Search Documents

By R. T. DeHoff, F. N. Rhines

The problem of determining the number of particles of a phase distributed randomly in unit volume c an opaque matrix from measurements made on random plane sections is closely investigated. A formalized derivation for the general case is presented. Applications of this formal result are made to specific types of aggregates dispersed in a matrix; circles, particles with flat circular ends, cylinders, and constant shape aggregates of ellipsoids of revolution. Quantitative measurements of the average geometric properties of these aggregates are developed. In 1953 R L. ullman' developed a technique for determining, from measurements made on random plane sections, the number per unit volume, Nv, of particles imbedded in an opaque matrix. Using meas urements previously developed for volume fraction, vV,' and surface area per unit volume, .SV,= he was able to determine average volume, average surface area, and average dimensions of particles, independent of size distribution, for spheres and circular disks,' as well as for uniformly sized cylinders of any axial ratio.4 The technique is mathematically rigorous for the cases studied. Recent studies of this problem have yielded rigorous solutions for the general case of cylinders, permitting evaluation of NV independent of size distribution and with all axial ratios intermixed. Generalization to shapes having flat, circular ends has been deduced. A solution has also been obtained for the ellipsoid of revolution, independent of size distribution, but requiring a constant axial ratio in the generating ellipses. In the following derivations, as in Fullman's work, it is assumed that the problem is purely a combination of geometry and probability theory,that is,eithe the particles to be measured are imbedded randomly in the matrix, or a sufficient number of random plane sections are taken to make the sample representative. A relationship exists between the number of par- ticles of a given size and shape that may be situated in unit volume (Ny) and the number of intersections a random plane can be expected to make with these particles (Na). It is the purpose of the following section to derive this relationship. The probability of Intersecting Convex Bodies. Consider a convex, closed surface, that is, one for which no two surface normals point in the same direction, situated at some arbitrary position and orientation in a cube of material that is one unit long at each edge. It is not necessary to assume that this surface is symmetrical. Let planes be constructed in the cube parallel to the top face and at random distances from it. Those planes which lie between the two planes which are just tangent to the top and bottom of the surface will intersect it. The probability that a plane will intersect the body may be defined as the limit of the fraction of planes that lie between the two tangent planes as the total number of planes constructed becomes infinite. This fraction, in the limit, is equal to the distance between the two tangent planes, DV, divided by the total length over which planes are constructed, which has been taken as unity. If the body is now rotated to a new orientation, applying the same argument, the probability of intersection is again numerically equal to the distance between tangent planes. Let DV ($, 0) be defined as the distance between tangent planes of a convex body as a function of orientation, Fig. 1. The probability, Pr, of intersecting the body with a randomly oriented and randomly positioned plane, is then equal to the average value of DV (0, $) over all possible orientations. It may be easily shown that for a system of spherical coordinates the probability of finding an orientation between Q and Q + de, $ and $ + d$ is (1/4n) sin $ dB d$, so that This relationship may be applied to each body in an aggregate of convex bodies. In particular, if the bodies in an aggregate are classified according to size and shape, and the ithsuch class is considered, Eq. 111 may be applied to this entire class, the probability of intersecting any body in this class being

Jan 1, 1962

By H. Buttner, F. H. Udin, J. Wulff

GRAIN boundaries, the interface between adjacent crystals differing only in respective orientation, have been the object of much experimental and moderate theoretical attention for many years. The earlier experimental work was primarily concerned with the effect that grain boundaries have on the physical behavior of metals, while the basic nature of the boundaries themselves drew only secondary attention. A complete turnabout has been witnessed in more recent years, however, as researchers are now undertaking investigations bearing as directly as possible on the character and physical properties of the grain boundary. Quite comprehensive reviews of the past literature on the subject of grain boundaries have been reported on numerous occasions by various authors, and together these references represent a thorough coverage.'-? It becomes evident in these reviews and other papers"" that the properties of grain boundaries can be explained by regarding a grain boundary as a crystalline region of transition between two unmatched grains. Studies show that grain boundaries have a definite interfacial energy. By measuring the geometry about a point where three interfaces meet in equilibrium, the energy of any two boundaries can be measured with respect to the third. If the absolute interfacial energy of one boundary is known, the absolute energy of the other two boundaries can be calculated. This apparently holds true for any system of phases and interfaces.""25 Most studies have been carried out on solid-solid or liquid-solid systems, or systems of both. However, if about the point of intersection one phase is vapor and the other two are disoriented grains, then two interfaces are surfaces, and one is a grain boundary. Investigations leading to relative grain boundary energy for this particular system have been made by Fullman,23 Greenough and King," and Bailey and Watkins." Herring" treats the system mathematically and arrives at an equation, which appears as Eq. 1, in a form consistent with the reference angles indicated in Fig. 1, ( d?1 ?n = x sina + cosa ( ------- 2 ?2 Sinß da + C0Sß (d?2/d As the system of interfaces comes to equilibrium, a groove of angle appears at the intersection of the grain boundary with the surface, an effect first observed by Rosenhain and Humfrey." Eq. 1 is essentially a summation of the vertical components in Fig. 1 set equal to zero, where ?1 and ?2 are surface tension vectors lying along the side of the groove at point 0 in the groove root. The differential vectors represent the forces tending to rotate the surface into a position that will expose a crystal cleavage surface of lowest energy. Chalmers, King, and Shuttleworth- simplify Eq. 1 by assuming the differential vectors equal to zero, ?1 = ?2= ?Surface and arrive at the following equation: ?B = 2?s cos [2] Greenough and King24 use Eq. 2 for determining the grain boundary energy of silver using prepared bi-crystals of known orientation. Bailey and Watkins25 have done the same for copper, as did Chalmers and others on lead. These authors report the grain boundary values as ratios relative to the surface energy. In the above work, the orientation of grains forming the grain boundary in all cases was controlled so

Jan 1, 1954

By G. W. Neighbor, M. S. Abrahams

The actice slip plane in InSb is found to be of the {111} type by using the method of two-tmce anulysis. Measurements of the rotation of the tensile axis with increasing plastic shear strain indicate that the tnacroscopic slip direction is a <110>. Recrystal-lization is first observed at a sheav strain of about 70 pct for a crystal deformed at a temperatzcre of 473° C and at a shear strain rate of 2.90 times 10- %ecc-I; the dislocation density corresponding to this value of the strain is 1 x 108 em-&apos;. At a shear strain of about 170 pct, the recrystallization process is complete. This appears to be the fimt time re-crystallization has heen observed in a semiconductor. V HE operation of the {Hi} glide plane in Ge was first noted by Gallagher.&apos; This finding was confirmed by Treuting who also extended the result to Si. Using X-ray techniques, Treuting3 demonstrated that the macroscopic slip direction in Ge was of the <110> type. The choice of the (111) <110> slip system in Ge is exected since in the diamond-type structure the {Ill7 plane is the plane of highest density and the <110> direction is the direction of closest packing.4 There is little reason to expect that the glide elements of InSb are different from those of Ge, since the zinc-blende structure characteristic of InSb is quite similar to the structure of Ge. With regard to InSb, Allen5 has indicated that the glide plane is the {111} type. There appears to be no quantitative determination of the glide direction, however. The present work was undertaken to establish both the glide plane and glide direction in InSb and also to see if recrystallization occurs at large shear strains. EXPERIMENTAL PROCEDURE Mechanical Deformation. The single crystals of InSb used in this investigation were deformed in uni-axial tension. The tensile axis of the specimens was [145] and the periphery was bounded by (111) and (321) planes as shown in Fig. 1. The metallographi-cally polished specimens were deformed in a dynamic atmosphere of He, using an Instron machine. The tensile tests were performed at temperatures in the range of 300°C to the melting point (525OC) and at shear strain-rates varying between 1.94 X 104 sec-&apos; to 1.94 x 103 sec-&apos;. Additional details concerning the preparation and deformation of the specimens have been presented previously.&apos; Analysis of the Glide Elements. The traces of the active slip plane in both (111) and (321) planes were analyzed according to the usual method of two-trace analysis." The macroscopic slip direction was determined by measuring the rotation of the tensile axis as a function of plastic strain,8 These measurements were made from Laue back-reflection X-ray photographs. The X-ray beam was perpendicular to the (111) plane. A given crystal was used for only one measurement of the rotation of the tensile axis after it had been strained to a predetermined value. Therefore, care was taken at all times to insure that the X-ray beam was parallel to the [Ill ] direction of all the samples. RESULTS AND DISCUSSION Glide Plane. After 3.4 pct plastic shear-strain, at a temperature of 369°C and a shear strain-rate of 1.94 x 104 sec-&apos;, the slip traces on the (111) and (3.21) planes appear as in Figs. 2(a) and 2(b), respectively. The results of an analysis of these glide markings demonstrate that the slip plane is (111). The operation of this plane is expected on the basis of the orientation of the specimen, since glide on this plane and in the [I011 direction will result in

Jan 1, 1962

By Barbara A. Thompson

A previously reported high-resolution method for the location of boron-rich areas in metallurgical and biological specimens was been adapted for general use on a routine basis. The rnetlzod utilizes neutron activation and autoradiograpizy. Alpha-particles emitted by boron nuclei upon neutron capture are recorded on a photographic emulsion. The resulting a-particle tracks show the location of boron-rich areas. Experimental techniques, interferences, and limitations of the method are discussed in detail. The method is most useful where there is marked segregation of boron. In this type of sample, the segregation can be observed when the nominal boron concentration is as low as 0.0006 pct. THE positive identification and location of boron-rich areas in metals is frequently of great interest in metallurgical work. Unequivocal identification is often difficult to make by conventional metallo-graphic methods. Recently, a method has been described which accomplishes this objective by neutron activation and autoradiography.l-3 The method can be described briefly as follows. Upon neutron capture, a -particles are emitted by boron nuclei according to the following reaction: ,Blo + n - ,a4 + 3Li7 + 2.4 mev The energy is dissipated as kinetic energy of the products. By irradiating a boron-containing sample in contact with a photographic emulsion and subsequently developing this emulsion, a-particle tracks are obtained whose location corresponds to the location of boron-rich areas in the sample. Two factors combine to make the reaction extremely specific for boron. The first is the unusually high (755 barns) cross section of boron for thermal neutron capture. The second is the higher neutron energy required to produce (n, a ) reactions in essentially all other nuclei except lithium. These two factors make the method specific for boron by six to seven orders of magnitude when a predominantly thermal neutron source such as the Brookhaven reactor is used. The reported limit of detection of this method is of the order of 0.01 pct B., The present work was originally undertaken to determine whether this limit could be lowered by use of a thinner emulsion. However, initial experiments showed that in order to use the method at all, it was necessary to reestablish the optimum experimental conditions in terms of the available irradiation facilities. It is the purpose of this paper to describe these experimental conditions in detail, to discuss the factors influencing sensitivity, and to evaluate several techniques for increasing sensitivity. EXPERIMENTAL A) Preliminary Experiments—The first measure-ments were made using samples of crystal oriented silicon steel containing various concentrations of boron. In the later experiments, samples of various high-temperature alloys such as M-252, hcoloy 901, Nichrome V, and so forth, were used. Faraggi, et al.,2 reported that the lower limit of sensitivity in this type of sample was about 0.01 pct B using nuclear emulsions of 50- u thickness. but that it should be possible to extend this limit by the use of thinner emulsions. Accordingly, we first used Kodak Auto-radiographic Stripping Film (Permeable Base) which has an emulsion thickness of only 5 µ. This was mounted on the metallographic specimens according to the technique described by Boyd.4 The emulsion remained in contact with the metal surface throughout exposure and development. Since the emulsion is transparent after development, the autoradiograph and metal surface can be viewed simultaneously and any correlation between film blackening and structure of the metal can be made directly with no problems of realignment. Because the silicon steel is readily attacked by moisture alone, it was necessary to apply a protective coating to the metal surfaces before mounting the emulsion. The coating was made extremely thin in order to absorb as few a-particles as possible. Boyd4 and Gomberg5 have discussed various plastics used for this purpose; however, none was sufficiently impermeable to prevent chemical attack of the steel during the developing process. This attack resulted in the production of gross chemical artifacts in the emulsion. It was, therefore, necessary to use the method of Wolfsberg and John6 as follows. A very thin (approximately 1 µ) coating of Plexiglas II was applied by dipping the sample in a 2 pct solution of Plexiglas II in dichloroethylene. Then, because the emulsion will not adhere to Plexiglas 11, a thin coating of Parlodion was applied in a similar manner using 2 pct Parlodion in iso-amyl acetate. No protective coating was necessary with the high-tem-

Jan 1, 1961

By H. F. Webster, C. G. Dunn

Rolling md annealing procedures are described for developing the (110) preferred orientation in tantalum for use as a thermionic-emission materinl where electrodes of a uniform high work function are needed. The development of a (110) [001] annealing texture is also reported. A preferential growth process due to low (110) gas-metal surface energy is considered in explanation of the observed growth of (110) grains. ThE growth of (110)-oriented grains in thin strips of rolled zone-refined tantalum, previously discussed briefly,&apos; is reported together with more recent information on the formation of a (110)[001.] annealing texture, i.e., one where the [001] direction of each grain lies close to the rolling direction. The development of tantalum sheet with a (110) preferred orientation (position of [001 ] directions unspecified) is of importance as a thermionic-emission material in thermionic-energy converters. Since therm ionic-emission properties of cesium-coated crystals vary over a wide range depending on (hkl) a2-7 a single (hkl) preferred orientation is more useful than either several preferred orientations or none. Furthermore, the (110) preferred orientation yields the highest thermionic-emission density for a cesium coverage of less than a monolayer.5-7 A consideration of the stability of the grain structure needed at high operating temperatures suggests that the preferred orientation itself should be produced by annealing processes. Information on annealing textures in tantalum, however, is rather limited. Pugh and Hibbard 8 obtained (111)[112] and (lll)[112] components in a 1400°C recrystallization texture in 95 pct cold-rolled foil; these components were enhanced on annealing at 2500°C. Muller9 rolled 99.97 Ta to 90 pct reduction in thickness in the final stage and noted only a slight sharpening of the rolling texture, with some intensification of the (111)[112] component upon annealing at 1200°C. The experimental procedure followed in the present work stemmed from the idea of using possible differences in (hkl) surface energies to provide a small driving force for growth of grains having the lowest (hkl) surface energy.l0* This would mean using high-purity tantalum processed into thin sheet form and annealed at high temperatures either in a noncontaminating atmosphere or in one producing additional sample purification. EXPERIMENTAL Zone-Refined Tantalum. National Research Corp. tantalum in the form of 1/8-in.-diam rod and of 99.98 purity (53 ppm 0, 15 ppm N, and 8 ppm C) was zone-melted four times. The traversing speed was 4 mm per min. Such treatment increased the resistance ratio, R293k/R20K , where R293K is the resistance of the specimen at 20°C and R20K the re-resistance at liquid-hydrogen temperature, from a value of 25 for the initial rod to 175 for the zone-refined rod (the increase in the resistance ratio provides a measure of purification11,12). Part of a zone-refined rod was rolled unidirec-tionally to a strip 0.003 in. thick and 0.28 in. wide. Samples were given anneals in the range 1300" to 2100°C using ac conduction heating in an oil-free vacuum system. Recrystallization to small grains appeared to be complete at 1300°C. Large grains were formed at higher temperatures, but these were not entirely strain-free, some grains showing asterism in Laue X-ray patterns and some displaying a range in orientation. Part of the difficulty was caused by thermal expansion of the tantalum strip as it was held between fixed supports. However, in a different arrangement with the tantalum freely suspended in a tantalum can, high-frequency induction heating in vacuo at about l0-6 Torr produced

Jan 1, 1964

By Matti J. Saarivirta

Two new precipitation hardening copper-base alloys, Cu-1.5pct Ti-2.5 pct Sn and Cu-1.5 pct Ti-2.5 pct Sn-0.4 pct Cr were developed. High strength and medium conductivity are obtained by solution annealing and subsequent heat treatment. Combination of solution annealing, cold working, and heat treatment increases the strength further. Good elevated temperature strength and ductility makes these alloys superior to copper-beryllium-cobalt alloys. A considerable amount of research work is being done throughout the world to develop new and improved copper-base alloys. In many cases, the object is to develop alloys having a combination of high electrical and thermal conductivities, and high strength. The problem is complicated by the fact that any alloying element added to increase strength, invariably decreases conductivity. Among the few commercial precipitation hardening type copper-base alloys, the Cu-Be-Co alloys are best known. These alloys are used in many applications that require good spring characteristics cornbined with high strength and moderately high electrical conductivity. They can be formed readily after solution annealing, then heat treated to obtain high strength and conductivity. Because of the relatively high cost of the Cu-Be-Co alloys, considerable research has been aimed at developing new lower cost alloys having similar properties. As far as is known today, the binary Cu-Ti alloy is the only one whose properties are

Jan 1, 1962

By R. W. Fountain, J. F. Libsch

ONSIDERABLE time and effort have been ex- pended recently in research designed to provide a better understanding of the solid state transformations leading to the permanent magnet qualities of many commercial alloys. The knowledge derived from these investigations has resulted in the development of new magnetic materials as well as improvements in some of the older materials. The permanent magnet qualities of the newer alloys having high residual inductions and high coercive force values, have been attributed to the precipitation of a second phase or an order-disorder transformation in a solid solution. The ternary Co-Fe-V alloys provide an interesting group of magnetic materials having a wide range of magnetic properties. The low vanadium alloys&apos; possess soft magnetic properties similar to the binary Co-Fe alloys, while at higher vanadium contents (about 10 pct) interesting permanent magnet properties are obtained.2,3 Previous work by Nesbitt&apos; on the ternary Co-Fe-V permanent magnet materials, resulted in the conclusion that the precipitation of a second phase, ?, was responsible for the permanent magnet qualities of these alloys. However, unlike the conventional precipitation hardening alloys, the ? phase, which is stable at high temperatures, is precipitated in the low temperature a phase upon aging. Recent work by Geisler and Martin"." has established that the order-disorder reaction inherent in the binary Co-Fe alloys is also present in the ternary Co-Fe-V alloys. On the basis of their investigation, it was proposed that ordering of the a phase is also a possible explanation of the mechanical and magnetic hardening in these alloys. Thus, it appears that the Co-Fe-V alloys are rather unique in that two solid state transformations are taking place which may account for the properties of these alloys. In view of these facts, the present investigation was undertaken to study the mechanism of mechanical and magnetic hardening in the 10 pct V-Co-Fe alloy with particular reference to the role of the ? phase; and to correlate the results of this study with other magnetic properties. In addition, the influence of increasing vanadium contents on the transformation of Co-Fe alloys was also undertaken to supplement the main objective of the investigation. Experimental Procedure The Co-Fe-V alloys used for study consisted of: 1—a commercial lot of Vicalloy I, a 10 pct V alloy, and 2—alloys prepared by powder metallurgy and containing 52 pct Co and 0 to 14 pct V, balance iron. The commercial alloy was received in the cast and forged condition and was used primarily for the detailed study of hardness, electrical resistivity, and magnetic properties of the 10 pct V alloy. The powder metallurgy alloys were prepared from electrolytic iron, commercial cobalt, and ferro-vanadium powders. The processing cycle was as follows: 1—Elemental powders were weighed and mechanically mixed for 24 hr. 2—Mixed powders were pressed at 60,000 psi into the form of bars 5/8 x 5/8 x 5 in. 3—Bars were sintered in an atmosphere of purified dry hydrogen for 24 hr at temperatures ranging from 1380" to 1340°C, depending upon the vanadium content. The as-sintered bars were cut in half; and one half was machined to 7/16 in. diam for thermal analysis. The remainder of the bar was hot forged to approximately 0.050 in. for X-ray diffraction analysis. Additional details of sample preparation are presented subsequently according to the type of test performed. Thermal Analysis Measurements One of the simplest and most widely used tools for determining phase relations as they change with temperature and composition is to study the rate of change of temperature of a material as heat is supplied or extracted at a constant rate. The thermal analysis method used in this investigation is the inverse-rate method described by Smith.&apos;

Jan 1, 1954

By Frances H. Clark

IN wartime, the fabrication and use of metals assumes increased importance, for a modern war of sizable proportions cannot be undertaken with- out a vast supply of this material. Light alloys of aluminum and magnesium, which were comparatively unknown in the last war, will play an important role in this one. Due to research on a vast scale in the interim of years, industry is today well equipped to fabricate metals on an enormous scale. Research for the past year has been most productive and a comprehensive survey in the space of this article would be impossible. It is hoped that the stressing of a few items here and there out of a great quantity of able work will indicate in a small way the advances made within recent times. Specific references to the literature will be omitted at the Editor&apos;s re- quest but some authors of papers that I have found of special interest are mentioned.

Jan 1, 1940

By W. C. Leslie, R. M. Fisher, K. G. Carroll

J. S. Bowles and J. K. Mackenzie (Commonwealth Scientific and Industrial Research Organization, Mel- bourne, Australia)—The authors are to be congratulated for giving the first direct experimental proof that the observed macroscopic strain accompanying marten-site transformations is not necessarily a shear. We

Jan 1, 1953

By Ulf S. Landergren

Diffusion coefficients and marker movements have been determined in brass using welded couples. Three different concentration ranges were employed at 750°C, while a fourth concentration range was measured at 500°, 600°, 700°, and 800°C. Marker displacements toward and porosity development in the high zinc side of the couple were observed in all cases. The results were interpreted as favoring a vacancy diffusion mechanism. IN the large amount of diffusion literature which has accumulated during the past half century, very few experimental data have been presented on body-centered-cubic metallic phases. The impressive combination of experimental and theoretical studies which has given the vacancy mechanism a fairly secure position as the preferred mechanism for most face-centered-cubic metals has no counterpart for body-centered-cubic phases, although the theoretical picture seems to be in a better state than the experimental work. The attempts to select diffusion mechanisms by the theoretical calculation of the activation energy for the various conceivable unit diffusion processes and the energy of formation of the type of lattice defect required have been based on models of two kinds. The first kind utilizes potential functions fitted for certain properties of pure crystalline copper; the second kind is fitted for sodium. The latter is an example of an open metal in which the metallic ion is small compared with interatomic distance,&apos; while the former is characteristic of the larger class of metals in which the ions have diameters about as large as the interatomic distance. brass should resemble copper in this fundamental property. It should be evident that results on the mechanism of diffusion obtained for one of these classes need not necessarily apply to the other. Huntington and Seitz, using a model of the copper type, showed that the activation energies to be expected for interstitial migration or a two-atom direct interchange were much higher than that for vacancy diffusion. Zener pointed out that a four-atom ring rotation would involve an activation energy considerably lower than the two unlikely mechanisms although still higher than the vacancy mechanism in the face-centered-cubic structure. Furthermore, he argued, the four-atom rotation would be relatively easier in a body-centered-cubic lattice and might be competitive with the vacancy mechanism. Paneth showed that a body-centered-cubic crystal-like sodium could form interstitial defects of special type, called a crowdion, by crowding an extra atom into < 111 > directions where the defects would affect a line of about eight atoms. The defect was constrained to move along the defining line. Unlike the face-centered-cubic case, therefore, the theoretical calculations for the body-centered-cubic mechanism do not provide a very clear-cut answer, and the absence of adequate experimental studies is the more serious. This investigation was undertaken for the purpose of obtaining useful information on diffusion in body-centered-cubic phases. brass was chosen for several reasons: 1—the phase field is sufficiently wide at high temperatures to permit the use of diffusion couples with adequate concentration ranges for satisfactory chemical analysis, 2-—thermodynamic activity data are available for the calculation of mobilities from diffusion coefficients, 3—the phase is as close to the theoretical model of the copper -type on which the calculations have been done as any body-centered-cubic metallic phase is likely to be. Measurements of the movements (if found) of inert markers during the diffusion process would indicate that a vacancy or Or mechanism played a major role in diffusion in brass. If no marker movement should be found, a ring mechanism would be indicated but not assured. Experimental Procedure The alloys used in these experiments, listed in Table I with analyses, were supplied by the Naval Research Laboratory. They were from ingots of 3x3 in. cross-section, heated to 760°C, reduced under a

Jan 1, 1957

By Frank W. Clinard, Oleg D. Sherby

The effect of multiple an transformations on diffusion in a dilute iron alloy was studied. Inter-penetration between iron and an Fe-Co alloy was evaluated, under thermal-cycling conditions chosen so as to transform only the iron-rich side of the diffusion couple. It is shown that a small but consistent diffusion enhancement by a factor of two results when data from the transforming side of thermally cycled specimens are compared with predicted inter diffusion coefficients based on isothermal data. No enhancement is found when data from the mntransforming side of such couples are compared with predicted diffusivities. It is suggested that short-circuiting by moving pain boundaries is the most likely mechanism contributing to the observed diffusion enhancement from multiple phase transJormations in iron. Transformation-induced Point defects, dislocation-pifie short-circuiting, and point defects generated by dislocation motion are less effective in contributing to enhanced diffusion under the experimental conditions utilized. SO far as is known, no studies have been made of the effect of allotropic transformation on diffusion. previous to this investigation. Studies have been made, however, of the influence of transformation on the diffusion-controlled processes of spheroidization and sintering. Cullen1 has described a method of accelerating spheroidization in an AISI 1030 steel. Partial transformation between a and ? phases was uti-

Jan 1, 1965

By D. Y. F. Lai, R. J. Borg

Tracer diffusion of Fe59 has been measured in the a-stabilized Fe-1.8 at. pet V alloy from 700° to 1500°C. The activation energies are obtained in both the presence and absence of magnetic order. Furthermore, it is established that diffusion in the alloy is identical to that in pure iron and consequently the values of Do and Q accurately represent the temperature dependence for self-diffusion. The purpose of this investigation is to obtain an accurate estimate of the temperature dependence for self-diffusion in bee iron in both the presence and absence of magnetic order, and, in so doing, to establish the temperature range of the magnetic effect.&apos;" As the temperature interval suitable for diffusion measurements is severely limited in both bee phases of pure iron because of the intervening fee ? phase, the experiments were performed on an a-stabilized alloy containing 1.8 at. pet V. This alloy is bee over the entire range from room temperature to the melting point. Although there have been several independent investigations of self-diffusion of iron in a, iron,1, 3-6 there still exists considerable disagreement regarding the values of Do and Q for the paramagnetic region. The two systematic studies of diffusion in 6 iron6, 7 previously reported are also only in fair agreement; but in view of the extremely small temperature range available for diffusion studies, i.e., 1390o to 1535oC, this is not surprising. It is comparatively easy to obtain accurate values of Do and Q for the a-stabilized alloy inasmuch as measurements can be made over the entire temperature range -700o to 1500°C. However, in order to assume that these same values apply to pure iron requires careful comparison of the data in the a, and 6 regions in both the alloy and pure iron. We have made several measurements in the appropriate temperature ranges and are unable to establish any systematic difference between the diffusion coefficients of iron in pure iron and in the alloy. We therefore conclude that the values obtained for the alloy are truly applicable to pure iron; the complete evidence favoring this conclusion will be discussed later in this paper. EXPERIMENTAL The experimental methods will be given here only in barest detail since they have been thoroughly de- scribed elsewhere.l, 7 The alloy was prepared by induction melting and chill casting under argon. Diffusion samples were machined from the ingot and annealed in hydrogen for several days at 900°C to give an average grain diameter of 1 to 2 mm. The penetration profiles of the tracer were established by a sectioning technique, the residual activity being counted after the removal of each section. The tracer used is Fe59 which emits two high-energy ? rays of 1.098 and 1.289 Mev, respectively; these were detected by a ? scintillation counter equipped with a pulse-height analyzer. For the measurements in the temperature range -700o to 1130°C the samples were vapor-plated with Fe59, encapsulated in quartz under vacuo, and annealed in resistance-heated furnaces which are controlled to ±1°C. The specimens diffused at higher temperatures are prepared as edge-welded couples, the two halves being separated by a thin washer of the alloy to prevent sintering. The diffusion anneal is then carried out by inductive heating under a dynamic vacuum. The temperature is monitored pyrometrically. RESULTS The diffusion coefficients are obtained from the penetration profiles in the usual way using the error-function complement relationship. The results over the entire temperature range are shown in Fig. 1. In the linear region, 900o < t > 1500°C the least mean squares (lms) values of the diffusion coefficients are given by D = 1.39 exp[-(56.5 ± 1) x 103/Rt] cm2/sec [l] The average departure of the measured diffusion coefficients from the values given by Eq. [1]In order to determine whether or not the slope is truly constant over the entire range from 900" to 1500°C, the data are arbitrarily divided into two groups, the first containing values between -900" and 1133°C and the second between -1133o and 1500°C. The lms values for the two groups are given by Eqs. [2] and [3]: D = 0.519 exp[-55.7 x 103 /RT](900° to 1133°C) cm2/sec [2] D = 1.45 exp[-56.7x103/RT] (1133° to 1500=C) cm2/sec [3] Thus, there is no significant difference between the high- and low-temperature segments of the linear region. This not only assures us of the consistency of the values obtained by induction heating as compared to those obtained from the resistance-heated

Jan 1, 1965

By R. A. Wolfe, H. W. Paxton

Self-diffilsion coefficients for cr51 and Fe55 in 12 pct Cr-Fe and 17 pct Cr-Fe for Fe55 in chromium, and for Cr51 in vanadium have been measured. The results are compared with other values for the Fe-Cr system, and with the various theories of diffusion in hcc metals. Some empirical correlations are discussed between Do and Q in hcc systems, or, expressed differently, the constancy of ?G*/T solidus for seveval bcc metals and alloys is noted. It appears very probable that a vacancy mechanism is operative in bcc metals, hut this cannot he stated with certainty. THE great bulk of work on diffusion in metals, both experimental and theoretical, was for many years concentrated on those with close-packed and, in particular, fcc lattices.1,2 There appears to be little doubt that the mechanism of diffusion in these solids is vacancy migration, leading to mass transfer and in substitutional solid solutions to a Kirken-dall effect.3,4 For bcc metals, the picture is much less clear. The Kirkendall effect certainly occurs in several alloys.5-10 However, attempts to understand the factors contributing to the pre-exponential in the usual expression for the diffusion coefficient D =D, exp {-Q/RT) by extension of ideas useful in close-packed lattices have not always been successful. Zener,11 Leclaire,12 and Pound, Paxton, and Bitlerl3 have suggested that various forms of ring diffusion may be important in some bcc metals. For close-packed metals, Do is usually about 1 sq cm per sec and Q - 35Tm kcal per mole (Tm = melting temperature in OK). The theory of Pound et al. suggests for ring diffusion that Do may be about 10-4 and Q, although difficult to calculate with any precision, would be significantly less than 35 T,. The experimental results on self and solute diffusion in ? uranium14,15 and ß zirconium,10 and for solutes in 0 titanium,17 and possibly for self-diffu- sion in chromium below about 0.75 T,," gave some credence to this theory. However, not all bcc materials display low values of DO and Q, and the exceptions were not predicted by any theory. Furthermore, it has recently become apparent that, in bcc materials, log D is not always linear with T-l if a sufficiently wide range of temperature is studied.16,18 This variation may be such that Q may increase18,19 or decrease20 with increasing temperature. The present work was undertaken in an attempt to provide further diffusion data on bcc metals, and to try to understand the factors which contribute to differences in behavior between the various elements. For part of this work, the Fe-Cr system was chosen since it is of considerable technological importance, and data on 12 pct Cr and 17 pct Cr alloys appeared well worthwhile to supplement that existing for the remainder of the stern.18,22 The diffusion of Fe55 in chromium was studied as an example of a more or less "normal" tracer element in a possibly abnormal host lattice. Finally, no data were available for vanadium, the neighbor of chromium in the periodic table, because of lack of a suitable isotope so cr55 was used as a tracer in a few preliminary experiments. For convenience, we shall refer to elements whose Do and Q are low compared to those predicted by Zener&apos;s theory as "anomalous". PROCEDURE This investigation determined self-diffusion rates by means of radioactive tracers and the integral-activity method first utilized by Gruzin.23 In this method a thin layer of radioisotope of the diffusing element is plated or coated onto a planar surface of the diffusion sample, which is then given an isothermal-diffusion annealing treatment. The determination of an activity-penetration curve involves measuring the residual activity of the specimen after each successive layer or section has been removed parallel to the original planar surface. The method used here is essentially the same as that used by Gondolf18 and Kunitake.21 Two radioactive tracers, cr51 and Fe55, were used in this investigation. Diffusion coefficients were determined for the diffusion of one or both of these tracers in four different materials, viz., Fe-12 wt pct Cr alloy, Fe-17 wt pct Cr alloy, chromium, and vanadium. The diffusion samples had nominal dimensions of 1.5 cm diameter and 0.5 cm thickness. The grain size was several millimeters for the Fe-Cr alloys and at least 1 mm for the chromium and vanadium samples. Accurately planar surfaces

Jan 1, 1964

By L. D. Hall, S. J. Rothman

The diffusion of lead and of trace amounts of bismuth in liquid lead have been investigated by the capillary method, using ROD and ROE as tracers. The results are compared with existing theories of diffusion in liquids, the agreement with theory being fair. The heat of activation for self-diffusion in lead is found experimentally to be close to the corresponding activation energy obtained from viscosity data. The pioneer data of Groh and von Hevesi for the self-diffusion of liquid lead, using ThB as a tracer, fit in with the present results. OMPARED with the volume of effort expended on solid state diffusion, there has been little attention paid to diffusion in liquid metals. A review of experimental work in this field, covering most of the studies up to 1951, was given by Jost.&apos; More recent papers include those by Hoffman2 on self-diffusion in mercury, Morgan and Kitchener" and Stark, Shchilishchev, and Kazachkev4 on the diffusion of various elements in liquid iron, and Grace and Derge on the diffusion of bismuth in bismuth-lead alloys. The need for more experimental data to aid in the evaluation of existing theories prompted the

Jan 1, 1957

By R. E. Ogilvie, J. I. Goldstein, R. E. Hanneman

The interdiffusion coefficients for the Fe-Ni system were determined as a function of composition in both the a and y phases at 1 atm pressure. The inter diffusion coefficients were also determined in the y phase at 40 kbar pressure. The concentration gradients were measured with an electron probe, and the diffusion coefficients were calculated by the Matano analysis. At 1 atm, Ey increases with nickel content up to 65 at. pct Ni. The relationship of Dy with temperature and nickel concentration, up to 50 at. pct Ni, is given by the equation DY-l atm = exp(.0519 CN~ +1.15) MANY of the transformations that occur in the solid state are diffusion-controlled, especially those involving precipitate growth in most metallurgical systems. Multicomponent diffusion is usually expressed in terms of a chemical or an inter diffusion The activation energy Q at 1 atm decreases from pure iron to 60 to 70 at. pct Ni and decreases from pztre nickel to 60 to 70 at. pct Ni. The Kirkendall marker movement indicates that DFe is greater than DNi zcp to about 60 at. pct Ni. Above 60 at. pct Ni, however, DNi becomes peater than DFe The inter diffusion coefficients were also compared with those calculated from self-diffusion data. A comparison between calculated and measured coefficients shows only rough agreement. The effect of 40 kbar pressure is to decrease DyPl atm by an order of magnitude. The activation volumes for high-pressure diffusion are in favorable agreement with theoretical models developed for self-diffusion. coefficient which determines the rate at which material is transferred from the matrix phase to the precipitate and hence determines the growth rate of the precipitate. Since the numerical values of the diffusion coefficients differ widely as a function of the metallic system, temperature, and pressure, the diffusion coefficients must be determined for each particular environmental condition. Of long-standing interest is the growth of the classic Widmanstatten pattern in metallic (Fe-Ni) meteorites. The growth rate of the pattern is controlled by the diffusion coefficients in the Fe-Ni system. Since these coefficients are essential to calculating the growth rate of the pattern, we found it necessary to determine the diffusion coefficients

Jan 1, 1965

By T. Kunitake, H. W. Paxton

The self-diffusion coefficient of chromium in various alloys in the iron-chromium system has been measured. A variation in Dofrom 10-4 for pure chromium to a maximum of 102 near 60 pct Cr appears with a range of activation energy from about 50 to 80 kcal per mol. The general pattern of behavior observed is not consistent with present theories of diffusion by a vacancy mechanism. Suggestions are made for further clarifying experiments. RECENT work on the self-diffusion of chromium,&apos; ?-uranium2-4 and ß-zirconium,23, 26 has indicated that these elements have very low Do (10-3 to 10-4) compared to that conventionally expected for vacancy diffusion. Pound et al.5 have suggested that this may be due to ring diffusion being operative, and have proposed a model which gives Do of the correct order of magnitude. The essential difference between this model and Zener&apos;s earlier model6 is that whereas Zener assumes tight coupling between atoms in a ring to establish a vibrational frequency, Pound&apos;s model allows the atoms to vibrate essen- tially each in their own normal modes. The probability of each atom moving in the correct direction to cause a ring of n atoms to rotate by 2p/n must then be calculated as a "synchronal entropy" which may be -20 e.u. for a 4 atom ring. In iron, the latest determinations of self-diffusion yield Do in the range 18 to 522.7-9 The most accurate is probably that of Borg and Birchenall8 which is 118 ± 3. This value is not inconsistent with that expected for a vacancy mechanism. Since iron and chromium are very closely similar in size, are both transition metals, and their solution is thermodynamically ideal at temperatures of interest in diffusion, it appeared that a determination of self-diffusion coefficients, Do and Q as a function of composition would be of substantial interest to diffusion theory. The following measurements were made with this purpose in mind.

Jan 1, 1961

By S. Yukawa, M. J. Sinott

The low solubility of bismuth in copper and its segregation at copper grain boundaries with resulting embrittlement is well known.1-3 The heterogeneous diffusion of liquid bismuth into polycrystalline copper has also been studied.4 This work presents some further observations on these phenomena using high-resolution auto radiography to detect the presence of bismuth in copper polycrystalline and bicrystal specimens. The auto radiographs were made by the stripping film technique which produces an autoradiograph that is intact with specimen surface and can be examined simultaneously with the under-lying microstructure. The details of the technique have been described previously.3 In one series of experiments, high-purity copper was alloyed with 0.02 to 0.03 wt pct radioactive bismuth obtained from the Oak Ridge National Laboratory. Melting and casting was done under a

Jan 1, 1960

By John Chipman, Helen Towers

DEVELOPMENT of a simple radioactive tracer technique for measurement of the diffusion coefficient of calcium ion in liquid slag has already been reported. The investigation was of a preliminary nature designed to explore the possibilities of using radioactive isotopes and to establish a method, and no great accuracy was claimed for the results. It was shown, however, that Ca15 could be used successfully, and the values obtained for the diffusion coefficient appeared to be of the right order of magnitude. However, it was considered that a better technique could be devised for carrying out the diffusion. The experiments were, therefore, continued to find a more accurate method, to check the calcium results, and to extend the investigation to include measurements of the diffusion coefficient of silicon in the same slag. There were two flaws in the original technique. Convection was the cause of too many failures, and the interface between the two slags in the diffusion cell was curved rather than plane, probably due to a surface tension effect caused by the addition of the tracer Ca45 in the form of chloride. Two methods of overcoming these defects were tried. Since it was planned to study diffusion of the short-lived isotope Si"&apos; also, the development of methods was restricted to those which gave promise of being reasonably rapid. Method 1—The principle was identical to that previously used with respect to the diffusion conditions, but there were three differences in technique. Temperature control was more rigid, the tracer was added to the slag as Ca45O rather than the chloride and, finally, an effort was made to translate to high temperatures the elegant method developed by Cohen and BruinV or bringing together the two liquid layers and beginning the diffusion. A master slag was melted in graphite, ground to a fine powder, and analyzed (38.5 pct Ca0-20.9 pct Al2O3-40.5 pct SiO2). The tracer was received as Ca45C12 in an aqueous solution of CaC12 + HC1. From this, calcium oxalate was precipitated, ignited to calcium oxide, and added as such to the slag to give an activity of approximately 5 µc per g. The mixture was then melted into the form of a cylinder 3/16 in. diam and about 1 in. long. A solution identical to the tracer solution without Ca45 was given exactly the same treatment. The resulting Ca40O was mixed with that slag which was to form the tracer-free half of the diffusion cell, and this was likewise melted into a cylinder of the same size. Slag cylinders, two with and two without Ca15, were set in graphite blocks as shown in Fig. 1, and the crucible was lowered into position in the fur-

Jan 1, 1958

By T. S. Lundy, R. E. Pawel, F. R. Winslow, C. J. McHargue

The volume-diffusion coefficients of Nb-95 and Ta-182 in niobium have been measured over the temperature range 878° to 2400°C. High-temperature specimens (T 21500°C) were sectioned by conventional lathe and grinding techniques while low-temperature specimens were sectioned by an anodizing and stripping technique previously described. Over a range of more than ten orders of magnitude in diffusivity, the self-diffusion data is well-represented by the equation: D = (1.1+ 0.2) exp [- (96,000 * 900)/RT ] cm2/sec Upward deviations from this line are explained by short-circuit diffusion in combination with the loss of resolution of the sectioning techniques. Oxygen contamination within the range 100 to 450 ppm was found to have little or no effect on the self-diffusivitj 01 niobium. In a limited number of measurements, the diffusivity of Ta-182 in niobium was found to be about half that of Nb-95 in niobium. These data may be described by D = (I.0+0.90.5)exp [- (99,300 * 2400) /RT ] PREVIOUSLY reported studies of self-diffusion in niobium were by Resnick and Castleman1 and Peart, Graham, and Tomlin.2 The data of Resnick and Castleman over a temperature range 1585" to 2120°C show considerable scatter but. can be represented by D = (12.4 *0.8) exp [-(105,000+3OOO)/RT] cm2/sec The limited data of Peart, Graham, and Tomlin follow the equation D = 1.3 exp (-95,00O/RT) cm2/sec In view of recent diffusion data for ß zirconium3 and 0 titanium4 in which Arrhenius-type plots do not yield straight lines and of the potential high- temperature uses of niobium and its alloys, we thought it important to investigate the diffusion behavior of this bcc refractory metal over a large temperature range. This has recently been done for molybdenum,5 chromium,6 vanadium,"" and tantalum.9 Niobium was a particularly suitable material for this study largely because of the development of an anodizing and stripping sectioning technique which enables diffusivities as small as l0-19&apos; sq cm per sec to be determined accurately.10 In addition to self-diffusion data, data were obtained for the diffusion of Ta-182 in niobium and for the effect of varying oxygen content on self-diffusion. I) EXPERIMENTAL PROCEDURE Materials. The niobium was obtained from three sources: 1) melting stock from Shieldalloy Corp., 2) electron-beam melted single crystals, 1/2 in. diameter, from Semi-Elements, Inc., and 3) oxygen-doped, electron-beam melted single crystals from Materials Research Corp. The melting stock was electron-beam melted and specimens 5/8 in. diameter by 1/2 in. long were machined from the center of the bar. The analysis supplied by the vendor and an interstitial analysis of one of the specimens are given in Table I. The single crystals (type 2 above) were cut with a jeweler&apos;s saw to specimens about 3/8 in. long. The interstitial analysis of this material is given in Table 11. This material was used

Jan 1, 1965

By Cyril Wells, Paul E. Busby, Donald P. Hart

EARLY workers in the field have established that the diffusion of nitrogen follows normal diffusion laws. Concentration-penetration data from layer analyses of reasonably pure iron specimens nitrided in an atmosphere of CO-NH, are given in the literature,&apos; and were utilized to calculate diffusion constants for nitrogen. The diffusion of nitrogen in y-iron may be represented approximately by the equation D - 1.07x10-&apos; e-:".&apos;m&apos;"&apos;r The diflusion coefficients, D, given by this relation represent mean values, since carburization occurred simultaneously with nitriding. The rate of diffusion in a-iron has been determined from calculations based on data obtained by measuring the internal friction of samples at each of several frequencies and temperatures.&apos; The expression for the diffusion of nitrogen in a-iron derived from these experiments is D= 1.4x10-1 e-it tudent Increasing carbon content decreases the diffusion of nitrogen in a-iron.3 Oxygen, either dissolved or combined, has been shown&apos; to exert a marked influence on the diffusion of nitrogen from a nitriding atmosphere. D values increase rapidly with increasing amount of dissolved oxide in the metal, but diffusion is completely inhibited if preliminary oxidation has produced the slightest trace of iron oxide on the surface. Diffusion coefficients have also been calculated from permeability and solubility data,&apos; but these values are of questionable accuracy since the boundary conditions assumed may not actually exist. Experimental Method The method selected for determining the diffusion coefficient of nitrogen in a-iron in the present investigation was the so-called fractional saturation method. This method consists of heating the specimen in an atmosphere of H2-NH8 mixtures of known composition for a predetermined time. Following this treatment, the specimen is analyzed for nitrogen; and the diffusion coefficient is calculated from a known relationship of the fractional saturation, time of diffusion, and dimensions of the specimen. This relation is based on a solution of Fick&apos;s second law and assumes that the surface of the specimen immediately reaches a given concentration which remains constant during the diffusion anneal. Tabular solutions are available4-5 and provide the relationship between fractional saturation and the quantity \/Dt /a, where D is the diffusion coefficient in cm2 per sec, t is the time of the diffusion anneal in seconds, and a, in the present experiments, is the half-thickness of the slab specimen in cm. The method is adequately discussed elsewhere.9-10 Material The material for this study consisted of 3/32 in. thick (0.24 cm) carbonyl iron plate. The major impurities in this iron are oxygen and carbon, which are readily removed by long time high temperature treatment in an atmosphere of purified hydrogen. In order to reduce the diffusion anneal time for saturation runs, some of this material was rolled to a thickness of 0.015 in. (0.038 cm) prior to the hydrogen treatment. Hydrogen treatments were carried out in a hydrogen-purified Armco iron tube heated in a global

Jan 1, 1957