Search Documents

By E. T. Turkdogan, W. L. Worrell

The rate of reaction of hydrogen sulfide with ferrous sulfide was studied by measuring the initial rates of sulfidation of iron strips in hydrogen sulfide-hydrogen-argon mixtures at 670°, 800°, and 900" C. The time-dependent surface sulfur activity is derived from the instantaneous rate of sulfidation with the assumption that diffusion in the sulfide layer is in a pseudo-steady state with the gas-sulfide chemical reaction. The rate of sulfur transfer from hydrogen sulfide to the surface of iron sulfide is proportional to the partial pressure of hydrogen sulfide and inversely proportional to the activity of sulfur at the surface of the sulfide layer. The derived rate equation is based on the assumption that most of the surface sites on the chemisorbed layer of iron sulfide are occupied by sulfur atoms and that the slow rate-controlling reaction is the dissociation of hydrogen sulfide on the chemisorbed layer. The experimental results are in reasonable accord with this reaction model. 1 HE slow approach to parabolic growth rate in the sulfidation of iron in hydrogen sulfide-hydrogen mixtures is a manifestation of slow approach to surface equilibrium between the gas and the surface of the sulfide layer. For example, at 800°C and Fig. 7 in Part I, the parabolic growth rate begins after approximately 1 day of sulfidation time. With decreasing temperature and decreasing partial pressure of hydrogen sulfide, the time necessary to reach gas-sulfide surface equilibrium is much longer. These observations are similar to those reported previously by Turkdogan et al.' on the oxidation of iron to wustite in hydrogen-water vapor mixtures. The slow approach to gas-sulfide equilibrium is well-demonstrated by the results in Fig. 1 where the square of the weight gain per unit area, ( g S per sq cmI2, is plotted against time for gas mixtures having PH g//>H = 1.0 with 0, 25, 50, and 67 pct Ar at 800°C. The value of 1.5 X lo-' (g S per sq cm)' corresponds to almost complete sulfidation of the iron strip (- 0.05 cm thick) to iron sulfide. The points fall on S-shaped curves. If the inflection parts of the curves are considered to be linear, thus indicating parabolic growth rate and the establishment of gas-sulfide equilibrium, their slopes would have to be the same for a fixed ratio in the gas mixture. Such is not the case, and the slopes for the "linear" parts of the S-shaped curves are lower than the value when gas-sulfide surface equilibrium is established. The subject matter of this paper is the kinetics of the surface reaction of hydrogen sulfide with iron sulfide during the early stages of sulfidation of an iron strip. EXPERIMENTAL RESULTS The apparatus and materials used were the same as those discussed in Part I. In the present experiments of relatively short duration, a purified iron strip, 5 by 2 by 0.05 cm, which was suspended in the uniform hot zone of a vertical zircon tube, reacted with a flowing gas mixture of hydrogen sulfide-hydrogen-argon. The sample was suspended from a gold chain attached to a calibrated silica spring. The amount of sulfur picked up by the iron, which formed a layer of iron sulfide, was determined by measuring the displacement of a reference point on the silica spring, using a cathetom-eter. In all cases, the degreased sample was first annealed in a stream of oxygen-free dry hydrogen for several hours to remove any impurities such as oxygen, nitrogen, or carbon which might have been present on the surface of the sample.' The rate measurements were carried out at 670°, 800°, and 900° C in hydrogen sulfide-hydrogen-argon mixtures with pH2 s/Ph2 ratios from 4 to 0.1. In a few experiments of short duration no sulfur deposition was observed in gas mixtures with (comments concerning sulfur deposition at high pH s/ph ratios were made in Part I.) Typical examples of the results obtained are shown in Fig. 2 for two temperatures and two gas compositions. With the present experimental technique, reproducible rate data could be obtained only after a uniform

Jan 1, 1969

By Richard B. Reese, George R. St. Pierre, Robert A. Rapp

The vapor pressures of pure iron and pure chromium have been measured using a Knudsen cell coupled with a mass spectrometer. The experimental results agree well with some previously reported data; heats of sublimation and vaporization have been calculated. From vapor pressure rr~easurernents, the activities of iron and chromium in their binary alloys have been determined for NcY = 0.10 to 0.65 over a 1400" to 1700~ temperature range. The activity of chromium in the liquid alloys was found to exhibit positive deviation from ideal behavior when referred to solid chromium as the standard state; however, for the activity of chromium referred to nonequilibrium undercooled liquid chromium as standard state, the activity behavior is almost ideal. ALTHOUGH the Fe-Cr system is important as the basic system of alloy steels, the thermodynamic properties of this system have not been thoroughly investigated at elevated temperatures. In view of the potential practical applications of such data as well as their contribution to alloy theory, further study of the Fe-Cr system was undertaken. In addition, further development of experimental techniques for equilibrium measurements at high temperature is required. The present research involved an application of the Knudsen effusion method in conjunction with a time-of-flight mass spectrometer. Belton and ~ruehan' have measured the activities in Fe-Ni and Fe-Co alloys by a similar method. Speiser, Johnston, and lackbburn' have determined the vapor pressure of pure solid chromium between 1283" and 1561°K by a Langmuir free evaporation technique. The vapor pressure of chromium was described by the expression: over the temperature range 1283" to 1561°K. McCabe, Hudson, and paxton3 have measured the vapor pressures of pure iron and pure chromium by the Knudsen cell technique around 1500°K. The vapor pressure and heat of sublimation at O°K, AH,", for pure chromium agreed within experimental error with the measurements of Speiser et al. at 1200°C. Kubaschewski and HeYmer~ determined the vapor pressure of pure solid chromium in the temperature range 1170" to 1400°C by the Knudsen effusion technique combined with a radioactive tracer analysis of crS1. Their results agreed with the work of previous investigators. Edwards, Johnston, and Ditmars5 have measured the vapor pressure of pure solid iron by the Langmuir technique between 1356" and 1519°K. The vapor pressure of solid iron in the temperature range studied was given by: The data of McCabe, Hudson, and paxton3 on the vapor pressure of pure iron around 1500°K by the Knudsen effusion technique were in good agreement with those obtained by Edwards et a~.~ Hultgren, Orr, Anderson, and Kelley give the selected values for the AH: of pure iron and pure chromium as 98,990 *200 and 94,340 * 500 cal per mole, respectively. These values are the average values determined in the several investigations by different methods. The Fe-Cr phase diagram7 is shown in Fig. 1. Previous determinations of activities in this system have been carried out by several investigators, but most of the work has been done on solid alloys. Using a Knudsen effusion method with cr5', Kubaschewski and Heymer4 reported that the Fe-Cr system exhibits positive deviation from the ideal behavior for compositions ranging from NCr = 0.0398 to 0.699 at 1170"

Jan 1, 1969

By R. M. Douthwaite, G. Birkbeck

The effect of grain size and specimen dimensions on the occurrence and magnitude of a yield point drop in pure iron has been examined. The tests were done at room temperature. The results indicate that gain size is the critical factor determining whether a yield drop occurs. The specimen diameter:gain size ratio does not influence the magnitude of the yield drop for the range of gain size and specimen diameters investigated. HUTCHINSON,' petchY2 and worthington3 have investigated the grain size dependence of the upper yield stress (uUYP) and the yield point drop, ouyp - aLYP, in Armco iron, mild steel, and Si-Fe, respectively. These authors used center-annealed wires to avoid stress concentration from specimen shoulders and to ensure uniformity of the applied stress. Coarsegrained specimens show no yield point drop; the yield point drop then increases as d-'I2 increases (where d is the grain diameter). These results can be understood in terms of Petch's analysis2 in which the magnitude of amp depends on the strain rate in the (limited) number of grains deforming plastically at the upper yield point. If N grains per unit volume are undergoing plastic deformation, the strain rate in the deforming grains is l/Nd3 times higher at the upper yield stress than when the whole specimen deforms homogeneously. If N is constant and independent of grain size at uUyp, then all the grains are deforming when ~d~ = 1, so there is no yield point drop. At finer grain sizes, the yield point drop is given by: where AU, is the friction stress increase corresponding to a tenfold strain rate increase. Christ, Smith, and Burton studied the upper yield point in 0.050-in.-diam wire specimens prepared from zone-refined iron. Each end of the 2; -in. specimens was soldered to a & -in. diam steel ball and the load applied through the two balls. There was a yield point drop in specimens with grain diameters less than 0.1 mm; no drop was observed in specimens of grain diameter greater than 0.1 mm that were free from veining substructure, but specimens of grain diameter 0.129 mm, annealed in the y range, showed an upper yield point. and this was associated with the presence of a veining substructure in these specimens. Christ et al. suggested the upper yield point occurs when more than a specific number of grains fill the specimen cross section so a multiplicity of slip is required to initiate plastic deformation. Table I summarizes the published data on the grain size at which the yield drop disappears. On the basis of this data the ratio of specimen diameter:grain diameter = 5 to 27 appears to be the critical value for the occurrence of a yield point drop. Thus, petch2 interprets the appearance and magnitude of the yield point drop in terms of the limited number of grains deforming plastically at uuyp, whereas Christ et a1. consider that the occurrence of the yield drop depends on the ratio specimen diame-ter:grain diameter. The published data cover a wide variety of materials so some variation between the results is to be expected, but the size range covered in the data of Table I is rather limited. In an attempt to discover whether the specimen diameter:grain size ratio or the grain diameter is critical in determining

Jan 1, 1969

By C. T. Liu, J. Gurland

The deformation behavior in tension of spheroidized carbon steels was studied at room temperature as a function of carbon content, 0.065 to 1.46 wt Pct, and carbide particle size, 0.88 to 2.77 p. It was found that the Hall-Petch strength-grain size relation is directly applicable to the yield and flow stresses of the two lower-carbon steels , 0.065 and 0.30 pct C. The strength data for the medium- and high-carbon steels, 0.55 to 1.46 pct C, also satisfied the Hall-Petch relation, provided that these data are based upon the particle spacing. Beyond 4 pct strain, the flow stress data of all the steels studied could be represented by the same Hall-Petch relation with dinerent spacings for grain boundary and particle strengthening. The behavior of the higher-carbon steels was consistent with the postulated formation of a dislocation cell network during processing and initial deformation (up to 4 pct strain). The cell size was assumed to be equal to the planar particle spacing. The true stress at the ultimate tensile strength was also found to be a function of the particle spacing. At a given temperature and strain rate, the yield and flow stresses of carbon steels depend on the type and dimensions of the microstructure. Starting with the work of Gensamer et al. in 1942,' experimental studies on pearlitic and spheroidized carbon steels revealed that the strength of steels is a function of two main parameters: the ferrite grain size2'3 and the carbide particle spacing;1'4'5 on this basis, two different strengthening mechanisms have been developed to apply to steels of low and high carbon contents, respectively. In polycrystalline iron and mild steels the grain boundaries are regarded as the major structural barriers to slip. The relation between strength and grain size is generally represented by the Hall-Petch equation which is based on a linear proportionality between strength and the inverse square root of the average grain size.2'3y677 However, Gensamer et al.' and Roberts et related the yield strength of medium -and high-carbon steels to the carbide particle spacing alone, and they found a linear relation between the logarithm of the mean free path in the ferrite and the yield strength in both spheroidized and pearlitic steels. By means of the electron microscope, Turkalo and LOW' extended the study to finer structures; they concluded that the logarithmic relation is not valid for the entire range of microstructures unless grain boundaries are also included in the measurement of the mean free path. For the specific case of spheroidized steels, Ansell and aenel' found that the yield strength data,4'5 when plotted as a function of mean free path, fit the Hall-Petch equation; however, T'ysong found that the same data fit the 0rowanl0 relation if a planar inter-particle spacing is used. Recently Kossowsky and ~rown" studied the strength of prestrained spheroidized steels, 0.48 and 0.95 pct C, and concluded that the strength due to the carbide dispersions varies linearly with the reciprocal of the square root of the mean free path between carbide particles and dislocation networks. Such networks were first observed by Turkalo." The conclusion common to all these studies is that the available slip distance in the ferrite is the most important variable in determining strendh. Previous work on carbon steels is restricted to limited composition and strain ranges. The mechanism which governs the flow properties is not clearly understood, and, in particular, little is known about the composition dependence of the transition between grain boundary strengthening and particle hardening. The purpose of the present work is to investigate the strengthening mechanism in spheroidized steels over a wide range of carbon content, 0.065 to 1.46 wt pct, and plastic strain, yielding to necking. The spheroidized structure was chosen because of its relative simplicity and the relative ease of control and measurement of the structural parameters. The experimental work is limited to tensile testing at room temperature at constant extension rate. The effects of the carbide particles on the fracture behavior of spheroidized steels are discussed elsewhere.13 EXPERIMENTAL PROCEDURE Eight different grades of vacuum-cast carbon steels were supplied in the form of forged and rolled plate by the Applied Research Laboratory of the U.S. Steel Corp. The compositions furnished with these steels are given in Table I; the carbon content ranges from 0.065 to 1.46 wt pct, or from 1.0 to 22.3 vol pct of carbide. The steel plates were cut transversely into rods a little larger than the test specimens, 1 in. gage length, i in. diam. The rods were austenitized in air (enriched with CO by a consumable carbon-rich muffle) at 50° C above theA, orA., temperature for 2 hr and then quenched in oil with vigorous stirring. The as-quenched rods were tempered in two stages in order to obtain the desired distributions and sizes of carbide particles. The rods were first tempered at 460° C for 10 hr and then at 700" C for periods ranging from 4 hr to 3 days, in vacuum. After final machining, all specimens were vacuum-annealed again at 650°C for 1 hr in order to relieve residual stresses. The tension tests were carried out in two steps. The initial part of the load-strain curve, up to about 2 pct strain, was determined on a Riehle testing machine with an extensometer of small strain range, 4 pct strain, in order to obtain the yield and initial flow piopertiesi As soon as the first part of the test was finished, the specimen was placed in an Instron testing machine equipped with a strain gage extensometer with a maximum strain range of 50 pct. The load-strain curve to fracture was

Jan 1, 1969

By E. Miller, K. Komarek, W. Johnson

The activity of aluminum in solid Cr-A1 alloys has been measured by an isopiestic technique between Cr-A1890' and 1126" and 13 and 80 at. pct Al. The integral free energy of mixing has a minimum value of —5600 cal per g-atom at 59 at. pct Al. The maximum solid solubility of aluminum in chromium was determined to be 43 at. pct Al, and the composition limits of the compounds CrA14, Cr4A19, and Cr5Al, at 1000"~ were found to be 79 to 80, 66 to 70, and 59 to 63 at. pct Al, respectively. The thermodynamic properties of the Cr-A1 system have been investigated as part of a thermodynamic study of aluminum-transition metal systems.172 Little information is available on the equilibrium properties of the Cr-A1 system. The heats of formation of solid Cr-A1 alloys have been determined by Kubaschewski and Haymer at 600" and low-temperature specific heat data have also been obtained.~ More extensive work has been performed on the phase diagram, and a compilation has been provided by Hansen and Anderko,~ their phase diagram at elevated temperatures being essentially based on the work of Bradley and LU.~ The high-temperature portion of the phase diagram shows an intermediate phase CrA14 decomposing peritectically at 1018°C and existing at 82 at. pct A1 at 1000°C. They also identified the compounds with solubility limits of 72 to 75 at. pct A1 at 1000°C, and Cr5A1,, existing at 61 at. pct A1 at 1000°C. The maximum solid solubility of aluminum in chromium at 1000°C was found to be 46 at. pct Al. These elevated-temperature data were obtained by examination of quenched samples and were considered as less precise than the lower-temperature data. Koester, Wachtel, and Grube7 have revised the phase diagram as a result of their magnetic susceptibility and X-ray study. The results of this work differ appreciably from those of Bradley and Lu at temperatures above 800°C. The CrA1, compound is given as existing between 79 and 81 at. pct A1 at 1000°C, and they do not indicate the presence of a CrA13 phase reported by Bradley and Lu. They also report the compound Cr4Alg as having solubility limits of 66 to 70 at. pct A1 at 1000°C, while Bradley and Lu show this compound stable only up to 870°C. Koester et al. state that the high-temperature modification of the compound Cr5A18 is stable down to 1125"C, and not 980°C as stated by Bradley and Lu, and that the low-temperature modification of Cr5Al, has a range of homogeneity of 58 to 63 at. pct A1 at 1000°C. They also report that the maximum solid solubility of aluminum in chromium is 43 at. pct A1 at 1000°C. APPARATUS AND EXPERIMENTAL PROCEDURE An isopiestic method was employed which has been successfully applied to the determination of aluminum activities in solid ~e-All and Ni-Al alloys. Alloy specimens were held at different positions in a temperature gradient and were equilibrated with aluminum vapor from an aluminum reservoir kept at the temperature minimum of an impressed thermal gradient in a closed alumina system. Diffusion of aluminum into the specimens occurred until equilibrium was reached, at which the partial pressure of aluminum in each of the specimens was given by the vapor pressure of the pure aluminum reservoir. The activity of aluminum referred to liquid aluminum as the standard state in a given equilibrated sample at temperature T could therefore be expressed by: vapor pressure of pure aluminum at _ the temperature of the reservoir Vapor pressure of pure liquid aluminum, at specimen temperature T Since both the temperature of the aluminum reservoir and the specimen temperatures were determined experimentally, and the vapor pressure of pure aluminum is known as a function of temperature,' the activity of aluminum in a given aluminum alloy of known composition could be calculated. Initial runs were made with samples consisting of pure chromium chips placed in alumina crucibles. These runs exhibited large inconsistencies, indicating that equilibrium was not attained. High aluminum content Cr-A1 alloy powders were therefore substituted for the pure chromium specimens. The starting composition of the alloys was adjusted through experimentation until the concentration change necessary to attain equilibrium was small. In this manner, consistent results were obtained in reasonable times. SPECIMEN PREPARATION Alloy specimens were prepared from chromium of 99.997 pct minimum metallic purity: with 0.028 to 0.038 pct H, 0.0002 pct N, and 0.27 to 0.46 pct 0 (Aviquipo, Inc.). The aluminum had a purity of 99.99+ pct and the following impurities: 0.003 pct Cu; 0.002 pct S; 0.002 pct Fe; 0.001 pct Pb; 0.001 pct Ga (Aluminum Corp. of America). Alloy powders were prepared from weighed mixtures of chromium and aluminum by double-arc melt-

Jan 1, 1969

By K. M. Myles

The vapor pressure of palladium over a series of Rlz-Pd alloys has been measured by the torsion-effusion method. The thermodynamic properties of the alloy system at 1575=K have been calculated from the vapor pressure data. The activities and the free energies of formation exhibit large positive deviations from ideal behavior. The enthalpies of formation are endother-mic. The entropies of formation are positive and larger than the ideal entropy of mixing. All of the thermodynamic properties suggest that a strong tendency toward phase separation exists in the solid solutions. The possible origin of the phase instability and the various factors that influence the thermodynamic properties are discussed. RECENT studies of the thermodynamic properties of alloys of palladium with nontransition elements have indicated that a significant contribution to the enthalpy of formation is related to the redistribution of the conduction electrons upon alloy formation.'-' The present work was undertaken to ascertain the importance of this contribution in Pd/transition-metal alloys. The Rh-Pd system was chosen for this investigation for several reasons: 1) The thermodynamics of the system were unknown. 2) Rhodium and palladium are completely soluble at high temperatures; below 1118OK the solid solution becomes immiscible.', 3) The difference in magnitude between the vapor pressures of rhodium and palladium permitted the use of an existing effusion apparatus. 4) Additional information was known about the alloy system that would facilitate the interpretation of the thermodynamic results. EXPERIMENTAL PROCEDURE The thermodynamic properties of the Rh-Pd alloy system were calculated from the vapor pressure of palladium over solid palladium and over several solid Rh-Pd alloys. The vapor pressure was measured by means of the torsion-effusion apparatus that has been described previously. In this method, an effusion cell is suspended from a tungsten filament inside a high-temperature furnace. Two orifices are located eccentrically such that the effusion of the vapor creates a rotational torque in the filament. The angle of rotation is directly related to the total vapor pressure within the cell. As the vapor pressures of rhodium and palladium are approximately five orders of magnitude apart," the total vapor pressure was considered to be effectively equal to the equilibrium vapor pressure of palladium. The effusion cells were made from high-purity alumina since auxiliary experiments indicated that essentially no reaction occurs between alumina and solid Rh-Pd alloys. Unfortunately, because the orifices were irregular, an accurate calculation of the Free- man-Searcy correction factors" could not be made. The constants were determined in an independent experiment where the vapor pressure of copper, as measured in the alumina cell, was compared with an accurate value of the vapor pressure, which had been determined previously.4 Depletion of palladium from the surfaces of the specimens was minimal as the deflection of the cell remained constant, for at least 15 minutes, at each of the experimental temperatures. Lattice parameter measurements of the postrun alloys also indicated that no changes in the composition of the surfaces had occurred. The alloys were prepared by arc melting the requisite amounts of the 99.99 pct pure elements. The four most palladium-rich alloys were remelted in a levita-tion furnace since complete melting of the components had not occurred in the arc furnace. All of the alloys were subsequently homogenized at elevated temperatures in sealed alumina thimbles. After heat treatment, the alloys were analyzed chemically and were in essential agreement with the nominal compositions. The thermal histories and nominal compositions of the alloys are given in Table I. Lattice parameters of the heat-treated alloys were computed, by means of the method described by Mueller et a1. ,I2 from X-ray diffraction powder patterns obtained with filtered copper radiation in a 114.6-mm-diam Straumanis-type Debye-Scherrer camera. The results, which are tabulated in Table I, exhibit a slightly greater negative deviation from Vegards' law than the values reported by Raub et al.13 The diffraction lines were sharp and well-resolved and thus indicated that the alloys were homogenous. RESULTS The logarithms of the individual values of the vapor pressure of palladium were fit, by the least-squares method, as a linear function of the reciprocal of the absolute temperatures. The constants of the equations are given in Table I along with the temperature range over which the data was accumulated. The latent heat of vaporization at 298.15"K for pure palladium, calculated by the third-law method,14 showed no systematic temperature dependence. The average value of 88,920 * 20 cal per g-atom agrees favorably with the average of the results obtained in the most reliable previous investigations.15"19 The activities of palladium were computed from the vapor pressure data at 1525', 1575", and 1625OK. Consistent with the mass spectrometric study of the atomicity of palladium vapor,lg the vapor was assumed to be monoatomic. The activities of rhodium were determined by integrating the Gibbs-Duhem equation with the aid of the a function.20 In the calculations, the activities of the pure solid metals were assigned the value of unity. From the activities, the partial and integral free energies, entropies, and enthalpies of formation at 1575° K were computed; they are assembled in Table 11.

Jan 1, 1969

By W. F. Chiao

Ultrasonic attenuation, a, measurements in the frequency range of 5 to 55 mc per sec have been studied to determine their quantitative relationship with the following three variables of mixed microstructures in steels: 1) the volume percent, XF, of polygonal fer-rite in mixed structures of martensite and polygonal ferrite in Fe-Mo-B alloys: 2) volume percent, XA, of retained austenite plus martensite aggregates in high-carbon steel; and 3) substructural differences between 100 pct bainitic ferrite structures formed at various temperatures. The quantitative relationship obtained in the first two conditions by plotting a us the known structural parameters can be expressed, respectively, as: where al, a 2 and C1, Cz are constants. In the third condition the nature of the attenuation depends on the state of dislocations generated at the transformation temperatures and also on the alloy composition. From these measured results, the mechanism of ultrasonic attenuation caused by these mixed microstructures can also be studied. MUCH interest has recently been shown in the application of ultrasonic attenuation and wave velocity measurements to the study of the microstructural characteristics of steels. The general aims of most of the investigations in this field can be grouped into two categories: one is to study the mechanisms of ultrasonic losses caused by the characteristic phases in the microstructure of steel,''' and the other is to develop nondestructive test methods and applications for quality control.~' 4 Apparently no work has been done on the evaluation of ultrasonic attenuation meas -urements as a means of quantitative determination of a given phase in the microstructure of a steel. It is well-established that the decomposition of austenite results in four main microstructural constituents—polygonal ferrite, pearlite, bainite, and martensite—and that each phase has different mechanical properties. Thus, when a steel consists of mixed microstructures, the mechanical properties can often be related to a quantitative measure of the volume percent of each phase present. This study relates ultrasonic attenuation measurements to: 1) the volume percent of polygonal ferrite in mixtures of martensite and polygonal ferrite in Fe-Mo-B alloys; 2) the substructural differences between 100 pct bainitic ferrite structures formed at various temperatures; and 3) the vol- ume percent of austenite in austenite plus martensite aggregates in a high-carbon steel. The choice of the specimen materials was based on the laboratory stocks which were suitable to produce the required mixed microstructures for this study. EXPERIMENTAL PROCEDURES Materials and Heat Treatment. Polygonal Ferrite Plus Martensite Structures. This mixture of phases was produced in a vacuum-melted Fe-Mo-B alloy. The alloy was hammer-forged at 1900" ~ to a -f-in.-sq bar. By isothermally heat treating the alloy at 1300° F for various times and then water quenching, variations in the amount of polygonal (or proeutectoid) ferrite can be controlled in a microstructure in which the balance of the material is martensite. In the present work, four different times of isothermal transformation were adopted; after heat treatment, the four specimens were machined for ultrasonic measurements. The compositions, heat treatments, and dimensions of the four specimens are listed in Table I. 100 pct Bainite Structures Formed at Different Temperatures. It has been well-established by Irvine et al.= that the presence of molybdenum and boron in ferrous alloys can retard the formation of polygonal proeutectoid ferrite and expose the bainitic transformation bay, so that a more acicular or bainitic ferrite can be obtained over a wide range of cooling rates. Their investigation6 also showed that the mechanical properties of fully bainitic steels are usually closely dependent on the substructural characteristics of the steels. For studying the substructural characteristics in completely bainitic structures, six Fe-Ni-Mo alloys, of which five were free from carbon addition and one with 0.055 pct C addition, were selected so that a wide range of hardness values for 100 pct bainitic ferrite structures could be produced by normalizing at 1900" F followed by air cooling. The different bainitic transformation temperatures were recorded during air cooling. All of the alloys were vacuum-melted and then forged at 1900" F to square bars. Data on the six specimens of these structure series are summarized in Table 11. Austenite Plus Martensite Structures. The high-carbon steel used to study austenite plus martensite structures was vacuum-melted and then forged into Q-in.-sq bar. The series of mixed structures of austenite plus martensite was produced by quenching the specimens from the austenitizing temperature to room temperature and then refrigerating them at various temperatures within the range of martensite transformation to produce different amounts of retained austenite. Data on the four specimens of this series are listed in Table 111. Quantitative Analysis of the Microstructures. The microstructures containing martensite plus polygonal ferrite were analyzed by the point-counting technique.

Jan 1, 1969

By C. L. Magee, H. W. Paxton

PREVIOUS workersl'% ave found that as-quenched ferrous martensites can be plastically deformed at low stresses. They have also found that the stress to obtain small strains can be significantly increased by aging at temperatures up to 150°C. This behavior has been interpreted in two ways: 1) that the deformation at low applied stress is due to internal stresses and thus the aging phenomenon is due to relaxation of the internal stresses;' 2) that the microdeformation arises through the motion of unpinned dislocations and the aging process is the diffusion of carbon to the dislocations to pin them.293 In a recent paper,4 we have shown that under certain conditions the dominant microdeformation mechanism of as-quenched Fe-31Ni marten-site-austenite aggregates is the transformation of retained austenite. Our results on resistance changes during deformation and temperature dependence of the deformation also established that, at temperatures above 300° K, the microdeformation was not due to this transformation. The purpose of this note is to report results of experiments which indicate that explanation 1 is appropriate for this deformation. All results presented here are from conditions where transformation does not make a measurable contribution to the microplastic deformation, in Fe-31Ni-0.002C, or in Fe-29Ni-0.02C-0.25Ti. The experimental techniques that were used for measurement of the plastic deformation are described4 and analyzed5 elsewhere. Fig. 1 shows the microplastic response of 100 pct austenite at 300" and 350° K. The microplastic response of an austenite-martensite aggregate is also given in Fig. 1 and it is seen that, after formation of martensite, the initial plastic deformation occurs at lower stresses. Thus, the softness is not entirely due to retained austenite—the quenched structure can be deformed at lower stresses despite the fact that the macroscopic strength is significantly higher. A result which allows a distinction to be drawn between the alternative explanations for this behavior is derived from aging experiments on the low-carbon alloy containing titanium. We have previously shown4 that the presence of titanium in this alloy reduced the free carbon to a sufficiently low level that thermal stabilization of the austenite-martensite transformation was not detected. Despite this, Fig. 2 shows that the microplastic response is decreased by aging treatments. A similar aging effect in Fe-ZONi, with some titanium, has previously been reported by . Thus, carbon is not necessary for the aging process and explanation 2 is at best incomplete. Since internal stress relief can occur independently of carbon (any temperature dependence of the flow stress can result in decreased internal

Jan 1, 1969

By P. S. Kotval, H. Hahell

PRECIPITATION of M23C6 carbides at grain boundaries in austenitic steels and nickel base superalloys has received considerable attention. Hatwell and Berghezan,1 in a study using carbon extraction replicas, observed that dendritic growth of the M23C6 carbide phase occurred at grain boundaries in austenitic steels. Merrick and Nicholson2 have observed that in thin foils of "Nimonic 75" alloy the precipitation of M23C6 carbide was discontinuous at grain boundaries. Lewis and Hattersley3 have suggested that the type of M23C6 precipitation observed in Ni-Cr austenites represented the early stages of discontinuous precipitation. In a more recent study, Singhal and Martin4 have concluded that grain boundary M23C6 precipitation involves the migration of an austenite grain boundary by a mechanism very similar to that proposed by Tu and Turnbull.5 Tu and Turnbull's suggestion essentially involves grain boundary migration under the driving force arising from the difference between the surface free energies of the opposite sides of the plate. Thus, Singhal and Martin's observation4 implies that the first stage of discontinuous precipitation has taken place and that the grain boundary has provided a diffusion short circuit for the sub-stitutional solute atoms (principally chromium) which assist the preferential growth of the M23C6 plate in contact with the grain boundary. However, no evidence of the characteristic cellular morphology typical of discontinuous precipitation was found in Singhal and Martin's studies. Presumably, this is due to the low volume fraction (i.e., low supersaturation of solute) of the precipitating carbide reported by Singhal and Martin.4 Recently, Raymond6 has studied precipitation of the M23C6 carbide in Incoloy alloy 825 and shown that the growth of this phase is perpendicular to grain boundaries at 1400°F and in the plane of the grain boundary at 1700°F. The former aging condition resulting in discontinuous or "cellular" precipitation. Raymond has suggested an explanation for this behavior based on the relative rates of chromium diffusion by volume and grain boundary diffusion at the two different aging temperatures. In the course of a recent investigation on precipitation behavior in several ' strengthened superalloys, precipitation of M23C6 carbide at grain boundaries was studied. Under conditions of a high degree of supersaturation, the M23C6 carbide precipitates in the discontinuous or cellular morphology in Waspalloy (Ni-18.51 Cr-2.0 Fe-3.53 Mo-1.16 Al-2.78 Ti-0.09 C). This note presents electron microscopy observations of this mode of precipitation and its effect on mechanical properties. Specimens of Waspalloy after "solution" heat treatment at 1950°F show essentially single phase matrix with a few residual particles of "primary" MC carbide. In normal commercial heat treatments, precipitation of Ni3(Al,Ti) ordered fcc r' phase occurs throughout the matrix and M23C6 precipitates as globular particles at the grain boundaries. In these commercial heat treatments the degree of supersaturation during precipitation is low. However, if specimens of Waspalloy are quenched from the solution heat treatment temperature of 1950°F and then aged at an aging temperature where the degree of supersaturation is high, discontinuous precipitation of M23C6 occurs. Figs. 1 and 2 show thin foil micrographs of Waspalloy, solution heat treated at 1950°F, water quenched, and aged at 1292° and 1598°F, respectively, for 500 Fig. 1—Thin foil of Waspalloy, solution heat treated at 1950°F for 1 hr, water quenched, and aged for 500 hr at 1292°F. Magnification 15,500 times. Fig. 2—Thin foil of Waspalloy, solution heat treated at 1950°F for 1 hr, water quenched, and aged for 114 hr at 1598° F. Magnification 13,000 times.

Jan 1, 1970

By D. de Fontaine

A convolution procedure is described for obtaining an analytical expression for the composition profiles of solute concentration:; in binary solid solutions. The free energy of solutions containing such compositional inhomogeneities can then be evaluated and the stability of spherically symmetric clusters (G.P. zones, for short) us a function of their radii and central compositions can be studied accordingly. Two characteristic zone radii are defined: the minimum radius (RM) corresponding to the smallest zone for which growth is energetically favored over dissolution, and the critical radius RK) corresponding to zones associated with the minimum work of formation. The values of RM and RK coincide at the two classical limits of nucleation and growth and spinodal decomposition. As the spinodal is approached, however, RM and RK diverge increasingly and it is proposed that the initial] decomposition morphology is governed more closely by the minimum, rather than by the critical radius since the former possesses no singularity within the m eta stable phase boundary. Reversion is briefly discussed in the light of the foregoing and it is shown that a set of reversion temperatures located between the coherent miscibility gap and the coherent spinodal must correspond to a given distribution of low temperature zones. SINCE the early works of Guinier1 and preston,2 the kinetics and morphology of GP zones have been studied extensively, see Ref. 3, for example, for a recent review. Nevertheless, a completely satisfying theoretical treatment of the decomposition of meta-stable or unstable binary solid solutions is still lacking due to the almost unsurmountable mathematical difficulties that one encounters in attempting to solve the pertinent diffusion equation which, in principle, describes the whole process. This equation, first derived by Cahn for the initial stages of spinodal decomposition,4 is a nonlinear partial differential equation of the fourth order in anisotropic space, a general analytical solution of which is completely out of the question. From this equation, some prediction can be made, however, concerning the later stages of spinodal decomposition5 and a recent numerical solution6 of the equation in the case of one- and two-dimensional composition modulations of a solid solution inside the spinodal has led to the following considerations: the sinusoidal morphology7 is to be expected only for solutions of average composition c, located near the center of the coherent miscibility gap.8 For solutions with co closer to the spinodal decomposition, depleted shell zones9 should constitute the first recognizable decomposition product, later evolving to iso- lated, enriched preprecipitates in a uniformly depleted matrix as this must, in all cases, constitute the final metastable morphology, barring, of course, loss of coherency and change in crystal structure. In agreement with the qualitative description of Bonfiglioli and Guinier,l0 it thus appears that the various models discussed by previous authors are, in fact, idealized successive stages of the general decomposition process. Unfortunately, corresponding calculations for a solid solution outside the spinodal are not yet available because of two additional difficulties: it is essential that the diffusion equation be solved in three-dimensional space, and furthermore, one must include in the equation an appropriate "source term" describing the statistics of the creation of composition fluctuations required to trigger the nucleation events. At present, a general treatment of this problem does not appear to be tractable. Nevertheless, by using a modelistic approach, i.e., by selecting a priori certain composition variation profiles and by evaluating their associated free energies, one can draw certain conclusions relative to the stability of these isolated composition modulations or "zones", both outside and inside the spinodal. This is the object of the present investigation. THEORY We shall now describe a standard four-parameter spherically symmetric GP zone, evaluate its free energy according to the Cahn and Hilliard11 formula (Eq. [5], below), and plot the corresponding energy surfaces as a function of zone amplitude and radius as

Jan 1, 1970

By R. J. Fruehan, F. D. Richardson

Electrochemical measurements have been made of the activity of oxygen in copper and its alloys with silver and tin at 1100" and 1200°C. The galvanic cell used was Pt, Ni + NiO/solid ellectrolyte/[O] in metal, cermet, Pt The results do not support any of the equations so far designed for predicting the activities of dilute solutes in ternary solutions from activities in the corresponding binaries. If, however, a quasichemical equation is used with the coordination number set to unity, agreement between observed and calculated activities shows that this empirical relationship can be useful over a fair range of conditions. SEVERAL solution models have been proposed for predicting the activity coefficients of dilute solutes in ternary alloys from a knowledge of the three binary systems involved. Alcock and Richardson1 have shown that a regular model, and a quasichemical model,' in which the dissolved oxygen is coordinated with eight or so metal atoms, can reasonably predict the behavior of both metal and nonmetal solutes when the heats of solution of the solute in the separate solvent metals are similar. But when this is not so, neither model gives useful predictions unless coordination numbers of one or two are assumed. Wada and Saito3 subsequently adopted a similar model to derive the interaction energies for two dilute solutes in a solvent metal. Belton and Tankins4 Rave proposed both regular and quasichemical type models in which the oxygen is bound into molecular species, such as NiO and CuO in mixtures of Cu + Ni + 0. However, their models have only been tested on systems in which the excess free energies of solution of the solute in the two separate metals differ by a few kilocalories. Ope of the reasons why more advanced models have not been proposed is because few complete sets of data exist for ternary systems in which the solute behaves very differently in the two separate metals. For this reason measurements have been made of the activities of oxygen dissolved in Cu + Ag and Cu + Sn. Measurements on both systems were made by means of the electrochemical cell, Pt, Ni + NiO/solid electrolyte/O(in alloy), cermet,Pt [1] The activity of oxygen was calculated from the electromotive force and the standard free energy of formation of NiO, which is accurately known.5 Before investigating the alloys, studies were made of oxygen in copper to test the reliability of the cell and to check the thermodynamics of the system. Of the previous studies those by Sano and Sakao,6 Tom-linson and Young,7 and Tankins et al.8,7 have been made with gas-metal equilibrium techniques; those by Diaz and Richardson,9 Osterwald,10 wilder," Plusch-kell and Engell,12 Rickert and wagner,13 and Fischer and Ackermann14 have been made by electrochemical methods. EXPERIMENTAL The apparatus employed was the same as described previously,9 apart from slight modification. The molten sample of approximately 40 g was held in a high grade alumina crucible 1.2 in. OD and 1.6 in. long. The solid electrolytes were ZrO2 + 7½ wt pct CaO and ZrO2 + 15 wt pct CaO; the tubes 4 in. OD and 6 in. long were supplied by the Zirconia Corp. of America. They were closed (flat) at one end. In one experiment with Cu + O, both electrolytes were used in the cell at the same time. The reference electrodes inside the electrolyte tubes consisted of a mixture of Ni + NiO. They were made by mixing the powdered materials and pressing them manually into the ends of the tubes, with a platinum lead embedded in them. The tubes were then sintered overnight in the electromotive force apparatus at 1100°C. By sintering the powders inside the tubes (instead of using a presintered pellet9) better contacts were obtained between the electrolyte, the powder, and the platinum lead. Troubles arising from polarization9 were thus much reduced. The electromotive force was measured by a Vibron Electrometer with an input impedence of 1017 ohm; the temperature was measured with a Pt:13 pct Rh + Pt thermocouple protected by an alumina sheath. The couple was calibrated against the melting point of copper. The cermet conducting lead of Cr + 28 pct Al2O3, previously found to be satisfactory9 for use with Cu + 0 was also found satisfactory with Cu + Ag + 0 and Cu + Sn + 0. Superficial oxidation was observed, but it did not interfere with the working of the cell. The reaction tube containing the cell was closed at each end with cooled brass heads and suspended in a platinum resistance furnace. The tube was electrically shielded by a Kanthal A-1 ribbon which was wound round it, and the ribbon was protected by a N2 atmosphere between the furnace and the reaction tube. The cell was protected by a stream of high purity argon which was dried and passed through copper gauze at 450°C and titanium chips at 900°C. All the metals used were of spectrographic standard. Procedure. In studies of the system Cu + 0, be-

Jan 1, 1970

By G. L. F. Powell

g samples of Cu-20 wt pct Ag alloy have been mdercooled to a maximum of 197°C by melting under a slag of commercial soda-lime glass in a vitreous silica crucible. No grain refinement of the primary copper was observed in samples undercooled to the maximum of 197°C. When the samples contained a small amount of oxygen, the copper dendrites were partially recrystallized at undercoolings greater than 97°C. In previous papers'-3 reporting the grain structure of undercooled silver and copper, it was observed that grain refinement was dependent on both undercooling and oxygen content. Grain refinement occurred in undercooled silver when the degree of undercooling exceeded the range 153" to 175"C, while in Ag-0 alloys (0.12 wt pct) fine equiaxed grains were exhibited when undercooling was greater than 50°C. Similarly, copper samples undercooled as much as 208°C displayed fan-shaped growth from a single nucleation site, while the grain structure of Cu-O alloys (0.08 wt pct) was fine and equiaxed at undercoolings larger than 150°C. Thus the presence of oxygen greatly reduced the undercooling at which grain refinement occurred. It was also observed that the change in grain size resulted from recrystallization and was not due to an enhanced nucleation rate in the liquid-solid transformation. It is possible that the influence of oxygen on recrystallization is due primarily to its presence as a solute element. walker4,' reported that, although a grain size change did not occur in pure nickel until the undercooling exceeded 150°C, small grains were observed in samples of Ni-Cu alloy solidified at small and large degrees of undercooling. Jackson et al.6 suggested that the fine grained structure of the Ni-Cu alloy resulted from the melting off of dendrite arms during recalescence. This remelting process may occur in alloys as a result of segregation during freezing which causes a variation in liquidus temperature from point to point within a dendrite. It was therefore decided to undercool copper with a metallic alloying element to ascertain whether the presence of a metallic solute would have a similar effect to oxygen in inducing grain refinement. A Cu-Ag alloy was chosen, since both metals had been shown to behave similarly on undercooling. The alloy Cu-20 wt pct Ag was selected since the eutectic constituent outlines the initial growth form of the primary copper, so that the as-frozen grain structure is not obscured if subsequent recrystallization occurs. This paper describes the results of undercooling experiments carried out with Cu-20 pct Ag samples undercooled to a maximum of 197°C and the effect of oxygen content on the grain structure of the undercooled samples. EXPERIMENTAL Melting was carried out in a small cylindrical resistance furnace using "fine" silver granulate and oxygen-free high conductivity copper. The procedure adopted was to melt the required quantity of silver in air in a clean vitreous silica crucible for approximately 15 min, freeze, and add granulated commercial soda-lime glass to form a complete surface slag cover, after which the sample was melted and frozen several times to reduce the oxygen content. The glass slag cover was approximately 3 in. thick. Pieces of copper (=50 g) were added to the crucible until the required quantity to make 350-g samples of alloy had been charged. Each piece was added quickly to the crucible which was held at a temperature slightly above the melting point of silver. The piece was quickly pushed beneath the glass to minimize oxidation and any oxide coating usually decomposed before the piece had settled down into the silver. After the full quantity of copper had been added, the melt was stirred with a silica rod to hasten homogenization and a Pt/Pt 13 pct Rh thermocouple enclosed in a vitreous silica sheath inserted for temperature measurement. Heating and cooling curves were recorded on a potentiometric chart recorder fitted with a zero suppression unit. The milli-voltage range of the recorder was adjusted so that temperatures could be read to 1°C. Heating and cooling curves were taken every hour until three consecutive readings gave the same solidus-liquidus range, consistent with the solidus-liquidus range for this alloy composition by reference to Hansen and Anderko.7 Metallographic examination of samples frozen at this stage, failed to show any variation in composition from bottom to top of the ingot. Consequently, it was considered that the melt was homogeneous at this stage and undercooling experiments were then car-

Jan 1, 1970

By G. R. Love, M. L. Picklesirner

THE use of color adds a new beauty, power, and versatility to metallography. This has been amply demonstrated in a number of public exhibits and on the walls of corporate, government, and university metallographic laboratories around the world. Indeed, in a number of cases, the color not only is of esthetic value but is scientifically essential to the competent examination of particular samples and systems. Optical strain analysis,' determination of true grain size in U-A1 alloys: phase identification in Nb-base alloys,3 analysis of transformation rates in heat-treated Nb-Zr alloys,4 measurement of solidification segregation in cast Nb-Zr alloys,' study of the intermediate phases in Nb/Sn diffusion couples,6 investigation of the oxidation kinetics at various surface orientations in single crystals,7'8 and a number of other research efforts9'10 would have been vastly more difficult without the extra clarity and distinction introduced by color—either the colors resulting from particular polarized light techniques or the interference colors obtained by oxidation. It is unfortunate that the beauty of colored micro-structures cannot be shared more widely, but the economics of publishing a technical magazine apparently prohibits color plates except under extraordinary circumstances. It is doubly unfortunate that no convenient, general method exists for communicating the information available in colored microstructures to the scientific community. True, in some instances the color changes observed are dramatic enough that direct black and white photography is adequate. In photoelastic stress analysis, several orders of interference fringes may be produced and it is frequently desirable to suppress the color effects (use monochromatic light) to simplify interpretation. On the other hand, relatively minute color variations may be readily detectable by eye, be of considerable assistance in interpreting a structure, and yet be enormously difficult to reduce to an unambiguous black and white print. In order to illustrate a possible compromise or solution to the above problems, we wish to show three black and white photographs of the same region of an Nb/Sn diffusion couple whose interpretation depends heavily upon use of color techniques. These micro-structures were obtained from a section through a rod consisting of commercial-purity niobium tubing filled with 99.995 pct Sn (melt-filled in 10"6 Torr vacuum). The composite rod, measuring 500 mm long, was bent into a hairpin shape and inserted into water-cooled copper grips in a vacuum high-temperature furnace. The rod was outgassed for 2 hr at a furnace temperature of 800°C and was further heated by self-resistance using a 2.2-kva ac power supply. The rod was raised to 1910°C over a 4+-hr period, held for 15 min, slow-cooled to 1500°C over a -hr period, and free-cooled to room temperature (all temperatures were measured with W5Re/W26Re thermocouple). Throughout the heat treatment the ambient pressure was kept below 8 x lo-' Torr. Metallographic preparation of the sample was completely conventional except that it was necessary to extend the grinding steps to remove damage in the brittle NbsSn phase introduced by sectioning. No satisfactory chemical or electrochemical etch has been found for these diffusion couples; of those which have been tried, any that attack the niobium-rich phases at all attack the tin-rich phases disastrously. However, the final polishing steps did produce a slight relief effect and this, coupled with anodizing, yielded excellent clarity in the microstructures for direct observation. The anodizing voltage was chosen to give maximum change in color per unit change in composition. In this case, indeed normally, this maximum sensitivity occurred in the blue-to-red color range. The pure niobium tube wall anodized to a relatively bright blue

Jan 1, 1967

By A. W. H. Morris, G. H. Laurie, J. N. Pratt

Using- galvanic cells of the form Sn(liq)/Sn" (LiCl-KC1-SnCl,)/Sn-Ag (alloy), measurements have been made of relative thermodynamic properties of the a, C, E, and liquid phases of the Ag-Sn alloy system. Partial heats of solution of the components in the liquid alloys lzave also been observed by direct cal-orimetric measurement in an isoperibol calorimeter. The thermodynanzic quantities are disczlssed in relation to structural and other properties and the existence of anomalous minor fluctuations in the partial heats and entropies of solution in liquid alloys is tentatively suggested. In the course of a recent electro motive-force study of the thermodynamic properties of Sn-Ag-Pd liquids,' some measurements were also performed on the Ag-Sn binary system. Most previous thermodynamic studies of this system have been concerned with the liquid state. Measurements confined to the limiting heat of solution of silver in liquid tin have been made by many calorimetric workers2 while high-temperature calorimetric measurements of the heats of formation of the full range of liquid alloys are reported in the early work of Kawakami~ (1323°K) and more recently by Wittig and Gehrin~(1248°K). Electromotive-force studies on liquid alloys have been made by Yanko, Drake, and Hovorka' (606" to 686°K; 86 to 99.4 at. pct Sn) and by Frantik and Mc Donald' (900°K; 30 to 90 at. pct Sn). The only known measurements on the solid state are of heats of formation of the a, £, and c phases; these values were obtained using tin-solution calorimetry, at 723"K, by Kleppa,~ whose investigation also yielded heats of formation of liquid alloys containing more than 64 at. pct Sn. The present experiments on the Ag-Sn alloys include a re-examination of the liquid phase and, because of the dearth of free-energy data for the solid state, attempted measurements on the a, c, and E phases. The principal new feature of electromotive-force results obtained for the liquid phase was an indication of anomalous fluctuations in the partial heats and entropies of solution of tin at certain compositions. However, since the values for these thermodynamic quantities were determined from the temperature coefficients of cell potentials, which are commonly subject to considerable error, confirmation by calorimetric measurements was considered desirable. A detailed investigation of the partial heats of solution of the components in the binary liquids was made using a liquid metal solution calorimeter. I) GALVANIC CELL STUDIES a) Experimental Details. Measurements were made, as a function of alloy composition and temperature, of the potentials of reversible galvanic cells of the form: ~n(liq)/~n++/~n-Ag (solid or liquid alloy) Details of the apparatus and experimental techniques have been given elsewhere.' so that a brief account will suffice here. Molten salt, 58 mole pct LiC1-42 mole pct KC1, containing small amounts (1 to 2 mole pct) of stannous chloride was used as the electrolyte. The salts were dehydrated by pre-electrolysis and vacuum -drying techniques. Cells were established under an argon atmosphere by immersing tin and alloy electrodes in the molten salt contained in a large silica tube, heated in a vertical resistance furnace. The tube was sealed at the top by a head plate provided with openings permitting the simultaneous insertion of six electrodes, a central thermocouple sheath, and connections to vacuum and argon lines. Temperatures were controlled to *0.2"C over prolonged periods, with maximum variation over the electrodes at any time of 0.l°C. Temperatures were measured with a standardized Pt/13 pct Rh-Pt couple. The electromotive force of this and the cell potentials were measured on a Cambridge Vernier potentiometer and short-period galvanometer. Alloys were prepared from Pass "S" tin (99.999 pct) and Johnson-Matthey high-purity silver (99.999 pct) by melting in evacuated silica capsules and quenching in cold water. For liquid phase experiments, pieces of the resulting alloys were remelted into prepared silica electrode units, while solid electrodes were prepared by remelting into 3-mm bore tubing, inserting a cleaned molybdenum lead wire, and quenching to produce uniform rods about 3 cm in length, with soundly attached leads. In all cases remelting was done under an argon atmosphere. The solid electrodes were subsequently annealed in evacu ated silica tubes for 14 days at about 20°C below the solidus and quenched. Analyses showed that these techniques produced uniform electrodes with no significant change from weighed out compositions. b) Results and Discussion. Measurements were made on about forty alloys in the solid and liquid states, over varying ranges of temperature between 550" and 1050°K. Stable, mutually consistent, and reproducible electromotive-force data were obtained with most liquid alloys and these are shown in Fig. 1. Investigations were extended below the liquidus tem-

Jan 1, 1967

By Harry L. Brown, James M. Dickinson, Philip E. Armstrong

The elastic stiffness coefficients of niobium have been rneaslrred from -150°C to over 650 C and ave in accord with the anomalous temperature dependence previously observed for Young's modulus. This unusual behavior has been shown to be the result of changes in the relative values of the shear elastic coefficients. The bulk modulus remained essentially constant over the temperature range studied, while changes in the shear coefficients produced a more nearly isotropic behavior as the temperature increased. A number of investigators have reported data on the temperature dependence of Young's modulus of polycrystalline niobium. The values cited for the room-temperature modulus ranged from 0.71 to 1.17 x 10" dynes-cm-' (10.3 to 17.0 x lo6 psi) and in each case the modulus was observed to decrease with increasing temperature. In our recent study,8 both polycrystalline and single-crystal samples of essentially strain-free high-purity niobium showed a Young's modulus whose temperature dependence as well as magnitude was strongly dependent on specimen orientation. This was shown to be the result of changes in the relative magnitude of the two shear moduli, C and C'. Young's modulus of polycrystalline samples with strong [110] textures and Young's modulus of single crystals whose orientations were away from the [loo] actually increased as much as 12 pct between room temperature and 1000°C. While studies on the residual effects of previous fabrication, thermal treatments, and impurity contents will probably resolve the problem of the widely differing Young's modulus values observed in polycrystalline niobium, a more complete description of the elastic properties requires an evaluation of the single-crystal elastic constants. A knowledge of these constants can provide a clearer picture of the nature of the temperature dependence, as well as establish a basis for predicting the modulus behavior of polycrystalline niobium, both in the randomly oriented condition and with various fabrication textures. The effects of residual strains on the elastic behavior then can be separated from orientation effects. EXPERIMENTAL METHOD Thin rod-resonance techniques were used to obtain both the Young's modulus and shear modulus of single-crystal specimens. The determination of these moduli as functions of the crystallographic orientation of the specimen axis provided sufficient information to permit calculation of the single-crystal elastic constants. The necessity for torsional as well as longitudinal measurements led to the design and assembly of a vertical-tube vacuum furnace incorporating probes for the simultaneous determination of both longitudinal

Jan 1, 1967

By Noel Jarrett, Allen S. Russell, Bernard M. Starner, Stanley C. Jacobs

A process has been developed to refine high-grade commercial aluminum to 99.99 pct purity. This enzploys precipitating titanium, vanadium, and zirconiu~ as borides. The upgraded liquid is partially crystallized under conditions that prevent massive freezing. The resulting crystals are compacted and washed free of adhernig impure liquid aluminum by melting the top crystals while draining off the liquid. Finally, the bottorn portion of the crystal bed is melted and removed as the pure product. Leakage from the crystallizat~on vessel is prevented by a layer of alumina granules that are not wetted by the aluminum. ALUMINUM of 99.99 pct purity is desired in a variety of applications. These include electrolytic capacitors, catalyst carriers, decorative trim for automobiles, aluminum powder, and high-strength alloys. This market was for many years filled by refining commercial-purity aluminum in a three-liquid-layer electrolytic cell of the type originally proposed by Hoopes. To meet the expanding need for lower-cost high-purity aluminum, a refining process based on chemical precipitation and fractional crystallization was developed. The availability of this super-purity metal was announced in November 1961, and three patents1-3 relating to the process were issued in 1965. An alternate process, comprising continuous fractional crystallization of a portion of a liquid-aluminum feed stream under conditions of strong agitation at the surface of the growing crystals, was published by Dewey4 in 1965. Earlier proposals by Larsen 5 and by Regner6 also emphasized the advantage of rapid agitation of the melt in the region of the crystallization zone. johnson 7 more recently proposed extracting purified aluminum from an alloy by mixing crystals of pure aluminum into the molten alloy, allowing them to grow, and removing the pure aluminum as by casting. Aamot's 8 recent patent describes a freeze refining process that involves controlled partial freezing against cooled moving retort walls. The Alcoa process to produce high-purity aluminum is schematically described in Fig. 1. Relatively pure aluminum is treated with boron in a furnace (a) to precipitate high melting point impurities, the supernatant liquid is transferred to a crystallization vessel (b), the liquid is partially crystallized with mechanized "tamping" (c), the remaining impure liquid is tapped out and the crystal bed partially and directionally re-melted to wash the underlying crystals (d), and the balance of the crystals is melted and withdrawn as high-purity aluminum (e). PRINCIPLES OF FRACTIONAL CRYSTALLIZATION Commonly encountered impurity elements form two types of systems with aluminum. Titanium, vanadium, and some other transition elements in periodic groups IV, V, and VI are the high melting point impurities. Initial solidification of aluminum containing small percentages of these elements will produce solid markedly enriched in them at temperatures slightly above the freezing point of high-purity aluminum. Most other elements, such as silicon and iron, produce eutectic type reactions with aluminum. In this case the first solid formed on initial solidification at temperatures slightly below the freezing point for high-purity aluminum is markedly depleted while the remaining liquid is enriched in these elements. The equilibrium change in composition during solidification of aluminum contaminated with silicon, iron, or titanium is illustrated in Fig. 2. Here, for clarity, the temperature scale is expanded beyond the

Jan 1, 1968

By E. A. Gulbransen, A. Pebler

Pressure-composition isotherms have been determinedfor the systems ZrMo2 -H2 between 0" and 900°C at hydrogen pressures between 10-4 and 760 Torr. Tkese studies plus X-ray diffraction analyses of selected compositions show that these intermetallic compounds form a complete series of hydrogen solid solutions within this temperature and pressure range. The partial molar enthalpies and entropies of solution of hydrogen in the three intermetallic compounds were calculated. The solubility of hydrogen in these Laves phases can be related to their free electron concentration. ThE reaction of hydrogen with inter metallic compounds of zirconium with Laves phase-type structure has both theoretical and practical interest. The in-termetallic compounds have structures different from any of the pure metals. In addition the electron to atom ratio of the intermetallics varies appreciably. In the nuclear materials field, these materials are of interest since intermetallic compounds may exist as precipitates in commercial zirconium alloys with transition metals. From studies on the hydrogen reaction, one can establish the stability of the intermetallic compounds relative to the hydrides of the component metals. The intermetallic compounds may also be used as getters for hydrogen and as materials for hydrogen storage. This paper will present the results of equilibrium and structural investigations of the ZrCr2-H2, the ZrV2-H2, and the ZrMo2-H2 systems for the temperature range of 0" to 900°C and the pressure range of Torr to 1 atm. A complete thermochemical analysis will be made of the measurements. Some of the results have been referred to in a survey paper' covering the types of hydrogen behavior observed in zirconium alloys. LITERATURE ZrCr2 is the only intermetallic phase in the Zr-Cr system.' It exists in two allotropic modifications. Between 900" and 1000°C the hexagonal MgZa (C14) structure transforms into the cubic MgCu2 (C15) structure. Beck4 has previously shown that the intermetallic ZrCrz takes up hydrogen into solid solution up to a composition H/ZrCr2 = 1.14 at room temperature and 1 atm hydrogen. A ZrCr2 hydride was stated to have a fcc structure. The binary Zr-V system has been investigated by williams5 although Rostoker and Yamamoto 6 had given earlier a partial phase diagram. ZrV2 was the only intermediate phase found in the system. wallbaum7 reported earlier that the compound ZrV2 hp the MgZm (CJ4) type structure with a = 5.288A and c = 8.664A. This was not comfirmed by Elliot8 nor by Matthias, Compton, and corenzwit9 who reported that ZrV2 has a cubic MgCua LC15) type structure. The latter authors gave a = 7.429A. Also, several authors do not agree on the temperature of the peri-tectic horizontal. It appears that ZrV2 may exist in several modifications. A high capacity for occluding hydrogen was reported by Beck4 who found H/ZrV2 = 1.38 at room temperature and 1 atm of hydrogen. ZrMoz is the only intermetallic phase in the Zr-Mo system.10 ZrMoa is formed by a peritectic reaction of the melt with a molybdenum-rich solid solution. ZrMo2 has the C15-type structure with a = 7.59A. MATERIALS High-purity chromium, vanadium, and molybdenum were used together with grade 1 crystal bar zirconium to prepare the alloys. Table I lists the chemical analysis and the sources for the metals. Ten-gram alloy compacts were levitated and melted in an inert argon

Jan 1, 1968

By M. C. Flemings, F. R. Mollard

Two-phase Pb-Sn alloys, ranging in compositiotz from 12 to 26 at. pct Pb, were unidirectionally solidified in a convection-fvee system, with thermal gradients in the liquid of up to 480°C per cm. Plane front solidification Loas achieved in all alloys at sufficiently steep gradient and slow growth rate. Conditions causing breakdown of the interface to a cellular or dendritic structure were roughly as predicted by a sinlple constitutional supercooling criterion. Structures of alloys of near-eutectic composition, solidified with plane front, exhibited a lamellar structure; the alloys furthest from the eutectic exhibited a rod-like morphology. Assumptions used in a previous paper for calculations of solute redistribution in two-phase plane front growth are shown to be reasonable for Pb-Sn alloys. Results of calculations are in good agveement with experiment. These show that variations in growth rate during solidification of the composite produce marked changes in average composition of the solid. IN a previous paper, Part I,' conditions necessary for plane front solidification of two-phase alloys were discussed and calculations given on solute redistribution in this type of solidification. This paper, Part 11, describes experiments on Pb-Sn alloys, designed to check the assumptions of the analyses presented, and to demonstrate their applicability to a given alloy system (Pb-Sn). APPARATUS AND PROCEDURE General. A vertical crystal growing unit was built for the study as described below, and "normal solidification" employed throughout (i.e., the ingot was completely melted except for a portion near the bottom end and then solidification initiated). The furnace was designed to permit achieving very low interface velocities, and steep gradients, in essential absence of convection; it is sketched in Figs. 1 and 2. Simple heat flow considerations indicated choice of a small-diameter specimen (ingot) to achieve high thermal gradients; ingot size chosen was 0.3 cm in diam. The ingots were held inside silica tubes 0.06 cm wall thickness. The requirement for an essentially convection-free system was met by use of the vertical furnace with the heat source over the heat sink, as well as by use of the small diameter ingots. In addition, by selection of tin-rich Pb-Sn alloys, the diffusion boundary layer (lead-rich) was denser than the bulk liquid and would not cause convection. The only apparent driving force for convection in this system was the small transverse temperature gradient resulting from heat flow into the samples from the furnace. Such convection must have been small, however, since 1) no temperature fluctuations in the melt were noted in the recordings, which would have indicated turbulent convection, and 2) experimental results obtained and described below could not have been achieved in presence of significant convection. Ingot Preparation. Pb-Sn alloys ranging from 12 to 26 at. pct Pb were prepared from high-purity lead and tin (99.9999 pct pure metal). Each alloy was prepared by first melting the pure metals in the proper ratio in air, in a Pyrex container. The melt was then stirred thoroughly with a graphite rod and cast into a preheated graphite mold. Ingots were then prepared from these castings for use in the actual experiments; each ingot consisted of a rod of alloy, 15 cm long in a 25-cm-long transparent fused-silica tube, 0.3 cm ID and 0.06 cm wall thickness. One chromel-alumel thermocouple, made with 0.004-in.-diam wire and protected by a two-hole alumina tube, was inserted into each ingot. The 1.5 cm of the thermocouple nearest the bead were protruding from the alumina tube and were coated with a very thin layer of boron nitride paste which provided electrical insulation while allowing good heat conduction. The electrically welded junction size was about that of each wire and the cross section of the two wires represented about 0.3 pct of the sample cross section. Including the coating, the thermocouple cross section still represented less than 0.5 pct of the ingot cross

Jan 1, 1968

By R. Gibala

Internal-friction measurements on hydrogen-charged iron over the temperature range 4° to 300°K are reported. Two relaxation peaks, the hydrogen Snoek peak at 48 °K and the hydrogen cold-work peak in the range 145" to 2200K, are observed at 80 kc per sec. The effects of hydrogen-charging, aging, and deformation on the internal-friction results are determined and are used to show evidence of partitioning of hydrogen between solid solution and dislocations. The hydrogen-dislocation binding energy is estimated from the behaviors of the relaxation tiwzes and relaxation strengths of the Snoek peak and the cold-work peak. Its value is approxi,mately 6400 cal per mole. MANY investigators have demonstrated the usefulness of internal-friction measurements in determining the manner in which interstitial impurities interact with dislocations in bcc metals. Part of this success has resulted from the existence in bcc interstitial alloys of two anelastic relaxation peaks—the Snoek peak' and the cold-work or Koster peak2 (designated hereafter by SP and CWP, respectively). The SP has its origin in the redistribution of interstitials in solid solution among sites made energetically nonequivalent by an applied stress. The rate at which the redistribution occurs, measured by the relaxation time for the process, is related in a simple way to the diffusion rate of the interstitial in the free solution. The relaxation strength is proportional to the concentration of interstitial in solid solution. The CWP appears to have its origin in the relaxation of an interstitial-dislocation complex for which the relaxation time at a given temperature is much longer than that for the SP. The relaxation strength depends strongly on the concentration of interstitial and the amount of deformation at relatively low concentrations and deformations.2-8 From the dependence of the relaxation strength on interstitial concentration, the SP and CWP may serve as measures of the concentrations of interstitials in free solution and bound at dislocations, respectively. Thus their behaviors have been used to investigate interstitial-dislocation interactions in Fe-C, Fe-N, and several refractory metal systems.'-' However, the Fe-H system has not been studied extensively,? even though disputes in recent years over the mechanism of hydrogen embrittlement in steels involve argument over the extent to which lattice hydrogen and hydrogen-dislocation interactions play an important role.?-'' Previous work on Fe-H indicates the internal-friction behavior to be qualitatively consistent with that for other systems. Apparently, both a SP and a CWP are observed in the temperature ranges 30" to 50°K and 100°to 140°K, respectively (at 1 cps). However, there is neither sufficient nor accurate enough information to identify convincingly these peaks with their supposed counterparts in other systems. Nor have there been any attempts to make quantitative use of these peaks to study hydrogen diffusion, aging, or interaction processes. In this paper, an attempt is made to obtain more complete information on internal friction of hydrogen in iron than presently available. Internal-friction measurements in the range 4° to 300°K on hydrogen-charged iron show that the two peaks discussed above are in fact the hydrogen SP and CWP. Several effects of hydrogen charging, aging, and deformation on these peaks are observed and are used to determine the partitioning of hydrogen between solid solution and dislocation sites in iron. EXPERIMENTAL Materials. The starting material for this investigation was a high-purity iron supplied by National Research Corp. as a 1-in.-diam rod. The following analysis was given (inwt pct): C, 0.004; N, 0.0003; 0, 0.007; P, 0.002; S, 0.005; Si, 0.006; Mn, 0.001; Cu, 0.002; Ni, 0.005; Cr, 0.001; Mo, 0.001. After an initial vacuum anneal (-5 XTorr) at 700°C, the rod was cold-

Jan 1, 1968

By P. R. Sahm, T. V. Pruss

Mixtures of fine GaSb and Gds as well as preal-loyed GaSbl,As, powders were hot-pressed at 690°C and 25,000 psi. Dense alloys with compositional gradients of less than 5 pct were obtained from mixtures containing about 20 mol pct GaAs. For x < 0.2, there were increasing compositional gradients, and for x > 0.2 a GaAs-rich second phase appeared in the microstructure. Densification as well as alloying wlechanisms were enhanced by dissociation and, possibly, oxidation reactions of the powders. Densification of coarse prealloyed material, however, primarily depended on plastic flow phenomena and required temperatures just below solidus and pressures of 50,000 psi Thermal and electrical properties were measured. Although in no case was the figure of merit of melt growth materials approached closer than to within 25 pct, the better overall thermoelectric properties were found in coarse-grained, prealloyed materzals which had been compacted to near theoretical density and where grain regrowth had been induced. Similar results are believed to hold in other binary 111-V compound systems if processed under similar conditions. The densification of unalloyed GaSb powders during pressure sintering (= hot pressing) was shown to depend strongly on powder particle size.&apos; Fine powders displayed a "liquid skin effect" that enhanced compaction through the presence of liquid gallium, whereas coarse powders compacted predominantly by plastic flow. The liquid skin effect, in particular, appeared attractive for alloying GaSb with other, higher-melting, III-V compounds, and to densify these in a one-step operation. This is of special interest in the case of III-V compound alloys as the conventional techniques of melt growth or long-time annealing of powders2 is very time-consuming, especially in cases where a large separation of liquidus and solidus can be expected, such as in GaSbl alloys.2 It was felt that experimentation with this system, a particularly unfavorable example, would allow us to extrapolate to several other, more favorable, cases. Uses of Ill-V alloys are most evident in thermoelectric energy conversion devices.&apos; For this reason certain thermal and electrical properties were measured and compared to those of melt-grown material and to hot-pressed prealloyed powders. EXPERIMENTAL PROCEDURE The pressure-sintering apparatus has been previously described.&apos; Using this equipment, both powder mixtures of GaSb with GaAs and prealloyed GaSbl-,As, powders were hot-pressed. The mixtures were pre- pared from cast GaSb and melt-grown GaAs. The prealloyed material was obtained from stoichiometric melts, initially heated to 1200°C in an evacuated quartz ampoule, and then annealed in the solidus-liquidus interval. A typical annealing cycle consisted of a heating to 850°C (1 hr), extended successive annealing at 750°C (65 hr), and slow stepwise cooling (25 hr) to below solidus. The GaSb, GaAs, and GaSbl-,As, materials were ground in a vibrational mill to particle sizes below 500 . A jet mill reduced these further where necessary. Fine powders were analyzed by Coulter counting for their size distribution. Average sizes by volume were calculated from the data. Mixing of the powders, where necessary, was carried out through rapid vibrational motion. The hot-pressing operation consisted of a degassing period of 15 hr, in most cases at 690°C, followed by several hours of compression, normally 25,000 psi at 690°C for powder mixtures and 50,000 psi at 710°C for prealloyed powders. The chemical compositions were confirmed by X-ray fluorescence analysis. To estimate the degree of solid solution achieved, lattice parameters were determined and interpreted according to Vegard&apos;s rule. In addition, optical microscopy helped to correlate the relative amounts of the phases present as well as the degree of porosity to the measured density. To reveal grain boundaries, polished surfaces were etched4 with H 2 O:H 2 O:HCl = 2:l:l. In several cases microscopic concentration gradients were monitored by electron-probe analysis-. RESULTS AND DISCUSSION Alloying and Densification of GaSn-GaAs Powder Mixtures. After preliminary experiments showed that no appreciable alloying took place in mixtures of coarse GaSb (22.5 p) and fine GaAs (2.5 p) powders, hot pressing of mixtures was confined to fine powders only (3.1 and 2.4 p, respectively). Alloying and densification apparently occurred simultaneously with a grain regrowth mechanism which depended on the presence of a liquid phase.&apos; The liquid phase was provided by the dissociation of GasbS particles into liquid gallium and antimony above 555°C. This not only enhanced grain regrowth and densification, as in unalloyed GaSb,&apos; but took on additional importance here for the alloying process. Alloying was speeded up by the resulting liquid-solid interaction as compared to the very slow solid-state diffusion process normally expected at these temperatures.&apos; The results obtained with a series of powder mixtures have been compiled in Table I. It is seen that for mixtures with less than 20 mol pct GaAs, x < 0.2, the compositional ranges increased and for x > 0.2 a GaAs-rich phase and free antimony appeared in the microstructure in sizable amounts.

Jan 1, 1968