Search Documents

By L. D. Jaffe

From the limited experimental data in the literature, preliminary values were derived for the thermal diffusivity of titanium alloys and for the quenching severity of various mediums used in heat treating them. The thermal diffusivity, under quenching conditions, was found to be approximately 0.006 sq in. per sec. The quenching severity H for brine is approximately 8, for water 4, for oil 0.9, for air 0.08, and for argon 0.02 in. -' The quenching severity of the Jominy-Boegehold end quench water jet against titanium alloys appeared effectively infinite. Using these results, graphs were prepared showing ideal round sizes and Jominy positions having cooled conditions equivalent to those at the center and surface of titanium alloy rounds, plates, and sheets quenched in various media. THE importance of proper quenching of steel parts has long been recognized, and much quantitative work has been done on the subject. Recently, it has become evident that for titanium alloys, too, the rate of cooling from high temperature is important in obtaining high strength and hardness. Maximum hardness and strength may not be attained in the as-quenched condition but rather after a subsequent tempering or aging treatment. Nevertheless, if cooling from temperatures near the ß transus is too slow, for the hardenability of the alloy used, a soft a-ß structure will be produced which will not harden on subsequent aging.'-" The quenching of parts of practical interest can usually be approximated by simple shapes, such as cylinders (rounds) and plates (or sheets) of various thickness and combinations of these. A standard that has been used extensively for comparison purposes is the center of an ideal round: a cylinder quenched in a medium that instantly cools its surface to room temperature." For steel, considerable practical use has been made of the Jominy-Boegehold end quench hardenability specimen,".' and its value for titanium alloys is becoming evident. It therefore appeared worthwhile to attempt comparison of cooling conditions in titanium alloy rounds and plates of various sizes quenched in various media with cooling conditions in ideal round and Jominy quenches. These comparisons are based upon measurements reported in the literature. Quenching Severity of Jominy Water Jet and of Brine As one starting point, there were available the data of Phillips and Tobin8 for positions on a Jominy specimen having the same hardness as the centers of round bars of various size quenched in 5 pct NaOH brine, shown in Table I. The Jominy speci- men is a round bar, normally 1 in. diam and 4 in. long, quenched by a water jet at one end and air cooled on the other surfaces.' After quenching, hardness is measured vs distance from the quenched end on longitudinal flats ground 0.015 to 0.10 in. deep. The data of Table I were based on a subsized specimen of 3/4 in. diam. Since the positions tested are close to the jet quenched end, cooling there may be taken as due solely to the jet, and air cooling can be neglected. These positions are then equivalent to various locations near one surface of an 8 in. plate jet quenched on both sides. Assuming that the thermal diffusivity is constant and that Newton's law of cooling holds, temperature vs time curves at corresponding positions in Jominy specimen and round bar then may be calculated for various assumed quenching severities of the water jet and of the brine. Cooling curves are generally taken to be metallurgically equivalent when they show the same time to reach a temperature halfway between the initial (heat treating) temperature and the final (quenching medium) temperature.5 On this basis, using Russell's tables,9 upplemented by heat flow tables derived in similar fashion by the author for positions near the surface of severely quenched plates, it was found that the experimentally observed equivalence, Table I, was best matched if the quenching severity H of the Jominy water jet was taken as infinite (H = co). This means that the jet quenched surface should be considered as instantly cooled to the temperature of the water

Jan 1, 1956

By M. E. de Morton

Low frequency-internal friction measurements on annealed chromium have shown a marked increase in damping below - 40°C which is strongly strain amplitude dependent. An interpretation of these results is given in terms of the motion of antiferromagnetic domains assumed to exist below - 40°C. A time effect on internal friction was measured after slight overstraining. After heavy deformation a pemzanent, five-fold increase in internal friction was observed at -45°C and the character of the internal friction us temperature curve markedly changed. Full recovery of damping was not attained until after heating to 800°C. A qualitative interpretation of some of these effects is given. THE aim of this preliminary investigation was to study the internal-friction characteristics of chromium containing comparatively small quantities of interstitial solutes in the fully annealed and heavily worked condition. Marked anomalies in a number of physical properties of chromium at temperatures near 40°C have been observed1 and an investigation by neutron diffraction' has shown that chromium at room teyperature is weakly antiferromagnetic with a Nee1 temperature at approximately 200oC. No irregularities in the degree of antiferromagnetic ordering have been observed at -40°C, the transition temperature, and a complete explanation of this transition in chromium is still awaited. Recently weiss3 has suggested that an ordering effect of nitrogen in the face-centered interstices of the bcc lattice of chromium could be responsible for the superlattice detected below 200 °C in the neutron diffraction experiments. EXPERIMENTAL MATERIAL AND METHOD Specimens 10 by 0.030 in. diam were prepared from ingots of arc-melted electro-deposited chro-

Jan 1, 1961

By Nick Holonyak

Halogen vapor-transport synthesis of Ga(As,-,Px) and its preparation into laser junctions are described. Electrical and optical properties of Ga(As,-,PX) laser junctions are discussed. The present limitations in these properties are related to material proble?vzs and the very early state of development of Ga(As,-,PX), and are discussed in this context. Inhonzogeneity problems, fluctuation of As :P ratio, and problems with deep-level contaminants are described and related to Ga(As,-,PX) junction performance. Laser junctions are demonstrated which operate to wavelengths as short as 6470 A (77°K). The prospect that Ga(As,-,P,) will operate to wavelengths perhaps 100A shorter is mentioned. IN this paper we wish to describe some of the more important features of almost 1 year of experience in the preparation and properties of Ga(As,-,P,) p-n junction lasers. Because much data so far available are far from complete, some of the results still must be considered tentative and subject to correction and revision as work on Ga(As,-,P,) crystal problems progresses. Although the electrical, optical, and device properties of Ga(As,-,P,) junction lasers are understandably of considerable interest, the work to date points out that by far the most serious problems extant and those deserving first attention are problems involving the growth, doping, and perfection of Ga(As,-,P,) crystals. One of the most formidable of these is to grow Ga(As,-,P,) sufficiently free of deep levels, either due to contaminants or to structural imperfections, to allow fabrication of laser p-n junctions which can be cooled and not suffer from "freeze-out" of carriers on deep levels (and consequent double-injection behavior). In addition to the problems of growing crystals adequately free of deep levels and generally of the highest perfection, so that the best possible laser junctions can be realized, the further goal remains of increasing the phosphorus content well beyond 40 pct in an attempt to determine the practical lower wavelength limits of laser action in the Ga(As,-,P,) system. Even though the laser junctions prepared in Ga(As,-,P,) are of importance in their own right, we have considered them at this point mainly a means to study the growth and crystal problems of the material. It can be concluded that the laser junctions now available are almost completely limited by the purity and structural deficiencies of Ga(As,-,P,) crystals, and cannot be fairly compared with, for example, GaAs p-n junction lasers until Ga(As,-,P,) crystal growth has been more extensively developed. The basis for the statements above will become evident in the following sections which describe Ga(As,-,P,) crystal and junction preparation, and laser-junction properties. G(As,-,P,) CRYSTAL GROWTH As previously reported,2 we have found it convenient to employ the halogen vapor-transport process3 for synthesizing Ga(As,-,P,) crystals for use in laser junctions. Commercially available GaAs and Gap, separately prepared by halogen vapor tranport, are sealed in the desired ratio in a quartz vessel with a quantity of column-VI donor and a small quantity of metal halide which supplies the halogen for transport. The sealed ampoule, as shown in Fig. 1, is heated at a temperature in the range from 1025" to 1100°C at the position of the source GaAs and Gap. The GaAs and Gap are simultaneously vapor-transported and deposited as Ga(As,-,P,) at one end of the ampoule

Jan 1, 1964

By G. Wakefield, A. H. Daane, D. H. Dennison, F. H. Spedding

Preparation of pure scandium metal was accomplished by calcium reduction of the fluoride by two methods; a low-temperatzdre alloy process and direct reduction with subsequent distillation of the product. The following properties were determined: melting point: 1181OK; boiling point (calculated): 3000°K; lattice constants at 298°K (hexagonal lattice): a = 3.308 * 0.001 A, c = 5.267 * 0.003A; calculated density at 298°K, g per cm3: 2,990 + 0.007; electrical resistivity, ohm-cm: 299°K, 66.6 ± 0.2 x 10 -6; 373oK, 77.4 * 0.2 x 10 -6; thermal coefficient at 299°K, ohm-cm per deg: 5.4 X x; heat of sublimation at 298°K, kcal pel- mole: 80.79. The vapor pressure was determined as a function of temperature between 1505o and 1748°K, with the data fitted to a straight line shielding the equation: Log Pmm = -1.718 X 104/ToK + 8.298. SCANDIUM, element number 21, was first discovered by Nilson in 1879 and was recognized as the Ekaboron as predicted by Mendeleff. As it is in group III of the periodic table, the general properties are a little like aluminum and also resemble quite closely the properties of yttrium and the rare-earth metals, in both the metallic and ionic form. Although the earth's crust contains approximately 5 ppm of scandium (the element is as abundant as arsenic and twice as abundant as boron) it generally occurs so widely distributed that it has earned the reputation of being very rare. The one exception to this is the mineral thortveitite, which has been found in Madagascar (20 pct Sc2O3) and in Norway (35 pct Sc2O3). Scandium also occurs in small but distinct amounts in uranium and rare-earth ores; the recent larger scale processing of these materials has made some scandium available from these sources. As with other naturally-occurring monoisotopic elements (except Be), scandium contains an odd number of protons and an even number of neutrons. Scandium metal was first prepared by Fischer and coworkers1 in 1937 by electrolysis of scandium chloride in a molten eutectic mixture of lithium and potassium chlorides, using molten zinc as a cathode and collector of the scandium metal produced. The zinc was removed from the Zn-2 pct Sc alloy by vacuum distillation, leaving a product reported to be 94 to 98 pct Sc, with the main impurities being iron and silicon. They reported a melting point of 1400° C for this material. Scandium has also been prepared by the reduction of scandium chloride with potassium metal in a glass apparatus by Bommer and Hohmann in 1941,' resulting in a mixture of metal and potassium chloride; these workers did not isolate the metal proper, but the X-ray diffraction of the slag-metal mixture showed it to be hexagonal with a = 3.30A, c = 5.45A. petru3, 4 and coworkers have recently reported the preparation of the metal in a compact form by the reduction of either ScF3 or ScC13 with calcium metal and subsequent distillation of the product. This process probably yielded a metal of high purity, but they list no chemical analysis nor do they list any of the properties of their product. Previous related work in this Laboratory has been concerned with the production of yttrium and the rare-earth metals and the determination of their physical properties. Because of its similarity to these metals, scandium is being included in this study. PREPARATION OF SCANDIUM METAL The preparation of yttrium and the rare-earth metals may be accomplished by reduction of their fluorides with calcium metal in tantalum crucibles.5 This process leads to the introduction of tantalum (up to 0.5 pct) as an impurity in the higher melting rare earths, but since the tantalum occurs as dendrites, uncombined with the rare-earth metals, its presence is not objectionable in some cases. The preparation of scandium metal in this manner, however, was found to yield a product containing 2 to 5 pct Ta. To obtain a purer product, the following two methods were developed for the reduction of scandium fluoride with calcium metal: i) a low-temperature process utilizing zinc to form a low

Jan 1, 1961

By G. W. Cullen

Niobium-tin has been prepared on insulating suhstrates hby simultaneous hydrogen reduction of gaseous niobium and tin halides. Stoichiometric material is greater than 98.8pct theoretical density, appears to he single-phase, and has a total impirity level less than 0.3 at. pct. Transition temperatures and lattice constants may he related to composition. Typical current densities in various fields and orientations are given. IT is known that Nb3Sn carries large current densities in high fields, 1 and exhibits an unusually high transition temperature2 and critical field.3 For these reasons high-field solenoids have been constructed of this material.4 Because of the emphasis on solenoid fabrication, much of the materials research has been directed toward development of wire geometries wherein the brittle Nb3Sn is supported by a more malleable metal or alloy. There are, however, basic studies and applications that can be carried out only on Nb3Sn supported by an insulator. Insulating substrates are desirable for a number of reasons: the substrate does not interact with applied or induced currents and fields; diffusion of metallic species into the de- posit is avoided, both during the preparation and during subsequent heat treatment; irradiated samples are not more active than the deposit itself. Unsupported Nb3Sn geometries may also be prepared by preferential chemical dissolution of the insulating substrate. This report describes the preparation and properties of Nb3Sn deposits on ceramic substrates. Experiments currently utilizing material prepared in the manner described in this report include the following: dependence of current-carrying properties on field strength and orientation 5,6 radiation damage,7,8 tunneling,9 thermal conductivity (substrate removed),1° microwave surface impedance," tube magnetization,6,12 resistivity as a function of temperature,13 and current density and transition temperature as a function of annealing.14 Possible practical applications include magnetic shielding, superconductor generators, and frictionless bearings. PREPARATION Methods for the deposition of elemental niobium and tin by hydrogen reduction of the gaseous metal halides have been known for some time,15,18 but only recently a procedure was developed in this laboratory17 for the codeposition of these two metals on a hotwire as the intermetallic compound, Nb3Sn. This method of simultaneous hydrogen reduction of the gaseous niobium and tin chlorides is used for the preparation of Nb3Sn On ceramic substrates. The procedure differs from that employed on a hot wire in that insulating substrates cannot conveniently be heated to temperatures greater than their surround-

Jan 1, 1964

By Martin Weinstein

A technique is described for preparing finegrained lead telluride by ultrasonic agitation of a solidifying melt. Material prepared by this technique is dense and chemically homogeneous. N-type PbTe, containing 0.03 mole pct PbIz, has been made and its thermoelecCric Properties have been measured as a function of temperature. For comparison, the properties of similarly doped PbTe made by cold pressing and sintering, and by extrusion, have also been measured. The ultrasonically prepared material compares favorably with the others in its thermoelectric-efficiency factor. Comparison of compressive-flow curves for the three kinds of PbTe show that the ulhasonically prepared material has much superior ductility. The cleavage strength of the ultrasonic material is related to the grain size, D, by the Petch relation:Extrapolation to infinite grain diameter yields a fracture strength of 5000 psi for single-crystal PbTe. WHEN metals and semiconductors solidify in a mold they normally assume one of three cast structures illustrated in Fig. 1.' Solidification in a casting starts at the mold walls where a large number of fine crystals of random orientation can be nucleated to form the chill zone. Because of the anisot-ropy of the growth rate, those crystals whose direction of maximum linear growth is parallel to that of the heat flow grow faster and crowd out the crystals of less favorable orientation. A second zone, having a preferred orientation, called the columnar zone, thus forms adjacent to the chill zone. The speed of growth of the columnar grains will depend, in the case of limited undercooling of the melt, on the rate at which heat can be abstracted. The columnar zone can extend to the center of the casting, Fig. l(a), but it is frequently followed, in solid-solution alloys, by another zone of equiaxed randomly oriented grains in the central portion of the casting, Fig. l(b). The equiaxed structure in Fig. l(c) represents the completely grain-refined case where the equiaxed grain size is small. The third type of structure is the most desirable because it imparts a somewhat higher strength for all except high-temperature applications, it reduces segregation of the doping constituent, and it reduces the possibility of hot tearing (the separation of grains due to the melting of a low-melting point constituent in the grain boundaries). In order to obtain fine-grained castings, the nu-cleation frequency must be increased and grain growth must be restricted. Various techniques have been employed for obtaining this end, namely: rapid cooling, vibration, and inoculation of the melt. The success of these techniques depends upon the characteristics of the metal species and on factors controlled by heat transfer through the crucible walls. Recent studies2" have shown that the application of ultrasonic energy to the solidifying ingot is an extremely powerful tool in the improvement of the crystalline structure of metallic castings. Both the mechanical properties and homogeneity of the ingots are improved. Ultrasonic oscillations may affect the crystallization process by 1) permitting subcritical nuclei to agglomerate and thus form nuclei of sufficient size to cause crystallization, 2) changing the state of surface adsorption of various catalysts found in the liquid melt, 3) breaking pieces of solid from the solidified surface near the supercooled region thereby increasing the number of possible nuclei, 4) causing temperature fluctuations to occur in certain parts of the melt thereby increasing the supercooling at specific points in the liquid, and 5) decreasing the critical nucleus size by increasing the pressure at specific points in the melt.' Various investigators have concluded that the effectiveness of the ultrasonic oscillations results from cavitation which disrupts the growing solid-liquid interface The present paper describes the application of ultrasonic oscillations to the crystallization of the semiconductor N-type PbTe. The purpose of this study is to prepare crystalline

Jan 1, 1964

By Ronald Gibala

THE Au-Ni alloy system is one of many binary systems in which the supersaturated solid solutions may exhibit a multistaged precipitation behavior during aging. In addition to the equilibrium phase separation, which was studied extensively years ago in the range 400" to 600",' there exists at least one earlier stage of precipitation, i.e., preprecipitation, which occurs below 300°C in quenched alls.-This note reports on some effects of composition on the preprecipitation in Au-Ni alloys as determined from resistivity measurements. The kinetics of the preprecipitation were studied in alloys of 15 to 85 at. pct Ni which were water- were found in alloys which after various quenching-and-aging treatments were slowly heated and cooled (1Z"C per hr) between room temperature and about 300°C. Two such heating-cooling curves for a 40 pct Ni alloy are shown in Fig. 2. Similar effects were observed in other 15 to 50 pct Ni alloys. These curves are characterized upon heating by two changes in slope indicated by A and B in the figure and upon cooling by a resistivity hysteresis of approximately the size of the original decrease in resistivity after the aging treatment. The inflection at B generally falls at about the same temperature for all heating curves for a given alloy and appears to be a critical temperature (T,) for the preprecipitation for that alloy. The values of T, for most of the 15 to 50 pct Ni alloys fall in the range 200" to 220°C. The position of the inflection at A depends upon the heat treatment prior to the heating-cooling run and perhaps the heating rate, which was held constant in these runs. The preprecipitates grown in by the appropriate aging treatments seem to decrease in size as the alloys are heated from A to T,, dissolve at T,, and do not reappear after cooling below T,. When similar heating-cooling treatments were performed on 60 to 85 pct Ni alloys, an entirely different behavior was observed. Typical results are shown in Fig. 3 for an 85 pct Ni alloy. In these

Jan 1, 1964

By A. U. Seybolt, R. A. Oriani

The pressure and composition of the solid solution of nitrogen in chromium has been investigated as a function of temperature. The partial molal heat of solution, the activity of nitrogen in solid solution, and the partial molal free energy have been computed. The enthalpy of three reactions is summarized: N + 2Cr(3) = Cr2N(4), H° = —28,500 cal per mol 1/2N2(p) + 2Cr(8) = Cr2N(3)1 ?HO = —23,500 cal per mol 1/2N2(p) = N, ?H° = 5,000 cal per mol. WORKERS who have investigated the problem of the preparation of ductile chromium have been concerned with the effects of minor amounts of impurities upon this refractory metal. In particular, gaseous impurities such as oxygen and nitrogen have been suspected of causing trouble because of many known cases where brittleness in transition elements has been definitely traced to these elements. In order to evaluate the possible effects of the nitrogen impurity in chromium, it appeared desirable to learn something of the dependence of the nitrogen solubility upon temperature and pressure. Previous Work Maier3 in 1942 in his treatise on sponge chromium reviewed the thermodynamics of the chromium nitrides Cr2N and CrN, both of which are well-estab- lished compounds, but which apparently have not been studied in much detail. Maier presented some free energy-temperature expressions for the reactions 2 Cr,N = 4 Cr + N2 [1] and 4 CrN = 2Cr2N + N2 PI based mainly on the work of Satoh2 and Sano." Pearson and Ende3 in 1953 briefly reviewed the thermodynamics of Cr-N using the same data available to Maier, and also derived a free energy-temperature expression for reaction 1. Of the two nitrides mentioned previously, the present investigation was only concerned with Cr,N, since this nitride is in equilibrium with the metal phase, at least over most of the temperature range, Blix4 in 1929 established the structure of Cr2N (close-packed-hexagonal) by X-ray diffraction methods, and indicated that its homogeneity limits extend from 11.2 pet to 15.1 wt pet N. At a temperature of 1200°C, 38 mm Hg from the Maier expression and 43 mm Hg from the Pearson

Jan 1, 1957

By C. G. Dunn, P. K. Koh

The primary recrystallization textures of 11 single crystals cold rolled 70 pct were determined. The effect of crystal orientation on recrystallization behavior was investigated also and found to be large. An exception was found to the relationship that strong components of the recrystolliza-tion texture are rotated from the cold rolled texture by 30° about <110> axes, but generally this relationship was found to hold quite well. Evidence supporting oriented nucleotion was found and discussed in terms of probable active slip systems and <110> edge dislocations. THE present paper is one of a series&apos;.&apos; on recrys-tallization phenomena in cold rolled 31/4 Si-Fe single crystals and is mainly concerned with the textures produced on recrystallizing from stable end orientations. Data for the cold rolled textures of the crystals used have been reported by Koh and Dunn." Their results showed that the sharpness of the cold rolled textures varied throughout the range of stable end orientations and that the tendency to re-crystallize was related to the sharpness of the cold rolled texture, wherein the rate of recrystallization increased and the size of recrystallization grains decreased with decreasing sharpness of the cold rolled texture. In the project under consideration the cold rolled crystals involved a range of initial orientations that, together with those studied earlier,&apos; provide information on several points. The first point is whether the recrystallization texture from a (111) [112] cold rolled texture, a stable end orientation that can be approached from different directions, correlates in any way with the initial orientation before cold rolling, or whether the recrystallization texture is entirely determined by the observable features of the cold rolled texture. The second point involves a preliminary study to see if the main components of the recrystallization textures correlate with the most probable active slip systems. The third point is to determine what effect departure from slip on one slip system has on the tendency to recrystallize. Experimental Procedure Nine specimens from a previous investigation and two additional ones from thicker stock (lot D) but of similar composition to the others were used in the present work. The latter two crystals, like the nine, were cold rolled to a reduction in thickness of 70 pct and the amount of widening was determined. Depending on their primary recrystallization behavior, the cold rolled crystals were given short or long anneals at temperatures of 980°C or higher. Table I summarizes the data on initial orientations, the textures obtained by cold rolling, and the anneals given to provide samples for the present work. Fine grained specimens were thinned to 0.002 in. thickness and placed in an integrating specimen holder. Chart recordings using a transmission X-ray method and MoKa radiation were obtained for diffraction from (110) planes. This method,&apos;,3 which gives data for all of the pole figure except a central region of 30" radius, proved sufficient to give generally five, and sometimes six, of the six possible pole concentration areas of any one component of the texture. These data, therefore, were generally adequate for a positive identification of each component present. The first texture obtained for specimen B-2 was poorly defined because of too coarse a grain structure and therefore too few crystals in each component for satisfactory X-ray reflections. The same

Jan 1, 1957

By J. J. Gilman

Rates of prisnzatic plastic glide ( {1010}<2110>) in pure Cd crystals have been measured at temperatures from 158° to 276°C. The glide rate is proportional to the 2.75 power of the applied shear-stress. The activation energy is 29.2 kcal per mol. It is suggested that the rates of prismatic glide in Cd and Zn crystals are determined by the strengths of the atomic bonds across the prism planes. ALTHOUGH zinc and cadmium crystals both have hexagonal structures with nearly the same c/o ratios, their mechanical behaviors are quite different. Whereas zinc becomes quite brittle below room temperature, and cleaves readily on (0001) planes, cadmium is very ductile even at 4.2°K, and cannot be cleaved at any temperature. As part of a search for the explanation of this difference in behavior, a study has been made of prismatic glide in cadmium. It complements a previous study of the same phenomenon in zinc.1 The primary glide plane in cadmium is the basal (0001) plane, and the primary glide direction is the direction of atomic close-packing <2110>. However, if a tensile stress is applied parallel to the basal plane, then the shear stress on the basal plane equals zero so basal glide cannot occur. At low temperatures, deformation by twinning occurs; but at high temperatures, it has been found in this study that gliding on prismatic planes (1010) occurs in the close-packed direction. This is completely analogous with the case of zinc, except that lower temperatures are required. EXPERIMENTAL The experiments were done in nearly the same way as for zinc.&apos; Therefore, only the differences in the two sets of experiments will be described here; for other details the reader is referred to the previous paper. Crystals 6 by 6 mm sq were used that were grown from "Super Purity" 99.999 + pct Cd purchased from the New Jersey Zinc Co. The crystals were grown with their basal planes parallel to both the rod axis and one of the flat sides of the square cross section, A close-packed direction in the basal plane was oriented parallel to the rod axis. Gage sections were carefully machined into the crystals, and the surface damage caused by machining was removed with the same chemical polishing reagent as had been used for zinc. Thus, specimens of the same form and size as the previous zinc ones were produced. The crystals were heated to the test temperatures in a silicone oil bath (fused salts were tried first, but they reacted badly with the cadmium). An Instron tensile-testing machine was used for applying stresses to the crystals at various constant strain-rates. RESULTS The general behavior of cadmium crystals is much the same as for zinc. Only the constants that measure specific properties are different. Just as

Jan 1, 1962

By D. C. Jillson

BRIDGMAN1,2 melted metals in a graphite or hard glass tube and lowered the tube through a furnace to make it cool from one end only. Because of the difficulty of eliminating all vibration of the mold, others have raised the furnace instead. Neither way seems as desirable as a setup with no motion of either the mold or the furnace. A furnace in which both the shape and the slope of the temperature gradient could be controlled and varied over a wide range and in which such a gradient could be moved from one end of the furnace to the other, freezing the specimen from one end to the other, at any desired rate, all by means of isolated external control, seemed to offer definite advantages. In addition to eliminating the possibility of vibration due to moving parts, it might also give better control over the direction and rate of heat flow. Such an apparatus was assembled and is shown schematically in fig. 1. Equipment Furnace: A clay-carborundum tube of 3 1/2 in. ID with a % in. wall thickness was used as the core. This had a main winding of No. 12 B.&S. gauge Nichrome V with some compensation for heat loss at the ends. Taps to this winding were made at points 12 in. from each end. Fig. 2 shows details of this winding and of the two end windings. Each end winding was wound over Micanite over the main winding. No. 15 B.&S. gauge Nichrome V was used for these, and again taps were supplied at two inter-mediat,e points on each. The tube was supported in an 18 in. OD black iron shell with brick insulation. The input to each of the windings was regulated by means of a variable transformer. Temperature Regulation: A chromel-alumel thermocouple located near the furnace wall in the center of the furnace was connected with a Celec-tray controller, type PIC, 300" to 900°C range, which opened and closed a shunt across an external variable resistance in the main winding circuit. An "anticipatory response" device also was incorporated to minimize the cyclic effect of the control mechanism. Eleven survey couples in a common pyrex glass protection tube were located near the furnace wall and spaced 2 in. apart. The couple leads went to a common reference junction box where they were connected to copper leads from a selector switch which, in turn, was connected to a type K potentio-meter. This arrangement permitted rapid temperature surveys. The knob on the controller dial was replaced with a threaded brass spool turned at a controlled rate by a thin iron wire one end of which was attached to a gear train and the other to a small weight heavy enough to prevent slipping of the wire on the spool. The gear train, driven by a 1-rpm Telechron motor, contained a series of different sized interchangeable gears and threaded drums permitting a range of travel speeds from 0.10 in. per hr up to a hypo-

Jan 1, 1951

By H. P. Leighly, F. C. Perkins

THERE have been several statements in the literature about the difficulty of producing single crystals of high-purity (99.99pct) by the strain-anneal method. Consequently, investigators tend to employ low-purity aluminum for their single-crystal experiments, or else resort to the Bridgman or other techniques which depend on solidification for the production of single crystals. The following paragraphs describe a solid-state method for the manufacture of single crystals of high-purity aluminum which should provide crystals with greater perfection than those formed by solidification. The starting material for this method of producing single crystals was secured from the Aluminum Corp. of America and has a purity of 99.99 pct, the balance being trace amounts of impurities. Specimens approximately 1 by 4 in. are sheared from sheet having a nominal 0.050 in. thickness. The specimens are given a preliminary anneal at 640°C for approximately 3 hr in order to remove fabrication strains and to produce an average grain size of about 1/4-in. diam. The specimens are then etched in Tucker&apos;s etchant (45 pct HC1, 15 pct HNO3, 15 pct HF and 25 pct H2O) to remove the oxide film. Critical strain is applied by wrapping each specimen about a 1 3/4 in. round and subsequently straightening it against a flat surface. The high-purity aluminum sheet is sufficiently soft that this operation can be accomplished by ordinary finger pressure. The specimens are immediately annealed again at 640°C for 3 hr and reetched for examination. Ordinarily, considerable growth of certain of the grains will have occurred, and occasionally a single crystal will be produced on the first attempt. The procedure of alternately straining, annealing and etching is repeated until the majority of a batch of specimens contains usable crystal sizes. Typical examples are illustrated in Fig. 1. The greatest changes in crystal sizes are produced in the initial treatments. As the average crystal sizes get coarser in the later treatments, the sever- ity of the strain must be increased in order to produce grain boundary- migration. This increase in severity is effected by decreasing the diameter of the round used for straining (to 1 1/2 in., for example) and/or wrapping the specimens about the round twice, with opposite faces in contact with the round, before flattening. Usually the strain treatments described are not severe enough to produce nucleation in coarse grain high-purity aluminum. The growth of grains occurs by strain-induced grain-boundary migration. It has been observed that the grain boundaries move most readily during the first hour or so of each annealing treatment and that the rate of movement decreases with extended holding times at temperature. Prolonged annealing treatments are therefore not usually beneficial. Similarly, the rate of growth of each crystal appears to depend upon the orientation of the crystal with relation to those of its neighbors. Frequently island grains are formed after the initial heat treatment as the result of slow grain-boundary migration. These sometime become stationary during later heat treatments. Twin orientation interfaces are frequently developed during annealing. These imperfections can usually be removed by increasing the severity of strain to produce actual nucleation of new grains of more favorable orientation at the imperfection interfaces. The largest single crystals produced in our laboratory by the above method measured 4 by 1 by 0.050 in. Examination of Laue back-reflection patterns from a limited area of the specimens, gave no evidence of polygonization. Probably there is some indication of polygonization in the original grain area provided a more sensitive technique is used for detection. Experiments to produce wider specimens were less successful, possibly because wider sheets increase the complexity of the strains induced by deformation and promote widespread nucleation. Grain boundary migration occurs preferentially in a direction parallel to the longitudinal axis of the specimen. The choice of specimen geometry with respect to the rolling direction of the sheet appears to be immaterial in regard to the production of single crystals.

Jan 1, 1961

By O. N. Cole, R. E. Davis, T. D. Brotherton

Pvfoperties of rubidium and cesium are compared with those of other alkali metals. Methods for the preparation of rubidium and cesium metal are reviewed briefly, ad hazards and safety precautions required in handling rubidium and cesium metal are described. SimPle methods for handling laboratory quantities of elemental rubidium ad cesium under oil and under argon are illustrated. CESIUM and rubidium were both discovered by Bunsen and Kirchoff in the years 1860 and 1861. Cesium was discovered while Bunsen and Kirchoff were making a spectrographic study of spring water from Durkheim, Germany. It was the first element discovered by this technique. The element was given the Latin name "caesius" for sky blue because of the color it imparted to flames. Elemental cesium was first prepared by Setterberg in 1881 through electrolysis of a molten mixture of cesium and barium cyanides.&apos; In 1861, Bunsen and Mirchoff discovered rubidium while making a spectrographic study of the alkali content of lepidolite ore from Saxony. Two prominent dark red lines appeared in the spectrum and so the element was named "rubidius" or dark red. In 1861 Bunsen and Kirchoff first prepared elemental rubidium through the electrolysis of molten rubidium chl~ride.~ The alkali metals are very similar in their chemical and physical properties. They are the most active metals, and all exhibit a single oxidation state of one. They are so reactive that they are never found in nature in the elemental state. The reactivity of he metals increases with atomic number. Rubidiurn and cesium ignite spontaneously in air and react explosively with water. This extreme reactivity makes necessary special equipment and techniques for the safe preparation and use of the metals. The cation: of Group IA can be separated from the metallic cations of other groups with relative ease, but the separation of the individual alkali cations is much more difficult because of their close chemical similarity. This close chemical and physical similarity of alkali metals can be inferred from the atomic radii, oxidation numbers, and ioni-zation potentials shown in Table I. A difficult but the most practical physical method of separating the elemental alkali metals from one another is by distillation. Only lithium has a boiling point sufficiently far-removed from the boiling points of the other alkali metals to make this method of separation feasible without elaborate and expensive equipment. This method is not used, however, for lithium because it possesses sufficient chemical dissimilarity from the other alkalies for it to be separated relatively easily as lithium hydroxide monohydrate or as lithium carbonate. Francium, the last member of the group, has no known stable isotopes. The naturally occurring

Jan 1, 1962

By F. V. Lenel, M. V. Rose, A. B. Backensto

WHEN aluminum-flake powders are compacted and hot pressed and the resulting compacts are extruded or hot forged, a group of materials with unique properties is obtained. Not only do they exhibit high room-temperature strength of the order of some of the aluminum alloys, but they maintain this strength even after prolonged heat treatment at temperatures up to 900°F. Moreover, their strength decreases much less with increasing temperature than that of aluminum alloys of comparable room-temperature properties. Irmann,&apos;" who discovered and first described this group of materials, which was called SAP, showed that the room-temperature mechanical properties of the extrusions are a function of the oxide content of the powders and therefore the oxide content of the extrusions. The higher the oxide content, the higher are their tensile strength, yield strength, and hardness, and the lower their ductility. Lyle" investigated the relationship between mechanical properties at elevated temperature and oxide content in the course of the development work of the Aluminum Co. of America on materials, called APMP, similar to those developed in Switzerland. He found that at 600 °F, just as at room temperatures, tensile strength and yield strength increase and elongation decreases with oxide content. A careful study of his data (Fig. 3, loc. cit.) shows, however, that the scatter is more pronounced than at room temperature. Gregory and Grant&apos;s data on the creep5 and stress-ruptureb roperties of the one commercial grade of the Swiss SAP and the various grades of APMP indicate that creep strength and stress to rupture also generally increase with increasing oxide .content. Ever since this new group of materials was discovered, attempts have been made to explain the mechanism by which they obtain their high room-temperature and particularly their elevated-temperature strength. In most of the attempts the new material was considered as a dispersion of aluminum oxide particles in a matrix of aluminum. Ir-mann, Von Zeerleder, and Rohner&apos; applied a theory on dispersion strengthening developed by Rohners to the flake-powder extrusions. According to this theory the elastic limit S, of a dispersion would depend upon the average distance between the dispersed particles L according to the equation: 2 E ¦ a where E is the modulus of elasticity and a the distance between nearest neighbors in the lattice of the matrix material. Irmann and co-workers assumed

Jan 1, 1958

By S. J. Sindeband

J. WULFF*—It seems to me that the author could improve the quality of his high temperature material by using less nickel as a cementing agent in hot pressing. Furthermore, to avoid the presence of undilfused nickel which would not be oxidation resistant, permit me to suggest that he use a nickel-chromium powder containing a small percentage of boron. This would give him a liquid phase at a lower temperature and accelerate both sintering and homogenization. The commercial alloy known ascolrnonoy contains sufficient chromium to be heat resistant and sufficient boron to be of low melting point. Such an alloy may be applicable as a cementing agent in increasing the quality of Mr. Sindeband&apos;s alloys. Finally 1 believe the author unfair to himself in expecting a porous material to have properties equivalent to one of 100 pet density. S. J. SINDEBAND—Dr. Wulff&apos;s point is well taken. Early in the program we realized the fact that we were having difficulty with these low-melting alloys, and we had hoped, since the colmonoy alloys of which you speak are at the nickel-rich end of the diagram, that by going to the other end of the diagram (chromium boride-rich) we would not have this difficulty. Thus we tried to use very small percentages of binder; but for reasons which are not exactly apparent we found that when we did this it was necessary to raise the pressing temperature to a point which we were unable to attain. For all compositions which we tried using nickel, or for that matter any nickel iron or cobalt containing material as a binder, we found that we would get low melting boride phases, which as you point out are the commercial "strength" of the hard-facing materials. 1 personally believe these owe a great deal of their behavior and even hardness to the nickel boride which is considerable evidence when an alloy of chromium boride is made with nickel. Nickel boride itself is very hard. We did try pure chromium as a binder and found there was considerable difficulty in handling it, particularly since we could not just hot press under

Jan 1, 1950

By P. W. Ramsey, D. P. Kedzie

INCREASING demand for improved strength-weight ratios made on aircraft structures has resulted in a gradual increase in the tensile strength requirements for steels used in such applications. As the cyclic loads often are more critical than static loads, and the fatigue properties are generally the design criteria, the question arises as to whether an improvement in fatigue properties accompanies the increased tensile strength. The present investigation was undertaken to determine the relationship between fatigue and tensile properties for an aircraft quality steel in the 155 to 280 ksi (1000 psi) tensile strength range, using the Prot method of fatigue testing. Procedure and Results After preliminary work to establish proper heat treating procedures for this heat of aircraft quality electric furnace steel (a Ni-Cr-Mo-V steel," AMS- Prot Fatigue Testing Method—The determination of the endurance limit by the conventional Wohler method can be time-consuming and easily require 20 specimens when statistical variance is to be established. This is also true when endurance limit loads are determined on actual parts or subassem-blies, and the desirability of a reliable accelerated method has been evident for some time. The primary advantage of the Prot method of fatigue testing is in reducing the number of specimens and the time required for each test. Unlike the Wohler method, every Prot specimen tested con- tributes to the determination of the endurance limit, and the authors have been able to duplicate Wohler results obtained with 20 specimens with only half the specimens and a fourth of the total cycles. This new technique for defining the endurance limit was proposed in 1948&apos; by Marcel Prot and consists of progressively increasing cyclic stress until fatigue failure occurs. By varying the load rate, a, a useful relationship is found to exist with failure stress, so that for steel the failure stress appears to

Jan 1, 1958

By R. D. Reiswig

An increasing amount of high-temperature metall~?~gical research is carried out in resistively heated tube furnaces in which a bare specimen is suspended by a fine wire at the midpoint of the tube. It is shown that the assumption of near-blackbody conditions on the surface of the sample may give rise to errors in optical pyrometry of 50°C or more. A number of papers published in the metallurgical literature in recent years have described the use of high-temperature tube furnaces for heat treatments, melting-point determinations, and other work. Typically, the tube is of a refractory metal with open ends, is heated by the passage of an electric current, and has concentric cylindrical radiation shields outside it. In at least one of these,&apos; the specimen is contained within a crucible having a small hole in its lid through which optical-pyrometer measurements are made under essentially blackbody conditions. In another design,* the specimen is inside a muffle which is concentric with the heater tube and which contains a number of washerlike longitudinal radi- ation shields to reduce axial temperature gradients and to give conditions near the specimen closer to those of a black cavity. Yet another design3 employs a wire specimen coaxial with the heater tube. Blackbody conditions are simulated by passing current from a separate supply through the wire until it "disappears" against the background (the inside of the heater tube) as seen through a radial spy hole in the tube. Good agreement with known melting points is achieved when an optical pyrometer is used under such disappearance conditions as the melting point is approached. That this latter method works so well is due, evidently, to considerations such as those mentioned by Dike4 and others.5-8 A more common method than the above, however, has been the suspension of an uncovered specimen by a fine wire in a resistively heated tube with open ends. A radial spy hole about midway between the ends of the tube permits optical-pyrometer observations on the specimen, the inside of the tube, or both. So observed, the specimen typically appears to differ but little in brightness temperature from the inside of the tube. According to De vos9 the radiation from such a spy hole in an empty tube of reasonable 1ength:diameter ratio (e.g., 10 or greater) should be essentially blackbody radiation. For these reasons, apparently, descriptions of work carried out in such furnaces usually include a statement to the effect that conditions surrounding

Jan 1, 1964

By D. H. Shaffer, W. F. Harris, W. M. Hickam, M. H. Loeffler

Experiments have been performed on the quantitative addition to niobium of enriched isotopes 018 and 017 at the 0.1 - 40 micro-gram level and their subsequent recovery. A measured quantity of molecular oxygen of known isotopic concentration is added to niobium metal by oxidation. Recovery of the enriched isotopes as carbon monoxide is by the vacuum fusion technique. Quantitative determinations of the enriched isotopes are made in standard volumes at constant temperature using accurately calibrated micromanometers and the mass spectrometer as pressure measuring instruments. THE presence of low concentrations of nonmetallic impurities in metals can often adversely affect the physical properties of the metal. The increase in hardness and the embrittlement of refractory metals due to oxygen is well known. In metallurgical studies dealing with the diffusion rates or the effect of oxygen on the physical properties of refractory metals, it is often necessary to add small quantities of oxygen to the metal and later recover the total quantity present. It is desirable, therefore, to establish the accuracy with which microgram quantities of oxygen can be added to niobium and recovered by vacuum fusion. The successful addition of microgram quantities of oxygen to metals by the usual manometric techniques and the subsequent recovery by vacuum fusion is difficult because of the uncertainty in the accurate determination of the quantity of oxygen originally present in the metal and because of the chance of contamination by atmospheric oxygen. In order to minimize these uncertainties earlier workers have added quantities of O16 at levels much higher than the oxygen concentrations that are of interest.&apos;-3 In order to eliminate the uncertainty associated with the O16 additions to metals at low microgram levels, experiments have been performed on the quantitative addition of oxygen highly enriched in O17 and O18 at the 0.1 to 40 pg level to niobium and the subsequent quantitative recovery of the isotopes by vacuum fusion. Since the natural abundance of O17 is 0.04 pct and of 0" is 0.2 pct this enriched oxygen provides a uniqueness from interference not achieved in earlier work.&apos;-= A measured quantity of molecular oxygen of known isotopic concentration is diffused into the metal. Recovery of the isotopes in the form of carbon monoxide is by vacuum fusion. The total quantity of gas is determined in standard volumes at constant tempera- ture using accurately calibrated Consolidated Electrodynamics micromanometers and the mass spectrometer for pressure measurements. The molecular and isotopic composition is determined mass spectrometrically. APPARATUS A schematic of the oxygen addition apparatus is shown in Fig. 1. Take-offs are provided for the attachment of the isotope container and for sample bulbs which are used to transfer samples to the mass spectrometer. Metal valves are used throughout the system including the sample bulbs in order to achieve a vacuum of 10"6 mm Hg. Evacuation is accomplished with a 3-stage oil-diffusion pump backed by a mechanical pump. A liquid nitrogen trap is used to remove condensables and to prevent oil vapor from entering the system. A micromano-meter is used for pressure measurements. The furnace assembly is shown in Fig. 2. It consists of a quartz tube joined to the apparatus through a 40/W joint with Apiezon W low-pressure wax. The furnace cap, also a 40/W pyrex joint, is provided with a pyrex hook. The sample is suspended from this hook with a platinum wire. The total volume of the furnace section is 435 ml. All volumes were determined by expanding a known volume of gas at a known pressure into the evacuated system and re-measuring the pressure. Both the micromanometer used on the oxygen addition system and the micromanometer used on the mass spectrometer were compared to a precision

Jan 1, 1961

By R. W. Kraft, F. D. George, F. D. Lemkey

From a conszderation of the geometrically possible ways in which an array of lines or linear features in three-dimensional space can depart from a statistically random arrangement, a system was developed to describe and measure different types of anisotrypic arrangements of lines in opaque solids. The technique was applied to a specimen of unidirectionally solidified Al-Cwll, eutectic which contained microstructural imperfections termed lamellar faults. Experimental data obtained from photomicrographs of different facets of a grain of this alloy were then analyzed using this system. It was deduced that the defects, which are essentially linear in natuve, were predominantly parallel to one another but not parallel to the nominal growth direction probably because of some lateral heat flow during solidification. InGOTS of lamellar eutectic alloys which have been unidirectionally solidified under growth conditions which prevent colony formation consist of large elongated grains in which the lamellae are approximately parallel to one another. Deviations from a parallel lamellar structure within a grain are caused by extra lamellae which form occasionally during the solidification process. Three-dimensional analyses of such specimens have revealed that one extra platelet produces two defects, termed 1amellar faults, which can and do grow more or less independently of one another.&apos; The faults are analogous in many ways to edge dislocations in crystals. If a lamella of one phase is considered to be basically two-dimensional, e.g., a surface, then a fault can be defined as the locus of points which delineates an edge (or half the periphery) of one extra lamella as it grows and forms bridges and saddles in an otherwise parallel arrangement oi lamellae, Fig. 1. Were it not for the fact that bridges and saddles occur in the structure, the extra platelet could be defined as the unit defect and the concept of linear faults of positive and negative character around the periphery of the extra platelet would be unnecessary. If an extra platelet is merely inserted (actually it is grown in) between other parallel platelets, the fault line around the periphery of the extra platelet would circumscribe one and only one platelet, namely the extra one, and the faults would lie in a plane parallel to the lamellae. But when an extra platelet merges with adjacent ones and each side grows independently of the other, it is no longer topologically possible to delineate "the extra plate." A fault circuit which either closes on itself or passes out of the specimen still occurs although it no longer will be coplanar with the lamellae. In fact, it is usually not even Planar even if the lamellae are. Under these conditions a concept that a specimen contains N extra platelets of a given size per unit volume would be very artificial and consequently meaningless. Therefore, because of the topological complexity of the alloys, it was deemed necessary to define one-dimensional

Jan 1, 1962

By Stanley L. Lopata, Eric B. Kula

The problem of preferred orientation has been considered in quantitative phase analysis by X-ray diffraction techniques. The average intensity of a diffraction peak can be obtained by integration over the whole pole figure, or by random rotation of the specimen in the X-ray beam, or by a combination of both. For rod or wire, where the pole figure shows rotational symmetry, the intensity is measured along a radius of the pole figure. Where no symmetry is present, rotational symmetry is imparted by rapid rotation of the sheet around its surface normal during measurement. These methods are applied to the determination of aus-tenite in an Fe-31 pct Ni rod and Fe-28 pct Ni sheet. X-RAY diffraction techniques are frequently used for quantitatively determining the relative amounts of two or more phases present in a structure.1,2 Briefly, the method requires a measurement of the intensity of a Bragg reflection for each phase. The method is based on the principle that the intensity of a Bragg reflection, IA, is directly proportional to the volume fraction of the given phase present, CA: T KRACA/2µ [1] where K is a constant independent of the diffracting phase, RA can be calculated and depends on the Bragg angle ? and the particular reflection measured, and µ is the linear absorption coefficient of the mixture.&apos; For a mixture of two phases A and B, IA = RAcA [2] IA/IB=RAcA/Rbcb and CA+ CB = 1 [3] This allows the volume fraction of a phase to be calculated directly from the measured integrated intensities of the two phases present. This method is frequently used for quantitatively determining retained austenite in steels, but can be used for the quantitative determination of in titanium alloys, as well as phases in other alloy systems. The intensities are generally measured by diffracting only from planes whose normals are perpendicular to the specimen surface, with the implicit assumption being made that this intensity is typical of reflections from planes whose normals lie at other angles to the surface. If the specimen contains a texture or preferred orientation of the crys- tallographic planes by virtue of previous mechanical working, however, this measured intensity is correct only if the intensity level fortuitously is one times random. Examination of pole figures for actual rolled sheet shows that the intensity level at the center of the figure is seldom at the random level, and can very easily differ by a factor of five or more, leading to a commensurate error in the calculated amount of the phases. Textures have been commonly found in materials which are cold-rolled or recrystallized after cold rolling. This problem is becoming of greater importance since structural materials in use today, such as the high-strength stainless steels, frequently are further strengthened by cold rolling, and hence contain textures. Although high-strength carbon-containing alloy steels generally do not contain textures, some of the newer thermomechani-cal treatments will impart a texture even in these cases. In the case of hexagonal metals, such as titanium, probably every wrought sample will show some anisotropy. Since the standard X-ray techniques are inapplicable in all these cases, there is a clear need for some method of taking this anisotropy into account. The specific problem is to obtain the integrated intensity of a Bragg reflection for each phase, averaged over all orientations of the sample with respect to the X-ray beam. This can be done either by purely mechanical means or by a mathematical averaging. Mechanically, a small spherical speci-

Jan 1, 1965