Search Documents

By J. P. Denny

The constitution of the Ga-In system was determined by thermal methods. An experimentally determined metastable equilibrium line (an extension of the indium-rich liquidus) was obtained. The various alloys were studied metallographically using polished samples obtained by a casting method. These low melting alloys required a special dry-ice assembly to maintain a suitable temperature. RECENT interest in alloys that are liquid at room temperature has led to rather extensive investigation of gallium-base alloys. Widely distributed over the earth, gallium could be produced in substantially larger quantities than at present, if a significant demand existed.' One study' has established its presence in 12 out of 14 zinc blends, in all of 15 aluminum ores, in 4 out of 12 manganese ores, in 35 out of 91 iron ores, and in all of 7 magnetite ores. It occurs as a rule in minute amounts, however, leading to high extraction costs. Recent quotations run from $2.50 to $7.50 per g. During the course of the present investigation, portions of the system Ga-In have been redeter-mined, and the results of this study are presented herein. Thermal and metallographic methods have been employed. Lecoq de Boisbaudran, the discoverer of gallium, conducted the first investigation" on Ga-In alloys in 1885. The temperatures of incipient melting, and of completion of melting, were determined at four alloy compositions. In 1936, Hansen' constructed a eutec-tic-type phase diagram for the system Ga-In, based on . work. The existence of a Ga-In compound was regarded as improbable by Hansen, and subsequent investigations are in agreement. French, Saunders, and Ingle3 conducted a more complete study of the system in 1938, using thermal methods. Their phase diagram is a eutectic type, containing a unique concave-upward liquidus. The solid-solution range of gallium in indium was reported as 9.5 pct by weight, and that of indium as less than 1 pct, at the eutectic temperature. The eutectic composition, determined as being bracketed by the compositions showing a true horizontal at the eutectic temperature (16°C), was reported as 76 5-0.5 pct Ga and 24 i 0.5 pct In. Experimental Procedure The preparation of Ga-In alloys is simplified by the low melting points involved. Various compositions were prepared by melting in pyrex tubes, using a cover of distilled water or parafin to prevent the alloys from wetting the glass wall. In all cases, the melts were homogenized. by stirring. Where possible, both cooling and melting curves were determined. The extensive undercooling of gallium was found to prohibit a satisfactory cooling-curve analysis of gallium-rich alloys, however, and transition points on the gallium side of the eutectic could be determined only by melting curves. The inverse rate method of thermal analysis proved to be most satisfactory and was used to a great extent. Various heating and cooling rates were used, ranging from 0.2" to 5.0°C per min. Low temperature melting analyses were conducted within a constant temperature bath, maintained at about 70 °C. The alloys were solidified (under water or paraffin) within a pyrex tube, using dry ice; the tube was then sealed within a cold Dewar flask, the unit transferred into the constant temperature bath, and periodic temperature readings taken. The high temperature melting-curve determinations and all cooling-curve determinations were made in a vertical tube furnace. At near-eutectic compositions, the furnace was placed within a refrigerated room held at —20°C. Accordingly, the furnace on cooling approached —20°C asymptotically and permitted the determination of those phase transitions occurring below room temperatures. Temperatures were measured with a 30-gage iron-constantan thermocouple, immersed directly in the alloy. To prevent contamination of the melt, the leads and junction were coated with Lucite, applied by painting with a solution of Lucite in ethylene dichloride. Electromotive force measurements were made with a Leeds and Northrup precision potentiometer, type 8662. The couples were calibrated against the boiling point. of water and, at lower temperatures, against a calorimeter thermometer having a Bureau of Standards certificate. The melting points of gallium and indium used in the present investigation were determined as 29.-77° and 156.1°C, in good agreement with previously reported values of 29.78°C° and 156.4"C.' The spec-

Jan 1, 1953

By I. W. L. Finlay, H. R. Ogden, R. Jaffee, D. J. Maykuth

Aluminum has been found to be soluble in a titanium to about 26 pct, and to raise the temperature range of transformation from a to 8. Two intermediate phoses exist in the system, a new face-centered tetragonal phase, designated as 7, which occurs between 34 and 46 pct Al, and TiAI Metallographic and X-ray diffraction data were used to determine the diagram. ALUMINUM is one of the useful alloying additions for titanium-base alloys. It is an important constituent in two commercial alloys released in 1950, one containing 4 pct A1 and 4 pct Mn, and the other containing 3 pct A1 and 5 pct Cr as the metallic alloy content. It is probable that, in the future, other titanium alloys containing aluminum will be released. In the course of alloy development work, some of the features of the constitution of titanium-rich Ti-A1 alloys have been developed. These are described in the present paper. The only available information on the Ti-A1 system concerns the aluminum-rich end of the diagram. In summarizing the early work, Hansen' indicated that there is little solubility of titanium in aluminum, and that a two-phase region exists up to the intermediate compound, TiA1a (37.2 wt pct Ti), which has a congruent melting point at 1355°C. Fink' has reported that a peritectic reaction occurs at 665°C between the melt and TiA1³ to form the terminal solid solution of titanium in aluminum. The solubility of titanium in aluminum has been established at 0.24 pct at 510°C and is estimated to be between 1 and 1.5 pct at the peritectic temperature. No information has been published on the diagram above TiAl,, other than a statement- hat aluminum raises the a to ß transformation temperature of titanium. Materials and Preparation of Alloys Iodide-type titanium and high-purity aluminum were used in this investigation. Typical analyses of these materials are given in Table 1. Alloys were prepared by arc melting on a water-cooled copper hearth, using the techniques described previously for contamination-free melting.' Weight losses during melting were very low, ranging from 0 to 0.02 g in a 10 to 15 g charge, indicating that little or no loss of aluminum occurred during melting. This permitted the use of nominal compositions for these alloys. Alloy ingots containing from 0 to 7.5 pct A1 were rolled in air at 850°C to 0.040 in. sheet. Alloys containing from 7.5 to 16 pct A1 could only be rolled about 25 pct reduction at temperatures up to and including 1200°C. Alloys containing above 16 pct A1

Jan 1, 1952

By J. L. Taylor, P. Duwez

The phase boundaries in the ternary system Ti-Cr-Al have been established at 1800° and 1400°F for alloys containing more than 60 pct Ti. The martensite transformation temperature has been measured for the titanium-rich alloys. THE studies of binary alloys of titanium with the transition elements which have been published so far1-0 ndicate clearly that the solubility of these elements is always much greater in the high temperature body-centered cubic form of titanium (ß phase) than in the low temperature hexagonal form. In this category of alloying elements are iron, nickel, chromium, manganese, vanadium, molybdenum, tungsten, columbium, and tantalum. Although studies of the nontransition elements have been rather limited, early work on titanium binary alloys'" has shown that most of them were not very soluble in the a form, with the exception of aluminum, for which the solubility in a is about 25 pct.13, 12 In addition to this large solubility in a, aluminum is also soluble in ß titanium to the extent of about 35 pct and the transformation temperature from ß to a is raised with aluminum content. This effect is also typical of aluminum, in contrast with that found for the transition elements which lower the ß to a transformation temperature. Therefore, it may be anticipated that tprnarv titanium alloys with aluminum and any one of the transition elements will constitute a class of ternary alloys having characteristic features absent in alloys involving two transition elements. Chromium was chosen as a typical transition element for the present study because ternary alloys of Ti-Al-Cr commercially produced have been found to possess very interesting physical properties (MST-3 pct A1-5 pct Cr of Mallory-Sharon).'" Binary Systems The binary systems Ti-A1 and Ti-Cr have been established recently by several investigators. The data used in this investigation were taken from the complete diagram of Bumps, Kessler, and Hansen." This diagram agrees quite well with earlier investigations." For the Ti-Cr system, the phase boundaries at 1800" and 1400°F were taken from refs. 6 and 7. The binary system Cr-A1 has been studied recently by Bradley." This system is rather complex and contains not less than seven intermediate phases. The present study was limited to the titanium corner of the ternary system (less than a total of 40 pct Cr plus Al) and none of the phases present in the Cr-A1 binary alloys were found in the ternary alloys investigated. Experimental Technique Microscopic observation and X-ray diffraction measurements were used to identify phases and de-

Jan 1, 1954

By F. Wald, A. J. Rosenberg

The general features of the constitution of the B-Pt system were determined using standard rnetal-lograph~c, thermoanalytic, and X-ray diffraction techniques. Three compound were found. Two of these, Pt3B and Pt,B, are formed by peritectic reactions at 523° and 890°C, respectively. The third, Pt3B,, is congruently melting with a flat maximum at 940°C but decomposes eutectoidally in to Pt,B ant1 boron nt - 600° to 650°C. THE low-temperature allomorph of boron (red, simple rhombohedra1 a boron) is of scientific and technological interest as an elemental semiconductor.' However, the studies of this material have been hampered by its reported instability above 1200"~ which precludes crystal growth from the melt (mp - 2200°C). Crystallization from platinum solutions has been suggested as an alternative crystal-growth technique, but has met with only limited success.' The technique depends upon the existence of a significant difference between the eutectic temperature and the transformation temperature of boron. In order to clarify the conditions for further crystal-growth experiments, we found it desirable to redetermine the main features of the B-Pt phase diagram since previous reports on the system1'5'6'7 are in marked disagreement. EXPERIMENTAL The experimental methods used were thermal analysis, metallography, X-ray analysis, and, to a lesser extent, measurements of microhardness. Most of the alloys were prepared from spectrograph-ically standardized boron obtained from Johnson-Matthey &Co., Ltd. (212 ppm impurities, exclusive of carbon and oxygen) and platinum powder obtained from F. Bishop & Co. (200 ppm impurities, mainly of other platinum group metals). Some alloys were also prepared with very high-purity, float-zone refined boron (99.9999 pct obtained from "Wacker Chemie" and extrahigh-purity platinum (99.999 pct) obtained from Johnson-Matthey & Co., Ltd. The reported results did not depend on the choices of these starting materials. Five-gram alloy specimens containing 10, 20, 25, 27.5, 30, 33.3, 34, 35, 37, 37.5, 38, 39, 40, 41, 42, 43, 45, 50, 55, 60, 70, and 80 at. pct B were made by melting the elements together in boron nitride crucibles using rf heating of a graphite susceptor, either in vacuum or under high-purity argon. All alloys were heated to at least 1800°C for -5 to 15 min. Most of the alloys did not wet the crucibles when the latter were outgassed by preheating under vacuum. In any event, no weight loss was detected after melting, and the nominal composition was assumed for all specimens. Thermal analysis on 2.5-g samples were carried out in boron-nitride crucibles under a vacuum of 5 x X torr. The apparatus was heated in a "Kan-thal A 1" wound furnace, which limited the maximum temperature to about 1100°C. The output of the indicator thermocouple was fed to a dc recorder with a 1-mv full-scale span and an adjustable zero. The apparatus was calibrated repeatedly, using the freezing points of high-purity aluminum, silver, and gold. The results justified the use of the NBS voltage vs temperature tables for Pt/Pt 10 pct Rh thermocouples. All thermal analyses were run at least twice and both the heating and cooling effects were recorded. Most of the alloys had a very strong tendency to supercool. However, the use of mechanical vibration permitted reproducibility within *5°C for all alloys, except in the region around 40 at. pct B. Only the cooling effects are plotted in Fig. 2, since they appear to be more reliable. For metallography, the alloys were cut with a diamond cutting wheel, cast in a polymethacrylate resin, ground and polished with diamond paste, and etched with dilute aqua regia, a common etch for platinum alloys. Both copper and molybdenum radiation were employed to obtain X-ray diffraction data using Debye-Scherrer cameras and a "Norelco" diffractometer Diffractometry with high scanning speeds (1 deg per min) using nickel filtered CuK, radiation was used to identify the main regions of the diagram. However, molybdenum radiation was used for the detection of boron, since the latter showed very strong absorption and fluorescence effects with CuK, radiation. RESULTS AND DISCUSSION Three intermediate compounds, corresponding to the compositions Pt3B, Pt2B, and Pt3B2, were found in the system. Fig. 1 reproduces their X-ray diffraction spectra, together with those of pure boron and pure platinum. As can be seen from the thermal-analysis data in Fig. 2, Pt3B and Pt2B are formed by

Jan 1, 1965

By Norman L. Newbern, Robert E. Green

An investigation of the effect of size on the deformation properties of metal single crystals made it desirable to grow from the melt cylindrical aluminum single crystals possessing exactly the same crystallographic orientation but varying in diameter. Methods for the growth of rather large (up to 3 in. diam) aluminum single crystals in molds of the desired shape are well known. However, when one applies the same techniques to the growth of smaller (less than 1/8-in. diam) specimens many additional problems arise. Two of the major problems are the construction of the molds to be used and the removal of the thin specimens from these molds without deformation after growth has been completed. Also the problem of the constraint forces exerted by the mold on the specimen is much more pronounced for the smaller specimens. The technique described herein is a zone-melting method in which no mold is used since the surface tension of the molten metal and its oxide is sufficient to retain the melt. The specimen is suspended vertically in a manner similar to that of Keck and Golay.2 However, while Keck and Golay supported the specimen from both ends, the method used here supports the specimen only at the upper end. The surface tension of the molten aluminum and aluminum oxide contains the melt and suspends a length of newly grown crystal that has already passed through the molten zone until the weight of the newly grown crystal is equal to the component of surface tension holding the specimen together in the molten region. When the weight of the newly grown crystal is equal to the component of surface tension holding the specimen together the specimen pulls in two in the molten zone. The length of crystal which can be grown before this separation occurs varies with the diameter of the specimen and the size of the molten zone. In order to prevent the specimen from pulling in two a tube of glycerine is placed beneath the furnace in such a manner that the newly grown crystal passes into the glycerine which provides sufficient buoyancy to allow aluminum single crystals in wire form to be grown to lengths determined only by the length of the tube containing the glycerine. A schematic diagram of the experimental arrangement is shown in Fig. 1. The furnace used is constructed as follows. High-resistance nichrome wire is coiled on a 3/8-in. diam fused quartz tube. The distance between the windings is much less over a length of about 3/8 in. at the center so as to develop a hot spot. This core is located in the center of a 4-in. diam steel cylinder 5 in. in height and insulated from it by Kayolite. The temperature of the furnace is controlled by means of a thermocouple located near the central hot spot and a Minneapolis-Honeywell temperature controller. Temperature ins,ide the furnace is measured with a thermocouple and a standard potentiometer in the usual manner. By the above method it is possible to grow high-purity (99.99 pct Al) aluminum single-crystal wires over a range of sizes. Crystals varying from 1/8 to 1/64-in. diam were grown by the authors. Since the tube containing the glycerine cannot contain a crystal longer than approximately 3 ft, the growth of crystals longer than this presents another problem. It was found that one could cut the newly grown crystal without deformation in the region between furnace and glycerine tube try means of a cutting torch and then by moving the glycerine tube slightly aside continue

Jan 1, 1962

By W. G. Pfann

A simple method of obtaining multistage batch separations by crystallization was described recently. Known as zone-refining, it comprises passing short molten zones through a long solid charge. This technique can now be used on a continuous basis by means of the zone-void method described in this paper. Feed enters, at an intermediate point, a column down which molten zones travel, and waste and product leave at the ends. Materials move in the column through the agency of voids, which are introduced at the ends and travel toward the feed inlet. The voids and molten zones are moved by external heaters in a simple manner, and the principle of reflux is utilized. ANEW method of obtaining multistage separations by crystallization was described recently.' Named zone-refinina.,-, the method comwrises slowly passing a series of molten zones through a long solid charge. Solute becomes concentrated at one or the other end of the charge, depending on whether it raises or lowers the freezing point of the solvent. The separation increases with the number, P, of zone-passes, approaching a limit as P approaches infinity. Zone-refining has been highly effective in purifying germanium and other substances.2,3 and new applications are steadily increasing. Zone-refining is a batch method and as such it has certain limitations inherent in batch operation. If it could be made continuous, its scope and utility would be greatly broadened. This end has been achieved by the zone-void method described in this paper. In the zone-void method feed is introduced continuously at an intermediate point in a column down which molten zones travel, while impure waste and purified product leave at the ends. Both the flows of feed, waste and product, and also the travels of the zones, are actuated by external moving heaters in a simple manner; and the system utilizes the principle of reflux. The method provides, in the field of crystallization, the counterpart of the continuous fractiona-tion column in the field of distillation. The following will be discussed: apparatus and mode of operation, fundamental nature of the separation, design theory, and practical considerations. The method will be described in terms of a binary solute-solvent system in which the solute is an impurity to be removed and the solvent is the desired product. The distribution coefficient, k, defined as the ratio of solute concentration in the solid freezing out of a molten zone to that in the liquid in the zone, is assumed to be constant and less than one. The process is equally effective for k's greater than one and for ternary or higher order systems. Method and Apparatus The essential features of a continuous zone-refining process are represented in highly generalized form in Fig. 1. A series of molten zones, produced by moving heaters, travels slowly down the column or charge (to the left in Fig. 1). If there were no flows of feed, waste, or product, the process would simply be batch zone-refining, the action of the molten zones being to sweep solute down the column, solvent up the column. For the process to be con- tinuous, with stripping and enriching sections in the column, feed must enter, and waste and product must leave, as indicated. The zone-void process accomplishes both objectives, namely, the indicated movements of zones and the indicated flows of material. Zones are moved by moving heaters, just as in batch zone-refining. Materials are made to flow by creating voids at the waste and product exits and causing these voids to move to the feed inlet. Since there must be a net flow of material from the feed inlet to each of the outlets, the indicated movements of voids are in the desired directions, because movement of a void in a given direction corresponds to flow of material in an opposite direction. In order to produce the desired movements of voids, the column is folded into two vertical sections having the feed inlet in common at their upper ends. Voids are displaced upward by the liquids in the molten zones and their travel is actuated by the motions of the zones. Voids travel with the zones in the enriching section and move continuously. Voids travel opposite to the zones in the stripping section and move intermittently. Creation and travel of voids will now be examined in detail. The enriching section of a column in operation, with its void generator, is shown in Fig. 2. The column section is a vertical tube around which a series of closely fitting, regularly spaced heaters travel slowly upward. Each heater produces a molten region, the temperatures of the heaters and the cooling between heaters being controlled so as to maintain the molten zones approximately constant in size. A void is normally present atop the molten zone in each heater. As the heater rises, it continuously melts solid above it, which drips through the void into the molten zone and continuously freezes out solid below it, of concentration k times that of the liquid in the zone. When a molten zone and void reach the feed inlet, which is kept molten, the void is displaced by an equal volume of feed liquid. Generation of voids of controlled size in the enriching section is shown in Fig. 3. The void generator is a tube of small cross-section, provided with lateral heat-conducting fins which sense the position of the heater. Liquid can escape only when the entire outlet tube is within the heater. If any part of the outlet tube is outside the heater, liquid cannot escape,

Jan 1, 1956

By W. Chubb

By measuring the hardness of metals at temperatures just above and just below their allotropic change point, it has been established that crystal structure has a real effect upon the strength of metals. The DPH of cobalt, iron, titanium, uranium, and zirconium have been measured at temperatures up to 1000°C. It was found that the body-centered-cubic crystal structure is always the softer structure when it is involved in an allotropic transformation. The close-packed and more complex structures are inherently harder, and may be expected, therefore, to be better base materials for high strength alloys. IT has been observed that for metals with nearly the same melting point (i.e., tin and lead, magnesium and aluminum) the metal with the more complex crystal structure has a higher tensile strength and hardness. It might be asked whether this effect is really caused by the difference in crystal structure; and, if so, just what is the contribution of crystal structure to the strength of these metals. A qualitative answer to these questions can be obtained by measuring the change in hardness produced by the change in structure in metals exhibiting allotropic forms. The evidence for changes in mechanical properties during allotropic transformations is rather meager. In 1909, Rosenhain and Humfrey' observed that "ß" iron showed much less deformation at the transformation temperature than a iron. By means of a torsion test, Lee found that iron bars were "critically plastic" at the a-7 transformation temperature. Sauerwald and Knehans have shown that the dynamic hardness of iron increases between 900" and 1000°C. In 1930, Albert Sauveur' presented data on the hardness, tensile strength, and torsional strength of pui-e iron up to 1000°C. He concluded, "At 900°C, the a iron transforms into y iron, and this is accompanied by increased strength and stiffness and decreased ductility." The same year, Schischokin" obtained data on the hardness of thallium at temperatures above and below its transformation temperature. His data are replotted on semilogarithmic coordinates as Fig. 1. The fact that his data fall on straight lines in this type of plot is believed to attest to the purity of his metal and the accuracy of his results. In 1941, Nadai and Manjoine" obtained tensile data on pure iron with an indicated increase in strength at about 900°C. These experiments make it quite evident that changes in mechanical properties occur when a metal transforms from one crystal structure to another; but, except for Schischokin's data on thallium, none of the measurements seem accurate enough to determine the relative magnitude of the mechanical property change. Materials The metals that have been examined for changes in hardness during transformation include l—as- deposited commercial electrolytic cobalt about 99.7 pct pure, 2—Armco ingot iron, 3—are-melted iodide titanium of a purity in excess of 99.8 pct, 4—uranium of a purity in excess of 99.5 pct, 5—are-melted iodide zirconium of a purity in excess of 99.8 pct, 6—annealed, 0.08 pct C, rimmed steel, and 7—commercially pure RC55 titanium. While the purity of the first five of these materials leaves something to be desired, it will be noted that the observed hardness discontinuities occurred at the accepted transformation temperatures. (These temperatures are indicated by vertical arrows in Figs. 1 and 3 through 7.) The last two materials are commercial alloys introduced to show the effect of minor alloying elements upon the hot hardness curve. Apparatus Hardnesses at elevated temperatures were determined by means of a vacuum hot hardness machine. A diagram showing the basic features of this machine appears as Fig. 2. This machine is essentially a deadweight hardness .machine in a furnace within a vacuum chamber. The sapphire-tipped indentor

Jan 1, 1956

By J. L. Brimhall, R. A. Huggins, M. J. Klein

IN a recent paper1 it was shown that small additions of magnesium, copper, and oxygen decrease the stacking fault probability in plastically deformed silver. Correlation of :X-ray data with measurements of the twinning frequency and metallographic observation of slip lines indicated that this effect may be due to an increase in the relative stacking fault energy. It is the purpose of the present note to report experiments on these alloys of known stacking fault probability that relate to the stacking fault contribution to electron scattering in silver. There has been a considerable amount of disagreement in the literature concerning the contribution of dislocations to electrical resistivity in metals. Calculations relating to the scattering of conduction electrons by the dislocation core, its strain fields, and related effects have been made by several authors.2-5 In addition, experimental work on Cu,3,6-10 Ni,11 and Ni-Co alloys1:! has shown that dislocations in those materials exert an anomalously large effect on resistivity, this effect being related to the presence of extended dislocations. The problem of the scattering due to the stacking fault region between partial dislocations has been treated theoretically using several different procedures.13-'0 The results of these calculations show a wide range of predicted values, from rather large to completely negligible, due to the models and the approximations involved. Almost all of this work has been applied to the specific problem of scattering in copper. The more complete theoretical treatments depend rather critically. on the shape of the Fermi surfaces and the effective mass parameters. The scattering from stacking faults might therefore be expected to vary significantly from one metal to another. Unfortunately, sufficiently accurate information is not available .which would allow one to have any confidence in a theoretical prediction of the scattering in the case of silver. Resistivity measurements were made on small wire specimens of vacuum melted high purity silver and several dilute alloys. These measurements were made at approximately 20°C, with great care being given to temperature measurement and corrections. At such temperatures phonon scattering is significantly greater than imperfection scattering. However, the experimental technique was such that measurements on similarly treated samples were reproducible to better than 2.4 X 10-9 ohm-cm, which was satisfactory for this work. Values of the stacking fault probability and the increase in resistivity after severe plastic deformation are given in Table I. The increases in resistivity resulting from plastic deformation agree quite well with previous work on silver.14,21-26 In addition, it has been previously found that interetitials and vacancies in silver anneal out well below room temperature, so that the residual imperfection scattering in these experiments could be expected to be primarily associated with dislocations;. It is seen from the data in Table I that large changes in stacking fault probability are accompanied by differences in resistivity due to imperfection scattering of only a few percent. If the scattering of conduction electrons by extended dislocations is primarily due to the area of the stacking fault region independent of the stacking fault energy, and not the core and strain fields associated with partial dislocations, the resistivity increase should be proportional to the product of the dislocation density and the stacking fault width, as is the stacking fault probability. Therefore, the imperfection scattering caused by plastic deformation of the alloys should be considerably smaller than that found in the pure silver. Since this was not observed, these results sugegest the possibility that stacking faults do not themselves contribute appreciably to the electrical resistivity of silver deformed at room temperature. An alternative explanation of these observations is that the scattering by the stacking fault is roughly proportional to the stacking fault energy, rather than being determined by the total area of such faults. Sufficient evidence is not yet available to separate these two possibilities. This work was supported by the Office of Naval Research.

Jan 1, 1962

By L. Burris, C. H. Stockman, I. G. Dillion

Zone melting is a purification process in which separation of impurities is effected by slowly moving a narrow melted zone through a bar of solid material. Equations are presented which 1—predict the concentration profiles after successive passes of the zone and 2—give the limiting distribution of solute along the bar after an infinite number of passes of the molten zone. These equations take into account the effect of normal freezing in the last zone length. These equations were solved numerically using an electronic computer, and the concentration profiles are presented as a function of the solute distribution coefficient and the ratio of molten zone to total bar length. ZONE MELTING is a purification process in which separation of impurities is effected by slowly moving a narrow melted zone through a bar of solid material. Separation occurs if the solid freezing from a solution at any moment has a composition different from that of the liquid from which it is freezing. Solutes which lower the freezing point will tend to remain in solution and will be moved in the direction of zone travel. Solutes which raise the freezing point will preferentially freeze out, and they will be moved in a direction opposite to that of zone travel. After repeated unidirectional passage of a molten zone through a bar of material, impurities will be concentrated in the ends of the bar. A limiting concentration gradient will be reached in the bar after a large number of passes of the molten zone. Both the magnitude of this limiting concentration gradient and the rate of movement of solute in any pass depend on the zone length and on the equilibrium concentrations of solute in the liquid and solid phases. The basic theory of zone melting was thoroughly covered by Pfann.1 The purpose of this paper is to extend the theory of zone melting by presenting 1—equations which relate the concentration profiles in successive passes of the zone and 2—equations which give the limiting distribution of solute along the bar after an infinite number of passes of the molten zone. These equations take into account the effect of normal freezing* in the last zone length concentration of a given solute in the solid to the concentration of the same solute in the liquid. Analytical methods for determining the concentration profiles after any pass have been developed by Lord2 and Reiss. The method developed by Lord does not take into account normal freezing in the last zone and also becomes very complicated when the number of passes becomes large. The method developed by Reiss enables the direct calculation of the complete concentration profiles (including that in the last zone) after any pass for values of k close to unity (0.9 < k < 1.1). Both methods are useful and provide valuable insight into the zone melting process. The equations given in this article are based on the same assumptions used by Pfann. These are: 1—uniform composition in the liquid, 2—negligible diffusion in the solid, 3—constant distribution coefficient, 4—constant length of the liquid zone (except at the ends), and 5—solubility of the solute in the melt is not exceeded at any point. Equations for Concentrations in Successive Passes For a melted zone of length 1 at the left-hand end of a bar of uniform unit cross-section, the amount of solute in the liquid is J0Cn-1(x) dx [1] where the subscript is the number of the pass. Cn-1(x) represents the concentration (expressed as weight per unit volume) obtained in the previous

Jan 1, 1956

By H. Hansen, A. U. Seybolt, P. Yurcisin, B. W. Roberts

The Bi-Mn phase diagram in the region near BiMn was investigated, using principally thermal analysis and changes in magnetization with temperature. Of chief interest are the findings related to the magnetic transformation in Bi-Mn. THE Bi-Mn system is of both theoretical and practical interest because of the occurrence of the compound BiMn which exhibits unusual ferromagnetic properties. Because the Bi-Mn phase diagram reviewed by Hansen&apos; (see Fig. 1) in 1936 is based on work done by BekierY in 1914 and by Siebe" in 1919, which left considerable doubt about the nature of the diagram. it appeared desirable as an aid in understanding the alloying characteristics and magnetic behavior to examine the pertinent regions of the phase diagram. The most interesting reaction is the peritectic at 445°C which forms BiMn by the reaction: Liquid + Mn(sulfield) ? BiMn(sol field) There is also a eutectic at 262° C between pure bismuth and BiMn. Hansen draws in BiMn dotted, since in 1936 evidence for its existence was uncertain. Experimental Materials—Preparation of Alloys Electrolytic manganese of about 99.97 pet purity and bismuth of 99.99 + pet purity were used as starting materials.

Jan 1, 1957

By F. W. Glaser

Metal-carbon-boron powder mixtures were hot pressed and the resulting specimens were studied by X-ray diffraction. It was found that regardless of the starting combination of the metal, carbon, or boron powders, a metal boride phase was always the major component in these samples. In the absence of carbon the boride phase formed on hot pressing depended only on the amount of boron present. Two new phases of the system Ti-B were found. They are Ti2B and Ti2B5. The existence of a controversial face-centered cubic phase of formula TiB was confirmed. Electrical resistivities were measured for various boride phases. It was found that the diborides are generally better conductors than the monoborides of the same metal. THE carbides and borides of the transition elements have very high melting points, in the range 2500° to 4000°C, and are therefore of interest as high temperature materials. The literature on the stability or chemical reactivity of these carbides and borides is very scarce. Various investigators&apos;-" have demonstrated a relative instability of certain carbide phases in the presence of boron or boron-containing substances. In a recent publication, Glaserl demonstrated the stability of zirconium-boride (ZrB,) in the presence of carbon at temperatures in excess of approximately 2900°C, while during a preliminary investigation of boride phases, Steinitz&apos; concluded that the diborides are stable in the presence of carbon while the monoborides of the fourth and fifth group are not, forming diborides plus carbides instead. Nelson, Willmore, and Womeldorph" have elaborated on the reaction B,C + 2TiC = 2TiB, + 3C, which was known to occur because of a relative instability of B,C and the great tendency towards TiB, formation at relatively low temperatures (approximately 1200°C). A similar study, involving as starting materials TiO, and B,C and resulting in TiB,, was recently described by Honak4, who observed the beginning of an exothermic reaction of a Ti0,-B,C powder mixture, which, when preheated in a hydrogen atmosphere to approximately 950°C, was carried to about 1600 °C by the heat of reaction. To shed more light on reactions of this type (Metal-C-B), the final product apparently always resulting in a boride phase at the expense of a carbide phase," a systematic investigation was started * Boride phases of various metals, as reported to date, are listed in Table I. and the following is an account of some of the results that were obtained. Materials, Preparation of Samples, Testing Methods The raw materials employed for this work consisted of various carbide, boride, and metal powders. as well as of boron and graphite powders. In cases where commercial grades of carbides were considered unsuitable because of low purity or excessive amounts of graphitic carbon, such carbide powders were prepared by this laboratory. The procedure for the preparation of carbide powders (zirconium carbide, titanium carbide, tantalum carbide, and niobium carbide) consisted of mixing graphite and the respective metal hydride powders in stoichio-metric proportions and subsequent heating of such mixtures in a hydrogen atmosphere in carbon crucibles. The heating was by high frequency to temperatures ranging between 1700" and 2100°C. The resulting carbide was then comminuted and screened to the desired particle size. ZrB, and TiB, powders were produced by the electrolysis of fused salt baths, according to the method described by Andriex.. The borides of niobium, vanadium, tantalum, molybdenum, chro-ium, and iron were obtained by mixing the respective metal and boron powders in the desired proportions. Such metal-boron mixtures were heated in a high frequency furnace to form boride powders. For each metal-carbon-boron group (Tables I1 through XI) a metal, its hydride, carbide or boride were mixed with carbon, boron or boron carbide powders. The additions of carbon, boron or boron carbide powders to any of these metals or metal compounds were calculated to satisfy a particular carbide or boride phase that according to the literature (Table I) had definitely been established by X-ray diffraction work. Samples of powder mixtures were hot pressed in graphite molds that were heated by direct conduction. The specimen dimensions were approximately 2.5X1X1 cm. Hot pressing temperatures were measured optically and maintained for approximately 30 sec under a constant pressure of about 1.3 ton per sq in. Wherever possible, an attempt to obtain maximum specimen density was made by temperature variation. Electrical resistivity testing was done by measuring potentiometrically the voltage drop over a length of 1.5 cm for a current of 10 amp, at room temperature. To obtain electrical resistivities for specific carbide or boride phases, values were plotted as a function of the respective sample densities

Jan 1, 1953

By H. J. Fisher, J. L. Walter

Two types of Si-Fe columnar ingots with preferred crystal orientation have been produced: a) ingots with (0011 directions of the crystals parallel to the longitudinal ingot axis and b) seeded ingots in which all crystals have the (100) [001]orientation with respect to the sides of the ingot. Measurements of freezing rates and temperature gradients indicated that columnar freezing occurred as long as the temperature gradient ahead of the solidification front was greater than zero. THE development of preferred orientations or textures in wrought metals may be influenced by the presence of a preferred crystal orientation in the ingot from which the product was made. The deliberate introduction of preferred orientation in the ingot for the express purpose of producing a particular preferred orientation in the wrought product has not, to the authors&apos; knowledge, been reported in the literature.* This report describes the prepara- tion of ingots of 3 1/4 pct Si-Fe in which preferred orientations near (100) [001] were produced. These ingots were subsequently rolled to strip, 0.012 in. thick, to provide a sufficient quantity of polycrystal-line sheet having the (100) [001] orientation or "cube-texture" to evaluate the magnetic properties. The properties; and test results showing the advantage of the cube-textured Si-Fe magnetic sheet over magnetic sheel having the (110) [001] texture or "Goss" texture have been reported elsewhere.&apos; The rolling and recrystallization textures of single crystals of silicon-iron having orientations near (100) [001] have been studied by Walter and Hibbard.2 They found that if the crystal has an original orientation such that the (100) plane is within 30 deg of the rolling plane, the recrystallization orientation will approach (100) [001]. If the (110) plane is nearly parallel to the rolling plane,

Jan 1, 1962

By Murray Kaufman

Using en; 41 for a base alloy, 15 modifications were made in sheet form. The elements Al, Ti, C, Mo, Co, and Cb were varied. The effects on the phases formed and their temperature dependence, the aging characteristics, tensile and stress-rup- ture properties, and weld cracking were determined. The A1-Ti content and balance is the most important factor influencing the major sCrengthening phases [Ni,Ti and Ni,(Al, Ti)] and, consequently, mechanical properties. THE mechanical properties, fabricating and joining characteristics, dimensional stability, and so forth, of alloys containing secondary phases are largely determined by the phases present, their distribution, and their temperature dependence. The phases in the basic ternary and quaternary Ni-Cr-Al-Ti combina- tions have been investigated thoroughly.&apos;* The phase behavior of many commercial nickel-base alloys has been determined also. In some cases, for example U500, 5 the effect on mechanical properties has been found. However, little systematic work has been done to clarify the action of each of the important alloying elements in controlling the phases formed in complex commercial alloys and their behavior. The effect of molybdenum on the carbides formed in M252 was investigated by Beattie and verSnyder.6 Application of the results provided by that work

Jan 1, 1963

By Eric R. Morgan, J. C. Shyne

STRAIN aging is the name given to time-dependent changes which occur in the properties of cold-worked metals and alloys during storage. These changes are best observed through a study of mechanical properties, as illustrated in Fig. I for low-carbon steel. The effects of strain aging are usually regarded as being detrimental, particularly in deep-drawing steels. For example, an increase in yield strength and loss in ductility lead to buckling and tearing during drawing. At the same time the return of the yield point results in the formation of Luders bands or stretcher strains during the pressing operation. As a result of these effects many attempts have been made to eliminate strain aging from low-carbon steel. The problem still remains, however, because all known methods of preventing strain aging result in a more costly steel. The present work is an assessment of the situation in terms of theoretical, practical, and economic factors and suggests a new solution to the problem. Previous Work According to the Cottrell theory,&apos; strain aging in a-iron may be regarded as the migration of interstitial atoms to free dislocations and the anchoring of dislocations by the resulting atmospheres. It may be concluded that the important factors which will govern the degree of strain aging are the diffusivity, the effective amount of interstitial element in solution, and the time of aging. The diffusivities and effective amounts of carbon and nitrogen in solution in low-carbon rimming steel are such that significant aging occurs in a few hours&apos; storage at room temperature. Strain aging in deep-drawing steels cannot be controlled by controlling the time of aging because in commercial practice this is determined by economic factors concerned with inventories and may be several weeks. The diffusivity of nitrogen is twice that of carbon at room temperature but the diffusivities can be controlled by controlling the storage temperature. The retardation produced by only a 25 °C drop in temperature is significant but the cold storage of large steel inventories is an impractical solution to Table I. Analysis of Base Materials Alloy Pet C Yet 0 Pet N Vac 0.0037 0.052 0.0003 H 0.0044 0.013 0.0008 C 0.06 0.0014 0.0001 the problem. By far the most simple and effective ways to control strain aging are through control of the amount of carbon and nitrogen in solution. This may be effected through alloying, controlled precipitation, or by eliminating the carbon and nitrogen. Consider first the control of the amount of carbon in solution. The simplest approach would be alloy -ing in which strong carbide-forming elements are added. Examples of this approach are the additions of titanium&apos; and niobium9 0 steel. These elements also combine with nitrogen and oxygen and when added in sufficient quantity result in a completely killed and non-aging steel. They are, however, impractical additions because they result in a costly steel. Carbon and nitrogen are often removed from iron on a laboratory scale by annealing in wet hydrogen, but because of cost this process has never been adopted on a commercial scale. In practice the amount of carbon in solution is controlled by the rate of cooling from the annealing temperature (710°C or 1310°F). In commercial box

Jan 1, 1958

By R. M. Rose, C. S. Tedmon, J. Wulff

It is frequently of interest to study the influence of small quantities of interstitial impurities on the properties of various refractory metals and alloys.

Jan 1, 1964

By R. L. Fleischer, R. S. Davis

BICRYSTAL specimens grown from the melt have been used extensively for investigations of the influence of grain boundaries on plastic deformation. The bicrystals are usually of predetermined orientation prepared by seeded growth from the melt in milled graphite boats. This technique has been described in detail by Chalmers.&apos; Since the direction that the grain boundary follows during growth is affected by both thermal conditions and crystallography, a difficulty arises in maintaining the grain boundary position constant along the length of the specimen. This problem is particularly

Jan 1, 1960

By J. B. Clark

In multiphase ternary diffusion studies, plots of the variation bf the composition on the isotherm constitutes one of the most effective methods of presenting experimental results, especially if such plots are made in a manner which relates the composition path on the isotherm with the diffusion layer structure in the couple. Unfortunately, if different conventions for plotting are used serious misinterpretations occur readily, causing apparent disagreement between investigators who actually share the same viewpoint. A common convention for mapping these diffusion paths is needed at this time because with the increasing use of the electron-probe X-ray microanalyzer, a much greater number of ternary diffusion studies can be expected in the near future. In an effort to prevent unnecessary confusion in the future litera- ture, a convention for these plots is proposed, as outlined below. A hypothetical example of the use of these conventions is given in Fig. 1 where the path of composition through the diffusion structure of the 100 pct A/50 pct B, 50 pct C couple is plotted on the appropriate isotherm of the ABC system. The lower case letters relate the region in the diffusion layer structure to the appropriate composition on the ternary isotherm. 1) The diffusion path is plotted on the isothermal section corresponding to the diffusion temperature of the couple. 2) The terminal compositions of the plotted diffusion path are the initial compositions of the couple halves—provided infinite boundary conditions are maintained. If, in either couple half, infinite conditions are not maintained, then the path terminates at the composition of the exterior surface of the couple. If, at this surface, the composition varies with time or if the plotted path is time dependent, then the diffusion time should be noted beside the plotted path. 3) A solid line crossing a single phase field (e.g., line a-b in Fig. 1) denotes an existing layer of that phase in the diffused couple. 4) A dashed line crossing a two-phase field parallel to the tie lines (e.g., line g-h) represents an interface between the two phases with interfacial compositions designated by the ends of the tie-line. The line represents no spatial extent. 5) A solid line crossing a two-phase field so as to cut tie lines (e.g., line b-c or line j-k) represents a locally equilibrated columnar two-phase layer. The solid line should cut those tie lines which correspond to the lateral interfacial equilibria along the columnar

Jan 1, 1963

By Nicholas J. Grant, Klaus M. Zwilsky

EVER since the unusual high temperature creep resistance and structure stability of SAP (Sintered Aluminum Powder) and similar aluminum-alumina alloys were reported,&apos;," there has been a need to determine whether other metal-metal oxide (or other hard stable particle) alloys, not so ideal as the Al-Al 2 O 3 system, would show similar high temperature strength and stability. Of the many possible methods of obtaining a metal-metal oxide alloy,a one that offers many practical and theoretical advantages is the simple mechanical blending of metal powders with hard stable particles, followed by pressing, sintering, and hot extruding. With this purpose in mind copper powder and silica and alumina were selected for experimentation because these materials are readily available, cheap, reliable, and lend themseives te the simplest processing and testing. Starting Materials—Two sizes of electrolytically deposited copper powders were used, a —200 mesh powder referred to as A powder in this investigation and a —325 mesh powder referred to as C powder. Silica was added in the form of pure Potter&apos;s flint from which a —10µ fraction had been separated. Alumina of the a variety was used in two size ranges. The first was a —10µ fraction separated from a coarser aggregate, the second a commercial product of 0.3µ average diam manufactured by Linde Air Products. Preparation of Powders—Five-hundred gram batches of A copper with 1, 3, and 10 volume pct of each of the three oxides mentioned above were prepared, as well as C copper containing 1, 3, and 10 volume pct of —lop silica. Powders were weighed out carefully and blended by ball milling in air for 24 hr. Since this treatment resulted in surface oxidation of the copper, the charge was hydrogen reduced at 260°C for 5 hr. The next step consisted of hydrostatically compacting the powder into a shape suitable for the required hot working. For this purpose a slug about 11/2 in. in diam and 31/2 in. in length was produced by placing the mixed powders into a rubber sleeve, holding in a perforated steel canister, and rubber stoppering at the ends; the assembly was then subjected to a hydrostatic pressure of 50,000 psi. The air was evacuated from the assembly prior to pressing in order to prevent

Jan 1, 1958

By C. S. Smith, L. H. Beck

KNOWLEDGE of the Cu-Zn constitution diagram was critically examined in 1944 by G. V. Ray-nor,1 who suggested the diagram reproduced in Fig. 1 as the best evaluation of available data. No pertinent work on the constitution has been published since that time, although a number of authors have discussed transformation kinetics and theory. Despite the large amount of work that has been done on the constitution of these industrially important and scientifically interesting alloys, minor uncertainties still exist. For example, it is still not possible to say definitely whether a two-phase region exists between the ordered and disordered forms of the 8 phase. A careful re-examination of the phase boundaries in this part of the diagram was therefore undertaken. Two methods were selected for the new investigation—quantitative metallography for the higher temperatures and precision lattice parameter measurement for lower temperatures. Because of the relatively narrow two-phase regions flanking the ß phase in these alloys, quantitative metallography is particularly suited for their examination, since the relative amounts of the two phases are very sensitive to their compositions. The method, however, is useless below about 350°C because of the inadequate resolution of the two constituents. Below 400°, however, the X-ray diffraction method is applicable, although it is of somewhat lower sensitivity in this system. The changes of composition that occur during quenching from above about 500°C entirely vitiate X-ray measurements at room temperature though they do not affect metallographic area measurements. The two methods overlap and supplement each other admirably. Both methods depend on the attainment of equilibrium composition of the two phases in a more or less intimate mixture and thus avoid the difficult problems of nucleation and long range diffusion involved in the conventional metallographic bracketing method. The alloys were melted from a master alloy made from re-electrolyzed cathode copper and Horsehead 99.99 pct Zn. Melting was done under a cover of either borax or charcoal and the alloys cast into a 7/8 in. sq graphite mold for the a-ß alloys or a 3/8 in. diam cylindrical graphite mold for the ß-? alloys. The a-ß alloys were rolled at 750°C to 5/8 in. sq; annealed for 24 hr at 800°C and quenched; then rolled and swaged at 200°C to 3/8 in. diam rods, annealed 24 hr at 200°C, quenched and cold rolled 2 pct reduction. This treatment was intended to obtain a small grain size and to produce many nuclei of the minor phase for growth on subsequent heat treatment. The unworkable ß-? alloys were given preliminary treatment by sealing in Vycor tubes in vacuum and annealing for 4 hr at 800°C followed by cooling in air. Samples of homogenized alloys were checked spectrographically for impurities, but the principal components were determined on samples subsequent to heat treatment. The impurities analyzed averaged approximately as follows: Ag, 0.0002 pct; Fe, 0.004; Si, 0.0008; Cd, 0.005; Mg, 0.0002; and Pb, 0.002. For metallography, samples of appropriate composition were cut into rods 5/8 in. long, sealed in pyrex tubes in vacuum and annealed for eight weeks at temperatures of 470°C and below, two weeks at 480°, 500°, and 600°C, and 4 hr at 700°C, followed by quenching in cold water. Temperatures were accurate to ±2°C. The cylinders were then machined to 8.7 mm diam and a slice weighing 1.5 g cut with a thin saw transversely from the center of each. This

Jan 1, 1953

By W. R. Johnson, M. Hansen, John M. Parks

AS part of a research project sponsored by the Signal Corps Engineering Laboratories, which had the objective of obtaining a magnet wire of good conductivity and low temperature coefficient of resistance, a comprehensive literature survey, supplemented by experimental work, revealed that a linear relationship exists for solid-solution alloys between the electrical conductivity, K, and the temperature coefficient, a, as expressed by the equation: k = a/ß where ß is constant for the solvent metal and a - per°C. tR0 It has long been known that the electrical conductivity at a given temperature and the temperature coefficient of resistance of binary alloys are related to their constitution by the LeChatelier-Guertler rules, as illustrated by Figs. 1 and 2: Rule 1—In alloy systems consisting of a heterogeneous mixture of two phases, the electrical conductivity and temperature coefficient vary as a straight line, if the composition of the alloys is given in volume percentage, Fig. 1. Rule 2—In solid solutions, the electrical conductivity and temperature coefficient are always lower than those of the solvent metal, and usually a considerable decrease occurs with the first small percentage additions of the solute metal, Fig. 2b. If the two components form a continuous series of solid solutions, the electrical conductivity and temperature coefficient curves take the form of a U-curve, Fig. 2a. That some kind of proportionality exists between electrical conductivity and temperature coefficient of resistance is implied by the LeChatelier-Guertler rules and by study of data such as those given by Smith and Palmer,&apos; Fig. 3. Dellinger2 in 1910 first recognized this relationship in a number of commercial copper wires, ranging in conductivity from 94 to 100.7 pct I.A.C.S. He gave the ratio of temperature coefficient to the percentage of conductivity as a constant, C = 0.003939. A year later, Lindeck observed that the linear relationship was valid for a wider range of conductivity in "various sorts of copper," and concluded that the product of resistivity and temperature coefficient was 6.78x10-5 1.5 pct. Since then, this relationship seems to have been entirely ignored. Conductivity-Temperature Coefficient Relationship If the electrical conductivity of copper solid-solution alloys with phosphorus, silicon, arsenic, antimony, aluminum, nickel, manganese, and iron is plotted against the temperature coefficient of these alloys, the diagrams in Figs. 4 and 5 are obtained. Similar data were found with copper-rich solid solutions containing magnesium, cobalt, and titanium,&apos;" as shown in Fig. 6. In all cases, a linear relationship exists. The point representing the highest

Jan 1, 1952