Search Documents

By R. E. McNulty, R. D. Heidenreich, C. H. Gerould

TIIE electron microscope techniques and their application to magnesium alloys that are to be discussed in this paper are the result of research at The Dow Chemical Co. over the past three years. The view- point underlying the work is not wholly metallurgical, but has evolved through a variety of problems related to the physics and chemistry of surfaces. It can be said that the interest and stimulus were closely associated with a desire to employ effectively the electron microscope in research concerning the properties of solids as related to their physical structure. Only a portion of this work will be described here, and, consequently, many points of interest and debate are of necessity either only briefly recognized or are completely omitted. As has been pointed out by Desch,1 a great deal of the knowledge of the structure of metals has been obtained through microscopic studies of the surfaces of metallic specimens prepared in a particular manner. This information, when added to that obtained by means of thermal, electrical, X-ray, dilatometric and other methods, has been of great value both from a theoretical and a practical view- point. With the introduction of new instruments such as the electron diffraction camera and the electron microscope, it would be expected that still more information could be obtained. These two instruments have in many instances been very effective when directed toward the investigation of metallic surfaces, provided the proper methods of surface preparation were employed. Experience shows that the problems of surface preparation greatly limit the general application of these instruments to metallographic studies. It was found necessary to discard the standard metallographic procedures in applying the electron microscope to magnesium alloys and to work out new methods, which were required to satisfy certain conditions. These methods are unique to magnesium only in the specific chemical reactions occurring during etching. The presentation of the methods and results will be ordered as follows: I. Electron Microscopy of Surfaces. 2. Surface Preparation of Magnesium Alloys. 3. Discussion of Microstructures. a. Pure Magnesium and Single-crystal Studies. b. Types of Precipitation. c. Some Common Structures. d. Special Micrographs and Techniques. e. Structure of Mg-A1 Solid Solutions. 4. Correlation of "Fine Structure" with Corrosion Behavior of Magnesium Alloys.

Jan 1, 1946

By Ashok K. Sinha

During the course of an in~*estigation into the occurrence of ordered AB3 structures, the following new phases have been found —CrRh3 (AuCu3 type), CrCo3 (MgCd3 type), HfCo4 (Ths Mn23 type), and WPt, MoPh type). The composition of the TiPt3-x phase (TiNi, type) is close to Ti23Pl77. The alloy chenzistry of transition rnetal AB3 structures is rezliewed in the light of electron concentration correlations of hex-agonality recently obtained for quasi-binary alloys. The relatizte colurne contraction in the AB3 structures increases with increasing difference in volume of the conzponents. A family of ordered close-packed layered structures is formed by stacking identical layers of composition AB, in various sequences, such that the coordination is twelvefold throughout and there are no A-A contacts. Previous work' on quasi-binary AB3 alloys has led to the conclusion that the stacking sequence of the AB, structures changes with increasing radius ratio RA/RB from a purely cubic, through different mixtures of hexagonal and cubic stacking to a purely hexagonal stacking. However. for binary AB3 alloys, a correlation between the type of the crystal structure and the position of the components in the various volumns of the periodic table has been noted.2-5 It has been noted6 that this correlation appears to hold even though the radius ratio RA/RB may vary over a considerable range with the location of the components in the three long periods. Another study7" of several quasi-binary systems led to the conclusion that an increase in hexagonality of the stacking is associated with increase in the electron concentration e/a. as defined by the average per atom of the total number of electrons outside the inert gas shells. In apparent conflict with this conclusion, it is known that seven binary alloy structures isotypic with TiNi3 which is 50 pct hexagonal occur at a higher electron concentration (e/n = 8.5) than that (e/a = 8.25) for the 100 pct hexagonal MgCd3 type structure present in seven binary AB3 alloys. Table 111. In the present work, an investigation into the occurrence of binary AB3 structures in transition metal alloys was made, and a survey of binary AB3 structures is presented. EXPERIMENTAL The starting materials were pure metals of 99.9 wt pct purity. The alloys were arc-melted under partial pressure of argon and annealed in sealed silica capsules lined with molybdenum foil under argon at- mosphere. The total weight loss upon melting and subsequent annealing was always less than 1 pct and hence the alloys will be referred to by their intended (unanalyzed) compositions. Wherever the constitution permitted. the alloys were given a homogenizing treatment at 1200°C (3 days) prior to annealing. Unless otherwise stated all alloys were annealed at 900°C for 1 week and water-quenched. Sometimes the final annealing treatment was carried out on powders to accelerate the attainment of equilibrium. X-ray powder patterns were taken using a Guinier-de Wolff focusing camera (CuK, radiation) or an asymmetrical focusing camera (Co or CrK, radiation). For lattice parameter determination. internal silicon standards were employed. The intensity calculations were made using a Fortran IV program written by Jeitschko and parthe.9 RESULTS Twenty AB3 and three AB4 alloys were investigated. Table I lists the crystallographic data on some of the intermediate phases encountered in the present work. Table II contains the X-ray data for HfCo, (Th,,Mn,, type). The positional parameter, x. was assumed to be 0.378. the value for Th6Mnn2310 The X-ray pattern of ZrCo, was very similar to that of HfCo, and the previous structure determination of ZrCo, by Kuzma el al." was confirmed. Ordering in the alloy CrCo could be ascertained by the presence of only one weak super lattice line (101). the others being too weak presumably owing to the small difference in the scattering powers of chromium and cobalt. This line was observed in the X-ray pattern of powder from the massive sample annealed at 830°C (7 days) after the powder had been reannealed at 600°C (24 hr). The diffraction pattern of the powder similarly reannealed at 830°C (24 hr) contained only the lines due to a mixture of hcp and fcc Co(Crj solid solutions. Therefore, it appears reasonable to assume that O2 and/or N2 contamination which would be less likely to occur during the 600°C anneal was not responsible for the observed weak reflection. Also. this reflection cannot be identified with any of the strong lines of the neighboring s phase which is present in the Co-Cr system at higher chromium contents. The composition corresponding to the TiNi3 structure observed by Raman et al.12 in the two-phase alloy Ti,zt,, has been established in the present work as being between There was satisfactory agreement for the low-angle lines (up to d = 1.997A) between the observed diffraction pattern of TiCua and that calculated assuming the ZrAu, structure. as recently proposed by Pfeifer-et a1.I3 However. some of the superlattice lines. e.g., at d = 1.937 and 1.919A. predicted by the ZrAu, structure were not actually observed eve? though neighboring lines. at d = 1.947 and 1.986A. of comparable calculated intensity were present. The ZrAu

Jan 1, 1970

By E. S. Machlin

The possibility that formation of voids under creep-rupture conditions may take place by the condensation of vacancies has been investigated theoretically. It has been concluded that nucleation of voids under creep-rupture conditions by vacancy condensation is highly improbable. However, growth of pre-existant voids by vacancy condensation is probable. A number of predictions made in this theory have been verified by the data. It has been predicted and checked that the product of rupture life and steady-state creep rate for preannealed metals and single phase alloys is an approximately invariant quantity, independent of stress, temperature, and atomic number for a given type structure. The direction of the effect of cold work on this product has been predicted and found in agreement with experiment. A number of experiments to evaluate the vacancy condensation mechanism further are described. SEVERAL papers have appeared recently which speculate on the origin of voids formed at grain boundaries under stress.' ' The object of this paper is to examine quantitatively the proposition that the voids produced in a creep test are a result of vacancy condensation. A result of this paper is a theory of creep-rupture. Void Nucleation Application of standard nucleation theory" to the problem of void nucleation leads to the following conclusions: 1—Homogeneous nucleation of voids requires a supersaturation ratio (concentration of vacancies in supersaturated to that in saturated solution) of 400 for a reasonable surface energy of 1000 erg per cm-and 1.4 for the improbably low surface energy of 10 erg per cm. 2—Heterogeneous nucleation of voids at plane interfaces between two phases requires a supersaturation ratio of 2.5 for a typical contact angle of 145 3-—Void nucleation about a solid particle may be accomplished at a supersaturation ratio of 1.17 for a typical value of work of adhesion? of 60 erg per The work of adhesion is the surface work 10 replace two solid-vauor surfaces by a solid-solid interface. enr ' between an oxide and a metal in the presence of a surface active element such as sulphur. Estimates of the supersaturation ratio at which voids are produced in diffusion experiments yield a maximum of 1.01. Inasmuch as the foregoing mechanisms of void nucleation probably will not operate at this level—too low a surface energy is required—the investigatol. is led to the conclusion that voids must already exist. That is, nucleation of voids probably does not occur. Rather, existing submicroscopic voids grow out to visible size. Already existing voids might be produced during solidification or working. Supercritical sized parlicles which contain cracks may act as heterogeneous void nuclei. Gas pockets may act as void nuclei. Experiments are desired to determine the nature of the heterogeneous void nuclei which grow out to voids in both diffusion and creep experiments. Void Growth Void growth might occur in at least two possible ways, depending upon whether the already existing void nuclei are at grain boundaries or within the grains. In the case of a spherical void far from a crystal boundary, vacancies are generated during creep as a consequence of the migration of suitable dislocation jogs' and are also annihilated at sinks. Under these conditions, a steady-state concentration of vacancies is built up in the crystal, defined by the condition that for any differential volume the rate of generation of vacancies in that volume equals the rate of annihilation of those vacancies." This equality would lead to the development of a gradient of vacancy concentration radially outward from the void surface up to a radius where the vacancy lifetime becomes equal for all directions of vacancy migration. The distance over which this vacancy concentration gradient extends equals about 2vD,T* where D, is the vacancy diffusivity and T:' the vacancy lifetime in a crystal outside the gradient in a zone of constant vacancy concentration. The vacancies generated in the region over which the gradient exists will annihilate more often at the void than elsewhere. Approximately a little over one-half the vacancies generated in the gradient zone will annihilate at the void. Hence, the growth rate of the void is given by on where R is the radius of void in centimeters, is the atomic volume, and R is the rate of generation of vacancies, number per centimeter" per second. R D and T* may be estimated in terms of other physical parameters." In particular, R = n.j e/b [3] where n is the average number of vacancy produc-

Jan 1, 1957

By S. Nakajima, H. Okazaki

Quantitatiue studies of the deformation texture in drawn tungsten wives were made by the X-vay dif-fractottletetr. Experimental results show that the diffraction Intensities are equal to tilose pvedicted from the (1 10). fiber lexlure but the angxla), spreads of. diffraction peaks in the pole distribution curres are different for different diffraction planes and directions. For this reason a modified (110) fiber lextuve model, in which a kind of anisotropy is assumed, is proposed to explain the results. According to this model the poles lying on a line directing front the (110) to the (110) poles in the (1 10) standard stereograpllic projection should show spreads which are different from those lyitlg on a line directing from the (001) to the (001) poles, which is confirmed by the experiments. The anisolvopy and the spveads of the pole positions are large at the outer part of the wires and decrease gradually lowards the inside of the wire. The possibilily of occurrence of such anisolropy in irrelals with fcc stvuctures is discltssed. THE deformation texture of drawn tungsten wires has been assumed by different investigators to be the simple ( 110) fiber texture.' Recently, however, Leber2,3 has shown that a swaged tungsten rod has a cylindrical texture. It changes gradually to the (110) fiber texture by drawing through dies. However, even after drawing to 0.25 mm in diam, the cylindrical texture can still be found in wires together with the (110) fiber texture. This was deduced from the pole figures obtained from the longitudinal section of these wires. Use was made also of quantitative measurements of the pole distribution curves. Leber stated that the angular spread of the pole distribution curves (henceforward called dispersions) are quite different for (400) 45 deg and (400) 90 deg: the former is always larger than the latter. This inequality is accompanied by deviations of the diffraction intensities from the theoretical values for the ( 110) fiber texture. Bhandary and cullity4 have reported similar results on iron wire and explained them by assuming a cylindrical texture. Both Leber3 and Bhandary4 used only the results of the (400) reflection for the determination of the dispersion. The pole figures found by Leber3 and by Rieck5 are largely different. The model given by Leber to explain the effects is in the authors' opinion in some respects unsatisfactory, especially if one looks at other than the (400) reflections. Intensities and dispersions of diffraction peaks are conclusive factors for the determination of the fine structure in wire textures. For this reason we studied them extensively to come to a model which is more suitable to fit the facts. In the following, after giving the experimental set-up, we report about measurements of X-ray diffraction on drawn tungsten wires. Different models to describe the experimental results will be discussed. EXPERIMENTAL GO-SiO2-A12O3 doped tungsten wires drawn to 0.18 mm in diam were used for the measurements. The wires were chemically etched to various diameters down to 0.03 mm. Measurements were carried out for the different wires in order to determine the dependence of the texture on the radius. The wires were cut to pieces of 10 mm length and fixed with paste closely against each other on a flat, polished glass plate. Parallelism of the wires with the surface of the glass plate should be adequate. For the diffraction studies three different X-ray sources were applied, respectively, giving the CuK,, FeK,, and FeKp emission. The measurements were carried out with a diffrac-tometer with a GM counter. The latter was fixed to a certain diffraction angle 20hkl and the diffraction intensity was recorded as a function of the angle of rotation of the specimen around the axis, lying in the specimen surface and perpendicular to the wire axis, as shown in Fig. 1. Measurements were also done with the detector at angles slightly deviating from the diffraction maxima The measured intensities in this case were taken to be equal to the background level. The deviations were chosen as small as possible but large enough to eliminate the influence of the diffraction maxima. The useful range of the rotation angle x of the specimen is generally limited by the wavelength of the X-rays. We have: where and cp is the angle between the wire axis and the normal of the diffraction plane. Intensity measurements were made to find the necessary corrections for counting loss of the GM counter and for distortion resulting from such effects as absorption of X-rays and from inclination of the reflection plane under study with respect to the surface of the specimen. The counting loss was esti-

Jan 1, 1968

By A. L. Bailey, B. A. Landry

W. W. ANDERSON and G. E. KELLER*—We want to compliment the authors on this very thorough paper. It gives information which the coal industry has needed for some time. We hope that the additional information which the authors are collecting will he available shortly. The mixing and riffling procedure that was followed for experimental purposes is obviously not practical in routine float-and-sink testing because of the particle size degradation which would result in handling the sample so many times. It is important to obtain our tloat-and-sink fractions with a minimum amount of handling of material. A statement is made in the paper (p. 80) that "the variable most likely to affect the size of sample required to meet a given preassigned accuracy would be the state or degree of mixing of the coal." We agree that this is a large factor, but do not believe it is the most important factor. Our own opinion is that the most important single factor governing the total gross weight of sample that must be collected is the percentage of the weight of material in the smallest fraction that results from the screening and float-and-\ink operations. In other words, size of sample is governed by the total number of fractionations that must he made, and the distribution of material within the fractions. We can imagine a coal with perfect mixing, but with such a small amount of material in some float-and-sink fraction in one of the coarse sizes that a much larger sample would have to be taken than would be the case with very poorly mixed material, but with a large percentage of coarse material more evenly distributed in all float-and-sink fractions. Our own observation of many float-and-sink tests that we have run in our own organization on many types of coal is that the size of sample that must be used on fine size float and sink is governed more by the requirements for weight of material to be used for analysis in the laboratory than by weight of material necessary to obtain accurate float and sink percentage of weight values. In other words, it is our opinion that very small samples can be used for float-and-sink fractionation in the fine sizes, but that accurate analysis of the fractions will depend on a larger weight of sample being pulverized for the laboratory than is necessary to establish the float-and-sink distribution with respect to weight. A. L. BAILEY and B. A. LANDRY (authors' reply)—The authors thank Messrs. Anderson and Keller for their comments based on long experience. It is agreed that the involved mixing and riming technique used may be disadvantageous from the standpoint of degradation. Fortunately, the paper does point out that the extended riming was unrewarding in causing further mixing. Two large unknowns remain, however: (1) how much of the mixing from the presumed highly unmixed state in the bed was achieved toward the random state during blasting, loading, transportation, screening, and further transportation to the point where the gross sample was taken, and (2) how much of the mixing took place during the preparation described preceding riming. As has been pointed out by one of the authors.6 the degree of mixing has a very large effect on the size of sample required and there are still too few experimental data to show at what stage of coal handling most of the mixing occurs. The discussion states that the weight of material in a screened fraction, or in a float-and-sink fraction, is more important than the mixing factor. We do not believe that these factors are comparable in this instance inasmuch as our purpose was to give minimum sampling requirements to achieve a preassigned accuracy in the percentages of float, middlings, or sink, and nothing more. The gross sample had already been screened and no further division by screening was made or contemplated; also, it was not intended that the middlings and sink fractions would necessarily be adequate for percentage ash or other determination. In other words, the sample obtained by the method outlined is not intended for washability studies but only for preparation plant control. Further experimental work has been done, since the paper was prepared, to investigate the effect of increasingly larger top and bottom sizes on the variability of float, etc., of a double-screened coal from Western Pennsylvania. Results will be published and eventually attention is to be given to the preparation of sampling specifications. E. H. M. BADGER*—I should like the authors to explain more fully the fundamental assumptions on which their Eq 4 is based. The equation is of the form s2 = p(l - p) which is the usual expression for the (standard deviation)2 when the chance of finding a particular kind of particle in the sample is proportional to the number fraetion, p. But instead of the number fraction, the authors have used the weight fraction, WF/W. The chance of finding a particular kind of particle in the sample can only be proportional to the weight fraction, if the average ?eig?ts of all kinds of particles, that is, float, midlings, or sink, are the same. Surely a much more justifiable assumption would be that the average volumes of the particles are the same, and, if this is so, Eq 4 would not be true. This may be demonstrated as follows: Let be the weight fraction of float, middlings, or sink, dl the density of this fraction, and d2 the density of the rest of the coal. Then assuming that the average volumes of the pieces in the three classes are the same, the number fraction, p, is given by ? P = d1/l-?/d2 + ?/d1 = ?d2/d1 + ?(d2-d1) The weight fraction, w, in terms of p is given by ? = pd1/(l-p)d2 + pd1 = pd1/d2 + p(d1-d2) _____ [61

Jan 1, 1950

By Earl Ingerson

BORE surfaces of used guns commonly show a pattern of cracks in various degrees of development. It has been suggested that these cracks may aid erosion by providing channelways for the gases, eventually leading to the tearing out of blocks of metal. Under the general studies of gun erosion conducted at the Geophysical Laboratory it appeared desirable to consider this factor, and to determine the resistance to surface cracking offered by gun steel and by various other materials that might be used in the bores of guns. METHOD AND APPARATUS In order to make such a determination it is necessary to have (I) an arrangement in which the surface to be studied can be subjected to conditions similar to those in a gun with respect to temperature, pressure (developed by explosion gases), composition of the gases and length of time the surface is exposed to the gases; and (2) a specimen that can be examined readily after any number of shots, and then fired again, if necessary. The first of these conditions can be met by using an explosion vessel designed by H. S. Roberts of the Geophysical Laboratory. Fig I shows a cross section of the apparatus. The "gun" G from which the charge is fired is made from the receiver of a caliber 0.30 army rifle, M1903, and the charges are prepared in caliber 0.30 cartridge cases. The test specimen S is held in place by a cone-shaped plug P and by a steel ,bridge B which is bolted to the outside of the explosion vessel on either side of the section shown. The plug, in turn, is held in place by a screw collar C. Maximum pressure is controlled by a brass rupture disk R. Part of the explosion gases flow past the two flat areas on S. st the two flat areas on S. Temperature and composition of gases, and the time before the rupture disk breaks, can be controlled by amount and kind of powder used, and pressure can be controlled by using rupture disks of different thickness. While the explosion vessel used was not designed so that temperature and pressure could be measured, in other vessels of closely similar size and shape containing proportionate charges of the same kind of powder as that used for standard testing procedure, values of 2700°K and 5000 psi were obtained. Since the orifices past the flat areas on the plug are small compared to the rupture disk hole (1:16), most of the gas goes out through the rupture disk when it breaks. The amount that goes by the plug depends on (I) the pressure attained before the rupture disk breaks and (2) the time it takes for this pressure to be built UP. The second condition (examination at will after any firing) is met by using for specimens short rods that have flat areas ground on opposite sides. These flat

Jan 1, 1947

By A. L. Benham, W. E. Dowden, W. J. Kunzman

The mechanism involved in creating a miscible fluid displacement of typical reservoir fluids through porous media by light hydrocarbon mixtures usually involves a number of mixing and separation stages in the contact area. This mechanism has been illustrated on ternary diagrams for miscible displacement of reservoir fluid by rich hydrocarbon gases. A relatively simple method has been devised for calculating the approximate conditions for a miscible displacement of reservoir fluid by rich, light hydrocarbon gases or by LPG mixtures. Miscibility is favored by increased pressures, decreased temperatures, light reservoir fluids and light hydrocarbon-displacing fluids rich in the heavier components. The conditions for miscible displacements of five typical reservoir fluids by various light hydrocarbon mixtures were determined for a range of temperatures from 70" to 260°F and pressures from 1,000 to 3,000 psia. A correlation of these conditions was obtained using the variables of temperature, pressure, C,+ molecular weight of the reservoir fluid, C,+ molecular weight of the displacing fluid, and mol per cent methane in the displacing fluid. This correlation has been applied to data appearing in the literature with very good success. A partial check of the correlation has been made using a step-wise procedure with a windowed PVT cell. INTRODUCTION The term, "miscible fluid displacement", may be defined as any oil-recovery displacement process where there is an absence of a phase boundary or interface between the displaced and displacing fluids. A water flood would not be a miscible fluid displacement, although cycling in a condensate reservoir would be. Another example is displacement of oil by gasoline. Not so obvious, however, are the conditions under which fluids (such as propane, mixtures of propane and methane, or other similar systems) will give a miscible fluid displacement of oil. There are four standard miscible-fluid-displacement processes being used for the recovery of oil from an "nderground reservoir—(1) high-pressure dry-gas miscible displacement, (2) rich-gas miscible displacement, (3) miscible slug using a slug of LPG followed by a lean dry gas under miscible conditions, and (4) gas driven-slug process using a slug of LPG followed by a lean dry gas under immiscible conditions. The first problem faced by the engineer designing a miscible displacement process is to determine the conditions for miscibility. A number of papers have included discussions of the process of obtaining miscibility in these processes.1-8 However, methods for calculating the conditions for miscibility have not been published. This paper includes a method for calculating the conditions for a miscible displacement of reservoir fluid by a rich gas (LPG diluted with methane or natural gas). In this paper, a rich-gas miscible displacement process is defined as an oil-recovery process in the reservoir. The cost of a rich-gas miscible displacement will be decreased by diluting the LPG with methane or natural gas as much as possible. Therefore, it is important to be able to predict the maximum dilution for miscibility. MECHANISMS The mechanisms for obtaining miscibility between a displacing mixture of light hydrocarbons and displaced reservoir fluid often involve a number of displacing and mixing stages. As stated previously, the cost of a rich-gas miscible displacement is decreased by diluting available LPG with natural gas. The problem is to predict the maximum dilution allowed while maintaining miscibility. To get a picture of the displacing and mixing stages, it is helpful to consider these complex hydrocarbon mixtures as pseu do-ternary systems. Graphical representation of a complex system as a pseu do-ternary system can be thermodynamically rigorous if enough variables are used to describe the character of each pseudo-component in one of the phases. However, even if the characterization of each pseudo-component is not complete, pseudo-ternary representation does serve to illustrate the mechanism for obtaining miscibility and can be used to make some general conclusions concerning the conditions for obtaining miscibility. A phase diagram for the ternary system, methane-

By J. D. Defilippi, E. M. Gilbert, K. G. Brickner

FCC chromium-manganese-nitrogen (Cr-Mn-N) steels differ from most other fcc materials in that these steels undergo a ductile-to-brittle transition. Transformation to martensite is considered to be responsible for this behavior in some metastable Cr-Mn-N steels. However, very stable Cr-Mn-N steels also exhibit a ductile-to-brittle transition. The results of this study indicate that deformation faulting is the probable cause of the brittle behavior of stable Cr-Mn-N steels. Deformation faulting accounts for the ductile behavior of these steels in a tension test at -320°F and brittle behavior in an impact test at -320°F. Deformation faulting also accounts for the toPological features observed on the fracture surfaces of impact specimens of these steels. FACE- centered- cubic chromium-manganese-nitrogen (Cr-Mn-N) steels differ from most other fcc materials in that these steels undergo a ductile-to-brittle transition. Many Cr-Mn-N steels transform to martensite during deformation,l-5 and several investigatorsl-3 have suggested that the brittle behavior of these steels is caused by martensite formation. However, very stable Cr-Mn-N steels also exhibit brittle behavior. Schaller and Zackeyl reported that a very stable Cr-Mn-N steel (less than 3 pct martensite formed at -320°F) exhibited a transition temperature higher than that for steels in which large volume fractions of martensite formed during testing. The explanation given by Schaller and Zackey for this observation was that in the very stable steel the martensite, because of its higher interstitial content, was more brittle than that formed in their other steels. This explanation was questioned by Tisinai and samans4 and Baldwin.6 Moreover, because the toughness of stainless martensite at cryogenic temperatures is generally very low, this explanation does not account for Thompson's7 observation that small additions of nickel (1 to 3 pct) greatly improve the toughness of high nitrogen (0.35 pct) Cr-Mn-N steels. The present paper summarizes the results of an investigation of the low-temperature brittleness in very stable Cr-Mn-N steels. The importance of the mode of deformation on the toughness of these steels is discussed. Table I. Compositions of the Steels Invertigated, Pet Steel C Mn P S Si Ni Cr N - A 0.09 14.70 0.018 0.011 0.47 0.22 18.40 0.54 B 0.12 14.90 0.001 0.008 0.48 0.14 17.80 0.38 C 0.12 14.95 0.004 0.005 0.62 3.95 18.43 0.38 MATERIALS AND EXPERIMENTAL WORK The compositions of the steels investigated are shown in Table I. Steels A and B had compositions within the limits of a proprietary Cr-Mn-N stainless steel,* whereas Steel C was similar in composition to the proprietary steel except for its 3.95 pct Ni content. All steels were hot-rolled to 1/2-in. thick plate. The plates were subsequently annealed for 1 hr at 2000°F and water-quenched. Standard longitudinal and transverse Charpy V-notch impact specimens were machined from the annealed plates. Duplicate longitudinal and transverse impact specimens were tested at 212", 80°, 32", 0°, -100°,-160°,-200°,-256", and -320°F. Longitudinal tension-test specimens were also machined from the plates and tested at a crosshead speed of 0.05 in. per min at the aforementioned temperatures. The fractured impact and tension-test specimens of all three steels were examined to determine whether martensite had formed during testing. Magnetic, X-ray, electron-diffraction, and electron-microscopy techniques were used to detect the presence of martensite in the highly deformed areas of these specimens. Metallographic examination of highly deformed areas of impact and tension-test specimens revealed the presence of dark-etching bands, such as those shown in Fig. 1. These bands were observed only in deformed samples and were thought to be associated with the low-temperature brittleness of the Cr-Mn-N steels. Accordingly, a sample 1 in. wide by 3 in. long was cut from the 1/2-in.-thick plate of Steel C. This sample was surface-ground to a in. and then cold-rolled 60 pct at -320°F. Thin foils were prepared from the cold-rolled sample and examined in a JEM electron microscope. Brightfield, dark-field, and selected-area diffraction techniques were used to determine the cause of the dark-etching bands. Fractographic experiments were also performed. Impact specimens Of Steels A, B, and C were broken at -320oF, and the fracture surfaces of these specimens were immediately shadowed with carbon. The carbon replicas were examined in a Siemens electron microscope, and attempts were made to correlate the topological features of the fracture surfaces with the deformation mechanisms that could be occurring during an impact test of these steels.

Jan 1, 1970

By Herbert J. French

Herbert 1. French (Member 1934), vice president of the International Nickel Co. Inc., assistant vice president of the International Nickel Co. of Canada Ltd., and AIME Howe Memorial Lecturer for 1956, died in Rochester, Minn. on Aug. 17, 1955, after a long illness. Before his death, Mr. French had laid the foundation of research on which the Howe Memorial Lecture (p. 770) was based. His associates at Inco, 0. 0. Miller and J. W. Sands, were able to complete the project, present the lecture, and prepare it for publication, following Mr. French's general outline. Born in New York on May 27, 1893, he graduated from the School of Mines, Columbia University, with a degree in metallurgical engineering in 1915. Mr. French gained his early engineering experience with the American Smelting and Refining Co., serving as a chemist and as-sayer at the company's Murray, Utah, installations. He later returned east and became a metallurgical engineer for the General Vehicle Co., Long Island City, IT. Y., a subsidiary of the General Electric Co. In 1917 Mr. French was appointed civilian inspector of materials for the U.S. Army Signal Corps, and continued this work when it was later absorbed into the Bureau of Aircraft Production. Subsequently he was appointed district inspector in Philadelphia for the Pittsburgh (Ordnance Metallurgical) office, later transferring to Pittsburgh. After World War I Mr. French became an assistant physicist on the metallurgical staff, U. S. Bureau of Standards, Washington, D. C. During his years with the Bureau, he was successively associate physicist, metallurgist, senior metallurgist, and assistant chief, division of metallurgy. In 1929 Mr. French joined International Nickel as a metallurgist at the company's Bay-onne, N. J., research laboratory. Transferring to the development and research division, New York, in 1931, he took charge of alloy steel and iron development, becoming assistant manager of this division in 1943. Mr. French's appointment as an Inco vice president came in March 1947. Previously, he was appointed assistant vice president of the International Nickel Co. of Canada Ltd. in January 1947. Mr. French served the War Production Board during World War II as senior technical consultant in charge of the metallurgical and specifications section of the iron and steel branch. When this section of the WPB was organized into the steel division, he was appointed chief of the metallurgical branch. From March 1944 to September 1945 Mr. French served as assistant director. Herbert French was well known not only for his prolific contribution to the metallurgical literature, for which he received numerous awards, but also for his prolonged service to the profession through its technical societies. He was a past president of ASM, and was the Campbell Memorial Lecturer for that group in 1933. Mr. French was honored by the American Society for Steel Treating in 1931 when he was awarded the Henry Marion Howe Medal. A series of lectures he gave in 1941 before the Fourth Western Metal Congress in Los Angeles was later published in book form under the title Alloy Constructional Steels.

Jan 1, 1957

By Carl A Zapffe, George A. Moore

In a previous publication from this laboratory1 the conclusion was drawn that the embrittling effect of occluded hydrogen on iron and steel must result from the precipitation of the gas within small openings through the crystal structure. A review of the literature then published indicated the real existence of crystal substructures of the mosaic type and gave presumptive evidence that the "rift openings" associated with hydrogen occlusion are the same as the L'disjunctions" of the mosaic theory. Inasmuch as no clear conception of the actual nature of the crystal fragments and openings inherent in these theories has been established, it was considered desirable to obtain direct micrographic evidence of their true configuration. In this direction the standard methods of metallographic attack have proved useless. Recourse was had, therefore, to the direct examination of the surfaces upon which failure by cleavage had resulted from the presence of hydrogen. It was expected that such a surface, unaffected by chemical attack and undis-torted by polishing, would reveal the fine openings penetrated by the gas and thus indicate the real nature of the substructures. Materials Studied In commercial steels, defects caused by hydrogen take various external forms according to such factors as structure and purity, and to the extent to which hydrogen either occludes or produces chemical changes. The present work relates only to the well-known transcrystalline hydrogen embrittlement—-"pickling brittleness "—in which elemental hydrogen occluded within the grain develops cleavage characteristics. For example, hydrogen embrittlement resulting from hydrogen reaction products is excluded from detailed study. It was also desirable, as far as possible, to have the study uncomplicated by other elements which themselves cause brittleness. Three types of iron were used. These included ingot iron, decarburized free-machining steel, and a specially prepared, aluminum-killed, electrolytic iron. This choice allowed a comprehensive selection of unavoidable minor impurities and inclusions, which could subsequently be shown to have no specific effects on the structures within the grains. Carbon, and any other impurities free to react with hydrogen, were first removed by a preliminary annealing treatment in hydrogen. Subsequent treatment in vacuo extracted much of this hydrogen and its gaseous reaction products. The resultant ductile iron was then embrittled either by a second anneal in hydrogen, followed by rapid cooling, or by cathodic electrolysis. Group 1: Prepared from Armco Ingot Iron As shown in Table 1, the principal impurity in this iron was oxygen, which survived the purifying treatment to a large extent. Two series were prepared. The first (Nos. SIO-24) was purified only one day in hydrogen and dehydrogenized one day in nitrogen. Containing both hydrogen and its

Jan 1, 1943

By Carl A. Zapffe, George A. Moore

In a previous publication from this laboratory1 the conclusion was drawn that the embrittling effect of occluded hydrogen on iron and steel must result from the precipitation of the gas within small openings through the crystal structure. A review of the literature then published indicated the real existence of crystal substructures of the mosaic type and gave presumptive evidence that the "rift openings" associated with hydrogen occlusion are the same as the L'disjunctions" of the mosaic theory. Inasmuch as no clear conception of the actual nature of the crystal fragments and openings inherent in these theories has been established, it was considered desirable to obtain direct micrographic evidence of their true configuration. In this direction the standard methods of metallographic attack have proved useless. Recourse was had, therefore, to the direct examination of the surfaces upon which failure by cleavage had resulted from the presence of hydrogen. It was expected that such a surface, unaffected by chemical attack and undis-torted by polishing, would reveal the fine openings penetrated by the gas and thus indicate the real nature of the substructures. Materials Studied In commercial steels, defects caused by hydrogen take various external forms according to such factors as structure and purity, and to the extent to which hydrogen either occludes or produces chemical changes. The present work relates only to the well-known transcrystalline hydrogen embrittlement—-"pickling brittleness "—in which elemental hydrogen occluded within the grain develops cleavage characteristics. For example, hydrogen embrittlement resulting from hydrogen reaction products is excluded from detailed study. It was also desirable, as far as possible, to have the study uncomplicated by other elements which themselves cause brittleness. Three types of iron were used. These included ingot iron, decarburized free-machining steel, and a specially prepared, aluminum-killed, electrolytic iron. This choice allowed a comprehensive selection of unavoidable minor impurities and inclusions, which could subsequently be shown to have no specific effects on the structures within the grains. Carbon, and any other impurities free to react with hydrogen, were first removed by a preliminary annealing treatment in hydrogen. Subsequent treatment in vacuo extracted much of this hydrogen and its gaseous reaction products. The resultant ductile iron was then embrittled either by a second anneal in hydrogen, followed by rapid cooling, or by cathodic electrolysis. Group 1: Prepared from Armco Ingot Iron As shown in Table 1, the principal impurity in this iron was oxygen, which survived the purifying treatment to a large extent. Two series were prepared. The first (Nos. SIO-24) was purified only one day in hydrogen and dehydrogenized one day in nitrogen. Containing both hydrogen and its

Jan 1, 1943

[RECEIPTS Magazine, Advertising $ 32,823.60 Sale of Magazines 3,933.29 Total Magazine $ 36,756.89 Dues, Arrears 5,866.97 Dues Current 100,223.80 Dues of New Members 6,143.29 Dues in advance 1,586.32 Initiation fees 6,820.00 Sale of Transactions 5,464.99 Sale of Binding 5,618.90 Sale of Pyrometry Volume 336.30 Sale of Indexes and Emmons 106.10 Sale of Pamphlets 575.32 Miscellaneous Receipts 105.60 Interest on Bank Balances 1,217.67 Interest on Life Membership Investment 512.18 Interest on James Douglas Library Fund 5,678.67 Total Receipts $177,013.00] [ ]

Jan 1, 1925

By J. C. Melrose, W. R. Foster, J. G. Savins, E. R. Parish

Many differences can be imagined between gas-oil flow in which the gas is supplied at the face of the core and gas-oil flow in which the flowing gas was originally dissolved in the oil. If capillary pressure characteristics and flow requirements control gas saturation distribution, the gas would be expected to be located at preferred sites within the porous medium as determined by pore sizes. On the other hand, during solution gas drive the gas first appears as bubbles through a nuclea-tion process. Nothing in self-nucleation theory specifies at which sites the first bubbles should be formed. In all probability they will be randomly distributed throughout the porous medium. Furthermore, it is not at all certain that even at low rates of production the gas will redistribute itself after nucleation to the channels normally occupied by gas in simple gas flow. Stewart, et at, have shown that at least for some limestone samples, oil recoveries could not be predicted for all rates of production using any one set of relative gas and oil permeabilities. An important factor in controlling recoveries during solution gas drive was the rate of bubble formation, higher rates giving higher recoveries. Stewart, et al, attributed the increase in recovery to a better distribution of the gas phase in heterogeneous limestone samples than is obtained by simple external gas drive. Differences in recovery from these causes were not reported for sandstone cores. In the experiments to be reported here, oil recovery, pressure and producing GOR history were measured during solution gas drive for a 5-ft sandstone core. The results were compared with predictions from the Muskat method for computing solution gas-drive behavior using external gas-drive relative permeability. The effects of changing the rate of production and oil viscosity were studied. At high laboratory rates of average pressure decline, two observations were made which would not have been predicted by Muskat's depletion theory: (1) oil recovery increased with increasing rate of production for a given viscosity oil, and (2) oil recovery increased with increasing oil viscosity for a given high rate of production. Both of these observations are explained as consequences of diffusion control of gas saturations superimposed on the normal gas-oil flow requirements, Fur- thermore, discontinuous gas phase flow appears to be significant during solution gas drive. The laboratory tests were performed at rates of average pressure decline many times greater than the maximum possible rate of average pressure decline in an actual oil field. It is, therefore, not possible to draw any direct conclusions regarding the effect of rate on recovery for the solution gas-drive mechanism under actual field conditions. However, at the lower laboratory rates, recoveries were nearly independent of rate and could be predicted by the Muskat method, using external gas-drive relative permeability data. These results suggest that at normal oilfield rates the effect of rate on recovery for the solution gas-drive mechanism is negligible. EXPERIMENTAL PROCEDURES The core material for the pressure depletion studies was Bandera sandstone from an outcrop in the Mid-Continent. This sandstone was selected because of its low permeability (about 10 md), which would permit the development of substantial pressure gradients in the corn at moderate flow rates. The core was 5-ft long and 2-in. in diameter. Its properties are listed in Table 1. Relative gas-to-oil permeability ratios were measured by an external gas-drive method2. The results are shown for a short 2-in. core and for the 5-ft core in Fig. 8. Oils used in the pressure depletion experiments were kerosene and a highly refined white oil (standard white oil No. 3) with gas-free viscosities of 1.8 and 25 cp, respectively. The gas was a naturally occurring methane from Gough field, Inglewood, Calif. The oil viscosities, gas solubilities and formation volume factors are plotted as functions of pressure at 75°F in Figs. 1 and 2. Methane viscosities and compressibilities were obtained from the literature"'. A core mounting was required which could withstand up to 2,500 psi internal pressure. This was obtained by first encasing the core completely in a plastic resin (Scotch Cast, manufactured by Minnesota Mining & Manufacturing Co.) The plastic covered core was then inserted into a steel pipe equipped with screw caps so that the plastic coating could be pressured from

By Alfred Boyles

"HEREDITY"' in cast iron has been a subject of much discussion. Numerous experimenters have found that the properties of gray iron may vary greatly without corresponding variations in composition or thermal treatment. The main structural difference in such cases is in the size and distribution of the graphite flakes and the explanations given usually attribute this difference to some chemical element not shown by ordinary analysis or to some physical factor, such as inclusions in the melt. It cannot be said that any of these theories is completely satisfactory. Much remains to be learned about cast iron and for this reason a study of the material was undertaken as a part of the program of fundamental research at Battelle Memorial Institute. A review of the literature indicated that very little systematic work had been done on the mecha-nism of freezing in cast iron. The experiments described below are largely concerned with that problem.

Jan 1, 1937

By Arne Bergholm

Australian rutile was chlorinated in the presence of CO or carbon. The chlorination velocity in CO was found to be strongly influenced by temperature and proportional to the CO concentration, but independent of the Cl, concentration. In the presence of carbon, the reaction velocity is much higher. The reactivity of the carbon and the distance between the carbon and the rutile surfaces are important variables. The reaction velocity is approximately proportional to the Cl, concentration and independent of the CO concentration of the surrounding atmosphere. Experiments with fluidized-bed chlorination of carbon-rutile mixtures indicate that the motion of the bed has little influence on the reaction velocity. At low temperatures, the chlorination velocity of dense tablets is much greater than that of TiO, coke mixtures suitable for fluidization. The reaction mechanism is discussed. In the industrial production of TiCl, rutile is chlorinated in the presence of carbon. Disregarding intermediate steps, the reaction may be expressed by the following equations: The ultimate object of this study was to find out whether the reaction velocity was higher in fluidized bed operation compared with chlorination of pelle-tized carbon-rutile mixtures. From a literature survey and preliminary experiments it was learned that some basic knowledge about the reaction mechanism was needed for a good experimental design. Therefore the following sets of experiments were carried out: 1) Studies on the reaction velocity in the chlorination of rutile with CO as the only reducing agent. 2) Chlorination of separate rutile-carbon tablets. 3) Chlorination of rutile-carbon tablets at different temperatures with various kinds of carbon, various grain sizes, and various tablet-making techniques. This series of experiments was carried out not only with pure chlorine but also with mixtures of chlorine with argon, CO and CO,. 4) Chlorination of rutile-carbon tablets made in a strictly standardized manner. 5) Chlorination of static-bed rutile-carbon mixtures. 6) Fluidized-bed chlorination of the same mixture. The experiments 4 to 6 formed the final and direct test of the main question: Static bed vs fluidized bedo. Although there are many patents and papers dealing with the general aspects of chlorination, only few experiments from which detailed information can be obtained have been reported in the literature. Pamfilov and coworkersl3 have studied chlorination of TiO, with CO or carbon as the reducing agent. They found that the weight decrease per hour at 600o to 800°C amounted to 11 to 14 pct with CO and 45 to 51 pct with carbon. They suggest that phosgene might be an intermediate in the chlorination of TiO,. Takimoto and Hattori have chlorinated reduced titanium oxide (TiO) and found a very high rate of TiC1, -production. They proved that the composition of the gas from the chlorination of rutile-carbon mixtures contained mainly CO. They reported for instance 74.7 pct CO, and 5.6 pct CO at 800°C at which temperature the Boudouard equilibrium composition is 12 pct CO and 88 pct CO. Seligman and Segerchano6 have studied the chlorination of TiO. They proved that Ti0 and chlorine react rather rapidly at temperatures as low as 300°C. Above 400°C, the reaction was complete. At 500°C the velocity of this reaction was much higher than was chlorination of TiO, + C. McTaggart,7 Nishimura, et al. and Wilskam have chlorinated mixtures of carbon and various kinds of rutile or beneficated ilmenite. Nishimura, et al. also report the results from reduction of TiO, with H2, CO, or carbon. At 900°C only 1 pct Ti O is formed in 2 hrs. At 1100°C 23.1 pct Ti, 0, was formed if elementary carbon was present. No carbide formation occurred below 1400°C. McIntosh and Cofferll have observed that the CO, content of the exit gases from chlorination of rutile and calcined petroleum coke is appreciably higher than found in the Boudouard equilibrium. At 900°C the ratio (CO, /CO + CO) is about 80 pct, whereas the equilibrium value is about 2 pct. W. E. Dunn12 has studied the chlorination rates of several TiO,-bearing minerals with CO + Cl, or COCl . Chlorinations were carried out either in a fluidized reactor or a fixed-bed reactor, both having a 30-mm diameter. The results obtained in both reactors were comparable. It was proved that benefi-ciated ilmenite (i.e., ilmenite from which the iron oxide had been removed by chlorination) was chlorinated 10 times faster thanrutile. Sore1 slag showed an intermediate rate. When phosgene is used, the

Jan 1, 1962

By M. H. Waxman, L. J. M. Smits

A simple physical model was used to develop an equation that relates the electrical conductivity of a water-saturatedshaly sand to the water conductivity and the cation-exchange capacity per unit pore volume of the rock. This equation fits both the experimental data of Hill and Milburn and data obtained recently on selected shaly sands with a wide range of cation-exchange capacities. This model was extended to cases where both oil and water are present in the shaly sand. This results in an additional expression, relating the resistivity ratio to water saturation, water conductivity and cation-exchange capacity per unit pore volume. The effect of shale content on the resistivity index - water saturation function is demonstrated by several numerical examples. INTRODUCTION A principal aim of well logging is to provide quantitative information concerning porosity and oil saturation of the permeable formations penetrated by the borehole. For clean sands, the relationships between measured physical quantities and porosity or saturation are well known. However, the presence of clay minerals greatly complicates log interpretation, particularly the electrical resistivity and SP logs, and considerably affects evaluation of hydrocarbon-bearing formations. The conductance and electrochemical behavior of shaly sands and their relation to log interpretation have been studied by many workers. wylliel and Lunch2 reviewed this work in some detail. Virtually all laboratory measurements of electrical resistivity and electrochemical potential of shaly sands published to date are the work of Hill and Milburn.3 Their measurements were made on about 300 cores covering a large variety of sedimentary rocks, and a wide range of equilibrating NaCl solution concentrations. Hill and Milburn described their conductivity data by an empirical equation in which the shaly sand conductivity Co was expressed as a function of the solution conductivity C and two parameters b and F01 . The quantity b was shown to be a measure of the effective clay content of the rock, being approximately proportional to the cation-exchange capacity of the rock divided by its pore volume. The latter ratio is designated as 9, in this paper and has the dimensions meq/ml or equiv/liter. Qv is identical with the term representing the concentration of fixed charges in the Meyer-Sieved and Teorel15 theory of permselective membrane behavior. F01 is a formation resistivity factor referred to a hypothetical equilibrating solution resistivity of 0.01 ohm m at 25C,* where clay effects presumably are minimized. F01 was correlated to porosity by an Archie-type equation.6 The Hill-Milburn equation describes their data with a standard deviation of approximately 1 percent and a maximum deviation of + 10 percent. Shaly sands behave as permselective cation-exchange membranes, their electrochemical efficiencies increasing with increasing clay content. The electrochemical potential data of Hill and Milburn were expressed graphically, and demonstrate that the membrane efficiency (or cation transport number) of these sands is a function only of the b value (i.e., Qv) and the respective salt concentrations of the two solutions forming the liquid junction in the sand. The diffusion potentials are not dependent on F01 or any parameter relating to the porosity or pore geometry of the rock. The Hill-Milburn resistivity equation correctly predicts a decreasing sand conductivity Co with decreasing solution conductivity C?. However, at some low value of C,, the calculated Co-C, function goes through a minimum; with further decrease in C?, the predicted sand conductivity increases sharply. As pointed out by Hill and Milburn, an increasing value of Co with decreasing C, is physically meaningless. Since this occurs below the range of practical values of C,, the usefulness of the empirical equation is not affected when

Jan 1, 1969

Jan 1, 1967

By AIME

J. B. Haffner, or "Barney" as he is known to his wide circle of friends, was born at Landskrona, Sweden, on Christmas Day, 1886. He attended grade and preparatory schools there and then went to the Royal Technical Institute at Stockholm from which he graduated as a mining engineer in 1910. During these school years he gained practical experience in his summer vacations by working in various mines in Sweden and other European countries. Upon graduation he set forth for the land of opportunity, the United States. Upon arrival in this country he found people everywhere talking about the fabulous gold fields in Alaska and thereupon set sail for Nome, to take part in placer gold mining operations. Returning to the States, Barney worked in Illinois coal mines for a time and then carried on some construction work in Oklahoma. In 1913 he accepted employment with the Ray Consolidated Copper Co. in Arizona and soon found himself in positions of responsibility in the operating and

Jan 1, 1948

RECEIPTS [Magazine Advertising $ 62,564.83 Magazine Sales 3,591.44 Totalm 66,156.27 Dues, Arrears7,032.88 Dues. Current96,839.74 Dues, New Members6,525.00 Dues, in advance 1,907.02 Initiation Fees7,584.00. Sale of Transactions3,909.03 Sale of Binding7,845.05 Sale of Special Editions 338.20 Sale of Year Book76.20 Sale of Pamphlets1,189.06 Sale of Authors' Reprints867.54 Miscellaneous Receipts62.93 Interest on Bank Balances1,712.89 Interest on General Reserve Fund337.50 Interest on Life Membership Fund557.84 Interest on James Douglas Library Fund....5,491.89 Total Receipts$208,433.04] [ ]

Jan 1, 1925

By Robert T. Ault

The nature of fracture in unnotched tensile and notched tensile sheet and round specimens and V -notched and precracked Charpy-type sheet specimens of both wrought stress -relieved and re-crystallized molybdenum was investigated over the temperature range of -78° to 300°C. The sharp rise in fracture stress, for unnotched tensile samples, as the temperature is increased above the brittleness transition temperature (Tb), is found to be a result of an increase in flow stress due to plastic constraint and an increased strain rate which results from the onset of necking at Tb . Quasi-brittle fracture in notched tensile samples is found to occur when the tensile stress at the elastic-plastic interface, beneath the root of the notch, reaches a maximum critical value, which is independent of test temperature over the range from —78° to 25°C, but dependent on microstructure. In the temperature range between 150° and 300°C, unnotched tensile and notched tensile samples alike are found to fracture by a ductile fibrous tearing process which is discontinuous in nature, as a result of the competition between the processes of tearing, through continued plastic flow, and local work hardening. Results from the V-notched and fatigue-cracked Charpy-type impact tests demonstrate that crack initiation is the governing factor in the low-temperature (-78° to 25°C) fracture process for molybdenum. In recent years, there has been considerable interest, and a commensurate number of investigations, concerning the duc tile-to-brittle transition in refractory metals. There has not developed, however, a satisfactory explanation for the uniaxial tensile fracture-stress transition which accompanies the well-known ductility transition. This matter therefore warranted investigation. In a similar manner, the increased importance of notched tensile strength values in evaluating materials and in design criteria requires a better understanding of the factors which control the fracture behavior of notched tensile samples. In a previous investigation1 of the nature of initial yielding and fracture in notched sheet molybdenum at room temperature, it was suggested that plastic constraint was the controlling factor in governing the fracture behavior of notched samples. The final portion of this investigation was concerned with the fracture toughness of molybdenum. Of particular interest was the comparison of effective surface energies for fracture in V-notched samples with fatigue-cracked, Charpy-type samples in order to ascertain the relative importance of the initiation and propagation phases of the fracture process. MATERIALS AND TEST PROCEDURE Sheet tensile, notched tensile, and V-notched Charpy specimens were prepared from 50-mil, stress-relieved molybdenum sheet.* Half of these specimens were vacuum-annealed for 1 hr at 1200°C and 7.5 x l0-5 Torr and furnace-cooled to produce a relatively uniform grain size of 0.30 mm diameter. Round notched and unnotched tensile specimens were prepared from warm-rolled and swaged 5/8-in.-diam bar.* These specimens were vacuum-annealed for 1 hr at 1350°C and 5 x lom5 Torr and furnace-cooled to produce a grain diameter of 0.11 mm. Material analyses in ppm were: C N O H Sheet 290 10 4 1 Bar 50 30 30 2 The unnotched tensile sheet specimens had a 1/4-in. gage width and a 1-in. gage length. The unnotched tensile round specimens had a 9/32-in. gage diameter and a 1-1/4-in. gage length. The sheet and round notch tensile specimens are shown in Fig. 1. All of the notched tensile specimens tested were tandem-notched in order to study the location and mode of fracture initiation. The tandem-notched specimens were carefully machined so that the largest variation in notch depth on a single specimen was 0.001 in. Thus, when fracture occurs, the extent of plastic deformation that exists in the un-fractured notch section is that which exists just prior to fracture. All of the tandem-notched sheet specimens were electrolytic ally polished and chemically etched prior to testing. The V-notched Charpy-type specimens machined from the 50-mil sheet material had the standard dimensions of 2.125 in. long, 0.394 in. deep, and a 0.010-in. root radius. The unnotched tensile and notched tensile tests were conducted over the temperature range of -78° to 300°C, in a 10,000-lb Instron Universal Testing Machine, at a constant crosshead speed of 0.020 in. per min. The V-notched Charpy-type

Jan 1, 1964