Search Documents

By W. Buchanan, O. Johari

STEREOGRAPHIC projection has been one of the most useful aids in crystal-structure studies. With the increasing use of electron microscopy in metallurgy and other fields, stereographic projection is being used in many ways1 to interpret the results from electron diffraction patterns and transmission micrographs. For many of these applications, e.g., trace analysis, twinning, and so forth standard projections of various indices are required.

Jan 1, 1967

By N. F. Fiore, M. F. McGuire

THIS communication describes a series of experiments in which dislocation etch pitting was used to ascertain the depth of damage in crystals which were mechanically polished or which were cut with either a commercial metallographic cut-off wheel,* a com- All specimens were cut from single crystals of American Smelting and Refining Co. 99.999+ pct pure copper which were grown from the melt by the Bridge-man technique. The test specimens were disks, with faces being (111) planes and the disk axis a (111) direction. The disks were annealed in a vacuum of 10-5 Torr for 24 hr at 950°C prior to cutting or polishing and were given a light electropolishing in an ortho-phosphoric acid bath before etch pitting. Both Youngs1 and Livingston's2 reagents were used for etch pitting and the reagents gave identical quantitative results; however Livingston's reagent produced less staining and so all photomicrographs shown are of {111) surfaces etched with this solution. The surface of the annealed crystals showed well-defined subboundaries and a density of dislocations of about 1 X 106 per sq cm excluding those present in subboundaries, Fig. 1. The depth of cold work produced during mechanical polishing was determined after an annealed specimen was mounted in a paraffin holding block to protect it from extraneous damage. The exposed (111) surface was ground on metallographic papers and then polished on billiard cloth through 0.05-p alumina abrasive. It was necessary to remove about 0.3 mm (-0.01 in.) to obtain a structure representative of the annealed sur- face, indicating that the crystal was damaged to this depth during mechanical polishing. The cold work caused by standard cutting methods was measured in another series of tests in which cuts

Jan 1, 1967

By R. Kossowsky

In the course of a study concerned with structural stability of precipitation hardening nickel base super -alloys after exposures to elevated temperatures, some interesting effects of hydrostatic pressure treatment on Rene were observed. The evidence presented below suggests that hydrostatic pressure treatment of the solution heat-treated alloy can affect the dispersion of the y' (Ni3Al) phase upon subsequent precipitation heat treatment and long time aging. Rene 41 alloy has the following nominal weight percent composition: Al B C Co Cr Fe Mn Mo Ti Ni 1.4 0.006 0.7 11.3 19.2 1.0 0.05 9.4 3.3 balance The effect of successive pressure, precipitation heat treatment, and long time aging on the microstructure of this alloy was determined by subjecting a number of alloy buttons to five different thermopressure treatments. Pressure and precipitation heat-treatment parameters are listed in Table I. Fig. 1 shows y' morphology* of specimens in groups I and 11. The treatment results in a fine dispersion with particle size on the order of 950 and 1350A, respectively, see Table I. Fig. 2 shows the morphology of y' precipitate in

Jan 1, 1967

By Donald L. Johnson

REVIEWS on the applications of optical ellipsometry to the study of polycrystalline metal surfaces and oxidation-corrosion processes have been presented by Kruger and Hayfield, and white.' In view of the ability of this method to measure differences in film thicknesses to below 10A, ellipsometry was used to study nitrogen chromium surface reactions during the early stages of growth. Chromium specimens in the form of a vapor-deposited film on glass substrate were exposed to purified nitrogen for selected time increments at 700°F. This temperature was selected because it gave the best combination of exposure time and thickness change. Ellipsometer readings were taken immediately after exposure to determine polarizer and analyzer angles using monochromatic light at 5000A.2 After each set of readings was obtained, the specimens were exposed for an additional increment of time in nitrogen and a new set of readings taken. Film thickness was computed from LUCy's3 equation for nonabsorbing thin films. In Lucy's derivation, two equations for computing thickness were obtained. Peterson and Bashara4 have discussed the validity for using one of these equations when an independent value of the refractive index is used. The film was assumed to be Cr2N (Ref. 5) and was determined to be optically isotropic. Optical isotropy was determined by rotating the films about the surface normal and noting that the azimuth and phase difference did not change. Since Cr2N is reported to be hexagonal, it is assumed that optical isotropy results from randomly oriented grains of very small size. The thickness was computed using refractive indexes varying from 1.6 to 2.0. The maximum variation in computed thickness over this range for a given exposure was 20 pct. The data reported are based on a refractive index of 1.8. Reference readings on the metal surface were taken just prior to nitrogen exposure. No attempt was made to deoxidize or denitride the metal surface after vapor deposition, although specimens were observed to gain approximately 2A per day in air at room temperature. Twelve experimental exposures were made on a single sample over a total accumulative time of 132 min. Additional experiments over longer time intervals were made to determine the effect of heat up and cool down during and after exposure. As a result of these experiments, the thickness computed for each exposure was increased by 4A. Time profiles based on the linear law are similar to the oxidation profiles for many metals.' The rate of accumulation of the nitride film is rapid at first and slows down as the film thickness increases. The parabolic law x2 =k1t +k2 [11 gives a straight line in Fig. 1; therefore, it is concluded that the kinetic behavior is parabolic. Sey-bolt and Haman5 reported similar results for CrlN on chromium at 1000°C although the order of the thickness was much greater (10-3 cm). Analysis of thickness data based on other values of refractive index shows that the curves are shifted vertically but retain the same shape. According to Mott,7 it is expected that for thin films k2 is zero in Eq. [I]. Since the plot in Fig. 1 is linear through the origin, k2 is zero and the parabolic nature of the film growth is further indicated. The parabolic rate constant, calculated from the slope in Fig. 1, is 4.4 x 10-16 sq cm per sec. The corresponding work (+) required to bring an electron from the metal into the oxide is estimated to be 1.5 ev.7 The author thanks professor N. M. Bashara and D. W. Peterson, Electrical Materials Laboratory, University of Nebraska, for making the ellipsometer available and providing technical assistance. The Electrical Materials Laboratory is supported by the office of Naval Research, NONR 3718 (01). This work was partially supported by the Engineering Experiment Station and the Mechanical Engineering Department, University of Nebraska.

Jan 1, 1967

By Robert M. Asimow, Bisuddhi Riddhagni

THE use of back-reflection Laue photographs for crystal orientation is standard in many research projects. The time required for the orientation of any given crystal depends largely on luck and the skill of the investigator. In any case, where large numbers of crystals are used, this time can be quite considerable when the standard methods involving a Greninger chart and Wulff net are employed. For this reason the technique described below was developed. This new procedure applicable for cubic crystals (without preparation of new charts) largely eliminates the skill and luck factors and makes possible the determination of orientation in minutes. The procedure works in the following way: the Laue photograph is examined and arbitrarily an important hyperbola is picked out. An important hyperbola is defined as one which contains some spots through which several hyperbolas pass. Such a hyperbola inevitably corresponds to a low-index zone axis. The angles between the reflection planes causing all the spots on this hyperbola can be read directly from a Greninger chart in the usual manner. If the hyperbola contains more than five spots, these interplanar angles will positively identify the zone axis and the indices of each of the reflecting planes. However, initially since neither the zone axis nor the Miller indices of a diffraction spot are known, there are a very large number of possibilities that must be considered. Using tables of interplanar angles to check out every possibility would be quite tedious, so that this approach is not generally used. Instead, various poles are plotted on a stereographic projection. In the cubic system, however, it is possible to plot out once and for all the angles between all planes of a particular zone axis up to some upper limit of Miller indices. This has been done for all low-index zone axes in Fig. 1. Given these charts, it is quite a simple matter to identify the spots of the unknown hyperbola. All that is necessary is to plot the angles read off the Greninger chart to the same scale as Fig. l(b) on a strip of paper. Then this strip can be slid along each zone axis of Fig. l(a) until a good fit is obtained. Once this is accomplished, the Miller indices of each spot are given and thus the orientation problem is solved. To check, it is possible to pick a second hyperbola and see whether a compatible solution is obtained. This is most conveniently done by picking for the second hyperbola one that contains a spot common with the first. If the specimen axis is chosen parallel to the incident X-ray beam, then usually the [ loo] hyperbola will appear on the film as one of the more important hyperbolas and consequently will be either chosen initially or as a check. Using the [100] hyperbola, after the (001) or (011) spots have been identified, the latitude and longitude of the specimen axis can be read directly from the Greninger chart so that it is unnecessary to use a Wulff net for the entire orientation procedure. This makes the operation not only quicker but more accurate. Of course, if neither the first nor second hyperbolas correspond to [loo] then a third hyperbola can be chosen.

Jan 1, 1967

By R. Kossowsky

BRIDGMAN has listed five possible transitions in bismuth at 25°C which were observed as volume changes. He was unable however to detect the last three transitions by electrical-resistivity changes. Kennedy and LaMori have verified the pressures at the Bi 1-11 and Bi 11-111 transitions. They suggested however that the 125-kbar transition previously reported by Bundy3 actually corresponds to the 90,000 kg per sq cm transition observed by Bridgman.1 Jura4 reported a well-defined drop in resistivity at 92 kbar obtained by a four-lead method. The Bi 111-IV and Bi IV-V transitions have not been reported in any of the above-mentioned references.'-' Bundy3 observed a small dent in the low-temperature boundary of the Bi Vll region which is consistent with the presence of a 111-IV boundary. Zeitlin and Bray-man5 were the only ones to report small but well-defined drops in resistance of bismuth at 44.0 and 63.8 kbar, corresponding to Bridgman's 111-IV and IV-V transitions, respectively. High-purity (99.99 pct) 0.020-in.-diam 0.5-in.-long bismuth wire was used to calibrate the Uniram tetra-hedral high-pressure apparatus used in the present work. The bismuth phase changes at 25°C were correlated with changes in electrical resistance as a function of pressure. The output of a sensitive bridge constructed of carbon resistors and kept at a constant temperature of ±°C during the measurements was fed to the Y axis of a MoseleyX-Y recorder and the X axis was used for time scans. Specimens were pressurized using a hydraulic press with maximum capacity of 200,000 lb. Load was applied in increments of 2000 lb and held for about 3 min. The resistance was recorded at the end of this time period. Good contact was established below 20,000 lb. Typical recorder charts are reproduced in Figs. 1 and 2 which show that resistance fluctuations were practically nil during each time scan. Well-defined discontinuities in resistance were observed at 58,000 and 82,000 lb, Fig. 1. Two small but definite "blips" are noticed at 92,000 and 122,000 lb, Fig. 2. At 156,000 lb another drop is observed which

Jan 1, 1967

By Günter Petzow, Siegfried Steeb

DURING an investigation of the U-Co-A1, U-Ir-A1, and U-Ni-A1 ternary systems, three previously unknown phases have been identified which correspond to the phases UCoA1, UIrA1, and UNiA1, and which be- long to the C22 (Fe2P) type (P 6 2m). In the following the structure determination of the new ternary phases is reported. The alloys have been prepared in an arc furnace under purified argon from 99.97 wt pct U, 99.83 wt pct Co, 99.99 wt pct Al, 99.95 wt pct Ni, and 99.97 wt pct Ir. The pellets obtained were bright and brittle; chemical analyses agreed with the nominal compositions. To obtain equilibrium, the samples were annealed in evacuated quartz capsules at 950°C for 30 hr and then quenched in water. Guinier photographs were processed and evaluated according to a method already described previously.1,2 The hkl's, measured and calculated sin2 6 values, and in the case of UCoAl only calculated and estimated intensities, are given in Tables I to 111. Lattice parameters refined according to the least-squares method and the respective mean-square variations are presented in Table IV. Lattice parameters, especially the c/a ratio, lead to the assumption that the three phases are crystallizing in the C22-type structure. For the calculation of the intensities given in Table I, the atomic positions recently reported by wyckoff3 have been accepted for the C22 (Fe2P) type. The uranium atoms were assumed to be in phosphorus sites; furthermore the scattering amplitudes of the boron atoms have been averaged in the stoichiometric formula UB2 according to atomic percentages; i.e., for these atoms a statistical distribution on the iron sites was assumed. The two smallest distances between a uranium and

Jan 1, 1967

By V. Patarini, D. Kuehl, F. Galasso, M. Salkind

In recent years, a great deal of research has been directed toward the production of continuous low-density, high-strength, high-modulus boron fibers for aerospace applications. A technique which has proven to be quite successful for forming these fibers involves the reaction of boron trichloride and hydrogen on a hot tungsten wire. In this process, a boron fiber is produced which consists of a tungsten boride core covered with amorphous or crystalline boron depending on the temperature of the preparation.2,3 The purpose of this paper is to present the properties which have been obtained to date from measurements and observations on these fibers. Selected samples of boron filament produced by reacting boron trichloride and hydrogen on a 0.5 by 10-3 in. diam tungsten wire substrate using different reacting temperatures were cross-sectioned, polished, and photomicrographed with cross-polarized light. The temperature of the wire was measured using a two-color optical pyrometer. Photomicrographs of these samples are shown in Fig. 1. Three general types of boron, in addition to the tungsten borides, have been identified by means of X-ray powder diffraction techniques: a) at low temperatures (below 1400°C) the boron is deposited in the amorphous form; b) at temperatures between 1400° and 1500°C the boron is deposited in the ß tetragonal form; and, c) above 1500°C the boron is deposited in the B rhom-bohedral form. In the depositing range of 1400" to 1500°C the boron initially deposits as amorphous and becomes crystalline as evidenced by the large equi-axed grains which then merge into columnar grains in later stages of boron deposition. This can be seen in Fig. 1 and has been observed in the X-ray patterns of these fibers, which contain both halos from the amorphous boron and lines from the ß tetragonal form. Above 1500°C the only boron reflections observed in the X-ray patterns were those of the ß rhombohedral form and only columnar grains are seen in the photomicrographs. Photomicrographs of the surface in Fig. 2 show the smooth-surface kernal-type deposition which corresponds to the amorphous form and the rougher surfaces which correspond to the crystalline form. Mechanical measurements were made on amorphous boron produced at 1300°C rather than the crystalline boron because the strengths of these filaments were greater and were more consistent. A tensile tester consisting of a fixed crosshead containing a double

Jan 1, 1967

By G. T. Higgins

PREVIOUS workers1,2 have found evidence of exaggerated grain growth or secondary recrystallization on annealing sponge zirconium at temperatures high in the a, phase field, but no details of the orientation relationship between the secondary grains and "as-recrystallized" matrix were reported. In other hexagonal metals, namely dilute Mg-A13 and Cd-zn4 alloys, secondary recrystallization has been observed in which the secondary grains are related to the primary recrystallized structure by rotations about the c axis. These rotations are thus comparable with the rotations about the (111) axis observed in the secondary recrystallization of fcc metals. In the current work secondary recrystallization has been observed on annealing iodide zirconium at 800°C and the associated changes in crystallographic texture have been examined. The material employed was arc-melted iodide zirconium of the following composition: Impurity O H Fe Hf Pb I Ni C1 A1 ppm by wt 130 5 230 98 23 28 200 39 30 This was received as 0.25-in. plate, and, prior to the annealing studies, was cold-rolled to 0.040 in. Specimens 1 in. sq were chemically polished and annealed for various times at 800°C in a vacuum better than 5x 10-6 Torr. For metallographic examination specimens were first chemically polished and then anodized8 to enhance contrast between the grains when examined under polarized light. Changes in crystallographic orientation associated with the development of secondary grains were followed by (1011) and (0002) pole-figure determinations using a conventional Schulz goniometer in reflection. Filtered Cr Ka radiation was employed and great care was taken to avoid spurious re- sults due to the close proximity of the Bragg angles for the (1011) and (0002) reflections. Figs. 1 and 2 illustrate the change in grain structure on annealing at 800°C. After 24 hr the structure has ceased to be equiaxed due to the development of certain grains in preference to others. It is notable that the secondary grains do not grow to the extent observed in sponge zirconium.1,2 This appears to be due in part to the fact that normal grain growth is inhibited in the less pure sponge zirconium, with the result that the driving force for secondary recrystallization remains high, while at the same time the number of grains able to develop in the presence of the impurities is also smaller. As a consequence, in sponge zirconium those grains that do develop preferentially are able to grow into the fine -grained matrix without impingement, while in iodide zirconium the preferential growth of many grains leads to early impingement. The pole figures determined for specimens annealed for 5 min and 24 hr at 800°C are illustrated respectively in Figs. 3 and 4. These indicate that the dominant feature of the initial recrystallized texture consists of basal planes at an angle of 321/2 deg to the sheet surface, with a (1120) axis lying parallel to the

Jan 1, 1967

By F. Galasso, A. Paton

ThE recent emphasis on producing boron fibers by reacting boron trichloride and hydrogen on a hot tungsten wire has brought about a renewed interest in the tungsten boride phase diagram. The reaction of the boron with the tungsten core results in the formation of 6 WB, W2B5, and another tungsten boride phase, which was suspected of being the WB4 phase reported by Chrbtien and Helgorsky.1 While the powder-diffraction data for the boride in the fiber and those reported

Jan 1, 1967

By Mark J. Klein

The behavior of nitrogen in Cr-0.12 at. pct Y was studied in specimens of varying interstitial concentration and thermal history. It was found that soluble nitrogen can be introduced into this alloy by increasing the concentrations of either oxygen or nitrogen. For low nitrogen concentrations the nitrogen in solution is low relative to that in unalloyed chromium. However, the nitrogen in solution increases with increasing nitrogen concentration, approaching that in unalloyed chromium. When the alloy is supersaturated with nitrogen, the nitrogen precipitates readily from solution as a chromium nitride. The stress -induced relaxation of nitrogen is influenced by internal oxidation or internal nitrification of this alloy. THE addition of small concentrations of reactive elements such as yttrium have been found to enhance the low-temperature ductility of chromium.&apos; However, the mechanism by which the ductility is increased is not known. It has been suggested that this effect is caused by a drastic reduction in the concentration of nitrogen retained in solution and therefore a decrease in the number of dislocations locked by nitrogen, a factor that has been associated with an increase in ductility.1,2 However, a change in the type of interstitial precipitate induced by yttrium could also contribute to enhanced ductility of the alloy, since chromium nitride precipitates appear to be embrittling.&apos; In this regard, it has been shown that yttrium scavenges nitrogen in chromium to form YN,4 thus reducing the embrittling chromium nitride phase. However, the change in nitrogen solubility induced by yttrium and the partition of nitrogen between chromium and yttrium nitrides are not known. The object of this investigation is to study the effect of yttrium on the behavior of nitrogen in chromium to obtain information relevant to the mechanism by which scavengers enhance the ductility of chromium. EXPERIMENTAL PROCEDURES The experimental procedures used in this investigation have been described in detail elsewhere.5,6 Wires of Cr-0.12 at. pct Y, 15 cm in length and 1 mm in diam, were drawn from rods 0.6 cm in diam supplied by the General Electric Co. The grain size of the wires after recrystallization at 1150°C was somewhat nonuniform, varying from 0.006 to 0.06 mm in diam. The interstitial analysis for the wires in ppm by weight was about 160 0, 80 N, 12 C, and <2 H. Heat treatments were carried out after encapsulation of the specimens in quartz tubes containing argon. Changes in interstitial concentrations, or interstitial "loading" as it is termed in this study, were generally accomplished by annealing in oxygen, ammonia or methane at 1150°C. The internal friction was measured using a conventional torsional pendulum enclosed in an evacuated chamber. Electrical-resistivity measurements were made at liquid-nitrogen temperature.

Jan 1, 1967

By M. S. Rashid, C. J. Altstetter

Allotropic transformations in cerium have been studied by dilatometric, resistometric, X-ray diffraction, and metallographic techniques. The dilatometric study indicated that, on cooling below O°C, the high-temperature fcc phase, y, transforms partly to the hexagonal phase, ß, and, on further cooling, to the collapsed fcc phase, a. The amount of $ phase present at room temperature is increased by repeated cycling through the a-y transformation. It has been shown metallograPhically that the y-ß transformation has many characteristics of a martensitic transformation. In contrast to the y-ß transformation the ?-a transformation does not give the manifestation of a shear transformation. Small cellular ? domains of random shape and size collapse to a in a short time with no apparent coordination with neighboring domains. The considerable confusion in the literature over the existence of more than one high-temperature fcc phase is discussed. Two such phases have been reported in the literature and an attempt is made in this study to clarify the situation. Twelve fcc and two hcp structures have been shown to be easily reproduced or eliminated. It is proposed that the two "additional" allotropes reported in the literature and fourteen of the phases detected here are not allotropes of cerium but are due to contamination. CERIUM exists in several allotropic forms, but there is some disagreement over what the forms are. Furthermore, the conditions favoring the presence of a particular allotrope and the nature of the transformations from one form to another are uncertain. The objectives of this research were 1) to ascertain the allotropic forms of cerium, 2) to establish the conditions under which the allotropes exist, 3) to study the effects of annealing and thermal cycling on the allotropic transformations, and 4) to study the transformation mechanisms. Dilatometric, resistometric, metallographic, and X-ray diffraction techniques were employed. The form of cerium commonly found at room temperature is fcc and is designated ?. A complex hexagonal phase, 8, forms when y is cooled to slightly below room temperature. At still lower temperatures the y fcc structure transforms to an fcc form with a much smaller lattice parameter, termed a cerium. A bcc form, 6, which exists just below the melting point (800°C), will not be considered further in this work. There is a substantial body of experimental evidence (reviewed by Gschneidnerl) which favors the acceptance of these four allotropes, though some investigators have tried unsuccessfully to observe the ß hexagonal form.&apos;-&apos; There is disagreement, however, over the phase-transformation temperatures, due, in part, to broad hysteresis and overlapping of the transformations between the a, ß, and ? forms. The transformations are also sensitive to prior thermal and mechanical treatment. The differing purity of cerium used by different investigators is undoubtedly a factor. Cerium is difficult to separate from other elements and is quite reactive, igniting spontaneously when it is filed in air. The highest purity of cerium to date is reported to contain several hundred parts per million by weight of impurities, and early investigations were carried out on cerium containing several percent of impurities. There have been reports of more than one fcc allotrope at room temperature. Gschneidner, Elliott, and McDonald5 obtained diffraction patterns of an fcc phase with a lattice parameter about 1 pct less than that of the ? phase, instead of the y phase, on slowly cooling cerium filings from 23° to -198°C and warming them back to room temperature. However, when the sample was heated to 447°C and cooled to room temperature it consisted of only the ? phase. They have designated this new fcc phase "a-? intermediate", and say it is quite sensitive to impurities. After prolonged high-temperature treatment of a powder specimen, Weiner and Raynor2 obtained a diffraction pattern of an fcc phase of lattice parameter about 1 pct less than the ? phase. This they called the y&apos; phase. It could not be reconverted to the y phase and is claimed to be different from the a-? intermediate phase.5,6 Dialer and Rothe3 reported two fcc phases* after cycling their powder specimens between room temperature and -192°C. Gschneidner, Elliott, and McDonald5 suggested that one of the fcc structures obtained by Dialer and Rothe was equivalent to their "a-? intermediate" phase. Table I presents some pertinent data on the proposed allotropes. For the ?(fcc)-ß(hexagonal) transformation

Jan 1, 1967

By R. D. Nelson, J. C. Shyne

The ß and a ß transformations in plutonium were studied with particular emphasis on the transformation kinetics and microstructure. Significant observations are: 1) The kinetic data show conclusively that the ß — a transformation in high-purity plutonium can proceed isothermally with no athermal component. 2) Plastic deformation of the stable (3 phase retards the subsequent (3 — a transformation. 3) Plastic deformation of the stable a phase accelerates the a — ß transformation; the acceleration is attributed only to residual stresses. 4) The a and a?a volume changes occur anisotroPically in textured plutonium. 5) An apparent crystallogvaphic relationship exists between the parent and the product phases of the and (3 — a transformations. 6) Both applied uniaxial compressive stresses and uniaxial tensile stresses raise the starting temperature for the ß — a transformation. 7) A given uniaxial tensile stress favors the a — ß transformation more than an equivalent applied uniaxial compressive stress opposes the transformation. These observations of the (ß —a and a — ß phase changes in plutonium are consistent with known mar-tensitic transformations. ThIS paper elucidates some of the characteristics of the a— ß and ß —a transformations in plutonium. Because considerable conjecture exists about the mechanisms by which the phase transformations occur in plutonium, experiments have been performed to provide indirect information concerning the mechanisms responsible for the a —ß and ß -a transformations. Indirect information is of particular value in the study of plutonium because of the experimental difficulties presented by the metal. Single crystals have not been produced in any of the allotropes. The large density results in high X-ray and electron-absorption factors and consequently complicating X-ray and electron diffraction. The kinetics of ß — a and a — ß transformations of plutonium and the behavior of the transformations under a variety of conditions have been investigated in detail. Information about the mechanisms of the allo-tropic transformations of plutonium was obtained indirectly from three sources: 1) the effect of plastic deformation of the stable parent phase upon the transformation kinetics; 2) the behavior of the metal transforming under applied stresses; and 3) the microstruc-tural and crystallographic features between parent and product phases. PHASE-TRANSFORMATION CHARACTERISTICS In characterizing solid-state phase transformations in metals and alloys, it is useful to define several types of transformations. An aim of the present work was to identify the low-temperature transformations in plutonium by type, i.e., as martensitic or nonmar-tensitic. Practical definitions for these terms follow. The terms commonly used to categorize phase transformations lack universally accepted definitions. This confusion arises doubtlessly because some terms specify crystallographic or morphological character while other words have a kinetic or a thermodynamic connotation. For example, martensitic specifies certain definite crystallographic restrictions. Unfortunately, martensitic is sometimes used in an ill-defined way to imply kinetic characteristics. Further confusion attends the use of such expressions as nucleation and growth, diffusional, and massive. From time to time new systems of phase-transformation nomenclature are suggested; unfortunately none of these has gained general acceptance.1,2 The authors of the present paper have no intention of entering the controversy. We recognize that some readers may object to the nomencliture used here. For exampie, the terms military and civilian have recently been used in much the same way as martensitic and non-martensitic are used in this paper. This paper is intended to describe several specific details of the low-temperature phase transformations in plutonium. The authors have found it useful to identify these transformations as martensitic; the term was chosen as the best available to describe the experimentally observed features of the transformations studied. A martensitic transformation is one that occurs by the cooperative movement of many atoms; the rearrangement of atoms from parent to product crystal structure occurs by the passage of a mobile semico-herent growth interface. The geometric features characteristic of a martensitic transformation are a specific orientation relationship between the product and parent phase lattices, a specific habit-plane orientation for the growth interface, and a shape change with a specifically oriented shear component. There is no alloy partition between the parent and product phases in a martensitic transformation. Martensitic transformations may display either athermal kinetic behavior or thermally activated isothermal kinetic behavior. Some martensitic transformations occur isothermally, although more commonly martensitic transformations are athermal. Isothermal martensitic transformations are suppressible by rapid cooling. In athermal martensitic transformations, nucleation and growth are not thermally activated and the transformations are essentially time-independent. Nucleation, growth, or both can be thermally activated in isothermal martensitic reactions. Transformation of the parent phase into a marten-

Jan 1, 1967

By W. A. Backofen, D. H. Avery, G. A. Miller

An evaluation has been made of the relative importance of yield strength (?) and stacking-fault energy (y) to the rate of fatigue-crack growth in materials of fcc structure. Pure copper and its solid-solution al-loys with aluminum and nickel were chosen for the study because they provided sufficient range in both quantities of interest that either could be varied independently of the other. Experiments involved alternating tension and compression of flat specimens which were prepared with sharpened internal notches so that most, if not all, of the crack-nucleation interval could be eliminated. Growth rate (dC/dN) was concluded to be proportional to the square of the plastic-strain amplitude (€,,) over a strain range of approximately 6x 10-4 to 6 x 10-3. The factor, k, linking dC/dN and ep in dC/dN = kEp2 increased and decreased with corresponding variations in y, but it did not respond syste?>/atically to change in ay, indicating that y is the significant variable in crack growth at constant plastic-strain amplitude. In polycrystalline material, k varied by a factor of 5 over the available range of y. In a few single-crystal experiments on Cu-A1 alloys the growth rate responded less strongly to change in y. It has been suggested that single crystals behave somewhat differently than poly crystalline material because there is more extetnsive substructure near the grain boundaries in the latter, and this facilitates crack advance by separation along subgrain boundaries. A point of some controversy in current work on fatigue relates to the effects of strength and stacking-fault energy on crack growth. In recent experiments a separation was made between the cycling intervals for crack nucleation and the subsequent growth that eventually ends a specimen&apos;s fatigue life.&apos; The study was carried out on Cu-A1 alloys primarily, fatigued in alternating four-point bending to constant deflection. A nucleation interval of about 10&apos; cycles (at a total strain amplitude = 0.2 pct) was found to be insensitive to aluminum content in the range 0 to 7.5 wt pct, while the growth period was increased approximately forty fold over the same compositional range. The increase was not in any sense linear, however. Rather, most of the change occurred below 4 pct A1 or a stacking-fault energy, ?, of about 15 ergs per sq cm. It was argued that the plastic-strain amplitude was approximately constant, and therefore the effect of composition must have grown out of the reduction in stacking-fault energy. Several studies have shown that, with high ?, cross slip is encouraged, subgrain structure is introduced during fatigue, and cracking is aided through propagation along subgrain boundaries.1-5 Therefore, lowering ? sufficiently to interfere with substructure formation would be expected to retard growth rate. On the other hand, it is a general rule that resistance to fatigue cracking increases as strength is raised. Accordingly, there might still have been some doubt that Y was the controlling variable, since strength would be increased as y was lowered by the aluminum additions. To help in dispelling that doubt, an experiment was made on a polycrystalline Cu-Ni alloy similar in strength to the Cu-A1 alloys but of higher ?; the crack-growth interval was found to be essentially that of pure copper.&apos; Further support for this position on stacking-fault energy as it relates to crack growth is derived from work by Boettner and McEvily,6 in which the actual crack-growth rate was measured on samples previously notched so as to minimize the nucleation period. Unfortunately, it was necessary in isolating strength level to compare different alloy systems and grain sizes. Recognizing the complication, it was still concluded that growth may be retarded by a reduction in y, per se. A related study has also been made by Roberson and Grosskreutz.7 The zinc content of a brass was systematically changed to alter strength and stack-ing-fault energy, although not below the 15 ergs per sq cm at which pronounced change in growth interval was found in the earlier work. The results were limited to more or less conventional S-N diagrams so that nucleation and growth events could not be separated. No definite conclusions were drawn, but

Jan 1, 1967

By W. O. Philbrook, R. H. Spitzer, F. S. Manning

A generalized mathematical model has been developed to describe the kinetics of the gaseous, topo-chemical reduction of porous hematite spheres. Gas-solid reduction is permitted at each of three advancing interfaces and is controlled by a complex series-parallel sequence of chemical and transport steps. Despite the fact that transport resistances are significant to the advance of each interface, little deviation from linear advance is predicted for the first 98 pct of reduction. Calculations to predict the movement of magnetite /wustite interface by a solid-state diffusion mechanism have been incorporated into the earlier dense-pellet model.&apos; The porous-pellet model is found to give a better description than the dense-pellet one for McKewan&apos;s measurements2 of interface movements in dense polycrystalline hematite reduced in hydrogen-water vapor mixtures. THIS paper reports a continuation of work"3" directed toward the better understanding of gas-solid reactions and the development of mathematical models that may be helpful in the design or control of ore-reduction processes. Laboratory studies of the reduction of single oxide samples in tube furnaces normally employ an abundant excess of pure reducing gas. Reduction may appear to be adequately described by a model assuming gas-solid reaction solely at the iron/wustite interface.&apos;&apos;&apos; However, when iron ore is reduced in either a fixed bed or an efficiently operated counter-current column, there are portions of the bed where the partially spent gas is no longer capable of reducing wustite to iron. Nevertheless, the formation of the intermediate reaction products, wustite or magnetite, can and does continue. In order to account for this aspect of the reduction of beds of particles, three semiempirical mathematical models were devised.&apos; One of these was for the limiting case of completely dense particles, and two alternatives were provided for porous particles. For purposes of the analysis, a dense pellet was defined as one in which the wustite and magnetite layers were impervious to the reducing gas when the atmosphere was capable of forming iron. In a porous pellet, the intermediate product layers were assumed to be sufficiently permeable to allow reducing gas to penetrate beyond the iron/wustite interface to form wustite and magnetite by removal of oxygen as a gaseous product rather than by a solid-state redfiction mechanism. Further progress required a better understanding of the reaction of the individual particles with the gas stream. A comprehensive mathematical treatment was developed&apos; to describe the continually changing interplay of transport and reaction steps

Jan 1, 1967

By J. W. Spretnak, J. N. Cordea

The gain boundary relaxation in high-purity aluminum, nickel, copper, and silver was studied by means of a low-frequency torsion pendulum. Both internal friction and creep at constant stress tests were conducted. A lognormal distribution in relaxation times was found to account for the relatively wide experimental internal friction peaks and the gradual relaxation behavior during the creep tests. This distribution was separated further into a lognormal distribution of relaxation time constants and a normal distribution in activation energies. A spread of up to ±6 kcal per mole in the activation energies accounted for the major part of the distribution. A "double-peak" internal friction phenomenon was observed in silver. The activation energies in kcal per mole derived from the grain boundary relaxation phenomena are 34.5 for aluminum, 73.5 for nickel, 31.5 for copper, and 41.5 for silver. It was found that the rain boundary relaxation strength in these metals increases with the reported stacking-fault energy. GRAIN boundary relaxation phenomena have been observed in a large number of polycrystalline metals and alloys. Numerous investigations have been conducted to study the structure of the grain boundary through this relaxation process. One of the first investigators was Ke1-4 who observed that the activation energy for grain boundary relaxation in aluminum, a brass, and a iron was about the same as that for volume diffusion. He concluded that the grain boundary behaved as if it were a thin liquid layer with neighboring grains sliding over one another. Leak5 conducted experiments on iron of a higher purity and observed that the grain boundary activation energy is comparable with that of grain boundary diffusion. He suggested that, in metals where this relationship holds, the damping may be caused by a reversible migration of grain boundaries into adjoining grains. Nowick6 has presented an interesting view of inter-facial relaxation with his "sphere of relaxation" model. A relaxed interface is represented as one where the shear stress is greater than the normal value along the edges and zero in the interior of the interface. The region of the stress relaxation is pictured as a sphere surrounding the interface. From his calculations Nowick concluded that the slip along an interface is directly proportional to its length. Therefore, the time of relaxation, T, depends on the size of the relaxation interface. This means that in the Arrhenius relationship, t = TO exp[H/RT], valid for atom movements, the relaxation time T is predicted to be proportional to the grain diameter through the pre-exponential term, TO. Since the internal friction can be given as Q-1 = ?j wt/(1 + w2r2), where ?J is the relaxation strength and w is the angular frequency, an increase in grain size at a constant frequency will shift the peak to a higher temperature. A great deal of work has been done to determine the exact relationship between the internal friction and grain size.1,5,7,8 In metals, the grain boundary peaks are found to be lower and broader than predicted theoretically.&apos; The above model can explain this by a distribution in the size of the interface areas, represented by a distribution in the parameter tO, and an overlap of spheres of relaxation, represented by a distribution in activation energies. Both these phenomena result in an over-all distribution in the relaxation time, which could affect the internal friction peak height, breadth, and also position. This relationship between the experimental data and theoretical calculations appears very promising in the study of interfacial relaxation mechanisms. THEORY A lognormal distribution in t can sometimes be used to adequately describe the spectrum of relaxation times governing an anelastic relaxation. wiechert9 originally suggested such a distribution to explain the elastic after-effect in solids. This choice is particularly applicable to grain boundary relaxation when considering Saltykov&apos;s work.&apos;&apos; He found a lognormal distribution in the grain sizes within a metal. Recently Nowick and Berry11 have introduced a log-normal distribution in T into the theoretical internal friction equations. The form of the distribution function is where z = In(r/rm), and Tm is the mean value of t. The parameter ß is a measure of the distribution and is the half-width of the distribution when is l/e of its maximum, IC/(O). Nowick and Berry have described the methods to obtain the parameters Tm, ß, and ?,J from experimental internal friction and creep test data. In the idealized case, where only one relaxation event occurs with one relaxation time, only ?J and T are necessary to completely describe the event, and 0 = 0. For the broader internal friction curves 6 is some positive number greater than zero. The larger the 6, the greater is the half-width of the distribution in In t.

Jan 1, 1967

By M. L. Rudee, R. A. Huggins

Samples of iron containing nitrogen or carbon have been given treatments similar to those used in cold-work peak (CWP) measurements and examined by transmission electron microscopy. It was observed that the unusual and nonreproducible behavior of the carbon CWP can be explained by a strong tendency for carbon to form grain boundary precipitates at temperatures below those used for CWP measurements. These precipitates dissolved at the temperatures used in the CWP measurements. There was no evidence for nitrogen precipitation at grain boundaries. There was no indication of precipitation along dislocations in either carburized or nitrided samples given treatments similar to those of CWP measurements. Although it is possible that subelectron-microscopic clustering had occurred, this observation supports the theories of the CWP that are based on continuous atmospheres rather than on individual precipitates. In an earlier paper,&apos; the present authors developed a new distribution function to predict the occupation of sites for interstitial impurity atoms around a dislocation. When this distribution was applied to the case of carbon and nitrogen in iron, it predicted that, if the temperature dependence of the concentration of solute atoms in the matrix was controlled by the presence of equilibrium carbide or nitride precipitates, the tendency for nitrogen to segregate to dislocations would be greater than that for carbon even though their binding energies to dislocations are identical. The cold-work internal-friction peak (CWP) is considered by most authors to be produced by the interaction of interstitial impurities with dislocations. Many investigators have studied the CWP in iron containing carbon and nitrogen and have observed a significant difference between its behavior in the two cases. In this paper a series of experiments will be reported that were initiated to determine whether the application of the new distribution function would explain the observed differences between the carbon and nitrogen CWP. Although it was found that the distribution function, pev se, did not explain the differences, the differences became clear, and some insight into the mechanism of the CWP was realized. Before reporting the present experiments, the literature pertaining to the differences between the carbon and nitrogen CWP in iron and the various mechanisms proposed for the CWP will be reviewed. LITERATURE REVIEW Snoek2 first observed the CWP in iron specimens containing nitrogen, but also reported a weak and unreliable peak in carburized samples. Later, Ke3 established that the CWP height was proportional to the degree of deformation. The presence of nitrogen alone would produce a peak of the same size as found in a sample containing both nitrogen and carbon, and KG concluded that the CWP was caused by nitrogen. In a discussion of G&apos;s paper it was reported that Dijkstra had investigated the CWP in samples that contained only carbon. He found it to be much smaller than the nitrogen peak and "unstable". KG et al.4 charged specimens of iron with both carbon and nitrogen. They observed that the carbon CWP was much smaller than that observed in nitrided samples, but that aging at 300°C caused the carbon peak to increase. A similar treatment produced no change in a nitrogen peak. Annealing at higher temperatures caused the height of the CWP in both the nitrogen and carbon samples to decrease. This behavior was also observed by Kamber et al. 5 who found that the activation energy for the annealing of the CWP was identical with the activation energy for the self-diffusion of iron. They concluded that the annihilation of dislocations by climb caused the reduction in the CWP height. Kamber et al. studied the "unstable" carbon peak in detail. They measured both the Snoek and CWP during various aging treatments. In carburized samples, aging at 100°C caused the Snoek peak to disappear, although the CWP peak remained small. However, a subsequent treatment for 5 hr at 240°C caused the CWP to reach a maximum. They proposed that an additional location for the carbon, other than whatever site produced the CWP, is present. In nitrided samples the CWP was completely developed as soon as a measurement was made; additional sites are not present. No explanation of either the additional site or the difference in the behavior of carbon and nitrogen was offered. petarra5 performed a systematic study of the effect of composition on the CWP. Using three kinds of "pure" iron, he showed that there was a residual CWP when the carbon and nitrogen concentrations had been reduced to less than that detectable by Snoek-peak measurements. He observed the same general annealing behavior and composition dependence as previous investigators, with the following exceptions. On first measuring the carbon CWP, it was found to be identical with the residual peak, and essentially independent of the carbon content. If the CWP was measured a second time in the same sample, it increased in size, but was still only about one-fourth the size of a CWP in a sample of the same iron nitrided to the same composition. On the other hand, a series of annealing experiments showed that the nitrogen CWP was not al-

Jan 1, 1967

By R. C. Ku, R. G. Davies

The stress-strain behavior of Fe-C and Fe-Ni solid-solution alloys has been investigated at 25°C. It is found that the rate of solid-solution strengthening is dependent upon the strain at which the yield stress is measured, the macroyield stress being much more concentration-dependent than the microyield stress. In fact the addition of nickel does not lead to any increase in the microyield stress of iron. The primary effect of increasing the carbon or nickel content is to increase the slope of the stress-strain curve in the microyield region; this effect has been correlated, by means of transmission electron microscopy, with a decreased ability of the dislocations to cross-slip. It is concluded that the Mott-Nabarro theory is a good description of true solid-solution hardening in these and other iron-base alloys. A.LL theories of the solid-solution strengthening1-4 indicate that the flow stress of an alloy increases both with solute concentration and "misfit" between the solvent and solute atoms. The functional relationship between strength, solute concentration, C, and misfit depends upon the particular theory. Mott and Nabarro1,2 and Schoeck and Seeger3 both predict a linear increase in strength with concentration, while Fleischer4 indicates that the strength should be proportional to C1/2; there are in the literature experimental observations to support both these compositional dependencies.1&apos;8 The experimental evidence used to support the theories is the change in macroscopic yield stress (strain, e, = 10-3) with composition. However, recent investigations have revealed that solid-solution strengthening in a Cu-Al single crystals7 and in Fe-V polycrystal-line" alloys depends upon the strain at which the flow stress is measured; a flow stress in the microplastic range (e = 10-6 to 10-5) is much less concentration-dependent than the macroyield stress. Since the theories of solid-solution strengthening are concerned in particular with the stress to move a dislocation in a dislocation-free matrix, they should only be compared with microyield-stress data; it will be shown that in a material deforming homogeneously the microyield stress (e = 10-6) is approximately the stress to move through the matrix those dislocations which have been nucleated by stress concentrations such as particles or grain boundary ledges.9-11 In contrast, the macro-yield stress is more complex in that it is affected by several phenomena: dislocation multiplication, dislocation interactions, and the ease of propagating slip across grain boundaries. For example, it is experimentally observed that when long-range order12&apos;13 or solute effects10 restrict cross alip the grain boundary

Jan 1, 1967

By P. L. Hendricks, J. W. Spretnak

The interstitial atom principally responsible for the yield point and strain aging in electron-beam-melted tantalum is identified by analysis of the kinetics of the return of the yield point after an increment of plastic deformation. Two sets of specimens contained two levels of oxygen with very low hydrogen contents and the third set had comparable oxygen and hydrogen contents. The activation energy for the return of the yield point agrees well with that for diffusion of oxygen for the first two sets of specimens. For the third set of specimens, the activation-energy value lies between those for diffusion of hydrogen and for diffusion of oxygen. The advent of the dislocation model of plastic deformation in metals has revitalized interest in the yield point and strain aging in bcc metals containing a certain minimum content of interstitial solute elements. Much theoretical and experimental work has been performed in recent years to elucidate the detailed mechanism of these phenomena. The purpose of this investigation is to attempt to identify the principal interstitial element responsible for the yield-point phenomenon in electron-beam-melted tantalum by analysis of the kinetics of the return of the yield point after an increment of plastic deformation. Some of the earlier theories of the yield-point phenomenon proposed a grain boundary film of iron carbide. Such models could not satisfactorily explain all features of strain aging and the yield-point phenomena. The most widely accepted explanation is that of Cottrell,1 later extended by Cottrell and Bilby.2 Strain aging is ascribed to "locking" of dislocations by interstitial solute "atmospheres". The yield-point phenomenon results when the dislocations are torn away from their atmospheres. The strain-aged condition is re-established after sufficient time to allow the interstitial atoms to diffuse to the dislocation lines and re-establish the locking atmospheres. Clearly, the Cottrell-Bilby model is concerned with the bulk of the grain and does not specifically involve the grain boundary. The recent modification of the Cottrell-Bilby model is a redirection of attention to the role of the grain boundary and the possibility of multiplication of a limited number of free dislocations rather than unlocking all of the dislocations. Theories have been advanced by Hahn3 and conrad4 which are modifications of the Cottrell-Bilby theory. The model proposed by Hahn indicates that, although the possibility of un- locking anchored dislocations is not excluded, it implies that unlocking is not necessarily required to explain yield drop. Locking of dislocations during the aging treatment is a necessary part of the theory; however, it assumes that dislocations once locked remain locked. It is suggested that the yield drop observed is a result of the following factors: 1) the presence of a small number of mobile dislocations initially, 2) rapid dislocation multiplication, 3) the stress dependence of dislocation velocity. In the case of bcc metals, locking is considered to be the means by which dislocations are immobilized. Cold working of the metal results in the generation of larger numbers of new dislocations and the stress dependence of the dislocation velocity accounts for yield drop observed. Conrad&apos; has proposed a model similar to the one just described which applies to strain aging of iron and steel and which logically could be extended to other bcc metals. This model also does not require large-scale unlocking of dislocations. It is proposed that, during initial loading of a specimen below the upper yield stress, a few dislocations are torn free of their Cottrell atmospheres at regions of stress concentrations. With an increasing stress, some multiplication of dislocations occurs by a double cross slip mechanism, thus giving a preyield microstrain. At some critical stress represented by the upper yield stress, sudden profuse multiplication of dislocations occurs, enabling plastic flow to proceed at a lower stress. In this model, microstrain, preyielding, and flow represent the movement of free dislocations. It should be noted that this model also requires the locking of dislocations by interstitial solute atoms for the occurrence of a yield point; however, unlocking of large numbers of dislocations is not required. If it is assumed that the yield point will return when some fraction of free dislocations produced during pre-straining are pinned, the number of solute atoms required to pin unit length of a dislocation line can be calculated when prestrain and reloading are done at the same temperature and strain rate. Since the migration of solute atoms back to the stress fields of dislocation lines is controlled by the diffusion rate of interstitial solute atoms, it is to be expected that the activation energy for strain aging would be identical to the activation energy for diffusion. It would also be expected that the strain aging observed will be controlled by the fastest diffusing species capable of producing locking over the temperature range investigated. The rate of yield-point return has been found to be adequately expressed by an empirical rate equation of the form: rate=Ae-Q/RT [1] where A = constant, Q = activation energy, R = gas constant, and T = absolute temperature. Cottrell and Bilby2 have expressed the number of atoms per unit length of dislocation line which arrive

Jan 1, 1967

By L. Seigle, R. Resnick

An inert-marker movement experiment indicates that the ratio of the intrinsic diffusion coefficients DC:DTa = 80:l in TaC at 2500°C. Measurements of the diffmion coefficient of carbon in nonstoichiometric TaC at temperatures from 1700° to 2700°C reveal that Dc increases with decreasing carbon content, but much less than expected from the probable change in vacancy concentration with carbon content. A diffusion process involving two simultaneously operating mechanisms is postulated, and shown to be theoretically feasible. The average value of the carbon diffusion coefficient is given by DC = 0.18 exp[(-85,000 ± 3000)/RT] sq cm per sec over the composition range 46 to 49.5 at. pct C. BECAUSE of their high melting points and hardness, the carbides of the IV, V, and VI group transition metals, along with those of uranium, have attracted considerable interest for applications at high temperatures. In these applications the reactivity of the materials is important, and, since rates of diffusion within the compounds influence reactivity, a knowledge of diffusion kinetics and mechanisms is desirable. While many investigations of the mechanical and electrical properties of these compounds have been made, only two fundamental investigations of diffusion in the carbides are known. Chubb, Getz, and Townleyl measured the diffusivity of carbon and uranium in UC, and Gel&apos;d and Liubimov2 measured the diffusivity of carbon and niobium in NbC. This paper describes an investigation of the diffusion of carbon in tantalum monocarbide and, in particular, the influence of carbon deficiency on this process. Tantalum carbide melts at approximately 3800°C, which makes it one of the highest melting materials known. The compound exists over a rather wide range of carbon Content.3-7 At the peritectic temperature, 3240°C, the phase extends from about 36 to 50 at. pct C. Although the compound can exist with a substantial carbon deficiency, the high carbon phase boundary remains near the stoichiometric composition over the entire temperature range; i.e., no carbon excess is observed. The structure of TaC is the NaCl type wherein carbon atoms normally occupy the octahedral sites in a somewhat expanded fcc lattice of tantalum. Decrease of the lattice parameter with decreasing carbon suggests that the removal of carbon introduces octahedral vacancies into the lattice. I) EXPERIMENTAL DETAILS AND RESULTS Inert-Marker Experiments. In a compound such as TaC the interstitial element would be expected to diffuse more rapidly than the metal. This was confirmed by an inert-marker experiment, following Srnigelskas and irkeendall.8 Ideally, the markers should be placed at the interface between a slab of low-carbon TaC and graphite, and their movement during subsequent inter-diffusion measured. Unfortunately, no solid could be found which is unreactive in contact with carbon at the high temperatures employed in these experiments. In order to circumvent this problem, a specimen was designed in which the markers consisted of several small canals running just below the surface of a tantalum slab. This specimen was prepared by machining grooves on the surface of the tantalum slab and then diffusion-bonding a thin plate of tantalum to the slab over the grooves. The surface of the plate was then ground down until the distance between the canals and surface was as small as possible (about 0.01 cm). Thus, the canals would lie entirely within the TaC phase after a short period of diffusion. The diffusion anneal consisted of immersing the metal sample in high-purity graphite powder and heating for approximately 10 hr at 2500°C under vacuum. At this temperature, the vapor pressure is sufficiently high and the transfer of carbon from graphite sufficiently rapid to allow the surface of the diffusion sample to attain the stoichiometric carbon concentration very quickly. Conclusions regarding the relative diffusion rates of carbon and tantalum in the compound layers (TaC and Ta2C) can be drawn from the location of the canals after the diffusion anneals. If the growth of the layers is governed mainly by the diffusion of carbon, as expected, the canals should remain close to the sample surface. If the diffusion of tantalum contributes appreciably to formation of the compound layers, the distance from the markers (canals) to the surface should increase. Fig. 1 shows, diagrammatically, the appearance of the specimens after diffusion, and Table I presents the depth below the surface at which the

Jan 1, 1967