Search Documents

By M. van Swaay, C. E. Birchenall

A. S. lBrling (Johnson, Matthey & Co. Ltd., Laboratories)— Because of its initial emphasis upon the production of membranes by powder metallurgy and by normal casting and rolling techniques, this paper is of considerable interest to those concerned with the design and application of industrial diffusion units. As the subsequent performances of the three grades of palladium were not compared in detail, it can perhaps be concluded that specific permeabilities were very similar. A short table in which the absolute diffusion rates obtained by the Authors with their three grades of palladium were compared with those of other investigators18-25 would, however, have been of particular value in view of the very variable results obtained with this membrane material. The Authors state early in the paper that the permeability of the thoriated samples decreased in much the same way as that of the nonthoriated samples. Later, they emphasize that sintered palladium without thoria was very susceptible to poisoning and did not respond as well as the other materials to the decontamination treatment. These results, which suggest that sintered membranes are less effective than those produced by normal methods, have considerable industrial implications, and amplified experimental details would be very welcome. The results of the Authors differ in one important respect from those of other investigators. Thus, no "bleeding" was required to maintain a constant rate of diffusion, although Davis, Darling, and De Rosset found this to be necessary. Davisz3 found that "bleeding" was necessary even when the hydrogen had already been purified by diffusion through a palladium tube, and it would be of value to know the purity of the gas stream emerging from the purification train shown on Fig. 1 of the Authors' paper. The Authors show that diffusion rates varied according to the square root of the pressure difference providing that the palladium membrane was not contaminated in any way. This finding is in agreement with the results of those workers2,18,19,23,24 who diffused their hydrogen through the membrane into a partly or completely evacuated system.Those workers who did not find this square-root law dependence did not have a partial vacuum on the downstream-side of the membrane.'' The difference between the two experimental con- ditions appears to be significant, as the absolute diffusion rates per sq cm of diffusion area are lower when the square-root law dependence is not observed. While the effect could be attributable to poisoning of the membrane it is conceivable that a positive pressure on the downstream side of the membrane might influence the mechanism of diffusion. In view of the capricious behavior of palladium, factors of this type, which are likely to influence its industrial significance, should be carefully considered. M. van Swaay and C. E. Birchenall (authors' 9-eply) — Our Eq. 191 may be compared with several others for hydrogen diffusion in palladium: D = 5.95 x exp (-5770/1)' D = 3.55 x 10"3 exp (-5550/)' D = 4.3 x 10"3 exp (-5620/1')' Other values are given in the literature, but in many cases surface poisoning may have affected the results. No analysis was performed on the hydrogen gas stream after purification. It should be noted that the presence of liquid nitrogen traps near the heated membranes may have set up convection currents and provided the equivalent of "bleeding." The question of whether a partial vacuum existed on the downstream side of the membrane is not clear. Does it imply that there is something unique about the permeability measured near one atmosphere pressure? We do not believe that there is. When square root pressure dependence is not observed, diffusion rates are not measured. Is "absolute diffusion rate per sq. cm. of diffusion area" equivalent to permeability? The terminology is new to us. In any case we doubt very much that pressures of the order of several atmospheres can affect the mechanism of diffusion.

Jan 1, 1962

By G. W. Doepel, W. P. Sibley

A three-dimensional, multilayer, mathematical model has been developed for predicting performance by mis-c.ible displacement. Areal and vertical coverage, total gas-oil ratio, propane-oil ratio and oil-recovery factor for the multilayer system are computed as functions of hydrocarbon pore volumes of solvent injected. The technique has been programed for use on the IBM-7090 digital computer and was used in calculating perfortnance for continuous miscible displacement of a five-spot, 50-layer .system. Results reported are based on the use of published laboratory data for a single, uniform five-spot layer. However, the technique is sufficiently generalized to calculate performance of any system of layers of any pattern configuration and mobility ratio, provided data are available to describe perfortilance of a single layer. The technique has also been adapted for predicting oil recovery resulting from various gas-driven slug sizes. The results presented for both continuous and slug-type injection show, that permeability variation has 0 significant effect on oil recovery but that solvent composition, within the range studied, has a lesser effect on oil recovery. INTRODUCTION Experience from many of the miscible-phase displacement projects currently in operation indicates that the injected solvent flows preferentially through the more permeable zones in the reservoir. This preferential solvent flow in the high-permeability streaks results in a reduction of vertical coverage (C,) and an attendant decrease in oil recovery. The multilayer technique presented in this paper uses areal coverage (C.,) data for a single layer from the literature and from laboratory measurements of displacement efficiency (herein referred to as microscopic coverage, C,,) to calculate areal coverage (C,) and vertical coverage (C,) for the multilayer system. Oil recovery for the multilayer system is obtained as the product of (Ca) (Cm) (Cv). It is assumed that C, in a single layer is 100 per cent; therefore, (C,) (Cm) is a measure of oil recovery from that layer. Data presented in the literature for a single layer generally deal only with areal coverage and/or microscopic coverage —but seldom specifically with oil recovery. Nevertheless, this technique is flexible in that either Ca and C,, of an individual layer or a direct measurement of oil recovery for the layer may be used. In developing this technique, the problem arose concerning a method of determining the relative amounts of solvent that will enter each layer during any increment of solvent injected. To solve the problem, a dimensionless equation was derived for expressing solvent injected into each layer as a fraction of total injection. A finite-difference method was applied in integrating the equation to determine oil recovery for the system of layers as a function of pore volumes injected. The technique subsequently was programed for solution on a digital computer. The method was applied to predict performance of a deep California reservoir under continuous miscible displacement and has been extended for calculating oil recovery as a function of slug size for gas-driven slug injection. The technique may be applied to any petroleum reservoir that can be adequately represented as a layered system, by considering permeability variation,1-3 permeability range and average permeability. Each layer is then assigned its appropriate permeability, porosity and hydrocarbon saturation. RESERVOIR VOLUME CONTACTED BY SOLVENT It has been recognized for some time that high microscopic coverage (C,) can be obtained by injecting a miscible agent, such as propane, into an oil reservoir. These high values of C,,, have been demonstrated in the laboratory and are normally between 80 and 100 per cent, depending on solvent and reservoir oil composition as well as on reservoir temperature and pressure. In miscible-phase displacement, because there is no interface between solvent and oil, effective permeability to solvent is the same as that to oil. (This assumes that oil saturation has not changed and that water is immobile.) Consequently, mobility ratio (M) simply reduces to that of viscosity ratio . Perrine4 and Habermann5 have shown that large

By H. Schwartzbart, J. R. Low

Annealed, .Poly crystalline, low carbon steel exhibits a phenomenon known as the "yield point." If such a steel is loaded in tension, the load increases steadily with elastic strain, drops suddenly, fluctuates about some constant load value for a time and then rises with further strain. The load at which the sudden drop occurs is called the "upper yield point"; the load at which deformation proceeds at constant load is known as the "lower yield point"; the amount of elongation at constant load is termed the "yield point elongation." Deformation which occurs during the yield point elongation is heterogeneous in nature, deformed metal existing next to undeformed metal. The drop in load from the upper yield point is concurrent with the formation of a band of deformed metal which usually appears at a stress concentration such as a fillet and at an angle of approximately 45" with the direction of pull. As deformation proceeds during the yield point elongation, the deformed region grows until the entire specimen has been strained an amount equal to the strain in the first band. A phenomenon very often associated with the yield point is that of strain aging. Aging in steels may be of two types: quench aging, or precipitation hardening, and strain-aging. Strain aging differs from precipitation hardening in that if the metal is merely strained instead of given a solution anneal and quenched, it then undergoes changes in properties similar to those observed in quench aging. The yield point effect and strain aging are most markedly exhibited by low carbon steel, but have been observed in other alloys. Numerous explanations for the mechanisms of both phenomena have been advanced, but the evidence for the acceptance of any of the hypotheses is incomplete. One of the explanations of the yield point phenomenon frequently proposed is that there exists a honeycomb of grain boundary material that supports the load above the yield strength of the ferrite and that when this network ruptures, the ferrite flows with no increase in load. This notion was first advanced by Dalbyl in 1913 and has been since put forward in more concrete form by many investigators. The testing of this hypothesis was, in part, the object of the present investigation. Similarly, the mechanism by which strain aging occurs remains to be explained. Presumably, it is a precipitation hardening phenomenon similar to quench aging, the precipitation being initiated by strain. However, it is not known whether the process is a solution and precipitation effect or a change in solubility due to strain. One obvious method of determining whether or not the yield point is a grain boundary phenomenon is to determine if single crystals exhibit a yield point. Low and Gensamer2 have shown that carbon and nitrogen are responsible for the yield point and aging in steel and that these effects can be eliminated by the removal of these elements through wet hydrogen purification. Single crystals that have been grown in wet hydrogen treated material must then be recarburized or re-nitrided in order to evaluate the effect of grain boundaries. In the present investigation, single crystals of iron which had been annealed in wet hydrogen to remove the carbon and nitrogen were cut into two specimens having the same orientation with respect to the tension axis. Of these, one specimen was pulled in tension in the wet hydrogen treated state and the other was carburized or nitrided and tested similarly. All the crystals were examined for a yield-point and strain aging. The grain boundary hypothesis for the explanation of the yield point in steel has been tested only indirectly previously. Arrowsmith,3 Andrew and Lee,4 Edwards and Pfeil,5 Fell,6 Ludwik and Scheu,7 Winlock and Leiter8 and others found a decrease in proportional limit and yield point effect with an in-

Jan 1, 1950

By J. B. Cohen

EARLY investigators1-4 reported that additions of beryllium to silver caused embrittlement. The in-termetallic phases which are now known to exist5,' were not known at the time. Solidification appeared to occur by a simple eutectic reation.4,7 (The primary beryllium-rich phase appeared globular, when the liquidus temperature was near the melting point of pure beryllium, and angular when the liquidus was much lower; in discussion of Sloman's work,4 it was suggested that this difference was due to an intermetallic compound which was not revealed by the polishing technique.) In the investigation reported here, it was found that if the compounds now known to exist were avoided, certain beryllium-silver alloys containing large amounts of the primary beryllium phase possessed sufficient ductility to be cold formed. Alloys containing 14 and 20 at. pct Ag were prepared from vacuum-cast beryllium containing 1.0 to 1.2 pct impurities and silver, 99.99+ pct pure. The elements were melted together by induction in beryllia or alumina crucibles under argon at less than 1 atm pressure. The cooling rate was adjusted by regulating the temperature of the melt and by reducing the power to zero without stopping the water flow in the induction coils, which were in contact with the crucible. Ingots were about 0.75 in. in diam with a grain size of 0.12 to 0.25 in. Microscopic investigation of transverse or longitudinal sections of an ingot cast from a temperature well above its melting point showed that, in general, the compounds did not form, presumably because of rapid cooling; the alloys consisted of beryllium-rich particles in a silver-rich matrix as shown in Fig. 1. A eutectic structure was not observed, probably .due to the low beryllium concentration of the eutectic composition and eutectic divorcement. Within a given grain, the small globules of the primary phase were all oriented in the same direction, as indicated by examination under polarized light; the orientation varied from grain to grain. This suggested that the globules were branches of the same dendrites. If cooling was too slow, a third constituent was observed in the microstructure; because it formed around the primary beryllium-rich globules it was Fig. 1—Ag-Be alloy, 18.6 at. pct Ag, as-cast. Bright field illumination. Etched with an aqueous solution of NH,OH and H2O2. X500. Reduced approximately 21 pct for reproduction. probably the 6 compound. A thin ring around an ingot often contained this constituent. When this compound was present in large quantities, or the silver segregated so that large regions occurred with the beryllium-rich globules in contact, sections from the ingot were brittle. When the compound was initially absent, it was possible to carry out reductions in thickness of 50 pct by cold rolling with only limited edge cracking; failure did not occur until reductions of 80 pct. A typical cross section after cold rolling is shown in Fig. 2. After simple bend tests, microcracks were found; these apparently were difficult to propagate in the two-phase structure, as the specimens continued to behave in a ductile manner after the cracks appeared. Similar cracks would probably be closed during rolling. It is apparent in Fig. 2 however, that the primary beryllium particles undergo extensive plastic deformation. Hardness values of the alloys indicated that some work-hardening occurred during cold working. Fig. 2—Ag-Be alloy, 18.6 at. pct Ag, reduced 83 pct in thickness by cold rolling. Bright field illumination. Etched with an aqueous solution of NH,OH and H2O2. X500. Reduced approximately 21 pct for reproduction.

Jan 1, 1961

By Heribert Wiedemeier, Ali Khan

The tetnperature-conposition phase diagram of the syste,n MnS-CdS has been investigated over a range fro)n 600° to 1000° C. The small degree of solid solution on the MnS rich slde cornpared to the large CdS rich single-plzase region could be esplained by the difference in strzrctures of the cotnponents and the relative sizes of the metal ions. The change in lattice parameter as a function of cot)zposition is linear atzd jollous Vegavd's rule within experittzental limits of error. Mns and CdS have gained considerable attention in recent years because of their interesting solid state properties. In order to investigate such properties and thermodynamic quantities of mixed compositions of transition metal chalcogenides, it is necessary to establish their phase diagrams. Two earlier publications are concerned with the MnS-CdS system. Schnaase' found complete miscibility when MnS and CdS were co-precipitated in aqueous solutions at room temperature. However, there is evidence that the mixed crystals were not stable over the entire range of composition. roger's' investigation of the ternary system ZnS-CdS-MnS at 900" C indicates a solution of 48 mole pct MnS in CdS. During the course of this work which covers a temperature range from 600" to 100O° C, emphasis has been placed on the use of high-purity materials and well-defined modifications of starting products. EXPERIMENTAL Starting Materials. The stable green modification of MnS with NaCl structure was prepared by direct fusion of the elements and by chemical transport reaction. Stoichiometric amounts of manganese of 99.99+ pct purity and of sulfur of 99.999 pct purity from Gallard-Schlesinger Corp. were sealed in an evacuated quartz tube at a pressure of 10"5 mm Hg and annealed for 48 to 60 hr at 800°C. The quartz tubes used throughout this work were outgassed under high vacuum at a temperature of 90OU to 1000° C for 12 to 18 hr. For the synthesis of MnS by chemical transport3 stoichiometric ratios of elemental manganese and sulfur and small amounts of iodine were sealed under vacuum in a quartz tube, which was placed in the temperature gradient, 900/800-C, of a two zone furnace. Single crystals of MnS in form of platelets and octahedra were deposited at the cooler end of the tube. The manganese content was determined by EDTA titratin. A typical result of several analyses was 63.05 pct Mn which agreed well with the computed value, 63.14 pct. Commercially available CdS of 99.9+ pct purity, a mixture of the cubic and hexagonal forms, was converted into the hexagonal high-temperature modification by annealing CdS under vacuum in quartz tubes at 700° C for about 36 hr. As a second method, cadmium sulfide was transported in a temperature gradient, 70O° /650" C, with in vacuo sublimed iodine from the hot to the cooler zone of the tube. Large sized crystals in form of hexagonal columns and some platelets were obtained. An average result of the cadmium analysis by EDTA titration4 was 77.93 pct in good agreement with the computed value, 77.81 pct. Only the NaCl modification of MnS and the wurtzite modification of CdS were used for the solid solutions. Preparation of Solid Solutions. Finely powdered MnS and CdS samples were thoroughly mixed in molar ratios and compressed into pellets of about 6 mm diam and 2 mm in height at a pressure of approximately 13,500 atm.

Jan 1, 1969

By V. Patarini, M. Salkind

BECAUSE of its high strength and stiffness coupled with low density, continuous boron filament is showing great promise as a potential reinforcement for both metals and nonmetals. Of primary consideration for structural applications is the fatigue behavior of boron fiber composites. Composite fatigue behavior is dependent upon three factors: 1) fiber fatigue behavior, 2) matrix fatigue behavior, and 3) the effect of the interface and fiber ends. The purpose of this note is to describe the behavior of boron filament in rotating beam fatigue. The filament, described by Talley' and Galasso, Salkind, Kuehl, and Patarini,' was made by the hydrogen reduction of boron trichloride on a hot tungsten filament. Testing was performed using a commercial rotating beam wire fatigue tester* described by Votta.~ One end of the test sample is held in a chuck in a synchronous motor. The sample is bent around in a "U" shape, and the other end inserted into a carbide bushing. The maximum strain occurs at the bottom of the "U" where the radius of curvature is a minimum. The strain is adjusted by adjusting the length of the specimen and the location of the bushing. Samples were rotated at 3600 rpm and an automatic shutoff device used to determine the number of cycles to failure. The measurements were accurate to the nearest 360 cycles. The fatigue behavior is summarized in Fig. 1 for fifty-one samples of 4.8 x 10"3 in. diam fibers and ten samples of 3.2 X 10"3 in. diam fiber. The two sample groups represent two different filament production runs and two different nominal size groups. Both groups exhibited nominal tensile strengths of 400,000 psi. No significant difference in the fatigue behavior of the two sample groups is apparent. The scatter in the data is not unexpected for a brittle filament which exhibits a tensile strength distribution with a standard deviation of 10 to 15 pct. Because the test procedure provides a known outer fiber strain in the sample, the maximum stress was calculated assuming a Young's modulus of 58 x 10' psi. This value was selected as being representative of several nominal fiber sizes.2 All of the samples which failed exhibited sudden, catastrophic fracture, and no permanent deformation. One sample was stopped after 1.66 x lo8 cycles at *200,000 psi. The fiber exhibited no apparent damage or deformation. It is apparent that the fatigue resistance of boron filament is quite good. It can therefore be speculated that the fatigue behavior of boron fiber composites would be controlled by the "weaker links", namely matrix fatigue resistance and interface behavior. In fact, preliminary investigations have revealed that this is true for boron-reinforced aluminum? It should be noted that the data presented herein may not necessarily reflect the tensile fatigue behavior of boron filament. It is well-known that boron filament has a radial residual stress pattern with the outer surface in ~om~ression.~ Because of this fact, tensile failure is normally nucleated in the interior of the filament.5 Since the data presented herein is for a rotating beam test in which only the outer surface of the fiber is stressed to the maximum value, it would be incorrect to assume that the failure mechanism and fatigue behavior would be the same if the fiber were tested in tensile fatigue. The fact that many of the data are well above the nominal 400,000 psi tensile strength would tend to support this argument. The authors would like to express their appreciation to the United Aircraft Research Laboratories for supporting this investigation and for permission to publish the results. A portion of this effort was supported by the Air Force Materials Laboratory under contract AF33(615)-2125.

Jan 1, 1968

By E. S. Machlin, C. W. Chen

RECENTLY Chen and Machlin' proposed a mechanism for intercrystalline cracking in metals during high-temperature stressing. According to this mechanism the formation of voids at grain boundaries is primarily responsible for intercrystalline cracking. A model has been provided to demonstrate the initiation of submicroscopic cracks and voids at jogs of grain boundaries where grain boundary sliding is inhibited. Submicroscopic cracks can grow as a result of either further grain boundary sliding or localized tensile stress. In the former case, it is believed that grain boundary sliding produces more voids which link together to produce macroscopic cracks and which may lead to the eventual failure of the metal. Concurrently, Nield and Quarrell2 reported their interesting results obtained in a systematic investigation of intercrystalline cracking in two aluminum alloys under various creep conditions. Their observations are generally in agreement with the above mechanism, especially the conclusion that voids are formed as a direct consequence of grain boundary sliding. On the basis of the volume fraction occupied by voids, they could find no confirmation of vacancy condensation as the mechanism for void growth. Furthermore, an important feature of intercrystalline cracking has been likewise pointed out by them that the majority of cracks are not formed at grain corners. Nield and Quarrel1 meanwhile emphasized the importance of grain boundary migration by designating it as one of two main processes operative in the mechanism of crack formation during creep. The designation seems unnecessary because overemphasis of grain boundary migration only complicates the mechanism of void formation. There are two reasons which militate against the inclusion of grain boundary migration in the mechanism of void formation. In the first place, the formation of intercrystalline voids per se does not require grain boundary migration. Secondly, in addition to grain boundary migration, void formation may be minimized by any other process such as plastic deformation which is capable of relieving localized stresses around jogs. Chang and Grant3 have shown that intercrystalline cracking is prevented in aluminum by lattice bending in high stress regions. The influential role played by grain boundary migration in creep has also been realized in our studies. In several cases, voids did not appear at all in grain boundaries whenever grain boundary migration took place prominently during creep. The purpose of this paper is to discuss the effect of grain boundary migration on the formation of intercrystalline voids and other related rupture properties. Experimental evidence indicates that in an unstable grain structure void formation may be markedly modified by structural changes. The abnormally high ductility and the transcrystalline mode of fracture observed in copper at elevated temperatures reveals that effective grain boundary migration minimizes void formation. Grain boundary migration has long been known as a common high-temperature phenomenon in metals with the grain configuration in a nonequilibrium state. Its effect on ductility and mode of fracture has not been recognized until recently.4 Since the introduction of the term "Equicohesive Temperature," the conventional belief has been that below the equicohesive temperature the ductility of metals increases moderately with decreasing temperature until the ductile-to-brittle transition temperature of fracture is reached. In this temperature range, metals fracture in a ductile mode with a "cup-and-cone" fracture surface. Above the equicohesive temperature, ductility remains invariably low and decreases slightly with increasing temperature. The mode of fracture is strictly brittle and intercrystalline. Typical examples have been shown for a brass and copper in Fig. 1 and 9, respectively, in the paper by Greenwood, Miller, and Suiter.5 The observation in the upper temperature range that ductility of metals is low and insensitive to temperature may be explained in terms of void formation at grain boundaries. It is also in accordance with the findings of McLean6 and Fazen, Sherby, and Dorn7

Jan 1, 1961

By Y. Kondo, A. M. Gaudin

A combined method for sizing in the sub-micron range was developed using centrifugation and gamma emission. Its essential features are: 1) Centrifugation of suspended mineral in water carried out at 4°C where water has its maximal density and zero temperature gradient of density 2) Gamma activition of the sample suspension induced by irradiating it in a nuclear reactor prior to centrifugation 3) Measurement of the activity of a beam of suspension as it is sedimenting in the centrifuge, using a scintillation counter provided with a discriminator. Albite, the sodium feldspar, was used as feed to the comminution step, and the induced gamma activity of Na24 was measured. Dissolved sodium ion caused a a non-settling activity in the measurement. Other "noise" is due to scattered gamma radiation from albite particles in the centrifuge outside of the line of sight. These two activities were obtained sep arately and subtracted from the total activity counted. The relative solid concentration in the suspension under the collimating slit is proportional to this net activity. The cumulative weight per cent of undersize particles was calculated and plotted against size on logarithmic scales. The relationship was found to be linear down to about 30 mµ. The specific surfaces calculated from particle size distribution agree with those measured by krypton gas adsorption. The measurement of particle size distribution of comminution products has been of interest in clarifying the mechanism of the process. The Gaudin-Schuhmann equation1,2 relates the cumulative percentage undersize, y, to relative particle size x/k. The particle size, x, and the size modulus, k, together with the distribution modulus, m, characterize the particle size distribution, m is the slope of a plot of log (y) against log (x/k). This equation has been verified for ground particles down to 1 micron in size.3'4 For particle sizes smaller than a few microns, the precision and facility of sedimentation and microscope techniques become questionable. Use of a centrifuge which shortens settling time from weeks to hours is attractive. Many centrifugal methods have been proposed; however, determination of the solid concentration of the suspension at a reference level as a function of time is difficult, and sampling the suspension is likely to introduce errors. Measurement of particle size distribution by X-ray absorption in gravitational5 and in centrifugal6 sedimentation have also been tried. In these methods, the absorption coefficient for the solid is independent of size, since the wave lengths of the X-rays and gamma-rays are orders of magnitude smaller than the particles. Radiometric techniques7 in size distribution measurement have been developed within the last several years. They offer advantages over usual gravimetric and optical methods. PRINCIPLE OF METHOD The combined method of centrifugation and gamma emission for sizing in submicron range described in the present paper has the following essential features. 1) Centrifugation is carried out at 4 ± 0.1°C. Convection currents in the water phase due to density changes caused by temperature variation may disturb completely the settling system. To avoid this difficulty, the unique property of water having zero temperature gradient of density at 4°C is used. 2) Prior to size measurement the sample suspension is irradiated in a nuclear reactor and gamma activity induced. The gamma activity of the particles sedi-mented radially is measured by a scintillation counter provided with a discriminator. By this method, interruption of centrifugation by withdrawing samples is avoided and the counting rate at any time represents the solid concentration in the suspension at the level of the collimating slit of the centrifuge at the particular time at which the measurement is made. 3) By proper selection of the feed material and of the comminution device, gamma activities of foreign

Jan 1, 1964

By A. W. Schlechten, C. M. Hsiao

The apparent vapor pressures of a number of metal sulphides were determined by measuring their rate of weight loss when they were heated under vacuum. The calculated pressures are due in some instances to the volatility of the sulphide and in others to the decomposition pressure. FIE direct reduction of metallic sulphides, as in the precipitation method for lead or antimony, is not a widely used process today. However, it might be employed for the production of a number of metals if the operation were carried out under reduced pressures. Two general procedures are possible; the reducing agent could form a stable nonvolatile sulphide from which the valuable metal could be removed by volatilization or liquation. The other approach would be to choose a reducing agent which would form a sulphide that could be volatilized away from the valuable metal. A vacuum operation would serve to protect the sulphides and metals from oxidation and would increase the evaporation rates. A related process is the separation of metallic sulphides by selective volatilization either at atmospheric pressure or under vacuum. This procedure is already used commercially for the elimination of lead and cadmium as PbS and CdS from zinc calcines; it might well have other applications. To predict such processes it is necessary to know the relative volatilities of the metallic sulphides and to know whether they volatilize as sulphides or whether the predominant weight loss is due to decomposition. Such data would also be of value in studying problems such as the use of sulphides for refractories, as described by Brewer,] or in the evaluation of the theory of ore deposition from the vapor state, as proposed by Brown.' Qualitative statements concerning the volatility of metallic sulphides can be found frequently in the literature and are well summarized by Mellor, but accurate determinations of rates of evaporation or vapor pressures are very scarce. Kelly" lists vapor pressure data only for PbS; his reference being to the work of Schenck and Albers,4 who measured the vapor pressure of PbS by a dynamic method between 850" and 995°C. More recently Veselovskii5 has determined the vapor pressures of the sulphides of antimony, lead, cadmium, and zinc at temperatures ranging from 360" to 1000°C using Knudsen's method. A second Russian worker, Pogorelyi,6 determined the vapor pressures of CdS and ZnS at temperatures from 900" to 1250°C. The results obtained by these workers are shown in subsequent graphs in comparison with the data of this paper. Brewer and coworkers1 have studied some of the metallic sulphides to determine their suitability as refractories in vacuum. Veselovskii's determinations on PbS checked well with those of Schenck and Albers, and his determinations of CdS check almost exactly with Pogo-relyi. For the vapor pressure of ZnS the data of Veselovskii and that of Pogorelyi are in fair agreement, as shown in Fig. 1, but it should be pointed out that Lumsden7 has discovered that the equation given by Veselovskii for the vapor pressure of ZnS: - 14,200 log P9.495 is in error by a factor of ten when compared to the experimental data. In contrast with the scarcity of data on vapor pressures, the decomposition pressures of many metal sulphides have been determined and are summarized by Smithells.8 Theoretical Considerations The experiments described in this paper consisted of determinations of the rate of weight loss of certain metallic sulphides under very low pressures at different temperatures. These data plus observation and analyses of the products reveal the mechanism of the weight loss and also provide a basis for the calculation of the vapor pressure or apparent vapor pressures of the sulphides.

Jan 1, 1953

By S. R. B. Cooke, H. S. Choi, D. H. Harraway, I. Iwasaki

The mineralogy of natural iron ore samples and their slimes have been investigated and the nature of slime interference in iron ore flotation has been studied. The results have been correlated with the flotation behavior of a natural ore sample. Various conclusions have been reached as a result of these studies. In the flotation of iron oxide ores, desliming is generally considered to be an essential step to successful separation, irrespective of whether the iron oxide or siliceous gangue minerals are to be floated, although some workers have found empirically that certain collectors and reagent combinations are less sensitive to the presence of slime than others. The detrimental effect of slime is considered to be twofold. First, the presence of slime presupposes a high reagent consumption because of high specific surface. When clays are present and when cationic collectors are used abstraction by ionic exchange may augment this condition. Attempts to compensate by the addition of large amounts of reagent result in only partial correction. Second, slime coatings on granular particles tend to interfere with the bubble-mineral contact. The nature of slime coatings in flotation is quite complex and has been the subject of considerable study; an attempt was recently made by Fuerstenau et al.' to relate the slime-coating density with flotation recovery and with the properties of the electrical double layer at the mineral surfaces. Although major constituent minerals in the oxidized iron ores of the Mesabi and Cuyuna ranges are hematite and goethite (limonite) as the iron oxides, and quartz and chert as the siliceous gangue, very little is known of the mineralogy and of the chemical and physical characteristics of the slimes, and still less of the interaction between slimes, collector, and the mineral surfaces. For pragmatic reasons iron ore slimes are defined as — 20p material, based upon the calculated settling velocity of a quartz sphere of the stated diameter, and of all other equal-settling mineral particles. Active slime interference with flotation is usually attributed to particles 5p in size or finer, the extent of such interference appearing to depend not only on the mineralogy but also on the size distribution of each mineral in the slime fraction. The objective of this work was to acquire a more specific knowledge of the constitution of iron ore slimes and of their role in flotation. The mineralogi-cal composition of natural slimes was obtained by differential thernal analysis (hereinafter referred to as DTA), electron microscopy, and chemical and X-ray analyses. Some flotation tests were made on natural ore samples. The surface properties of several constituent minerals were investigated by electrophoretic measurements and by determination of the adsorption of some typical flotation collectors. Finally, simplified flotation tests were made on granular goethite and quartz in the presence of various artifically prepared slimes; the results were compared with laboratory flotation tests made on natural ores. MATERIAL AND METHODS Mineralogy of Some Iron Ore Slimes: Samples of two wash-ore tailing::, hereinafter identified as A* and B, *Sample A of the present investigation 18 the same am ore A in ref. 3. of a wash ore C, and of a paint rock D, all from the Mesabi range, were selected for investigation on the basis of prior information concerning their flotation behavior. The two wash-ore tailings, when deslimed, were known to respond well to flotation, whereas the wash ore was high in goethite, tended to be slimy, and responded poorly to flotation. The paint rock also exhibited sliming tendencies, and was suspected to contain relatively large quantities of clay minerals because of its high A1203 content. Gruner states that the paint rock of the Mesabi frequently contains kaolinite. To ascertain the relative sliming tendencies of each sample a standardized desliming method was employed. All samples were ground through 65 mesh using jaw and rolls crushers. A 500-g sample of each

Jan 1, 1962

By T. M. Geffen, J. L. Sanderlin, F. F. Craig, D. W. Moore

Scaled reservoir models have been used to study the effect of gravity on oil recovery performance in frontal-drive operations; namely, water, gas, or solvent flooding. The difference in density between the reservoir oil and the injected fluid causes their segregation, resulting in a non-uniform advance of the fluid front. In the laboratory flow tests, which simulated both five-spot and linear injection operations in flat reservoirs, the viscous, capillary, and gravity forces present in these operations were scaled. Dyed fluids were used so that the gross movement of the injected fluid could be observed. The studies covered a range of injection rates, formation thicknesses, and rock and fluid properties normally encountered in field operations. The results of the model tests indicate that the volume of the reservoir contacted by the injected fluid at its breakthrough into the producing well is less than that expected based on information which neglects gravity effects. This difference can often be as much as 80 per cent by gas or water injection in uniform sand bodies. Preliminary flow tests on a non-uniform sand body indicate that the uniformity of the flood fronts may in some situations be influenced to a much greater degree by permeability variations within the rock body than by gravity effects. The magnitude of fluid segregation due to gravity is controlled by the average injection rate, rather than day to day or week to week variations. INTRODUCTION One of the important factors controlling the oil recovery from a frontal-drive operation, such as water, gas, or solvent injection, is the volumetric sweep efficiency. This factor is a measure of the gross portion of the reservoir that is contacted by the displacing fluid. The volumetric sweep efficiency is influenced by gravity effects, well arrangements, and variations in rock permeability within the reservoir. The gravity effects are due to the displacing fluid being of different density than the reservoir oil. This causes the displacing fluid to move preferentially toward either the top or bottom of the formation. Non-uniform advance of the flood front can be caused by the relative positions of the injection and production wells. Variations in rock permeability also result in an uneven advance of the displacing fluid. Gas zones, representing paths of low resistance, may result in non-uniform movement of the injected fluid. The more uneven the advance of the displacing fluid, the lower is the volumetric sweep efficiency and thus the lower the oil recovery efficiency. The effects of well arrangement and permeability variations, upon the gross movement of the injected fluid have been the subjects of extensive investiga-ti0ns.1,2,3,4,5,6 Gravity effects have been recognized as being present in frontal-drive operations'.' but there has been little quantitative measure of their magnitude. Very likely the reason gravity effects have not been studied is that this problem does not lend itself to simple experimental or mathematical analysis. The development of reservoir modeling techniques has made possible a quantitative determination of the effect of gravity on fluid segregation and thus upon the oil recovery performance of frontal-drive operations. This paper presents a progress report on a laboratory investigation of gravity effects in fully liquid-saturated, horizontal, uniform as well as non-uniform systems. These results were obtained from scaled experiments on both linear and five-spot systems within the range of conditions normally encountered in secondary recovery operations. This study is concerned with the gross movement of the injected fluid to the time it first reaches the production wellbore. Therefore, it should be emphasized that the results of this paper can not be used to determine the effect of gravity segregation on oil recovery efficiency at abandonment conditions. MODEL SCALING The science of scaling models is relatively new as applied as petroleum production problems, Many of the scaling techniques have been adaptions of those used in basic studies of heat and fluid flow. There are

Jan 1, 1958

By J. Husson, R. Iffly, M. Gondouin

A systematic variation of well deliverability, as reflected from isochronal back-pressure tests performed at regular intervals, has been observed in some gas condensate wells producing at high rates. The same effects have been obtained using a numerical model of gas and condensate flow which takes into account secondary gasoline deposited in the pore space as a result of pressure reduction, and non - Darcy flow of gas in the vicinity of the wells. Matching calculated values with previous test results has been possible, and future predictions have been obtained. An application of this method to the Hassi Er R'Mel gas-condensate field in Algeria is tentatively shown. INTRODUCTION Flow capacity of gas wells is generally derived from an analysis of back-pressure tests. The empirical equation q = C (?p2)n used by Rawlins and Schellhardtl can be derived rigorously assuming that steady-state radial flow of a dry gas of constant viscosity and compressibility is established during each flow period of the well tests. Furthermore, when Darcy's law applies in the entire flow region, the theory predicts that the exponent n is equal to 1. In low-permeability reservoirs, it was soon discovered that the time required to reach a stabilized flow often exceeded the duration of the flow periods normally available for testing wells. Consequently, transient gas flow had to be considered instead of the steady-state assumption previously used. This led to the isochronal testing procedure established by Cullender2 which has largely replaced conventional back-pressure testing. For dry gas fields, this method yields definite values of C and n equivalent to those of the empirical equation. These values should remain constant for each well as long as the permeability of the formation and the characteristics of the gas (viscosity and compressibility) do not change appreciably. This is the case when reservoir pressure remains -close to the original value and when the formation near the wellbore remains free of plugging. Under those conditions, stabilized flow potential curves of gas wells can be established from a single sequence of isochronal flow and shut-in periods. An analysis of a pressure build-up following a longer production period provides additional data on the transmissivity (k h/ u) of the reservoir, and eventually on the 'drainage radius rd of the well, which can be related to the value of C so that future performance of the well can be predicted using the concepts developed by A. Houpeurt.3 At high flow rates, Darcy's law no longer applies in the vicinity of the wellbore, and inertial effects in the high velocity gas flow introduce additional pressure drops. As a consequence, exponent n of the backpressure tests becomes smaller than 1, and a slight curvature of the log-log plot of ?p2 vs q can be predicted when going from very low to high rates of flow (Elenbaas and Katz).5 The effects of variations of viscosity and compressibility with pressure on the radial flow of dry gas in an infinite reservoir were taken into account by Jenkins and Aronofsky.7 Numerical solutions of the transient flow of an ideal gas in finite radial reservoirs were presented by Bruce, Peaceman, Rachford and Rice.8 In the case of gas condensate wells, however, the presence of gasoline in the pore space as soon as reservoir pressure is reduced below the dewpoint pressure further complicates the interpretation of flow tests so that the prediction of stabilized well performance becomes very difficult. Field observation shows that both C and n derived from isochronal tests vary in time, even when reservoir pressure has not changed appreciably. Such a variation cannot be attributed to any change of the gas characteristics, and must result from the effect of a gasoline saturation on the gas flow.

By Paul B. Crawford, Melvin B. Burton

A procedure is given which describes how the gelatin model may be-used to study the effect of mobility ratio on areas swept before and after breakthrough in fluid injection programs. Differences in mobility ratio are achieved through the proper selection of electrolytes for the invading and displaced fluids. The progress of the front is followed visually and photographed or traced onto plates (it desirrd positions before and after breakthrough. Example applications to standard flooding pattrrrls, vertically and horizontally fractured reservoirs and the in-situ combustion recovery process ore presentrtl. INTRODUCTION In water flooding, gas cycling and other fluid injection programs, the sweep efficiencies and swept areas after breakthrough are known to depend on the mobility ratio. Previous investigators have described the application of the X-ray technique, stepwise use of the poten-tiometric model and numerical procedures for studying these problems.1,3,4,6 Others have described the application of the electrolytic and gelatin models for studying sweep efficiencies at a mobility ratio near one.2,5,7,8,9 It is a purpose of this paper to describe a procedure for using the gelatin model to study swept areas at mobility ratios other than one. PROCEDURE The technique developed for gelatin model studies is composed of four phases: the construction of the model, the selection and preparation of electrolytes, the running procedure, and the recording of the swept areas. Each phase is discussed independently in order that the over-all procedure will not be confused. Construction of the Model Since some of the advantages of the gelatin model are its economy, and speed, it is desirable to have as small a model as is possihle without creating error through the inability to follow visually the progress of the colored front. For this work, it was found that a distance of 4 to 6 cm between salt bridges is convenient to use. Since the size is relatively small, it follows that the gelatin layer must be thin. For this study, 1/8 in. was used as the gelatin thickness in each case. The model was constructed on a flat 1/2-in. plastic (Lucite) base with boundaries constructed from 1/8-in. plastic strips. The well pattern to be studied was outlined with the 1/8-in. plastic strips on the base with narrow apertures (- - 2 mm) in the boundary to represent the wells. Behind each well opening, a rectangular area of some 8 to 12 sq cm was outlined by additional plastic strips. The plastic strips were fastened to the base by any suitable non-conducting cement so that the boundary system was liquid tight. The electrodes were constructed from sheet copper in a fashion that would allow easy removal. In this work, it was found that by simply forming a large U-shaped clip from a strip of the copper that would slip over the boundary was satisfactory and easily removed. The electrodes should be as wide as one of the sides of the rectangular area behind the well opening. If a fractured pattern is studied, fractures inside the well pattern may be represented by strips of copper of the desired shape mounted in some easily removable fashion. The reason that apertures (or salt bridges) were used for wells instead of metallic electrodes is that the plating of the electroactive ion at the cathode and the introduction of an electroactive ion at the anode has a detrimental effect on the swept area outline if these operations \rere permitted to be carried on within the well pattern itself. The anode may be used within the well pattern

Jan 1, 1957

By J. L. Moriarty, Irving Johnson, H. M. Feder

The crystallization of ternary alloys has been studied as a coprecipitation phenomenon by measuring the substitutional carrying of trace metals with CeCd11 precipitating from liquid cadmium. The Doerner-Hoskins logarithmic distribution law has been found to apply in all cases; the coprecipitation coefficients obtained were: 1.48 (La), 1.06 (Th), 0.63 (Pr), 0.24 (Gd), 0.17 (Sm), 0.13 (U), 0.10 (Er), 0.09 (Sr). Li, Na, K, Ba, Y, Sc, and Zr were not measurably carried. It has been observed that coprecipitation was most extensive for those tracers which form with cadmium an intermetallic compmdnd isostructural with CeCd11 or closely related thereto. For these tracers an inverse relation between Solubility in cadmium and coprecipitation coefficient was noted. THE partition of impurities between the liquid and solid phases of a metal or a congruently melting intermetallic compound has been extensively studied by the methods of zone melting.' However, when the composition of a liquid alloy and of the solid phase crystallizing from it differ grossly, partition of an impurity cannot be investigated in this manner. Partition of the impurity under equilibrium conditions may be studied by investigation of the appropriate phase diagram and partition under nonequi-librium conditions may be studied by observation of coring in the solidified alloys.2 The nonequilibrium case is of special interest in metallurgical practice because many complex alloys are cooled too rapidly to permit the uniform diffusion of impurities throughout the crystallizing phase(s). With this in mind, the present study was undertaken to obtain quantitative data on the phenomenon of coprecipitation via substitutional solid solution in certain metallic systems. It will be noticed that rather than adopting metallurgical ter- minology we have chosen to refer to the partition of impurities during alloy crystallization as coprecipitation.3 In the terminology of coprecipitation the impurity is referred to as the tracer and the precipitating phase as the carrier. From aqueous solution, coprecipitation may occur as a result of adsorption on the surface of the carrier, by occlusion of the liquid within the crystals, or by solid solution of the tracer in the carrier. By slow cooling of liquid metallic solutions it is relatively easy to form large, regular crystals of the carrier; the effects of adsorption and occlusion may be minimized and the influence of substitutional solid-solution formation may be made to predominate. Previous studies of the crystallization of some intermetallic compounds4 and of copper, silver, and gold5 have shown that the coprecipitation or partition behavior of trace metals is entirely analogous to that observed in aqueous systems when ionic substitution occurs, e.g., coprecipitation of Ra++ with BaSo4. The precipitation of CeCd11 from liquid cadmium was chosen for study. This choice was influenced by the following consideration. If the relative atomic sizes of the tracer atom and the atom for which it substitutes determine the tendency to coprecipi-tate, then a lattice having a substitutional site of well-defined size would be desirable. In CeCd116 (structure type cubic BaHg117) the cerium atom is completely surrounded by a cage of twelve nearly equidistant cadmium atoms. Thus, atoms much larger or much smaller than cerium should be poorly accommodated in this cage. The choice of a lanthanide central atom allowed the testing of the size hypothesis with a number of lanthanide tracers which differed in size but not in valence or electro-negativity. EXPERIMENTAL The general technique of this investigation was to prepare a slightly undersaturated solution of carrier in liquid cadmium to which a tracer was added. The concentration of tracer was always so low that saturation over the entire temperature range investigated was never exceeded. The solution was cooled in a stepwise fashion. After each step a por-

Jan 1, 1964

By F. J. Fayers, J. W. Sheldon

A theoretical discussion is presented of the behavior of the interface between two fluids of different physical properties when displacement is occurring along a thin tilted bed. An approximate equation of motion is deduced for the motion of the interface, suitable boundary conditions are derived and the conditions necessary for the approximation to be valid are determined. Steady-state behavior and the circumstances under which a steady-state interface can develop are discussed. Finite-difference solutions of the equation of motion for two characteristic water-drive problems are presenled. The theory of displacement in a thin bed may be regarded as being the porous-medium-equivalent of the classical "shallow-water wave" theory of hydrodynamics. INTRODUCTION The solution of moving interface problems are important in predicting the behavior of oil reservoirs under various conditions of production. Kidder4,5 has obtained analytic solutions for the fingering problem and for another particular case of the motion of the interface between two immiscible fluids of unequal density in a porous medium. In the present analysis, we analyze the problem of the motion of an interface along a slightly dipping bed whose length is long compared with its thickness. The requirement of a thin bed is necessary in order to give a simple differential equation for the motion. The two fluids are of unequal density and mobility. In the derivation of the equation of motion, it is assumed that the lighter fluid is being injected at constant rate at the high end of the porous layer, while heavy fluid is produced at the low end of the layer. The analysis applies equally well, however, to the reverse procedure in which the heavier fluid is injected at the low end of the tilted layer. It will be shown that the behavior of the system is critically dependent on which fluid has the higher mobility and that the interface can in certain circumstances achieve a simple progressive steady-state shape. The fluids may be considered to be immiscible or miscible. It is assumed, however, that the thickness of the interface region is negligible in comparison with other dimensions of the system, and that the strength of the saturation or concentration discontinuity remains constant during the motion. In conjunction with Buckley-Leverett theory and areal sweep-out calculations, the tilted interface problem is one of the fundamental problems of reservoir engineering; it defines the very important underpassing and overpassing phenomena associated with water-drives, gas-drives and secondary recovery operations. The tilted interface problem has been dealt with previously by Dietz.2 In the present paper, we endeavor to derive the pertinent mathematical results more rigorously, and in doing so we obtain modifications in both the differential equation and boundary conditions for those given by Dietz. We also attempt to show the connection between these results and some of the recent work on the stability of interfaces. Finally, we attempt to demonstrate the practical importance of the thin bed theory by showing the difference in behavior of two characteristic waterflood problems. In the favorable mobility case recovery is improved by increasing production rate, while in the unfavorable case undercutting or overcutting is minimized by reducing the displacement rate. PROPERTIES OF THE EQUATION OF MOTION Fig. 1 illustrates a porous layer of constant thickness H and tilt 8 with Fluid 1 of mobility ?l and density pl below Fluid 2 of mobility ?2 and density p2. The fluids are incompressible and flowing with volumetric flux q along the layer. Both the slope a of the interface with respect to the bed and the curvature of the interface are assumed to be small. What is meant by "small" is indicated in the derivation given in Appendix A of an approximate equation of motion for the interface. Other symbols are defined in the "Nomen-

By G. Sachs, E. J. Riping

All ferrous alloys can be made brittle by straining at sufficiently low temperatures. However, the changes in mechanical properties for different ferrous materials with decreasing testing temperature do not appear to follow any universal law. In particular, complex effects of testing temperature have been observed if cold worked steels were subjected to tensile tests at low temperatures. EFFECTS OF PRESTRAINING AT A TEMPERATURE ABOVE THE TESTING TEMPERATURE ON THE FRACTURE CHARACTERISTICS‡ A considerable amount of data has been presented by a number of investigators on the fracture characteristics of various cold worked steels.' These data usually relate to a two-step procedure first used by Davidenkov,² consisting of stretching by tension to a certain strain or " prestraining" at room temperature and then completing the tensile test, or "testing" at the temperature of liquid air. On the basis of data of this type made available to date, ferrous alloys can be classified into two groups with respect to their fracture characteristics observed on testing at a low temperature after prestraining at room temperature. Very complex phenomena have been observed for all annealed steels. On the other hand, some observations indicate that steels in certain other conditions represent a fundamentally simpler relation. In considering the effect of prestraining at one temperature on the properties obtained at some lower temperature, it should be expected that any cold work reduces the ductility retained in proportion to the magnitude of the cold work. However, the test data presented to date indicate that only heat treated (quenched and tempered) steels appear to conform to this expectation, according to the very limited test data in Fig 1, presented by McAdam, Geil and Mebs. This series of tests, Fig 1, shows that the retained ductility became smaller the larger the prestrain. However, this decrease is less than the amount of prestrain or, in other words, less than the decrease in ductility at the prestraining temperature. These effects of prestraining are fundamentally identical with those observed by Bridgman3 when a steel was prestrained (in tension) under hydrostatic pressure and then subjected to a regular tension test. It appears, therefore, that the basic effect of prestraining under conditions which yield a higher ductility than the (subsequent) testing procedure, consists of gradually increasing the total ductility of the metal from the initial low value of testing to the higher value of prestraining. On the other hand, all annealed steels, if cold worked by tension at room temperature and then tested at a lower temperature, provided that some ductility is retained, showed a far more complex behavior than that discussed above. This relationship is exemplified by Fig 2, for three pearlitic steels.4,5 These particular steels were selected because of the various shapes of their ductility-prestrain curves. All three steels, and any other investigated so far, suffered a rapid decrease in retained (and consequently also in total) ductility when subjected to small pre-strains. Then, after exceeding a certain prestrain it was generally observed that the ductility recovered.* This rather complex behavior of annealed steels may be tentatively correlated with the presence of stretcher strains at small restrains. Heat treated steels do not exhibit stretcher strains. This fact might possibly explain their simpler prestrain-retained ductility relationship. The destructive effect of these stretcher strains may possibly be associated with the presence of triaxiality, which may occur in any highly non-uniform stress and strain state (and which is retained as residual stress after unloading). Even a small degree of triaxiality may then cause a reduction in ductility, or even embrittlement at a sufficiently low testing temperature. The discussion presented above deals with the effects of cold working at a high temperature on the fracturing

Jan 1, 1950

By M. R. Tek, R. L. Nielsen

The scaling laws as formulated by Rapoport relate dynamically similar flow systems in porous media each involving two immiscible, incompressible fluids. A two-dimensional numerical technique for solving the differential equations describing systems of this type has been employed to assess the practical value of the scaling laws in light of the virtually unscalable nature of relative permeability and capillary pressure curves and boundary conditions. Two hypothetical systems — a gas reservoir subject to water drive and the laboratory scaled model of that reservoir — were investigated with emphasis placed on water coning near a production well. Comparison of the computed behavior of these par ticular systems shows that water coning in the reservoir would be more severe than one would expect from an experimental study of a laboratory model scaled within practical limits to the reservoir system. This paper also presents modifications of the scaling laws which are available for systems that can be described adequately in two-dimensional Cartesian coordinates. INTRODUCTION Present day digital computing equipment and methods of numerical analysis allow realistic and quantitative studies to be carried out for many two-phase flow systems in porous media. Before these tools became available the anticipated behavior of systems of this type could be inferred only from analytical solutions of simplified mathematical models or from experimental studies performed on laboratory models. To reproduce the behavior of a reservoir system on the laboratory scale, certain relationships must be satisfied between physical and geometric properties of the reservoir and laboratory systems. Where the reservoir fluids may be considered as two immiscible and incompressible phases, the necessary relationships have been formulated by Rapoportl and others.2-5 Rapoport's scaling laws follow from inspectional analysis of the differential equation describing phase saturation distribution in such systems. It will be recalled that these scaling laws presuppose three conditions: (1) the relative permeability curves must be identical for the model and prototype; (2) the capillary pressure curve (function of phase saturation) for the model must be linearly related to that of the prototype; and (3) boundary conditions imposed on the model must duplicate those existing at the boundaries of the prototype. These three requirements seldom if ever can be satisfied in scaling an actual reservoir to the laboratory system because: (1) The laboratory medium normally will be unconsolidated (glass beads or sand) while the reservoir usually is consolidated. Relative permeability and capillary pressure curves are usually quite different for consolidated and unconsolidated porous media. (2) The reservoir usually will be surrounded by a large aquifer which could be simulated in the laboratory only to a limited extent. (3) Wells present in the reservoir would scale to microscopic dimensions in the laboratory if geometric similarity is to be maintained. In view of these considerations, rigorous scaling of even a totally defined reservoir probably would never be possible. The purpose of this paper is to assess the practical value of the scaling laws in the light of the unscalable variables. This has been done by carrying out numerical solutions in two dimensions to the differential equations describing the flow of two immiscible, incompressible fluids in porous media for a field scale reservoir and a laboratory model of that reservoir. While both the reservoir and the laboratory model were purely fictional, each has been made as realistic and representative as possible. The field problem selected as the basis for the investigation was an inhomogeneous, layered gas reservoir initially at capillary gravitational equilibrium and subsequently produced in the presence of water drive. The laboratory model of this reservoir was designed to utilize oil and water in a glass bead pack. The numerical treatment employed was similar to that of Douglas, Peaceman and Rachford6 and it included both capillary and gravitational forces as

By R. M. Fowler, H. L. Hamner

DURING the last 25 years, there appeared in the literature a number of papers describing equipment and operating techniques for the determination of total oxygen in iron and steel. In the early papers, comparatively little attention was paid to side reactions that occur when iron is saturated with carbon at a high temperature in a high vacuum. As more experience was gained with the new method, the limitations appeared. Metals, with a boiling point lower than iron, distilled out of the crucible, condensed in the cooler part of the furnace, and subsequently reacted with carbon monoxide to form oxides and carbon. Aluminum, manganese, and the alkaline earth metals proved troublesome in this respect. Modifications have been devised to overcome the difficulties due to aluminum and manganese, while the alkaline earth metals are not of too much concern as yet to the steel chemist. The reactions of sulphur have received less attention, although it is generally recognized that some carbon disulphide is probably formed. Liquid air traps have been used in some cases to prevent any carbon disulphide from entering the analytical train. Sulphur was not a particular problem at the Union Carbide and Carbon Research Laboratories until the analysis of the irons of higher than usual sulphur contents became necessary. In order to describe the procedure developed for such irons, it will be necessary to first describe the type of vacuum-fusion equipment employed because the operation of vacuum-fusion equipment is an art to a considerable extent. At the Laboratories, vacuum-fusion equipment has been in daily operation since about 1930. Over the years, the evolution of this apparatus has been governed by the versatility required to analyze a wide variety of materials. Since all types of materials from high purity metals to slags are analyzed and, since seldom more than a few samples of one type are analyzed in succession, the apparatus was designed for accuracy and flexibility rather than for speed. A general view of the equipment is shown in Fig. 1, while Fig. 2 shows the relationship of the various parts. In Fig. 2, A is a vertical double-glazed 24X27/8 in. OD silica tube heated with a 20 kva mercury-arc high frequency generator. The tube is closed with a water-cooled brass head attached with de Khotinsky cement. A turret cap fits on this head and is sealed with a 1/8x3 in. "0" ring. The head contains a sight glass and two 8 in. horizontal glass tubes, for sample storage, cemented in with de Khotinsky cement. B is a four-stage Gaede mercury diffusion pump connected to the water-cooled head with a 11/2 in. copper tube through a flange with an "0" ring. The output of the vacuum pump feeds into an automatic Toepler F pump which compresses the gas sample over mercury in the measuring burettes, G and H. There are bypasses in this line for a McLeod gage C and a Piranni gage D, a manometer E, and for initial pumping out of the system with a Cenco Megavac backing-up pump. The reduction crucible shown in Fig. 3 is machined of Tungar Bulb quality (A.G.K.T.E.-2) graphite. The sample is guided into the crucible with the graphite funnel depicted on this figure. The arrangement of the crucible, radiation screens, etc. in the furnace is as follows: A 1 15/16x61/2 in. Alundum thimble, porosity RA-84, with two 1/8 in. holes near the top and opposite each other which take a pair of special tongs used for lowering it in place, is prepared. A 3/4 in. layer of powdered graphite is placed in this Alundum thimble, and the graphite crucible containing 20 g of ingot iron and 0.5 g of carbon is carefully centered on this bed of graphite powder and graphite powder is packed around it to the top of the graphite cru-

Jan 1, 1953

By A. J. May

THE Star shaft is near the north boundary of the group of mining claims belonging to the Ground Hog Unit of the American Smelting and Refining Co., near Vanadium, N. Mex. The shaft bins and surface plant are served by a siding from the Atchison, Topeka and Santa Fe railroad, and this factor of railroad transportation influenced the choice of site. This paper describes the shaft sinking only, and does not cover the surface plant or other parts of the mine or its equipment. A. J. MAY is Engineer in Charge of Shaft Sinking, Ground Hog Unit, American Smeling and Refining Co, Vanadium, New Mexico. El Paso Meeting, October 1948. TP 2737 A. Discussion of this paper (2 copies) may be sent to Transactions AIME before Feb. 28, 1950. Manuscript received Nov. 29, 1948; revision received June 22, 1949. Experience with this shaft seems of interest because of the mechanical mucking, good rate of progress, and the costs obtained at the present level of labor and material costs. The shaft has four compartments, as shown in fig. 1. Equipment: Riddell Mucker: The mucking was done by a clamshell bucket and a Riddell mucker, which, as in other shafts sunk with this machine, consisted of a carriage mounted on axles with double flanged wheels that traveled back and forth on a swinging track, parallel to the long axis of the shaft (fig. 2 to 6). Three air motors were mounted on the carriage; one hoisted the clam, one operated the bucket opening and closing device, and the third moved the carriage along the track. The carriage was designed with guide shoes and could be hoisted to the surface in either skip compartment without trouble (fig. 4). During blasting operations the clam was tied on the rails and the carriage was hoisted up a few sets to protect it from flying rock. No attempt was made to hoist the carriage and clam together lest the clam catch in the timbers. The track on which the carriage operated was built of extra heavy 6 in. pipe with 1x2 in. strap welded on top for rails. A Yzx4 in. angle was welded on the bottom of the 6 in. pipe with a 1 in. round rod welded in the angle for a filler and additional strength. The end or cross pieces of the track which kept the track to gauge, were also made of reinforced, heavy 6 in. pipe and were bolted to the long members on which the carriage traveled (fig. 2 and 3). This section of track was used, instead of heavy railroad rails, for two reasons: (1) this section would bend before it snapped and was much safer for the men in the bottom; and (2) it was a better blasting shield, and protected the timbers better, in fact, only five wall plates were broken by blasting in 1926 ft of shaft, although 150 to 225 lb of 40 pct powder was loaded per round, depending on the ground conditions. The swinging track was supported by four % in. hoisting cables anchored to the end plates. There were also four safety cables fastened two to three sets above the carriage, which would catch the carriage if one of the supporting cables should fail for any reason (fig. 7). Clamshell Buckets: Two clams were used, each of % yard capacity, both manufactured by Blaw-Knox, Nos. 662 and 663. The No. 663 clam was 4 in. wider and 5 in. shorter than the No. 662, but seemed to dig and load more satisfactorily. It was easier to hoist and lower when it came up for repair. It did not turn over as easily, had less spill

Jan 1, 1951

By A. J. May

THE Star shaft is near the north boundary of the group of mining claims belonging to the Ground Hog Unit of the American Smelting and Refining Co., near Vanadium, N. Mex. The shaft bins and surface plant are served by a siding from the Atchison, Topeka and Santa Fe railroad, and this factor of railroad transportation influenced the choice of site. This paper describes the shaft sinking only, and does not cover the surface plant or other parts of the mine or its equipment. A. J. MAY is Engineer in Charge of Shaft Sinking, Ground Hog Unit, American Smeling and Refining Co, Vanadium, New Mexico. El Paso Meeting, October 1948. TP 2737 A. Discussion of this paper (2 copies) may be sent to Transactions AIME before Feb. 28, 1950. Manuscript received Nov. 29, 1948; revision received June 22, 1949. Experience with this shaft seems of interest because of the mechanical mucking, good rate of progress, and the costs obtained at the present level of labor and material costs. The shaft has four compartments, as shown in fig. 1. Equipment: Riddell Mucker: The mucking was done by a clamshell bucket and a Riddell mucker, which, as in other shafts sunk with this machine, consisted of a carriage mounted on axles with double flanged wheels that traveled back and forth on a swinging track, parallel to the long axis of the shaft (fig. 2 to 6). Three air motors were mounted on the carriage; one hoisted the clam, one operated the bucket opening and closing device, and the third moved the carriage along the track. The carriage was designed with guide shoes and could be hoisted to the surface in either skip compartment without trouble (fig. 4). During blasting operations the clam was tied on the rails and the carriage was hoisted up a few sets to protect it from flying rock. No attempt was made to hoist the carriage and clam together lest the clam catch in the timbers. The track on which the carriage operated was built of extra heavy 6 in. pipe with 1x2 in. strap welded on top for rails. A Yzx4 in. angle was welded on the bottom of the 6 in. pipe with a 1 in. round rod welded in the angle for a filler and additional strength. The end or cross pieces of the track which kept the track to gauge, were also made of reinforced, heavy 6 in. pipe and were bolted to the long members on which the carriage traveled (fig. 2 and 3). This section of track was used, instead of heavy railroad rails, for two reasons: (1) this section would bend before it snapped and was much safer for the men in the bottom; and (2) it was a better blasting shield, and protected the timbers better, in fact, only five wall plates were broken by blasting in 1926 ft of shaft, although 150 to 225 lb of 40 pct powder was loaded per round, depending on the ground conditions. The swinging track was supported by four % in. hoisting cables anchored to the end plates. There were also four safety cables fastened two to three sets above the carriage, which would catch the carriage if one of the supporting cables should fail for any reason (fig. 7). Clamshell Buckets: Two clams were used, each of % yard capacity, both manufactured by Blaw-Knox, Nos. 662 and 663. The No. 663 clam was 4 in. wider and 5 in. shorter than the No. 662, but seemed to dig and load more satisfactorily. It was easier to hoist and lower when it came up for repair. It did not turn over as easily, had less spill

Jan 1, 1951