute fields and demonstrates that the Mullins-Sekerka stability criterion for binary systems (with capillarity accounted for) can be readily extended for application to ternary systems. SOME time ago, Tiller et al.&apos; calculated the solute concentration distribution ahead of the planar solid-liquid interface of binary alloys undergoing steady-state unidirectional solidification. An earlier qualitative proposal that the transition from planar to nonplanar growth morphologies is associated solely with the onset of constitutional supercooling in the liquid layer ahead of the moving interface2 was used in conjunction with this calculation to put the now well-known constitutional supercooling (C-S) stability criterion into quantitative terms. Mullins and Sekerka,3 in a recent and very elegant analysis, established a more complete criterion (hereafter referred to as the M-S criterion). Interfacial stability was investigated by determining the time derivative of the amplitude of a sinusoidal perturbation of infinitesimal amplitude which had been introduced into the originally planar shape of the moving interface. Of particular importance is the fact that capillarity was included in the boundary conditions of their calculation. The purpose of the present paper is to extend all of this earlier work on dilute binary systems for application to dilute ternary alloy solidification. The analysis is divided into three sections. In the first the two solute distributions ahead of a moving planar interface are considered. Mathematical solutions are de- termined for situations in which: a) diffusional interaction is negligible, 6) diffusional interaction must be considered but circumstances permit use of constant diffusion coefficients, and c) the concentration dependence of off-diagonal diffusion coefficients can be described by first-order dilute solution approximations. In the next section, a stability criterion analogous to the C-S criterion is developed and the influence of diffusional interaction on interface stability is analyzed. Finally, the perturbation formalism of Mullins and Sekerka, with capillarity included in the boundary conditions, is extended for analysis of ternary systems in which diffusional interaction is negligible. The study of interface stability in binary systems usually commences with the assumption that the equilibrium distribution coefficient and the slope of the liquidus line are constant at values corresponding to infinite dilution. Similar assumptions have not been introduced into the present treatment; that is, we do not assume planar solidus and liquidus surfaces joined by tie lines which yield constant distribution coefficients. The latter involves the assumption of no ther-modynamic interaction between solute species in both the solid and liquid. We consider a ternary phase diagram for which the solidus and liquidus surfaces are, in general, nonplanar and of course pass through the corresponding binary solidus and liquidus lines. These lines are not assumed to have constant slope. In the dilute regions we are concerned with, the following assumptions are made: i) The solidus and liquidus surfaces are of a form such that both the solidus and liquidus temperatures are monotonically varying functions of each solute concentration. ii) The tie lines are such that the equilibrium distribution coefficient of a given solute is greater than unity for every point on the solidus (or liquidus) surface or it is less than unity for every point. STEADY-STATE SOLUTE DISTRIBUTIONS IN THE LIQUID As will be demonstrated in the next section, a knowledge of the steady-state solute profiles is not a necessary prerequisite for the formulation of a ternary C-S stability criterion. However, in that details, such as the complete description of the equilibrium liquidus temperature profile, require an evaluation of the solute distributions, the overall treatment is enhanced if these distributions are determined. Consider a ternary system (solvent plus solutes 1 and 2) for which a planar solid-liquid interface is in unidirectional motion at constant velocity V. At this stage it is unnecessary to limit ourselves to dilute solutions. For a stationary frame of reference the generalized forms of Fick&apos;s equations are:

Jan 1, 1969

Jan 1, 1969

By Richard R. Zupp, David A. Stevenson

A theoretical treatment of austenitic Fe-C-Z solutions (where Z represents a substitutional solute) was developed by extending a solution model which was previously applied to austenitic Fe-C alloys. The model assumes that each carbon atom in the solution excludes H interstitial sites from occupancy by other carbon atoms rmd thai otherwise the assembly of carbon atoms is equivalent to a modified Langmuir lattice gas. The results of the treatment provide theoretical justification for a phenomenological expression previously developed to represent the influence of substitutional solutes on the thernlodynamics of carbon in austenitic Fe-C-Z alloys. The expression is: where Kz is a parameter which characterizes the Fe-C-Z system at a given temperature and Xz and xj are the concentrations of substitutional solute and interstitial carbon expressed as atom ratios. The parameter, Kz, was associated with the differ- A number of investigators have measured the influence of various substitutional solutes on the carbon activity in austenitic Fe-C-Z alloys, where Z represents any substitutional solute. Experimental information is available for alloys containing the following substitutional solutes: Mn,&apos; Si,&apos; Ni,2-5 CO,&apos;-&apos;, Cr,5, 9-11 CU,4 Al,= v,11, 12 and Ti.13 This information is essential for studying many phenomena in Fe-C-Z alloys, particularly diffusion and diffusion-controlled phase transformations For example the pffprts nf various ence between the interaction energies between carbon atorns and host-lattice atoms in the Fe-C-Z solution and the corresponding energies for the Fe-C solution. The carbon activity in a ternary Fe-C-Z alloy, ac, was obtained in terms of the composition of the alloy and the single parameter, Kz: where y° is the Henry&apos;s law constant for the binary Fe-C solution. Both y° and H crm be determined from the data for Fe-C solutions, without recourse to any experimental data for Fe-C-Z solutions. The theoretical treatment was extended to alloys containing several substitutional solutes. By neglecting interactions between the substitutional solutes, an expression for the carbon activity in multicomponent alloys was obtained in terms of the Kz values characterizing ternary Fe-C-Z systems. substitutional solutes on the rate of pearlite growth is explained14 on the basis of their effects on carbon diffusion, which result from their influences on the carbon activity in the solid phases. In addition, recent studies have emphasized the dependence of the pro-eutectoid ferrite transformation15 and other diffusional processes in austenite16 on the influence of substitutional solutes on the carbon activity in austenite. They have indicated the need for suitable analytical expressions to represent such influences. The present authors previously developed general phenomenological expressions to represent the equilibrium partitioning of carbon between Fe-C alloys and Fe-C-Z alloys.&apos;2 The expressions are applicable to all the austenitic Fe-C-Z alloys for which experimental data are available, and are sufficiently simple to recommend their general application. For alloys in

Jan 1, 1969

By Owen F. Devereux

Two discrete types of microstructure me seen in cermets formed by the reaction: 4Al (liquid) + 3SiO2 (glass) — 2A12O3(a) + 3Si (solid) Relatively coarse-grained macroscopically isotropic specimens resulted from reaction with aluminum containing appreciable silicon, while macroscopically anisotropic fine-grained specimens were formed by reaction with aluminum low in silicon. Optical microscopy, electron microscopy, and X-ray diffraction suggest that the anisotropic material contains gross areas in which submicroscopic grains of a A12O3 are oriented along radial lines drawn from the specimen surface. These lines are presumed to originate at defects in an otherwise protecting surface layer. A unique cermet is formed by the reduction of high-silica glass by molten aluminum according to the reaction: 4A1 (liquid) + 3SiO2 (glass) — 2A12O3 (a) + 3Si (solid) The free-energy change accompanying this reaction is -150 kcal at 800°C; the volume change (volume of A12O3 and silicon minus volume of SiO2) is +6 pct. The reaction product, however, contains aluminum metal as well as Al2O3 and silicon because some of the product silicon is dissolved in the aluminum bath and is replaced by aluminum in the structure. The product is a hard gray solid that retains the external dimensions of the original silica article. The system has been studied in the past, the investigations culminating in patents to George1, 2 and Saw-chuk., The former emphasize that this reaction is but one of a relatively broad class of such systems, while the latter demonstrates that a marked improvement in the gross physical properties of the cermet can be obtained by dissolving appreciable quantities of silicon in the aluminum reactant. The system has also been studied by workers at Rolls Royce Ltd.,4-6 who have patented the addition of small quantities of arsenic, antimony, bismuth, or phosphorus to the reactant bath to SO impede the reaction that glass fibers may be metal-coated by immersion in molten aluminum.

Jan 1, 1969

By Shiro Ban-ya, John Chipman

Equilibrium in the reaction was investigated at temperatures of 1500°, 1550°, and 1600°C for sulfur concentrations up to 7.2 wt pct. Multisample crucibles contained the liquid alloys in a resistance-heated furnace using a technique especially designed for the study of more complex alloys to be reported separately. Modern free-energy data are used to correct the H2S:H2 ratio for dissociation of H2S and calculalion of the partial pressure of S2. Published data on the equilibrium are similarly corrected. Thermodynanzic treatment of the data employs the composition variable zs = nS/(nFe — nS) and the activity coefficient Gs = as/zs The data at 1500" and 1550°C are fitted by the equation log s = —2.30zs. Within the limits of experimental error the same coefficient is applicable to the data at higher temperatures. Equations are given for the free-energy change in Reaction [I] as well as for the solution of S, gas in the metal. The heat of solution of 1/2 s2 is -32.28 i2.5 kcal. Uncertainty in the free energy is very much smaller. For dilute solutions of interest in steelmaking, the activity coefficient of sulfur is unchanged from that listed in Basic Open Hearth Steel-making. DETERMINATIONS of the thermodynamic properties of sulfur in liquid iron by Morris and williams1 and by Sherman, Elvander, and chipman&apos; provided a basis for control of sulfur in steelmaking processes. From the standpoint of understanding the chemistry of metal plus nonmetal in liquid solution they left several questions unanswered. The activity of sulfur in dilute solution at about 1600°C was well-established but temperature coefficients were uncertain, due at least in part to the use of the optical pyrometer and uncertainty regarding the effect of sulfur on emissivity. It appeared that deviation from Henry&apos;s law increased with increasing temperature, a most unusual behavior requiring either confirmation or disproof. These studies were based on experimental determination of equilibrium in the reaction: At high temperatures H2S is partially dissociated so that the gas mixture contains HS, S2, and S in addition to HS. At the time of the earlier studies the free energies of these constituents were unknown and it was therefore impossible to make adequate correction for dissociation. Observations on the effects of alloying elements by Morris and coworkers1, 3 and by Sherman and Chip-man4 enable us to assess the effects of alloying elements on the activity and to make corrections for incidental impurities in the binary liquid. These studies as well as a number of more recent investigations will be reviewed in detail after out own experimental results have been presented. It was our purpose in planning this study to avoid uncertainties regarding the emissivity of alloys and the errors of thermal diffusion which plagued some of the early attempts,5 by using a resistance furnace and thermocouple in preference to induction heating and optical pyrometer. Modern data on free energies of the gaseous species are to be applied to our data and to those of other investigators to obtain corrected values of K1 and of the activity coefficient and ultimately to relate the sulfur content of the bath to the equilibrium partial pressure of S,. Extension of the study to include ternary and complex solutions will be described in a later section. EXPERIMENTAL METHOD a) Preparation and Calibration of H2-H2s Gas Mixture. The source of hydrogen sulfide was a preparer mixture of 43 pct H2S, balance hydrogen, contained in a large aluminum cylinder. This was passed through anhydrone and through a microflowmeter. Hydrogen was passed through platinized asbestos, ascar-ite, and anhydrone, and through a capillary flowmeter. Argon was passed through copper wool at 500°C, then through ascarite, anhydrone, and a flowmeter. The flow rate of hydrogen was kept constant at 200 ml per min, to which an arbitrary amount of the hydrogen-hydrogen sulfide mixture was constantly added and then the prepared gas mixture was introduced into the reaction tube through a gas mixer. In certain experiments 200 ml per min of argon was added to the hydrogen-hydrogen sulfide gas mixture to increase the total flow rate of gas. The ratio of hydrogen-hydrogen sulfide in the inlet gas was checked for each run by chemical analysis. A sample of the gas taken from a bypass was bubbled through zinc and cadmium acetate solution (4 pct zinc acetate, 1 pct cadmium acetate, and 1 pct acetic acid) to remove hydrogen sulfide from the gas mixture, and the flow rate of the remaining hydrogen was measured by a soap bubble method to determine the volume of hydrogen. The amount of hydrogen sulfide absorbed in solution was determined by titration with iodine against sodium thiosulfate, with starch used as the indicator. The ratio of hydrogen sulfide to hydrogen in the inlet gas could be kept within ±2 pct in the range from 10-2 to 5 x 10"4 which corresponds to from 0.2 to 7.0 wt pct sulfur in liquid iron. b) Furnace Arrangement. Fig. 1 shows the furnace arrangement and the shape of the alumina crucible used in this experiment. A vertical-tube silicon carbide electric resistance furnace contained the reaction tube which consisted of two parts, the gas-tight

Jan 1, 1969

By M. A. Dayananda, R. E. Grace, P. F. Kirsch

Ternary diffusion experiments were carried out at 750°C with vapor-solid diffusion couples in single-phase copper-rich Cu-Zn-Sn alloys. The Philibert-Guy method was used to calculate intrinsic diffusion coefficients at compositions corresponding to the plane of inert markers imbedded in the diffusion couples. Inter diffusion coefficients were determined at common composition points of intersecting diffusion paths. Intevdiffusion coefficients were also obtained at compositions corresponding to maxima in the tin concentration pvofiles Relations between the interdiffusion coefficients and the expansion of vapor-solid diffusion couples are described in detail. Such relations permit the calculation of constant interdiffusion coefficients in binary couples from expansion data directly. In ternary couples expansion data provide an additional constraint on the interdiffusion coefficzents. ThE phenomenological analysis of diffusion in ternary solid solutions has been developed extensively, and various approaches taken in the determination of diffusion coefficients have been reviewed recently by Guy et al.&apos; Finite,&apos; semi-infinite,3 and infinite4 diffusion couples have been used in studies of ternary diffusion, but most of the work has been done with infinite diffusion couples. Semi-infinite diffusion couples3 (vapor-solid couples) have been used to determine both intrinsic and interdiffusion coefficients in the Cu-Zn-Mn system, in which two diffusing species, zinc and manganese, were supplied from the vapor phase. If only one component has a sufficiently high vapor pressure to be supplied from a vapor source, vapor-solid diffusion couples may still be used. In this case the boundary conditions at the gas/solid interface may be governed by partial equilibria rather than by virtually complete equilibrium between the phases. In the present work diffusion measurements were carried out at 750°C with vapor-solid couples of Cu-Zn-Sn alloys. Only zinc was transported between the vapor source and the diffusion disks, because the vapor pressures of both copper and tin are at least 108 lower than that of zinc. Both intrinsic and interdiffusion measurements were made at several compositions. A secondary objective was to relate expansion of the diffusion couples with the interdiffusion process. Such measurements for binary couples yield average values of the interdiffusion coefficients over composition increments. In ternary diffusion couples ex- pansion data provide an additional constraint on the interdiffusion coefficients. COMPOSITION-DEPENDENT INTRINSIC AND INTERDIFFUSION COEFFICIENTS As shown by Philibert and GUy5 the determination of intrinsic diffusion coefficients in a ternary system requires two independent diffusion couples containing marker planes with identical composition. In the case of vapor-solid couples such a requirement torresponds to two couples set up with the same vapor source exposed to two diffusion disks of different composition. Inert markers placed initially on the disk surfaces become imbedded during diffusion. If the pair of couples exhibits appreciable marker burial with marker planes of virtually the same composition, the intrinsic diffusion coefficient may be evaluated at the composition of the marker planes. Fig. 1 is a schematic diagram of the concentration of of a Cu-Zn-Sn ternary vapor-solid couple after diffusion of zinc from a vapor source into a binary Cu-Sn alloy. The areas, Ai correspond to the cumulative intrinsic fluxes of the three diffusing species across the marker plane:5 where Ji is the intrinsic flux of component i, t is the diffusion time, and D3i are the intrinsic diffusion coefficients. For the determination of interdiffusion coefficients a pair of diffusion couples with intersecting diffusion paths is needed. At the composition point of intersection (C1, C2) the coefficients are evaluated from the equation:3

Jan 1, 1969

By Nestor Torres-Acuña, Fathi Habashi

Metallic copper of purity equal to commercial electrolytic copper is deposited during the anodic dissolution of technically available white metal, Cu2S, in m acidic solution of&apos; copper(II) sulfate as electrolyte and copper sheets as cathodes. The process may be applied on commercial scale since it has the advantage of by passing the converting and poling steps, decreasing pollution problems due to SO, and recovering the sulfur of the white metal in the elemental form. The anodic dissolution is believed to take place in three steps: The working conditions that lead to 98 pct cathodic current efficiency are: 30 gpl Cu2+ and 100 gpl H2SO4 in electrolyte, current density 1.0 amp per sq dm, and tetnperature of bath 40°C&apos;. DURING the smelting of copper, nickel, lead, and zinc sulfide ores, large amounts of sulfur dioxide are produced. Because the release of this gas in the atmosphere causes pollution problems, many attempts were made in the past for its recovery either in liquified form or processed to sulfuric acid. Although these two forms of sulfur are the usual forms that are consumed by the chemical and metallurgical industries, their storage and transport usually raise economic problems. It was realized, long ago, that a process by which sulfur can be recovered directly in the elemental form would be most attractive. Elemental sulfur is easily stored and transported, and is readily converted to SO2, H2S, or H2SO4 when needed. Processes aimed to recover the metals from their sulfide ores, and at the same time by-product sulfur in the elemental form, were recently reviewed.&apos; As early as 1882, Marchese2 patented a process in which a matte having the composition: Cu 15, Pb 14, Fe 41, and S 25 pct was electrolyzed. According to cohen3 the process was applied on a pilot scale by the Aktiengesellschaft fur Bergbau-, Blei-, und Zinkhiitten-betriebe at Stolberg, Rheinland, Germany. The process was unsuccessful, however, owing to the contamination of the electrolyte during electrolysis. Studies in this field were reported by Bernfeld4 and Egli.5 Borchers et a1.6-8 developed the process further by using white metal (Cu 78.2 pct, S 19.6 pct) instead of the matte, and a pilot plant was operated by the Mansfeldschen Kupferschiefer Co. at Eisleben, German~.~ It appears, however, that the process never went into commercial operation. Further studies in this direction were carried out by Russian workers.10-23 The International Nickel Co. of Canada erected a pilot plant at Port Colborne in 1951 to apply the process to nickel matte which is essentially pure nickel sulfide."&apos;- 26 Full-scale plant went into operation at the Thompson Refinery in Manitoba in 1964.27 In this method nickel sulfide matte is melted and cast into anodes analyzing 76 pct Ni and 20 pct S with small amounts of copper, cobalt, and iron. These anodes are then electrolyzed in an aqueous solution of NiS04 at pH 4. Further work on this method was reported in the literature.28-31 In this paper, experiments on the anodic dissolution of white metal in aqueous copper sulfate solution are reported. EXPERIMENTAL White metal, Table I, obtained from The Anaconda Co., was cast in form of anodes 10 by 7.5 by 1 cm, Fig. 1. The anodes are mechanically strong and do not break easily; when polished they gave a shiny gray metallic appearance. From the chemical composition shown in Table I, it can be seen that the atomic ratio Cu:S corresponds approximately to the formula CU2.06S. The copper in excess to the formula Cu2S is most probably present as a separate phase since metallic copper could be seen under the microscope, Fig. 2. Fig. 3 shows the X-ray diffraction pattern of white metal which matches fairly well with that of reagent-grade Cu2S.

Jan 1, 1969

By H. R. Thresh, A. F. Crawley, D. W. G. White

The densities of liquid tin, lead, and Sn-Pb alloys have been measured over a range of temperature above the liquidus. In all cases, data can be adequately represented by an equation of the type p = a + bT. For pure tin, the values of a and b are, respectively, 7.139 and 7.125 x 10-4, and for pure lead, 11.060 and 12.220 x 10-4. At constant temperature, molar volumes in the Sn-Pb system vary linearly with composition. DURING the course of a long-term program at the Department of Energy, Mines and Resources, Ottawa, Canada, the densities of liquid tin, lead, and Pb-Sn alloys were studied. The work reported herein draws on two sets of results independently achieved by Thresh, on the one hand, and by Crawley and White, on the other. EXPERIMENT The method and apparatus were originated by Thresh and have been described in detail in Mines Branch reports.172 In operation, the 14 cu cm graphite pycnometer is put in the furnace as shown in Fig. 1 after a machined charge has been placed inside the pycnometer and the lid screwed down. The dimensions of the charge are such that, upon melting, a small excess of liquid metal is discharged through the capillary hole to displace the weighted float above. When the pycnometer has been equilibrated at the experimental temperature, the

Jan 1, 1969