Search Documents

By R. E. Reed-Hill, W. H. Hartt, W. A. Slippy

An unusual form of double kinking has been observed at the ends of {1121} twins in deformed zirconium. These kinks lie partly outside of the twin and partly inside. While they are undoubtedly closely associated with accommodating the twinning shear into the matrix, they are apparently closely related to other important aspects of twin growth. A mechanism is proposed for {1121} twin growth involving interactions between 1/3 (1120) dislocations in the bend planes and twinning dislocations. It is also shown that boundaries without long range strain fields develop at the tapered ends of the twins. The {1121} twins can, therefore, grow without forming high-energy noncoherent twin boundaries containing large numbers of closely spaced twinning dislocations. NUMEROUS investigators have indicated that slip is often closely associated with deformation twinning. An example is twin accommodation kinking,1-3 where dislocations form bend planes and permit accommodation of a twinning shear within the parent crystal. Another is the incorporation of slip dislocations into twins, as proposed by Sleeswyk and verbraak4 and Ishii and Kiho5 for iron and ß-tin, respectively, and observed in hcp metals by Price.6 In addition, it has been shown that shear discontinuities exist at the points of twin intersections in zinc,7 magnesium,8 titanium,9 ß-tin,10 and bismuth," where the continuity conditions for twin intersections, originally proposed by cahn,12 are not satisfied. A similar situation occurs for zig-zag (1121) twins in zirconium13 and ?-phase silver-aluminum alloys,14 where the twinning shears of the two component twins are not parallel. In all these latter cases additional deformation by slip is believed to accommodate the twinning shear discontinuity. The present paper is concerned with a newly observed aspect of the slip and twinning interrelation, in which two kinks, one inside the twin and the other external to the twin, are formed. These kinks, observed in conjunction with (1121) twinning in zirconium, form to accommodate the twinning shear. A model has been constructed explaining the growth and accommodation of (1121) twins in terms of these kinks. This mechanism may be significant since it is concerned with an important twin mode, which occurs not only in zirconium but also in titanium and hafnium. The practical significance of (1121) twinning has been adequately demonstrated. At 77° K the high ductility of zirconium containing a large number of grains unfavorably oriented for slip is primarily due to (1121) twinning.'5 Also, (1121) twins nucleated at 77°K may grow during further deformation at room temperature. This growth can greatly increase the transverse ductility15 as well as produce large mechanical hysteresis effects.= EXPERIMENTAL PROCEDURE Arc-melted, sponge zirconium plate previously described17 was prestrained 0.65 pct at 77° K by cold-rolling to introduce {1121} twins into the structure. Tensile specimens were cut with axes normal to the plate rolling direction. These specimens were pulled in tension to near the point of failure in an Instron machine at room temperature at a cross-head speed of 2 x10-2 in. per min. This additional deformation caused growth of the twins nucleated by prestrain. Deformed specimens were then prepared for metallo-graphic examination. This included an anodizing treatment18 that made the surfaces sensitive to polarized light, permitting small lattice rotations to be revealed. EXPERIMENTAL RESULTS Nature of the Twin-Kink Band Structure. Fig. 1 is a photomicrograph of a typical specimen after prestrain at 77° K. The structure is characterized primarily by long, narrow, parallel sided (1121) twins which norma1ly traverse the entire grain. Fig. 2 shows several (1121) twins after additional straining at room temperature. These lie near the center of the polarized light photomicrograph and are inclined upward to the right. It is evident that this latter deformation has resulted in appreciable twin growth. The dark regions extending from the ends of these twins are accommodation kinks. The difference in reflected light intensity indicates the

Jan 1, 1969

By L. C. Brown, A. J. Stewart

Phase transformations have been studied in Ag-Zn alloys quenched from the high-temperature bcc 0 phase and lying in the composition range 37.4 to 50.0 at. pct Zn. On quenching, the alloy with the lowest zinc content transformed to massive a (fcc); all other alloys formed the ordered (3' phase. On deformation, a martensitic reaction took place forming a close-packed structure. At Lou' zinc contents transformation to this structure was complete. However, above 43 at. pct Zn a p.rogressively increasing amount of ß' was retained following deformation. The close-packed structure was fcc at lower zinc contents (under 47 at. pct Zn) and either hcp or a mixture of fcc and hcp at high zinc concentrations. The structures produced by cold working were unstable and on aging transformed to the ß' phase or the complex hexagonal to phase. For low zinc contents (under 43 at. pct) the fcc st-ructure transformed directly to the equilibrium To phase. With higher zinc contents, some 13' was retained following defownation, and on aging the amount of this phase relative to the close-packed structure increased. On further aging both structures transformed cornpletely to PO. Metallographic examination indicated that the nature of the deformation product depended very markedly on the mode of deformation. Tensile deformation gave a microstructure characteristic of thermal martensite. KING and Massalski1 have shown that an alloy of Ag-50 at. pct Zn quenched from the high-temperature bcc ß phase undergoes a low-temperature ordering reaction forming ß' which can be retained to room temperature. On deformation an hcp phase is produced, although this is unstable and the structure reverts completely to ß' on aging. In the present paper, the work of King and Massalski is extended to cover a wide composition range in the p phase, observations being made of the structure of the quenched 13 and of the effects of deformation and aging. The relevant portion of the Ag-Zn phase diagram is shown in Fig. 1. King and Massalski have shown that close to stoichiometry the ß' phase is the equilibrium structure at low temperatures. The approximate composition range of this phase, taken from a recent paper by Pops and Massalski,3 is shown on the phase diagram. The locations of atom sites are very similar in the ß' and to structures, and the 4, structure can be derived from the 0' structure by a small movement of atoms mainly in the direction of the cube diagonal.4'5 However, the probability of filling any particular atom site by a silver or a zinc atom is markedly different in the two structures. The structure can be considered partly ordered but the arrangement of the ordered atoms is completely different from that in the p' structure and transformation from ß' to requires a large fraction of the atoms to change position. It has been shown that the ß' transformation is thermally activated involving normal nucleation and growth processes.5,8 Massalski and Barrett7 have shown that in the Cu-Zn binary the (3 phase can be retained in the ordered state to room temperature over a wide composition range. On deformation a new close-packed phase is produced, the structure varying from fcc to a mixture of fcc + hcp to hcp alone with increasing zinc concentration. It appears that an ordering reaction is necessary to retain the bcc phase to room temperature. In CU-Ga,8 the bcc phase cannot be retained to low temperatures as it transforms by a massive reaction above the ordering temperature of the ß phase. The structure of the massive phase is very similar to that in Cu-Zn, varying from fcc to fcc + hcp to hcp with increasing zinc concentration. Similar structures have also been observed in CU-A1,9 Ag-Cd,10,11 and Ag-Al.12 EXPERIMENTAL The alloys were prepared by melting fine silver (99.99 pct purity) and electrolytic zinc (99.999 pct

Jan 1, 1969

By H. W. Mead, C. E. Birchenall

SELF-DIFFUSION in pure gold has been studied by Sagrubsky,1 McKay,2 Gatos and coworkers,3,4 and Okkerse.5 Sagrubsky used a sectioning technique but reported results far lower than any of the others. The work seems to be carefully done, and it appears that a factor of 60 must have been omitted in his calculations or the units misstated by an equivalent amount. McKay used a surface activity method, and Gatos and Kurtz, following Gatos and Azzam, applied autoradiography to the measurement of activity on an oblique section through the diffusion zone. These two investigations agree well with each other but give results below those reported here. Okkerse employed a very careful sectioning technique. Chemical interdiffusion in Au-Ag alloys has been measured under several concentration conditions. Average values of the diffusion coefficient for the range of 0 to 18 atomic pct Au were obtained by Braune.6 since the diffusion coefficient probably varies markedly over this range, the values may not be too useful. Also, a plot of these data suggests that grain boundary contributions may have been appreciable at the lowest temperature employed. Ebert and Trommsdorf7 give average values for 0 to 9 atomic pct Ag, and Johnson8 reports average values for 45 to 55 atomic pct Au. Johnson also measured silver and gold self-diffusion coefficients in 49.2 atomic pct Au alloys, and self-diffusion in pure silver. The self-diffusion of gold in pure silver has been determined by Jaumot and Sawatsky.10 Other measurements which do not cover a range of temperature or are of low precision are not mentioned in this account since a review by Kubaschewski is available." This contribution reports diffusion coefficients for self-diffusion in gold, self-diffusion of gold in Au-Ag alloys, and chemical diffusion in Au-Ag alloys. Although it had been intended to obtain silver self-diffusion data in the alloys it now appears unlikely that these measurements can be done soon, so a table has been worked out for a self-consistent scheme of diffusion coefficients which should not be too far from the true values. Experimental Details Experiments on pure metals were conducted using fine gold of 99.99+ pct Au and silver of the same purity. Alloys were prepared by melting the two metals together in a graphite crucible in vacuo and allowing the slug to solidify in the crucible. After machining to a cylinder about 0.85 in. diam each slug formed one half of a diffusion couple. Compositions were determined later by analyzing chips taken from the extremities of the diffusion zones. The gold used in the preparation of alloys was sometimes pure gold but was normally either in the form of clean pickled alloy from previous experiments or gold recovered by oxalic acid precipitation from warm aqueous solution. Chips from the various stages of experimentation were dissolved in aqua regia and the solution taken to low bulk, diluted with water, and again taken to low bulk before diluting for the precipitation. This precipitated gold contained a small amount of silver as the only impurity detectable by spectrographic analysis. Disks for use in diffusion couples had opposite plane faces abraded on emery paper as far as zero grit, were annealed for 12 hr at 900°C in argon, and had one face of each disk polished to 000 emery paper. Radioactive gold was plated at about 2 ma per sq cm from a small volume of cyanide solution containing 1 mg of inactive gold as a carrier. Specimens , were freshly abraded on 4/0 emery paper immediately prior to plating. In the case of the active couples disks of a similarly prepared but unplated alloy or pure metal were placed on top and welded to the active disk. With the concentration couples disks of different compositions were welded. Welding was performed after a pressure of 5,000 lb had been applied at

Jan 1, 1958

By Michel Elle, John Chipman

The equilibrium of gaseous H20-H2 mixtures with liquid iron-columbium alloys in the range 0.2 to 2.4 pct Cb shows that the activity coefficient of oxygen is strongly diminished by Cb. The oxide in equilibvium with the melt is identified as CbO2. The equi-librium constant at 1600 deg C is [%Cb][fo. % 0]2 = 2.9 x 10-4 where log fo= —0.14 [%Cb]. The deoxidizing power is comparable to that of vanadium, COLUMBIUM (niobium) is finding increasing use as an alloying element in steel. It forms several stable oxides and is removed from the steel bath by oxidizing slags. The specific oxide formed in steel-making is not known although Sazanov snd Shvarts-man,' in their study of its partition between slag and metal, have assumed this to be Cb2O5 It is the purpose of this study to determine its deoxidation equilibrium at 1600°c, the nature of the oxide formed and the interaction between columbium and oxygen in the melt. The investigation was based on the reaction H2 + O=H2O; Kl=ph2o/(pH2[%O]) [1] in which 0 signifies oxygen dissolved in the liquid Fe-Cb alloy. The method and equipment of recent studies2,3 were utilized. In brief, the molten alloy, contained in a small crucible, is held in a controlled mixture of steam, hydrogen, and argon until equilibrium is reached. It is then cooled quickly in a stream of helium and analyzed for columbium and oxygen, the latter by vacuum fusion.4 The solubility of the oxide in liquid iron is determined by visual observation of scum formation on the surface of the melt. Crucibles of MgO proved unsatisfactory because of partial reduction at the metal interface. Pure A12O3 crucibles were heavily blackened by the melt and could not be used at high concentrations of columbium. Silica crucibles were more satisfactory but these introduced some silicon into the melt thus necessitating an uncertain but small correction for its effect on the activities of the other solutes. In silica it was possible to observe the scum formation when the melt became saturated and there was no apparent reaction between scum and crucible. The procedure for saturated melts was as follows. The alloy charge composed of electrolytic iron and columbium grain was melted and held at constant temperature. The gas ratio H2O:H2 was progressively increased by raising the saturator temperature until a point was reached at which a bright film appeared on the surface and covered it completely. This film could be dissolved by raising the temperature of the melt and it reappeared when the temperature was lowered. Under a partial film covering, utilizing bare areas for temperature measurement, the melt was brought to equilibrium with the atmosphere by holding at temperature for about 1 hr as suggested by Dastur.5 INTERACTION IN THE MELT Several experiments were made without columbium to establish a base in silica crucibles. The results are shown in Table I. Except for one run, they agreed well with Floridis1 comparable results in other crucibles. The value of log K1 is taken from Table I as 0.58. This is slightly larger than Floridis' result, probably because of less elaborate safeguards against thermal diffusion. Results obtained in the Fe-Cb-0 series in which no scum formed are shown in Table 11. The equilibrium ratio Kl is defined as ph2o/(ph2[%O]). It is related to the equilibrium constant K1 as follows:

Jan 1, 1962

By R. V. Smith

The theory for unsteady radial flow of gas, as simplified by Aronofsky and Jenkins, has been reviewed and compared with actual well performance. This study indicated that the theory should be modified by the introduction of an empirical "rate of flow" function "Y(q)". The expansion of the theory on the flow of gas to include a Y(q) term bridges the gap between the theory of unsteady-state flow of gas and actual gas-well behavior. Apparently, the Y(q) term is a function only of the rate of flow for a given well. The completion factor or skin effect is associated with the Y(q) function in such a manner that at least two or more sets of drawdown or pressure build-up tests are needed to separate the Y(q) function and the completion factor. Since the Y(q) function used in this report represents an energy loss dependent on rate of flow near the well-bore of a gas well which is in addition to the loss required by Darcy's law, the Y(q) function is related to the exponent of the back-pressure curve for a gas well. Techniques and equations are presented which permit the estimation of stabilized and "short-time" deliver-abilities. INTRODUCTION The published theory of radial unsteady-state flow of gas through porous media may be divided into papers1-" which use Darcy's law as a premise and other papers4-7,16-10 which recognize that Darcy's law may not hold for flow into gas wells. Miller8 compared theory based on Darcy's law with actual gas-well behavior and concluded that flow into gas wells deviates from Darcy's law. Houpeurt4,5 in a thorough investigation of flow into gas wells concluded that the flow rate into a gas well is proportional to the difference between the squares of upstream and downstream pressures elevated to a power between 1.0 and 0.5. Houpeurt also concluded that the deviation from Darcy's law is caused by irreversible, kinetic energy exchange between the flowing gas and the porous media. Tek9 presented a complete evaluation of the flow problem in gas wells. Several investigators state that the inclusion of gas compressibility into flow equations based on Darcy's law does not begin to explain the flow problems suggested by gas-well behavior. For many years, natural-gas engineers have realized that gas wells by actual test exhibit behavior which cannot be explained entirely by Darcy's law. Many tests on gas wells have indicated that a value of 0.85 for the exponent in the back-pressure equation is a fair average for many wells. However, Darcy's law either for steady or unsteady-state conditions indicates that the exponent should be 1.0. The purpose of this paper has been to examine this deviation from Darcy's law and to present a reasonable mathematical means of evaluating the deviation. The method presented in this paper combines previous theory with an empirical function to provide a description of gas-well behavior. All equations in this report are for constant terminal-rate-of-flow conditions. While present methods of testing wells usually are not based on this concept, they are believed to approach the constant terminal-rate conditions closely in many tests. However, several gas wells were tested by holding the rates of flow constant in order to approach more closely the constant terminal rate of flow; the results confirm the concepts presented in this paper. REVIEW OF THEORY FOR UNSTEADY-STATE FLOW OF GAS The basic partial differential equation for the flow of natural gas has been given in technical literature; the equation for a cylindrical reservoir with the well in the center is

By H. R. Hose

The total crude fluorspar reserves in Western Europe, including the USSR and the Soviet sphere, Thetotalare estimated to exceed 17 million metric tons. The total fluorspar production in 1951 in Western Europe, excluding the USSR and Soviet sphere, amounted to 371,109 metric tons. The chief fluorspar-producing countries in Europe, excluding Russia and the Soviet sphere, are France, Western Germany, Italy, Spain, and the United Kingdom. TOTAL crude fluorspar reserves in western Europe, containing more than 35 pct CaF2, are estimated at 12 million metric tons, while reserves in the USSR and Soviet sphere may exceed 5 million metric tons. The bulk of these reserves are concentrated in the USSR, Spain, United Kingdom, Western Germany, and Italy. Since published estimates of the tonnages of individual mines and explored deposits are not available for most countries in Europe, the estimates given in Table I for individual countries are necessarily approximate, as they have been compiled from various sources and modified and combined according to the author's judgment. The majority of fluorspar deposits in Europe belong to the late Paleozoic metallogenetic epochs and have been found within or along the borders of the Hercynian massifs. Deposits in the Central and Esterel massifs of France, those of Western and Eastern Germany, Sardinia, the Catalonian coastal range and the Sierra de los Santos in Spain, and those of Devonshire and Cornwall belong to the Hercynian metallogenetic epochs. The deposits in southern Norway and Sweden belong to the middle Paleozoic or later, those of northern England to the late Paleozoic or later, those of northern Spain and Greece to the late Mesozoic, and those of northern and probably central Italy and Switzerland to the early Tertiary. Fluorspar occurs typically in veins, lenses, and stringers, but also in bedded replacement deposits and pockets, and may be colorless to white, yellow, green, pink, brown, blue, or purple to almost black. Except for the colorless variety, the only kind which can be used for the optical industry, the color of the fluorspar does not affect its ultimate use. Crystalline cubic or octahedral fluorite of CaF2 composition, with 48.9 pct fluorine and 51.1 pct calcium, has a specific gravity of 3.18. The ore in place, however, usually averages 3.0 metric tons per cubic meter owing to the quantity of other minerals asso- ciated with the fluorspar in the veins and the physical state of the fluorspar, which may vary from sugary to massive crystalline or microscopic inter-growths with barite or quartz. Barite of 4.5 sp gr, quartz of 2.66 sp gr, and chert and calcite are commonly associated with fluorspar. Other minerals usually present in lesser amounts are celestite, wad, silver, lead, zinc, copper, iron, and uranium minerals. In some cavity fillings the rhythmic deposition of the fluorspar, barite, and quartz into successive crusts may make it possible to handpick the crude ore on belts or to separate the fluorspar from other minerals while the actual mining is carried on. Where metallurgical grade lump spar is required for foundries, separation of quartz and barite may be made also by jigs and heavy media, and in some mines the lead and zinc sulphides may be separated and recovered by vibrating tables. However, except in the case of unusually high grade deposits, acid grade fluorspar is obtained by flotation. By this process fluorspar deposits grading down to 35 pct CaF2 in the crude ore can be worked at a profit when fluorspar, lead, zinc, and silver prices are high, as the sulphides of these ores can be separated from the fluorspar by flotation and concentrated.

Jan 1, 1956

By R. G. Schaal

The Magma Mine at Superior, Arizona had an unimpressive beginning as a. worked out silver mine that was purchased in 1910 for $130,000 and then incorporated into the Magma Copper Company which has been continuously mining the deposit since 1915. Copper is the principal metal produced. However, zinc, gold, silver and iron have also contributed to the total metal product. Prior to 1973, the production rate of 1500 tons per day was maintained by hoisting 4 and 5 ton pay-loads at speeds up to 1100 FPM through four small hoisting shafts. The hoists used included 8' and 7' diameter double drum, double clutched doc. and a.c. motor driven geared units built as early as 1917. One old timer that did yoeman's service for many years was an 8' Bollen Pulley with a.c. geared drive that utilized an endless rope system with mechanical grip and release mechanism built into the pulley's throat. Primary power for all of these old production hoists was 25 Hertz. Exploration that was completed in 1967 developed a reserve of 10 Million tons of 5 per cent copper which was sufficient to justify new capital expenditures for enlargement of productive facilities to 3300 tons per day. To reach the new reserves a 22' I.D. concrete lined vertical shaft was sunk to an initial depth of 48431. A 12 ½ diameter adit was bored from the mill side of the mountain to intercept the shaft at a point about 1100' below the collar. This 9200' long adit provides the right-of-way for the mine to mill railway and egress for heavy supplies from rail head as well as a conduit for potable water, electric power, compressed air and mine drainage. Mine operating levels begin 3582' below collar and thereafter have been cut at 100' vertical intervals. Skip loading stations have been installed 4282' and 46821 below collar with an in-shaft skip dump station located 800' below collar. This provides for 4500 tons of storage above the mine to mill train loading station located 1100' below collar. Ore and waste are moved horizontally from stope chutes and drift headings by rail haulage to transfer raises which channel all material to the two mentioned skip loading stations. The 22' diameter shaft crossection is arranged with the two skip compartments on the south side separated from the manway and utility compartments by the cage compartment. The cage counterweight operates in a portion of the southwest quadrant of the shaft. The hoisting plant consists of two conventional Nordberg double drum, single clutch, geared, d.c. motor driven units that are identical except for the drives and automatic controls. The drums are 15' diameter x 105" face with LeBus grooving for 2-1/8" diameter wire rope wound in three layers. Mechanical design emp1o::ed is the same as has been used for the past twenty- five years. One drum is attached to the main shaft by 2 large bolted integral shaft flenge. The loose drum is engaged by a 4-arm positive tooth clutch meshing with radially cut teeth of the clutch ring which is bolted to the drum. The drums are one-piece steel plate weldments having a finished weight of 60 tons each and are mounted upon a 36" diameter shaft that also weighs 60 tons. With the addition of brake rings, clutch, bull gear and rope, the total dead weight adds up to nearly 250 tons that is supported by just three double row spherical roller pillow block bearings. There is no bearing between the drums. All pinion shafts as veil as all head sheave journals are supported in the same type of antifriction bearings.

Jan 1, 1975

By Donald A. Corrigan, John Chipman

It is shown that the heat of solution of nitrogen in liquid-iron alloys is Proportional to the interaction coefficient. This proportionality forms the basis for a method of predicting nilrogen solubility at any temperature when the interaction coefficient is knoun at 1600°C ov when the solubility is known at some one temperature. A method for predicting the solubility of nitrogen in liquid steels, including alloy steels, was proposed some years ago by Langenberg.' Since the data available at that time were restricted to temperatures of about 1600°C, his predictions were similarly restricted. More recently, Nelssonz devised an approximate method for estimating nitrogen solubilities at higher temperatures. It is the purpose of the present paper to present a method by which the solubility of nitrogen at any temperature within the steelmaking range may be predicted from the known interaction coefficients at 1600°C. Nitrogen in iron and in all of the alloys to be considered here is known to obey Sieverts' Law; hence the activity of nitrogen in an alloy of fixed composition is proportional to its concentration, here expressed in weight percent. The activity of nitrogen is defined as equal to its weight percent in the binary Fe-N solution at any temperature. Its activity coefficient in any alloy is the comparison being made at equal values of T and PN2. In all alloys the activity coefficient changes with composition, log fN being a linear function of the percent of alloying element at low concentrations. In the following calculations the alloy compositions are restricted to those in which this linear relation is exhibited. For most systems this is a fairly wide range, for example up to 5 pct V and to 10 pct Cr or Ni. For this composition range. complex alloy is the same as that in the Fe-N-j ternary solution. Langenberg's method of prediction was based on Eq. [3], the values of log fN for various additions being taken from a graph. Instead of using the simple addition indicated in Eq. [3], he used a more complex method suggested for Fe-S-j alloys by Morris and Buehl.3 For the composition ranges in which Eq. [2] is valid the two methods are equivalent. Nelson extended the prediction to other temperatures on the assumption that logfN is inversely proportional to absolute temperature. It may be shown that inverse proportionality would result if the partial molal entropy of nitrogen were unaffected by addition of the alloy element, an assumption which Nelson designated as a "regular" solution. It has been shown recently by the authors4 that this assumption is not a good approximation. The heat of solution of nitrogen (N2) in an alloy may be obtained if the solubility is known as a function of temperature. Measurements of this nature have been reported recently by Beer5 for Fe-Mn alloys, by Nelsson2 for Fe-C, by El Tayeb and parlee6 for Fe-V, and by Pehlke and Elliott7 for many alloys embracing a wide range of compositions. In connection with the work of the last named authors it is noted that their Fig. 22 gives the heat effect per half mole rather than per mole of N2. In using their data we have shifted some of their values slightly by using the best straight line in preference to their curve. Their data, supplemented by others from the literature, are summarized in Table I. In Fig. 1 the heat of solution of nitrogen is shown as a function of composition for the several alloy systems. The difference between the heat effect in the alloy. AH; and that in pure

Jan 1, 1965

By C. L. Meyers, A. V. Levy, G. Y. Onoda, R. J. Kotfila

This paper presents the hypothesis that solid-solution hardening of regions in the order of tens of angstroms thick along grain boundaries is the most important mechanism controlling the ductile -brittle transition behavior of poly crystalline bcc refractory metals. Further, a physical model is developed and a general equation is derived to describe the dependence of the solute concentration at grain boundaries on temperature, nominal impurity concentration, thermal treatments, vain size, and prior plastic deformation. Analytical curves of the solute concentration at grain boundaries us temperature, calculated by means of the derived equation, are found to have the same general form as curves of ductility us temperature plotted from experimental data. Senziquantitative relationships between the transition temperature and metallurgical variables can he determined by using the physical model. The room-temperature brittleness of the poly-crystalline refractory metals (tungsten, molybdenum, and chromium) is a result of their high ductile-brittle transition temperatures. Recently studies on single crystals have revealed that, when sufficiently pure, these crystals have much lower transition temperatures than polycrystalline specimens of similar purity.'-3 For example, the transition temperature (measured by reduction in area in tensile tests) for very high-purity tungsten crystals has been reported to be as low as -50°F whereas polycrystalline specimens of similar purity had a transition temperature near 400" ~.' These results suggest that the ductile-brittle transition behavior of these polycrystalline metals can be attributed principally to the influence of grain boundaries on plastic flow and fracture. This paper contains a discussion of the manner in which grain boundaries influence the ductility of the poly-.crystalline bcc refractory metals and proposes that the segregation of impurities to grain boundaries is the most important mechanism for controlling the ductile-brittle transition behavior of these metals. Further, this paper contains an analysis of the effect of the compositional and structural variables on the grain boundary-ductility relationship, and introduces a semiquantitative model that can provide a foundation for more theoretical treatment to describe the influence of these variables on ductile-brittle transition behavior. PHYSICAL MODEL FOR PREDICTING DUCTILE-BRITTLE TRANSITION BEHAVIOR Grain Boundary Segregation of Impurities. Based on their own observations and on theoretical grounds, numerous investigators have suggested that impurities segregate to the grain boundaries. Such segregation would be expected because regions exist where both large and small foreign atoms fit with less strain than within the undistorted crystal structure. Therefore, the concentration of impurities should be higher at boundaries than within the grain body. TO estimate the degree (i.e., the ratio of grain boundary to nominal bulk concentration) to which an interstitial solute tends to segregate at boundaries within a solvent metal, the solubility rather than the size misfit could be used as a guide because solubility expresses the joint effect of both elastic and electronic factors.5 The solubilities of interstitial elements such as carbon, hydrogen, nitrogen, and oxygen are low in the group V-A metals and very low in the group VI-A metals.6 These low solubilities suggest that segregation to grain boundaries should occur for each of these solutes in both refractory metal groups. The difference between the ductility of single-crystal and polycrystalline bcc refractory metals can be described in terms of two distinct effects.4 The first of these is the general observation that grain boundaries hinder slip within grains and slip transfer between grains, and the second is that deformation is more complex in each grain of a polycrystalline specimen than it is in a single crystal. Mc Lean has pointed out, furthermore, that strong slip-hindrance effects at grain boundaries should occur in metals having strong segregation of impurities to grain boundaries. Therefore, the effect of grain boundaries on the hindrance of slip is expected to be the dominating influence on the ductile-brittle transition behavior of these polycrystalline refractory metals. cahn5 has suggested that temperature and solute mobility should influence solute equilibrium segregation at grain boundaries, and Fig. 1 illustrates

Jan 1, 1965

By T. O. Anderson, E. J. Stahl

Tests to increase knowledge concerning the process of hydraulic fracturing have been performed on three wells in the Allegheny field in western New York. The fracture orientation and its physical appearance at the well-bore was determined by impression packers run at appropriate times throughout the program. Fract~tre azi171uth and its dimensions thus determined were compared to the rock properties and treatment conditions. Logs to provide data concerning rock properties and to improve identification of vertical fractures were run before each step of the fracture treatment. Pressure osc~illu~iot~s were studied after each phase of tile treatments surface pressures during treatment. These measurements show that subsurface pressures exhibit many of the same characteristics as surface pressures during treatment of a well; however, transient pressures are dampened and deelayed in time at the bottom of the well when COmpared with surface pressures. Data obtained from this study provides a means to help locale and determine orientation of vertical fractures made in the open wellbore. Although these rests were performed in only one area, much of the information gained may be applied to fracturing programs in orher areas. INTRODUCTION In this study, various tests were made to provide data to better understand formation fracturing in situ. The reservoir studied is the Richburg oil sand which is a part of a series of predominantly grey sandstones and shales with occasional brown sandstone layers, known as the Caneadea unit of the Canadaway group of Upper Devonian age. Average physical characteristics of 12 cored wells in this area are 13-ft thick gross sand formation, 12-ft thick net sand. 1.7 md permeability, 11 percent porosity, 44 percent oil saturation (before secondary recovery) and 16 percent water saturation (before secondary recovery). This particular area of the Allegheny oil field was originally drilled prior to 1890. The first waterflood secondary recovery attempt was made in 1929. The present sustained waterflood development began in 1951 and is being carried out on a 4.5 to 5-acre spacing. Objectives of these tests were to (I) study fracture azimuth and dimension in relation to rock properties and treating conditions, (2) improve fracture identification techniques, (3) compare surface and subsurface pressure during all phases of treatment and (4) compare the periods of the fluid column oscillations occurring after each phase of treatment. Although the areas of investigation listed appear divergent in purpose, the program was designed primarily to increase knowledge of down-hole conditions leading to rock failure and extension of the fracture thus created. TEST PROCEDURES Wells upon which tests were conducted are designated as Wells 1 through 3. Well 2 was located 440 ft north of Well 1, and Well 3 was located approximately 880 ft west of Well 1. Wells 1 and 2 were completed with cable tools, while the producing interval of Well 3 was rotary cored using fresh water as the drilling fluid to obtain oriented cores. Tests were divided into separate phases on each well as the fracturing program progressed and consisted of measurements performed prior to treatment, during and after the formation breakdown operation. during fracture treatment and after treatment. To verify location and azimuth of the fracture created, and to obtain an impression of wellbore irregularities, a packer was used (Fig. 1). This packer was lowered on tubing into the wellbore opposite the interval under study and inflated in stages to allow the impression material to conform to the wall of the wellbore. A magnetic compass was attached to the tool for orientation reference. Impressions made prior to treatment showed no existing fractures in the wells under study. Impressions made after both formation breakdown and treatment pro-

By L. W. Reynolds, R. F. Campbell, K. G. Carroll, S. H. Ballard

Based on metallographic and X-ray data probable equilibrium conditions from 1340" to 2400°F are presented for the composition range investigated. These are correlated with investigations of Takei and Kuo. Provision is made for existence of Mo, C and (Fe, Mo)2,C, carbides heretofore omitted in diagrams in this composition range. These carbides and (Fe, Mo),C, in equilibrium with ferrite and austenite, restrict the austenite and austenite + ferrite fields to lower carbon contents and higher temperatures than in the Fe-C system. OK groups of steel of different molybdenum content, but with approximately the same carbon variation in each group, Table I, were studied. The alloys were melted in an acid-lined 30-lb high-frequency induction furnace, and the ingots forged to l-in. rounds, which were then ground to 0.75-in. diam. To minimize chemical segregations a short length of each bar was given a high-temperature heat treatment as listed in Table I. Specimens 0.062 to 0.125-in. thick were thereafter cut from the homogenized bars for heat treatment to determine phase changes in the alloy. Heat treatments were performed in either a lead bath or muffle-type electric furnace, depending upon the temperature required. Maximum temperature variation was * 2" F in the lead bath and * 5"F up to 1900°F and +8"F above 1900°F in the electric furnace. Adequate protection of the specimen from the oxidizing atmosphere in the electric furnace was obtained by sealing the specimen in either a pyrex or a silica capsule which was evacuated or not as circumstances required. For austenitizing followed by quenching to an isothermal temperature, the specimen was placed in a closed-end refractory tube the mouth of which, extending beyond the furnace, was connected to a vacuum pump. The specimen could be withdrawn from this tube and quenched as desired. IDENTIFICATION OF PHASES The principal method of identification of phases was metallographic examination conducted on surfaces well removed from exposure to lead or atmosphere effects during heat treatment. Hardness measurements (Dph/lOkg) were also made on these surfaces. The etching reagents used and their effects are listed in Table II. Martensite and ferrite were readily recognized by their customary appearance after etching with saturated picral or 1 pct nital. The presence of liquid at high temperature was indicated by curved interfaces characteristic of boundaries between solid and melt, by intergranular pools of dendritic areas in a carbide matrix, and by a black network coincident with austenite grain boundaries. Examples of these structures are shown in Figs. l(a) and (0). The types of carbide which were attacked by the electrolytic chromic acid etch and the alkaline KMnO, etch, Table 11, were identified by X-ray analyses of extracted carbides of five alloys. Carbide residue was obtained from a metallographic specimen by etching the specimen for 24 hr in hydrochloric-picric acid in alcohol (grain-size reagent), after which the carbides, standing in relief, were scraped from the surface of the specimen and mounted on a glass slide. Diffraction patterns were obtained with a Philips spectrometer using iron-filtered cobalt radiation. Figs. 2, 3, 4, and 5 illustrate typical etching behaviors with chromic acid and alkaline KMnO,. Chromic acid attacked Mo,C as shown in Figs. 2(a) and (b)-but not (Fe, Mo),C, Fig. 3(b). Alkaline KMnO, attacked (Fe, Mo),C, Fig. 3(c), and also attacked M0,C. In accordance with these behaviors, a typical differentiation of coexisting Mo,C and (Fe, Mo),~ is shown in Figs. 4(a) and (b). The carbide (Fe, Mo),,C, could not be obtained free of other coexisting carbides for etching experiments. The behavior. however. of coexisting carbides identified by x-ray diffraction as Mo,C ind (Fe, Mo),,C, indicate that (Fe, Mo)23C6 is not etched by chromic

Jan 1, 1961

By G. H. Bruce, C. E. Williams

The trend toward deeper drilling, together with the attcndant increase in power requirements for circulation of the drilling fluid, has emphasized the need for a critical examination of the factors affecting the removal of bit cuttings from the hole by the drilling fluid. The ability of drilling fluids to lift cuttings is called their carrying capacity. A series of laboratory and field experiments has been conducted to determine the minimrim annular velocity necessary to remove cuttings, and to investigate the effects of properties of drilling fluids on their carrying capacities. Consideration of the results of these experiments led to the following conclusions: 1. Turbulent flow in the well annulus is most desirable from the standpoint of cutting removal. 2. Low viscosity and low gel are advantageous in removing cuttings. 3. Increase in mud weight is effective in increasing carrying capacity. 4. The carrying capacity is higher when the pipe is rotated than when it is not. 5. If turbulent flow can be maintained, an annular velocity slightly higher than the slip velocity of the largest cuttings to be transported should keep the bore hole clean. This implies velocities of 100 to 125 ft per minute rather than the presently used 175 to 225 ft per minute. INTRODUCTION Power Savings by Reduction of Annular Velocities A large portion of the power expended in drilling operations is consumed in circulating the drilling fluid. An important factor in establishing the rate of mud circulation is the minimum velocity in the annulus necessary to remove bit cuttings. Empirically, it has been found that average annular mud velocities of about 200 ft per minute will remove cuttings. It was not definitely known, however, whether annular velocities of about 200 ft per minute were just above the minimum necessary to remove cuttings, or whether such velocities could be materially reduced without sacrifice of the ability of the mud to remove cuttings. It is apparent that if annular velocities could be reduced without impairment of cutting removal, a considerable saving in power requirements would result. Need for Research on Carrying Capacity The ability of a drilling fluid to transport cuttings is called its carrying capacity. Although it has been recognized that the carrying capacity of a mud is affected by mud properties such as viscosity and density1,2,3,4 there have been various views in the industry as to the effects of these mud properties on carrying capacity. The economic importance of the problem of carrying capacity and the scarcity of information on the subject indicated that research on the problem was needed. THE FACTORS AFFECTING CARRYING CAPACITY Qualitative Determination of Factors The mechanism of cutting transport is closely related to that involved in the separation of material by settling processes. A considerable amount of research has been done on settling problems, and discussions of sedimentation theory can be found in standard texts.5,6 Consideration of the information available from these sources, together with consideration of the mechanism of cutting transport, leads to the conclusion that the factors affecting carrying capacity are the dimensions of the system, the physical properties of the cuttings. and the physical properties of the drilling fluid. system Dimensions and Their Effect The dimensions of the fluid circulating system of importance to carrying capacity are the bore hole size, drill pipe size, pump capacity, and pump speed. These dimensions determine the annular velocity of the drilling fluid. Physical Properties and Their Effects The physical properties involved in the interaction between mud and cuttings are the density and shape of the cuttings and the density, viscosity and gel strength of the drilling fluid. The effect of the density factor on carrying capacity is fairly obvious; high density difference between cuttings and fluid results in a low buoyant force and therefore decreases carrying capacity. The effect of cutting shape is less obvious. Although

Jan 1, 1951

By Keith R. Bock, Norman Parlee, Albert M. Barloga

Coils of pure iron and iron-carbon alloy wire (0.05 to 0.80 pct C) and sufficient sulfur to saturate the solid phase were equilibrated in evacuated or argon filled tubes. After rapid cooling, and removal of the outside nonmetallic layer, the wires were analyzed for carbon and sulfur and the data used to construct an Fe-S binary and isotherms of the Fe-C-S ternary in the range 1149° to 1427°C. THE solid solubility of sulfur in steel is of interest in connection with such phenomena as hot shortness, burning, and so forth. "Burning," the more or less permanent damage that some steels suffer when heated for forging or rolling, has been shown to be related closely to the behavior of sulfur and less closely to carbon and oxygen.' Attempts to interpret burning phenomena in steels fail because of lack of data on the Fe-C-S and the more complex systems in this family. Rosenqvist and Dunicz 2 and Turkdogan, Ignatowicz, and pearson3 have largely elucidated the Fe-S diagram in the region of interest but no information on the Fe-C-S diagram in this region appears to be available in the literature. This paper deals with the elucidation of the Fe-C-S diagram in these interesting ranges. The method employed is different from those used by previous workers2'3 on the Fe-S system. EXPERIMENTAL METHOD Pure iron wires (Ferrovac E) or iron carbon alloy wires of 1 mm in diam were cleaned with acid and acetone, coiled, and placed in silica tubes (7 mm OD and 5 mm ID) previously closed at one end. Enough sulfur was added to assure saturation of the solid iron phase. The filled tubes were either simply evacuated and sealed, or filled with argon at a reduced pressure and sealed. The argon was required at the higher temperatures to prevent collapse of the tubes. The filled and sealed tubes were placed in the middle uniform temperature zone of a Globar tube furnace and equilibrated at temperatures ranging from 2100°F (1149°C) to 2720°F (1493°C). After equilibration the tubes were removed from the furnace and quenched in air or water, the form of quenching being found to have no effect on the results. The tubes were broken open and the coils were placed in a 1 : 1 HC1 solution to remove the sulfide-rich layer. The coils were then cut into small pieces and analyzed for sulfur and carbon. In the early stages of the investigation different equilibration times ranging up to 17 hr were tried and the cores of the wires were analyzed to test for saturation. One hour appeared to be sufficient to reach maximum sulfur content at 1454°C and 2 hr sufficient at 1149°C. The practice adopted was to use at least 3 hr at the higher temperatures and at least 5 hr at the lower temperatures. The iron-carbon alloy wires used were made by carburizing pure iron wire with carbon monoxide gas in a one inch diameter ceramic tube at about 1204°C. Differing carbon contents were obtained by allowing fairly large coils to react with the gas for varying lengths of time at a flow rate of 800 cc per min. The reacting times ranged from 15 min to 4 hr depending upon the amount of carbon desired. The furnaces used were controlled by means of Leeds and Northrup Speedomax Type H instruments with D.A.T. Control attachment. Thermocouples were calibrated against the melting points of gold and copper. The temperatures recorded appear to be accurate within i2.7OC. Starting with run No. 85 and continuing with the

Jan 1, 1962

By A. R. Marder

Fe-C alloys have been investigated at temperatures below the eutectoid transformation to determine whether the superplasticity phenomenon exists for these materials. As a result of void formation at the carbide-ferrite interfaces that led to void coalescence and premature failure, the large elongations characteristic of superplustic matem'als did not develop. It had generally been assumed that the conditions for obtaining superplasticity include: a) T/Tm =0.5, b) a strain rate sensitivity of the flow stress that results in m s 0.3, and c) fine-grain structure. Although superphsticity was not observed in the alloys studied, stress-strain rate measurements indicated that an additional condition for the existence of superplasticitv in Fe-C allovs is that the fine-wain structure be equiaxed in agveement with previous results on Zn-Al al1oys. MANY investigations have been made on the hot deformation of iron-base alloys in order to obtain an understanding of the variables that may affect high-temperature strength and ductility.'-6 However, only a few papers have studied the effect of carbon content on the high-temperature properties of iron-base alloys. Peretayat'ko and Zaikov and Wallquist and carlen4 studied the effect of carbon on strength in the y temperature range, and recently Robbins, Shepard, and sherby6 studied the influence of carbon on the torsional behavior of iron in the a and cu-plus-cemen-tite temperature ranges. In addition to studying hot deformation behavior, current investigations have also been concerned with the phenomenon of superplasticity in nonferrous materials, and, in particular, ferrous alloys. Superplasticity is identified by the increased resistance to necking that results from a high strain rate sensitivity of the flow tress. ackofenr proposed the strain rate hardening coefficient m = d log a/d log e as a measure of strain rate sensitivity and showed that m must be "-0.3 for superplasticity to exist. In addition, a fine-grained microstrcture,- a eutectic or eutectoid ornosition,- and a test temperature in the creep range (TITm > 0.5)" have also been advanced as necessary conditions for superplasticity. Although many workers""23 investigated superplasticity associated with the allotropic transformation in iron and the eutectoid transformation in steel, and recently chadler " and orrison have studied superplasticity in the a - y phase region, no one has measured the superplastic effect in the a + cementite temperature region. Therefore, the purpose of this investiga- tion was to determine the effect of carbon content, microstructure, and test temperature on the strain rate sensitivity of Fe-C alloys at temperatures below the eulectoid transformation. EXPERIMENTAL PROCEDURE Fe-C alloys with carbon contents of 0.2, 0.4, 0.6, 0.8, and 1.0 wt pct were prepared as 100 lb 54 in. ingots by vacuum-induction melting. The compositions are listed in Table I. All the ingots were soaked at 2100°F for 2 hr prior to rolling to 4 in. plate. The minimum finishing temperature for rolling was 1800°F. The plate was sheared in 10-in. lengths and tensile bars were rough-machined prior to heat treatment. The samples were heat-treated according to the schedule in Table I1 in order to obtain a fine-grained microstructure consisting of finely dispersed carbides in a matrix of small ferrite grains. All the tensile specimens were heat-treated prior to finish machining, and tensile testing was accomplished on an Instron machine in conjunction with a clamshell furnace. The buttonhead specimen configuration is shown in Fig. 1. An inert atmosphere, helium, was maintained during testing with a steel tube shroud in a bell jar arrangement. Predried helium was continually purged over the buttonhead specimens and exhausted through ceramic fiber at the bottom of the tube to keep oxidation at a minimum and to prevent

Jan 1, 1970

By K. C. Hong, R. B. Jensen

The problem of determining the optimum set of steam volumes and cycle lengths for a single well undergoing multicycle steam stimulation in order to maximize the cumulative discounted net income has been formulated mathematically and programmed for a digital computer. The mathematical fonnulation of the problem and the method for its solution are discussed in this paper. The oil production performance during each stimulation cycle was simulated by either a constant percentage or a harmonic decline. Using simple analytical expressions for production performance, the cumulative discounted net income and cumulative time of operation were related to the pertinent process and cost parameters and two principal process control variables (cycle length and steam injection volume). The discrete maximum principles was used to transform the equations for cumulative time of operation and cumulative discounted net income into a set of simultaneous equations. The simultaneous equations then were solved by trial and error on a digital computer to determine the set of cycle lengths and stem injection volumes that gives maximum cumulative discounted net income over the project life. INTRODUCTION Steam stimulation is a process for improving the oil recovery rate from wells producing high-viscosity crudes. The process is applied on an individual well basis and is executed in a series of cycles, each consisting of three phases: steam injection, soaking (steam condensation), and production. Significant increases in production rate following the stimulation operation result from heating the reservoir around the wellbore. As heat is removed with produced fluids and dissipated into nonproductive formations, the production rate declines, usually to near the prestimulation value. Typical production responses are given in case histories reported by Owens and Suter,l and are depicted in Fig. 1. Duration of the production phase is equal to the time for the oil production rate to decline to some specified value and is called the cycle length. (Cycle length is defined as the producing portion of the cycle and does not include the downtime required for steam injection and soaking operations.) Termination of the production phase coincides with the start of steaming for the next cycle. The process is continued, cycle by cycle, until it becomes unprofitable. Models 2-4 have been developed to simulate behavior of a single well during one cycle of steam stimulation and have been used to investigate the effects of system and operating conditions on the production responses. However, these investigations have not shown how the models can be used to determine a most profitable set of operating conditions for a multicycle stimulation project. Perhaps these models are too complicated mathematically to be adapted readily to an optimization study. This paper presents a method for optimizing a multicycle steam stimulation project based on simple production performance models. First, the performance of a steam-stimulated well during each cycle was simulated by either a constant percentage or a harmonic decline model. Using these simple analytical expressions for production performance, the cumulative discounted net income (before tax) and cumulative time of operation for each cycle were related to the pertinent process and cost parameters and two process control variables—the cycle length and the steam injection volume. Finally, the profit function to be maximized was formed. The analogy between the above problem and problems of optimizing multistage decision processes that have been reported extensively in the chemical engineering literatures5-10 in recent years has led this investigation to the use of the discrete maximum principle.5-7 This principle was used to transform the equations for cumulative discounted net income and cumulative time of operation into a set of simultaneous equations characterizing the optimum operating conditions in terms of the two control variables. These equations

Jan 1, 1970

By C. E. Williams, G. H. Bruce

The trend toward deeper drilling, together with the attcndant increase in power requirements for circulation of the drilling fluid, has emphasized the need for a critical examination of the factors affecting the removal of bit cuttings from the hole by the drilling fluid. The ability of drilling fluids to lift cuttings is called their carrying capacity. A series of laboratory and field experiments has been conducted to determine the minimrim annular velocity necessary to remove cuttings, and to investigate the effects of properties of drilling fluids on their carrying capacities. Consideration of the results of these experiments led to the following conclusions: 1. Turbulent flow in the well annulus is most desirable from the standpoint of cutting removal. 2. Low viscosity and low gel are advantageous in removing cuttings. 3. Increase in mud weight is effective in increasing carrying capacity. 4. The carrying capacity is higher when the pipe is rotated than when it is not. 5. If turbulent flow can be maintained, an annular velocity slightly higher than the slip velocity of the largest cuttings to be transported should keep the bore hole clean. This implies velocities of 100 to 125 ft per minute rather than the presently used 175 to 225 ft per minute. INTRODUCTION Power Savings by Reduction of Annular Velocities A large portion of the power expended in drilling operations is consumed in circulating the drilling fluid. An important factor in establishing the rate of mud circulation is the minimum velocity in the annulus necessary to remove bit cuttings. Empirically, it has been found that average annular mud velocities of about 200 ft per minute will remove cuttings. It was not definitely known, however, whether annular velocities of about 200 ft per minute were just above the minimum necessary to remove cuttings, or whether such velocities could be materially reduced without sacrifice of the ability of the mud to remove cuttings. It is apparent that if annular velocities could be reduced without impairment of cutting removal, a considerable saving in power requirements would result. Need for Research on Carrying Capacity The ability of a drilling fluid to transport cuttings is called its carrying capacity. Although it has been recognized that the carrying capacity of a mud is affected by mud properties such as viscosity and density1,2,3,4 there have been various views in the industry as to the effects of these mud properties on carrying capacity. The economic importance of the problem of carrying capacity and the scarcity of information on the subject indicated that research on the problem was needed. THE FACTORS AFFECTING CARRYING CAPACITY Qualitative Determination of Factors The mechanism of cutting transport is closely related to that involved in the separation of material by settling processes. A considerable amount of research has been done on settling problems, and discussions of sedimentation theory can be found in standard texts.5,6 Consideration of the information available from these sources, together with consideration of the mechanism of cutting transport, leads to the conclusion that the factors affecting carrying capacity are the dimensions of the system, the physical properties of the cuttings. and the physical properties of the drilling fluid. system Dimensions and Their Effect The dimensions of the fluid circulating system of importance to carrying capacity are the bore hole size, drill pipe size, pump capacity, and pump speed. These dimensions determine the annular velocity of the drilling fluid. Physical Properties and Their Effects The physical properties involved in the interaction between mud and cuttings are the density and shape of the cuttings and the density, viscosity and gel strength of the drilling fluid. The effect of the density factor on carrying capacity is fairly obvious; high density difference between cuttings and fluid results in a low buoyant force and therefore decreases carrying capacity. The effect of cutting shape is less obvious. Although

Jan 1, 1951

By Henry Leidheiser, Melvin C. Jr. Hobson

The rate of formation of the intermetallic compound, indium antimonide, at the interface between iudium and antimony at 100°C is greatly increased when a composite electrode of indium electrode -posited on antimony is made the cathode in an electrolytic cell. The amount of indium antimonide formed during electrolytic treatment at constant potential appeared to he proportional to the square root of time, but the scatter of the data, particularly at long times, makes it impossible to state this conclzlsi1:ely. The rate constant for the electrochemical synthesis was estimated to he 1.5 x 10-12 sq cm per sec. A similar- set of experiments was conducted in the absence of an applied potential but under otherwise similar conditions. The rate constant lor the thermal synthesis was less than 2 x 10-14 sq cm per, sec, at least two orders of magnitude less than obtained during electrolytic treatment. In a series of electrochemical studies it was observed that intermetallic compounds formed on the surface of polarized electrodes.1'2 The possibility of using this technique for the electrochemical preparation of intermetallic compounds was investigated and subsequently applied to the synthesis of the semiconducting III-V compounds. Gallium antimonide, indium arsenide. indium antimonide, and indium bismuthide were successfully formed by electrodepositing the Group III element from a boiling acid solution onto an electrode made of the Group V element. The following kinetic studies were undertaken on indium antimonide to obtain quantitative information on the rate of formation of this representative III-V compound under a potential gradient. EXPERIMENTAL The electrolytic cell for these experiments was a modified, three-neck, round-bottom, 500-ml distillation flask. Sealed into the side of the flask was a Luggin probe which monitored the surface of the working electrode and extended out of the flask to a small reservoir containing a saturated calomel electrode. The counter electrode was platinum. The three electrodes were connected to a Wenking po-tentiostat which maintained a preselected constant potential near the surface of the working electrode with reference to a saturated calomel electrode. All potentials noted in this paper are with reference to a saturated calomel electrode at 45° ± 5°C. The potential and the current were both monitored using Sargent Model MR recorders. The antimony electrodes were cast in graphite molds to form spheres about 15 mm in diameter with a shank 6 mm in diameter and 10 mm long. A heavy copper wire (12B & S gage) was soldered to this shank. Collodion was applied for protection to the shank and the connecting part of the copper lead. A snug-fitting O-ring was slipped over the shank and the electrode was then fitted into a teflon holder such that only the spherical portion was exposed. The electrode surfaces were cleaned and chemically polished prior to assembly and introduction into the electrolyte by etching in a modified CP-4 solution (50 parts glacial acetic acid, 10 parts concentrated nitric acid, 2 parts 48 pct hydrofluoric acid). The electrolyte was 10-2 M In+3 in 1 MH2So4 and its temperature was held constant by refluxing. The reflux condenser was open to the atmosphere and no temperature correction was made for changes in barometric pressure. The indium concentration in the electrolyte was checked from time to time by the analytical method outlined below for the determination of the amount of indium antimonide formed electrochemically. A schematic of the recorded potential is shown in Fig. 1. At point A the antimony electrode was introduced into the cell and the open-circuit potential recorded. At point B the potentiostat, preset at the desired potential, was switched on. It was switched off at point C and the potential fell to the open-circuit value of the electrodeposited indium. The

Jan 1, 1965

By Wayne L. Worrell

Thermodynamic data for the stable carbides and oxides of chromium, molybdenum, and tungsten have been critically eualuuted and are used to determine the stable condensed phases at 1 atm total pressure in each metal-carbon-oxygen system. Pourbaix-Ellingham diagrams have been constructed and are used to estimate the minimum temperatures neces.see)*y to oblain these metals by corbolhermic redlrction under reduced pressures. USING recently reported data for the carbides and oxides of chromium, molybdenum, and tungsten, the stable phases in the Group VI metal-carbon-oxygen systems at specific temperatures and pressures can be determined. The thermodynamically stable regions in each metal-carbon-oxygen system can be completely pictured using a Pourbaix-Ellingham diagram in which the two coordinates are oxygen potential (-RT In Po2) and temperature. Worrell and chipman' have described and constructed such diagrams for the metal-carbon-oxygen systems of vanadium, columbium, and tantalum. Downing2 has used similar diagrams to provide a description of the reactions in submerged arc furnaces. One of the examples which he cited was the Cr-C-O system; however, an improved diagram is presented in this paper using recently obtained data for the carbides of chromium. Although some of the oxides and carbides of the Group VI metals exist over slight ranges of composition,3 all solid phases are represented as stoi-chiometric compounds in the subsequent diagrams. Any error introduced into the calculated equilibria by neglecting the compositional variation of these phases is less than the uncertainty in the thermodynamic data. In a Pourbaix-Ellingham diagram for a three-component system, the phase rule requires that at equilibrium four condensed phases specify a point, three condensed phases determine a line, and two condensed phases define an area. To obtain a well-defined area at high oxygen potentials and teniperatures, the initial carbon to oxygen mole ratio is specified to be more than sufficient to reduce the oxide but not enough to convert all the metal to carbide. Thus, the lowest regions in the subse- quent Pourbaix-Ellingham diagrams represent areas in which the metal and its carbide are the stable condensed phases. Cr-C-O SYSTEM A) Thermodynamic Data. To simplify the subsequent calculations, all thermodynamic data are presented in the form of linear Gibbs-energy-of-formation* equations which were derived for the temperature range 505° to 2000°K and are reasonably accurate over the entire temperature ranges of the diagrams. The thermodynamic data for the compounds occurring in the Cr-C-O system are summarized in Table I. Several volatile chromium oxides have been reported,4 but their influence is negligible in the Cr-C-O system. Thermodynamic data for CO and CO2 were taken from the compilation of Coughlin.5 The only solid chromium oxide sufficiently stable to exist below 1800°K in the Cr-C-O system is Cr2Os.11 The Gibbs-energy-of-formation equation for Cr2O3 was obtained by revising the data tabulated in Coughlin,5 using the heat of formation of ah' and the high-temperature thermal data for chromium.7 In 1953 Richardson8 re-evaluated the original work of Kelley and coworkers,9 who measured the thermodynamic properties of the chromium carbides. However, the high-temperature equilibrium results of Boericke9 yield values for ?G°f of Cr2O3 which are 2 kcal more positive than the accepted data. His results for the chromium carbides are based on the same questionable equilibria. Gleiser10 recently determined the Gibbs energy of formation of Cr3C2 and the equation in Table I was calculated from her data and the thermal data of Kelley and coworkers.9 Thermodynamic data for the Cr7C3 and Cr23C6 carbides were obtained by combining the Cr3C2 data

Jan 1, 1965

By M. L. Hammond, G. M. Bowers

Silicon dioxide films have been prepared on silicon wafers by the oxidation of SiH, near 400°C. Results previously reported by Goldsmith and Kern7 have been qualitatively corrobovated using a system of different geometry, and significantly greater deposition rates have been achieved. The deposition rate is independent of thickness and nearly independent of the O2/SiH, ratio over a wide range; however, the rate is strongly temperature-dependent. The structure of the deposit depends upon the cleanliness of the substrate, and the electrical field strength of the deposited oxide depends upon the nature of the electrodes. MULTILEVEL interconnects in integrated circuits using aluminum metallization require an insulator that can be deposited at low temperature. Because the A1-Si eutectic is only 577°C, the maximum deposition temperature must be below about 500°C to prevent reaction of the aluminum with the silicon and with other materials in the integrated circuit. Many methods of depositing insulating films of SiO, and other dielectrics have been reported'-6 but the technique described by Goldsmith and Kern7 offers the simplest method of preparing SiO, films. The technique involves the reaction of SiH, and 0, to form SiO, films in the 300" to 500°C temperature range. Previous work using this gas system has been reported8 and the initial development work at Philco Microelectronics was done at the East Coast R&D Laboratories. 9 This paper corroborates some of the results of Goldsmith and Kern,7 shows that much higher deposition rates can be obtained than previously reported, and presents some observations on the structure of deposited SiO, films. EXPERIMENTAL CONDITIONS Fig. 1 is a schematic drawing of the apparatus used in this study. It differs from that of Goldsmith and Kern in that 1) the gases are mixed before they enter the reaction chamber, 2) there is no fritted glass diffuser, and 3) the wafers are stationary during deposition. The control thermocouple was in a well located 0.06-in. from the substrate, and three gas supplies were used: 1 pct SiH4 in N 2, O 2, and N 2. Water vapor must be excluded from the plumbing to prevent the formation of colloidal SiO, and the SiH4 must be diluted with an inert gas because SiH4-0, mixtures are extremely flammable above about 4 pct SiH The wafer temperature was determined by attaching a thermocouple to a test wafer and the temperature was maintained to +5 C. The conditions used in this investigation are listed in Table I. The temperature range was chosen for compatibility with the A1-Si system. The volume ratio of 0, to SiH4 is the actual volume of 0, delivered to that of SiH4, taking into account the initial dilution of the SiH4. This ratio, which was used by Goldsmith and Kern,7 is a useful indicator of the experimental conditions. The active/inactive ratio is the ratio of the volume of active gases, SiH4 and O2, delivered to the volume of dilution gas. The active/inactive ratio is an important parameter to specify for deposition rate determinations since it determines the amounts of reactants that arrive at the sample per unit time. Fairly high dilution ratios were necessary to achieve uniform deposits of SiO,. Most of the process characterizations were performed on the polished (111) face of 3 to 5 ohm-cm, n-type silicon wafers which were 0.009 in. thick; but the data approximately apply to deposition of SiO, on thin films of aluminum. It is difficult to obtain accurate deposition rate data over aluminum films because of coarsening the aluminum surface during exposure to 300" to 400°C. The deposited SiO, film fol-

Jan 1, 1969

By J. P. Hammond, C. J. McHargue, S. E. Adair

Cold-rolled sheet textures have been determined for binary solid solutions of aluminum, columbium, tantalum, and zirconium in titanium. An alloy containing 3.8 pct Al had a (0002) [10101 texture, whereas alloys with the other elements had rotated (0002) [10101 textures similar to that of iodide titanium. THERE have been relatively few studies of the effect of alloying on deformation textures. A second element may be present in the base metal in solid solution and/or as a second phase. The literature offers no satisfactory explanation for the effects of solid solution alloying on textures noted thus far. For example, there is no satisfactory explanation of the difference in the textures of copper and a brass. The most complete study of the effects of solid solution alloying on deformation textures has been that of Brick, Martin, and Angier.¹ These investigators reported that the copper texture was retained by alloys of copper with up to 30 pct Ni, less than 5 pct Zn, less than 2 pct A1 or Co up to the limit of solubility. Copper containing more than 5 pct Zn, more than 2 pct Al, or the maximum soluble concentration of tin, silicon, or manganese showed the deformation texture of a brass. Bakarian² reported that additions of 0.20 pct Ca to 99.98 pct Mg or 2.0 pct Mn plus 0.15 pct Ca to magnesium resulted in entirely different textures than for the pure magnesium. The texture obtained for high-purity magnesium sheet showed the basal plane to be concentrated parallel to the rolling plane. The texture obtained for the alloy containing 0.20 pct Ca had the basal planes concentrated in elliptical areas whose centers were tilted 15" from the center of the pole figure toward the rolling direction. This split basal orientation was even more pronounced in the alloy containing 2.0 pct Mn and 0.15 pct Ca. Fuller and Edmunds" studied the rolled sheet textures in a zinc alloy containing 1 pct Cu and 0.01 pct Mn. The basal pole figure of this alloy contained a region of strong concentration not found in the pole figure for pure zinc.' In the above cases, results were discussed in terms of possible changes in slip and/or twinning systems and critical shear stresses. However, there were no experimental data to relate these speculations with the observed textures. It has been suggested that the deformation textures of titanium, zirconium, and beryllium could be directly ascribed to their low c/a ratios. The complex deformation mechanisms of titanium" also have been suggested to be characteristic of hexagonal close-packed metals with low c/a ratios. The purpose of this investigation was to study the cold-rolled sheet textures in binary titanium-base a alloys in which the c/a ratios were greater than, less than, and about the same as the c/a for iodide titanium. The annealed textures of these alloys will be reported in a subsequent article. Materials and Methods The chemical analyses of the metals used in this investigation are given in Table I. The alloys were prepared as 15 g buttons on a water-cooled copper block in a vacuum arc-furnace. The furnace was evacuated to pressures of the order of 10.' mm Hg, flushed with 99.98 pct pure argon, and melting was carried out in 0.2 atm pressure argon. Each button was remelted until radiographs and chemical analyses indicated complete solution of the solute elements. All buttons were forged and homogenized 72 hr at 500°C. The buttons to be rolled were ground and etched to sizes and shapes suitable for the rolling

Jan 1, 1954