Search Documents

By W. D. Robertson, E. G. Grenier, V. F. Nole

The electrical, mechanical, and corrosion cracking properties of an age-hardenable Cu-Ni-Si alloy have been studied over a range of time, temperature, and deformation states for the purpose of determining the relationship between the properties and the structural state. The precifitate has been identi3ed as and the sites of preferred precipitation have been located by electron microscope studies of the structures developed by various combinations of heat treatment and plastic deformation. An extreme form of deformation banding has been observed in the aged alloy that results in high strain concentration in bands lying parallel to {111} planes. These bands are the structural paths along which transcrys-talline cracks propagate in the deformed alloy. The observations provide a basis for a general mechanism of transgranu-lar corrosion cracking of face-centered cubic alloys in terms of stacking fault probability. AMONG the age-hardening copper alloys the copper-nickel-silicon (Silnic Bronze) type is outstanding in exhibiting a high strength combined with a considerable capacity for cold working, a relatively high conductivity, and a low to negligible susceptibility to stress corrosion cracking. It is not a new alloy and it has been the subject of a number of investigations,1-5 which were made primarily to establish practical composition and heat treatment limits. The structural characteristics of the aging processes in this alloy have not previously been studied in detail. This particular investigation was undertaken to explore, in detail, and over a wide range of conditions, the properties obtainable by various combinations of working and heat treatment and to correlate structural changes with the observed mechanical, electrical and corrosion properties. MATERIAL AND PROCESSING Eight different commercial heats of the alloy were made and used in the investigation. The range of composition among the eight heats is shown in Table I. Billets, 8 by 24 in., were cast into water-cooled molds, extruded to 1.687-in. rod, and quenched in water after extrusion. Extruded material was rod rolled to 0.750 in., annealed at 1450° F for 2.5 hr, and quenched in water. Subsequent working operations were performed by tandem rolling to various sizes depending on the degree of reduction required. Final cold-working operations were performed by drawing to a uniform size of 0.187 in. diam and straightening by a roller straightening machine (Lewis). The combinations of heat treatment, working and aging that were investigated are summarized in Table 11; the various treatments were performed in the order given in each row of the table, and they will be subsequently identified by the symbols shown in the first column. SOLUTION TREATMENT All material investigated in the cold-worked condition, or cold worked prior to aging, was solution

Jan 1, 1962

By Nicholas J. Grant, Laszlo J. Bonis

Two dilute nickel alloys in each of the systems Ni-Al, Ni-Ti, Ni-Cr, and Ni-Si u:ere internally oxidized at 700° to 900°C for time periods up to 100hr to establish the oxide particle size, depth of oxide penetration, and the change in oxide particle size with depth. Fine (-270) mesh alloy powders were then intevruzlly oxidized, hydrogen sintered, compacted, and extrmded, yielding bar stock for tensile tests at room temperature and creep-mpture tests at 815° to 1093°C. Recrystallization studies were made up to 1400OC. ReCENT studies have clearly indicated that oxide dispersion strengthened alloys exhibit outstanding elevated temperature properties, espe- cially in long-time tests. Certain structural parameters must be observed to achieve optimum properties, namely, micron or submicron interpar-ticle spacing, submicron oxide particle size, and a high level of stored energy.'-5 A number of processes have been utilized to obtain the desired structures, but certain of them appear to offer greater promise. Thus far the most promising method is the internal oxidation of dilute alloys consisting of a relatively noble solvent metal and a readily oxidizable solute element.3'6)7 Several important principles governing internal oxidation have been established in earlier studies;' however, these principles are as yet not adequate to permit alloy development studies. Much of the earlier work was performed on relatively coarse wires or in idealized systems. As a result, there are still a number of unanswered questions which require considerable additional work: a) How does oxide particle size change with depth of penetration of the oxide? How does this vary with the specific type of solute element? b) What is the effect of solute content on the average oxide particle size?

Jan 1, 1962

By S. J. Matas, R. F. Hehemann

The existence of two distinct forms of bainite—upper and lower bainite—in hypoeutectoid steels is confirmed by a systematic study of the structure of the product resulting from this mode of austenite decomposition. The transition between these two forms occurs at about 650° F regardless of the alloy content of the steel and their structural differences can be interpreted in terms of the kinetics of carbide precipitation from supersaturated ferrite. In the past few years, a renewed interest has been displayed in the kinetics and mechanism of the bainite reaction in steels. Several rather divergent models for the transformation mechanism have evolved from these studies,'-' however, it is difficult to evaluate these models critically on the basis of the existing experimental data. In particular, a more detailed knowledge of the structure of bainite should be helpful in providing a framework for the construction of an appropriate model. Metallographic studies7-l4 have led to a recognition of at least two distinct forms of bainite—frequently termed upper and lower bainite. Although the morphology of bainite changes with reaction tem- perature, there appears to be a relatively sharp transition between the two forms. This occurs in a narrow temperature range near 650° F. Transformation at temperatures below about 650° F results in the presence within the bainite of small plates of carbides oriented at an angle of 60 deg with respect to the direction of growth of the plate, while at higher temperatures carbides are oriented parallel to the growth direction. There is no generally accepted view regarding the significance of this structural transition. Some authors feel that carbides precipitate from austenite at all reaction temperatures1' while others claim, at least for low-temperature bainite, that carbides precipitate from supersaturated ferrite.14.15 It was recognized many years ago that the nature of the carbide in bainite depends on the reaction temperature. The carbide hase in upper bainite is unmistakenly cementite.16-19 The exact nature of the carbide in low-temperature bainite is less certain. Magnetic measurements of lower bainite reveal no cementite Curie point21'22 and dilatometric studiesz0 produce total expansions which are larger than those predicted on the basis of a ferrite-plus cementite structure. These effects have been attributed to the

Jan 1, 1962

By J. L. Taylor, P. Duwez

PARTIAL phase diagrams of titanium with iron, cobalt, and nickel have been established by previous investigators.1-3 These diagrams seem to be reliable, at least for concentrations of titanium ranging from 0 to about 50 at. pct. More recently, Long and his collaborators have presented a tentative titanium-nickel diagram covering the titanium-rich side.' These various studies indicate that in each system three intermediate phases exist and correspond to the compositions Tia, TiX, and TiX2 (or TiX3). The purpose of this paper is to review the available information on the crystal structure of these phases and to give the results of additional X-ray diffraction studies. Previous Work A summary of the work of Wallbaum and his collaborators"-' is presented in table I. The first intermediate phase on the titanium-rich side of the three systems, with iron, cobalt, and nickel, is of the Ti2X type and has a face-centered cubic structure. Although it is stated that the unit cell contains 96 atoms: no lattice parameters are given. The second phase is centered around the compositions TiX for the three systems. The crystal structure of this phase is mentioned in a short note" as being body-centered cubic (CsCl type), but no parameters are given. The analogy between the three systems, which is perfect as far as the two phases of the type Ti2X and TiX are concerned, disappears at the titanium-poor end of the diagrams. In the system titanium- iron, TiFe2 has an hexagonal (MgZn2 type) struc-ture with the parameters given in table I. The existence of TiFe3 has also been reported8 (table I) but was questioned by other investigators2 and is also refuted by the present study. In the system titanium-cobalt, a phase corre-sponding to the composition TiCo2 has been investi-gated by Wallbaum and Witte6 who claim that the structure is cubic (MgCu2 type) for alloys slightly deficient in cobalt and hexagonal (MgNi2 type) for alloys slightly deficient in titanium. The parameters of the two phases are given in table I. In the system titanium-nickel, the nickel-rich phase corresponds to the composition TiNi3 and has been described as hexagonal7 with the parameters given in table I. The present investigation was initiated for the purpose of filling in the gaps in the knowledge con-cerning the structure of the various phases sum-marized in table I. Since the titanium used by Wallbaum and his collaborators was only 95 pct pure,' it was found desirable to redetermine the lattice parameters of the alloys prepared with a

Jan 1, 1951

By R. W. Kraft

The normal Mg-Mg,Sn eutectic is a classic example of a Chinese script eutectic. When the alloy is unidirectionally solidified, a much simpler topo-logical arrangement of the phase particles can be produced although individual grains still usually consist of two or more interpenetrating lamellar systerns. Specimens from unidirectionally solidified ingots were analyzed metallographicall~ and crys-tallographically, and it was found that the interfaces between the lamellae had a common crystallograPhic character describable by: Lamellar interfaces II {220}Mg2Sn 1) {1011}Mg Arguments are presented to explain why this particular orientation relationship tends to occur and it is shown that this criterion can be nearly satisfied simultaneously on up to four different lamellar systerns between two interpenetrating single crystals. Finally, on the basis of this work ad other similar work which has previously been done on Al-CuAl2 eutectics, a general principle pertaining to low-energy interfaces between two crystalline Phases is proposed. WITHIN the past few years several investigators1-4 have demonstrated that the microstructures of many eutectic alloys can be significantly altered by uni- directional solidification. If appropriate control is exercised over the purity of the alloy, the growth rate, and the thermal gradient during unidirectional solidification, ingots can be made in which the phase particles are aligned predominantly parallel to one another throughout substantial portions of the ingot. The present author has also shown for at least one of the alloys that the parallel lamellae tend to assume a unique crystallographic relationship during

Jan 1, 1963

By J. Gordon Parr, A. J. Goldak

OgdEN et al.1 and Bumps et al.2 suggested that the solubility of aluminum in a titanium extended to 30 pct.* Sagcl,3 Clark and Terry,4 Anderko et al.,5 Ence and Margolin6 and Saulnier and croutzelles,7 subsequently proposed the existence of an intermetallic compound in the region of 25 pct Al. Although Anderko et al.5 suggested that the structure of this phase, Ti3Al, was of the DO,, (Ni3Sn) type, their conclusion was based only upon the results of a powder photograph that included reflections from ? = 0 to ?= 27° (CuK),some of which were reported to be indistinct. Ence and Margolin,6 referring to a 25 pct alloy (but without indicating whether the figure referred to weight or atomic pct) stated: "The structure of Ti2Al is isomorphous with Ti3Sn DO19, as pointed out by Pietrokowsky". But the point was not proven, because Ence and Margolin did not report values of calculated intensities of Ti2Al with the DO,, structure. In fact, the observed intensities they reported are in poor agreement with the values we calculate for DO19 (see Table I). Saulnier and croutzelle7 concluded, on the basis of electron micrography and microdiffraction studies, that the structure of the 25 pct A1 alloy was ordered hexagonal. Our purpose has been to unequivocally determine the Ti3Al structure. Alloys containing 25 pct Al were prepared from iodide titanium and spectrographic standard aluminum by levitation melting. The levitation chamber was evacuated to 10-6 mm Hg before filling with 99.999 + pct A. After casting, the alloy was chilled to liquid air temperature and crushed to -200 mesh in a percussion mortar. The powder, wrapped in molybdenum foil, was annealed (by induction heating) at 1000 "C for 2 hr in the argon atmosphere of the levitation chamber. No sublimate was seen on the foil or elsewhere, and therefore the assumption is

Jan 1, 1962

By J. F. Radavich

Many heats of steel of low carbon value have been known to produce brittle pieces of steel. The brittleness is believed to be due to the impurities located within the grain boundaries. Such brittle steels have been examined with an optical microscope to ascertain the nature and the amount of the impurities present at the grain boundaries. Due to the relatively low resolving power of the optical microscope, the impurities are not visible in fine detail. The writer obtained some sheet steel and proceeded to determine the location of the impurities and to show the application of the electron microscope to the study of grain boundaries. One sample was known to be capable of becoming embrittled, whereas another sample was believed to be much less susceptible to embrittlement. Treatment of Specimens The specimens were embrittled by annealing above the A3 point under mildly oxidizing conditions. One piece of ingot iron could not withstand a 90" bend, whereas another piece of ingot iron was not affected and could withstand a 90" bend. The brittle piece was then annealed at a high temperature in a hydrogen atmosphere. The annealed ingot iron was termed cured and could withstand a 90" bend very easily. The three specimens examined will be designated as brittle, good. and cured in the discussion that follows. Procedure The sizes of the specimens were as follows: one piece of brittle ingot iron-3/8 by 35 in.; one piece of good ingot iron-96 by 1/8 in.; one piece of cured ingot iron-36 by 54 in. The specimens were too small to be polished by hand and therefore were mounted in bakelite. The polishing procedure was carried out in the conventional manner with the use of 1/0 through 3/0 papers, and the final polish was done with alumina on a billiard cloth. The specimens were then etched in a 4 pct solution of picral in alcohol, and then they were examined through an optical microscope. An area was chosen that showed distinct grain boundaries, and an effort was made to keep near this area when pulling the replicas REPLICA TECHNIQIJE The replica technique used in the preparation of the replicas for examination under the electron microscope is described in Electron Metallography.' It consists essentially of the following steps: 1. Obtaining a suitably etched specimen. 2. Applying a swab of ethylene di-chloride on the surface. 3. Applying a formvar solution on the surface. 4. Placing a screen on any desired spot. 5. Breathing on the fornivar layer. 6. Applying scotch tape on the screen and film. 7. Pulling the film and the screen up with the Scotch tape. 8. Separating the screen from the Scotch tape. This replica technique is very similar to the one described by Harker and Shaefer. However, with the added step, the percentage of replicas removed is very much higher regardless of the length of the time from the etching of the specimen to the actual pulling of the replica. The replicas were then shadow cast with manganese at a filament height to replica distance ratio of 1 1/2:7. This produced a very high contrast replica for use in the electron microscope. One of the dificulties encountered with this study was the restricted area of the specimen. The width of the specimens was the same as that of the 200 mesh nickel supporting screen. In order to increase the effective area, the screens were cut down as shown in Fig 1. The arrow indicates the direction in which the replica was pulled. This operation made it possible to obtain a large percentage of good replicas. Fig 3 shows an electron micrograph of a brittle piece of ingot iron and a grain boundary that was polished mechanically. The surface is very rough probably due to the incomplete removal of the flowed layer by the picral etchant. The grain boundary does show evidence of impurities. It was decided to electropolish the specimens to obtain a much smoother surface than the one obtained by mechanical polishing. ELECTROPOLISHING The specimens were cut in half to expose the metal on the back side. The exposed metal had sufficient area to make good electrical contact and electropolishing was carried out easily. The conditions for electropolishing were 0.9 amp, 35 volts, and 25 sec. in an electrolyte composed of 850 cc of ethyl alcohol, 100 cc distilled water, and 50 cc of perchloric acid. The polished specimens were then etched in the 4 pct picral solution for a shorter time than was necessary for

Jan 1, 1950

By T. E. Davidson, A. P. Lee

It is known from the early work of Bridgman that the two lowest-pressure transitions (I-II and II-III) are accompanied by substantial and abrupt changes in resistivity and Volume. However, unlike the temperature -induced allotropic transformations observed in such elements as lithium, cobalt, tin, and so forth, there is little actually known about many of the characteristics of the pressure-in&ced transitions. This current work involves an examination of the structural and transformation characteristics of the bismuth I-II and II-III transitions under hydrostatic pressures. The relationship of initial structure to the transformation pressure, rate, resistivity change, and resultant structure is discussed. It is shown that the transition pressure and transformation rate are independent of the presence of grain boundaries and associated anisotropy-induced deformation. An observed hysteresis in both the I-II and II-III transitions is shown. BISMUTH is one of the most interesting of the elements exhibiting pressure-induced polymorphs since it undergoes several allotropic transformations at pressures below 90,000 atm. It is known from the early work of bridman1,2 that the two lowest-pressure transitions (1-11 and 11-111) are accompanied by substantial and abrupt changes in resistance and volume. However, unlike the temperature-induced allotropic transformations observed in such elements as lithium, cobalt, tin, and so forth, there is little actually known about many of the characteristics of these pressure-induced transitions. It is the purpose of this work to examine some of the structural and transformation characteristics of the bismuth 1-11 and 11-111 transitions under hydrostatic pressures. Another interesting characteristic of bismuth is that, in its polycrystalline form, hydrostatic pressures of sufficient magnitude will induce severe progressive plastic deformation in the region of the grain boundaries.3 This deformation, which has also been observed in several other metals, is attributed to the high degree of anisotropy in the linear compressibility of bismuth, resulting in shear stresses in the grain boundaries when it is exposed to hydrostatic pressure. Most thermally induced allotropic transformations in metals, whether of the diffusionless ather-ma1 (martensitic) or isothermal nucleation and growth types, are dependent upon structure and prior history,4 viz., grain boundaries, deformation, and so forth. One logically wonders then whether the transformation characteristics of the pressure-induced polymorphs in bismuth might also depend upon initial structure, particularly with respect to the presence of grain boundaries and associated plastic deformation. In this investigation, the role of grain boundaries and plastic deformation on the characteristics of the bismuth I-II and 11-111 transitions will be established. The rather unique residual microstruc-tural changes associated with these transitions will be presented and discussed. The occurrence of a measurable hysteresis in both the I-II and 11-111 transitions will be demonstrated. The type of transformation mechanism based on the observed transformation rate will be discussed. EXPERIMENTAL PROCEDURE A) Apparatus. The hydrostatic pressure system utilized in this investigation is similar to that previously reported by Bridgman' and Birch and Robertson,5 and has been previously described.3 For the purpose of this work, the pressure medium utilized was a 1:l mixture of pentane and isopentane. Pressure measurement was by means of a manganin coil in conjunction with a Foxboro Recorder. The manganin coil was mounted in the bottom closure and inserted inside the pressure cavity. Based on calibration against a controlled clearance piston gage at approximately 10,000 atm, the estimated error in the pressure measurement was +2 pct. Assuming the nonlinearity in the pressure coefficient of resistivity between 10,000 and 28,000 atm to be not greater than 1 pct, then the estimated error in the range of the I-II and 11-111 transitions was +3 pct. B) specimen Material and Preparation. The bismuth utilized throughout this investigation was of

Jan 1, 1964

By D. Turnbull

RECENTLY it has been shown that aggregates of small liquid droplets of tin,' mercury' or gallium' kept from intercommunicating by suitable films do not solidify at an appreciable rate unless the supercooling is very much greater than that usually necessary to cause a large continuous mass of the metal to solidify. For example oxide-coated tin droplets1 (1 to 10 micron diam) must be supercooled 100" to 110°C before their rate of solidification becomes rapid, although large continuous masses of liquid tin usually begin to solidify when supercooled 30" or less." These experiments have been interpreted' on the hypothesis that nucleation of crystals in large continuous metal samples is almost always catalyzed by accidental inclusions. Therefore, if the metal is dispersed into a number of isolated droplets large in comparison with the number of inclusions, most of the droplets should not crystallize until a temperature sufficiently below the thermodynamic melt- ing point, T0, has been reached to permit an ap-preciable rate of homogeneous (noncatalyzed) nu-cleation. In general this temperature is very much less than the temperature at which the rate of heterogeneous (catalyzed) nucleation is appreciable. If this interpretation is correct then investigation of the kinetics of crystallization of small droplet aggregates should be one of the most fruitful methods of obtaining information about the homogeneous formation of crystal nuclei in liquids. In this paper the results obtained on gallium and mercury are reported more fully. In addition, results on the supercooling of aggregates of liquid bismuth and lead droplets are included. Experimental Procedure Some care must be exercised in the choice of a barrier to prevent the liquid droplets from coalescing lest the barrier itself catalyze the formation of crystal nuclei. From this standpoint, the most desirable barrier is vacuum or inert gas, but, because of the difficulty of experimental arrangement, solid compounds and adsorbed films have been selected. There are two guiding principles in the selection of compounds as barriers in addition to the requirement that they prevent coalescence of liquid droplets. First, the compound should be almost insoluble in the liquid metal and second, its lattice structure should be quite unlike that of the forming metal crystal. Even so, there is no a priori assurance that the solid compound film will not catalyze to some extent the formation of metal crystal nuclei. An adsorbed protective monolayer is less likely to be catalytic than a crystal compound film, but it is not often possible to find a monolayer that is stable under the experimental conditions. Also, it is desirable, though not generally essential, that the aggregate of droplets be prepared by breaking up the massive metal rather than from a powdered compound of the metal.

Jan 1, 1951

By P. G. Shewmon, J. Y. Choi

The surface-diffusion coefficient for , has been measured on (111) and (100) surfaces of copper from 1000" to 620°C. D,(Ge) on the (111) is two to three times that on the (100) as was found earlier for copper and gold on copper. D, (Ge) on both (111) and (100) is about thirty times that found for gold or copper tracers on the same surfaces, while in the temperature range where both D,(Ge) and Ds(Au) have been measured the activation energy for germanium is the same as that for gold. The apparent activation energy for Ds(Ge) is not a constant, but decreases with temperature. In a recent paper we reported on the application of a new tracer technique to the determination of the surface-diffusion coefficient (D,) for gold and copper on (111) and (100) surfaces of copper.' In this paper, we shall give the results obtained when the same technique was used to measure Ds for germanium on the (111) and (100) surfaces of copper. In addition to providing data on D., for an additional element on copper, the longer half-life of germanium (275 days) meant that Ds could be measured down to a much lower temperature and thus over a much wider temperature range (1000" to 620°C) than was used in our gold-tracer work (1060" to 780°C) or copper mass-transport studies (1060" to 800). The apparent activation energy for D, (Ge) is essentially the same as that found for gold or copper in the same temperature range, though the values of D, (Ge) are thirty times greater than those found for gold or copper tracers. The most interesting new result found is that the apparent activation energy for Ds(Ge), and thus Do, definitely increases with temperature. Such a trend was much more apparent in Gjostein's mass-transport work on copper than in our own, though in an earlier paper we mentioned the difficulty of fitting all of our points with one straight line:= No real explanation of this is given though experiments done on grain boundary groove growth at various ambient pressures demonstrate unequivocally that vapor transport plays no part in the increase in activation energy with temperature.

Jan 1, 1964

By Brian F. Dyson

The surface tensions at 1550°C of some Fe-S alloys (in the range 0.008 to 0.052 wt pct S), Fe-Sn alloys (0.31 to 48.4 wt pct Sn), Fe-P alloys (0.038 to 2.38 wt pct P), Fe-Cu alloys (2.15 to 22.8 wt pct Cu), and Fe-1 pct C-S alloys (0.005 to 0.076 wt pct S) along with the surface tension of the base iron have been measured by the sessile-drop method. A mean value of 1754 dynes per cm was found for the surface tension of the base iron. Sulfur was found to be highly surface-active, the surface-tension results being in quantitative agreement with existing data. Tin and copper were found to be less surface-active than sulfur while phosphoms was completely nonsurface-active. The surface tensions of Fe-1 pct C-S alloys were found to be lower than those of the Fe-S alloys containing the same sulfur content. This was shown to be a mmzifestation of the increase in the thermodynamic activity of suZfur by carbon. It is only in recent years that attempts have been made to measure the surface tension of liquid iron of known high purity.1-3 Earlier measurements4-7 were made on liquid iron containing variable amounts of what are now known to be surface -active solutes. The exact value of the surface tension of liquid iron is still, however, open to some doubt. Halden and Kingery' reported a value of 1720k 34 dynes per cm at 1570°C, Kozakevitch and Urbain8 gave 1790k 25 dynes per cm at 1550°C, while Van-Tszin-Tan et al. obtained a value of 1865k 37 dynes per cm at 1550°C. The first systematic investigation into the effect of controlled solute additions on the surface tension of iron was made by Halden and Kingery.' They showed that sulfur and oxygen were highly surface-active, whereas nitrogen was only slightly active, and carbon inactive. A subsequent investigation by Kingery indicated that two other group-6B elements, selenium and tellurium, were also surface-active. This highly surface-active nature of sulfur and oxygen has recently been substantiated by Kozakevitch and Urbainla and Van-Tszin-Tan et al. l1 Kozakevitch and Urbainl2 have also conducted an experimental survey of the effects of a number of metals on the surface tension of liquid iron. Their surface-active nature was, in all cases, less than that of the group 6B elements. The present investigation was undertaken to study in more detail the surface tensions of dilute Fe-S alloys and to measure the surface tensions of binary alloys of iron containing phosphorus, copper, and tin. The effect of sulfur additions on the surface tension of Fe-1 pct C alloys was also determined. EXPERIMENTAL PROCEDURE The sessile-drop method was employed in the present investigation. An apparatus was built similar in principle to that described by Humenik and Kingery.lS It consisted of a horizontal silica tube, which could be evacuated to pressures less than 10-5 torr, with its central portion surrounded by a water jacket within which was a high-frequency coil. This generated heat in a tantalum susceptor placed inside the silica tube, which in turn radiated heat to the specimen mounted on a recrystallized alumina plaque. Temperatures were measured by an optical pyrometer and photographs of the molten drop were taken on a fixed-focus plate camera giving a magnification of X2. Surface-tension values were determined from the resultant drop using the method described by Baes and Kellogg.l4 The high vapor pressure of molten iron made it necessary to conduct all the experiments under a 1/4 atm of argon (greater than 99.995 pct purity). The analysis of the base iron used in the investigation is given in Table I. Each sample was approximately 3 g in weight and had a hemispherical base to ensure a uniform advancing contact angle on melting. The iron alloys were prepared individually in the sessile-drop apparatus by drilling a hole in the top of each sample and adding the required amount of solute, the drops being analyzed after the experiment. This method of preparation had the advantage of ensuring a consistent minimal contamination by oxygen due to refractory attack and also allowed surface tension to be measured at the same time. Every precaution was taken to ensure that the specimen was not contaminated by grease when it was introduced into the apparatus, the samples being cleaned in acid, dried in alcohol, and rinsed in petroleum ether. All handling was done with tweezers. Once the specimen had been placed inside the susceptor, the furnace was evacuated and the Sample leveled. The furnace was then degassed at approximately 1000"C before the argon was introduced. In every case the surface tension was determined at 1550" C.

Jan 1, 1963

By Benjamin C. Allen

The surface tensions of liquid chromium and manganese were determined by a modification of the dynamic drop-weight method and found to be, respectively, 1700 * 50 and 1100 * 50 dynes per cm at their melting points. Molten drops were formed LITTLE is known about the surface tension of liquid chromium and manganese, even though they are technologically important metals and widely used in steelmaking. In the past, direct determinations have been hampered by their high reactivity on the lower end of a vertically held rod inductively heated under argon. The technique essentially reproduced surface tensions reported for liquid nickel and zirconium using static drop-shape methods. and vapor pressure. In this work, surface tensions were measured by the drop-weight method, using induction heating under argon. EXPERIMENTAL WORK Chemical analyses of rods used are given in Table I. Chromium rods were prepared from iodide-process crystals which had been are-melted, hot-extruded, warm-swaged, and centerless-ground.'

Jan 1, 1964

By B. C. Allen

Liquid surface tensions of copper and 18 Group IV-A to VIII transition metals (Ti, Zr, Hf, V, Cb, Ta, Mo, W, Re, Ru, Rh, Pd, Os, Ir, Pt, Fe, Ni. Co) have been measured by the static pendant-drop and dynamic drop-weight methods on the same rod sample. The drops were formed by electron-bombardment heating in high vacuum. Application of necessary corrections enabled determination of surface tensions to k2 pct and agreement between methods to * 1 to 4 pct for all the metals studied. Evidence is presented that the liquid surface was well screened by metal vapor, suggesting negligible adsorption of gaseous impurities and that the surface tensions measured are characteristic of the metals involved. Correlations between liquid surface tension and melting point, molar volume, heat of vaporization, and atomic number are presented and discussed. Temperature coefficients of surface tension were calculated using Eötvös' law. SURFACE tension and interface energies of metals are the result of incomplete atomic coordination resulting in a force perpendicular to the boundary, tending to minimize its area. These energies are important in many phases of metallurgy, including microstructure, sintering, joining, electronic emission, and lubrication, which in turn affect many physical and mechanical properties of metals. Liquid surface tension refers to the interface between the liquid and its own vapor or nonreactive atmosphere, and is considered a physical property of the liquid. Since equilibrium shapes can be readily obtained with liquids, the units of surface tension (force per length) and interface energy (energy per area) are interchangeable. Thermo-dynamically, the surface tension ?LV is given by the change in free energy F with surface area A at constant pressure P, temperature T, and composition N: for commercially important refractory metals, vanadium, columbium, tantalum,13,14 molybdenum,14 and tungsten,l5 and none for the rare platinum-group metals and rhenium. Measuring the surface tensions of transition metals is difficult because of their high reactivity and melting points. Of the many techniques available,16-18 the pendant-drop 19-21 and drop-weight methods17,22 are considered superior because they can be modified to eliminate persistent sources of contamination such as supports and capillaries necessary in the popular sessile drop, capillary rise, and maximum-bubble-pressure techniques. The necessary modification is to melt a drop on the tip of a vertical rod, which provides support through a solid of the same composition.13,15 The object of the work was to study the surface-tension behavior of copper and transition metals, having reasonably low vapor pressures. Liquid surface tension of each pure metal was systematically determined by using a combination of the static pendant-drop and dynamic drop-weight methods on the same rod. Liquid drops were formed by electron-bombardment heating in high vacuum. EXPERIMENTAL WORK As indicated in Table I, the metals studied were high purity and generally were obtained in rod form. The exceptions were titanium and zirconium, which were machined from crystal bars, and molybdenum (tot 11, osmium, and ruthenium, which were sintered and arc cast into rods. All the metals were ground or swaged to desired sizes between 1- and 7-mm diam, and centerless ground round and smooth to a finish better than 50 µ in., rms. Each rod was thoroughly cleaned with steel wool, degreased, acid etched, and dried. In a typical run, the rod was vertically clamped in a modified floating-zone electron-bombardment furnace designed after Calverley23 and Carlson.24 The specimen was outgassed and maintained positive at several kilovolts in a dynamic vacuum of 10-5 to 10-7 mm. The bottom end was enclosed by a tantalum pillbox and heated by electrons emitted from a hot concentric tungsten filament mounted on a movable bracket. The bombardment power was slowly raised until melting was observed. Power requirements ranged from 30 w for copper to 1300 w for 4-mm tungsten. The stabilized and out-gassed drop of near-maximum size was photographed at 4. 1X on panchromatic film at desired time inter-

Jan 1, 1963

By A. J. Shaler, H. Udin, J. Wulff

In the study of the sintering of meta powders, we have come to the conclusion in this laboratory that further progress requires a more basic understanding of the operating mechanisms. This is emphasized in detail by Shaler. He has shown that a knowledge of the exact value of the surface tension is imperative for a solution of the kinetics of sintering. This force plays a principal role in causing the density of compacts to increase.2 Furthermore, a knowledge of the surface tension of solids is also applicable to other aspects of physical metallurgy. C. S. Smith3 points out the relation between surface and interfacial tension and their function in determining the microstructure and resulting properties of polycrystal-line and polyphase alloys. This paper describes one group of results of an experimental program designed for the study of the surface tension in solid metals. As a by-product of this work, considerable information has been obtained on the rate and nature of the flow of a metal at temperatures approaching the melting point and under extremely low stresses, a field of mechanical behavior heretofore scarcely touched by metallurgists. The importance of this additional information to students of powder metallurgy need not be stressed. Theoretical Considerations Interfacial tension arises from the condition that an excess of energy exists at the interface between two phases. Gibbs proves that this energy is a partial function of the interfacial area; thus: ?F/?s = ? where ?F/?s is the rate of change of free energy of the system with changing surface area, at constant temperature, pressure and composition, and ? is the interfacial tension, or interfacial free energy per unit area. If one of the phases is the pure liquid or solid, and the other the vapor of the substance, ? may properly be termed "surface tension," and is a characteristic of the solid or liquid. The attempt of a body to lower its free energy by decreasing its surface gives rise to a force in the surface which is numerically equal in terms of unit length to the free energy per unit area of the surface. Thus ? may be expressed either in erg-cm-² or in dyne-cm-1. Similarly, surface tension may be determined either by a thermo-dynamic measurement of the surface energy or by a mechanical measurement of the surface force. We have chosen the latter approach. Tammann and Boehme4 determined the surface tension of gold by measuring the amount of shrinkage or extension of thin weighted foil at various temperatures and interpolating to zero strain. The method is of questionable accuracy because of the tendency of foil to form minute tears when heated under tension. Their assumption of F = 2W?, where W is the width of the foil, is unsound, as the foil can decrease its surface area by transverse as well as by longitudinal shrinkage. Although their experimentation was meticulous, the paper does not include details of the sample configuration required for recalculating ? on a correct basis, even if such a calculation were possible. In the experimental procedure chosen here, a series of small weights of increasing magnitude are suspended from a series of line copper wires of uniform cross-section. This array is brought to a temperature at which creep is appreciable under extremely small stress. If the weight overbalances the contracting force of surface tension, the wire stretches; otherwise, it shrinks. The magnitude of the strain is determined by the amount of unbalance, so a plot of strain vs. load should cross the zero strain axis at w = F?. If balance is visualized as a thermodynamic equilibrium, the critical load is readily calculated. At constant temperature, an infinitesimal change in surface energy should be equal to the work done on or by the weight: ds = wdl [A] For a cylinder, s = 2pr2 + 2prl [2] If the volume remains constant, r = vV/pl [31 s = 2vpl+2V/l [4] ds = vpv/l - 2V/l²) dl [5] Substituting [5] into [I] gives for the equilibrium load, w = ?(z/rV- 2V/12) [6] and, again expressing V in terms of r and l, w = pr?(1 - 2r/l [7] Here the end-effect term, 2r/l, is neglected for thin wires in subsequent work. Eq 7 can be confirmed by means of a stress analysis. If the x-axis is chosen along the wire, then the stress is 2pr? - w pr² pr2 [8] A cylinder of diameter dis equivalent to a sphere of radius r, insofar as radial surface tension effects are concerned.³ Thus xv = 2?/d = ?/r = sz [9] For the case of zero strain in the x direction, the strain will also be zero in the y and z directions. Since the wire is under hydrostatic stress, Eq 8 and 9 are

Jan 1, 1950

By H. Udin

G. KUCZYNSKI* and B. H. ALEXANDER*—This paper represents a most noteworthy attempt to evaluate experimentally the surface tension of a solid metal. Because of the great importance of such measurements, any proposed method should receive the closest scrutiny before the results can be considered reliable. In regard to the experimental method, we think that the marking of the gauge length by means of tieing knots in the wire may be the cause of some of the spread in the results. Such a knot may be expected to tighten slightly, and thus increase the gauge length, when placed under stress at high temperature. Although this effect would be very small, amounting at most to only a few times the wire diameter. A fairly tight knot in a wire will decrease the wire length by about ten times the wire diameter, thus only a slight tightening of the knot would cause considerable spread in the results. Upon plotting the stress strain curves from the authors' data, the writers found that there was a fairly consistent tendency towards an S-shaped curve, instead of a straight line. Such an effect could be caused by the tightening of the knots. The writers think, however, that the experimental results are fairly reliable, but that there may be other methods of interpreting them depending upon what mechanism is assumed to be responsible for the shrinkage of the wires. The authors have assumed that the stress due to surface tension results in viscous flow. It should be made clear that it has never been demonstrated that viscous flow can occur in metal crystals even at very high temperatures. The experiments of Chalmers13 on tin, which are so frequently quoted as giving evidence of viscous flow at low stresses are by no means satisfactory. In his experiments, Chalmers found that only the initial rate of flow was approximately proportional to stress. He also found that the rate of flow varied markedly with time which, in his experiments, was less than 2 hr. Inasmuch as there is no proof of viscous flow in metals, and the authors have brought forth no conclusive evidence on this point, it may be worth while to investigate other possible mechanisms of material transport which would account for the shrinkage of the wires. The writers wish to point out that in these experiments the shrinkage of the wires can be adequately explained, according to a self diffusion mechanism. Thus, if we assume a concentration gradient for self diffusion which is a function of the radius of curvature of the wires, and assume that diffusion will occur so that the total surface area is decreased, we find the following expression for the self diffusion coefficient: where k = Boltzmann constant r0 = initial radius of the wire T = absolute temperature ? = surface energy 8 = interatomic spacing t = time e = strain at zero applied stress Eq 19 may be used to evaluate the self diffusion coefficient of copper, using the strain measurements obtained by the authors for zero stress as obtained by extrapolating their curves for 5 rail wires. By inserting a reasonable value for the surface energy (1500 ergs per cm2) we find: -66,000 D = 5 X 10e RT [20] The activation energy is of the correct order of magnitude, but the frequency coefficient is much too high, indicating that surface diffusion may be playing an important role. This discrepancy in the action constant is much smaller than the corresponding discrepancy obtained by the authors for the viscosity coefficient. The writers by no means propose that this proves that the shrinkage of the wires is due to self diffusion but we merely wish to point out that there are explanations other than that given by the authors. In this, as in any kinetic phenomena, it is necessary to study the rate of the process before anything can be said about the mechanism. The determination of surface tension given by the authors is based upon an interpretation of the data which embody the concept of viscous flow. The final proof of this concept will be obtained only after the time relationships confirming the authors' Eq 15 have been conclusively established. The rough linearity of the stress strain curves obtained by the authors for experiments run the same length of time should not be considered as proving that viscous flow is occurring. H. UDIN (authors' reply)—All of the test specimens were annealed at 1000°C for an hour or more before preliminary measurements were made. During this anneal the wires recrystallize, and the greatest part of grain growth takes place. Also, the knots sinter at the cross-over points. This does not in itself eliminate the possibility of end errors, although it greatly decreases their probable magnitude. It is still possible that some extension occurs due to creep in shear at the sintered points. If so, this effect would be quite independent of and superimposed on the normal shrinkage or extension of the wire itself. Within the precision of the experimental results, straight lines satisfy the data as well as do any other simple curves. Until data of greater precision are obtained, it is futile to discuss any possible trends away from linearity. The disagreement between Kuczynski and Alexander's Eq 19 and our Eq 18 is one of semantics and mathematics, not mechanism of flow, since Eq 18 is based on the self-diffusion concept of viscous flow. It would be interesting to learn how the mathematics leading to Eq 19 deviates from that of Eyring and of

Jan 1, 1950

By Michael Hoch, Seong Kwan Rhee

The ternary system Zr-Cr-0 was investigated at 1200°, 1500°, and 1700°C. The isotherms at these temperatures were determined by metallographic and X-ray diffraction analysis of carefully selected alloys. AS a part of a series of investigations1-4 of the ternary and quaternary phase diagrams on refractory metals and their compounds, the ternary system Zr-Cr-O was investigated at 1200°, 1500°, and 1'700°C. The binary diagrams Zr-Cr, Zr-O, and Cr-0 as reported by Hansen5 and Levin et a1.' and the existence of a ternary coompound ZrsCrsO (q-carbide structure a. = 11.96A) provided a good base for this investigation of the Zr-Cr-O phase diagram. MATERIALS The materials used and their suppliers were: Zirconium powder, 99.4 pct pure, Charles Hardy, New York; Zirconium dioxide, Fairmount Chemical Co., Newark, N.J.; Chromium powder, 99.85 pct pure, Fairmount Chemical Co., Newark, N.J.; Chromium sesquioxide, 99.86 pct pure, J. T. Baker Chemical Co., Pillipsburg, N.J. Though the materials used were not "hyperpure", the impurities present do not affect the results (lattice parameters, phase boundaries) within the experimental accuracy. EQUIPMENT AND PROCEDURE The detailed procedures and equipment used for this investigation were similar to those described elsewherel' "ith the exception that in this study an induction-heated argon-atmosphere furnace was used. The latter furnace was used to prevent the vaporization of chromium metal in samples in the higher temperature range (above 1200°C). The tungsten crucible was induct ion-heated in a vertical quartz tube filled with argon gas. The quartz tube was 1-1/2 in. in diameter and 10 in. long; it was connected at the top and bottom to water-cooled copper plates. The argon gas of a purity 99.995 pct was further purified by being passed through zirconium metal chips which were maintained at 800°C in a silica tube 1 in. in diameter and 30 in. in length. The furnace was flushed with argon gas for an hour before starting heating; during heating, a slight (3 cm oil) overpressure of argon was maintained in the system. Power was supplied by a 20-kw Ther-Monic Induction Generator. The samples rested on a tungsten-wire basket; the wire basket was connected to a tungsten wire which passed through the tungsten crucible lid, and entered the system on top through a Wilson-type seal. By pulling on the wire, the sample was removed from the crucible and dropped to the water-cooled copper base of the furnace. Because of the presence of argon gas, the cooling of the sample was even more rapid than in the vacuum furnace.' EXPERIMENTAL RESULTS AND DISCUSSION 1) Isothermal Section at 1200°C. The compositions investigated have been numbered within the phase diagram, Fig. 1, and all the pertinent information about the samples can be found in Table I. This table lists the composition of the samples, the phases present in the room-temperature diffraction patterns, the phases detectable in the microstruc-ture, and calculated lattice parameters of the a-zirconium phase. The shapes of the various phase

Jan 1, 1964

By N. J. Grant, D. S. Bloom

THE development of high temperature, high stress alloys had proceeded with such rapidity during the war, and for a short time afterward, that our knowledge of the constitution of the alloys had become seriously inadequate. To interpret correctly alloy behavior it is not only necessary to recognize the phases present, but also to know the properties of the phases under all varieties of conditions and treatments. Chromium in these high temperature alloys holds the unique distinction of being the only one element which must be present when an oxidizing or generally corrosive atmosphere is present and when a long life is desired in service. Carbon and nitrogen must also be considered where chromium is present. Carbon is always present, either as an impurity in the raw alloying elements, or as a deliberate addition agent to increase the strength. Nitrogen is generally picked up from the atmosphere during melting or may be added as an alloying element. The effect of nitrogen additions is not too dissimilar to that of carbon additions. As such, chromium carbides and nitrides become of great interest since they are present in a very large number of alloys, including alloy steels, stainless steels, super alloys, and the like. In the chromium-carbon system there are three accepted, carbides: Cr,C, Cr7C8, and Cr3C2. The existence of one more carbide, a still higher carbon form CrC, has been advanced. The phase diagrams of chromium and carbon as published by Friemann and Sauerwald² and by Tofaute, Kuttner and But- tinghaus³ -do not extend beyond the carbide Cr3C2, which contains 13.3 pct carbon; but Hatsuta has extended the diagram up to 20 pct carbon, and in doing so hypothesized, with the use of some questionable experimental results, the compound CrC. This carbide would contain 18.75 pct carbon, 81.25 pct chromium, by weight. It is further known that as the chromium carbides increase in carbon the acid resistance increases but the oxidation resistance decreases4,5 This is of great importance in high temperature alloys where time-temperature stability, as well as oxidation resistance, is of prime importance. The possible existence of a carbide of the formula CrC thus becomes of prime importance and it was considered important to recheck the chromium-carbon binary constitutional diagram for such a compound. In view of the wide spread of values listed for the melting point of chromium (1500° to 1850°C) and 1930°C8, it was considered .that the liquidus temperatures might well be redetermined more ac-

Jan 1, 1951

By N. J. Grant, D. S. Bloom

P. Coheur and L. Habraken—We read this paper with great interest and are glad to congratulate the authors for their valuable work, supplying an important contribution to the mechanism of tempering on the carbides in vanadium steels. We are in a good position to appreciate this work since we have been studying, for more than a year, Cr-Mo steels of a composition similar, but less highly alloyed (namely 0.1 pct C, 3 pct Cr, 1 pct Mo). Our results are quite in concordance with the authors', however, we wish to point out that it is very interesting to use equally the electron microscope in order to observe the variations of microstructure in steels. We have studied simultaneously the evolution of the carbides in situ with the electron microscope, and, after extraction, again with the electron microscope, and then with X-rays. The first results which we have obtained have been communicated at the Congress of Delft (July 1949) and at the "Journees de la Metallurgie" in Paris (October 1949). These may be summarized as follows: After tempering with varying either the tempering temperatures or the tempering times, the carbides show two evolutions. The first one is connected to the carbides existing in the steels, principally in the boundaries of the grains. For some groups of steels these carbides may coalesce and for other groups they may

Jan 1, 1951

By E. S. Lenker, H. J. Van Hook

A complete swies of solid solutions has been found between the compounds InAs and GaAs Below the solidus. The melting relations determined by differential thermal and static quenching technzques, indicate that the liquidus and solidus are of the simple ascendant type. No evidence of unmixing of the homogeneous solid solutions was observed at sub-solidus temperatures. A quenching method for locnting the solidus, using X-ray diffraction techniques, is described in some detail. THE semi-conducting properties of compounds formed by various combinations of the less familiar intermetallic elements have created an increasing interest in the phase relations in systems involving these elements. The complete phase diagram of such a system, containing a compound of special interest, will often provide information as to the most favorable conditions for preparing large crystals from a melt, as well as specifying the temperature stability of the crystalline phase or phases produced. In view of the current importance of gallium arsenide in semiconductor research and of the structural similarities of indium arsenide, it was thought that the complete phase relations of the system InAs-GaAs should be investigated. Previous work on the system Ga-As by Koster and homa' and on In-As by Liu and peretti2 has shown that GaAs and InAs are the only compounds formed, having melting points of 1238° and 942°C, respectively. In a survey of various systems of elements from groups IIIb and Vb, spengler3 also re- Mines Ltd., for his continuing encouragement and for permission to publish the results of this work. ports complete miscibility of gallium, indium, and arsenic in the liquid state. In a partial investigation of the InAs-GaAs system, Woolley and smith4 found extensive solid solution throughout the alloy range. Their subsolidus work, however, was inconclusive at temperatures below 900" and they did not determine the liquidus relationships of the system. PRESENT INVESTIGATION The starting materials used were high-purity gallium, indium, and arsenic. The method of preparation of the compounds follows that devised by M. Shafer and is described in detail in a paper by Shafer and eiser.' The end-member compounds, GaAs and InAs, were first prepared by reaction of the elements in the desired proportions in evacuated silica glass vials. The vials were rotated in a furnace as the temperature was slowly raised to insure rapid and complete reaction at low temperatures. Thereafter the starting materials, GaAs and InAs, were melted by heating the sealed vials to 1300" and 1000°C, respectively; the vials in this case were filled as completely as possible to keep the vapor space small and thereby minimize the loss of arsenic to the vapor phase. Ten binary mixtures of these compounds were then prepared by reaction of the powdered material in the solid state just below solidus temperatures. At 900C preparations containing 3, 5, 8, 10, 25, 35, 50, and 75 mol pct GaAs were found to be single homogeneous phases, indicating a complete series of solid solutions between GaAs and InAs. Moreover, it was discovered that the homogeneity of any crystalline binary preparation could be very accurately checked by X-ray diffraction because of the large differences in unit cell dimensions of the end membe5s. The cubic lattice constant ofor GaAs is a = 5.652A and that of InAs is a = 6.061A

Jan 1, 1963

By Michael Hoch, Michel Desjardins

The isothermal section of the Mo-Ti-Zr-0 system at 1500°C was investigated using X- ray diffraction and mefallographic techniques. Up to 66.7 at. pct 0, the system contains ten four -phase regions. Isopleths at 10, 30, 45, 55, and 60 at. pct 0 were constructed. THE purpose of this investigation was to determine the general shape of the quaternary equilibrium phase diagram of molybdenum, titanium, zirconium, and oxygen at 1500°C. The system was truncated at 66.7 at. pct o. PREVIOUS INVESTIGATIONS The Ti-Zr-0 system was investigated in this laboratory.' The binary systems of interest have been compiled and discussed by Hansen' and Levin, McMurdie, and Hall.3 Farrar, Stone, and Margolin studied the titanium-rich corner of the Ti-Mo-0 ternary. MATERIALS The titanium powder, 99.6 pct pure, the zirconium dioxide, and the molybdenum powder were supplied by the Fairmont Chemical Co. The zirconium powder,

Jan 1, 1962