Search Documents

By A. T. Corey, C. H. Rathjens

INTRODUCTION Although the oil industry has been aware of the directional variability of permeability in porous rock, the directional variability of relative permeability has been largely ignored. Yet it is apparent that such an effect must be present in a system in which the distribution of oil and gas within the porous matrix is controlled by capillary forces. It is easy to visualize a rock composed of layers of fine and coarse material such that gas flow across the bedding planes would take place only after the average oil saturation had been reduced to a very low value. The fine layers, because of their greater capillarity, would remain saturated and act as barriers to the flow of gas after the coarse layers had been desaturated. Flow of gas parallel to the bedding planes would obviously take place at a much greater liquid saturation. Without more complete information concerning the geology of a reservoir than is generally available, it is not possible to predict exactly how such phenomena would affect the over-all performance of an oil field. It is possible, however, to predict qualitatively the effect of stratification on relative permeability measurements made on laboratory cores. In this investigation the effect of stratification was studied analytically by assuming that two porous materials with different capillary pressure-desaturation curves (but identical relative permeability curves) were in contact and in capillary equilibrium. As a qualitative check on the analytical results, cores having various degrees of visible stratification were used for relative permeability measurements made with fluids flowing both parallel and perpendicular to the bedding planes. A quantitative check was considered impractical because of the difficulty of devising models in which two materials of predetermined properties could be joined without the plane of contact becoming a discontinuity. THEORETICAL CONSIDERATIONS AND ASSUMPTIONS The assumption of capillary equilibrium in an oil-gas system implies that the difference in pressure between oil and gas is everywhere the same. This means that the curvature of the interfaces must be everywhere the same in order to satisfy the equation where PC is the pressure difference between phases, y the interfacial tension and r, and r2 are the major and minor radii of curvature. Depending on the pore size distribution of coarse and fine layers, the volumetric percentages of oil and gas in these layers will differ when equilibrium exists. The exact relationship can only be determined by obtaining the complete capillary pressure-desaturation curves for each of the porous materials in contact. It has been pointed out elsewhere' that the capillary pressure-desaturation curves of sedimentary porous materials can often be approximated by the relation where C is a constant and Soe is the effective saturation to oil based on a percentage of the pore volume effective to flow. In the same paper it was indicated that, as a first approximation, the values of oil relative permeability are given by and the values of gas relative permeability by For this analysis Eqs. 2, 3, and 4 were assumed to apply to each of two components of a hypothetical porous rock in capillary equilibrium. It was also assumed that each of the components had a residual wetting phase saturation of 20 per cent so that 80 per cent of the total pore volume was effective to flow. The permeability of the coarse stratum was taken as 100, and its displacement pressure was such that C in Eq. 2 had the numerical value of 1. The corresponding values for the fine stratum were 10 for the permeability and 10 for C. Units are not specified because they do not enter into the final results. The choice of the permeabilities and displacement pressure ratios was made to expedite the calculations. Any reasonable rock properties could have been chosen without changing the results qualitatively. Several arrangements of the two components were studied. Table 1 summarizes the resultant permeabilities obtained for four types of arrangement. RELATIVE PERMEABILITY CALCULATIONS The first step in the computation of relative permeability for the composite cores was the plotting of the capillary pressure-desaturation curves and the relative permeability curves for the individual components according to Eqs. 2, 3, and 4. At arbitrary

Jan 1, 1957

By E. N. Smith, J. C. Redmond

The increasing need for materials capable of withstanding higher operating temperatures for various applications such as gas turbine blading and other parts, rocket nozzles, and many industrial applications, has brought consideration of cemented carbide compositions. The well known usefulness of cemented carbides as tool materials is attributable to their ability to retain their strength and hardness at much higher temperatures than even complex alloys. However, it has been found that the temperatures encountered in cutting operations do not approach by several hundred degrees1 those involved in the applications mentioned above where the interest is in materials possessing strength and resistance to oxidation at temperatures of 1800°F and above. At these latter temperatures, the tool type compositions which are made up essentially of tungsten carbide are found to oxidize very rapidly and to produce oxidation products of a character which offer no protection to the remaining body. As a further consideration, the density of the tungsten carbide type compositions is high, from about 8.0 to 15.0. The refractory metal carbides as a class are the highest melting materials known as shown by Table 1 which summarizes the available data from the literature for the carbides of the elements which are sufficiently available for consideration for these uses. The density is also included in the table, since as mentioned above it is an important consideration in many of the applications for which the materials would be considered. It has been established that in the tool compositions the mechanism of sintering with cobalt is such as to result in a continuous carbide skeleton and that the properties of the sintered composition are thus essen- tially those of the carbide.2 On the hypothesis that this mechanism holds to a greater or less degree in cementing most of the refractory metal carbides with an auxiliary metal, it appears from Table 1 that titanium carbide compositions would offer possibilities for a high temperature material. Titanium carbide has extensive use for supplementing the properties of tungsten carbide in tool compositions. Although the literature contains several references to compositions containing only titanium carbide with an auxiliary metal,3,4,5,6 it may be inferred from the meager data that such compositions were deficient in strength and were considered to have poor oxidation resistance.7 Kieffer, for instance, reports the transverse rupture strength of a hot pressed TiC composition at 100,000 psi as compared to up to 350,000 psi for WC compositions. The work described herein was undertaken to determine the properties of compositions consisting of titanium carbide and an auxiliary metal and to improve the oxidation resistance of such compositions. It appeared possible that the inclusion of one or more other carbides with titanium carbide might improve the oxidation resistance and also that this might be more desirable than other means from the point of view of maintaining the highest possible softening point. Consideration of the available carbides in Table 1 suggests tantalum and columbium carbides because of their high melting points and general refractoriness. The work on improving oxidation resistance was concentrated on the addition of tantalum carbide or mixtures of tantalum and columbium carbide. The auxiliary metals used included cobalt, nickel and iron. It was also desired to learn the general physical properties of these compositions. Experimental Procedure The compositions used in this study were made by the usual powder metallurgy procedure applicable to cemented tungsten carbide compositions. The powdered carbide or carbides and auxiliary metal were milled together out of contact with air. In some cases cemented tungsten carbide balls and in other instances steel balls were used to eliminate any effect of tungsten carbide contamination. A temporary binder, paraffin, was then included in the mix and slugs or ingots were pressed with care to obtain as uniform pressing as possible. The ingots were presintered and the various shapes of test specimens were formed by machining, making the proper allowance for shrinkage during sintering. Thereafter the shapes were sintered in vacuum at temperatures of from 2800 to 3500°F. Final grinding to size was carried out by diamond wheels under coolant. The titanium carbide used contained a minimum of 19.50 pet total carbon and a total of 0.50 pet metallic impurities as indicated by chemical and spectrographic analysis. It was found by X ray diffraction examination with

Jan 1, 1950

By Terkel Rosenqvist

J. Chipman (Massachusetts Institute of Technology, Cambridge, Mass.)—The fact that the experimental work has been applied to copper rather than iron and that the paper is presented to the Iron and Steel Division, I regard as rather significant. It shows the unity of metallurgy and the fallacy of trying to cut it up by metals. This result for the solubility of sulphur in molten copper correlates with Professor Schuhmann's finding that the published data on the other side of the copper-sulphur miscibility gap are also in error. I should like to ask the author to say a little bit more about the sulphur capacity of the slag. T. Rosenqvist (author's reply)-—I hope that Dr. Chipman will find the derivation of the expression for sulphur capacity more clearly explained in the printed version of the paper than in the oral discussion at the meeting. I feel that this quantity, which actually is the ratio of two activities, can be measured more easily than the individual activities. Even if the ratio CaO/CaS is chosen as the standard state, the expression can be used for any slag, even for slags completely free of lime, and it represents a way to put the desulphurizing power of all slag constituents into one bag. Some doubt has been expressed as to whether oxygen ions really exist in calcium oxide and in molten slags. From a thermodynamic view point that question is of minor importance. The term oxygen ion activity, or any activity for that matter, is defined rigorously by the equation: activity = exp p/RT, where p is the change in free energy connected with the transfer of one mol of ions from the standard state into the slag. Whatever happens to the ion in the slag is of no concern to the thermodynamicist. Regardless of whether the ion is "free" in the slag or not, or whether it is present in a very small amount, its activity can always be expressed, and for a thermo- dynamic calculation that is all we need. However, ionic activities will only be of some real value if they are simple functions of the slag composition, or can be measured easily. Concerning the real nature of the oxygen in the slag, my feeling is that the oxygen atom has a rather multiplex nature depending on how strongly it is tied by covalent forces or polarized by the other atoms or ions present. The oxides of iron, cobalt, and nickel differ from calcium oxide and blast furnace slags as to the amount of free electrons that can give rise to electronic conductivity. In slags we know that the conductivity is mostly ionic. The fact that reversible emf's can be obtained with oxygen electrodes in certain salt melts, indicates a significant amount of oxygen ions in these melts. But extended work, e.g. polaragraphic studies and measurements of transference number; are needed to obtain quantitative information about the real structure of the slags. D. E. Babcock (Republic Steel Corp., Youngstown, Ohio)—-With reference to the ion, it might be well to remember Dr. Moses Gomberg. All of his life he had no use for the ionization theory and he contributed greatly to the field of chemistry on the assumption there was no such thing as ions. I do not think we have to worry about whether the oxygen is ionic or not. I think one thing specifically should be brought to your attention and this I think is one of the important contributions of Dr. Rosenqvist. He pointed out what we know as oxygen potentials or what is described as oxygen potentials. I have used this concept for a long period of time and I want to state that if this concept is properly applied, it vitiates much of what we have in the literature, or makes our usual ideas regarding oxidation seem primitive. That one thing is more valuable than almost all the rest of the discussion as a fundamental basis on which to build a reasonable in-

Jan 1, 1952

By Joseph Lindmayer, Karl M. Busen

If silicon is in contact with silicon oxide, a heterojunction is formed which induces an inversion layer ("channel"). Influences of impurities and structural parameters on the channel are discussed. There seems to be no relation between surface damage in silicon and the channel. Phosphorz~s doping of the silicon oxide has no pronounced effect on the channel. Channels which were well above the range predicted by the heterojunction model were related to the action of H* at the silicon/silicon oxide interface. Channels which were well below tlze range predicted by the heterojunc-tion model were related to a transition region in the oxide adjacent to the interface. In support of the assumed transition region and the H+ are the ex-periments described on the paper. Furthermore, the concept of metastable OH groups in silzca, known from other publications, has been used to explain certain channel characteristics by the formation and the accommodation of hydrogen in the silicon oxide. THE electrical characteristics of a semiconductor device in many cases are influenced by the conditions of its surface. Because these conditions often are changing with temperature, time, ambient, or other parameters, numerous attempts have been undertaken to stabilize them by suitable processes. Several years ago Atallal studied the silicon/silicon dioxide interface and since then increasing attention has been given to the silicon dioxide as a means to stabilize silicon surfaces. This led to the evolution of the planar silicon devices where oxide layers are used to protect places at which the junctions are intersecting the surface. The formation of the oxide layers on silicon can be carried out by a variety of processes: high-temperature thermal oxidation (exposing the silicon surface to oxidizing agents such as 02, HzO, COz), evaporation of SiOz, or anodic oxidation of silicon. Originally the oxides were considered able to "passivate" the silicon surface and it was understood that they 1) protected the surface against any influence from the ambient and 2) stabilized surface conditions which were necessary for proper device performance. It soon was realized that the second quality was not always obtainable: Atalla and Tannenbaum showed that, in the process of forming a layer on doped silicon by thermal oxida- tion, electrically active impurities in the silicon near the interface are redistributed, thus giving rise eventually to the formation of unwanted charge distributions. The degradation of planar junctions during reoxidation of silicon was observed by Barson et a. Surface studies with silicon planar junction structures indicated that the breakdown voltage depended on the kind of procedure which was applied during or after oxidation.*" Yamin and worthinge report on charge storage and dielectric properties of SiOz films at elevated temperatures. Another influential parameter on the silicon surface is discussed by Lindmayer and usen,' namely the appearance of a surface potential on silicon as a consequence of the heterojunction formed by the silicon/silicon oxide interface. In the present paper the influence of impurities and structural parameters in silicon oxide in connection with the heterojunction is a particular object of our investigations. I) EXPERIMENTAL The material used was 10 ohm-cm p-type silicon slices with either damaged surfaces (mechanically lapped or polished) or damage-free surfaces (chemically polished). The majority of the slices were oxidized for 1 hr and 20 min at 1250°C in a steam + oxygen atmosphere generated by passing O2 through HzO at 90°C (wet oxygen). This led to oxide layers of about 9000 A. On some slices oxide layers were grown in pure oxygen (dry oxygen). By a photoresist process a set of 20 by 200 mil windows was then etched out from the oxide layer. The pattern of the windows is shown in Fig. 1. In a subsequent process n-type pockets with surface concentrations between 2 x 10" and 5 x 10" cm"3 were formed by phosphorus diffusion. This allowed ohmic contacts between the silicon surface and tungsten probes which for electrical measurements were placed into two adjoining windows and

Jan 1, 1965

By J. W. Snavely

A practical mathematical treatment is presented for the determination and control of conveyor belt acceleration, particularly for conditions of starting where vertical curves are involved. A typical sample problem is analyzed, with required calculations, to clarify the procedure. THE part that acceleration plays in starting a belt conveyor and its effect on belt conveyor design are well understood in a general way. Its practical importance is easily overlooked, however, and under some conditions, it is absolutely necessary to give the problem of acceleration detailed study. Most handbooks on conveyor belting design adequately present basic data for the determination of acceleration values. This paper will only attempt to present practical thinking and a convenient method of treatment of acceleration in belt conveyor design. Mathematical Analysis In working out the various problems of conveyor belt acceleration, the starting point, as presented by the handbooks, is the familiar formula of "force of acceleration is equal to the mass times acceleration." By expressing these fundamental quantities in terms of belt conveyor design, it is possible to arrive at the unsuspected conclusion that the acceleration time for horizontal belt conveyors is independent of the load, and instead, dependent upon the belt speed, the type of drive arrangement and drive pulley, and the idler coefficient of friction. The mathematics leading to this conclusion are shown in Table I, which has been prepared to show this derivation. While at first the conclusion just given may not seem to be reasonable, further reflection indicates that obviously the type of drive pulley and the type of drive do affect materially the tension in the conveyor belt, and thus, as clearly shown, the time of acceleration is dependent upon the factors mentioned. Inasmuch as all of the factors except time are predetermined by the belt conveyor design, it becomes relatively easy to establish the accelerating time and to reduce further this time determination to a simple graph from which the time in seconds can be read directly. Such a graph is given in Fig. 1. The table appearing on Fig. 1 should be explained further. For a given belt speed, the time of acceleration can be expressed as a percentage of the belt speed. The time of acceleration is also dependent on the drive arrangement, and changes in the drive arrangement consequently change the time of acceleration. It further follows that for a given belt speed, the time expressed as a percentage of that belt speed also changes with the type of drive. Obviously then, it becomes possible to graph the percentage of speed for .each type of drive against the belt speed and accelerating time, after which, for a given belt speed and type of drive, the time can be read directly in seconds. Two constants were established for Fig. 1, the first one being the limiting of the maximum acceleration tension to 35 pct of the full load operating tension in the belt. The purpose of this is to limit the total tension imposed upon the belt during the acceleration period to 135 pct of the full load operating tension, 'which is the amount required to start or breakaway the fully loaded belt conveyor from rest. The other constant is the friction factor used for the idler equipment, which has been established as 0.022. For installations where it is necessary to establish the values of acceleration, invariably high grade idler equipment is used, and it has been established from field experience that 0.022 for the idler friction factor is amply conservative. The use of this friction factor for idlers must be tempered with judgment, of course, for occasions will arise where more power than indicated is required to start, even with the very best of equipment, such as low temperature operations that tend to congeal the grease in the bearings and thus produce additional friction drag. An inspection of the table in Fig. 1 affords a convenient rule of thumb method for determining the acceleration time, which conveniently can be 5 pct of the belt speed in seconds. The 5 pct of belt speed figure is close to the average for most types of drives. In using Fig. 1 it must be emphasized that it applies accurately to horizontal belt conveyors only.

Jan 1, 1953

By J. W. Snavely

THE part that acceleration plays in starting a belt conveyor and its effect on belt conveyor design are well understood in a general way. Its practical importance is easily overlooked, however, and under some conditions, it is absolutely necessary to give the problem of acceleration detailed study. Most handbooks on conveyor belting design adequately present basic data for the determination of acceleration values. This paper will only attempt to present practical thinking and a convenient method of treatment of acceleration in belt conveyor design. Mathematical Analysis In working out the various problems of conveyor belt acceleration, the starting point, as presented by the handbooks, is the familiar formula of "force of acceleration is equal to the mass times acceleration." By expressing these fundamental quantities in terms of belt conveyor design, it is possible to arrive at the unsuspected conclusion that the acceleration time for horizontal belt conveyors is independent of the load, and instead, dependent upon the belt speed, the type of drive arrangement and drive pulley, and the idler coefficient of friction. The mathematics leading to this conclusion are shown in Table I, which has been prepared to show, this derivation. While at first the conclusion just given may not seem to be reasonable, further reflection indicates that obviously the type of drive pulley and the type of drive do affect materially the tension in the conveyor belt, and thus, as clearly shown, the time of acceleration is dependent upon the factors mentioned. Inasmuch as all of the factors except time are predetermined by the belt conveyor design, it becomes relatively easy to establish the accelerating time and to reduce further this time determination to a simple graph from which the time in seconds can be read directly. Such a graph is given in Fig. 1. The table appearing on Fig. 1 should be explained further. For a given belt speed, the time of acceleration can be expressed as a percentage of the belt speed. The time of acceleration is also dependent on the drive arrangement, and changes in the drive arrangement consequently change the time of acceleration. It further follows that for a given belt speed, the time expressed as a percentage of that belt speed also changes with the type of drive. Obviously then, it becomes possible to graph the percentage of speed for each type of drive against the belt speed and accelerating time, after which, for a given belt speed and type of drive, the time can be read directly in seconds. Two constants were established for Fig. 1, the first one being the limiting of the maximum acceleration tension to 35 pct of the full load operating tension in the belt. The purpose of this is to limit the total tension imposed upon the belt during the acceleration period to 135 pct of the full load operating tension, which is the amount required to start or breakaway the fully loaded belt conveyor from rest. The other constant is the friction factor used for the idler equipment, which has been established as 0.022. For installations where it is necessary to establish the values of acceleration, invariably high grade idler equipment is used, and it has been established from. field experience that 0.022 for the idler friction factor is amply conservative. The use of this friction factor for idlers must be tempered with judgment, of course, for occasions will arise where more power than indicated is required to start, even with the very best of equipment, such as low temperature operations that tend to congeal the grease in the bearings and thus produce additional friction drag. An inspection of the table in Fig. I affords a convenient rule of thumb method for determining the acceleration time, which conveniently can be 5 pct of the belt speed in seconds. The 5 pct of belt speed figure is close to the average for most types of drives. In using Fig. 1 it must be emphasized that it applies accurately to horizontal belt conveyors only.

Jan 1, 1952

By Jun Hino, P. G. Shewmon, P. A. Beck

THE investigations of Crussard,' of Guinier and Tennevin,' and of Dunn and Daniels," indicate that the subgrains formed in a cold worked and annealed metal are capable of growing at each other's expense during annealing, if the temperature of annealing is sufficiently high and the time is long enough. The results of Dunn and Daniels are particularly convincing in showing that subgrain growth is essentially the result of the free surface energy associated with the subboundaries. In subgrain growth, as in ordinary grain growth, energy is gained as a result of the decrease in total subboundary surface area per unit volume. Recently, Wood and Scrutton4 found that the rate of subgrain growth upon annealing increased very considerably, if simultaneously a creep strain was applied to the specimen at a low strain rate. Working with 99.98 pct pure fine-grained aluminum, these investigators found that the continuous X-ray diffraction back-reflection rings of material strained at room temperature remained continuous after heating for ten days at 250 °C. However, when heated at the same temperature under a stress of 1000 psi even for only three days, the formerly continuous X-ray diffraction rings broke up into numerous dots, which were fairly clearly separated from each other. The continuous X-ray diffraction rings were interpreted as indications of a very small subgrain size, not resolved by the X-ray diffraction method used. The breaking up of the continuous X-ray diffraction ring during annealing under stress was taken as an indication of a great increase in subgrain size, so that the individual subgrains could then be resolved. The effect of simultaneous strain at a low strain rate in accelerating subgrain growth, discovered by Wood and Scrutton and designated by them as "cell growth," is of fundamental importance. The experiments described in this note were carried out in order to confirm Wood and Scrutton's results by direct metallographic observation. Also, information was sought as to the minimum creep strain necessary to produce this effect. A fine grained high purity aluminum strip was prepared by alternate 33 pct rolling and annealing treatments for 1 hr at 350°C. Specimens cut from this strip were subjected to a relatively fast creep strain of 7.2 pct in 3.5 min at 300°C under a constant load initially corresponding to 1185 psi. The subgrains set up were large enough (about 0.015 to 0.03 mm) to be clearly observed (Fig. 1) at X200 magnification with polarized light, after electrolytic polishing and anodic etching, producing a fine oxide film." The corresponding X-ray diffraction pattern is shown in Fig. la. After the fast creep strain treatment a portion of the specimen was subjected to a creep strain of 8.3 pct in 44 hr at 350°C under a constant load initially corresponding to 320 psi. Another portion, annealed under the same conditions, but not strained, served as control specimen. The subgrain size of a typical area of the specimen annealed under strain and of the one annealed without strain is shown in Figs. 2 and 3 (X-ray diffraction patterns, Figs. 2a and 3a). Annealing without strain produced clearly observable subgrain growth (subgrain size: about 0.03 to 0.05 mm). The effect of simultaneous strain was to increase greatly the rate of subgrain growth (resulting in a subgrain size of approximately 0.05 to 0.13 mm), in accordance with Wood and Scrutton. In another experiment, the effect of the amount of simultaneous strain was studied. The initial subgrain structure was set up by fast creep, as described previously. The specimen was then heated at 350 °C under a constant load initially corresponding to 320 psi, as above, for 2.2 hr (0.24 pct strain), and 8.8 hr (0.58 pct strain). Comparison with the corresponding unstrained control specimens by means of X-ray diffraction showed that 0.58 pct strain definitely had an effect, but the effect of 0.24 pct strain was doubtful.

Jan 1, 1953

By Howard Martens, Pol Duwez

IN two papers published in 1949, alloys of chromium with the refractory metals tungsten, molybdenum, tantalum, and columbium were investigated in view of their possible use as high temperature resisting materials. For the Cr-Ta system, a partial phase diagram was presented and the only intermediate phase was identified at Ta2Cr3. A phase of the same composition was also observed in the Cb-Cr system. The X-ray diffraction data presented in these papers, however, were insufficient for crystal structure determination. It is shown in the present study that the only intermediate phase in both the Ta-Cr and the Cb-Cr systems corresponds to the ideal stoichiometric ratio TaCr2, or CbCr2. Both structures are cubic, MgCu, type. At high temperature, however, TaCr2 has a hexagonal MgZn, type structure, which can be retained at room temperature by fast cooling. The alloys were prepared by melting in a helium arc furnace on a water-cooled plate. The design of the furnace was essentially the same as that described in ref. 3. Some alloys were also obtained by sintering compacts made of the mixed powders pressed at 80,000 psi. The sintering was carried on for 4 hr at 1375°C. The tantalum and columbium powders were supplied by Fansteel Metallurgical Corp., North Chicago, 111. The tantalum powder was the reagent grade, with a particle size smaller than 400 mesh and a total impurity content less than 0.1 pct. The columbium powder was smaller than 325 mesh and contained approximately 0.1 pct C and traces of Fe, Ti, and Zr. The electrolytic chromium powder from Charles Hardy, Inc., New York, was smaller than 300 mesh and contained about 0.1 pct Na, 0.05 pct Ca, and traces of Cu, Al, Mg, Si, and Co. Powder diffraction patterns were obtained with a 14.32 cm camera, using copper Ka radiation filtered through nickel foil. The powder pattern of the TaCr2 alloy obtained by sintering at 1375'C was different from that obtained on the same alloy rapidly cooled from the melt. Contrary to this result, the powder pattern of CbCr2 was the same, whether the alloy was made by sintering at 1375°C or by melting, and was similar to that of the TaCr, sintered. It was also found that the structure of the TaCr2 specimen obtained by melting was retained after heating for 4 hr at 1590°C, but transformed into the structure found in the sintered specimen after heating for 4 hr at 1375°C. Hence, the structural change of TaCr2, appears to be a reversible polymorphic transformation. CbCr2 and ToCr2 Structure, Low Temperature Form By using large scale Hull-Davey charts, the powder pattern of CbCr, and of the low temperature form of TaCr2 were readily interpreted on the basis of a face-centered cubic lattice with a parameter of approximately 6.95 kX. The indices of the reflections together with the values of sin' 0 are given in Tables I and 11. From this list of observed reflections, it appears that the (200), (600), (024), (046), and (028) reflections are missing. The lack of (h00) reflections for h 4n indicates a four-fold screw axis. The missing (Okl) spectra for k + 1 An indicate the existence of a diamond glide d. The combination of these symmetry elements can be found in the O— Fd3m space group, which is therefore the most probable one. After having determined the approximate density of TaCr, by the immersion method, the number of molecules per unit cell was calculated and found to be nearly eight. This information, added to the fact that the most probable space group is O leads to the consideration of a structure of the MgCu2 type, in which the atoms have the following positions: 8 magnesium in a and 16 copper in d. On the basis of this structure, intensities were computed by means of the usual formula: 1 cos'20 I a sin2 cos where F is the structure factor; 8, the Bragg angle: and p, the multiplicity factor. As shown in Tables I

Jan 1, 1953

By P. R. Celmer, Leonard R. Weisberg

THE principle of fractional crystallization has been successfully used to prepare high-purity (99.999 pct) Ga. Hoffman and Scribnerl removed single crystals of gallium solidifying in a gallium melt, while Zimmerman2 grew single crystals from the melt by the Kyropolous technique. In contrast, attempts at purifying gallium by zone refining have been less successful. ichards, reported that despite the passage of 40 zones through a gallium ingot, there still remained 5 to 70 ppm each of Cu, Fe, Ca, Mg, Si, Al, and Ag. Previously, Detwiler and Fox4 detected only one impurity, Pb, segregating in zone-refined ingots. These surprising results prompted an investigation of the factors controlling impurity segregation in gallium. Possible reasons for this were insufficient diffusion of impurities in the melt; recontamination of the melt by its oxide film which is not affected by the passage of the zone; reaction of gallium with the boat; sudden freezing of gallium following supercooling, especially since gallium easily supercools to and trapping of impurities at grain boundaries. Impurity segregation tests were carried out by directionally freezing gallium using the Bridgman rnethd, modified in that the molten gallium is lowered out of a furnace into a slush of dry ice and trichloroethylene, thus minimizing supercooling. Since the gallium is contained vertically, the oxide film is in contact only with the tail end of the melt. It was found that Teflon makes an excellent crucible for gallium since it is quite pure, non-reactive, translucent, flexible, machinable, and is not wet by gallium. The gallium crystals could be grown at various speeds, and the melt could be vigorously stirred by a Teflon rod moving through the gallium in a vertical reciprocating fashion. Single crystals could be grown by placing a solid Teflon plug at the bottom of the melt drilled out in such a way to cause the solidifying gallium to follow a winding path. Thus, even if many crystals are originally nucleated, only one grain will predominate. The grain structure of the gallium crystals was revealed by an etchant composed of equal volumes of HC1, HNO3 and HF, diluted with water to half strength. Emission spectrographic analyses were carried out on samples removed from the front and tail ends of the resulting gallium ingots. Typical results of this study are summarized in Table I. The rate of freezing in all three cases was about 1 in. per hr. It can be seen that even though stirring of the melt does help, it is even more important to grow a single crystal of Ga in order to obtain good segregation of impurities. The effect of crys-tallinity on the segregation of impurities was previously observed6 in the directional freezing of germanium; however, in this case, the effect was much less pronounced. This dependence of impurity segregation on the crystalline perfection of Ga may be related to its thermal conductivity which is the most anisotropic of all metals.7 The anisotropic thermal conductivity can cause the solid-liquid interface to be nonuniform, thus leading to trapping of impurities during freeing, and therefore reduced segregation. In conclusion, it is indicated that zone refining of gallium would be more successful if seeding and similar precautions are taken to insure single crystal growth. The authors are indebted to Mr. H. H. Whitaker for the spectrographic analyses and to Drs. B. Abeles and F. D. Rosi for helpful advice and encouragement throughout the course of this work. This research was supported by the Electronics Research Directorate, Air Research and Development Command, under Contract No. AF33(616)-5029. REFERENCES 'J. 1. IToffmon and B. T. Scribrer: I. Research h'atl. Bur. Standards, 1935, "01. 15, p. 205. 'W. Zirnmerman: Science, 1954, vol. 119, p. 41. %J. L. Richards: Nature, 1956, vol. 117, p. 182. 'D. P. Detwiler and W. M. Fox: I. Metals, 1955, "01. 7, p. 205. 5P. W. Bridgman, Proc. Am. ilcod. Sci., 1925, vul. 60, pp. 305,385,423. "S. L1. Christian: private c ommuni cation. 'K. W. Powell: roy. Sac., 1951, vol. 209, p. 525. 'W. G. Pfann: Zone Re fining, p. 20. John Wiley and Sons, Inc., New York, L058.

Jan 1, 1962

By J. W. Snavely

A practical mathematical treatment is presented for the determination and control of conveyor belt acceleration, particularly for conditions of starting where vertical curves are involved. A typical sample problem is analyzed, with required calculations, to clarify the procedure. THE part that acceleration plays in starting a belt conveyor and its effect on belt conveyor design are well understood in a general way. Its practical importance is easily overlooked, however, and under some conditions, it is absolutely necessary to give the problem of acceleration detailed study. Most handbooks on conveyor belting design adequately present basic data for the determination of acceleration values. This paper will only attempt to present practical thinking and a convenient method of treatment of acceleration in belt conveyor design. Mathematical Analysis In working out the various problems of conveyor belt acceleration, the starting point, as presented by the handbooks, is the familiar formula of "force of acceleration is equal to the mass times acceleration." By expressing these fundamental quantities in terms of belt conveyor design, it is possible to arrive at the unsuspected conclusion that the acceleration time for horizontal belt conveyors is independent of the load, and instead, dependent upon the belt speed, the type of drive arrangement and drive pulley, and the idler coefficient of friction. The mathematics leading to this conclusion are shown in Table I, which has been prepared to show this derivation. While at first the conclusion just given may not seem to be reasonable, further reflection indicates that obviously the type of drive pulley and the type of drive do affect materially the tension in the conveyor belt, and thus, as clearly shown, the time of acceleration is dependent upon the factors mentioned. Inasmuch as all of the factors except time are predetermined by the belt conveyor design, it becomes relatively easy to establish the accelerating time and to reduce further this time determination to a simple graph from which the time in seconds can be read directly. Such a graph is given in Fig. 1. The table appearing on Fig. 1 should be explained further. For a given belt speed, the time of acceleration can be expressed as a percentage of the belt speed. The time of acceleration is also dependent on the drive arrangement, and changes in the drive arrangement consequently change the time of acceleration. It further follows that for a given belt speed, the time expressed as a percentage of that belt speed also changes with the type of drive. Obviously then, it becomes possible to graph the percentage of speed for .each type of drive against the belt speed and accelerating time, after which, for a given belt speed and type of drive, the time can be read directly in seconds. Two constants were established for Fig. 1, the first one being the limiting of the maximum acceleration tension to 35 pct of the full load operating tension in the belt. The purpose of this is to limit the total tension imposed upon the belt during the acceleration period to 135 pct of the full load operating tension, 'which is the amount required to start or breakaway the fully loaded belt conveyor from rest. The other constant is the friction factor used for the idler equipment, which has been established as 0.022. For installations where it is necessary to establish the values of acceleration, invariably high grade idler equipment is used, and it has been established from field experience that 0.022 for the idler friction factor is amply conservative. The use of this friction factor for idlers must be tempered with judgment, of course, for occasions will arise where more power than indicated is required to start, even with the very best of equipment, such as low temperature operations that tend to congeal the grease in the bearings and thus produce additional friction drag. An inspection of the table in Fig. 1 affords a convenient rule of thumb method for determining the acceleration time, which conveniently can be 5 pct of the belt speed in seconds. The 5 pct of belt speed figure is close to the average for most types of drives. In using Fig. 1 it must be emphasized that it applies accurately to horizontal belt conveyors only.

Jan 1, 1953

By M. R. Tek, K. H. Coats, D. L. Katz

The nature and the limits of validity of Darcy's law US applied to the flow of natural gas through reservoirs has been considered in order to resolve some controversial aspects of the effect of turbulence on pressure drops. The equivalence between various concepts and viewpoints advanced in the past by several investigators to explain how and why a gas well does not necessarily perform according to Darcy's law is shown. Starting with generalized equations of flow of fluids through porous media, a partial differential equation has been derived which accurately represents the flow at all rates. This equation has been numerically solved using an IBM 704 digital computer. The results permit plots of unsteady radial pressure distribution curves from which specific isochronal backpressure curves may be constructed. These back-pressure curves show the effect of the ,8 factor on the slope of the back-pressure curve. The calculations further indicate that the drainage radius for a gas well in turbulent flow propagates at a rate dependent upon the rate of production at the wellbore. This is quite different from the case with liquid flow or natural-gas flow in laminar regime. Additionally, the effect of reservoir inhomogeneities and crossflow between layers of different permeability on the back-pressure performance of gas wells has been conridered. In light of the current numerical results the significance and linlitation of the rate of flow function Y proposed by Smith" has been discussed. INTRODUCTION The relationship between the pressure drop and flow rate in problems of fluid flow through porous media is known to be affected by the nature of flow through the porous matrix. It has been observed by many that, for a range of flow rates, the pressure drop remains proportional to the rate of flow. When some flow rate is reached, however, it is usually observed that the pressure drop gradually begins to increase more than proportionally to the flow rate. It is well known that this phenomenon was first observed by Osborne Reynolds in 1901 in experimenting with flow through pipes. In his classical experiments, Reynolds made visual observations on the condition of streamlines evidenced by injecting a dye into water flowing through glass tubes. In these experiments, the abrupt transition between steady, "streamline, laminar" flow and unsteady random turbulent flow was found to be a function of the dimensionless group (Dvp/u), now known as the Reynolds number. During these experiments, in addition to observations on the nature of flow regimes, the proportionality between flow rate and pressure drop in laminar flow was contrasted with the nonlinearity between these variables in turbulent flow. Fancher and Lewis' reported data on various consolidated and unconsolidated sands in 1933. Their conclusions were that ". . . the flow of fluids through these porous materials closely resembles that through pipes; that there is a condition of flow in porous systems which resembles viscous flow, another which corresponds to turbulent; that the change from one type to the other takes place at a definite and reproducible condition for each system". In 1947, Brownell and Katz published a method to predict the laminar and turbulent flow behavior from the particle size, bed porosity and the particle sphericity, employing the friction factor-Reynolds number charts for pipes. Several investigators have verified the work of Fancher, Lewis and Barnes and presented their data as friction factor-vs-Reynolds number plots.' The equation which would represent the pressure gradient over the whole range of velocity must have an added term over that represented by Darcy's law. Accordingly, the pressure gradient necessary to sustain flow at the velocity (v) through a porous medium may be represented by the following equation, suggested by Forscheimer." The nature and the range of validity of Darcy's law has been the subject of studies by many investigators over the past years. While everyone seemed to agree on the need for a quadratic correction term to Darcy's law to make it effective over the range of velocities, the concept of inception of turbulence and the use of the term "turbulent flow" remained controversial. Some fluid dynamicists define turbulence as a flow

By Lyman H. Hart

Some of the significant facts that will affect the supply and demand for metals during the next few decades are given in this presentation. This is important because the only hope for intelligent guidance of the national destiny depends upon presenting a consistent and accurate picture of basic problems. The metals problems now fomenting are only slightly less important than those precipitating the energy crisis. Admittedly, it is difficult to generate serious thinking on this subject when the world is enjoying easy access to all metals at modest prices. This is possible because of the current world capacity to over-produce, and United States credit standing, at least up to now, has permitted buying freely on world markets. Although both these vital conditions are subject to change, there is little doubt that if the U.S. financial house is kept in order, we can muddle along for a while on a generally expanding, metal-based economy. But should our credit standing be permitted to collapse, or should there be a tightening of world supplies of metals, the nation would immediately find itself in trouble. Unfortunately, clear signals indicate that both of these conditions are threatening. The metal age started many centuries ago, and with the exception of the noble metals, barely a dent was made in world reserves until the present century. Then, as a result of the pioneering ideas of D. C. Jackling and Paul Gemmel, and the willingness of the McNeals, Penroses and Guggenheims to back ideas with money, a virtual explosion of metal production was propagated throughout the world. It is important to examine the causes of this expansion to determine whether it is a trend that may be sustained or whether it is anomalous and therefore attributable to circumstances not likely to be repeated or continued. This period of plenty is a once only phenomenon. Vast metal resources have been sitting around here for eons, while the world was unaware of or unconcerned with their existence. But when population began to grow exponentially, and ways were discovered to use metals to develop machines to do the work of men and animals, the exploitation of raw materials increased at a staggering rate. And now the realization is dawning that minerals are a wasting asset and are beginning to present supply problems that should begin to be of concern. Two main factors have caused this condition; One is the population growth, and the other is the expanding demand for irreplaceable raw materials as nations become more developed. A group of systems analysts, at Massachusetts Institute of Technology in a book Limits of Growth, has warned that at present rates of population growth and resource consumption together with environmental constraints, a point will be reached within 100 years when the world population will become totally incapable of supporting itself. The authors conclude that this end result could be avoided if a planned equilibrium could be established at a lower rate of population growth and resource consumption. Fortunately, there is evidence that the population problem is moderating. The growth rate in the U.S. has decreased from 2.3% per year in 1900 to 1.1% in 1970. But, even at the lower rate, population will double in 70 years. The world rate is greater, and if it continues, world population will rise from 3.5 billion to nearly 10 billion in 50 years. All of these projections are questionable because the dangers of growing population are now being recognized. A marked decrease in growth rate has been noted in the more developed countries-in fact, it was zero in the U.S. in 1972. The other factor affecting consumption of raw materials is the increase in per capita use of most items as individuals become more affluent. Many economists believe that developing countries such as China and India represent a vast future potential market; but there are others, with whom the author agrees, who believe over-population tends to stagnate development. Nevertheless, it is a fact that vigorously developing societies increase consumption at a rate far above their population growth rate.

Jan 1, 1974

By J. P. Hirth, A. H. Clauer

A simple graphical method is presented for the determination of climb forces on dislocations exerted by uniaxial stresses. In conjunction with standard stereographic projections, the technique is applicable for any crystal system. Climb forces on glide dislocations in bcc crystals are determined as an application. In connection with a study1 of elevated temperature creep of molybdenum single crystals, a detailed analysis of the climb force on a dislocation was necessary. In general, such forces can be calculated for an arbitrary dislocation segment from the Peach-Koehler equation2 (or one of its variants3"6) where Fk are the components of the force per unit length of dislocation line, ?i of the unit sense vector tangent to the line, ojl of the stress tensor, bl of the dislocation Burgers vector, and is the Einstein permutation operator with components ?123 = ?231 = ?321 = —?132 = — ?321 = —?—213 = unity, and all other components zero. It is understood in the suffix notation of Eq. [I.] that repeated indices are to be summed over 1, 2, and 3. For the most general coordinate system, the matrix manipulations required to solve Eq. [I] are fairly complex. However, a simplification can be achieved by specifying a coordinate system based on the dislocation in question. The pertinent system involves cartesian coordinates xi, x;, xi, with attendant unit vectors e'i; ei being parallel to the Burgers vector b, and e'3; being parallel to the climb directions (b x ?), i.e., normal to the slip plane. In this coordnate system b = (bi, 0, 0) and ? = (?'1,?'2, 0). Hence, Eq. [I] yields for the climb force per unit length F3 = ?'2s'11b'1—?'1s'12b'1 [2] The first term on the right side of Eq. [2] is the climb force on the edge component of the dislocation, while the second is the (cross-slip) force normal to the glide plane acting on the screw component. Thus, the problem of determining the climb force reduces to that of resolving the stress components —il and s'12 in the xi coordinate system. Several authors7- l0 have shown, for the glide case, that such stress resolution can be performed most readily by graphical methods. Therefore, we present a similar method for the climb case. Of the suggested methods, we follow that of Hartley and Hirth,9 which can be performed directly with a standard stereographic projection for the crystal system in question. We treat the case of simple tension, or compression, this being the most common creep test in which climb is important. A set of cartesian coordinates (x1, x2, x3) is fixed with x3 parallel to the specimen axis. This set is connected to the x'i set by the factors defined in Table I. A perspective view of the coordinate systems relating them to the glide ellipse, is given in Fig. 1, while a stereographic projection of the coordinate systems is presented in Fig. 2. The transformation matrix connecting the xi and xi coordinates is x'i =aijXj [3] where an a11 a13 aij = a21 a22 a23 a31 a23 [4] s The corresponding transformation for the stresses is i [5] The only nonzero component of okl for the simple tension case is s33,. Hence, using Eqs. [4] and 151, we find that Eq. [2] becomes F3 = b'1s33 sin2 cos ?[cos ? sin a - sin ? cos a] [6] = m1b'1s033 sin a- rn2b's33 cos a where sin a = 5: and cos a = ?i, with a the angle measured from b to ?, i.e., a polar angle in the x1, x2 plane. The same equation with a sign change applies for compression. A graphical plot of the "Schmid" factor m1, relating to the climb force on the edge component, is presented in the stereographic projection of Fig. 3. In this projection the glide plane normal points toward the top of the page and the glide direction points normal to the plane of the page. To use the graph, one

Jan 1, 1970

By E. S. Machlin

An extension of the thermodynamic theory has been made for the case of irreversible growth processes occurring by the motion of an interface. The theory is applicable to such diverse phenomena as diffusion, growth in recrystallization, continuous grain growth, growth of carbides, etc., growth of eutectoid products, growth in solidification, recovery, "slipless" flow, etc. THE publication of a recent book1 has served to focus attention on a very powerful means of treating certain irreversible processes. This method which has been formally described as the thermodynamic theory of irreversible processes is applicable to processes which involve an approach to equilibrium and for which the deviations from equilibrium are small. The irreversible processes can be classified into three groups: chemical reactions, transport processes (diffusion, heat and electrical flows), and relaxation phenomena involving a degradation of internal energy to more stable states. The limits of applicability of the theory can be precisely defined1,2 for each type of irreversible process. For example, in a chemical reaction at constant temperature and pressure, the Gibbs free energy released per mol in the process should be less than the thermal energy, RT. Another type of irreversible process, which sometimes involves a combination of the first two of the above-mentioned groups, is treated in this paper. This process is one characterized by the motion of an interface separating two regions having different values of free energy and may be briefly described as a growth process. Prigogine2 and Herring3 have treated special cases of this type of process. Previous conscious applications of the theory in metallurgy have been limited to the field of diffusion, one of the transport processes. Darken,4 Bardeen,5 Prigogine2 and others have made significant contributions in this respect. Some other applications of the theory are described in this paper. The Theory The thermodynamic theory of irreversible processes is based on the work of Onsager,6 DeDonder,7 Prigogine2 and others. A resume of the theory has been given by Prigogine2 and DeGroot.1 A brief description of the theory follows, although for a complete understanding the reader is urged to read the references. It is first assumed that the change in entropy, even for a system removed from equilibrium, is given by dS = [dU + pdV - Si µi, dn, - Sk Pk dxk]T-1 Using this expression, the irreversible entropy production in the system is calculated by subtracting the contribution to the change in entropy of the system by transfer of heat, work, or matter from the surroundings. Thus, for example, the irreversible entropy production of a system at constant temperature and pressure is given by dtS = — — dG (DeDonder7) Now it has been shown by Prigogine,2 DeDonder,7 and DeGroot1 that the rate of irreversible entropy production diS/dt, when calculated in the manner suggested, can be written as a sum of products of conjugate forces XK and fluxes Jk, i.e., d,s/ = Sk Jk Xk [I-1] dt The assumption is next made that a linear relation exists between a given flux and the forces, obtained from Eq. I-1, i.e., Ji = Sk Lik Xk (i = 1, 2, . . . n) [I-2] Generally, there is a degree of freedom in choice of the fluxes and forces. However, for all the choices consistent with Eq. I-1 the Onsager8 relations, based on the principle of microscopic reversibility, namely, Lik = Lki, are valid. The coefficients Lik, depending upon the choice of the conjugate fluxes and forces, may be more or less dependent upon the parameters defining the state of the system. In general, the proper choice of fluxes and forces utilizing Eq. I-1, makes the coefficients Lik independent of time. Also, because many of the three classes of irreversible phenomena have already been treated,1,2 a proper choice can be made by analogy. In any case, the validity of Eq. 1-2, from whatever choice made using Eq. I-1, must be tested by experimentation. The theory, thus, comprises the following steps: 1—The irreversible entropy production is calculated to yield Eq. I-1. 2—The terms in Eq. I-1 are grouped

Jan 1, 1954

J. Chipman (Massachusetts Institute of Technology, Cambridge, Mass.)—The fact that the experimental work has been applied to copper rather than iron and that the paper is presented to the Iron and Steel Division, I regard as rather significant. It shows the unity of metallurgy and the fallacy of trying to cut it up by metals. This result for the solubility of sulphur in molten copper correlates with Professor Schuhmann's finding that the published data on the other side of the copper-sulphur miscibility gap are also in error. I should like to ask the author to say a little bit more about the sulphur capacity of the slag. T. Rosenqvist (author's reply)-—I hope that Dr. Chipman will find the derivation of the expression for sulphur capacity more clearly explained in the printed version of the paper than in the oral discussion at the meeting. I feel that this quantity, which actually is the ratio of two activities, can be measured more easily than the individual activities. Even if the ratio CaO/CaS is chosen as the standard state, the expression can be used for any slag, even for slags completely free of lime, and it represents a way to put the desulphurizing power of all slag constituents into one bag. Some doubt has been expressed as to whether oxygen ions really exist in calcium oxide and in molten slags. From a thermodynamic view point that question is of minor importance. The term oxygen ion activity, or any activity for that matter, is defined rigorously by the equation: activity = exp u/RT, where u is the change in free energy connected with the transfer of one mol of ions from the standard state into the slag. Whatever happens to the ion in the slag is of no concern to the thermodynamicist. Regardless of whether the ion is "free" in the slag or not, or whether it is present in a very small amount, its activity can always be expressed, and for a thermo- dynamic calculation that is all we need. However, ionic activities will only be of some real value if they are simple functions of the slag composition, or can be measured easily. Concerning the real nature of the oxygen in the slag, my feeling is that the oxygen atom has a rather multiplex nature depending on how strongly it is tied by covalent forces or polarized by the other atoms or ions present. The oxides of iron, cobalt, and nickel differ from calcium oxide and blast furnace slags as to the amount of free electrons that can give rise to electronic conductivity. In slags we know that the conductivity is mostly ionic. The fact that reversible emf's can be obtained with oxygen electrodes in certain salt melts, indicates a significant amount of oxygen ions in these melts. But extended work, e.g. polaragraphic studies and measurements of transference number; are needed to obtain quantitative information about the real structure of the slags. D. E. Babcock (Republic Steel Corp., Youngstown, Ohio)—-With reference to the ion, it might be well to remember Dr. Moses Gomberg. All of his life he had no use for the ionization theory and he contributed greatly to the field of chemistry on the assumption there was no such thing as ions. I do not think we have to worry about whether the oxygen is ionic or not. I think one thing specifically should be brought to your attention and this I think is one of the important contributions of Dr. Rosenqvist. He pointed out what we know as oxygen potentials or what is described as oxygen potentials. I have used this concept for a long period of time and I want to state that if this concept is properly applied, it vitiates much of what we have in the literature, or makes our usual ideas regarding oxidation seem primitive. That one thing is more valuable than almost all the rest of the discussion as a fundamental basis on which to build a reasonable in-

Jan 1, 1952

By A. U. Seybolt

Experiments on both suljidation and hot corrosion have been carried out using ternary Ni-Cr-X alloys and commercial nickel-base superalloys. It has been shown that lhere are certain micro structural similarities between sulfidation using H,S/H2 mixtures and hot corvosion using Na,S04 in "crucible" experiments or hot corrosion tests using a flame tunnel apparatus. It is proposed that two key processes in the hot cowosion process are: 1) formation of chromium sulfides which deplete the alloy matrix, and 2) oxidation which can be accelerated because of chromium depletion. Other processes play an important role in the hot corrosion process, such as accelerated oxidation attack due to the presence of a N& S04 or N&SO,- NaC2 film which inhibits the formation of a protective external oxide scale. However, this latter phenomenon was not studied in this work. THE term hot corrosion has come to mean a combination of sulfur and oxygen attack on nickel-base or cobalt-base superalloys used as gas turbine blades or partitions. This attack can be initiated by any combustion process in air involving fuel containing sulfur if the sulfur level is sufficiently high. Levels of 5 pet SO2l for example, in the combusting gas have been observed to produce hot corrosion attack on nickel-base alloys. The important immediate corroding medium in aircraft gas turbines, however, has been established to be Na2S04, formed from reaction between NaCl from sea air and sulfur in the fuel.' The Na2S04 concentrates the sulfur sufficiently so that superalloy parts heated in the range 1400" to 1900°F can become sulfided by reducing the Na2S04 film which forms on the parts, and this reaction allows sulfur to gain access to the alloy. The sulfur thus released diffuses into the superalloy and forms primarily chromium sulfides of various compositions, but frequently Cr2S, has been identified. Figs. 1 and 2 show the morphology typical of hot corrosion of a cobalt-base superalloy X-45 (25 Cr, 10 Ni, 7.5 W, 0.25 C) and a nickel-base alloy, Udimet 500 (19 Cr, 19 Co, 4 Mo, 3 Ti, 2.9 Al, 0.07 C). Both alloy samples were subjected to the same hot corrosion conditions: 616 hr at 1600°F in a combusted 2.9 pct S distillate oil containing 125 ppm Na as NaC1. It will be noted at once that the nickel-base alloy is much more severely attacked. The depth of corrosion is much greater and the number and size of the sulfides present are much greater. Both alloys show a duplex oxide scale, which is more noticeable in the X-45 alloy. The external scale is probably a spinel in both cases, with CrQ3 lying next to the alloy. Many investigators1 -"ve been concerned with hot corrosion phenomena, but there has been a lack of unanimity of opinion as to the details of the hot corrosion mechanism. More than one definition of hot corrosion is possible. Some investigators con- sider as a measure of hot corrosion only the weight gain due to an accelerated oxidation. However another point of view is that the total depth of affected alloy may be used as a measure. While it is recognized that surface oxidation attack is greatly accelerated by the presence of salts like Na2S04 or Na2S04-NaC1 mixtures, sulfur penetration into the alloy also undoubtedly produces very important effects. The region of sul-fide penetration which is so prominent in Fig. 2 is one of the outstanding characteristics of hot corrosion. For highly stressed parts, such as first-stage buckets, this sulfur-penetrated region can be more significant in mechanical property loss than the reduction of cross section produced by the formation of the massive external scale. In addition, evidence will be presented that this sulfided region can be less resistant to subsequent oxidation attack. Since it appears that the sulfided region below the surface oxide scale plays an important role in the overall corrosion process,

Jan 1, 1969

By C. E. Kemp, A. B. Dyes

Results of a field research project on the thermal recovery of oil by movement of a combustion front are presented. This field test was conducted in the South Belridge field, This The war paitern was a Southsing1e 2.5-acre five-spot pattern in the 700-ft deep Tulare sand. Results of this project indicate the technical feasibility of producing high viscosity, low-grallity crude oils by thermal means. It was demonstrated that heavy oils could be moved rapidly over considerable distances by thermal drive, and that high percentage recoveries of oil could be effected. The bulk of the oil recovery came after arrival of the burning front at a production well. However, rapid corrosion and high-temperature failure of conventional pipe and equipment both below and above surface occurred. Stainless steels gave more satisfectory performance under these conditions than did conventional steels. Oxygen was seldom observed in gas from wells ahead of the burning front, and it appears that the front was continuous throughout the life of the experiment. INTRODUCTION During late 1953 and early 1954, industry-wide attention was focused on the thermal recovery of oil by movement of a combustion front through an oil zone. Two separate Oklahoma field tests were described in publications by Kuhn and Koch, and Grant and Szasz. To obtain basic information needed for making engineering and economic appraisals of commercial recovery of heavy oil by this thermal recovery method, a new field research projecta3,4 was started May 31, 1955. Twelve oil companies contributed to this project with General Petroleum Corp. as operator. Since combustion recovery experiments* with heavy oils in other areas had been conducted in small patterns,"' an important purpose of this test was to determine whether the process could be operated in a pattern of near-commercial well spacing. The south Belridge field, located 30 miles north of Taft, Calif., was chosen because its oil sands are similar to many heavy oil sands which typically have low primary recovery and leave a large amount of oil in place. such fields are not particularly suited for common secondary recovery methods. DESCRIPTION OF TEST SITE The test interval was a Tulare sand at a depth of about 700 ft. Nine test wells were drilled, cored and completed from Oct., 1955, through Jan., 1956. The initial pattern shown on the map, Fig. 1, consisted of an injection well (II)', four producing wells (1P-4P), and four observation wells (IT-4T). An old well (81) was also recompleted for use as an observation well. The area enclosed by the original four production wells was 2.5 acres; the distance from the injection well to each production well was 233 ft. During the combustion period, Well 5P was drilled to replace Well 2P, and Well 6P was drilled to replace Well 3P. The new wells were 25 ft beyond the old production wells along a diagonal from the injection well, and resulted in an increase in pattern area to 2.75 acres. Isopachs are shown in Fig. 1. The average thickness of the oil sand in the test pattern was 30 ft. Fig. 2 presents two cross sections through the test pattern. Oil sand is represented by white and claystone by gray. The dashed lines indicate layers of conglomerate grading to claystone. The sand thickness was fairly uniform throughout the test pattern except in the 1P area. The sand body dips gently from the northwest to southeast, about 3 degrees from horizontal. The sand is unconsoli-dated, and grades from very fine material to pebbles as large as 2 in. in diameter. Table 1 presents the average initial properties of the formation. The in-place porosity of 37 per cent was estimated from the porosity of coke-consolidated cores taken after combustion. Original core analyses data were adjusted to 37 per cent, giving pore volume saturations of 60 per cent oil, 37 per cent water, and 3 per

By N. F. Schulz, H. A. Lex

Representative samples of typical manganiferous iron ores from the Cuyuna Range, Minn., were reduced to the metallic iron state at temperatures just short of fusion with an excess of coke in the manner prescribed by the R-N Direct Reduction Process. The magnetic separation characteristics of the reduced ores were determined at several grinds by means of the Davis magnetic tube concentrator. The efficiency of metallic iron concentration by magnetic separation was a direct function of the degree of metallization of the ore sample and of the degree of liberation at-tained in grinding. Some improvement of liberation characteristics was observed if a small amount of sodium chloride was present in the ore during reduction. Despite the extremely fine-grained structure of the crude ores, iron concentrates containing more than 90% Fe at iron unit recoveries of 90% or more were obtained from well-metallized samples of three of the four ores tested. The rejection of manganese to the tailings corresponding to such separations varied from 80% to 95%. The manganiferous iron ores of the Cuyuna Range in Minnesota constitute a major potential source of manganese in the United States provided that a technically and economically feasible process can be devised for its recovery. The iron and manganese of these ores occur chiefly as oxides or carbonates in the form of very small, irregular, discrete grains embedded in a matrix of quartz or other siliceous gangue. The present investigation was undertaken as a sequel to an earlier one conducted jointly by the U.S. Bureau of Mines and the R-N Corp. Those results indicated that the R-N Direct Reduction Process might be a means for recovering both a metallic iron concentrate and a marketable manganese by-product from Cuyuna manganiferous iron ores.' The R-N Process is one of many proposed schemes for the treatment of crude iron ores involving the di- rect reduction of the iron content to metallic iron followed by physical extraction of the iron. In general, it consists of the following stepsz2 1) Iron oxide in crushed natural ores or iron concentrates is about 95% reduced to metallic iron at elevated temperatures but without fusion in a rotary kiln in the presence of a solid carbonaceous fuel (usually coke), limestone, and combustible gases. 2) The kiln product, consisting of metallized ore, excess coke, lime, and coke ash is cooled. The excess coke is recovered and recycled to the kiln feed. 3) The metallized ore is ground for liberation and magnetically separated to yield a metallic iron concentrate. 4) The metallic iron concentrate is agglomerated by a briquetting process. The reduction kiln temperatures must be high enough for the desirable chemical reactions to proceed at practical rates without causing fusion of any of the ore constituents. The time of exposure of the ore to the reducing conditions must be ample for 90% to 95% metallization of the iron oxides present. If more than 95% of the iron is metallized, however, difficulties may arise in the briquetting step, presumably because of accompanying higher carbon content of the metallic iron. Although not always necessary, powdered limestone or dolomite may be added at some time during reduction to aid in eliminating sulfur and phosphorus from the final product.3,4 Other additives, notably sodium chloride, when present during reduction, result in improved liberation characteristics of the metallic iron from certain reduced ores.'v6 Under the reducing conditions normally employed, P(co + H2)= 1 atmosphere, and 1830' to 2150°F (1000° to 1170°C), the R-N Process is claimed to be adaptable to a wide variety of low, medium, and high grade iron and manganiferous iron ores. The manganese oxides and carbonates are reduced mostly to nonmagnetic MnO which is rejected to the nonmagnetic tailings. Manganese-bearing pyroxenes of the original ore, however, may be rejected poorly, if at all, apparently because these minerals occur within the iron lattice or are entrapped in the metal particles.7 The investigation was limited to studies of the Cuyuna ores listed in Table I, which were reduced to the metallic iron state under various conditions

Jan 1, 1969

By H. W. Winkler, Discussion, V. L. Forsyth

The development of a new gar lift valve has removed many of the obstacles limiting over-all gas lift efficiency. The valve is pressure charged in place in a well, and operating pressure can be changed without pulling. Temperature and gas gradient compensation have been eliminated. Intermitting and constant flow installations, using conventional pressure-charged valves, are compared with designs incorporating the new automatic stabilizer controlled valve. Specific well performance data are presented. The means of obtaining and controlling multiple-point injection of gas are explained and contrasted with single point injection. The effect on small diameter strings and multiple string installations is discussed. The effect of flowing temperature gradient on a design using conventional pressure-charged valves, the limitations imposed by such temperature, and the obvious benefits which result from the use of automatic stabilizer controlled valves are shown. Reduction of gas requirements is stated mathematically and demonstrated with specific examples which are identified as to oil company and well. Increased fluid recovery resulting from greater drawdown is illustrated. The economic advantages are mentioned with emphasis on reduction of capital expenditures, as well as reduction in operating expenses. INTRODUCTION Much of the engineering and research work performed with gas lift has been based upon the supposition that under producing conditions gas would be injected at a single point in a well. Many tests were performed in determining the optimum size of the opening through which gas would be passed and the exact placement of the valve equipment. The objectives in a gas lift installation are to unload the well and to efficiently produce the well after it is unloaded. Until recently, the function of upper valves was simply to unload the well. An operating valve located at a proper place in a well served as a single injection point for lifting liquid to the surface. The basic problem in using such equipment is to arrive at operating depth with sufficient operating pressure to efficiently lift the liquid. Few gas lift installations are designed with all required well data available. However, regardless of the quality and availability of data, and regardless of the accuracy of the design, there is the practical problem of preparing and placing gas lift equipment in the well in exact accordance with the plans and intentions of the engineer. In recognition of these and many other problems, attention was directed toward a means of removing the need for temperature and gas gradient compensation, a way to change the gas lift installation to meet the changing requirements of the well itself, and a manner of compensating for changes in oil-water content, fluid volumes, PI, and bottom-hole pressure resulting from flood or repres-suring activity. A device which could actually set and change the operating pressure of gas lift valves in the well would meet most of these needs. There now exists a new gas lift valve which fulfills the original objectives and in so doing actually accomplishes much more. With this device, called an Automatic Stabilizer Controlled valve, there is an exact matching of valve operating pressure and well conditions. Gas is injected at more than one point in the well, thereby making controlled multiple-point injection of gas a reality. Installation calculations are materially simplified and the need for basic temperature and gas gradient data no longer exists. Section I—The Automatic Stabilizer Controlled Valve The operation of the ASC valve is more easily understood by first studying a conventional pressure-charged valve and then inserting the stabilizer element to observe the changes which result. Fig. 1 is a schematic of a conventional precharged pressure-operated valve. The pressure in the dome pd, acts downward, against the area of the bellows A,, to hold the valve closed. The resulting force Bf plus the spring effect of the bellows Sf represents the total force acting to close the valve. Correspondingly, the tubing pressure ptbg works upward against the port area Atbg and the casing pressure pc works upward against a portion of the area of the bellows

Jan 1, 1965

By W. K. Godfrey

The highest compressive strength cement has the highest bond strength in tests in which the cement is subjected to a confining pressure. After perforating, the bond strength is reduced to nearly zero when the pipe is supported by weak cement. Perforating does not affect the bond strength, however, when the pipe is supported by strong cement. Pipe supported by weak cement is damaged by perforating with expendable capsule jets, but is not damaged by perforating with the hollow carrier gun. High strength cements are recommended for oil wells that are to be perforated. Introduction The purpose of this work is to determine the effect of jet perforating on the bonding strength of oilwell cement at simulated down-hole conditions. Bond strength is a good criterion for determining the effectiveness of cement in performing its main function, i.e., that of isolating zones. According to the literature on the subject, the optimum cement strength for perforating is low and should not exceed 2,000 psi compressive strength. This conclusion was based on the appearance of bullet and jet perforations in neat and gel cements largely done in the absence of a confining pressure more than 15 years ago. Discussion Literature A relatively weak cement has been recommended in the literature as having optimum properties for perforating. In 1947, Oliphant and Farris tested bullet perforating of neat cements at various curing times.' Some of the tests were made at down-hole conditions. They reported that the optimum cement strength for perforating was about 2,000 psi compressive strength. Enlarged holes occurred when strength was too low, and the cement shattered and cracked when the strength was too high. In 1951, Morgan and Dumbauld tested bullet and jet perforating of neat and gel cements at surface conditions.' They reported that optimum cement has a compressive strength less than 2,000 psi. In their tests at atmospheric pressure, high strength cement shattered and cracked when perforated. Thus, the optimum cement strength was based on the appearance of neat and gel cement after perforating. In 1953, Clark showed a relation between cement strength and fluid shutoff.' He concluded that a 500 psi compressive strength is required for good fluid shutoff and that maximum bonding ability is developed at approximately 1,000 psi compressive strength. In 1964, Bell and Shore invalidated most of the conclu sions by the previous authors when they demonstrated that atmospheric pressure tests cannot be used to determine casing deformation and damage that will result from perforating under down-hole conditions.' They showed that the higher the hydrostatic pressure, the more restricted will be the expansion of gases generated by the charge. Less casing deformation and damage occurs under down-hole conditions than at atmospheric pressure. In recent work, Bearden et al. reported that jet perforating at surface pressure with the hollow carrier gun caused the cement sheath to shatter and crack.' The perforating, however, did not seriously affect the pressure required to leak past the casing cement bond. Only three cementing mixtures were tested, and presumably they had strengths in excess of 2.000 psi. Bearden et al. also reported that the hydraulic bonding strength increased as the compressive strength increased. Their hydraulic test showed that approximately 1,000 psi pressure is sufficient to break the casing cement bond for a well aged neat Portland cement sheath. Unfortunately, the tests were not run with a confining pressure, so Bearden et al. were merely measuring the strength of their outer 8%-in. casing. In our investigation, we measured the hydraulic bond strength of cement to pipe under a confining pressure and found that the bond strength was not limited to 1,000 psi. Cement Bond The bond strength is a good criterion for measuring the effectiveness of cement in performing its main function, that of preventing interzonal communication of fluids behind the casing. Measuring bond strength after perforating should be a good quantitative method of determining optimum cement properties, and should be a better method than that used by previous investigators,11 ' who simply observed the appearance of the cement after perforating.

Jan 1, 1969