Search Documents

By J. B. Cheatham, J. W. McEver

A laboratory investigation of the behavior of casing subjected to salt loading indicates that it is not economically feasible to design casing for the most severe situations of nonuniform loading. When the annulus is completely filled with cement, casing is subjected to a nearly uniform loading approximately equal to the overburden pressure, and, although the modes of failure may be different, the design of casing to withstand uniform salt pressure can be computed on the same basis as the design of casing to withstand fluid pressure. Failure of casing by nonuniform loading in inadequately cemented washed-out salt sections should be considered a cementing problem rather than a casing design problem. INTRODUCTION Casing failures in salt zones have created an interest in understanding the behavior of casing subjected to salt loading. The designer must know the magnitudes and types of loading to be expected from salt flow and he must be able to calculate the reaction of the casing to these loads. In the laboratory study reported in this paper, short-time experimental measurements of the load required to force steel cylinders into rock salt are used as a basis for computing the salt loading on casing. These results must be considered to be qualitative only since rock salt behaves differently under down-hole and atmospheric conditions and also may vary in strength at different locations. The beneficial effects of (1) cement around casing, (2) a liner cemented inside of casing, and (3) fluid pressure inside of casing in resisting casing failure are considered. ROCK SALT BEHAVIOR UNDER STRESS The effects of such factors as overburden loading, internal fluid pressure, and temperature on the flow of salt around cavities have been studied extensively at The U. of Texas. Brown, et al.1 have concluded that an opening in rock salt can reach a stable equilibrium if the formation stress is less than 3,000 psi and the temperature is less than 300°F. At higher temperatures and pressures an opening in salt can close completely. These results indicate that calculations based upon elastic and plastic equilibrium for an open hole in salt should be applied only at depths less than 3,000 ft. In most oil wells the tem- perature will be less than 300F in the salt sections, therefore no appreciable temperature effects are anticipated. Serata and Gloyna2 have reported an investigation of the structural stability of salt. .They assume that the major principal stress is due to the overburden. Other stresses can be superimposed if additional lateral pressures are known to be acting in a particular region. In the present analysis an isotropic state of stress is assumed to exist in the salt before the hole is drilled, since salt regions are generally at rest. This assumption is partially verified from formation breakdown pressure data taken during squeeze-cementing operations in salt. Experimental measurements of the elastic properties of rock salt indicate a value of 150,000 psi for Young's modulus and a value of approximately 0.5 for Poisson's ratio. A value of % for Poison's ratio with finite Young's modulus would indicate that the material was incompressible. Values ranging from 2,300 to 5,000 psi have been reporteda for the unconfined compressive strength of salt. These variations may be due to differences in the properties of the salt from different locations or at least partially to differences in testing techniques. Salt is very ductile, even under relatively low confining pressures. For example, in triaxial tests reported by Handin3 strains in excess of 20 to 30 per cent were obtained without fracture. When casing is cemented in a hole through a salt section, the casing must withstand a load from the formation if plastic flow of the salt is prevented. To determine the forces which salt can impose on casing, circular steel rods were forced into Hockley rocksalt with the longitudinal axis of the rods parallel to the surface of the salt. The force required to embed rods 0.2 to I in. in diameter and 1/2 to 1 in. long to a depth equal to the radius of the rods was found to be F/DL =28,700 psi (± 3,700 psi) , .... (1) where D is the diameter, and L is the length of the rod. CASING STRESSES Since an open borehole through salt at depths greater than 3,000 ft will tend to close, cemented casing which prevents closure of the hole will be subjected to a pressure approximately equal to the horizontal formation stress after a sufficiently long time. As a first approximation the horizontal stress can be assumed to be equal to the overburden pressure. This is in agreement with the suggestion by Texter4 that an adequate cement job can prevent plastic flow of salt and result in a pressure on the casing approximately equal to the overburden pressure. He also advocated drilling with fully saturated salt mud

Jan 1, 1965

By J. B. Cohen, S. L. Sass

The nucleation process for y', Ni3Ti, is shou'n to change from heterogeneous to uniform as the undercooling within the phase boundary increases. As unifornz nucleation beconres copious, (100) alignment is observed in the early stages of aging; howecer. whether or not the alignment is due to spinodal decomposition could not be determined. The presence oi a large-scale reversible segregation of titanium and not the intermediate precipitate y' may be the cause of the initial strengthening in this alloy system and the discrete regions detected magnetically by Ben-Israel and Fine. Evidence is presented for the role of dislocations in the formation at long aging tines oJ the equilibrium precipitate . ThE nickel-rich portion of the Ni-Ti phase diagram consists of a terminal fcc solid solution of titanium in nickel called y.' For temperatures below 1290°C if the solubility limit is exceeded and the temperature is high enough, the excess titanium will combine with nickel to form q, Ni3Ti, a four-layer hcp ordered structure with stacking sequence ABAC."~ Since a change in structure is involved, formation of q is slow and a metastable precipitate referred to as y' forms. This metastable precipitate is an ordered cubic phase in the unconstrained state having a structure similar to Cu3Au (LIZ) with the composition Ni,Ti."'5 From X-ray studies,8, 7 the following sequence is indicated for the precipitation reaction at low temperatures: 1) The appearance of satellites around low-index diffraction spots and splitting of high-index spots, indicating the formation of a slightly tetragonal phase in an imperfectly periodic arrangement. The tetragonal phase is y' or the depleted matrix, depending on the volume fractions of matrix and precipitate (and hence, the alloy composition); tetragonality apparently arises from coherency strains. 2) The appearance of diffraction spots from the equilibrium second phase, q. Ben-Israel and Fine'" used magnetic methods to measure the changes in matrix composition during aging, and to estimate the precipitate composition in a Ni-10.1 at. pct Ti alloy. They observed that the alloy, solution-treated at 1270°C and quenched, contained heterogeneities with discrete compositions which gradually vanished upon aging. It was suggested that y' may initially form as a defect structure with several possible compositions deficient in titanium as compared to Ni3Ti; the composition Ni3Ti develops with prolonged aging. The initial composition was Ni6Ti which would require a very large deviation in stoichiometry for y'. It was noted that at 700°C the precipitation reaction was 80 pct complete after 2 hr and that the volume fraction of second phase was 0.2 after 1 hr. Mihalisin and Decker' suggested that the equilibrium precipitate, besides being nucleated at grain boundaries, could also form at stacking faults. ~errick" examined the formation of in Ni-Cr-Ti, and suggested that a collapsed vacancy cluster within y' acted as a nucleus for intragranular q at high temperatures, 820" and 900°C. After long aging times at lower temperatures, 650" and 750°C, diffraction patterns taken from regions containing ribbons of stacking faults showed spots associated with q. Aging greatly enhances the mechanical properties of these alloys, even at high temperatures, and the metastable precipitate y' is thought to be responsible.' At 600°C the hardness of a Ni-10.1 at. pct Ti alloy increases rapidly during the first 20 hr of aging and continues to increase slowly thereafter.'" At 700°C the hardness reaches a maximum at 1 to 2 hr and then decreases.l1 In this paper the kinetics of formation of y' and the mechanism of transformation to in a Ni-Ti alloy are described. Evidence is presented that suggests that large-scale regions of titanium segregation are present initially and these may have an important influence on the initial strengthening of the alloy: these regions may be the heterogeneities detected by Ben-Israel and Fine. I) EXPERIMENTAL TECHNIQUES The Ni-10.3 at. pct Ti-0.6 at. pct A1 alloy was prepared by vacuum melting. Its chemical analysis is given in Table I. A 0.007-in.-thick strip was sandwiched between foils of the alloy and placed in an open Vycor tube with titanium chips (to get oxygen) and solution-treated for ; hr at 1150°C in a Globar Furnace under a flowing argon atmosphere. Following a quench into an ice-brine mixture at -5" to -2"C, a sample to be aged at temperatures up to and including 700°C was sealed in an evacuated Vycor capsule before being put into a fused salt bath. Aging above 700°C was carried out on bare samples in a salt bath. When removed from this bath and quenched into an ice-water mixture, the sample was normally coated with a thin layer of oxide easily removed with emery

Jan 1, 1970

By Frank G. Miller

This report presents developments of fundamental equations for describing the flow and thermodynamic behavior of two-phase single-component fluids moving under steady conditions through porous media. Many of the theoretical considerations upon which these equations are premised have received little or no attention in oil-reservoir fluid-flow research. The significance of the underlying flow theory in oil-producing operations is indicated. In particular, the theoretical analysis pertains to the steady, adiabatic, macroscopically linear, two-phase flow of a single-component fluid through a horizontal column of porous medium. It is considered that the test fluid enters the upstream end of the column while entirely in the liquid state, moves downstream an appreciable distance, begins to vaporize, and then moves through the remainder of the column as a gas-liquid mixture. The problem posed is to find the total weight rate of flow and the pressure distribution along the column for a given inlet pressure and temperature, a given exit pres5ure or temperature and given characteristics of the test fluid and porous medium. In developing the theory, gas-liquid interfacial phenomena are treated. phase equilibrium is assumed and previous theoretical work of other investigators of the problem is modified. Laboratory experiments performed with specially designed apparatus. in which propane is used as the test fluid, substantiate the theory. The apparatus. materials and experimental procedure are described. Comparative experimental and theoretical results are presented and discussed. It is believed that the research findings contributed in this * paper should not only lead to a better understanding of oil-reservoir behavior, but also should be suggective in regard to future research in this field of study. INTRODUCTION In recent years much time and effort has been consumed in both theoretical and experimental studies of the static and . dvnamic behavior of oil-reservoir fluids in porous rocks. Although lack of sufficient basic oil-field data, principally concerning the properties and characteristics of reservoir rocks and fluids, largely precludes quantitative application of research results to oil-field problems, qualitative application has become common practice. In effect. oil-reservoir engineering research is serving as a firm foundation for oil-field development and production practices leading to increased economic recoveries of petroleum. This province of research. however, still poses many perplexing problems. The thermodynamic behavior of two-phase fluids moving through porous media constitutes one facet of reservoir-fluid-flow research that has not received the attention it deserves. This report embodies a theoretical discussion of this subject and a description of a series of related laboratory experiments. The significance of the problem to oil field operations is indicated but in articular the report centers around a theory and method for analyzing the steady. macroscopically linear, two-phase flow of a fluid (a single molecular species) through a horizontal column of porous medium. For simplicity in showing how the thermodynamic behavior of two-phase fluids moving through porous media affects oil-reservoir performance problems, attention is focused temporarily on a particular well producing petroleum from an idealized water-free solution-gas drive reservoir, the reservoir rock being a horizontal, thin, fairly homogeneous sandstone of large areal extent confined between two impermeable strata. The flowing hydrocarbon fluid is considered to exist entirely as a Iiquid at points in the reservoir remote from the well; however. the decline in fluid pressure in the direction of the well causes vaporization of the hydrocarbon to begin at a radial distance r from the well. Upstream from r the fluid moves entirely as a liquid and downstream from r it moves either entirely as a gas or as a gas-liquid mixture depending on the properties of the hydrocarbon and on the thermodynamic process it follows during flow. The distance r would be variable under transient flow conditions. but for purposes of analysis the flow is considered to l~e steady at the particular instant of observation during the flowing life of the well of interest. If the flow were isothermal and the hydrocarbon a pure substance, the fluid would be entirely gaseous downstream from r. Thus, this isothermal flow process for a pure substance would require that the heat of vaporization be supplied at r. over zero length of porous medium, at the precise rate necessary to maintain the constant temperature. This means that the solid matrix of the porous medium (reservoir rock) and the surroundings (impermeable strata confining the reservoir rock) would have to serve as infinite heat sources. Heat-transfer requirements would be somewhat less severe for the isothermal flow of a multicorn-ponent hydrocarbon as bubble and dew points at the same temperature correspond to different pressures. In this instance isothermal conditions would be sustained without complete vaporization of the fluid over zero length of porous medium. Nevertheless. as the flow is in the direction of decreasing

Jan 1, 1951

By Frank G. Miller

This report presents developments of fundamental equations for describing the flow and thermodynamic behavior of two-phase single-component fluids moving under steady conditions through porous media. Many of the theoretical considerations upon which these equations are premised have received little or no attention in oil-reservoir fluid-flow research. The significance of the underlying flow theory in oil-producing operations is indicated. In particular, the theoretical analysis pertains to the steady, adiabatic, macroscopically linear, two-phase flow of a single-component fluid through a horizontal column of porous medium. It is considered that the test fluid enters the upstream end of the column while entirely in the liquid state, moves downstream an appreciable distance, begins to vaporize, and then moves through the remainder of the column as a gas-liquid mixture. The problem posed is to find the total weight rate of flow and the pressure distribution along the column for a given inlet pressure and temperature, a given exit pres5ure or temperature and given characteristics of the test fluid and porous medium. In developing the theory, gas-liquid interfacial phenomena are treated. phase equilibrium is assumed and previous theoretical work of other investigators of the problem is modified. Laboratory experiments performed with specially designed apparatus. in which propane is used as the test fluid, substantiate the theory. The apparatus. materials and experimental procedure are described. Comparative experimental and theoretical results are presented and discussed. It is believed that the research findings contributed in this * paper should not only lead to a better understanding of oil-reservoir behavior, but also should be suggective in regard to future research in this field of study. INTRODUCTION In recent years much time and effort has been consumed in both theoretical and experimental studies of the static and . dvnamic behavior of oil-reservoir fluids in porous rocks. Although lack of sufficient basic oil-field data, principally concerning the properties and characteristics of reservoir rocks and fluids, largely precludes quantitative application of research results to oil-field problems, qualitative application has become common practice. In effect. oil-reservoir engineering research is serving as a firm foundation for oil-field development and production practices leading to increased economic recoveries of petroleum. This province of research. however, still poses many perplexing problems. The thermodynamic behavior of two-phase fluids moving through porous media constitutes one facet of reservoir-fluid-flow research that has not received the attention it deserves. This report embodies a theoretical discussion of this subject and a description of a series of related laboratory experiments. The significance of the problem to oil field operations is indicated but in articular the report centers around a theory and method for analyzing the steady. macroscopically linear, two-phase flow of a fluid (a single molecular species) through a horizontal column of porous medium. For simplicity in showing how the thermodynamic behavior of two-phase fluids moving through porous media affects oil-reservoir performance problems, attention is focused temporarily on a particular well producing petroleum from an idealized water-free solution-gas drive reservoir, the reservoir rock being a horizontal, thin, fairly homogeneous sandstone of large areal extent confined between two impermeable strata. The flowing hydrocarbon fluid is considered to exist entirely as a Iiquid at points in the reservoir remote from the well; however. the decline in fluid pressure in the direction of the well causes vaporization of the hydrocarbon to begin at a radial distance r from the well. Upstream from r the fluid moves entirely as a liquid and downstream from r it moves either entirely as a gas or as a gas-liquid mixture depending on the properties of the hydrocarbon and on the thermodynamic process it follows during flow. The distance r would be variable under transient flow conditions. but for purposes of analysis the flow is considered to l~e steady at the particular instant of observation during the flowing life of the well of interest. If the flow were isothermal and the hydrocarbon a pure substance, the fluid would be entirely gaseous downstream from r. Thus, this isothermal flow process for a pure substance would require that the heat of vaporization be supplied at r. over zero length of porous medium, at the precise rate necessary to maintain the constant temperature. This means that the solid matrix of the porous medium (reservoir rock) and the surroundings (impermeable strata confining the reservoir rock) would have to serve as infinite heat sources. Heat-transfer requirements would be somewhat less severe for the isothermal flow of a multicorn-ponent hydrocarbon as bubble and dew points at the same temperature correspond to different pressures. In this instance isothermal conditions would be sustained without complete vaporization of the fluid over zero length of porous medium. Nevertheless. as the flow is in the direction of decreasing

Jan 1, 1951

By A. L. Pranatis, G. M. Pound

Changes in length of copper foils of varying thickness and grain size were measured under such conditions of low stress and high temperature that it is believed that creep was predominately the result of interboundary diffusion of the type recently discussed by Conyers Herring. The surface tension of copper was calculated and results confirmed previous work within the limits of experimental error. Under the assumption of viscous flow, viscosities were calculated as a function of temperature and grain size. Predictions of the Nabarro Herring theory of surface grain boundary flow were borne out fully and the Herring theory of diffusional viscosity is strongly supported. ONLY a relatively few techniques for obtaining the surface tension of solids are presently available. Of these, the simplest and most straight forward is the direct measurement of surface tension by the application of a balancing counterforce. Thin wires or foils are lightly loaded and strain rates (either positive due to the downward force of the applied load or negative if the contracting tendency of surface tension is sufficiently greater than the applied stress) are observed. By plotting strain rates against stress, the load which exactly balances the upward pull is found and a simple calculation yields a value for the surface tension. The technique is of comparative antiquity, and solid surface tension values were reported by Chapman and Porter,' Schottky; and Berggren" in the early part of the century. Later, the filament technique became fairly well established as a method for determining the surface tension of viscous liquids, and Tammann and coworkers,'. " Sawai and co-worker and Mackh howed good agreement between the values of surface tension for glasses and tars obtained by the filament technique and by more conventional methods. With the increased confidence in the technique gained in these experiments, the method was applied to solid metals and the first reliable values of surface tension of solid metals were reported by Sawai and coworkers10' " and by Tammann and Boehme." More recently, Udin and coworkersu-'" have reported the results of experiments with gold, silver, and copper wires. Similar experiments with gold wires were carried out by Alexander, Dawson, and Kling.'" The excellent review articles of Fisher and Dunn" and of Udinl@ should be referred to for detailed criticism of the foregoing work and for discussion of underlying theory. In all the foregoing calculations, it is assumed implicitly that the material contracts or extends uni- formly along the length of the specimen and also that it flows in a viscous fashion, i.e., that strain rates are proportional to stress. For an amorphous material, such as glass, tar, or pitch, the assumptions are quite valid and good agreement is obtained with values of surface tension measured by other techniques. The values reported for metals, however, are occasionally regarded with misgiving, since it can be argued that, because of their crystalline nature, true solids can not deform in a viscous fashion. If this is true, then the results reported for solid metals over a long period of years are of only doubtful value. Thus it is clearly necessary that a mechanism be established that would explain both the viscous flow and the uniform deformation that has been assumed. Such a mechanism has been proposed by Herring."' Briefly, he suggests that, under the conditions of the experiment, deformation takes place by means of a flow of vacancies between grain boundaries and surfaces. This is a direct but independent extension of the theory proposed by Nabarro" in an attempt to explain the microcreep observed by Chalmer~.In a condensed form the Herring viscosity equation is TRL there 7 is the viscosity, T the absolute temperature, R and L grain dimensions, and D the self-diffusion coefficient. In its complete form, all constants are calculable and it includes such factors as grain shape, specimen shape, and degree of grain boundary flow. When applied to existing data, good agreement was obtained between predicted and observed flow rates. The theory received provisional confirmation from the work of Buttner, Funk, and Udin" who observed viscosities in 5 mil Au wire much higher than those in the 1 mil wire used by Alexander, Dawson, and Kling.'" More significant were the completely negligible strain rates found by Greenough" in silver single crystals. Opposed to these observations were those of Udin, Shaler, and Wulff'" who found indications of viscosity decreasing as grain size increased. Thus, complete confirmation of the theory was lacking in that the data to which it could be applied contained only a limited number of grain sizes. Hence, it was proposed that a series of experiments be carried out with thin foils of varying grain size up to and including single crystals, where, according to the Herring theory, deformation would occur only at almost infinitely slow rates.

Jan 1, 1956

By T. M. Morris, W. E. Horst

W. E. Horst—In regard to the flotation rate being described as "first orcler" for flotation of quartz particles below 65 p in size (or any size studied in this work) in this paper, it appears that the authors' conception of rate equations is not in agreement with cited references. A first order rate equation has as one of its forms the following: a In.=a/a-x=kt where a = initial concentration, a—x = concentration at time t, t = time, and k = constant. The constant, k, has the dimension of reciprocal time which is similar to the specific flotation rate, Q. described by Eq. 2 in the authors' article, as has been previously discussed by Schumann (Ref. 1 of original article). The plotted data presented in Fig. 4 of the article utilizes the specific flotation rate, Q (min.'); however, there is not adequate data given to indicate the order of the rate equation which describes the flotation behavior of the quartz system studied. Results from the experimental work indicate that the relationship between rate of flotation (grams per minute) and cell concentration (provided the percent solids in the flotation cell is less than 5.2 pct and the particle size is less than 65 p) is described by an equation of the first order (R, = k c+", n being equal to 1 in this size range) and the use of the first order rate equation does not apply. Similarly the relationship for other particle size ranges studied is expressed by equations of the second or third order depending on the magnitude of n. T. M. Morris—The authors are to be commended for the experiments which they performed. As they state in their discussion the concentration of collector ion In solution did change with change in concentration of solids in the flotation cell. Since for a given slze of particle, flotation rate increases with concentration of collector until a maximum is reached, the effect of concentration of particles in their experiments was to vary the concentration of collector ions. A collector concentration which insures maximum supporting angle for all particles eliminates the unequal effect of collector concentration on various sized particles and the effect of size of particles and concentration of particles upon flotation rate could be more clearly assessed. I believe that if the authors had increased the concentration of collector to an amount sufficient to attain a maximum supporting angle for all particles they would find that the specific flotation rate of particles coarser than 65 p would be constant with change in the concentration of solids in the flotation cell, and that a first order rate would apply to the + 65 as well as to the —65 p sizes. It might also be discovered when this change in collector concentration was made that the maximum specific rate constant would be shifted toward a coarser fraction than when starvation quantities of collector are used since this practice favors the fine particles and penalizes the coarse particles. P. L. de Bruyn and H. J. Modi (authors' reply)—The authors wish to thank Professor Morris for his kind remarks and for mentioning the influence of equilibrium collector concentration on flotation rate. With a collector concentration sufficient to insure maximum supporting angle for all particles, a first order rate equation may indeed be found to be generally applicable irrespective of size. Such a concentration would, however, lead to 100 pct recovery of the fine particles and consequently defeat the essential objective of the investigation to derive the maximum information on flotation kinetics. To establish absolutely the validity of any single rate equation for a given size range, the ideal method would be to work with a feed consisting solely of particles of that size range. Use of such a closely sized feed would also eliminate the possibility of the interfering effect of different sizes upon one another. The authors do not believe that increasing the collector concentration would shift the maximum specific flotation rate (Q) towards a coarser fraction. Experimentation showed Q to be independent of solids concentration for all particles up to 65 µ in size, whereas the maximum value of Q was obtained in the range 37 to 10 p. Professor Morris contends that the addition of starvation quantities of collector favors fine particles at the expense of coarse particles, but the reason for this is not entirely clear to the authors. The comments by Mr. W. E. Horst are concerned only with the concept of the term "first order rate equation." According to the usage of this term in chemical kinetics, time is an important variable, as is shown in the equation quoted by Mr. Horst. All the experimental results reported by the authors were obtained under steady state continuous operations when the rate of flotation is independent of time. To be consistent with the common usage of the "first order rate equation," it would be more satisfactory to state that under certain conditions the experimental results show that the relation between flotation rate and pulp density is an equation of the first order.

Jan 1, 1957

By Charles L. Mantell, James E. Hoffmann

Mechanical inclusion, electrophoretic deposition, and adsorption were studied as mechanisms for code-position of aluminas present in copper-plating electrolytes as an insoluble disperse phase. Mechanical inclusion was not a significant factor. That codeposi-tzon of aluminas by an electrophoretic mechanism was unlikely was substantiated by measurements of the potential of the aluminas. The alumina content of the deposits was studied as a function of the pH of the bath. These tests in conjunction with sedimentation studies demonstrated the absence of an isoelectric point for the alutninas over the pH range examined. Thiourea in the electrolyte (a substance known to be adsorbed on a copper cathode during electrodeposition) affected the amount of alumina in the electrodeposit. However, no adsorption of thiourea on aluminas in aqueous dispersions was detected. If it were possible to produce a dispersion-hardened alloy of copper and alumina by electrodeposition, an alloy possessing both strength and high conductivity at elevated temperatures might be anticipated. Investigation of the mechanism of codeposition of aluminas with copper was undertaken with the hope that knowledge of the mechanism would aid in the development of such an alloy. The word "codeposit" here does not necessarily imply an electrolytic phenomenon but rather that the materials codepositing, the various aluminas, are transported to and embedded in the electrodeposited copper by some means. Mechanical inclusion in electrodeposition implies a mechanism of codeposition which is wholly mechanical in nature; the only forces acting on a particle are gravity and contact forces. Such a particle is presumed to be electrically inert and incapable of any electrical interaction with electrodes in an electrolytic plating bath. Processes for matrices containing a codeposited phase by electrodeposition from a bath containing a disperse insoluble phase frequently state that code-position is caused by mechanical inclusion.10,2,12 If settling, i.e., gravity, be the controlling mechanism for codeposition of aluminas, then assumptions may be made that 1) the content of alumina in the electrodeposit should be enhanced by increasing the particle size, 2) the geometry of the system, that is, the disposition of the cathode surfaces relative to the di- rection of the falling particles, should affect the alumina content of the electrodeposit, 3) in geometrically identical systems the chemical composition of the electrolyte employed should exercise no effect on the alumina content of the deposit, that is, the alumina content should be the same in all cathode deposits irrespective of bath composition. A bent cathode19 evaluates the clarity of filter effluent in electroplating baths by comparing the roughness of the deposit on the vertical surface with that on the horizontal surface. Two difficulties are inherent in this technique: 1) the current density on the horizontal portion of the cathode would be substantially greater than that on the vertical surface; 2) should the deposit obtained be rough, projections on the vertical face could act as horizontal planes and vitiate the relationship between the vertical and horizontal surfaces. Bath composition should have no substantial effect on the alumina content of the deposit. Two different electrolytic baths were employed. They possessed variant specific conductances and substantially different pH ranges. The experimental tanks were rectangular Pyrex battery jars 6 in. wide by 3 1/4 in. long by 9 3/4 in. deep. The cathodes were stainless steel 316 sheet of 0.030 in. thickness, cut to 7.5 by 1.75 in. and bent at right angles to form an L-shaped cathode whose horizontal surfaces measured 1.75 by 3.0 in. All edges and vertical surfaces were masked with Scotch Elec-troplaters Tape No. 470. The anodes were electrolytic cathode copper 9 in. high by 2.25 in. wide by 0.5 in. thick. To eliminate inordinately high current densities on the projecting edge of the cathode, the anode was masked 1 in. above and below the projected line of intersection of the cathode with the anode. The exposed area of the anode was equal to that of the cathode, providing both with equal average current densities. The agitator in the cell was of Pyrex glass and positioned so its center line was equidistant from cathode and anode, and a plane passed horizontally through the center of the blade would be located equidistant from the bottom of the cathode and the bottom of the deposition tank. The assembled apparatus is depicted in Fig. 1. Hatched areas on anode and cathode represent the area of the electrodes wrapped with electroplaters tape. MATERIALS The chemicals were copper sulfate—CuSO4 • 5H2O— technical powder (Fisher Scientific Co.). Spectro-graphic analysis showed substantial freedom from antimony, arsenic, and iron. Traces of nickel were present.

Jan 1, 1967

By G. P. Airey, R. F. Mehl, T. A. Hughes

The coarsening of cementite particles in a ferrite matrix has been studied in a series of steels with 0.15 pct C only and 0.15 pct C plus 1 pct Ni, Mn, and Cr, respectively. Two initial states were employed: quenched nartensite, and quenched and cold-rolled martensite. A series of tempering temperatures between 500' and 700" and tempering times of up to 190 hr were used. The structures were studied by replica and transmission electron microscopy. Particle size distribution curves were determined. From the average size value coarsening curves were obtained. These were plotted in accordance with the Wagner analysis assuming diffusion control. A discussion of the significance of the results is given. L HE reactions occuring upon the tempering of martensite have long engaged the attention of metallurgists. The latter stages, when cementite particles coarsen in a ferrite matrix, have been studied both qualitatively and quantitatively. Studies of such coarsening processes have recently been spurred by the publication of the Lifshitz-Wagner theory1, and the extension of this to the a Fe-Fe3C system by Oriani3 and by Li, Blakely, and einold. Following Wagner the coarsening process is often designated as "Ostwald Ripening". The only quantitative data on the rate of coarsening, except for the work of Hyam and uttin,' in the a Fe-Fe3C system are those of Bannyh, Modin, and odin' for a commercial eutectoid steel and those of Heckel and ereorio" using a pure eutectoid steel. The data of Bannyh, Modin, and Modin have been employed by 0riani3 to derive the a Fe-FeE interface energy. The reaction is one of the most important ones in steel and is worthy of detailed study. This is the purpose of the present study. Laboratory heats were prepared; these were steels with approximately 0.15 pct C, selected so that the num ber of carbide particles would be relatively small and thus so that the overlapping diffusion fluxes would be minimized, presumably a desirable circumstance.'-3 In addition to Fe-C alloys, comparable heats containing 1 pct of Ni, Mn, and Cr, respectively, were included with a view of appraising the effect of alloying elements. This report includes an account of the micro-structures observed, primarily with the electron microscope, and of kinetic data and their interpretation. MATERIALS AND TREATMENT The alloys were prepared from electrolytic iron ("Plastiron") and high-purity graphite; these were melted in a zirconia crucible using a vacuum furnace. The alloy steels were made by adding electrolytic nickel, electrolytic manganese, and "vacuum grade" chromium, respectively, under a partial pressure of argon. Each melt was poured into a mold within the vacuum furnace and cooled in the mold. The ingots were 2 in. in diam. and 8 to 10 in. long. The analysis of the alloys is given in Table I. These ingots were hot-rolled to strip 0.1 in. thick, then cold-rolled to 0.05 in. and each alloy split into two batches. One batch was austenitized at 1200 for 1 min, quenched in cold brine, then cold-rolled to 0.02 in.; samples given this treatment are hereinafter designated as "worked". The other batch was cold-rolled to 0.025 in., austenitized at 1200" for 1 min, and quenched in cold brine; such samples are hereinafter designated as "quenched". These two batches were then tempered, as below. The purpose of the treatment given the first batch was to provide an initial structure of cold-worked martensite, with the expectation that the additional defect structure created by cold work would encourage a higher rate of nucleation of cementite on tempering and hence a more uniform distribution of cementite particles. Individual specimens were sealed in evacuated quartz or Pyrex tubes, then tempered in a muffle furnace. The temperature control was better than 3'C at 700. Tempering treatments wer: performed at 400°, 500°, 550°, 600°, 65o°, and 700C for time periods between 15 min and 190 hr. PREPARATION OF SPECIMENS Specimens for optical and replicalelectron microscopy were mounted, polished conventionally, and etched with 2 pct nital. For electron microscopy, single-stage "formvar" replicas were made, dry-stripped and rotary-shadowed with chromium at an angle of 30 deg. Carbide extraction replicas were prepared from electropolished specimens usirig the method described by Smith and uttin.' Thin foils for electron transmission microscopy were prepared by chemical thinning in an H202-HF bath prior to electropolishing in a chromium trioxide-acetic acid solution. The most

Jan 1, 1969

By Donald B. Evans, Robert D. Pehlke

The solubility of nitrogen in liquid Fe-A1 alloys has been measured up to the solubility limit for formation of aluminum nitride using the Sieverts method. The activity coefficient of nitrogen decreases slightly with increasing aluminum content in the range of 0 to 4 wt pct Al. Based on a nitride composition, AlN, the standard free energy of formation of aluminum nitride from fhe elements dissolved in liquid iron has been determined to be: ?F" = -59,250 + 25.55 T in the range from 1600º to 1750ºC. The solubility of nitrogen in liquid iron alloys and the interaction of nitrogen with dissolved alloying elements in liquid iron have been the subject of a number of research investigations.' Most of this work, however, has been reported for concentrations well below those necessary for the formation of the alloy nitride phase. Data in the concentration region near the solubility limit of the alloy nitride, particularly for systems exhibiting stable nitrides, are important in evaluating the denitrifying power of various alloying elements. They are also useful in determining the stability of a given nitride if it is to be used as a refractory to contain liquid iron alloys. In view of the importance of aluminum as a deoxidizing agent in commercial steelmaking and the fact that its nitride, AIN, is a highly stable compound and has merited some consideration as an industrial refractory, the following investigation was undertaken. The use of the Sieverts technique provided a measurement of the equilibrium nitrogen solubility in liquid Fe-A1 alloys as a function of nitrogen gas pressure up to 3.85 wt pct A1 in the temperature range of 1600º to 1750°C. The values obtained by the Sieverts method were checked by means of a quenching method in which liquid iron was equilibrated with an A1N crucible under a known partial pressure of nitrogen gas, and the solubility of A1N in liquid iron determined by chemical analysis. EXPERIMENTAL PROCEDURE The theoretical considerations involved in determining the solubility product of a solid alloy nitride phase in liquid iron by measuring the point of departure of the nitrogen gas solubility from Sieverts law have been discussed by Rao and par lee.' The principal problem is to determine the variation of nitrogen solubility in an alloy as a function of the pressure of nitrogen gas over it with sufficient precision to establish the break point in the curve at the solubility limit of the alloy nitride phase. A fairly large number of data points are required to do this. A second problem is the determination of the composition of the precipitated solid nitride phase. This is necessary in order to define completely the thermodynamic relationships. The Sieverts apparatus used to make the nitrogen solubility measurements in this investigation is of essentially the same design as that described by Pehlke and E1liott.l The charge materials were Ferrovac-E high purity iron supplied by Crucible Steel Co. and 99.99+ pct pure aluminum. Recrystal-lized alumina crucibles were used, and were not attacked by the liquid alloys. The hot volume of the system which was measured for each melt ranged from 46 to 50 standard cu cm and was found to decrease linearly with decreasing pressure and with increasing temperature. The temperature coefficient of the hot volume at 1 atm pressure of argon gas was essentially constant for all experiments at a value of -6 X 10-3 cu cm per "C. The melt temperature was measured with a Leeds and Northrup disappearing filament type optical pyrometer sighted vertically downward on the center of the melt surface. The temperature scale was calibrated against the observed melting point of pure iron taken as 1536°C. The emissivity of all melts was assumed to be that of pure iron, taken as 0.43. The charge weights ranged from 110 to 140 g and the range of aluminum contents covered was from 0 to 3.85 wt pct. Aluminum additions were made as 12 to 15 wt pct A1-Fe master alloys previously prepared in the system under purified argon. The compositions of the master alloys were checked by chemical analysis and found to be in agreement with the charge analyses. Vertical cross sections of the master-alloy ingots were used as charge material for the equilibrations in order to minimize the effect of any segregation which might have occurred during solidification of the master alloys. Determinations of the solubility product of

Jan 1, 1964

By E. M. Cox, A. S. Skapski, N. H. Nachtrieb, M. C. Bachelder

As a result of using copper-containing scrap in the steelmaking process, the copper content of steels has been steadily increasing for years. Consequently the possible role copper may play in the steelmaking process and in the finished product begins to attract the metallurgists' attention. Some time ago one of the present authors forwarded the idea—based on the results of the analysis of nonmetallic inclusions extracted electrolytically from steels—that sulphur in plain carbon steels is distributed mainly between copper and manganese, the amount of iron sulphide being very small; and that, consequently, the problem of copper and that of sulphur in steel cannot be treated separately.' At the time of the publication of the quoted paper little was known about the relative affinities of copper and manganese for sulphur at high temperatures except that at moderate temperatures (below 1000°C) the affinity of manganese for sulphur is much greater. To gather more experimental data on this subject, the present authors undertook the investigation of the equilibrium constants of the reactions: 2Mn(8 or 1) + S2(g) = 2MnS(s) 4Cu(s or 1) + S2(g) = 2Cu2S (S or I)* 2Fe(s) + S2(g) = 2FeS (s or 1) over a range of temperatures wide enough to establish the dependence of these equilibrium constants on temperature. From the equilibrium constants (K = l/Ps2) the free energy of formation (affinity) can be calculated from F° = -RTln 1/PSt (1) where the standard conditions chosen are: 1 atm of sulphur pressure and the activities of condensed components equal one. The decomposition pressure, Ps2, of sulphur over the respective sulphides is too small to be measured directly, but there is a way of eliminating this difficulty by measuring the equilibrium constant of the reaction between the sulphide and hydrogen. From the latter and from the equilibrium constant of the thermal dissociation of H2S we then calculate Ps2 for the respective sulphide. 2Mn + 2H2S = 2MnS + 2H, 2H2 + S2 = 2H2S_________ 2Mn + S2 = 2MnS The numerical values of the equilibrium constant of the thermal dissociation of H2S at different temperatures were taken from Kelley's paper, "The Thermodynamic Properties of Sulfur and its Inorganic Compounds."² In previous experimental work published by Jellinek and Zakowski3 and by Britzke and Kapustinsky4 the equilibrium constants of the reactions Metal sulphide + H2 = H2S + metal were determined by passing hydrogen, at different rates of flow, over the sulphide, analyzing the resulting H2S + H2 mixture and then extrapolating the H2S/H2 ratio (which is a function of the rate of flow) to the zero speed of flow, a method necessarily involving considerable uncertainty. In the present work the equilibrium ratio was actually measured instead of being extrapolated. The apparatus is shown in Fig 1. Experimental Procedure The sulphides were prepared by the following methods: FeS Powdered iron which had been reduced with hydrogen (ferrum reduc-tum) was mixed in stoichiometric ratio with sublimed sulphur and carefully ground. The mixture was put into an alundum crucible, covered with pure sulphur, and the reaction started by touching the mixture with a glowing iron rod. After the reaction was completed the product (still containing some metallic iron) was again ground with sulphur, put into a Rose crucible, covered with sulphur, and heated in a strong current of pure hydrogen. Analysis of the final product showed 62.46 pct Fe and 36.59 pct S. Theoretical for FeS: 63.53 pct Fe and 36.47 pct S. MnS Manganese sulphide (precipitated and carefully washed with distilled water containing H2S) was dried in a Rose crucible in an atmosphere of H2S and heated in a current of hydrogen for 2 hr at red heat. The product was then ground and ignited for several hours at 1000°C in a current of hydrogen sulphide. Analysis showed 64.53 pct Mn and 36.63 pct S. Theoretical: 63.15 pct Mn and 36.85 pct S. Some MnS samples were prepared from metallic manganese and sublimed sulphur by mixing and grinding them and then heating in a current of hydrogen sulphide in an alundum tube.

Jan 1, 1950

By D. W. Lillie

It has been demonstrated that considerable bend ductility exists in bulk specimens of polycrystalline high-purity silicon. The possibility of hot-forming at 1200°C is suggested. EXCELLENT corrosion resistance in many media and low cross section for absorption of thermal neutrons (0.13 barn) would make silicon of interest to nuclear engineers were it not for extreme brittle-ness and the difficulty of fabrication by any reasonable means. The use of silicon for structural purposes also has been considered in view of its light weight and oxidation resistance. Johnson and Han-sen' have investigated the properties of silicon-base alloys and concluded that there was no way of making pure silicon or silicon-rich alloys ductile at room temperature. In view of reports of appreciable ductility in germanium single crystals above 550°C'." and some plastic deformation in single-crystal silicon above 900oC,' the present investigation was undertaken to define more precisely the limits of high-temperature ductility in pure silicon. After this investigation was begun torsion ductility in both germanium and silicon was reported by Greiner." Through the courtesy of F. H. Horn, a small bar of cast extra high-purity silicon was obtained and small bend specimens were made from it by careful machining and grinding. All of the reported tests results were obtained from samples from this bar (bar No. 1) and one other of similar source (bar No. 2). No complete analysis was obtained but, based on analysis of similar semi-conductor grade material, metallic impurities were under 0.01 pct total. Vacuum-fusion analysis for oxygen showed a value of 0.0018 2 0.0003 pct for the first bar tested and metallographic analysis showed no evidence of a second phase. Bend tests were carried out on an Instron tensile machine using a bend fixture with a 1 -in. span loaded at the center. Supporting and loading bars were 0.250 in. round and the load was applied by downward motion of the pulling crosshead of the machine. Specimen thickness and width were approximately 0.10 in. and % in. respectively. Loading rate was controlled by holding crosshead motion constant at 0.02 ipm. In some cases a smaller specimen was used on a 5/8-in. span with a 0.129-in.-diam loading bar. The entire bend fixture was surrounded by a hinged furnace and all heating was done in air atmosphere. Temperature measurement was made with thermocouples fastened directly to the bend fixture within less than 1 in. from the specimen. Autographic stress-strain curves were recorded during each test, and breaking load, total deflection, and plastic strain could be obtained from these curves. Stress was calculated from the beam formula S = 3PL/2bh2, where P is the load in pounds, L the span in inches, b the specimen width in inches, and h the specimen thickness in inches. This formula is strictly correct only in the elastic range but has been used to calculate a nominal stress for convenience in the plastic range. The stress given is the maximum stress in the specimen. Results The results of the complete series of tests are shown in Table I. The first group of tests (specimens Nos. 1-6) showed the beginning of plastic flow at a test temperature of 900°C, so two additional tests (Nos. 8 and 9) were made at 950°C on small-size specimens from bar No. 2. Specimen No. 8 was tested in the as-machined condition, and No. 9 was heat-treated in hydrogen at 1300°C for 2 hr, cooled to 1200°C and held 1 hr, cooled to 1000°C and held 1 hr, cooled to 900°C and held 1 hr, and finally cooled to a low temperature before removal from the hydrogen. It is apparent that the heat-treatment had a significant effect on yield strength and ductility. In addition, the magnitude of the yield point was conslderably reduced in the heat-treated specimen as is shown m Fig. 1 by tracings of the stress-strain curves. After obtaining a furnace capable of reaching higher temperatures specimens Nos. 10 to 13 were tested at 1100 and 1200°C. Strain rate was increased by up to a factor of 10 to see whether the ductility observed was excessively strain sensitive. Specimen NO. 10, strained at 0.02 ipm and 1100oC, was still bending at a deflection of 0.322 in. when the load rate was increased to 0.2 ipm, resulting in immediate

Jan 1, 1959

By J. E. Hanafee, G. J. London

MANY studies of the deformation behavior of materials under a superimposed hydrostatic pressure have shown that materials brittle at ambient pressure behave in a ductile manner under pressure. Thus, with a metal such as beryllium which possesses relatively low ductility, but otherwise exhibits quite useful physical and mechanical properties, hydrostatic pressure may be particularly useful for both forming beryllium shapes and studying its deformation behavior. In fact, it has been found1"4 that polycrystal-line beryllium in both ingot and powder form appears to behave in a more ductile manner on a macroscopic level in a hydrostatic pressure environment, and it has been suggested2 that this is due to activating a new slip mode. Furthermore, Andrews and Radcliffe5 have found pressure induced nonbasal dislocation activity in hot pressed beryllium. Recently6 it has been shown in "c-axis" compression tests under hydro-static pressures up to 28 kbars that the shear stress needed to cause slip with a Burgers vector out of the basal plane (pyramidal slip) does not change with increasing pressure in beryllium with a purity of some 99.5 pct. This material is equivalent or more pure than the beryllium used in the previous pressure studies. Thus, it appears, as suggested by Inoue et al.,3 that the hydrostatic pressure affects the fracture stress rather than the stress necessary to activate pyramidal slip in beryllium. However, in "c-axis" pressure tests on high purity 12 zone pass beryllium (˜50 ppm total impurities) the macroscopic compression stress needed to cause pyramidal slip was considerably lower than that at ambient pressure.6 It has further been shown that alloying beryllium with nickel and copper in the range 2-5 wt pct also favors the occurrence of pyramidal slip in "c-axis" compression tests,7'8 while lower amounts of nickel and copper do not have significant effects. In the present study the combined effect of hydrostatic pressure and alloying high purity beryllium on the shear stress needed to cause pyramidal slip has been ascertained. A 2.5 wt pct Cu alloy was selected as the first alloy to study as this level of copper did favor pyramidal slip at room pressure. A high purity (12 zone pass) single crystal of beryllium 0.3 by 0.1 by 0.1 in. was cut and polished by an orientation and lapping technique8 so that the top and bottom compression surfaces were parallel and within 3 min of arc to the (0001) plane and the sides parallel to the {l010} and {ll20} planes. In these compression specimens, therefore, the resolved shear stress was nearly zero on both the basal and prism planes, and slip was restricted to pyramidal systems. Analysis of slip traces on the two lateral surfaces served to accurately identify the active slip planes.6'9 The pressure unit was a modified piston-cylinder device fitted with a manganin transducer coil arrangement which continuously monitored and recorded the hydro-static pressure. The load on the specimen was measured by a strain gage load cell which operated entirely within the pressure chamber. This load cell was calibrated before and after each pressure cycle at room pressure in situ and the calibration did not vary more than ±1 pet. These techniques and devices have been previously described in more detail.6 Successively higher compressive stresses were applied to the single crystal under a superimposed hydrostatic pressure until fracture occurred. The strain rate was (4.5 ± 2.0) x 10-6 sec-1 and the average rate of pressure application and release was approximately 0.3 kbars per min. As the load on the specimen was applied by the piston which was used to increase the hydrostatic pressure, the pressure increased during the compression test. This increase ranged from 0.0 to 0.8 kbars, and the maximum hydrostatic pressures are quoted in Fig. 1. The lateral surfaces of the specimen were examined in a light microscope after each pressurization/stress cycle so that the stress at the onset of {1122} pyramidal slip could be ascertained. Post compression height measurements allowed the plastic strain in the specimen to be evaluated to within 0.03 pct. The resulting compression stress-plastic strain curve is shown in Fig. 1 with results of a "c-axis" test on a similar Be-2.5 wt pct Cu single

Jan 1, 1970

By L. F. Bryant, J. P. Hirth, R. Speiser

The surface, gain boundary, and twin boundary energies, as well as the surface diffusion coefficient, of cobalt were determined from tests at 1354°C in pure hydrogen. A value of 1970 ergs per sq cm was calculated for the surface energy, using the zero creep method. It was possible to measure the creep strains at room temperature because the phase transformation was accompanied by negligible irreversible strain and no kinking. Established techniques based on interference microscopy were used to obtain values for the other three properties. The gain boundary and twin boundary energies were 650 ad 12.7 ergs per sq cm, respectively, while a value of 2.75 x l0 sq cm per sec was determined for the surface dufusion coefficient. In the course of a general study of cobalt and cobalt-base alloys, information was required about the surface energy of cobalt. Hence, the present program was undertaken to measure the interfacial free energy, or, briefly, the surface energy, of the solid-vapor interface of cobalt. The microcreep method was selected for this measurement because other surface properties could also be determined from the accompanying thermal grooving at grain boundaries and twin boundaries. A brief summary of the methods for determining the various surface properties follows. At very high temperatures and under applied stresses too small to initiate slip, small-diameter wires will change in length by the process of diffu-sional creep described by Herring.1 The wires acquire the familiar bamboo structure and increase or decrease in length in direct proportion to the net force on the specimen. For a specimen experiencing a zero creep rate, the applied load, wo, necessary to offset the effects of the surface energy, y,, and grain boundary energy, y b, is given by the relation: where r is the wire radius and n is the number of grains per unit length of wire. The first results obtained from wire specimens were reported by Udin, Shaler, and Wulff.' udin3 later corrected these results for the effect of grain boundary energy. The grain boundary energy is determined from measurements of the dihedral angle 8 of the groove which develops by thermal etching at the grain boundary-free surface junction. For an equilibrium configuration: Measurements of the angle 8 can be made on the creep specimens4'5 or on sheet material, as was done in this investigation by a method employing interference microscopy.= If the vapor pressure is low, the rate at which grain boundary grooves widen is determined primarily by surface diffusion and, to a lesser extent, by bulk diffusion. The surface diffusion coefficient, D,, is obtained from interferometric measurements of the groove width as a function of the annealing time, t. As predicted by Mullins~ and verified by experiment, the distance, w,, between the maxima of the humps formed on either side of the grain boundary increases in proportion to if grooving proceeds by surface diffusion alone. For this case: where fl is the atomic volume and n is the number of atoms per square centimeter of surface. When volume diffusion also contributes to the widening, the surface diffusion contribution can be extracted from the data by the method described by Mullins and shewmon.8 Where a pair of twin boundaries intersects a free surface, a groove with an included angle of A + B (using the groove figure and notations of Robertson and shewmong) forms by thermal etching at one twin boundary-free surface junction. If the "torque terms", i.e., the terms in the Herring10 equations describing the orientation dependence of the surface energy, are sufficiently large, an "inverted groove" with an included angle of 360 deg-A'-B' develops at the other intersection. The angles A + B and A' + B' are measured interferometrically. When the angle, , between the twinning plane and the macroscopic surface plane is near 90 deg, the twin boundary energy is calculated from the relation: 1) EXPERIMENTAL TECHNIQUES Five-mil-diam wire containing 56 parts per million impurities was used for making ten creep specimens. These specimens had about 15 mm gage lengths with appended loops of wire and carried loads (the specimen weight below the midpoint of the gage length) ranging from 3.7 to 149.8 mg. The wires were hung inside a can made from 99.6 pct pure cobalt sheet. Beneath the wires were placed small specimens of 20-mil-thick, 99.9982 pct pure cobalt sheet from which the relative twin boundary and grain boundary energies and the surface diffusion coefficient were measured. All the specimens were annealed at a temperature of 1354" i 3°C which is 92 pct of the absolute melting point of cobalt. The furnace atmosphere was 99.9 pct pure hydrogen that was purified further by a Deoxo catalytic unit, magnesium perchlorate, and a liquid-nitrogen cold trap. As a precautionary measure the gas was then passed through titanium alloy turnings which were heated to 280" to 420°C and replaced after every test period. The hydrogen was maintained at a

Jan 1, 1969

By C. C. Templeton, J. C. Rodgers

A key step in feed water treatment for generating wet steam for thermal oil recovery is the removal of calcium and magnesiunt hardness by cation-exchange series softening. Knowing the solubility of any scale forming salts in brines at elevated temperatures is necessary for fixing the level to which the feed water must be softened. Such calcium sulfate solubility data, previously not available above 392F, were determined by the authors in a flow equilibrium apparatus mud will be reported elsewhere. These data were used to develop a method for predicting the solubility of anhydrite in hot water or steam droplets for saturated steam pressures as high as 2,000 psig (637F). (The calcium sulfate solubility product is represented by a combination of two factors, one reflecting the effects of ionic strength and the other accounting for the effects of complex ion formation in either calcium-magnesium-rich or sulfate-rich brines.) The method is applied to a calcium-magnesium-rich brine If moderately high salinity from a pilot hot-water flood, I nd to several sulfate-rich, low-salinity feed waters and l lowdown (cooled steam droplets) samples from steam s ak operations. The predicted calcium hardness levels corresponding to the calcium sulfate solubilities agreed reasonably well with the results of laboratory solubility determinations run on the field samples. Further testing of the method is needed for brines of other composition classes. Existing field cation exchange softeners in the cases tested are performing adequately since all the samples were found to be undersaturated with respect to calcium sulfate at their operating temperatures. Introduction Prevention of scaling caused by precipitation of calcium sulfate (anhydrite) is of considerable concern in connection with thermal recovery processes using wet steam or hot water. To avoid anhydrite precipitation in a heated system, an engineer must keep the product of the calcium and sulfate concentrations in the water or steam droplets below the value of the solubility product of anhydrite for the temperature and brine composition in question. Usually it is most practical to keep the concentration product lower than the solubility product by keeping calcium low in the presence of high sulfate, or by keeping sulfate low in the presence of high calcium. This can be done by a choice of combinations of natural waters and water treatment processes (such as series cation exchange softening to remove calcium). Until recently, few anhydrite solubility data, particularly for solutions containing other salts, were available for temperatures above 392F (211 psig steam); Marshall, Slusher and Jones' studied the CaS0,-NaCI-H,O system up to 392F and surveyed the work of previous investigators. To model natural brines, one needs to study the solubility of anhydrite in aqueous solutions of sodium chloride, sodium sulfate, calcium chloride, magnesium chloride and their mixtures. Since steam pressures as high as 2,000 psig (637F) may be involved in thermal oil recovery projects, a solubility study was conducted between 482 and 617F.' Discussed in this paper is the application of these data to the prediction of anhydrite precipitation in some practical steam soak and hot-water injection projects. Any simple method for predicting the solubility of an inorganic compound over a wide range of temperatures and solution compositions must be based on some assumptions, and therefore must yield approximate results. On the one hand, natural brines contain too many ionic species for all to be included in a simple scheme; on the other hand, there is no adequate theoretical basis for the exact prediction of solubility in even simple solutions of mixed electrolytes. However, it is possible at a given temperature to base a reasonable prediction scheme on two phenomena:'-' the increase in solubility with increasing total concentrations of all ions (as measured by the ionic strength; see the Appendix), and an increase due to formation of cornplexes between calcium ions and sulfate ions and between sulfate and magnesium ions. Stiff and Davis' developed such a scheme for predicting the solubility of gypsum (CaSO, ¦ 2H,O) in brines at temperatures u~ to 212F. However, for the higher temperaturei of present interest, the stable solid phase is anhydrite (CaSO,). This study involves a two-part method for predicting anhydrite solubility products. First, one predicts a value K, for a given ionic strength I and a given temperature T corresponding to mo./msr, = 1 from data of the CaS0,-NaCI-H,O system. Second, one determines a group of factors F .= F(Ca) • F(Mg) • F(SO,), where the individual factors account for increases in the solubility product due to complex formation by high concentrations, respectively, of calcium, magnesium and sulfate. Combining the two parts, one obtains the solubility product in molalities as

Jan 1, 1969

By D. G. Westlake

Polycrystalline tensile specimens of various Zr-0-H alloys have been tested at 298°, 178°, and 77°K. Solute oxygen and hydride precipitates in quenched alloys made individual contributions to the yield strength at 0.2 pct strain which combined to produce a resultant strength increment, a,., Ductility changes which were ohserved can he interpreted in terms of the various oxygen and hydrogen concentrations, testing tem -peratures, and dispositions of the hydride. ADDITIONS of oxygen in solid solution were known to increase the yield and tensile strengths of polycrystalline zirconium as early as 1951.' More recently, the critical resolved shear stress (CRSS) for prism slip in zirconium single crystals was also shown to be affected by the solute oxygen impurity.' This latter work also demonstrated that large increments of strength could be contributed by the finely dispersed zirconium hydride precipitates that are present in quenched Zr-H alloys.3 It was concluded that the combined strengthening due to alloying could be expressed by where to is the increase in the CRSS due to solute oxygen alone and TH is the increase due to finely dispersed hydride precipitates. Eq. [I] is analogous to one used to express the combined strengthening effects of work hardening and neutron radiation damage.4 Eq. [1] was verified only indirectly and for only small amounts of the impurities—up to 0.14 at. pct 0 and 0.63 at. pct H. The present investigation was undertaken to obtain a more direct verification of the validity of the form of Eq. [1] for this system and also to determine the combined effects of oxygen and finely dispersed hydride precipitates on the tensile strength and ductility of polycrystalline zirconium. EXPERIMENTAL PROCEDURE Tensile specimens were machined from the same rolled billet of Kroll zirconium used in the earlier study.' These measured 38 by 4.7 by 0.5 mm and had 10-mm gage lengths which were 2.8 by 0.5 mm. Each specimen was ß-annealed in vacuo at 1173°K for 15.5 hr and a-annealed at 1073°K for 4 hr to D. G. WESTLAKE, Member AIME, is Associate Metal l ur-gist, Metallurgy Division, Argonne National Laboratory, Argonne, III. Manuscript submitted July 17, 1964. IMD______________ give an equiaxed structure with grain diameters averaging 0.06 mm. Oxygen was added by allowing the metal to react with a known quantity of oxygen during the 0 anneal and known quantities of hydrogen were added during the a anneal. Each alloy was encapsulated in Pyrex under vacuum, annealed at 873°K for 4 hr, quenched into ice water, and polished by immersion in a solution of 46.75 vol pct H2O, 46.75 vol pct concentrated HNO3, and 6.5 vol pct HF (49 pct) at 298°K. Special heat treatments given to a few specimens are described in the results below. Tensile tests were done on an Instron machine and were begun within 20 min after the quench, except where specified otherwise. Tests at 298°K were in air, at 178°K in acetone, and at 77°K in liquid nitrogen. All tests were at a strain rate of 8x sec-1. RESULTS AND DISCUSSION Yield Stress at 298°K. The compositions of alloys and the corresponding yield stresses (0.2 pct strain) are given in Table I. A plot of the yield stresses of the oxygen alloys, A, B, C, and D, indicates that varies linearly with CO1/2, where Co is the oxygen concentration, Fig. 1. This is in accord with Fleischer's6 theory for solution strengthening if the oxygen atoms do not cluster, or the cluster size remains constant with increasing oxygen concentration. In Fig. 1, it appears that if one could prepare some oxygen-free zirconium its yield stress would be very low. Therefore, we shall assume that for the oxygen alloys is equivalent to O0, the strength increment contributed by the presence of oxygen. The relationship between0.2and Co is expressed by 0.2 = 31.3 CO1/2, when the yield stress is in kg per sq mm and the concentration is in at. pct. Each of the hydrogen alloys, Al, A2, A3, and A4, contained 0.081 at. pct 0 as an impurity. In Fig. 1, it appears that this small amount of oxygen makes a significant contribution to the strength which cannot be ignored when we evaluate the contribution of the finely dispersed hydride. Let us assume the validity of the following equation: a0.2 = (a2o+a2R)1/2 [2] which is analogous to Eq. [I] for single crystals, and calculate values of UH for the hydrogen alloys by using the experimental values of 0.2 and o (0.081 at. pct) = 8.9 kg per sq mm. For 0.36 at. pct H, oH = 6.47; for 0.72 at. pct H, OH = 11.30; for 2.16 at. pct H, OH = 19.4; and for 3.60 at. pct H,

Jan 1, 1965

By Rudolph G. Wuerker

Charles T. Holland (Head, Dept. of Mining Engineeri*, Virginia Polytechnical Inst., Blacksburg, Va.) Mr. Wuerker has presented a very interesting discussion of the use of triaxial test methods for investigating the strength properties of rocks. Such methods, no doubt, eventually will develop considerable information of interest to those concerned with the design of mine layouts, particularly in the field of pillar design. From his discussion of my recent article, "Cause and Occurrence of Coal Mine Bumps" (Holland Mining Engineering 1958, p. 933-1002), it is evident that in one place at least I did not make my meaning clear to him and perhaps others. To clear the matter up I think it best to quote from the article, somewhat more fully than did Mr. Wuerker, as follows: "4) In actual operations — because rocksare not perfectly elastic, homogeneous, nor isotropic and because local yield does occur — the maximum stress as demonstrated by Phillips (Ref. 22, pp. 64, 65) and indicated by much experience in mining, does not occur at the walls of the opening but at a short distance inside the pillar. Furthermore, the maximum stress does not reach as great a value as theoretical considerations and laboratory experimental methods indicate.* Actual distance inside the pillar, measured from the wall, at which the maximum stress exists, has not been determined. Observations in many mines, however, indicate that this distance could have a mini-value of one to six or eight times the bed thickness and that it is probably affected by width and height of the opening, depth of cover, and relative values of the elasticity and plasticity of materials comprising the roof, floor, and coal seam. The actual value of the stress produced probably lies between the theoretical maximum and the average stress concentration that would be produced if the weight of the strata above the unsupported opening were evenly distributed over the pillars for a distance equal to the opening width." The footnote reference in the above quotation referred to the following: "*For example, the Pocahontas No. 4 coal bed in southern West Virginia is mined under cover up to 1800 ft thick. Development openings are driven 18 to 20 ft wide, and the bed is about 6 ft thick. According to the work of Panek, the tangential wall stress at mid-bed height under these conditions would reach values between 4000 and 5000 psi. Actual tests of 3-in. cubes of this coal show its compressive strength would be much less than this, perhaps as low as 400 psi. Yet the pillars usually show no evidence of failure in these headings. In this same bed at a depth of 800 ft, the author has seen an opening 225 ft between supports lying between two old groves approximately 1100 ft apart. According to the theoretical considerations, the stress in the pillar walls would have been about 18,000 psi, yet the pillar showed little or no evidence of weight. In view of these observations, it is clear that the wall stress does not attain the maximum values indicated by theory." (Underlining added to original wording.) By referring to Fig. 2A of my paper it will be noted that theoretically the maximum pillar stress would occur at the pillar wall, i.e., at the passageway surface of the pillar. Obviously this cannot be correct in the cases of stress ranging from 4000 to 18000 psi since the coal at the surface of the pillar is under no constraint and cannot have a strength much greater than 400 or 500 psi. Hence, my conclusion that the maximum stress does not occur at the wall but back in the pillar some distance from the wall. Since these stresses are pushed back in the pillar from the wall, it is also obvious that the loads transferred to the pillar from the opening will be spread over a greater area and hence Pillar stresses will not rise to the values postulated by theory and photoelastic experiment. Further since to visual inspection the coal along the pillar wall did not appear to be failed the conclusion was reached that the stress shift was caused by local elastic or plastic yield and by difference in the elastic modulus of the rocks composing the mine floor, mine roof, and coal bed. Later on under the heading "Strength of Mine Pillars" (pages 1000-1002) the effects of constraint is briefly described. Also a formula taking into account constraint is developed relating pillar strength to the uniaxial strength of coal and the L/T ratio of the pillar. Since my paper was written, reports of experiments conducted in South Africa (Denkhaus, et. al., 1959), in Sweden (Hast 19581, and in Canada (McInnes, et.al., 1959) reveal that the conclusion expressed relative to the existence of a low stress area existing around the edges of pillars and solid faces as described above is generally correct. But it seems possible that where the wall stress developed is less than the unconfined strength of the rock composing the pillar and where the roof, floor, and pillar

Jan 1, 1961

By H. H. Kaveler, Z. Z. Hunter

Variation of the horizontal permeability (parallel to the bedding plane) in the vertical section of reservoir rocks has long been observed as a characteristic of a normally heterogeneous system which reservoir rock represent. The use of a recently developed water injection profile device offered opportunity to measure with a high degree of reliability the rate of inflow of water into Burbank sandstone in wells previously cored. Water injection profiles were not correlative with core permeability profiles in such wells. Apparently the vertical permeability substantially influences the flow between strata in a formation in a manner as to void the usual conclusions that have been drawn from consideration of the horizontal permeability measurements alone. The results obtained in comparing water injection profiles with horizontal permeability profiles suggest that many of the usual production operations based upon "selective" behavior or treatment of rock exposed in well bores need to he re-valuated and re-examined. INTRODUCTION Petroleum reservoir rock are heterogeneous systems. Heterogeneity exists in respect to lithologic character insofar as such rock are composed of distinguishable solid phases. Heterogeneity also exists in respect to certain properties, such as porosity and permeability, that vary due to variation of the physi-cal structure of the rock. Except in exceptional cases, both the horizontal permeability (measured parallel to the bedding planes) and the vertical permeability (measured perpendirularly to the bedding planes) exhibit significant variation in any common source of supply. The variation in horizontal permeability. as reflected by con. analyses. has drawn the greatest attention of petroleum technologists probably out of the general notion that the mass movement of fluids in a reservoir is predimonantly in the horizontal direction. Furthermore, in the usual case, the rock permeability measured in the horizontal direction is greater than that in the vertical. The variation of horizontal permeability of reservoir rock has been the basis for developing a number of operating practices and procedures intended to improve the petroleum production operation. Many such procedures are referred to as "selective" in the sense that the practice is intended to control the flow to a more. or less. permeable interval within the common source of Supply. It is often said that such practices are "tailored" to the permeability profile. The practices referred to involve, among others, the following: selective perforation of casing; selective shooting, acidizing and plugging: plugging back to intervals of low permeability; and, regulation of flow to prevent coning of water or gas, or irregular encroachment of water or gas. Certain expressed notions involving a concept of "by-passing," or "trappingl" that are held to be particularly harmful in causing the avoidable loss of recoverable petroleum have grown from observed variations in the horizontal permeability. Oftentimes estimates of the reserve of a common source of supply are tempered by conclusions relating variation in horizontal permeability to recover-ability of the oil-in-place. Certain conclusions attributed to the significance of the variation of the horizontal Permeabilitv often extend to the design and operation of pressure-maintenance projects involving both water flooding and gas-injection. Many advocate increasing the number of injection wells, advocate maintaining uniform and equidistant input-output well patterns, or advocate so-called "dispersed" gas-drive techniques rather than gas-cap injection because the permeability profile of cored wells is supposed to indicate that "by-passing" or "trapping" would otherwise exist. It is important, therefore, to have an opportunity to test whether the variation in the horizontal permeability found through core analyses of a typical reservoir rock is sufficient to establish the paths of fluid flow in a reservoir. It is particularly important to have an opportunity to determine whether flow at the sand face of a well conforms to the permeability profile as established by core analyses. In that manner, the merit of certain 50-called "selective" operating procedures and other notions may be evaluated. The purpose of this paper is to compare horizontal permeability profiles of wells in the Bartlesville (Bur-bank) sandstone with water injection profiles, for the purpose of showing that there is no correlation between the horizontal permeability of a core and the water intake characteristics of a typical sandstone. GENERAL CHARACTERISTICS OF BARTLESVILLE (BURBANK) SANDSTONE The Bartlesville sandstones of Northeastern Oklahoma are off-shore bar deposits.' Although other reservoirs had different processes associated with their deposition or with the formation of their porous, permeable structure, the l!artlesville sandstones on which these field Fields were made are, in every respect. typical petroleum reservoir rock. The permeability of the Bartlesville sandstones shows a typical variation in both the horizontal and vertical direction. Furthermore, the permeability profile logs of wells in any pool are not correlative, even as between wells as close as 660 ft and 330 ft apart.'. The same condition exists in such sand-tones as the Jones Sand at Shuler' and is the ordinary and usual characteristic of reservoir rock. THE FIELD DATA The data reported herein are those obtained from coring of nine wells on the center of ten-acre locations for the purpose of providing water-injection wells in the Bartlesville (Burbank)

Jan 1, 1952

By Fred C. Bond

MOST investigators are aware of the present unsatisfactory investigatorsstate of information concerning the fundamentals of crushing and grinding. Considerable scattered empirical data exist, which andare useful for predicting machine performance and give acceptable accuracy when the installations and materials compared are quite similar. However, there is no widely accepted unifying principle or theory that can explain satisfactorily the actual energy input necessary canexplain commercial installations, or can greatly extend the range of empirical comparisons. Two mutually contradictory theories have long existed in the literature, the Rittinger and Kick. They were derived from different viewpoints and logically lead to different results. The Rittinger theory is the older and more widely accepted.'TheRittinger In its first form, as stated by P. R. Ritted.'tinger, it postulates that the useful work done in crushing and grinding is directly proportional to the new surface area produced and hence inversely proportional to the product diameter. In its second form it has been amplified and enlarged to include the concept of surface energy; in this form it was precisely stated by A. M. Gaudin' as follows: "The efficiency of a comminution operation is the ratio of the surface energy produced to the kinetic energy expended." According to the theory in its second form, measurements of the surface areas of the feed and product and determinations of the surface energy per unit of new surface area produced give the useful work accomplished. Computations using the best values of surface energy obtainable indicate that perhaps 99 pct of the work input in crushing and grinding is wasted. However, no method of comminution has yet been devised which results in a reasonably high mechanical efficiency under this definition. Laboratory tests have been reported- hat support the theory in its first form by indicating that the new surface produced in different grinds is proportional to the work input. However, most of these tests employ an unnatural feed consisting either of screened particles of one sieve size or a scalped feed which has had the fines removed. In these cases the proportion of work done on the finer product particles is greatly increased and distorted beyond that to be expected with a normal feed containing the natural fines. Tests on pure crystallized quartz are likely to be misleading, since it does not follow the regular breakage pattern of most materials but is regularrelativelybreakage harder to grind patternat the finer sizes, as will be shown later. This theory appears to be indefensible mathematically, since work is the product of force multiplied by distance, and the distance factor (particle deformation before breakage) is ignored. The Kick theory4 is based primarily upon the stress-strain diagram of cubes under compression, or the deformation factor. It states that the work required is proportional to the reduction in volume of the particles concerned. Where F represents the diameter of the feed particles and P is the diameter of the product particles, the reduction ratio Rr is F/P, and according to Kick the work input required for reduction to different sizes is proportional to log Rr /log 2." The Kick theory is mathematically more tenable than the Rittinger when cubes under compression are considered, but it obviously fails to assign a sufficient proportion of the total work in reduction to the production of fine particles. According to the Rittinger theory as demonstrated by the theoretical breakage of cubes the new surface produced, and consequently the useful work input, is proportional to Rr-l.V f a given reduction takes place in two or more stages, the overall reduction ratio is the product of the Rr values for each stage, and the sum of the work accomplished in all stages is proportional to the sum of each Rr-1 value multiplied by the relative surface area before each reduction stage. It appears that neither the Rittinger theory, which is concerned only with surface, nor the Kick theory, which is concerned only with volume, can be completely correct. Crushing and grinding are concerned both with surface and volume; the absorption of evenly applied stresses is proportional to the volume concerned, but breakage starts with a crack tip, usually on the surface, and the concentration of stresses on the surface motivates the formation of the crack tips. The evaluation of grinding results in terms of surface tons per kw-hr, based upon screen analysis, involves an assumption of the surface area of the subsieve product, which may cause important errors. The evaluation in terms of kw-hr per net ton of —200 mesh produced often leads to erroneous results when grinds of appreciably different fineness are compared, since the amount of —200 mesh material produced varies with the size distribution characteristics of the feed. This paper is concerned primarily with the development, proof, and application of a new Third Theory, which should eliminate the objections to the two old theories and serve as a practical unifying principle for comminution in all size ranges. Both of the old theories have been remarkably barren of practical results when applied to actual crushing and grinding installations. The need for a new satisfactory theory is more acute than those not directly concerned with crushing and grinding calculations can realize. In developing a new theory it is first necessary to re-examine critically the assumptions underlying

Jan 1, 1953

By Fred C. Bond

MOST investigators are aware of the present unsatisfactory investigatorsstate of information concerning the fundamentals of crushing and grinding. Considerable scattered empirical data exist, which andare useful for predicting machine performance and give acceptable accuracy when the installations and materials compared are quite similar. However, there is no widely accepted unifying principle or theory that can explain satisfactorily the actual energy input necessary canexplain commercial installations, or can greatly extend the range of empirical comparisons. Two mutually contradictory theories have long existed in the literature, the Rittinger and Kick. They were derived from different viewpoints and logically lead to different results. The Rittinger theory is the older and more widely accepted.'TheRittinger In its first form, as stated by P. R. Ritted.'tinger, it postulates that the useful work done in crushing and grinding is directly proportional to the new surface area produced and hence inversely proportional to the product diameter. In its second form it has been amplified and enlarged to include the concept of surface energy; in this form it was precisely stated by A. M. Gaudin' as follows: "The efficiency of a comminution operation is the ratio of the surface energy produced to the kinetic energy expended." According to the theory in its second form, measurements of the surface areas of the feed and product and determinations of the surface energy per unit of new surface area produced give the useful work accomplished. Computations using the best values of surface energy obtainable indicate that perhaps 99 pct of the work input in crushing and grinding is wasted. However, no method of comminution has yet been devised which results in a reasonably high mechanical efficiency under this definition. Laboratory tests have been reported- hat support the theory in its first form by indicating that the new surface produced in different grinds is proportional to the work input. However, most of these tests employ an unnatural feed consisting either of screened particles of one sieve size or a scalped feed which has had the fines removed. In these cases the proportion of work done on the finer product particles is greatly increased and distorted beyond that to be expected with a normal feed containing the natural fines. Tests on pure crystallized quartz are likely to be misleading, since it does not follow the regular breakage pattern of most materials but is regularrelativelybreakage harder to grind patternat the finer sizes, as will be shown later. This theory appears to be indefensible mathematically, since work is the product of force multiplied by distance, and the distance factor (particle deformation before breakage) is ignored. The Kick theory4 is based primarily upon the stress-strain diagram of cubes under compression, or the deformation factor. It states that the work required is proportional to the reduction in volume of the particles concerned. Where F represents the diameter of the feed particles and P is the diameter of the product particles, the reduction ratio Rr is F/P, and according to Kick the work input required for reduction to different sizes is proportional to log Rr /log 2." The Kick theory is mathematically more tenable than the Rittinger when cubes under compression are considered, but it obviously fails to assign a sufficient proportion of the total work in reduction to the production of fine particles. According to the Rittinger theory as demonstrated by the theoretical breakage of cubes the new surface produced, and consequently the useful work input, is proportional to Rr-l.V f a given reduction takes place in two or more stages, the overall reduction ratio is the product of the Rr values for each stage, and the sum of the work accomplished in all stages is proportional to the sum of each Rr-1 value multiplied by the relative surface area before each reduction stage. It appears that neither the Rittinger theory, which is concerned only with surface, nor the Kick theory, which is concerned only with volume, can be completely correct. Crushing and grinding are concerned both with surface and volume; the absorption of evenly applied stresses is proportional to the volume concerned, but breakage starts with a crack tip, usually on the surface, and the concentration of stresses on the surface motivates the formation of the crack tips. The evaluation of grinding results in terms of surface tons per kw-hr, based upon screen analysis, involves an assumption of the surface area of the subsieve product, which may cause important errors. The evaluation in terms of kw-hr per net ton of —200 mesh produced often leads to erroneous results when grinds of appreciably different fineness are compared, since the amount of —200 mesh material produced varies with the size distribution characteristics of the feed. This paper is concerned primarily with the development, proof, and application of a new Third Theory, which should eliminate the objections to the two old theories and serve as a practical unifying principle for comminution in all size ranges. Both of the old theories have been remarkably barren of practical results when applied to actual crushing and grinding installations. The need for a new satisfactory theory is more acute than those not directly concerned with crushing and grinding calculations can realize. In developing a new theory it is first necessary to re-examine critically the assumptions underlying

Jan 1, 1953

By E. Miller, R. Geffken, K. L. Komarek

The thermodynamzc properties oJ liquid Cd-Sb alloys were investigated using the cell arrangement measurments were obtained every 2°C at a heating and cooling rate of 12°C per hr and at equilibrium every 2O0C frorn 500°C down through the stable liquidus. The S-shaped asCd US composition curve was used in the cotnposition regzon near Cd,Sb, to calculate a tempeerture-dependent inleraction coefficient from quasichemical theory. Rapid changes in a scd were observed at a transition temperature varying from 400" to 465°C depending on con/kosition. It could not be determined if the changes in aScd were discontinuous, but tlze composition dependeke of the magnitude of the change is indicative of a second-order phase transformation in the liquid. The values of the experimental changes in ASCd are in agreement with calculations from the slope of the transition temperature, using the concept that a second-ovdev phase transition occurs in liquid Cd-Sb alloys. II is suggested that the transformalion is associated with the formation of Cd4Sb3 molecules in the liquid. ThE structure of liquid alloys is the subject of many investigations. X-ray, resistivity, and thermody-namic data have been interpreted as indicating varying degrees of short-range order in the liquid in alloy systems forming inter metallic compounds. In general, the melting process is not a transition from an ordered to a completely disordered state, but some degree of order is retained in the liquid. Maximum ordering in the liquid state occurs close to the melting temperature of the compound and the arrangement of atoms becomes more random at higher temperatures. Of special interest in this respect is the Cd-Sb system. It is one of the few metallic systems which form both stable and metastable compounds when liquid alloys are cooled at normal rates. The stable system exhibits an intermetallic compound, CdSb, melting at 459"c.l A second compound, CdrSbs, has also been reported,' melting close to this temperature. The metastable system has one compound, CdsSbz, melting at 420"c.I Resistivity measurements on liquid Cd-Sb alloys close to the liquidus temperatures have been interpreted in terms of a complex ordering behavior which changes rapidly with increasing temperature.3 The resistivity-composition curve is characterized by two maxima corresponding in composition to CdSb and CdsSbz. The resistivity-temperature plots show sharp breaks for alloys in the composition range of 45 to 70 at. pct Cd on cooling through a transition temperature close to the stable liquidus. Fisher and phillips4 investigated the influence of temperature and composition on the viscosity of liquid CdSb alloys. The viscosity of some alloys increases sharply on supercooling below the stable liquidus. A maximum in the viscosity-composition curve occurs at the composition CdSb. The thermodynamic properties of liquid Cd-Sb alloys have been investigated by Seltz and ~e~itt' and Elliott and chipmane by the electromotive-force method and their results are in good agreement. However, these investigations were carried out at temperatures well above the liquidus temperatures of the alloys, and the temperature coefficients of the electromotive force, dE/dT, were obtained from experimental points for each alloy at a few temperatures considerably above the liquidus temperature. Scheil and ~aach' investigated the thermodynamic properties of this system by the dew point method in the temperature range from about 100°C above the stable liquidus down into the supercooled liquid region. They reported several anomalies, i.e., the activity of a melt on heating differed from that on cooling, and the activity increased sharply in the limited temperature interval immediately above the liquidus temperature of the stable alloy, followed by a sudden decrease below the liquidus. Values obtained on heating and cooling were not in agreement. A reinvesti-gation of a few alloys by Scheil and Kalkuhl' by the electromotive-force method failed to confirm these observations. The authors concluded that the anomalies were due to inhomogeneities in the starting alloys and they discarded their previous results. The present investigation was undertaken in order to obtain thermodynamic data close to the liquidus temperature and in the supercooled region where the anomalies were originally reported, employing the electro motive-force method. This method is quite precise and will most easily permit observations of small changes in activity and partial molar entropy with temperature. Measurements were taken every few degrees so that the dE/dT values could be calculated over the entire temperature range and small changes in the thermodynamic properties close to the liquidus temperature could be observed. I) EXPERIMENTAL PROCEDURE Specimens were prepared from 99.999+ pct Cd and Sb (Cominco). Surface oxide was removed by scraping and then melting the metals under vacuum and filtering through Pyrex wool. Appropriate amounts of the metals were weighed on an analytical balance to k0.1 mg, sealed in double Pyrex capsules under vacuum,

Jan 1, 1968