Search Documents

By J. H. Jackson, P. D. Frost, C. H. Lorig, A. C. Loonam

THIS paper describes (1) the effect of sodium on the tensile ductility of magnesium-lithium base alloys, and (2) the precautions necessary to avoid sodium contamination. Effect of Sodium on Properties The harmful effect of sodium in magnesium-lithium base alloys has been recognized by other investigators.1,2 However, quantitative data on its effect have not hitherto been published. The effect of sodium on the ductility of an 87 pct Mg-9 pct Li-4 pct Zn alloy is shown in fig. 1. The sodium was introduced by adding sodium chloride to the flux, which consisted of 50 pct LiC1-50 pct KC1. Moderately high ductility was obtained at an indicated sodium content of about 0.05 pct. Actually, the value probably was less than 0.05 pct, be-cause analyses of the lithium, which were more accurate than those of the alloy, indicated that the highest ductility was obtained when its sodium content was 0.02 pct or less. In fact, in alloys having somewhat greater strength than this particular composition, the sodium content of the lithium must be under 0.01 pct to avoid embrittlement. The effect of sodium on the ductility of several alloys having fairly low strength levels is shown in fig. 2. These alloys were melted with a flux containing 75 pct LiC1-25 pct LiF, to which various quantities of sodium chloride had been added. Sodium had no significant effect in pure magnesium but markedly decreased the ductility of the magnesium-lithium-zinc alloy, even though the alloy contained only 3 pct Li. As the lithium content of the binary alloys was increased, the ductility of the sodium-free alloys also increased. However, the effect of sodium on ductility became progressively more pronounced. Sodium raised the minimum hot-rolling temperatures, particularly those of the high-strength alloys, but its effect, in this respect, was less critical than its effect on tensile ductility. For optimum cold roll-ability, a low-sodium content was required. No deleterious effect of sodium on the stability of properties at slightly elevated temperatures was found. Control of Sodium in the Alloys Optimum ductility is obtained in the magnesium-lithium base alloys when sodium-free materials are used. Commercial lithium was formerly the chief source of contamination. Now purified lithium is usually available. When lithium of ordinary purity is used, the effects of sodium can be minimized by using a flux consisting of 75 pct LiC1-25 pct LiF. Bubbling nitrogen through the melt also helps remove sodium. The damaging influence of sodium on the ductility of the magnesium-lithium base alloys was observed very early in the alloy development program. The discovery came about as a result of chemical analyses made on the lithium used in the alloys. Ductile alloys had been made from lithium containing about 0.2 pct Na, whereas the lithium used in the brittle alloys contained about 0.9 pct Na. The first magnesium-lithium alloys were melted under commercial fluxes ordinarily used in melting magnesium alloys. These fluxes were unsuited for scavenging the contaminating sodium, and the resulting alloys were not ductile. Next, a flux containing 50 pct KC1-50 pct LiCl was tried. This was somewhat more effective but still was not satisfactory. After a number of different experimental fluxes had been investigated, one containing 75 pct LiC1-25 pct LiF was adopted. This flux made it possible to obtain good ductility in the alloys. The

Jan 1, 1951

By R. I. Garrod, R. A. Coyle

The shapes and positions of X-ray reflections from specimens of copper, steel, and aluminum alloy haue been examined in the elastic and plastic ranges both while the specimen was under stress and in the unloaded condition. For the aluminum alloy the shape was unaltered by the application of stress either within the elastic limit or in the plastic range provided that no additional plastic strain was induced. In copper the broadening accompanying plastic deformation was very slightly reduced when the specimen was unloaded. A similay but more marked elastic component of broadening was also found for steel, but in this case below the yield stress. Line profiles corrected for instrumental and particle-size broadening indicate very large internal stresses in local regions of the plastically deformed metals. The results are discussed in terms of a recent suggestion that the heterogeneous dislocation distribution between the cells and their boundary walls plays a major role in the peak shifts and broadening of the X-ray reflections. STUDIES of the X-ray line profiles from strained polycrystalline aggregates concentrate usually on one or the other of two main parameters: a) the displacement of the peak of the intensity contour from its position for a strain-free aggregate, or b) the shape of the profile. From peak shifts data can be obtained either on the relation in both the elastic and plastic ranges between applied external stress and average lattice strains in a given (hkl) direction, or, alternatively, on the residual lattice strains which are present after a plastically deformed specimen is unloaded.' On the other hand, the shapes of the broadened profiles from cold-worked metals can be analyzed to separate the broadening produced by small particle size and by heterogeneous lattice strains.' In this paper the terms "size broadening" and "strain broadening'' are used in the general sense adopted by warren.' In the past, apart from two early qualitative observation, it has been customary to examine only the movements of the peaks of the profiles while the specimen is actually under load, since the line broadening induced by plastic strain remains after removal of the external stress. Consideration of the implications of existing data of this type suggests, however, that fruitful additional information on a number of fundamental aspects might be gained by careful examination of whether the X-ray line profile is in fact different in the loaded and unloaded states of the specimen. By taking advantage of the sensitivity and convenience of modern diffractometer techniques it is possible to explore with relative ease the magnitude and importance of any elastic effects which may be superimposed upon the well-known permanent changes in profile. The main aim of the work to be described was thus to investigate this point for typical metals and alloys. For this purpose annealed specimens were extended first elastically and then plastically and the positions and shapes of X-ray reflections were recorded. Initially it was anticipated that prime interest would center on observations within the plastic range; it has been found, however, that small changes in profile sometimes occur both before and after the nominal elastic limit of the material is reached. It is shown that the results obtained have important implications in relation to the structural changes and processes associated with deformation. I) EXPERIMENTAL To enable the diffraction lines to be recorded while the specimen was under uniaxial-tensile stress, a small hydraulic testing machine was designed and constructed for direct attachment to the goniometer of a Philips diffractometer. The specimens, which were machined from 1/2-in.-diam rod and had a central rectangular section 3/8 by 1/16 in. over a gage length of 1 in., were held in the machine by split collets mounted in grooves in the cylindrical ends of each specimen. No special precautions were taken to ensure precise axiality of loading. Constant oil pressure was maintained by a lever and weights system and transmitted to the loading rig by flexible pipe. The actual load on the specimen was measured by a load cell in the machine to an accuracy of * 1 pct. To enable smooth X-ray profiles to be obtained the specimen and machine were oscillated continuously during recording through *7-1/2 deg about the normal half-angle position of the goniometer. The three materials chosen for the investigations were high-purity copper as representative of a ductile fcc metal, a low-carbon steel for a bcc metal, and an aluminum alloy as a material in which the proof stress/ultimate strength ratio is high. Details are as follows. a) Copper. 99.999 pct purity. After machining the specimen surface was polished mechanically and

Jan 1, 1964

By G. B. Craig, J. Rezek

The critical resolved shear stress of zone-refined zinc single crystals deformed in tension is found to increase with increasing amount of striation substructure. The increase in strength with numbers of striations can be accounted for by a dislocation pile-up mechanism. A somewhat different means of plotting stress-strain data is introduced which allows a more unambiguous choice of yield stress to be made. FEW systematic studies of the role played by mac-romosaic or striation-type substructure in the de- formation of high-purity metal single crystals grown from the melt have been reported. McGrath and Craig' found that striations increased the critical resolved shear stress of zone-refined aluminum crystals, postulating a dislocation blockage by the striations as the strengthening mechanism. Lauri-ente and Pond2 found that, as the etch pit density increased in aluminum crystals grown from the melt, the critical resolved shear stress also increased. Fleischer and Chalmers3 also found changes in the critical resolved shear stress of aluminum crystals grown from the melt at varying rates, but could not be certain whether an increase of strength with rate was due to a higher concentration of impurity in the gage length or to an increase in the number of small-angled boundaries. An early work by Hibbard4 indicated that copper crystals containing lineage boundaries were stronger than those free from this type of imperfection.

Jan 1, 1962

By H. C. Fiedler

The development of cube-on-edge secondary re crystallization texture in Si-Fe strip depends upon the ability of inclusions, such as manganese sulfide, to restrain nomal grain gvowth. The ability of inclusions to perform this function is shown to depend upon their size and distribution. The desired uniform dispersion of many small inclusions is obtained by controlling the cooling rate from above the solution tempmature of the inclusions or, alternatively, by drastic 'quenching from above the solution temperature followed by heat treating at a lower temperature. The development of the cube-on-edge secondary recrystallization texture in Si-Fe strip depends upon both establishing the proper primary recrystallization texture and having inclusions to restrain normal grain growth. It has been suggested that the first requirement is met with a texture which, though weak, is near (110)[001]. A mechanism for secondary recrystallization originally proposed by ielsen' is consistent with the presence of this primary texture. It is visualized that a secondary grain appears where two adjacent grains happen to be sufficiently similar in orientation to be capable of behaving as a single large grain. This pair of grains will grow at the expense of its smaller neighbors as a consequence of its having lower grain boundary free energy. Other explanations of the origin of the secondary recrystallization texture have been offered,' and it can be concluded that the mechanism is not fully understood. Nevertheless, it is apparent that operationally the necessary primary recrystallization texture is obtained simply by interposing an anneal during the cold reduction to final gage. The role of inclusions in the development of the secondary recrystallization texture has been demonstrated by May and Turnbull4 in a comparison of a high-purity heat and a heat containing manganese sulfide. They showed that the driving force for secondary recrystallization is inversely proportional to the diameter of the matrix grains and that the function of the inclusions is to maintain a small matrix grain size. Other inclusions, such as various sul- fides,')' silicon nitride,8 and vanadium nitride,' titanium carbide,'' and vanadium carbide," have been shown to be effective in promoting secondary recrystallization to the cube-on-edge texture. There is also evidence that silica inclusions in Si-Fe strip made by powder-metallurgy techniques perform this function.' This report is concerned with the second of the two requirements for secondary recrystallization, namely the importance of the metallurgical control of the size and therefore number of inclusions. While the concepts involved are simple, they are fundamental to the making of this material. EXPERIMENTAL PROCEDURE A heat containing 3.27 pct Si, 0.026 pct S, 0.06 pct Mn, and 0.004 pct C was made in an air induction furnace from electrolytic iron, 98 pct ferro-silicon, electrolytic manganese, and iron sulfide. The heat was divided between graphite and baked-sand molds with a cavity 2-5/8 by 5 in. and with walls 2-3/4 in. thick. Slabs, 1 in. thick and 2-5/8 by 5 in., were cut from the ingots and hot-rolled either in the as-cast condition or after one of the heat treatments described later. After pickling, the 80- to 100-mil-thick hot-rolled band was, unless otherwise noted, heat-treated for 5 min at 900°C before being cold-rolled to 25 mils. After a heat treatment of 1 to 3 min at 860°C, the strip was cold-rolled to 12 mils, the final thickness. The percent cube-on-edge texture was determined by dividing the maximum torque as averaged from four samples by the maximum torque of a 3-1/4 pct Si-Fe single crystal having the (110) plane in the plane of the sample. The torque measurements were made in a field of 1000 oe on 1-in.-diam disks. RESULTS AND DISCUSSION The importance of the thermal history of the ingot is evident in Fig. 1 from the appearance of strips heat-treated in a temperature gradient. The sample

Jan 1, 1964

By Harry A. Lipsitt, Douglas Y. Wang

A fatigue study in completely reversed axial tension-compression has been perforried on high-purity titanium and on three high-purity alloys of titanium. The alloys each contain approxi7nately 0.75 at. pct of a single interstitial element; carbon, nitrogen, and oxygen, respectivley. The results corroborate a previously published theory which proposed that strain aging under alternating stress was responsible for the fatigue limit behavior of certain alloys. The present data indicate that in these alloys an increasing strain-talline aging effect under alternating stress is provided by oxygen, carbon, and nitrogen, respectively. CURRENT research on the nature of the fatigue limit in metals suggests that the presence of a fatigue limit in metallic materials is a manifestation of strain aging that occurs under alternating stress.lm5 A comprehensive theoretical model based on the above hypothesis has been developed to explain the existence of a fatigue limit.' This model also provides increased insight into several other fatigue phenomena as under stressing, overstressing, and coaxing effects. The theory, as well, provides equal understanding for those cases where no real fatigue limit is observed. Briefly, this theory assumes that the S-N curve for a pure metal is a smooth function of the applied stress, and the effect of adding an element that is soluble (or forms a precipitate) in the pure metal is simply to shift the S-N curve to the right. If the added element confers the power to strain age, the result is a further shift of the S-N curve, this time upward and to the right. Since strain aging is not expected to be a strong function of stress, and since damage per cycle is known to be quite stress dependent, it is to be expected that there will be some limiting lower stress at which the strengthening due to strain aging will balance the damage due to crack propagation. This stress is the fatigue limit. The position of the fatigue-limit knee was thought to be a function of the magnitude of the strain-aging effect on both the finite and infinite life portions of the S-N curve. Although the strain aging hypothesis seems to be reasonably valid for bcc materials,2'6 it needed to be tested for both fcc and cph metals. This report is the first of a series concerning the fatigue-limit behavior of titanium with varying amounts of the interstitial solutes (C, N,, and 4) that are known to cause static strain aging in titanium. Yield-point effects have been reported for polycrys-talline high-purity titanium alloys containing either carbon, nitrogen, or oxygen.7'9 These effects were observed at testing temperatures in the range 100 to 300'. In addition yield-point and strain-aging effects have been reported for single crystals of titanium containing 0.1 wt pct C plus N.' These yield points were observed over a wide temperature range, but no room-temperature aging occurred. Aging at 180' was required to cause the return of the yield point. The magnitude of the yield phenomena in titanium containing interstitials is not expected to be as large as is observed in bcc metals because of several factors. Titanium has a very high chemical affinity for oxygen and nitrogen. The thermodynamic stability of solutions of oxygen or nitrogen in titanium is recognized. Lattice parameter measurements of titanium containing arbon, oxygen,1° or nitrogen" show that the "c" parameter is expanded more than the "a" parameter, but that up to about 2 wt pct this results in an insignificant change of the axial ratio 'c/a." Ehrlich" has shown that the sites occupied by interstitial atoms in titanium are spherically symmetrical and therefore a lattice expansion, at a constant c/a ratio, results in a simple dilation of the interstitial site. Such a dilation involving no shear has been shown to react only with edge components of dislocations.13 This causes only a weak pinning action. Shear stresses would be anticipated locally when only one of the two interstitial positions was occupied. The carbon atom will cause a symmetrical distortion of the lattice whereas the oxygen and nitrogen atoms have, in addition, the previously mentioned chemical affinity of titanium for these elements. These factors will result in a considerably smaller reduction of free energy upon the association of interstitial atoms with dislocations, and therefore a much weaker pinning than has been observed for the bcc metals. These considerations would lead to the hypothesis that of the interstitial elements considered here carbon would cause the strongest pinning effect in titanium where the amount of interstitial in solution is constant. This hypothesis will be borne out in the analysis of the present results.

Jan 1, 1962

By John G. Simmons

A tutorial account of the tunnel effect between metal electrodes separated by a thin insulating film is presented. Energy diagrams of metal-insulator -metal sandwiches are briefly discussed, and the isothermal and thermal J-V characteristics for symmetric and asymmetric tunnel junctions are derived from a generalized expression. Experimental data from tunnel junctions are presented, and it is shown how, correlation between theory and experiment permits a determination of some of the oxide properties. Finally, the barrier heights in AT-Al2O3-A1 junctions determined by various inr3estigators are discussed. The object of this paper is to give a tutorial account of the electric-tunnel effect between metal electrodes separated by a thin insulating film and to illustrate its utility in the study of metal-oxide interfaces. This paper will emphasize the theoretical aspects of the electric-tunnel effect: however, a brief account of the more recent experimental work will be presented. A more comprehensive experimental study of the tunnel effect will be presented in the following paper by Drs. Pollack and Morris, see p. 497. We will open with a discussion of the energy diagrams for metal-insulator-metal junctions and what we mean by the electric-tunnel effect. Following this is an account of the theoretical studies of Sommerfeld and Bethe, and Holm, and the recently reported calculations on symmetric and asymmetric junctions using the generalized formula. The theoretical presentation is concluded with calculations on the temperature dependence of the tunnel J-V characteristic. In the experimental section, it is shown how barrier heights and thicknesses are determined by comparing the experimental data with the theory. Finally, the results of various investigators on the determination of the barrier heights in Al-Al2O,-A1 systems are discussed. I) POTENTIAL BARRIERS BETWEEN CLOSELY SPACED ELECTRODES Since it is necessary to supply energy to a metal in order to observe electron emission from its surface, it is evident that the potential energy of an electron within a metal is lower than that of an electron at rest outside the metal. It follows, then, that the electron must undergo a change in potential as it crosses the metal-vacuum interface. In the Sommerfeld1 free-electron theory of metals it is assumed that the free electrons, the electrons responsible for the electrical characteristics of metals, move in a constant potential environment in the metal, and are contained within the metal by a potential barrier at the metal-vacuum interface. It will be evident that the idea of a constant potential environment is a very idealized model, since the potential energy of an electron is influenced by all of the atoms and the other electrons within the metal; however, its use is justified by the many excellent results it gives.2 Fig. l (a) is the energy diagram of the model envisaged by Sommerfeld; the metal is represented by a potential-energy well of depth V,; the vacuum level represents the energy of an electron at rest outside the metal. At 0°K the electrons, which exist in pairs in discrete energy levels, fill up the well to an energy ?, the Fermi level. The distance, Vo - ?, from the Fermi level to the vacuum level is known as the work function of the metal, and represents the minimum energy required to free electrons from the interior of the metal at 0°K.

Jan 1, 1965

By A. G. Metcalfe, A. Siede

The hardening of annealed copper during fatigue testing appears to be independent of the applied stress and to occur largely within the first 4000 cycles. Copper hardened by fatigue is more resistant to annealing than copper hardened by static means, and softens by a multistage process. Creep studies at 300°C show that copper hardened by fatigue also has greater resistance to creep than statically hardened copper. Some theoretical explanations are discussed. RECENT interest in the behavior of metals subjected to fatigue stems from observations of phenomena which could not be explained by current theories of the structure of metals. In 1953 Bullen et al.l showed that the hardening caused by fatigue at low stresses did not cause severe bending of the lattice as did hardening by static methods. This was later confirmed by Wood, 'and Broom and Ham.3 Clarebrough et al4 showed that as the stress level used in fatigue hardening was increased, bending of the lattice was introduced, eventually approaching that of metals deformed by conventional cold-working methods. The type of hardening introduced by fatigue has other interesting characteristics. Kemsley5 showed that copper fractured by fatigue at low stresses resisted annealing as compared to copper hardened an equal amount by static means. The work of Clarebrough etal, shows that the manner of energy release during annealing differs for fatigue-hardened and statically hardened metal in both magnitude and temperature range of release. In addition to hardening of the lattice, fatigue has been shown to cause softening in materials previously hardened by static methods,5 and also in age-hardened alloys.7 The superposition of a small fatigue stress on the static loading of high-temperature alloys has been shown to increase their resistance to creep.' The present work was an attempt to determine the relationship between fatigue hardening, resistance to annealing, and the creep properties of pure copper. EXPERIMENTAL PROGRAM The program described here was designed to answer three questions: 1) Can stress-cycle combinations well below the failure values give significant increases in hardness? 2) Does hardening, under different stress-cycle combinations, give increase resistance to annealing? And if so— 3) Does superior resistance to annealing confer superior creep properties over statically hardened material? Three-quarter hard OFHC copper was annealed in vacuo at 550°C for 1 hr to produce a strain-free material having a fine grain size and uniform hardness. (The standard deviation of the mean hardness of the annealed samples was less than + 1.9 Dpn). Annealing was performed after machining the fatigue specimens so as to avoid work-hardened surfaces. Jigs were used to avoid distortion. The fatigue properties of the annealed copper were determined using an accurately aligned Baldwin-Sonntag SF1-U machine operating at 1800 cycles per min. This limited the minimum number of cycles which could be examined to about 4000 stress reversals. Since failure resulted in specimen damage, hardness increase with number of cycles was measured by testing specimens fatigued to different fractions of the failure values. The results are shown in Fig. 1, and are superimposed on the S-N curve of the copper in Fig. 2. The resistance to annealing of fatigue-hardened material was compared to that of material hardened by uniaxial tensile straining. Specimens hardened to equal values by both methods were progressively annealed for 1 hr at temperature in vacuo. After each annealing treatment the hardness was measured using a Leitz Durimet hardness tester with a 500-g load. At least five tests were made for each hardness determination. The standard deviation for individual specimens varied from 0.26 to 1.48 hardness points. The results are shown in Figs. 3 to 5. The discontinuous nature of the annealing process has been brought out by drawing idealized curves through the data points. This can be justified by the consistency of the results as well as by general agreement with results of Kemsley.5 The creep resistance of fatigue-hardened and

Jan 1, 1960

By J. Gurland

The strength variation ofWC-Co alloys with composition and particle spacing falls into two ranges. 1) Above a critical value of the mean free path, the strength follows a dispersion hardening relation; it is proportional to the volume fraction of WC and in-versely proportional to the particle size. 2) Below a critical value of the mean free path, the strength variation is based upon the Griffith criterion of fracture and is controlled by the case of crack propagation. The maximum strength is attained when the structure and load just satisfy the critical conditions for spontaneous crack propagation. THE relation between the strength and the structure of cemented carbides is of interest not only for its own sake but also for its contribution to the understanding of composite materials in general. The cemented carbides are essentially aggregates of a brittle carbide constituent and a ductile binder phase and, at certain structures of intermediate composition, the composite alloys attain strengths far higher than those of the separate constituents in bulk. It is this variation of strength with structure which will be discussed. The experimental procedures of sample preparation, mechanical testing, and quantitative metallo-graphic analysis have been described in Ref. 1. The reported strengths are the maximum fiber-stresses in bending calculated on the basis of completely elastic behavior. The plastic contribution to the over-all deformation is small and was neglected in the calculations (the elongation at fracture of the most ductile alloy, 50 pct WC-50 pct Co, is less than 0.5 pct).2 The pertinent structural parameters are: f, the volume fraction of tungsten carbide; p, the mean free path through the binder phase; the contiguity C, defined as the average fraction of surface area shared by a particle of wc with all neighboring particles of the same phase; and d, the average equivalent spherical diameter of the WC particles. The strength variation of WC-Co alloys with composition and particle spacing is shown in Fig. 1 and Table I, which are based partly on previously published data' and partly on new experimental work which extends the range to higher values of the mean free path. Each data point represents the average of five samples. For a given composition, the transverse rupture strength reaches a maximum

Jan 1, 1963

By J. C. Fisher

Martensite transformations in steels and other alloys are characterized in part by the absence of composition changes during the growth of a new phase. Transformation occurs rapidly, and there is insufficient time for long range diffusion or partition of alloying elements to take place; martensite reactions in alloys thus are similar to phase transformations in single component systems. A fundamental understanding of martensite transformations in steels is impossible without knowledge of the free energy change upon transforming austenite (face centered cubic iron containing alloying elements) to ferrite (body centered cubic iron containing alloying elements) of the same chemical composition. The present paper assembles the best information available concerning the influence of temperature and composition on this free energy change. Most of the material has been taken from the work of Johansson,' Mehl and Wells,2 Zener3,4 and Smith;5 and indirectly, through these authors, from the work of Austin.6 In agreement with the generally accepted viewpoint, martensite is assumed to be an ordered solution of carbon in ferrite of the same composition as the parent austenite; only at high temperatures and low carbon concentrations is the carbon in ferrite distributed at random. The properties of the disordered solution are estimated by extrapolating the known properties of iron-carbon solid solutions into the range of supersaturation, and the free energy change associated with ordering is estimated from the theory developed by Zener. By incorporating Smith's recent thermodynamic measurements and Zener's theory of ordering, the present analysis modifies previous estimates of the free energy change associ- ated with martensite transformations. Consider a two component system consisting of a solvent A and a solute B. Let Na and Nb represent mol fractions of A and B respectively, let aa = raNa and ab = YbNb represent activities, and let superscripts 1 and 2 refer to phases 1 and 2. The partial molal free energies of A and B in phases 1 and 2 can be summarized as follows: free energy standard state Fa' - Fao1 = RT In an1 pure A' Fa2 - Fa2 = RT In aa2 pure A2 Fb1 - Fbo = RT In ab1 pure B Fs2 - FB2 = RT In aB2 pure B. The free energy of a gram atom of phase 1 is* AF1 = Na'Fa1 + Nb'Fb1 and that of phase 2 is AF2 = NA2FA2 + NB2Fb2. A martensite transformation from phase 1 to phase 2 requires Na1 = Na2 = Na and NB1 = NB2 = Nb, and the free energy change per gram atom accompanying transformation is AFi-2 = NA(Fa2 - Fa1) + Nb(Fb2 - Fb1) = NA[RT In (aA2/aA1) + AFa1?2] + Nb RT In (aB2/aB1) = Na[RT In (ra2/ra1) + AFa1?2] + Nb RT In (TaVTB1). [1] where yn2, yn1, yB2, yB1 are activity coefficients, and where Ma'+* is the free energy change upon transforming a gram atom of pure A from phase 1 to phase 2 at the temperature in question. For martensite transformations in plain carbon steels, A = iron (Fe), B = carbon (C), 1 = austenite (7). 2 = ferrite (a), and Eq 1 is AFr ?a= NFe[RT In (rf.a/rfer) + AFF.r?a] + TVc RT In (eca/rc1)- L2] Nothing is known concerning the values of yFea and yca for carbon concen- , trations in excess of 0.025 pct. However, the approximation rf.7 = 7f." = 1 cannot be appreciably in error for small carbon concentrations, and Eq 2 reduces to Afr-a = NfeFfer?a1 + NC RT In (rca/rcr)- [3] Johanssonl and Zenera have calculated MFfr?a from the specific heat measurements compiled by Austin.6* Their calculated values agree closely, and are summarized in Table 1. The activity coefficients relative to graphite for carbon dissolved in iron vary with temperature according to the relationships d In rcr/d(1/T) = AHcr-R dlnyca/d(l/T) = AHCa/R where AHc? and AHca are the heats of solution of graphite in y and a iron. Assuming the values of AH to be independent of carbon concentration and temperature, In rCr = AHcr/RT + I1 In rca = AHCa/RT + 12

Jan 1, 1950

By L. Kaufman

An equation is derived relating AF a", the difference in free energy between austenite and martensite, to temperature and composition in the iron-chrmnium and iron-chromium -nickel systems. This equation is used to calculate at M,for low-carbon stainless steels. A method for utilizing the energetic equations to calculate M, in stainless steels is suggested. In addition, an explanation is given for the anomalous effect of chromium in forming a y loop concurrent with the lowering of M, in iron-base alloys. It has been demonstrated that the effect of alloying elements on the Ms temperature is strongly dependent upon the influence of the alloying element on the thermodynamic properties of the y and a phases. Notwithstanding the fact that alloying elements affect the Ms temperature by changing the lattice parameters, elastic constants, and inter-facial energies of the a and y phases, their primary role is a thermodynamic one. Thermodynamic analyses have been presented for the iron-carbon'-4 and iron-nickele systems in which equations are derived for Fa and FY, the free energies of the b.c.c. (or b.c.t.) and f.c.c. phases. These equations, which can be obtained with the aid of the equilibrium diagrams, can be used to help predict the Ms vs composition curves. This paper deals with a calculation of the thermodynamic properties necessary to compute AFffLY for iron-chromium alloys. In view of the lack of direct information concerning the activity of chromium in the a and y phases of iron, the regular solution approximationa,g was utilized to bridge the gap. After the necessary equations were derived for the iron-chromium system they were used to calculate AFff— Y for stainless steels by combining them with the previous results obtained for iron-nickel alloys. The results of these calculations offer a means of rationalizing the anomalous behavior of chromium in iron-base alloys. The latter depends on the regular solution approximation in contrast to the explanation offered by zener , which is based upon the "magnetic specific heat effect." THERMODYNAMICS OF THE IRON-CHROMIUM SYSTEM If y represents the atomic fraction of chromium in an iron-chromium alloy, then the free energy of the b.c.c. phase Fa, the f.c.c. phase FY, and the difference in free energy between the f.c.c. and b.c.c. phases of the same composition are given through [3] and Fmy represent the free energies of mixing of the a and y phases; while @era -' and uFea 4Y represent the difference in free energy between f.c.c. and b.c.c. chromium and iron respectively. The values of AEFe0ff4Y in the temperature range between 0" and 900K ( .

Jan 1, 1960

By E. V. Kleber, V. F. Novy, R. C. Vickery

The constitutional diagram has been determined in part for the sgstem Gd-Fe. Seven intermetallic compounds have been found at compositions corresponding to the following Gd-Fe ratios; 2:3, 1.2, 1:3, 2:7, 1:4, 1:5, and 2:17. The GdFe, com-pound has a congruent melting point above 18000C, whereas all of the remaining compounds melt incongruently, A eutectic exists at 13 wt pct Fe at a temperature of 860°C. ALTHOUGH an increasing amount of information is becoming available on the metallurgy of pure rare earth metals, relatively few studies have been made on alloy systems of the rare earths with the transition metals; iron, nickel, and cobalt. Even less data are extant on gadolinium alloys although it could be expected that such alloys might possess interesting structural and magnetic properties. In this paper, we report investigation of the gadolinium-iron alloy system. Studies on systems containing nickel and cobalt will be presented in subsequent papers. Previous data on gadolinium-iron compounds have been reported by Endter and Klemm1 and Jep-son and Duwez.2 Both reports agree on the existence of the GdFe2 phase, which crystallizes with the C-15type structure (face-centered cubic). This is in accord with results of the present work. EXPERIMENTAL The gadolinium metal used was prepared by met-allothermic reduction of high purity gadolinium fluoride. A typical analysis of the metal thus produced indicated the principal impurities to be oxygen, 0.24 pct, and tantalum, 0.2 pct. The iron metal used for alloying was a reagent grade material. Alloys were prepared by arc melting mixtures of the respective metals. Appropriate quantities of the two component metals were carefully weighed, blended and compacted. The compacts were then melted in an arc furnace consisting of a water-cooled copper hearth plate and a water-cooled tungsten-tipped counter electrode contained in a brass vacuum chamber. The furnace was evacuated to an absolute pressure of less than 1 with the leak rate established as being less than 20 per min. The furnace was then back-filled to a pressure of 25 mm of Hg, with a 75 pct He-25 pct Ar mixture. Arc melting at 25 v and 300 amp was maintained for about 30 to 60 sec until the entire visible area of the sample became molten, by which time surface tension had drawn the molten metal pool into a large bead. After allowing the furnace to cool, the alloy button was inverted and the melting process repeated. A series of melts was made covering entire composition ranges from 100 pct Gd to 100 pct Fe. All samples were examined metallographically in the as-melted condition. As the phase diagram became rationalized, and differential thermal analysis had given indications of temperature arrests, annealing was carried out in vacuo or in argon. Generally, compositions containing more than about 60 pct Gd were heat treated at 750°C and alloys containing less than this quantity of gadolinium were annealed at 1000°C. Prolonged annealing times were of the order of 100 to 200 hr. Metallographic examination was repeated on the annealed materials. Although little material loss was noted during alloy preparation, chemical analyses were made on all specimens prepared. The following standard, analytical techniques were used. Samples were dissolved in hydrochloric acid; insoluble material was filtered off, ignited and fused with potassium pyro-sulfate. The cold melt was leached with dilute acid and the solution thus obtained added to the main sample solution. Gadolinium was always determined gravimetri-callv as oxide from oxalate ignition. Iron was removed from an aliquot sample by electrolysis on a mercury cathode at approximately six volts. Gadolinium in the iron-free electrolyte was then precipitated as oxalate which was subsequently ignited to oxide. Iron was determined volumetrically in an aliquot solution by reduction with stannous chloride fol-

Jan 1, 1962

By C. J. Osborn

The temperature dependence of the free energies of formation of metallurgically important oxides, sulphides, chlorides, carbonates and sulphates is presented graphically, whereby the task of deriving practical metallurgical information is greatly simplified. The diagrams have been revised according to the latest available thermodynamic data. WHILE it is true that much careful investiga-tional work remains to be done, there are already available in the literature sufficient thermo-dynamic data to enable the metallurgist to predict with reasonable confidence the equilibrium conditions to be expected in most important, and potentially important, metallurgical reactions. In this connection, the careful, critical compilations of K. K. Kelley and his collaborators in numerous U. S. Bureau of Mines Bulletins are particularly valuable. However, the translation of reliable thermo-dynamic data into metallurgical language unfortunately involves lengthy manipulations of the cumbersome formulas which can be set up from heat of formation, entropy, and heat capacity data. Many years ago, Kelley20 predicted that thermo-dynamic calculations would be made in the future "with the aid of free energy, heat content, and entropy tables, rather than by means of the algebraic methods employed at present", and where such tables are available they greatly simplify the task of deriving practical metallurgical information. However, in addition to these tables, there are also available one or two reliable graphical presentations, particularly of free energy data, which may be used in the same way as the tables, but which have the added advantages that interpolation is never necessary, and that the behavior of different metals or compounds may be compared at a glance. Some years ago Professor Robert Lepsoe, Head of the Department of Metallurgy at Norways Institute of Technology (N.T.H.), began a compilation of curves showing the temperature dependence of the heat contents of the elements, and the heat contents and heats and free energies of formation of the common metallurgical compounds. With such a compilation, the thermal and free energy requirements of a reaction may be read from curves in a matter of seconds; and this statement applies not only to the formation of compounds from their elements, and to the reverse dissociation, but to any reaction involving substances whose thermodynamic properties are available. This follows from the fact that heats and free energies of formation are completely additive. For example, the free energy change associated with the chlorination of a metallic oxide at any temperature, T, is given by the free energy of formation of the chloride at temperature T minus that of the oxide also at T; both of these quantities may be read directly from curves of the type suggested above. Since this early, unpublished work of Lepsoe, both Ellingham7 and Richardson and Jeffes46 have

Jan 1, 1951

By R. W. Heckel, R. L. DeGregorio

The DeHoff method of determining the size distribution of ellipsoid-shaped, second-phase particles has been applied to the spheroidization of cementite in a eutectoid steel. The surface area of Precipitates determined from the various size fractions in the distribution was correlated with surface-area measurements based on the number of intersected interfaces on a random lest line. The precipitate particles were found to be oblate ellipsoids with an axial ratio of about 0.90. The size distributions were found to be log-normal. A method is proposed whereby the shrinkage of small particles and growth of large ones can be determined from the experimental data. The experimental data are compared to previously proposed mathematical models describing diffusion-controlled kinetics and various types of interface-controlled kinetics. The experimental growth and shrinkage rates are considerably slower than those predicted by diffusion-controlled kinetics. The best fit is obtained for a model describing interface-controlled kinetics limited by the rate of formation of cementite at the growing interfaces where the interfacial reaction rate is proportional to the solute thermodynamic activity gradient across the surface. THE subject of spheroidization of second-phase particles has been considered previously by other investigators. Livingston1 has studied the precipitation of the cobalt-rich phase from copper alloys containing 0.7 to 3.2 wt pet Co. His results indicate that the average particle size increases as the cube root of the heat treatment time in accord with diffusion-controlled kinetics.'-' Komatsu and rant' in their studies of the growth of SiO2 in a dispersion-strengthened copper-silica alloy found that the initial growth of SiO2 proceeded by diffusion-controlled growth and later stages were limited by interface-controlled growth. The values of the activation energy obtained for the interface- controlled process led them to the conclusion that the process was limited by the dissociation of SiO2 at the shrinking interface. The transition from diffusion-controlled growth to interface-controlled growth was characterized by a particle size which varied with the heat-treatment temperature. It is also interesting to note that the particle-size distribution as found by Komatsu and Grant exhibits log-normal behavior when replotted on log-probability coordinates. Dromsky, Lenel, and Ansell9 have observed the growth of Al2O3 particles having a mean free path of about 1.5 to 13 µ. Their photomicrographs indicate that the Al2O3 particle sizes were larger than those observed by Komatsu and Grant. Dromsky, Lenel, and Ansell concluded that the Al2O3 coarsened by an interface-controlled growth mechanism limited by the solution of A12O3 at the shrinking interfaces. Bannyh, Modin, and Modin10 have studied the spheroidization of a eutectoid steel (0.83 wt pet C). Their spheroidization (tempering) treatments were carried out in the range from 210o to 700°C for times between 1.5 sec and 20 hr. Measurements of mean particle size as a function of time indicated a cube root of time dependence of size at 700°C, in agreement with previous analyses of diffusion-controlled kinetics.2°7 At lower temperatures, the time dependence was less than the one-third power. It is important to note that, although mathematical models of growth* considering particle-size distribution have been available, measurements of only mean particle size have been carried out. DeHoff11 has presented a quantitative metallography technique which is applicable to the determination of the size distribution of ellipsoids of constant shape. This method is applicable to both oblate and prolate ellipsoids of all ratios of minor to major axis and is based upon an extension of Saltykov's analysis12 of the distribution of spheres of varying size. DeHoff's analysis is based upon measurements of the minor axis of the particles on a random plane of polish in a unit area. This method provides a measure of the number of ellipsoids in various size ranges per unit area. As pointed out by DeHoff, such measurements may be used to obtain surface area and total precipitate volume data. Thus, the accuracy of the distribution analysis may be checked by comparison of the surface and volume

Jan 1, 1965

By J. D. Grozier, W. W. Mullins, H. W. Paxton

The morphology of the nucleation and growth of austenite into high-purity iron in NH3-H2, gas mixtzires has been studied. The growth rate of an austenitic rim (planar interface) from scratched surfaces of- polycrystalline iron specimens was determined as a function of temperature and percent NH3 and found to he in reasonahle agreement with growth-rate calculations based on a model which assumes the diffusion of nitrogen to he the rate-controlling step. Under certain conditions, austenite pew as platelets; growth rates of these along the surface of single crystals were found to he dependent on surface preparation, orientation, THE development of microstructures can conveniently be thought of in terms of nucleation and growth. The complex topic of nucleation in solid-solid transformations will not be discussed at length, although a few qualitative experiments were performed. Growth is determined by the mobility of the various possible interfaces under the available driving forces in a particular situation. In homophase transformations, it has been clearly demonstrated that the mobility is very sensitive to small amounts of impurities, to the kind of impurity, and to the orientation relations across the boundary.' A degree of quantitative knowledge has temperature and percent NH3. On electropolished surfaces with a (110) pole, the maximum growth rate of platelets was determined and found to be in agreement within a factor of 2 to 3 with theoretical calculations of the edgewise growth rate of a platelet. The calculations are based on a model which assumes that the diffusion of nitrogen in the ferrite in front of the moving curved attstenitic interface at the tip of the platelet is the rate-controlling step. The growth direction on the surface of the single crystal and resztlting habit plane of the platelets were determined and found not to he unique but to vary systematically with the pole of the surface. been obtained. In the case of heterophase transformations, which are significantly more complicated to treat, a comparison of theoretical and experimental studies often shows rather poor agreement.'-' The main purpose of the experiments discussed below was to obtain additional quantitative data on crystallographic effects on boundary mobility in a system which was otherwise as simple as possible, viz. the growth of iron-nitrogen austenite into fer-rite. In the course of the experiments, it became necessary to re-examine the theory for the growth of Widmanstatten plates, especially the influence of anisotropic interfacial free energy. Some other minor experiments in the Fe-N system were performed, and are described briefly. EXPERIMENTAL METHODS A) Material Preparation. The floating zone purified iron used for this investigation was supplied by Battelle Memorial Institute. The nominal composi-

Jan 1, 1965

By H. R. Peiffer, R. Geckle

The silver phase of silver and finely divided alumina composites is shown to grow as single crystabs upon solidification from the melt. These crystals grow without the aid of an externally applied thermal gradient or a mold. The single-crystal growth is more readily disrupted as the alumina content varies from the optimum concentration of 20 wt pct. Pretreatment such as variation of fabrication pressure, and presintering of the specimens appear to play a lesser role in this phenomenon. In actual practice the growth of single crystals of metals from the melt is a tedious, painstaking process.' Care must be taken to insure the proper thermal gradients and rate of growth in a vibration-free environment. Excellent single crystals can be grown when these variables are controlled. When a second insoluble and finely divided phase is present in the melt the growth of single crystals is difficult, if not impossible, even under the most favorable conditions outlined above. During the past few years, a composite of silver and alumina has demonstrated an ability to resist flow even though the major phase, silver, is mo1ten. 2 A more surprising feature of this material is the ease with which single crystals of silver grow upon cooling the composite from temperatures above the melting point of silver to temperatures near room temperature. Observations have been made that demonstrate that single crystals of silver will grow even when conditions most unfavorable for growth of pure silver crystals exist. The crystals grown in this manner are imperfect in that a great deal of subgrain structure exists. It is the purpose of this manuscript to qualitatively describe this process and to attempt to isolate the influencing factors of this behavior. MATERIAL PREPARATION The composite material is prepared by powder-metallurgical techniques. Two items of great importance are the introduction of oxygen into the system and the obvious need for a thorough mixing of the metallic and refractory oxide phases. The oxygen is necessary in order that the silver wets the alumina. Without oxygen the phases separate and the material shows no unusual properties. In order to supply the needed oxygen the silver is added in the form of silver oxide which is readily decomposed to oxygen and silver upon heating. The silver oxide is readily obtained in a finely divided form that allows a uniform mixture of the alumina and silver oxide to be prepared. The correct ratio of alumina to silver oxide is mixed in the presence of alcohol in an ordinary food blender. The mixture has the consistency of a thin slurry and after mixing for approximately 20 min is dried and pressed into billets at various pressures. As is shown in Fig. 1, there is no limit to the shape of the specimens obtainable as single crystals. These billets are heated in air in a predeter-

Jan 1, 1964

By P. M. Robinson, M. B. Bever

THE heats of formation at 273°K of the compounds InBi, In2Bi, and TIBi2 have been determined by metal solution calorimetry with bismuth as solvent. The published information on the thermody-namic properties of these compounds is limited. The structure of the compound InBi (mp, 383°K) is tetragonal of the PbO (B10) type.1 The compounds In2Bi (mp, 362°K) and TIBi2 (mp of highest-melting composition, 486°K) are now believed to have structures of the NizIn (B8) type273 and not the A1B2 (C32) structures reported earlier.4, 5 The compounds In2Bi and TIBi2 have limited homogeneity ranges at room temperature, while the compound InBi has no detectable homogeneity range.' Samples of the compounds were prepared from 99.99 pet Bi (Mallinckrodt Chemical Works), 99.999 pet in (American Smelting and Refining Co.), and 99.99 pet Tl (Amend Drug and Chemical Co.) by melting the components in sealed, evacuated Vycor tubes. The compounds InBi and in2Bi were of the stoichiometric compositions. The compound TIBi2 contained 40 pet T1; this composition was chosen because at room temperature the alloy of stoichiometric composition does not lie within the homogeneity range.' The melts were held approximately 50°C above the liquidus for 8 hr and then quenched into iced brine. The samples were homogenized for 48 hr at temperatures 20°C below the temperature of complete solidification. Metallographic examination of sections taken from the center and the ends of the ingots did not reveal second phases or segregation . The heats of formation were determined as the difference in the heat effects on solution of a compound and of a mechanical mixture of its components added from the reference temperature of 273°K to liquid bismuth at 623°K. The details of the calori-metric procedure and method of calculation have been described elsewhere? The calorimeter was calibrated by adding bismuth at 273°K to the bismuth bath at 623°K. The calculated values of the heats of formation of the compounds, listed in Table I, are based on a value of 4.96 keal per g-atom for the difference between the heat contents of bismuth at 273" and 623oK7 No published values are available for the heats of formation of the compounds InBi and In2Bi. Hultgren et a1.7 have calculated a value of -0.50 i 0.10 keal per g-atom for the heat of formation at 423°K of the compound TIBi2 containing 40 at. pet T1 from heat-content measurements in the range from 398" to 700°K and the heats of formation of the liquid alloy. The difference between this value and the value of -0.66 * 0.01 keal per g-atom at 273°K is too large to be attributed to a difference between the heat capacities of the compound and the components over the temperature range 273o to 423oK7 However, the direct determination of the heat of formation by a calorimetric method should give a more accurate value. The heats of formation at 273°K of the compounds InBi and In2Bi lie on a straight line when plotted as a function of composition, Fig. 1. The compound In2Bi, therefore, appears to be barely stable with respect to its neighboring phases. In order for the free energy vs composition curve to be concave towards the composition axis, the entropy change on formation of In2Bi must be more positive than that on formation of InBi. As the compounds are ordered,2, 3, 8 there is no configurational entropy change on formation. The difference between the entropy changes on formation of the two compounds, therefore, is probably associated with the vibrational entropies. In view of the low heats of formation, the changes in vibrational entropy, due to changes in bond strength on formation of the two compounds, are likely to be small. Owing to the volume contractions on formation of the two compounds,5 the vibrational entropies probably decrease slightly but the decrease of the vibrational entropy of InBi is expected to be larger than that of In2Bi.

Jan 1, 1965

By G. S. Cole, G. F. Bolling

Recent experitnents on crystal growth have indicated that thermal convection in the liquid influences the conditions of solidification and subsequent crystal characteristics. However, the important columnar-equiaxed transition in castings has not been examined as a function of this free convection. An experiment is presented here to demonstrate a change in the transition as convection is suppressed. Lead- and tin-based alloys were cast in stainless-steel molds either without special procedures or with thin grids placed horizontally across the tnold diameter in portions of the castings. While the grids allowed some diffusion of heat and mass transport to occur during solidification, massive fluid flow was hindered in the vegiotz of the grids. A larger percentage of columnar grains appeared in the grid region than in other parts of the special castings or than in all regions of normal castings. This structural difference can be explained in terms of thermal convection. THERE is little information about the possible effects of natural convection which may occur during the solidification of a casting. Ruddlel-3 has stated that the movement of liquid metal produced by convection might assist in the transfer of heat, but in the absence of confirmatory evidence he suggested that this idea be treated with caution. chalmers4 considers, but does not assess, the importance of fluid motion in casting, while Tiller5 suggests that, in liquid metals which have extremely high thermal diffusivities, the effects of thermal convection may be negligible. In addition, throughout the literature, there are statements indicating the possible presence of thermal convection but without proofs or evaluations of its influence. Natural convection may be divided into two categories: thermal convection, where liquid flow is due to density differences caused by temperature gradients in the liquid (e.g., liquid near a cold wall falling with respect to hotter liquid further away from the wall); and solute convection, where flow is due to density differences as a result of concentration gradients, see Wagner.6 In most instances thermal convection predominates over solute convection. Nevertheless, even thermal convection has not been studied in any detail and is difficult to handle, since it can vary quantitatively from one experiment to another depending upon the exact positions of heat sources and sinks or the attitude of the container with respect to gravity. In spite of these difficulties it is possible to realize that a higher rate of heat transfer must result with convection than in any situation where conduction acts alone. Thus, the degree of increased heat transfer is the important undecided question in considerations of casting. It will be demonstrated here that a reduction in convective flow leads to a suppression of the columnar-equiaxed transition in a casting and that this suppression is related to a decrease in the degree of heat transfer. Fortunately it is possible to do this by comparative rather than quantitative measures. The insertion of horizontal plates or grids into a simple mold does not appreciably influence heat transfer per se, but in reducing the allowable height through which colder liquid may fall freely it reduces natural thermal convection. This leads to a reduced fluid velocity, a different flow pattern, and an altered heat transfer which can be measured. EXPERIMENTS The alloys were prepared from elements of 99.9 pct nominal purity and cast into an uncoated 2-in.-diameter stainless-steel mold insulated at the bottom with a suitable refractory so as to ensure predominantly radial heat flow during freezing. Casting was performed either in empty molds or in molds containing grids or plates in various configurations . The grids were constructed of aluminum screen (1/32 in. thick) which was cut to the desired shape, cleaned in aqua regia, and coated with a mixture of 2 parts aqua dag graphite -1 part Sauereisen cement mixture, thus supplying a hard, protective, and insulating layer approximately 0.006 in. thick. The

Jan 1, 1965

By G. R. Speich, L. Zwell, H. A. Wriedt

HORNBOGEN1 has recently shown that it is possible to measure the degree of segregation accompanying cellular precipitation in ferritic Fe-Zn alloys by measuring the lattice parameter of the ferrite phase within the cell. In extending this work to study a wider range of composition and to evaluate the free-energy change accompanying the growth of the cells, the lattice parameter and the boundary of the a! phase in Fe-Zn alloys were redetermined to check the earlier work of Schramm.2 This appeared to be necessary since the Fe-Zn diagram has been drastically revised since the work of Schramm. Stadelmaier and Bridgers3 have recently shown that, contrary to prior belief: the y iron field is closed and the a phase boundary extends to an a + T + L peritectic at 782°C where peritectic a contains 42 at, pet Zn. The phase T in equilibrium with the a solid solution has a composition near Fe3Zn10 and a cubic 081 type structure.' The alloys were prepared by two different methods. In the first method carbonyl iron powder (400 mesh, 99.8 pet pure) and high-purity zinc powder (200 mesh, 99.99 pet pure) were mixed in a ball mill and then pressed in a die at 40,000 psi pressure to make compacts with a diameter of 5/8 in. and a thickness of l/4 in. These were sealed in evacuated quartz capsules, gradually heated to 1100°C over a period of 1 week (3 days at 600°C) 3 days at 800°C, 1 day at (1100°C), and then quenched in iced brine. For the alloys containing 23.4 and 30.1 at. pet Zn, the final homogenization anneals were performed at 1000" and 850°C) respectively, to prevent melting. In the second method, alloys were prepared by saturating high-purity iron millings (0.001 in. thick, 99.9 pet pure) in zinc vapor. For each alloy the impregnation was conducted at a fixed vapor pressure of zinc and a fixed temperature between 750' and 800°C. When saturated, the millings were quenched in water. The compositions of the alloys prepared by both methods are shown in Table I. The lattice parameters of the ferritic solid solutions were determined by comparing the 26 values of their 310, peaks (using Co Ka) or of their 211 , peaks (using Cr Ka) with those of corresponding peaks for high-purity a iron. The 20 values were measured at room temperature with an X-ray dif-fractometer (G.E.-XRD5). By using a fiducial value of 2.8664Å for the lattice parameter of pure a iron at 25oC,5 the lattice parameters of the zinc-bearing

Jan 1, 1964

By A. M. Turkalo

IT is a well-known fact that martensitic steels show a greater resistance to brittle fracture than do pearlitic and bainitic steels. It was, therefore, thought worthwhile to investigate the mode of brittle fracture with respect to structure by studying the effect of various austenite decomposition products on the propagation of brittle fracture in steel by means of electron microscopy. Newly developed replication techniques together with the advantages of the electron microscope such as great depth of field, available high magnification and easy adaptation to stereophotography make electron microscopy very suitable for fracture studies. EXPERIMENTAL PROCEDURE An essentially plain carbon steel containing 0.56 pct C, 1.30 pct Mn, 0.02 pct P, 0.03 pct S and 0.22 pct Si was used in this investigation. The following table gives the heat treatments used to obtain the various austenite decomposition products: The resulting austenite grain size from the 870°C austenitization treatment was about an ASTM No. 7 grain size. The specimens which were 2 in. long, 1/4-in. diam bars were austenitized in a tube furnace through which argon gas was blown. The tube was not, however, air-free and the specimens were not completely oxide-free. Isothermal treatments were done in a lead bath. With the exception of the fully hardened specimen and the one as-isothermally transformed at 300°C the heat treated specimens were notched in the following fashion with a jeweler's saw. A cut about 0.08 in. deep was made perpendicular to the length of the bar. The specimen then was rotated about 45 deg and another cut about 0.08 in. deep was made. In the case of the hard martensitic and bainitic specimens, only one cut was made and the depth was not controlled. The design of the two-cut notch caused fracture to initiate at the point of intersection of the two saw cuts. The specimens were broken at -196°C in a Charpy impact testing machine. In order to hold the specimens securely in the impact machine two square pieces of mild steel were made about 5/8 in. long having the same cross section as that of a standard Charpy specimen but containing a hole 1/4 in. diam through the center of the cross-section. These adapters were slipped over the ends of the test specimen which was then secured tightly by means of set screws in the adapters. The specimen with the adapters was cooled to -196°C in liquid N,, then placed in the impact machine and broken within 3 sec. One half of the broken specimen was used for the electron microscope study of the actual fracture surface. The other half was nickel-plated and a cross section through the notch and fracture was polished for electron microscope examination. Carbon replicas of the fracture specimens for the electron microscope study were made essentially according to Bradley's evaporation technique.' Both direct carbon and two-stage carbon replication techniques were used. In the case of the direct carbon technique, a thin layer of carbon was evaporated first at an angle of about 45 deg to the mean fracture surface and then another film of carbon was evaporated normally to the fracture surface. The carbon replica was freed electrolytically.2 The specimen was made an anode in a 10 pct Nital polishing solution. It was etched intermittently by shutting off the power now and then. The freed replica was washed in 40 to 50 pct water solution of nitric acid for 10 min or so, then washed in water and picked up on a screen for examination. The two-stage carbon replication technique involved first making a primary replica of cellulose * acetate from which a carbon replica was then made. One side of a strip of cellulose acetate wet with acetone was pressed lightly against the surface of the specimen, allowed to dry and then stripped. Prior to the deposition of carbon, which was done at 90 deg to the cellulose acetate replica surface, chromium was evaporated at 45 deg. After the carbon evaporation the cellulose acetate was dissolved in acetone according to the Jaffe method3 leaving the carbon replica preshadowed with chromium for examination.

Jan 1, 1961

By H. M. Strong

Nickel and carbon form a metastable nickel-nickel carbide eutectic system which was ohsen)able by freezing latent-heat arvests at pressures 210 khars. The eutectic freezing temperature Is pressure had a slope of -0.9°Cper kbar. At 50 khars, the nickel-nickel carbide eutectic was at 1297° * 5°C and 0.34 mole fraction Ni3C. The composition of the eutectic increased at an initial rate of 0.004 mol fraction per kbar. The carbon content of the a phase also increased with pressure. Seveval illustrations of nickel carbide microstruc-tures are given and data on the lattice expansion of nickel in terms of carbon content are included. IRON, cobalt, and nickel form a variety of carbides with compositions varying between MeC and MC. Of the three, the carbides of iron are most easily recovered. The nickel carbides are so unstable at temperatures above 900°C that their occurrence in Ni-C alloys is unusual. Under sufficient pressure, however, the temperature range for the stability of the ferrous metal carbides is likely to be increased, or their decomposition suppressed (e.g., see Hilliard' on FesC), providing an opportunity to study the unstable carbides more closely. This was found to be true for nickel carbide which, under pressure, formed a eutectic with nickel. This paper describes the conditions of formation of nickel-nickel carbide eutectic compositions and their subsequent recovery. Despite the comparative instability of the nickel carbides, something is already known about them through various chemical reactions at temperatures not exceeding a few hundred degrees, and from their recovery by the very rapid quenching of molten Ni-C alloys from temperatures of about 1450°C or above. At least four carbides of nickel have been reported: A discussion of the state of knowledge about Ni3C is given by ansen.' The studies of Morrogh and williams7 who obtained Widmanstatten nickel carbide structures and of sidorenko8 who first obtained nickel carbide eutectic structures are also of interest. Sidorenko used short heating times for molten nickel in a graphite crucible followed by very rapid quenching to obtain the eutectic structures. Long heating times produced only the nickel-graphite eutectic probably due to the presence of graphite crystals in the melt which nucleated graphite crystallization. According to the early work summarized by anssen,' nickel carbide may form at temperatures far above the melting point and may be recovered by very rapid quenching. More recently, Giardini and Tydings, working on diamond synthesis, found evidence for "free nickel carbide'' in high-pressure reactions carried to temperatures far in excess of the limit for diamond growth, followed by rapid thermal quench. They stated that nickel carbide was not found in association with diamond growth. In the present work, observations were made on the melting and solidification temperatures of Ni-C alloys as a function of composition, temperature, and pressure. The nickel-nickel carbide eutectic was found at pressures above 10 kbar and its variation with pressure followed up to 60 kbar. Nickel-nickel carbide microstructures were recovered by quenching under pressure from temperatures of about 1400°C and higher. The formation of nickel carbide structures under ordinary diamond-growing conditions was usually observed but the role of carbide formation as an intermediate step in diamond growth needs more study. EXPERIMENTAL METHODS The melting and solidification temperatures of various nickel-graphite compositions were observed by latent-heat arrests in the sample holder illustrated in Fig. 1. The samples were composed of mixtures of carbonyl nickel powder and National Carbon Co.'s SP-1 graphite packed to densities of about 6.4 g per cu cm. They were contained within

Jan 1, 1965