Search Documents

By H. L. Luo, P. Duwez, C. C. Chao

BY rapidly cooling alloys from the liquid state, it is possible to obtain solid solutions beyond the equilibrium concentrations, provided that the components are miscible in the liquid state. Typical such cases include CU-Ag,1 Pt-Ag,2 CU-Rh,3 nickel-base alloys with silicon, germanium, and tin,4 and cobalt-base alloys with silicon, germanium, tin, aluminum, and gallium.5 In the present note it is shown that terminal solid solutions can be appreciably extended in A1-Mg alloys. The alloys whose compositions are given in Table I were prepared by induction melting in tantalum crucibles under an argon atmosphere. Both aluminum and magnesium were 99.99 pct pure. Weight losses after melting were less than 0.2 pct for the aluminum-rich alloys and less than 0.4 pct for the magnesium-rich alloys. From these weight losses it was estimated that the compositions of the ingots after melting were accurate within 0.3 at. pct for aluminum-rich alloys and within 0.5 at. pct for magnesium-rich alloys. As indicated in Table I, the compound AuCd10 in which there is an ionic contribution to the bonding. The authors thank Dr. J. H. Westbrook of the Research Laboratory, General Electric Co., for arranging the supply of materials and for his interest. This investigation was supported by the U.S. Atomic Energy Commission under Contract AT(30-1)-1002, Scope I. some of the alloys were also chemically analyzed and the results confirmed the limits of uncertainty based on weight losses. Rapid quenching from the liquid state was achieved by techniques previously described.8,7 The principle of these techniques consists of propelling a small globule of liquid alloy (about 20 mg) onto a copper substrate. The thin foils easily detached from the copper substrate were analyzed by X-ray diffraction using a Debye-Scherrer camera 114.6 mm in diameter and either nickel-filtered copper Ka or iron-filtered cobalt Ka radiation. Since the back-reflection doublets were resolved, the wavelengths used in the computatioas were CuKal = 1.54050A and CoKa1= 1.78892A. Lattice parameters were obtained by extrapolation of the Nelson-Riley function, and the tables compiled

Jan 1, 1964

By S. A. Mersol, C. T. Lynch, F. W. Vahldiek

The room-temperature Knoop hardness of single -crystal titanium diboride has been determined. Variations in microhardness have been correlated with orientation and structural imperfections in the crystals. Maximum differences in hardness values of 750 kg per sq mm with respect to orientation and 450 kg per sq mm with respect to substructure were found. The effects of etchants, load on indenter, and degvee of cracking of indentation were studied. The average hardness was determined as 2800 kg per sq mm on the as-grown crystals. Annealing increased the average hardness to 3375 kg per sq mm. ThE microhardness of polycrystalline titanium diboride approximating the stoichiometric composition TiB, has been studied by several investigators. The values for the Knoop hardness number range from 2710 kg per sq mm (Khn = 2710 kg per sq mm) to 3400 kg per sq mm.'-4 Samsonov et al.' reported a Vickers hardness number of 3400 kg per sq mm (Vhn = 3400 kg per sq mm) at 120-g loads for a 98 pct Ti + B material of 96 pct theoretical density. Recently Dunegan measured the Vickers hardness and obtained 2060 kg per sq mm with a 10-g load (Vhnlo = 2060 kg per sq mm), 1425 kg per sq mm for Vhnloo, and 1320 kg per sq mm for Vhnlooo. We have obtained a Khnloo = 3000 * 350 kg per sq mm for hot-pressed polycrystalline TiB2 that was 99.1 pct Ti + B and 96.6 pct of theoretical density. The present investigation has been made on single-crystal titanium diboride boules, and is believed to be the first such study on TiB, crystals of known orientation. EXPERIMENTAL The single crystals were prepared by a Verneuil-type process using an electric arc by the Linde Division of Union Carbide Corp. The largest specimens were 6 mm in diameter by 12 mm long. The crystals were of theoretical density (4.50 g per cm3) with a Ti + B content of 99.8 pct. The major impurities present were: 0.03 pct (by weight) 0, 0.06 pct C, 0.01 pct Si, 0.05 pct Fe, 0.01 pct Cr, and 0.01 pct Ni. The crystals were boron-rich, the average boron content being 31.S7 wt pct (stoichiometric value is 31.12 wt pct), and titanium-poor, the average titanium content being 68.2 wt pct (stoichiometric value is 68.88 wt pct). Titanium was determined gravimetrically by cupferron precipitation and boron by titration of H3B03 from a NaOH-fused specimen after removal of titanium with BaC03. The outer edge of the boule accounted for some of the excess boron. An electron-microprobe analysis showed that the titanium content increased from 60.0 wt pct at the edge to 68.5 wt pct at D, ~ 1000 p, where D, = distance from outer edge (see Fig. 5). X-ray measurements indicate that the boules are single crystals with considerable substructure. Upon etching this structure is seen in the form of a Widmanstatten-type rhombohedral line pattern in the basal (c) plane and a parallel "lines" pattern in the prism (a) planes. A typical WidmanstHtten-type pattern is seen in in Fig. 1. Electron microscopy reveals that the lines are less than 1 p across, which does not allow direct electron probe or microfocus X-ray determination of their composition. The identity of the second phase has not been established. It might be a higher boride or a nonstoichiometric TiB2. On annealing above 2200°C the precipitate is dispersed and dissolves in the matrix. The best etchants found were HF/HNO/HO, KF(CN)$NOH/HO (Murakami's Etch), and A large number of other etchants were previously Studied. The acidic etchant produced an acicular or "needles" pattern in the (a) planes which is an etchant effect on the "lines" pattern.

Jan 1, 1965

By Lawrence H. Van Vlack, Alfred S. Keh

The distribution of sulfur in iron was found to be dependent upon the time and temperature of the treatment as well as the chemical composition of the sulfide. With higher temperatures, the sulfide phase spreads more extensively between the iron grains. A complete network, however, does not form until approximately 1300°C (2372°F), ot which temperature the dihedral angle approaches zero. Silicon has little effect on this change. Aluminum, oxygen, and manganese all modify the temperature of sulfide network formation to a higher value. The microstructures which result and have significance in relation to hot-shortness are discussed. SULFIDE inclusions in steel were recognized by metallurgists before the turn of the century. Considerable attention has been paid to them, and numerous studies have been made on the working properties of steel as affected by them. It has been suggested that ho t-shortness of steel is due to the presence of an enveloping liquid sulfide film at the grain boundaries of the steel in the range of the hot-working temperatures. However, several facts have been observed in commercial steels which do not seem to be completely consistent with this explanation: 1) the frequent absence of continuous films of sulfide as observed under the microscope, 2) the absence of hot-shortness in some high sulfur and resulfurized steels, and 3) the limitation of hot-shortness to a definite temperature range. The role that manganese plays in steel to eliminate hot-shortness is well known, but the theory behind it is still under question. Moreover, the effect of de-oxidizers such as aluminum and silicon, and the effect of oxygen are still not clearly known. In this study, the effect of heat treatments and of additions of manganese, aluminum, silicon, and oxygen upon the distribution of sulfide inclusions in iron was investigated. Both the as-cast and the heat treated structures were studied, using microscopic and autoradiographic techniques. Part of the results were interpreted in terms of Smith's' concept of microstructure. Review of Literature Metallurgists used to consider sulfide inclusions as suspended particles. It was Wohrman' who first pointed out that sulfide inclusions are soluble in liquid iron. Many observed facts may be understood on the basis of this concept of solubility. Wohrman found that small castings contain small inclusions. and larger castings contain larger ones. Also, inclusion sizes are smaller near the chilled surface. For the same size ingot, the average inclusion size in a given heat of steel depends upon the rate of solidification. That is, with slow solidification, enough time for precipitation and agglomeration of these

Jan 1, 1957

By N. Brown, R. Kossowsky

Plain carbon steels with 0.48 and 0.95 pct C were quenched and tempered at 705°C to produce carbide dispersions with spacings on the order of 1 p. The morphology of the structure consisted of a carbide-dislocation network. The strengthening due to the dispersion was found to vary linearly with M-½ where M is the mean free-ferrite path determined by the entire network. No preyield microstrain preceded the upper yield point. After prestraining, the lowest stress at which dislocation movement could be detected was 104 psi; this frictional stress was independent of the dispersion and prestrains up to 6.5 pct. The Cottrell-Petch equation for grain-size strengthening was used to discuss the dispersion strengthening in this investigation. The results support an impurity mechanism for the upper yield point rather than one based on the Johns ton- Gilman theory. IT was pointed out by Gensamerl that the mean free path in the matrix is the important variable which controls the degree of dispersion strengthening. Gensamer's data on steels showed a linear relationship between the yield point and the logarithm of the mean free-ferrite path. The first theory was by Orowan,2 who suggested the mechanism of dislocations bowing between particles, with the resulting relationship that where a is the yield point, P is the distance between particles, and G is the shear modulus. The Orowan equation applies to that range of dispersion where P is large compared to the particle size and the particles are not coherent, so that the matrix is essentially free of internal stresses. As was pointed out by Orowan, there are different degrees of dispersion which, in turn, will influence the strength-controlling mechanism. However, in this investigation, we wish to confine ourselves to the region of coarse dispersion, where the Orowan mechanism should, intuitively, be applicable. It was soon evident that the data, which existed at about the time that the Orowan theory was proposed, did not agree with the Orowan equation in that the actual strengthening was always greater than the theoretical predictions. Thus, Fisher, Hart, and pry3 modified the Orowan theory by suggesting that the Orowan process took place in the microstrain region preceding macroscopic yielding and the subsequent, rapid work hardening in the form of residual dislocation loops around the particles largely determined the observed macroscopic yield point. The F-H-P theory stated that the increment of strengthening due to the work-hardening mechanism was proportional to f3/2 where f is the volume fraction of the precipitate. F-H-P used data by Shaw, Shepard, Starr, and Dorn4 on A1 3-5 pct Cu to support their theory. Roberts, Carruthers, and Averbach5 were the first to make a microstrain study of dispersion strengthening, and they found that for steel the Gensamer relationship was obeyed. Hayman and Nutting6 suggested that in the case of tempered steel 1) the ferrite grain boundary was the primary obstacle and 2) the carbide particles simply formed part of the grain boundary obstacles. They found that the strength varied as G"" where G is the ferrite grain size. Turkalo and LOW' determined the structure of quenched and tempered plain carbon steel using a replica technique; they concluded that the carbide particles did not necessarily lie in the ferrite grain boundaries. When they defined the mean free-ferrite path as being determined by both particles and the ferrite path boundaries, their data obeyed the Gensamer relationship. Meiklejohn and skoda8 showed that the strengthening from iron particles in mercury was a function of the particle size and the distance between particles. Dew-Hughes9 explained the Meiklejohn and Skoda results by the following theory: 1) grown-in dislocations which surrounded the particles were produced by the thermal stresses during the time the material was cooled to the testing temperature, and 2) the observed strengthening was associated with the cutting of the grown-in dislocations by the glide dislocations. Ansell and Lenel10 proposed a theory in which the glide dislocations pile-up or surround the particle until the number of dislocations in the pile-up is sufficient to yield or fracture the particle. The resulting theory says that the yield point varies as p-1/2, where P is the distance between particles. The Ansell-Lenel theory is almost identical to the

Jan 1, 1965

By J. H. Smith, H. W. Paxton

X HE recent investigation of Butler, McCabe, and paxton' on the vapor pressure of manganese gas in equilibrium with Mn-Fe alloys has been extended to include the iron-rich solutions and alloys in the two phase equilibrium region. The study of the phase equilibrium by Kurnakov and Troneva2 and the extensive reviews of the Mn-Fe system by Hansen,3 and Hellawe114 are not in agreement with regard to the location and extent of the equilibrium between the solid phases of fcc y and complex P in the temperature range above the eutectoid temperature of approximately 975Kwhere 0 decomposes to y and bcc a. An evaluation of the existing equilibrium data is confronted with the conflicting results presented by various investigators who used techniques requiring a quenching treatment to retain the high-temperature equilibrium or encountered temperature hysteresis during thermal analyses. The difficulty of suppressing the transformation of the manganese-rich concentrations of the y-solid solution during quenching has been well etablished.' To avoid this uncertainty the phase equilibrium was investigated at temperature by measuring the partial pressure of manganese gas in equilibrium

Jan 1, 1962

By R. A. Swalin, A. H. Geisler

THIS study represents an investigation of the effect of annealing temperature on the texture of drawn molybdenum, iron, and an iron alloy containing 14 pct Al. The effect of temperature on annealing textures has not been widely investigated in the past. One of the obstacles in making such a study using usual X-ray diffraction techniques results from the

Jan 1, 1957

By J. S. Kirkaldy, D. G. Fedak

The extra degree of freedonz introduced by a second independetzt concentration in a tenzary system gives rise to the possibility of unstable planar phase interfaces in semi-infinite diffision layer couples. Diffusion paths which pass continuously in and out of too-phase regions of the phase diagram but which are not coincident with tie-lines give rise to super saturated zones which may be either adjacent to or isolated from planar phase interfaces, with corresponding nonisolated or isolated precipitates. Nonisolated precipitates are equivalent to morphological development of the interface. Selected couples in the ternary systems Fe-Ni-Cr and Cu-Zn-Sn have been examined nzetallographically and ifz the latter case both isolated and uo~zisolated precipitates have been observed. The diffusion path in one of the couples showing interface morphological developtnent has been examitled using electron probe microanalysis. IGNORING the spurious effects of preferential grain boundary diffusion and anisotropic crystal growth7 it can be unequivocally stated that it is impossible for a binary isothermal diffusion couple constructed of two unsaturated phases to sustain a nonplanar phase interface. Consider for demonstrative purposes the schematic phase diagram of Fig. l(a). Given an initial flat phase interface, semi-infinite boundary conditions and the internal boundary condition fixed by the constitution diagram, xB = xB(T), the unique solution of the diffusion equation requires a persistent flat interface. No other solution is possible. If, however, we anticipate the involvement of surface tension, then the internal boundary condition must be written xB = xB (T,Y) where r is the radius of curvature of a hypothetical nonplanar interface. The introduction of a new parameter such as r can be equivalent to the release of an internal constraint, so that a stable nonplanar interface may be possible after all. Consider a perturbation to the interface as indicated in Fig. 1(b). According to the Ostwald-Freundlich equation: which is applicable to a system of this constitution, the solubility of A in the matrix ß ,xB (Y), at a point of curvature r is given by where Va is the molar volume of the a-phase, R is the gas constant, and a is the interfacial tension. Since the right hand side is positive, xB(r) >xB cm, as indicated in Figs. 1(a) and 1(c). Accordingly, the perturbation will carry the interface concentration in a direction which is opposite to that which would sustain local equilibrium. Such a perturbation is necessarily moderated and the system is returned to its initial condition. This result is general for all binary couples since we can regard an interfacial perturbation as an attempt to precipitate within un-

Jan 1, 1962

By J. E. Harris

By elimination of other possible nucleation processes, it has been demonstrated, for commercially pure magnesium and a Mg-Al alloy, that at stresses less than that necessary for triple-point cracking creep cavities are most probably nucleated heterogeneously at second-phase particles. Evidence for the existence of such particles is presented and the role of grain boundary sliding in the nucleation process is examined quantitatively. DURING creep deformation of polycrystalline magnesium and its dilute alloys in the temperature range 0.45 to 0.7 Tm (Tm = absolute melting temperature) cavities form mainly along grain boundaries oriented normal to the stress axis and failure eventually occurs by their growth and impingement.' The possibility of low-ductility creep failure of magnesium-alloy reactor cans during service has prompted a number of studies of cavity formation in these materials.2-5 In this paper divergent theories of cavity nucleation6-10 are examined critically and conclusions drawn regarding the most probable mechanism of formation of cavities in magnesium. The paper is divided into two parts. The first section consists of the review and quantitative assessment of the possible role of grain boundary sliding in the nucleation process. In the second part an experimental program is described where an electron-microscope technique has been used to determine if second-phase particles exist in as-received commercially pure magnesium and in a magnesium alloy 1) THEORETICAL REVIEW Possible nucleation sites for creep cavities are: a) high-angle grain boundary triple points,11-14 b) subboundary/grain boundary intersections,15-18 c) ledges in the grain boundary, 19-25 d) second-phase particles. 9,26-30 Each of these possibilities is now considered with special reference to deformation of magnesium. 1.1) Triple-Point Cracking. The stroh31 condition to nucleate fracture is: " > vr [1] where ss = applied shear stress, ? = surface energy, C = shear modulus, L = length of sliding interface. This relationship is independent of the nature of the interface and it is possible to calculate the stress concentration at the end of a sliding high-angle grain boundary and thus deduce, approximately,* the minimum applied stress necessary to nucleate triple-point cracking. Such theoretical predictions for Nimonic alloys have been made by McLean13 and found to agree quite closely with the minimum stress determined by experiment. Stresses capable of nucleating triple-point cracking should also be sufficient to nucleate fracture at a grain boundary due to intragranular slip, for the lengths of sliding interface in the two cases are approximately equal. Thus the minimum stress for triple-point cracking should represent a demarcation between failure due to general "cracking" and that due to cavitation. Hauser, Landon, and Dorn32 have demonstrated that for polycrystalline magnesium in the temperature range 78° to 298°K the fracture stress is inversely proportional to the square root of the mean grain diameter. As pointed out by stroh31 this result indicates that Eq. [I] describes the fracture characteristics of magnesium at low temperatures. Heal3 has reported that for rapid deformation (112 pct hr-1) of polycrystalline magnesium the transition from cracking to cavitation occurs in the temperature range 175" to 275°C. He also presented the fracture-stress values for the same material tested under similar conditions. From Eq. [I] taking G = 1.3 x 10" dynes , ? = 600 ergs . cm-2,33 and L = 0.017 cm, a Stroh-McLean13,31 tensile stress of -266 kg. cm-2 is calculated, corresponding to a transition temperature of -190°C; this is in

Jan 1, 1965

By E. Hornbogen

The deformation mechanisms of a iron-phosphorus alloys were compared at 20oand —196°C. Alloying with phosphorus increases the resistance to slip, but has little effect on the resistance to twinning. The stress for onset of twinning is nearly independent of temperature, but there is an increasing tendency for cleavage along (112) twin boundaries with decreasing temperature and increasing phosphorus content. FIVE iron-phosphorus alloys and one high-purity iron (0.004 pct C, 0.004 pct N, 0.05 pct O) were vacuum melted and cast, then homogenized for 6 hr at 1100oC and water-quenched. Single crystals were cleaved from the coarse grains produced in this manner. The hardness was measured* on polished (100) surfaces of the crystals at 20" and -196°C. Laue back-reflection patterns indicated a deviation of not more than 3 deg from coincidence of the {100} plane with the surface. The deformation mechanisms were studied by relief effects on the polished surfaces. Yield points in compression (0.5 pct offset) were measured with polycrystalline material. At 20°C the addition of phosphorus to a iron leads to a change from slip to twinning as the main deformation mechanism, Figs. 1(a), (b). Between 2 and 4 at. pct P slip and twinning occur simultaneously. In alloys with more than 4 at. pct P twinning is the main deformation mechanism and slip was observed only in the neighborhood of twin intersections. All cracks at 20oC were on (100) planes, some initiated by twins.' From measurements made at-196oC, hardness and yield points of all the alloys were nearly constant and the same as the hardness of high phosphorus alloys that deformed by twinning at 20°C, Figs. 2 and 3. In the high-purity iron, plastic deformation at -196°C occurred by both slip and twinning, but deformation in the iron-phosphorus alloys appeared to be entirely by twinning. In the high phosphorus alloys a large number of cracks formed along the (112) twin boundaries and at intersections of twins, in addition to the (100) cleavage, Fig. 1(c). DISCUSSION In , iron, plastic deformation can occur either by slip or by twinning, depending upon composition,

Jan 1, 1961

By D. W. Levinson, L. E. Tanner

DURING the course of an investigation of the zirconium-iron-tin system the zirconium-iron system was reexamined up to 55 pct* Fe. The alloys pre- visible melting3 and X-ray diffraction studies were made of samples of the alloys. The ferromagnetic properties of ZrFe2were determined by Professor P. W. Selwood* using a modified Faraday balance.4 The zirconium-rich portion of the phase diagram is presented in Fig. 1 and compared with the work of Hayes et al.1 The maximum solubility of iron in ß is somewhat decreased while the eutectic temperature is raised. There seemed to be no change in the eutectic composition of 16 pct Fe. Since the base material was so low in iron content, it was possible to establish the maximum solubility in a between 0.017 and 0.025 pct Fe at the eutectoid temperature. The observation of fine ZrFe, particles in the micro structure of a specimen of 0.012 pct Fe alloy quenched from 700°C indicated that the solubility drops below this composition at 700° C and lower. Further study was made of the intermediate phase ZrFe2. wallbaum5 identified the compound initially as ZrFe2 with a C 15 (MgCu2) structure, placing it at 55 pct Fe stoichiometrically. Hayes et al.1 believed they found the phase at about 47.2 pct Fe and called it Zr2Fe3 but reverted to ZrFe, in the discussion of their paper.6 Jordan and Duwez7 confirmed Wallbaum's work as well as finding the same single-phase structure to about 62.3 pct Fe or Zr0-61 Fe2-18. In order to establish the ternary phase relations in the zirconium-iron-tin systema the exact position of this phase had to be known. This was accomplished by X-ray diffraction, visible melting, and metallographic techniques. Elliott8 obtained a single-phase ZrFe2 X-ray diffraction pattern from an alloy of about 55 pct Fe. This was confirmed in the present study with an alloy of 54.81 pct Fe. However, extra lines (attributable to a or ß zirconium) were found in diffraction patterns of the 47.42 and 49.83 pct Fe alloys. The melting characteristics of these same alloys were studied by suspending 47.42, 49.83, and 54.81 pct

Jan 1, 1960

By H. P. Stüwe

HE recent papers of Aust and Rutter112 have again brought to mind the importance of certain "special" orientation relationships between the lattices of two fcc crystals on the mobility of the grain boundary separating them. No theoretical treatment of this problem has been offered till now. The only principle published so far that has given an indication why certain orientation relationships should be "special" is the counting of "coincidence sites" according to the suggestions by Kronberg and Wilson.3 The "coincidence plots" published by these authors show, for instance, that two fcc lattices rotated by 38 deg around a common (111)-axis have 1/7 of their lattice sites in common. Such an orientation relationship is also characteristic for high boundary mobility. The accuracy of the orientation relationship needed for high mobility is not too rigid. Deviations up to 10 deg from the ideal orientation seem to be tolerable as can be gathered from the experimental data.'j4 The auestion arises, then. how the density of coincidence sites will be affected by a deviatioi from mathematical accuracy. This can be checked by a simple model. Two sieves are prepared, the arrangement of the holes corresponding to lattice sites on a (111)-plane in a fcc crystal. They are placed upon each other, Fig. 1. One is kept fixed, the other is rotated around an axis pushed through one hole in common to both. The light falling through holes of both sieves is then a measure for the density of coincidence sites. If the diameter of the holes is very small, light intensity is contributed only by holes the centers of which coincide "exactly." If the diameter of the holes is rather large, some intensity will be contributed by

Jan 1, 1962

By J. P. Neumann, L. Raymond

At temperatures near 1000°C, only the (112) (111) slip systems are operatire in tungsten. Therefore, the substructure introduced into the billet upon rleforming a single crystal by rolling at 1000 "C can be controlled by proper selection 01. processing ral'iubles. Using a reduction of 10 pct pel- pass, the (001)[110] and the (110)[001] crystals will deform by duplex slip rip to a 50 pet reduction in thickuess. These rolled single-crystal sheels are characlerized by anisotropic hardening, retention of ductility with warm working, lattice bending which is related to the conlacl (angle, and macroscopic dimensional changecs in the length bid not the width of the billet. When several slip systems become operative, a greater degree of work hardening is itztvoducetl into the crystal. This rolling- condition is characlerized by isotropic hardening, loss of ductilily with warm working, complex lattice distortions, and macro-scopic dimensional changes in both the length rind Width of the billel. IN tungsten, the {1 10}(111) slip systems are operative below room temperature,' while above 700°C only the {112{ (111) slip systems are observed.2,3 In the transition region, i.e., room temperature. both systems appear. It is the purpose of this investigation to show that tungsten single crystals deformed by rolling at a temperature where only one slip system is operative generate dislocation substructure which depends on the crystal orientation. For a specific crystal orientation. the resulting dislocation substructure can be further influenced by modifying the rolling procedure. Crystals oriented for multiple slip display less substructural sensi- tivity to modifications in processing techniques than do single crystals oriented for duplex slip. Under simple slip modes of deformation, very little work hardening is introduced into a rolled single-crystal sheet. Mechanical properties such as room-temperature strength. ductility, and hardness are used to confirm the existence of the different work-hardened conditions generated within the crystal by modified warm-rolling techniques. EXPERIMENTAL PROCEDURE Material. Single-crystal billets of four different crystal orientations. grown by the Arc-Verneuil technique. were ground to a size of 0.25 by 0.375 by 3 in. In terms of their rolling plane and rolling direction, the billets had the following orientations: (1101 [001 ] (110) [110] (001) [110] (001) [100] The actual orientation of all crystals was within 5 deg of the ideal orientations. The typical impurity content of the crystals is given in Table I. Polishing and Etching. Polishing was done elec-trolytically in a 20 g per liter aqueous solution of NaOH at approximately 10 v. It was established by the Laue back-reflection X-ray technique that the surface damage due to grinding and cutting operations with a diamond wheel did not exceed 0.001 in. Therefore after each cutting or grinding operation at least 0.002 in. was removed electrolytically from the surface in order to eliminate surface effects. By lowering the voltage to 3 v, subgrain boundaries and individual dislocations are identified in agreement with Berlec 4 where the etch-pit traces on a cubic plane form a square. The etching time was about 5 sec at room temperature. Rolling. Rolling was carried out on a Loma 2-high mill (Model 600). The peripheral velocity of the 6-in.-diameter rolls was 500 in. per min, with the surface temperature of the rolls about 400° to 450° C The billets were heated before each pass for 30 min in argon to 1000 C. Oxidation of the crystals during rolling in air resulted in a thin layer of greenish-yellow WO 3 This oxide was removed after each pass by immersing the crystals for a few minutes in a boiling NaOH solution. Prior to the initial rolling pass, a surface layer of 0.001 in. was removed by polishing. The crystals were then annealed in vacuum (10-4 torr) at 2400 C for 30 min. All crystals were rolled in the same direction; i.e., the same end was always introduced into the rolls. The rolling program was conducted in

Jan 1, 1965

By F. N. Rhines, A. S. Nemy

The mode of high-temperature tertiary creep of 523-0 aluminum alloy was found to be strongly stress dependent. The occurrence of necking and/or fissures during tertiary creep exhibited a sequence with varying stress, wherein fissures with no necking occur at a minimum stress and necking with no fissures occur at a maximum stress. At an intermediate stress, tertiary creep begins without necking but with fissures in a region just beneath the external surface, followed by necking and the development of fissures along the axis of the specimen. The results are interpreted in terms of grain boundary gliding and the hydrostatic stresses produced by restraints to grain boundary gliding. THERE appear to be two possible reasons for the existence of a period of accelerating elongation prior to rupture during high-temperature creep in metals; namely, 1) a change in the structure of the metal itself, leading to a change in its response to loading, and 2) a reduction in the cross section of the metal resulting in a higher unit loading. Howe,' by annealing and retesting broken low-temperature creep specimens, found that irreparable damage is done once a metal is subjected to creep. Andrade2 first pointed out the association between the reduction in cross section and the increase in creep rate. The present research was designed to investigate both of these factors by retesting specimens machined from material which had been subjected to high-temperature creep testing short of rupture. EXPERIMENTAL METHOD An aluminum alloy, 52s-0, was selected for these studies because it is constitutionally simple, is strong enough to be relatively insensitive to mechanical damage during machining,3 and undergoes accelerating creep in a convenient temperature range. This alloy is hardened by 2.6 wt pct Mg in solid solution and is not subject to overaging effects; it has a very small quantity of an insoluble impurity phase, Mg2Si, finely dispersed throughout its micro-structure. Creep specimens, similar in design to ASTM 0.252-in. tensile specimens, were machined from 3/4-in. round stock and the surfaces were electro-polished. A 24-hr anneal at 500°C was then applied to stabilize the grain size to an average diameier of 0.27 mm. Subsequent heating for as long as 700 hr at the intended test temperatures produced no change in the tensile properties, indicating a well-stabilized state of the material. Corresponding to each of the typical testing conditions listed in Table I, one creep specimen was tested to destruction and this test was employed as a guide in the selection of times for subsequent tests. Duplicate creep tests were then run to various points along the creep curve to produce specimens which were used to determine changes in mechanical properties and microstructures. The latter were examined at the midsections of the specimens. Other specimens were tested to various stages of tertiary creep, after which they were remachined to new cylindrical test specimens, electropolished and re-tested in creep. RESULTS The major findings of this investigation may be illustrated by a comparison of the results of three sets of creep tests. These were run at 400°C at three stress levels, namely 750, 1000, and 1500 psi. In presenting these results in Figs. 1, 5, and 8, the creep curves are given in pairs: That curve of each pair which is drawn through filled circles corresponds to the initial test, and the curve drawn through open circles (and designated by - R) represents the result of retesting the same specimen after machining to a new cylindrical surface. The load was readjusted to the original unit stress at the beginning of each retest.

Jan 1, 1960

By Dogan E. Gücer

In the following note, with the help of a new parameter of strain hardening, the attention is drawn to the close relationship between the relative performances of steels under plastic range cycling and their strain hardening characteristics. If a cantilever beam of rectangular cross section is tested with the arrangement shown in Fig. 1 in reversed bending between constant deflection limits at the lever head, the longitudinal strain-deflection hysteresis shown in Fig. 2 is obtained. There is no lateral strain due to constraint at the rectangular test section.' If the absolute value of strain undergone between a maximum state of tension and maximum state of compression is called A, it will be seen that A decreases with cycling rapidly due to the strain hardening in the test section and the ensuing change in the bent beam profile. It will also be seen that A reaches an approximately stable value around 10th cycle. If we call the first cycle strain range, A,, and that of 10th cycle, A,,, then is an index of the strain hardening capacity at a particular deflection level. A plot of n vs ? is reproduced in Fig. 3 for two ASTM steels: A-302 and A-201. A-302 is a low alloy structural steel, A-201 a low-carbon steel. The specimens were first normalized at 1650°F for 75 min then grooved with a milling cutter; and the notches were ground to 50G finish. The final operation before testing was a stress relief at 1150°F for 75 min. The behavior of the two steels changes around ?1 = 0.8 pct or A,, = A, -n = 0.7. If these curves are compared with the endurance life curves given in Fig. 4, it will be seen that the relative performance of the two steels also change around A,, =. 0.70 pct. The steel with lower strain hardening capacity (7) at a certain strain level shows a longer endurance life, irrespective of their strength levels. Since the exponent governs strain hardening at lower strain levels than the Modulus, the relative values of both parameters also coincide with the change in n values and the relative performances of the two steels at different strain levels. These two observations indicate to the close relationship be-

Jan 1, 1962

By B. D. Cullity, S. S. Hsu

The stress distribution at the surface of a twisted cylinder is analyzed along the boundary of a slip plane of arbitrary orientation and this analysis is applied to the torsion of cylindrical crystals of magnesium. A criterion for slip is developed and the critical resolved shear stress of magnesium in torsion is found to be 0.038 kg per sq mm. The mechanism of torsional deformation in magnesium and aluminum is described, as well as the relaxation of a twisted crystal at elevated temperatures. THERE has been considerable fundamental study A of the mechanism of slip occurring during deformation in tension and compression but comparatively little attention has been paid to the mechanism of torsional deformation. This paper is concerned with the way in which a single crystal deforms in static torsion and with the phenomenon of untwisting which a twisted crystal undergoes during annealing. Particular attention is paid to the application of the critical resolved shear stress law to torsional deformation. Probably the most likely mode of torsional deformation that can be imagined consists of rotation of one part of the crystal relative to another on the active slip plane. Here the most interesting feature is the slip direction. If a cylindrical single crystal is twisted about its axis and if, to consider the simplest case, the active slip plane is normal to the specimen axis, then the direction of gross slip at any point must always be at right angles to the radius drawn to that point. It follows that the crystallographic direction of slip must be continually changing, in sharp contrast to slip in tension which occurs in a crystallographically constant direction. Extensive investigations of slip in torsion have been made by Gough and his associates.' They made torsional fatigue tests on metal single crystals belonging to a number of different crystal systems and observed, with a microscope, the appearance of slip lines on the surface and of strain markings on the transverse cross section of cylindrical specimens. They concluded that the critical resolved shear stress law was just as valid in torsion as in tension, in the sense that slip occurred on that plane and in that direction in which the resolved shear stress was a maximum. They also established that the operative slip planes and directions were the same as in tension. They did not, however, measure the critical resolved shear stress in torsion nor, in fact, did they establish that a critical value of the resolved shear stress was necessary for the initiation of slip. Their investigations were, moreover, restricted to torsional fatigue tests which necessarily involve small strain amplitudes. Wilman,' studying crystal growth and the abraded surfaces of crystals, observed crystal fragments

Jan 1, 1955

By R. L. Fullman, J. W. Cahn

A method is derived for obtaining the distribution of sphere diameters and plate thicknesses (or pearlite spacings) from size distribution functions obtained along randomly oriented lines. The method described offers several advantages over existing methods. IT is often of interest to measure not only the average particle size but also the distribution in sizes of particles dispersed in an opaque material, such as an alloy. Particles of a single size will exhibit a range of apparent sizes on the plane of polish. If the sample contains a distribution of particle sizes, the distribution of observed sizes on the plane of polish is related to the distribution of actual particle sizes by an integral equation, For example, a distribution of sphere sizes, N(r),

Jan 1, 1957

By R. S. Craig, W. E. Wallace, G. S. Kamath

The destruction of long-range order in Mg-Cd, has previously been thought to occur as a second order process. In the present work a variety of X-ray diffraction techniques are employed to show that in this case and in Mg-Cd alloys containing frorn 20 to 33 1/3 at. pct Mg the process more nearly resembles an ordinary first order phase transformation. It is found that the ordered alloy, which occurs as a single phase at lgw temperatures, transforms with increasing temperature into a Cd-enriched disordered (P) phase and an Mg-enriched ordered (a) phase. The interval over which these two phases coexist under apparently equilibrium conditiovs extends from 25" to 87°C content. At 33 1/3 at. pct Mg the two-phase region has disappeared. Measurement of the long range order parameter, s, for the ordered phase in MgCd, (from superlattice line intensities) shows that it disorders only very slightly before it is transformed into the p form. Kinetic experiments reveal that when the a phase is formed frorn the p phase by suitably quenching so as to suppress phase segregation, the transformation is essentially complete and the a phase has acquired its equilibrium value of S in less than 1 min. In an earlier investigation carried out in this Laboratory1 the rate of disordering of the MgCd3 superlattice was determined over an extensive range of temperature. The process was found to be first order and the rate constant, k, was established. At the lowest temperatures studied k was found to increase exponentially with temperature. At higher temperatures, however, it passed through a maximum and for the 100 deg interval below the point at which the superlattice was destroyed k was observed to decrease with increasing temperature. The various theoretical treatments1-3 of order-disorder phenomena lead to a negative temperature coefficient for k near the superlattice critical temperature but this region for MgCdB seemed unusually wide. Detailed analysis of the point indicated' that either the long-range order parameter, S, for MgCda varied with temperature in an anomalous way or the disordering in this instance is more complex than in the cases which have been treated theoretically. The original objective in the present work was simply to determine S vs temperature for MgCda to ascertain whether the disordering was in conformity with theoretical expectations.2° In an early stage of the work it was noted, however, that under equilibrium conditions ordered (a) and disordered (0) phases coexist in MgCd, over a considerable range of temperature. Since on thermo- dynamic grounds the coexisting phases must differ in composition, the investigation was enlarged to permit the compositions of the two phases in MgCd3 to be determined and to establish the extent of the two phase regions for alloys containing from 20 to 33 1/3 at. pct Cd.

Jan 1, 1963

By R. H. Hiltz, R. W. Douglass

Large-grain specimens of iodide titanium prepared metal-lographically were stain etched using the technique of New York University as modified by Watertown Arsenal Laboratories. Orientations of grains showing an abnormal response to stain etching were determined by back-reflection X-ray techniques. X-ray results showed that a-titanium grains, oriented such that their basal planes were parallel to the polished surface, have a slower reactivity to stain etching than any other orientation. Grains oriented such that their basal planes made small angles with the polished surface exhibited reaction rates between that of the true basal orientation and the normal rate. As the angle with the surface increases the reactivity increases up to a maximum angle of 14 deg. Basal planes making an angle with the surface of greater than 14 deg show normal reactivity to stain etching. These data, plus work on polycrystalline specimens, indicate that the oxidation rate of titanium might be reduced by correct alignment of crystal planes. ThE earliest observed phenomenon in the micrography of titanium was the occurrence of a grains which exhibited no birefringency. Even though a titanium possesses a hexagonal-close-packed structure and should be optically anisotropic, it can appear optically isotropic, when the crystal is so oriented that its intersection with the polished surface is perpendicular to the axis of symmetry. During the develoment of color staining techniques for titanium: it was noted that the a phase did not stain uniformly. Staining of specimens of commercially pure titanium resulted in a micro-structure of blue-colored grains, among which were randomly oriented a few grains exhibiting colors between the initial yellow and the final blue color. Under polarized light, these off-colored grains appeared to be optically isotropic. More intense study with polarized light, however, showed only the yellow grains to be isotropic. The intermediate colored grains exhibited weak birefringency, which at first glance is not detectable due to the strong birefringency of the surrounding structure. Aminoff2 and Straumanis3 have previously shown that in zinc the basal orientation is less reactive than other orientations. Leidheiser4 showed the same thing to be true for the hexagonal-close -packed configuration of cobalt. Thus, it would appear that the basal-plane orientation is responsible for the abnormality in stained titanium. Later work by Leidheiser5 on cobalt has shown that other planes besides the basal plane exhibited reaction rates slower than the normal rate. Reported planes were the (100), the (1011) and several planes making small angles with the basal plane. The last observance appears not to be due to the reactivity of the reported plane but due to the reactivity of the basal plane which is very close to being parallel to the surface. There are, then, two possible explanations for the range of colors observed in stain-etched a titanium. Since extensive research on the effect of orientation on surface reaction has failed to provide any definite trend, these two explanations are not adequate in themselves. Planes showing the retarded reactivity differ not only for different crystal structures but for metals of the same crystal structure and even for different reactions of the same metal. Tragert,6 Leidheiser,7 Tammann,6 and wagner9 have shown copper to exhibit retarded reactivity on the (111) plane. However, Leidheiser10 has shown that under other conditions the retarded reactivity of copper is associated with the (100)

Jan 1, 1960

By P. R. Gage, R. W. Bartlett

At high temperatues and reduced oxygen pressuves, molybdenum silicicles oxidize to form SiO(g) vathev than a passivating SiO2 film. This is a sevious problem for low-pressure applications of sili-cide-coated rejvactory tnetals. As oxiclalion pvoceeds, the higher oxidntion potential of silicon initially suppresses oxidation of molybdenum and silicon is depleted near the surface. Eventually, silicon depletion causes secondary passivation (SiO2) mid an active to passiue transition in oxidntion behavior at moderately low pressures. At very low pressures, silicon depletion continues until steady-state oxidation of both silicon and rnolyhdenum can proceed at rates propovtional to their initial stoicliiorrzetvies. Ohsel-vations of 'oxidatiou-rate belzcriliov at temperatures from 1101° to 1700°C and oxygen pressures from 10-7 to 1 at in were made using thermogravimetn'c, X-ray diffraction, and elect?*on-rtric?ro/,vobe techniqrres. The experimental results are in good agreement with calculated lirnits for the various types of behavior that were based on tlievtnochernicnl atzd dffusion rate data for the Mo-Si system. A protective oxide film normally forms during the oxidation of elemental silicon. However, at elevated temperatures and reduced pressures this film is not present and oxidation is rapid. wagnerl has shown that this behavior occurs because SiO(g) is more stable than SiO2. The behavior of the silicides at high temperatures is similar but more complicated. Molybdenum and several silicide compounds are involved and the chemical activity of silicon, or the equivalent silicon vapor pressure, can vary considerably in these phases. Also, diffusion in the silicides resulting from reactions in which a higher silicide is reduced to a lower silicide, or molybdenum containing dissolved silicon, must be considered. EXPERIMENTAL AND ANALYTICAL METHODS Isothermal oxidation experiments were conducted using hot-pressed MoSi2 and Mo5Si3 wafers. Each experiment was conducted at a fixed oxygen pressure by controlling the flow of oxygen in and out of the furnace chamber. Most of the experiments were conducted in a molybdenum-wound tube furnace. Special construction features, including use of a high-purity dense A12O3 refractory tube, permitted operation in high vacuum to 1500°C. Higher temperatures could be obtained at higher working pressures and with a sacrifice of tube lifetime. Temperature was measured with three Pt 6 pct Rh/Pt 30 pct Rh thermocouples and controlled with a proportional controller. Weight changes were monitored with a quartz spring balance. The few experiments, at higher temperatures and low pressures, were conducted in a cold wall vacuum chamber using induction heating with direct susceptance of the sample. Because of the variable levitation effect of the induction coil on the sample, weight changes could not be monitored continuously in these experiments. All samples representing both experimental methods were weighed before and after oxidiation and examined with a microscope and analyzed by X-ray diffraction. The analytical treatment assumes equilibrium at interfaces and uses existing thermodynamic data. For convenience, equilibria are expressed as chemical potentials or as vapor pressures of the gases involved. The resulting differences in equilibrium vapor pressure between adjacent interfaces and between the surface and external atmosphere drive diffusion processes in the solid and gas boundary film, respectively. These diffusion processes govern the rate of oxidation. A bibliography of the sources of thermochemical data for the reactions pertinent to the Mo-Si-O system is given in Table I. Reactions are listed with the sources of free-energy data referenced. Since all of these reactions involve one or more gas species, usually oxygen, a plot of the product-gas chemical potential, -RT In Pg, vs absolute temperature for each reaction is given in Fig. 1. These curves can be used to determine equilibrium vapor pressures or to compare the stability of different compounds in the various gases: O2, Si, SiO, and Moo3,. The data of Table I and Fig. 1 were used to determine the analytical results presented in this paper. EXPERIMENTAL RESULTS The weight changes for three MoSi, samples oxidized at 1500°C are plotted against time in Fig. 2. The treatments were identical except for the indicated variance of oxygen pressure and each curve represents one of three characteristic types of behavior that were observed. The upper curve is an example of passive oxidation, which is defined as

Jan 1, 1965

By M. van Swaay, C. E. Birchenall

Palladium has a large capacity for the dissolution or occlusion of hydrogen; the gas also diffuses very rapidly through the metal. Palladium thimbles are widely used in the laboratory for purification of hydrogen by diffusion. The use of palladium filters could be extended to several large-scale applications if greater stability of the permeation process could be attained. In this paper a number of aspects contributing toward this goal are discussed, and new data on solubility and diffusivity are given. Palladium sheets of 0.005, 0.010, or 0.020 in. thickness were prepared by three procedures: casting and rolling, sintering palladium powder, and sintering palladium powder to which 0.1 pct thorium oxide had been added. Half-inch diam circular specimens were cut and clamped in a stainless steel holder provided with knife edges to seal the palladium sheet with soft copper gaskets. The seals were tested for leaks at l0-5 mm Hg. No difficulties from this source were encountered even after many temperature cycles in the system. The holder was silver-soldered to Kovar-glass seals and attached to the system through glass ball joints. Each joint was sealed with Apiezon wax and was adjacent to a cold trap. In several experiments the holders were fused directly into the system, eliminating the ball joints, to show that the wax did not affect the filter. Three filter holders were supported in holes in a stainless steel cylinder, which also acted as a temperature equalizing block inside the resistance wound furnace. The measuring thermocouple was inserted in one of the holders. A potentiometric controller with control thermocouple on the furnace heater winding maintained ± 0.2oC. Electrolytic hydrogen was purified and stored as shown in the flow diagram, Fig. 1. It could be admitted to either or both sides of the palladium specimens. Only a mechanical pump was used, for even high pumping speeds would not bring the pressure on the off-gas side of the filter below 10 to 50p Hg owing to the rapid hydrogen flow through the filters. A series of steady-state permeability studies allowed identification of several conditions of permeability loss. A constant fore pressure was maintained while the low-pressure face was pumped. After enough time had elapsed for a steady flow to be established the pump was cut off, and the rate of pressure increase in a thermostated bulb was measured to determine the flow rate. The time delay method of Daynes1,9 was adopted since it offered the possibility of determining solubilities and diffusivities. For each measurement both sides of the filter were evacuated for 15 min, which gave about 10-4mm Hg residual pressure. Hydrogen was admitted to one side or the other of the filter at a fixed pressure. The pressure on the off-gas side was observed as a function of time. A

Jan 1, 1961