Search Documents

By J. C. Griffiths, D. W. Bennion

A mathematical model, which does not assume a priori that stratification exists, but was designed to test for the stratification was developed. The model segmented the reservoir horizontally into areas of common variance, then divided it vertically into strata (if strata were present). Next, trend surface techniques were used to determine the lateral extent and variation of each stratum. The model was tested on two reservoirs (sandstone and limestone) using porosity as the reservoir rock property. The sandstone reservoir contained approximately 60,000 samples from 2,000 cored wells, while the limestone reservoir had approximately 24,000 samples from 430 cored wells. Within the areas of common variance tested, the model was able to distinguish four separate lithological units or zones in the sandstone reservoir, four in the Marly section and seven in the Vuggy section of the limestone reservoir. INTRODUCTION To accurately predict the movement and production of fluids from a reservoir rock, it is necessary that the reservoir engineer have a knowledge of the rock properties and their distribution throughout the reservoir. For this study, a reservoir rock is defined as a solid containing interconnecting holes or voids which occur relatively frequently and are dispersed within the rock in a regular or random manner. The fraction of bulk volume that these interconnecting voids occupy is called porosity. The rock must also possess sufficient permeability to allow fluids to move through it and it may be composed of one or more strata or lithological units. A stratum or lithological unit as defined in this study is a body or volume of reservoir rock whose properties are so distributed that its lateral and vertical extent can be traced throughout or through a portion of the reservoir. Operationally, the existence of strata are question- able unless the variance within strata is less than among strata. Within the past few years computers have been widely used as a tool for calculating movement and production of fluids from reservoirs. Most prediction models require some type of a mathematical description of certain reservoir rock properties. In the past, most of these models have been relatively simple. Some have assumed vertical variation but no lateral variation, while others have assumed lateral variation but no vertical variation. In general, most of these models have not used functional relationships to predict reservoir rock properties as functions of position. The object of this study was to develop a model which would predict lateral and vertical variations of reservoir rock properties. Since samples are usually available from only a small portion of the total reservoir rock, it seemed logical that if measurements from these samples were to be used to infer the properties of the actual reservoir, the data should be treated statistically as a sample from the total population (reservoir). Therefore, the model developed in this study to predict reservoir rock properties was a stochastic model. The model was tested using porosity (a macroscopic reservoir rock property) measurements from two reservoirs — the first, a sandstone reservoir from which 60,000 samples were obtained; and the second, a limestone reservoir with 24,000 samples. Several investigators have proposed methods to determine the vertical and areal variation of a reservoir rock property. Stiles: Dykstra and Parsons 2 and Suder and Calhoun3 have all developed waterflooding prediction techniques which assume vertical variation but no lateral variation. Stiles developed a method to segment a reservoir arbitrarily using frequency distribution. Law4 suggested that porosity has a normal frequency distribution and that permeability has a log-normal frequency distribution. Elkins and Skov 5 proposed that when a reservoir is segmented vertically, the

Jan 1, 1967

By Robert Maes, Robert de Strycker

The solidifzcation reactions in the Cu-Sn-As system were determined by microscopic examination, thermal analysis, and X-ray diffraction. The system presents no ternary compound. There is a quasi-binary section, of a simple eutectic type, between the compounds Cu5As2 and SnAs; the system presents, in addition, one ternary eutectic reaction, one ternary peritectic reaction, and seven transitional reactions. The temp-peratures corresponding to each of these reactions were determined, most of them with an accuracy of *lLC. The isothermal lines of the liquidus surface were also established for almost the entire diagram. NO reference was found in the literature relative to the Cu-Sn-As phase diagram and this is believed to be the first study on this subject. The Cu-Sn diagram was taken from Hansen and Anderko;' it is represented by Fig. 1, with the phase designation which will be used in the remainder of this work. For the Cu-As diagram, the solution proposed by Hansen and ~nderko' must be modified to take account of recent data, principally from Heyding and Despault~ and from Hume-Rothery and Burns; this yields the diagram represented by Fig. 2. As for the Sn-As diagram, it is not established with certainty, particularly so far as concerns the solid solubility of tin in arsenic and the solidification reactions near the compound SnsAsz. For this last point, Hansen and Anderko propose two solutions: congruent melting of Sn3Asz with a eutectic SnsAsZ + SnAs or incongruent melting implying a peri-tectic reaction L + SnAs = Sn3As2. Microscopic examination of the few binary alloys prepared in this composition range made us prefer the second of these solutions, which gives the diagram shown by Fig. 3. EXPERIMENTAL METHODS The experimental techniques used in this work were microscopic observation and thermal analysis; a few X-ray diffraction patterns were taken to verify the

Jan 1, 1967

By Claude H. P. Lupis, John F. Elliott

QUITE generally an increase in the temperature tends to bring a system closer to ideality. It is reasonable as a first approximation to consider that the excess free energy will vary linearly with the temperature, or: where t may be defined as the "characteristic temperature" of the solution, or the temperature at which the system would become ideal if the results at T were linearly extrapolated. It must be emphasized that consideration of this hypothetical property is an appeal to intuition rather than an established physical concept. The order of magnitude of 7 for dilute solutions will depend on the nature of the solvent; a value of 3000° + 1000°K may be suggested for ordinary metallic solutions. In terms of free energy, entropy, and enthalpy first-order interaction coefficients, differentiation of the above equations with respect to composition leads to The concept of t offers a simple and plausible interpretation of Chipman and corrigan's3 recent empirical correlation between and a for a number of groups of ternary systems 1-24. The authors found that in each group the ratio was approximately constant. The corresponding calculated from their data by Eq. [4] are listed in Table I. The expected order of magnitude is obtained for 7. As noted by Darken,4 if a regular solution is not assumed, thus if utj) # 0, Chipman and Crrigan's results imply that, as T is increased, ultimately changes in sign, Fig. 1. Clearly Eq. [4] is valid only in a limited range of temperatures since at the temperature 7 at which the system becomes ideal not only should become null but also and . Thus, both and are functions of

Jan 1, 1967

By S. S. Marsden, S. A. Khan

Externally generated foam was injected continuously into short porous media. Both flow rate and pressure drop were measured. Liquid saturation was determined by electrical conductivity. Foam yuality I, expressed as the ratio of gas volume to total volume, was varied from 0.70 to 0.96. As measured with a modified fann VG Meter, apparent viscosity of this foam µa decreases with increasing shear rate but usually falls within the range of 50 to 500 cp. At a given shear rate, µa increases almost linearly with I-. When measured with a Bendix Ultraviscoson, kinematic pa is independent of V but absolute pa increases with T from about 3 to 8 cp. The effective permeability -apparent viscosity ratio ke/µa decreases almost linearly with V for porous media of high permeability, but the rate of decrease becomes less for tighter ones. The relative permeability-apparent viscosity ratio kr/µa us V data does not fall on a single line. The kr/µa ratio increases with liquid saturation in the porous medium and with surfactant concentration. Estimates of pa for foam in porous media vary from 30 to 100 cp. INTRODUCTION Although research on the development of a foam-drive, oil recovery process has been going on for almost a decade, most of the significant publications have appeared within the last several years. This illustrates well the rate at which interest in this process is accelerating. Bond and Holbrook 1 were the first to describe the use of foam to improve oil recovery in their patent of 1958. They proposed that an aqueous foaming agent slug be injected into the formation and that this be followed by gas to produce a foam in situ. Fried 2 studied the injection of foam into porous media which had already been subjected to conventional gas or water drives and found that gas could be used to drive a foam bank which would, in turn, displace additional oil in the form of an oil bank. He attributed the increased oil recovery to the high effective viscosity of foam flowing in porous media. His microscopic observations showed the importance of foam generation and regeneration within the porous medium. By injecting both air and aqueous surfactant solution, Bernard3 generated foams within the porous medium in which oil displacement was being studied. In a separate empirical test, he also measured the dynamic foaming characteristics of the same surfactants in water and/or oil. With some exceptions and for the seyen surfactants studied, there seems to be a qualitative relationship between the efficiency of liquid rlisplacrment and the dynamic foaming testtsed. This relationship was not consistent enough to eliminate the necessity of actual foam flood tests in porous media for surfactant selection. In a study basic to gas storage in aquifers, Bennett4 described the displacement of brine by foam in consolidated porous media. Among other things, he stated that the ability of a surfactant solution to foam is more important than the stability of its foam. The presence of a foam bank between the displacing air and the displaced brine improved both breakthrough and ultimate recovery. In a continuation of this work Kolb5 attributed the great reduction in surfactant solution production rate as displacement by air progressed to a decrease in relative permeability to gas. These several effects reported by both Bennett and Kolb can all be attributed to the high apparent viscosity of foam which was obviously flowing in the porous media. Bernard and Holm6 observed that gas permeability in the presence of foam was much less than in the absence of foam when both were measured at the same gas saturations. This reduction in permeability was proportionately greater for more permeable sands and sandstones than for the less permeable. Viscosities of the surfactant solution and air were used for their calculations rather than apparent foam viscosities. The foam stability increased with decreasing permeability, which was probably due to the natural regenerative properties of porous media, and the permeability reduction increased

Jan 1, 1967

By R. E. Reed-Hill

A simple accurate graphical method for orienting fcc crystals using (111) slip traces on a single surface is described. Solutions placing the pole of the surface in a unit stereographic triangle are obtained using a Wulffnet, a (111) standard projection showing the four (111) planes, and a (111) standard projection ruled into unit stereogmphic triangles. The multiplicity of solutions is described in relation to this method. AN accurate, rapid single-surface technique for orienting fcc crystals, using (111) slip or twin traces, has been a matter of continuing interest, as evidenced by the periodic appearance of papers on this subject.'-4 This note presents a new method possessing some advantages over those previously reported. It is an analog to the technique for hcp metals5 where the solution is based upon a stereographic rotation of the specimen surface on a (0002) standard projection that contains the planes of the possible features that might account for the observed traces. In the hexagonal solution, the axis of rotation is the basal-plane trace in the surface. In the fcc case, the axis of rotation is one of the observed ( 111) slip or twin traces. It has been observed that the solution is greatly facilitated if the indicated movements of the (111) traces during the rotation of the surface are plotted directly on a paper copy of a Wulff net as shown schematically in Fig. 1. The trace selected to represent the axis of rotation is assumed to fall at the north-south poles of the Wulff net. The directions of the other traces are then plotted as points at the proper angles around one-half of the basic circle of the net. The loci of points corresponding to the movements of these traces (directions) during a 90-deg rotation of the surface, about the polar axis of the net, are then easily and accurately drawn by following the appropriate small circles on the Wulff net. The solutions to the problem are obtained with the aid of a (111) standard projection of the (111) planes, see Fig. 2. This is normally a Xerox copy of a carefully drawn master made on high-quality tracing paper. The (111) standard projection is superimposed over the Wulff net and rotated about its center until the points of intersection of the loci of surface traces, drawn on the Wulff net, and the great circles representing the (111) traces, all lie along a common meridian of the Wulff net as demonstrated in Fig. 3. This latter great circle represents a possible position of the surface. The pole of this plane is then plotted on the (111) standard projection. Its location in a unit stereographic triangle may be readily obtained by superimposing (over the standard projection) another tracing paper (111) standard projection subdivided into unit stereographic triangles. The result of this last operation is shown in Fig. 4. MULTIPLICITY OF THE SOLUTIONS The maximum possible number of (111) traces on a single grain surface is four. Any one may be selected as the axis about which the surface is rotated. Corresponding to each such choice are three possible surface-pole orientations, related to each other by a simple 120-deg rotation about the pole of the (111) basic circle. This is a direct result of the threefold symmetry of the (111) standard projection. Since each of the four rotations will normally give a different set of three-pole positions, there are, accordingly, twelve possible solutions. Each of these twelve solutions places the surface pole in the same relative position in a stereographic triangle so that all are crystallo-graphically equivalent. Accordingly, only a single solution is required to obtain a basic orientation of the crystal. It should be noted, however, that the orientations obtained by reflecting* any of these twelve orien- *The reflected orientations may be obtained by a 90-deg rotation of the opposite sense to that shown in Fig. 1. This rotation places the correspond-ing trace loci in the lower hemisphere of the Wulff net rather than in the upper hemisphere. tations through the surface are also possible orientations that could account for the surface traces. Thus, a given set of four traces always corresponds to a pair of basic orientations related by reflection through the specimen surface. This ambiguity can only be re-

Jan 1, 1967

By J. E. Flinn, E. R. Gilbert

Recently, Green reported some high-temperature, 0.60 to 0.89 T3, creep results on unalloyed tantalum. From his study he determined an activation energy of 114 kcal per mole for steady-state creep using the T, differential-temperature method25 for testing. He accounted for the difference between this value and 110 kcal per mole for self-diffusion by the temperature dependence of the elastic modulus, 4 kcal per mole. This reported activation energy is in apparent disagreement with that which can be determined from Green's isothermal data, which we have plotted as log creep rate vs 1/T in Fig. 15. From this figure, it appears that there are two activation energies of -185 kcal per mole at the higher temperatures and -75 kcal per mole at the lower temperatures. It is not obvious that the data can be fitted to a straight line corresponding to an activation energy of 114 kcal per mole. One is led to two possible conclusions regarding Green's study: there are two discrete processes controlling the creep rate from 0.60 to 0.89 Tm or the apparent activation energy is not discrete but continuously increases with increasing temperature. The former speculation would be in closer agreement with the high-temperature creep behavior of tungsten.28 The apparent disagreement between the activation energies determined by the AT and isothermal methods certainly suggests that further study on the high-temperature creep of tantalum is needed.

Jan 1, 1967

By C. J. McHargue, J. C. Ogle

Lightly deformed columbiun single crystals which contained only parallel hoins or purullel and intersecting trains were annealed at 1000' and 1600"C. No re-crystallizntion occurred in specimens hawing only parallel twins. Only noncoherent twin boundaries nzipated at 1000°C but both coherent and noncoherent ones moved al 1600°C. Recrystallization occurred within a few minutes at twin intersections at 1000°C. The orientation 01 the recrystallized grains differed front that of both the matrix and deformation twins, but could he derired by (110) and/or(112) rotations. ALTHOUGH twinning in metals has been extensively studied, there have been no definitive studies of the annealing behavior of crystals containing deformation twins. Some effects observed after annealing deformation twins have been summarized by Cahn1 and Hall2. Any or all of these phenomena are observed: 1) The twins may contract so that the sharp edges of the lens become blunted, and eventually the twin may disappear entirely. 2) The twins may balloon out at an edge, giving rise to a large grain having the same orientation as the twin. 3) The specimen may recrystallize; i.e., new grains are nucleated and grow at the expense of the twins and the crystal immediately adjoining the twin. Such grains have orientations which are not present before. Contraction has been observed in iron,3 titanium,3, 4 beryllium,5 zinc,8, 7 Fe-A1 alloy,' and uranium.9 Long anneals at high temperatures are required to have any appreciable effect in these metals and only thin twins are absorbed. Lens-shaped twins are absorbed from the edges: the thin, almost parallel-sided twins are usually punctured in several places and each piece contracts independently. Absorption is very gradual and no sudden cooperative jumps have been observed. The expansion of a twin into a larger grain of identical orientation is unusual, but such growth has been observed in iron,"'" zinc,6 and uranium." Crystals which have been deformed simultaneously by slip and twinning recrystallize first in the area adjacent to the twin. New grains appear faster where the twins intersect: but isolated twins, especially if thick, can also give rise to new grains. This type of recrystallization occurs in zinc.6, 7, 12, 13 and beryllium.14 Reed-Hill noted, in a single crystal of magnesium, the nucleation of a recrystallized grain at a twin intersection which had the same orientation as the second-order twin and which grew into the highly strained matrix.15 Short-time annealing has been reported to cause no change in the deformation twins in vanadium,16 columbium, 17, 18 tantalum,19 tungsten,'' and zinc.7 The purpose of this investigation was to note the effects of annealing on the coherent and noncoherent boundaries of deformation twins in columbium and to locate the nucleating sites for recrystallization. The orientation relationships, which the new recrystallized grains have with the parent crystal and the deformation twins, were also determined. EXPERIMENTAL PROCEDURE Single crystals of columbium were obtained by cutting large grains from electron-beam-melted buttons which contained 10 to 50 ppm C, 10 to 100 ppm O,, 1 to 10 ppm H2, and 10 to 15 ppm N2. The crystals were hand-ground and chemically polished until all grain boundaries were removed. The specimens were mounted in an epoxy resin and a face of each crystal was mechanically polished on a Syntron polisher using Linde A and then Linde B polishing compounds. After all faces were mechanically polished, the crystal was electrolytically polished to remove all distortion due to cutting and grinding. Laue photographs were taken of all faces of the crystals to determine the quality and orientation of each crystal. The crystals were compressed about 10 pct at -196 C in a specially constructed compression cage with an Instron tensile machine. Each crystal was separated from the top and bottom anvils by teflon films which acted as a lubricant. With the specimen crystal in position, the entire cage was cooled to -196°C by being submerged in a Dewar containing liquid nitrogen. The crystals were compressed at a rate of 0.02 in. per min and the load was recorded on a strip-chart recorder. After deformation the crystals were mechanically polished on 600-grit paper and Pellon cloth with Linde A and Linde B polishing compounds. The crystal faces were chemically polished and then etched. The twin planes were identified metallographically from an analysis of the twin traces on two surfaces. Annealing was carried out by placing each crystal in a columbium bucket made from the same electron-beam-melted material as the crystal itself and suspending the bucket by a tantalum wire in a quartz tube. After a vacuum of 10-7 Torr was attained, a furnace at 1000" or 1600 C was raised into position and the crystals held for various lengths of time. The crystals were repolished and etched after annealing to remove any surface contamination. Approximately 0.010 in. was removed during this process. The resulting surface was examined metallographically for microstructural changes due to annealing. A microbeam Laue camera mounted on a Hilger Micro-focus X-ray unit was used to determine the Orientstions of the recrystallized grains. This X-ray micro-beam camera had a 0.002-in.-diam collimator and incorporated the ideas of both and and chisWik21 and Cahn.22

Jan 1, 1967

By T. F. Archbold, P. G. McCormick

Silica fiber 11arkevs and hardness irzdentatiozs have been used to study the thermal diffusion of racancies in zinc placed in temperature gradients of 460o and 706oK per Cm All markers mored toward the higher temperatures, and the data show that the heat of trarlsport Q is greater than the eizthalpy of z,acancj3 forrjalion E by an amount equal to 3.5 kcal per mole. These results ore not in agreement with those obtained in either of the two previous studies with zinc. The problem is discussed in terms of experimental differences and racancy source and sink efficiencies. THE general interest in thermal-diffusion phenomena over the past several years has prompted a number of marker movement studies using pure metals. In order to understand better the relationships between thermal and mass flow, a general goal of the experiments has been to obtain values for the net enthalpies carried by the diffusing atoms. For self-diffusion by means of the vacancy mechanism, the determination of relative marker positions before and after an anneal in a temperature gradient permits a calculation of a value for (Q - Ef). Q is the transferred enthalpy, and Ef is the enthalpy of vacancy formation in the metal being investigated. The initial experimental study of thermal diffusion in pure metals was accomplished by Shewmon who found essentially no movement of surface-hardness indentation markers in single and bicrystals of zinc held for 20 days in a temperature gradient of 310°C per

Jan 1, 1967

By V. B. Kurfman

The temperature, strain rate, and orientation deDendence of defbrnzation of single-crystal MgAg has been examined. The crystals exhibit a tendency to single glide and little or no hardening at 25°C for many orientations. A much higher hardening rate is observed when multiple glide occurs, such as can be initiated by surface defects. The tendency for easy glide becomes less dependent on surface preparation and orientation as T — 100°C and bars so tested often fail after one-dimensional necking-. At T > 200°C (transition temperature for single-crystal notch sensitivity and poly crystalline ductility) single glide diminishes and two-dirnensionul necking begins. The crystals do not strictly obey a critical resolved shear stress law, but show the influence of {loo) cracks in determining the slip mode. The results are correlated with the difficulty of sciperdzslocation intersection and semibrittle behavior of this compound in single-crystal and poly crystalline form. Comparisons are made with the slip selection mode observed in tungsten, with the reported observations of easy glide in bee metals. and with the mechanical behavior of poly crystalline MgAg. PREVIOUS work on tensile deformation of polycrys-talline MgAgl and bending deformation of single-crystal MgAg2 has shown that the compound is semi-brittle (i.e., notch and grain boundary brittle). If this semibrittleness is supposed to result from the difficulty of multiple glide (associated with the problems of superdislocation intersection) one might expect single crystals deformed in tension to show pronounced single glide and strong orientation dependence of hardening rate. These experiments were done to examine this supposition and to study the tensile deformation of a highly ordered system which may be considered bcc if the difference between the two kinds of atoms is ignored (actual structure: CsC1). EXPERIMENTAL Single-crystal ingots were grown by directional freezing as previously described.' These ingots were sliced into a by a by 2 in, rectangular bars by electric discharge machining, then round tensile bars were conventionally machined to 1/8-in.-diam by 1-in.-long reduced section. The bars were typically tested without an anneal because of the problem of magnesium vapor loss and they were typically tested as mechanically polished. The analyses are within the same limits as those reported earlier; i.e., the average composition for each specimen is within 0.5 at. pct of stoichiometry, while the total range from end to end in a given specimen varies from 0.7 to 1.4 at, pct. There has been no indication in the results of any variation in slip or fracture mode attributable to the composition fluctuations. The slip systems were determined by two-surface analysis of the bars after testing to failure at room temperature. Single glide was so dominant that there was little difficulty in identification of the dominant slip system even though the tensile elongation to failure often approached 7 to 8 pct in room-tempera- ture tests. Elevated-temperature testing was done in a silicone oil bath and low-temperature testing was done in liquid Np or a dry-ice bath. All stress measurements are reported as engineering stress unless otherwise specified, and crosshead travel is used as the strain measurement. RESULTS The tendency toward single glide is best seen in the pictures, Figs. 1, 2, and 3, which depict deformation at fracture as a function of test temperature. While it is possible to find regions of secondary slip by careful microscopy, such regions are very small. The development of a ribbon-shaped configuration from an initially round section bar pulled at 100°C is typical, occurred by single glide, and illustrates the degree to which such glide continues. At temperatures =100°C the bars typically show elongation of 20 to 50 pct by predominently single glide. Despite the large elongation, fracture even at 150°C occurs in a brittle mode, Fig. 2, in the sense that it is an abrupt failure which shows no discernible necking in the second dimension of the bar's cross section (i.e., there is no appreciable action of any slip modes which would decrease the broad dimension of the cross section). Near 200°C the fracture mode changes slightly. Although most of the sample extension is by single glide, after the bar develops the characteristic ribbon shape it begins to neck in the second (i.e., broad) cross-sectional dimension. The bar becomes very thin in the "necked down" region, Fig. 3, and the reduction in area approaches 100 pct. Often there oc-

Jan 1, 1967

By W. B. Braden

This paper presents a suitable method for predicting gas-free oil viscosities at temperatures up to 500F knowing only the API gravity of the oil at 60F and the viscosity of the oil measured at any relatively low temperature. The API pravity and the one viscosity value are used as parameters to determine the slope of a straight line on the ASTM slanaord viscosity-temperature chart. Then, knowing the slope of the line and one point on the line, the vrscosities at higher temperatures can be determined. The line slope correlations were developed at I00 and 210F since viscosity data are frequently measured at these temperatures. A procedure is given for predicting line slopes from measurements at other tetnperatures. A nomogram is furnished for solving the relationship. The correlation has been evaluated at temperatures up to 5OOF for oils varyzng in gravity from 10 to 33 " API. The correiution is applicable only to Newtonian fluids. Comparison at 500F of true viscosities and those predicted from values at 100F shows an average deviation of 3.0 per cent (maximum deviation of 6.0 per cent). Predictions from the values at 21 0F for the same oils how an average deviation of 1.5 per cent (maximum deviation of 3.4 per cent). INTRODUCTION Correlations have been developed by Beal' and by Chew and Connally' for predicting viscosities of gas-saturated oils at reservoir conditions. Each of these correlations requires a knowledge of the solution gas-oil ratio and the viscosity of the gas-free oil at the reservoir temperature. For temperatures below 350F, measurements of the gas-free oil viscosities can be made easily using commercially available equipment. In thermal recovery processes, however, reservoir temperatures well in excess of 350F are encountered. Viscosity measurements at such conditions are more difficult and time consuming and require modification of existing equipment or the construction of new equipment. Measurements are further complicated by the difficulty of handling highly viscous oils associated with thermal recovery processes. Therefore, it is desirable to have a correlation which allows accurate prediction of viscosities at high temperatures. A commonly used technique for predicting viscosities at high temperatures is to measure the viscosities at two lower temperatures, plot the values on ASTM standard viscosity-temperature charts and extrapolate to the temperatures desired. If either of the values is slightly in error, the extrapolated value can be significantly in error. To justify an extrapolation, three points are actually necessary. This procedure can consume much time, particularly with heavy oils. Considering the cost of viscosity measurements, it would be desirable to eliminate the need for direct measurements by having correlations which would allow viscosity predictions from other physical or chemical properties. Beal1 investigated the possibility of correlating viscosity with oil gravity at temperatures from 100 to 220F. While showing that a general relationship exists, he also found significant deviations. It is possible that correlations will be developed based on oil composition as more information becomes available. While not eliminating the need for viscosity rneasurements, the method presented herein requires that only one viscosity measurement be made. The API gravity must also be known. The theory is based on the fact that the viscosity of paraffins (high gravity) changes less with temperature than does the viscosity of naph-thenes or aromatics (low gravity). The gravity. therefore, is used as a parameter to determine the slope of a straight line on the ASTM standard viscosity-temperature charts. The correlation is applicable only to Newtonian oils, and deviations due to thermal decomposition and nonhomo-geneity cannot be predicted. Oils containing additives have not been evaluated. PROCEDURE Fifteen oils were used in developing the correlation; eight were crudes and seven were processed oils. Oil gravities varied from 9.9" API (naphthene base) to 32.7' API (paraffin base). The temperature range studied was 81 to 516F. Each oil used had a minimum of three viscosity measurements and each plotted essentially as a straight line on the ASTM charts. In all, 91 viscosity measurements were used in the correlation. Saybolt, rolling ball and capillary tube viscometers were used for the measurements. Viscosity data for Samples 1, 2, 4, 7, 10, 11 and 14 were obtained in Texaco, Inc. laboratories. The data for Samples 3, 5, 6, 8, 9, 12 and 15 were from Fortsch and Wilson,3 and data for Sample 13 were from Dean and Lane.' All data points used in the correlation are plotted in Fig. 1. It is seen that some of the viscosity values deviated slightly from the straight-line plots at the higher temperatures. Properties of the oils after exposure to the

Jan 1, 1967

By F. Galasso, A. Paton

ThE recent emphasis on producing boron fibers by reacting boron trichloride and hydrogen on a hot tungsten wire has brought about a renewed interest in the tungsten boride phase diagram. The reaction of the boron with the tungsten core results in the formation of 6 WB, W2B5, and another tungsten boride phase, which was suspected of being the WB4 phase reported by Chrbtien and Helgorsky.1 While the powder-diffraction data for the boride in the fiber and those reported

Jan 1, 1967

By F. W. Jessen, G. F. Sears

Cavities in massive salt for the purpose of storage of liquid hydrocarbons have assumed a prominent position in recent years. This paper describes a program to facilitate leaching operations for the formation of specifically shaped storage cavities. Various forms and sizes of cavities may be possible through use of the techniques developed. INTRODUCTION The creation of large under ground storage facilities for natural gas and liquified petroleum gases has been practiced for many years. Use of cavities obtained through solution of salt for this purpose is a fairly new and novel approach, but has gained increasing importance due mainly to the economics of this type of mining operation opposed to hard rock mining or surface and pit storage.1 The idea of controlling the shape of any cavity dissolved from massive salt has not been a prime consideration of companies engaged in the formation of storage space. This has been due mainly to an insufficient knowledge of the mechanics of the leaching process and a dearth of published information dealing with both the desirability and ease of control possible for this type of operation. Two distinct advantages are readily ascertainable. From a stability standpoint (i.e., the ability of the completed cavity to withstand stress imposed by overburden pressure and lateral tectonic stresses) a controlled cavity may be generated which will yield the most favorable attitude to these external forces and remain in operable use for longer periods of time.2 A second advantage is that for a given volume, a sphere (which is one controlIed shape possible) represents the minimum surface area exposed. This may become more important in the future when dealing with refrigeration and product losses in underground cavities. The basic solution mining process, without regard for controlling the final shape, is quite simple in that the equipment and materials required to dissolve a cavity in a salt dome or layered salt section are a source of fresh water, a circulating pump, several strings of tubing and a means of disposal of the return brine. To add control measures involves the use of an inert blanket material above the position at which solution proceeds. Initially the annulus of the largest wash pipe is filled with this blanket down to the top of the proposed cavity. The bottom of the proposed cavity is determined by the depth of the original drilled hole or a blanking plug. The blanket material is added incrementally in stages and displaces the water in the enlarging cavity downward. Where the blanket material has displaced the water, no further solution takes place. The rate at which to add this controlling blanket material so as to be able to form specifically shaped cavities of any particular size is of considerable importance. THE PROBLEM It is desirable to have as much information as possible as to the behavior of the mining operation since visual observation obviously is impossible. It would be advantageous, for example, to have a step-by-step program showing how much salt is to be removed, at what rate and the time required. This type of program would facilitate procurement of surface equipment and provide for proper management of manpower requirements, as well as yield a host of intangible benefits. As mentioned earlier, only rule-of-thumb estimates were available until recently and made this type of pre-planned operation haphazardous at best. Recent work3-' has done much to clear up the ambiguities and inaccuracies attendant to solution mining of salt and has helped to place the operation on a more scientific basis. This study attempts to correlate much of the information thus far developed into a complete mining program; to extend the idea of controlled solution mining to include not only spheres, but solid conic sections of the ellipsoidal variety; and to refine the computation of the rate of removal of salt during the various stages of mining. Dommers3 and IIemson4 made extensive laboratory

Jan 1, 1967

By A. O. P. Casbarian

An analytical method is developed to determine the variation of the ultimate lateral resistance of a plate in a cohesionless soil with depth. This analysis is based on a modification of Rankine's classical earth pressure theory and the theory of plasticity as applied to soils. The ultimate resistance is defined as the product of the effective stress at the midpoint of the plate, the area of the plate and a dimensionless variable termed the ultimate resistance factor. This variable has been plotted us the depth ratio: i.e., the ratio of of the depth of embedment us the height of the plate. The resistance of a plate may then be calculated using the values af the ultimate resistance factor from the chart provided or the equation may be programmed for use in an analysis of anchor systems in cohesionless soils. It is emphasized that the analysis is semitheoretical. The theory has been compared with experimental results reported in the literature and results indicate general agreement. Actual field tests are necessary to further verify this theory. INTRODUCTION With the exploration for oil offshore in waters all over the world, it is of importance to determine the behavior of the various soils in relation to their ultimate resistance to deformation. Examples of such problems are the holding capacity of anchors and the lateral resistance of piles. Very little information is available in the published literature on the design and performance of anchors, in either cohesive or cohesionless soils. The analysis developed in this report is the first step in obtaining a solution for the determination of anchor holding capacity in a cohesionless material. DISCUSSION The theory used in the analysis is based on the ultimate strength of the soil and is the maximum resistance developed by the plate against further movement. In such a state the elastic deformations are disregarded in comparison with the plastic deformations. Hence, the plate can be considered as completely rigid. The theory of plasticity determines the three unknown stresses at any point by means of two equilibrium conditions for a small earth element in combination with the failure condition. However, the exact solutions can only be carried out in a few simple cases such as, for example, when the rupture or failure lines are straight (Rankine theory) or with spiral and straight rupture lines (Prandtl theory).l Kiitter 2 derived a single equation expressing the variation of the stress in any given rupture line. To utilize this equation, it is necessary to know the stress in the rupture line at a certain point. Unfortunately, this is difficult to obtain unless the rupture line intersects the free surface at a certain angle or when the earth is cohesion-less and unloaded. A method to overcome this is to consider only the boundary conditions at both ends of a rupture line without investigating the equilibrium of the earth above the rupture line. This assumes that the rupture lines meet the surface at statically correct angles, so that boundary stresses may be determined. As Kiitter's equation furnished a relation between these stresses, the unit earth pressure may be calculated at the point where the rupture line meets the wall. Another way in which Kotter's equation may be applied is in investigating the equilibrium of a soil mass above a rupture line. This method assumes that the failure or rupture line is known and that the boundary stresses in the rupture line at the ground surface can be determined. In this case it is possible to determine the earth pressure from the equations of equilibrium.

Jan 1, 1967

By G. B. Clark

Dynamic rock mechanic as it pertains to blasting has involved several areas of active research effort. Among the important problems which have been the subject of research are wave mechanics in rocks, mechanisms of rock breakage, cratering, applications of similitude, energy partitioning of explosives, coupling, and other related factors. In some of these areas many problems have received fairly complete answers, while in others, investigations have raised new questions while attempting to answer existing ones. The slabbing mechanism explains many of the phenomena observed in blasting, but is hot the complete answer to observed breakage and fragmentation phenomena. There is currently a gap between coupling experimental results and theory. Applications of similitude to open pit blasting have been made unnecessarily complicated. Basic scaling laws for large true and apparent craters have not been derived for either chemical or nuclear explosives. There are few useful data and little developed theory on the partitioning of energy of explosives in all types of blasting. No suitable mathematical model has been developed which will describe the stress waves in rock which are responsible for rock breakage: Significant progress has been and is being made in these areas of research, however. ELASTIC AND NONELASTIC WAVES Complete solutions for spherical wave equations in viscoelastic media are not available in the literature. Equations for certain viscoelastic models do show limited promise, however, of being applicable to rock masses, particularly for pulse type phenomena. Seismic investigations, blasting operations, the design of protective underground structures, and the evaluation of other explosive effects in solids would be greatly bene-

Jan 1, 1967

By E. P. Dahlberg, R. E. Reed-Hill

On the assumption that stress or strain relaxation occurs as the result of a thermally activated process, equations are derived relating to tensile experiments that give the strain as a function of the time under the condition of constant stress, and the stress as a function of the time for constant strain. It is demonstrated that if the strain-rate equation i = previously proPosed by Kuhlmann., is used as a starting point, then the relaxation of strain at constant stress may be expressed by the equation c = (-RT/(Y) 1tz tanh (t + is the strain capable of being relaxed at any given instant. Similarly, it is shown that the relaxation of stress at constant strain may be given by a = (-RT/B) In tanh (t + t0)/27, where a is the instantaneois value of the relaxable stress. The fact that these relationships reduce to well-known empirical equations at both large and small values of the stress Or strain is also shozcn. The present theory is shown to agree well with experimental data obtained from tensile elastic aftereffect experiments on a zirconium specimen prestrained at 77 k as to make it strongly anelastic. It is also demonstrated that elastic aftereffect data obtained using torsional specimens ?,Lay agree reasonably well with the equation derived for the case of tension. RELAXATION experiments are often employed as a means of studying metallic deformation mechanisms.' The simplest and most commonly employed techniques involve stress relaxation at constant strain and strain relaxation at constant stress. In general, however, investigations of this nature have been seriously handicapped in the past by a lack of suitable equations giving the time dependence of the relaxing variable over an interval that extends from small strains up into the region where internal-friction experiments become strain-amplitude dependent. This paper presents a derivation of such a set of equations for the case where the time-dependent part of the strain is anelastic or recoverable and the specimens are loaded in simple tension. The relaxation of strain under the condition of constant stress will be considered first. Let us assume that strain relaxation occurs as the result of a reversible thermally activated process that occurs at a number of relaxation centers lying in an elastic matrix. Then, following Kuhlmann,2 we may express the rate of strain relaxation as follows: where C is the strain rate, AFx the free energy of activation of the process controlling strain relaxation, a, the effective or average resolved stress at the relaxation centers, u an activation volume, R the universal gas constant, T the absolute temperature, > a factor with dimensions of a volume that accounts for the strain contribution of a successful operation of a unit process, N the number of relaxation centers per unit volume, and v the Debye frequency. The first term on the right of Eq. [I] represents a strain rate in the direction favored by the stress, while the second term represents the rate in the opposite direction. It is implied in Eq. [I.] that both F and v are symmetrical with respect to the two basic directions of operation of a relaxation process. Eq. [I] may also be written where and S and Q are the activation entropy and activation energy, respectively, of the relaxation process. In the following, A will be considered a constant. This is compatible with a set of experimental conditions where the relaxation rate is controlled by a single basic reversible process in which it may be assumed that the temperature dependence of the product ?Nv is negligible in comparison with the temperature variation of the exponential term. It is also implied that v, 7, and N do not depend strongly on a, . In deriving a relationship for the strain as a function of the time from her equation, equivalent to Eq. [2], Kuhlmann2 chose to consider only the limiting cases where the time was either very small or very large. It will now be shown that it is possible to integrate Eq. [2] to obtain a single equation valid over a wide range of strains if the concept of relaxable strain is introduced. The use of this quantity, which is the difference between the instantaneous value of the strain and the value of the strain at complete relaxation, represents the primary point of departure of the present theory from that of earlier workers. Let us express the effective stress at the relaxation centers in terms of strain. For this purpose we may use the following equation derived by zener3 for the case of strain relaxation at slip bands: where M? and M are the relaxed and unrelaxed moduli respectively, go the applied constant tensile stress, and m an average orientation factor that takes account of the fact that a,, the effective stress at the relaxation center, may not be a tensile stress (i.e., a

Jan 1, 1967

By D. G. Westlake

NUMEROUS hydride habit planes in zirconium, Zircaloy-2, and Zircaloy-4 have been observed by various researchers and have been compiled by Louthan and Angerman.1 While only partial agreement exists on the identification of these habit planes, there is a growing consensus that a tensile stress applied on a polycrystalline rod during slow cooling will cause the hydride platelets to lie perpendicular to the tensile axis. walters2 agrees with Louthan and Angerman who reason that "The particular plane on which precipitation will occur is the plane that will result in the maximum relief of stress by precipitation of the low-density hydride." The experiment described below was initiated to test this hypothesis on a single crystal of zirconium. A single-crystalline tensile specimen of zone-refined zirconium, designated number 13 in a previous study,3 had a square cross section with one pair of parallel faces only 4 deg from the basal plane of the hcp structure. The tensile axis was about 86 deg from the (0001) pole, 14 deg from the (1010) pole, and 16 deg from the (1150) pole. The specimen was charged to 50 ppm H by allowing reaction with the correct amount of hydrogen at 1073°K, homogenizing at this temperature for 2 hr, and cooling in the furnace. Two- surface analysis revealed that hydride platelets had precipitated parallel to the three {1010} prism planes. The specimen was heated to 673°K, in air, for 16 hr to bring the hydrogen back into solid solution, and furnace-cooled to 573°K.* It was then stressed in ten- sion at 1.2 kg per sq mm. (This produced a resolved shear stress of 0.6 kg per sq mm for glide on the prism plane with the highest Schmid factor. A higher stress would have produced significant strain at this temperature because the critical resolved shear stress for prism-plane glide in this crystal was only 1.0 kg per sq mm at ambient temperature.) The poles of the three prism planes were at angles of 14, 46, and 74 deg to the tensile axis so that an applied stress would exert normal stresses on the three planes in the approximate ratios 10:5:1, respectively. As soon as the stress was applied, the sample was furnace-cooled to room temperature. It was ground on wet Sic paper and chemically polished in 50 parts HNO3, 50 parts water, and 7 parts HF before examination. The result of the experiment was negative; i.e., the three (1010) planes were of equal importance as hydride habit planes.* It must be admitted that the ex- periment is not entirely comparable to those performed by walters,2 Louthan,5 and Louthan and Marshall6 on polycrystalline materials. They have found that some critical stress (below the macroscopic yield stress) is required for stress-orienting the hydrides. The applied stress in the present experiment was, by necessity, smaller by an order of magnitude than the critical stresses they report for the various materials. If the critical stress was related to the microstructure, one might postulate that it should be decreased by increasing the degree of perfection of the metal lattice and by increasing the grain size. But the experiment above suggests that, for the extreme case of a zone-refined single crystal of zirconium, the critical stress lies above that which one can apply without introducing gross plastic strain as another variable.

Jan 1, 1967

By E. D. Levine

A study was made to delineate the dislocation mechanistns controlling prismatic and basal slip at low temperatu.ves in titanium containing approximately 100 ppm interstitial impurities. Mechanical tests were performed on single crystals prepared from electron-beam zone-refined crystal bar by a thermal cycling technique. Measurements were made of critical shear stress and activation volume as functions of temperature. The study revealed that, for prismatic slip, two overlapping thermally activated processes aye rate-controlling below 300°K. From absolute zero to 220°K, reasonable agreement is obtained between experiment and the mechanism which involves overcoming 01. the Peierl's stress by the formation of double kinks in dislocation lines lying in close-packed directions. Between 220°and 300°K prismatic slip is controlled by a difierent thermally activated process which is characterized by a stronger temperature dependence of critical shear stress. Above 300OK, prismatic slip is controlled by an athermal process. Basal slip is controlled by a thermally activated process between 195' and 400°K. An approximate analysis suggests that it is the Peierl's mechanism that is rate-controlling in this range. Below 195°K, the occurrence of mechanical twinning obscures the features of basal slip. Above '400 X, basal slip is controlled by an athermal process. TITANIUM, an hcp metal with a less-than-ideal axial ratio of 1.587, deforms primarily by {1010}( 1210) slip at low temperatures.'-3 The critical shear stress is strongly affected by the presence of interstitial impurities. For example, churchman3 reported that a decrease in the combined nitrogen plus oxygen content from 0.1 to 0.01 pct decreased the critical shear stress for prismatic slip from 9.19 to 1.4 kg per sq mm at room temperature. Basal slip, however, is less strongly affected by interstitials; Churchman's data for {0001}( 1210) slip show a smaller decrease in critical shear stress from 10.90 to 6.3 kg per sq mm at room temperature, for the same decrease in interstitial content. In the present study, an attempt has been made to delineate the dislocation mechanisms controlling prismatic and basal slip in titanium. In view of the strong effects that impurities may have on at least some of these mechanisms, experiments were performed on material of reasonably high purity, so that comparisons could be made with investigations on material of both lower and higher impurity contents. EXPERIMENTAL PROCEDURE Single crystals were prepared from 1/4-in.-diam electron-beam zone-refined crystal bar, furnished by Materials Research Corp. The analysis of the zone-refined material is given in Table I, indicating that the interstitial content is of the same order as the purer of the two materials investigated by Churchman. The Vickers diamond pyramid hardness number of this material was 85.6. This value agrees quite well with the correlation between hardness and interstitial content proposed by Ogden and Jaffee.4 On the basis of hardness, the material investigated in this study is, therefore, intermediate in purity between that studied by Churchman (Vpn = 97) and that studied by Spangler and Herman (Vpn = 62.4).5 Single crystals were grown by thermal cycling through the a-ß transformation. Two-in.-long by 1/2-in.-diam bars were heat-treated in vacuum for 4 hr at 1160°C. The temperature of the furnace was then lowered to 865°C, and the specimens were maintained at this temperature for 7 days. This procedure was repeated several times, resulting in the growth of individual crystals up to 14 in. in length. Specimens for testing in direct shear along either {l010}(1210) or {0001}(1210) were obtained by electrical-discharge machining. The specimens were in the form of cylinders 1/4 in. in diam by approximately 1/2 in. in length, with a reduced section 1/8 in. in diam and a in. in length. All specimens were electropol-ished subsequent to electrical-discharge machining. The ends of the test specimens were cemented in stainless-steel collars. Shear tests were conducted in an Instron Tensile Tester at various temperatures from 4.2° to 508°K. Measurements were made of critical shear stress and activation volume. The latter measurements were performed by measuring the change in flow stress upon cycling the strain rate between ?1 = 2.62 x 10-3 per sec and ?2 = 2.62 x 10-4 per sec. Activation volume, I)*, is defined as u* =ßkT [11

Jan 1, 1967

By D. J. Schmatz, H. I. Aaronson, H. A. Domian

Diffusion kinetics in disordered hcp Cd-Mg alloys have been investigated by means of the Kirkendall effect and concentration-penetration curves determined with an electron-microprobe analyzer. Self-diffusion coefficients of both species were determined at the three marker compositions obtained, averaging 27.6, 46. 7, and 78.1 at. pct Mg, by means of the Darken analysis. These coefficients were then corrected for the unequal and concentration-dependent partial molar volumes of the two elements with the Balluffi analysis, and for the vacancy flux effect by the Manning analysis. The latter correction reduced the Balluffi correction produced larger changes in the self-diffusiv-ities; neither, however, produced statistically significant changes in the Do's or the H'S. The most striking result of this investigation is that at all three compositions and at all temperatures studied both the uncorrected and corrected self-difjusivities of magnesium are higher than those of cadmium. The Cd-Mg system is the first one found in which the higher melting, lower vapor pressure element diffuses more rapidly. Both an empirical correlation due to Toth and Searcy and considerations of the atomic mechanism of diffusion indicate that this anomaly is probably due to a comparatively low value of the activation energy required for a magnesium atom and a vacancy to exchange sites, perhaps occasioned by the higher compressibility of magnesium atoms. KIRKENDALL effect studies have been previously reported for only two hcp solid solutions: the E phase of the Zn-Cu system1 and the a phase of the Cd-Hg system.' In neither investigation were the marker-movement studies supplemented with the concentration-penetration curve determinations necessary to evaluate self-diffusivities by means of the atano and the Darken analyses. The present program was undertaken to obtain both types of data on a hexagonal solid solution in order to provide more detailed information relevant to the mechanism of diffusion in this type of lattice. The Cd-Mg system was chosen for this study because the disordered solid solution extends across the entire phase diagram at temperatures above 253"c5 and the substantial difference in the melting points of the component pure metals promised that marker movements would occur at reasonably rapid rates should the diffusivities of the two species be as unequal as might be anticipated. The experimental convenience of the relatively low melting points of cad-miun and magnesium and the availability of extensive and accurate activity data6 (required for application of the Darken analysis) were additional reasons for selecting this alloy system. Since the anisotropy of diffusion is not large in either pure cadmium7 or pure magnesium,' the diffusion couples were prepared from polycrystalline components. The presence of a well-defined texture in the couples—the c axis of individual crystals tended to be normal to the diffusion direction— however, provides a fair degree of crystallographic definition to the data obtained. The principal (and entirely unexpected) finding of this investigation, that magnesium, the high-melting low-vapor pressure element, diffuses more rapidly than cadmium, in contradiction to a broad range of results in fcc and bcc alloys, as well as in the previously studied hcp alloys,172 makes the self-diffusivity determinations of immediate interest in understanding the origin of this anomalous result. EXPERIMENTAL PROCEDURE The cadmium (Belmont Smelting and Refining Co.) and magnesium (Dow Chemical Co.) used in this study were both of 99.99 pct purity. Alloys containing 51.0 and 65.6 at. pct Mg were prepared from these materials by melting under a MgC12-base flux in a high-purity graphite crucible. These alloys were subsequently hot-worked and then homogenized in a helium atmosphere at temperatures close to their solidus points. Sandwich-type diffusion couples of the type g/d/g were prepared from the pure metals by solid-state diffusion. Two-piece alloy couples of Mg/65.5 at. pct Mg (Mg/gCd) and Cd/51.0 at. pct Mg (Cd/CdMg) were welded by a liquation technique. The individual components of both types of couple were initially cylinders 1.27 cm in diam and in length; the ends of these cylinders were machined accurately flat and parallel. For both welding techniques, the pure cadmium cylinders and the alloys were chemically polished in a mixture of 40 pct ethyl alcohol, 40 pct hydrogen peroxide (30 pct conc), and 20 pct nitric acid,g while those of pure magnesium were polished in a solution of 10 pct nitric acid in ethyl alcohol.' Immediately afterwards, both metals were rinsed in freshly distilled acetone, and then in similarly purified methanol.' The Mg/Cd/g couples were assembled in a carefully cleaned stainless-steel welding fixture, in which a screw operating through a self-centering arrangement permitted a controlled pressure to be exerted upon a couple prior to welding.'' Tungsten marker wires 0.005 cm in diam were placed at the d:g interfaces of some of these couples, and imbedded in the couples during the application of pressure. As soon as a couple had been assembled, the welding fixture was inserted into a Pyrex capsule containing a packet of zirconium chips at each end. The capsule

Jan 1, 1967

By Robert K. Steunenberg, Irving Johnson, James B. Knighton

The activity coefficient of plutonium in liquid magnesium, over the temperature range 650° to 800°C, was obtained from measurements of the distribution of plutoninm between a 50 mole pct MgC12-30 mole pct NaCl-20 mole pct KC1 molten-salt mixture and liquid Zn-Mg alloys. For dilute solutions (0.08 at. pct Pu) the activity coefficient of plutonium was found to vary from 10.1 at 650°C to 12.2 at 800°C. The activity coefficients of plutonium in dilute liquid solutions of plutonium in uranium, silver, lanthanum. cerium, and calcium were estimated to be. The distribution data indicate a value of about 0.1 at 800°C for the activity coefficient of PuCl3 dissolved in the above ternary salt mixture. LIQUID magnesium and several liquid alloys of magnesium with metals such as zinc and cadmium have been shown to be useful solvents in pyrochemical processes for the recovery of uranium and plutonium from discharged nuclear fuels,' and for the separation of transuranium elements.' The present study was undertaken to determine the activity coefficient of plutonium in liquid Pu-Mg alloys in support of process-development work. The activity coefficient of plutonium in liquid magnesium was determined from experimental data on the distribution of plutonium between a liquid ternary MgC12-NaC1-KC1 salt mixture and various liquid Zn-Mg alloys. The distribution data were used to calculate the ratio of the activity coefficients of plutonium in liquid zinc and in liquid magnesium. The activity coefficient of plutonium in liquid magnesium was then computed from the known activity coefficient of plutonium in liquid zinc. It was not necessary to know the thermodynamic properties of the molten-salt system explicitly. The major features of the Pu-Mg system have been reported by Schonfeld.3 At the temperatures of interest in the present study, i.e., above about 600°C, the phase diagram indicates the existence of a wide liquid-miscibility gap, with the plutonium-rich liquid containing about 8 at. pct Mg and the magnesium-rich liquid containing about 10 at. pct Pu at the intersection with the solidus regions. Additional data on the compositions of the two equilibrium liquid phases obtained in this laboratory4 have defined the miscibility gap up to the consolute temperature (at about 1040°C). EXPERIMENTAL PROCEDURE AND RESULTS Materials. The 50 mole pct MgC12-30 mole pct NaC1-20 mole pct KC1 salt mixture was prepared by melting the required proportions of reagent-grade NaCl and KC1 with anhydrous MgC12. The molten salt was then purified by contacting it with liquid Cd-30 wt pct Mg alloy (at 450°C) to reduce oxidizing impurities, followed by filtration through a stainless-steel frit (pore size, 65 µ) to remove solid MgO formed during the reduction. The purity specifications of the zinc, magnesium, and plutonium were 99.999, 99.8, and 99.85 pct, respectively. Apparatus. The liquid salt and metal were contained in a tantalum crucible inside a graphite secondary vessel. The crucible assembly was located inside a resistance-heated stainless-steel furnace tube. The furnace tube was closed by means of a stainless-steel cover, which was attached by bolts, with a neoprene O-ring serving as a gas-tight seal. The top of the furnace tube was water-cooled to protect the O-ring. The furnace-tube cover was provided with a tantalum thermowell, a tantalum stirrer, and a port through which sampling tubes could be inserted and materials could be added to the melt without admitting air to the furnace tube. Vacuum and an argon atmosphere were available through a side-arm on the furnace tube. The furnace temperature was regulated by a proportional controller that was actuated by a chromel-alumel thermocouple between the furnace tube and the heating elements of the furnace. The melt temperature was measured by means of a chromel-alumel thermocouple in the tantalum thermowell. The accuracy of temperature measurement was ±3°C. The salt and metal phases were intermixed by a motor-driven tantalum paddle positioned at the liquid interface. The tantalum crucible was provided with four baffles to increase the turbulence. The sampling tubes consisted of 1/4-in.-OD tantalum tubing that terminated in a tantalum frit (Kawecki Chemical Co.; average pore size, 30 µ). Procedure. The zinc, magnesium, plutonium, and salt were charged to the tantalum crucible; then the system was evacuated and filled with argon. The melt was brought to the desired temperature, and agitated for 1 to 2 hr. After allowing the salt and metal to separate, both phases were sampled. Filtered samples were obtained by immersing the end of the sampling tube in the liquid and increasing the argon pressure sufficiently to force the liquid salt or metal through the frit into the tantalum tube. The sample was then partially withdrawn into the cooler portion of the furnace tube and permitted to solidify before being removed. The temperature sequence for sampling at each magnesium concentration was 800°, 700°, 600°, 650°, and 750°C. The composition of the liquid-metal phase was varied by incremental additions of magnesium in a series of experiments at low magnesium

Jan 1, 1967

By J. C. Trantham, J. W. Marx

From 1955 to 1958 the Phillips Petroleum Co. conducted a series of small scale counterflow combustion field tests in a tar sand about 60-ft deep and 6 to 12-ft thick near Bellamy, Mo. A total of seven different well patterns, including conventional five-spots and seven-spots plus a 15-well line drive pattern and a 10-well radial pattern, were employed. Electric heaters, gas burners and wellbore fuel packs were tested as counterflow ignition devices. Direct drive ignition followed by reversal to counterflow operation was not possible in this case because of formation plugging by the semi-solid tar during the direct drive phase. Sustained counterflow burning was achieved with wellbore fuel packs under conirolled conditions which permitted close correlation of fire front velocity, air velocity, air-oil ratios and yields for a wide variation in process parameters. The tar in place was about 10' APl with a reservoir viscosity greater than 500,000 cp. The composite oil prodlcct was 26' API with a viscosity in the range of 5 to 15 cp. Vertical sweep efficiency in the line drive pattern was essentially 100 per cent, with no evidence of gravity segregation. Original tar in place ranged between 800 and 1,000 bbllacre-ft, specific rock permeability averaged about 800 md before burning and effective air pemeability with residual oil and water in place (at the time of ignition) averaged 250 md. The volume of oil recovered during stabilized line drive counterflow burning represented about 67 per cent of the volume of tar originally in place in the burned out zones. A minimum was found when air-oil ratios were plotted against formation air velocities, and a limiting air velocity existed below which the counterflow fire front reversed and burned back over its own trajectory, using the residual coke left behind. INTRODUCTION Early in 1955 the Phillips Petroleum Co. began preparations to field test counterflow underground combustion as a method for producing oil from high viscosity hydrocarbon deposits, such as bituminous (tar) sands. At that time all published information on the use of underground burning io increase oil recovery was based on the direct drive process, in which the fire front moved through the formatim in the same direction as the injected air stream. Such direct burning is not applicable to many heavy oil sands or to most tar sands because the heat-thinned native hydrocarbon congeals in the cold rock ahead of the fire front forming a gas permeability block which smothers the combustion. Laboratory experiments first demonstrated that this permeability blocking would not occur if the fire front and the air stream moved in opposite directions. In this counterflow (reverse) burning process the heated heavy oil or tar is driven back through the combustion zone. Here thermal cracking produces relatively light hydrocarbon products which pass through the heated rock behind the fire front (largely in the vapor phase) and are incapable of causing a gas permeability block.' Exploratory wring revealed that consolidated bituminous sandstones with a combination of tar saturation, permeability, thickness and depth favorable for experimental field testing existed at several points in western Missouri. By the fall of 1955, a test site for determining technical feasibility of the counterflow process under field conditions had been selected near Bellamy, Vernon County, Mo. — about fifty miles north of Joplin. The pay zone was too thin and too shallow to constitute a commercial prospect, but these same factors permitted controlled experimental operation with a small compressor installation and with a large number of monitor wells. These were needed for adequate observation of the underground phenomena at reasonable cost. RESERVOIR CHARACTERISTICS The reservoir chosen for the experimental field tests was a 124 thick section of tar sand extending from 49 to 61 ft subsurface, with shale and siltstone laminations sealing the top and bottom. This section was part of a larger 30-ft thick tar sand deposit which extended both above and below the test zone. Core analyses showed that the 12-ft sand was effectively subdivided into two approximately equal sections, with the lower zone being consistently more permeable than the upper zone. The most significant laboratory core data are summarized in Table 1. The line drive test which is the principal subject of this report utilized only the lower 6-ft thick zone, with the upper zone regarded as part of the 55-ft overburden. WELL PATTERNS The over-all field test program involved a total of seven different well patterns, including conventional five-spots and seven-spots plus a 15-well line drive pattern and a 10-well radial pattern. In some of the tests the full 12-ft zone was employed; in other cases only the more perme-

Jan 1, 1967