Search Documents

By C. R. Johnson, R. A. Greenkorn, E. G. Woods

A new method of reservoir evaluation called pulse-testins has been developed for describing formation properties between wells. Pulse-testing utilizes a sensitive differential-pressure gauge at a responding well to measure and record the response generated by a series of flow rate changes (pulses) at an adjacent or pulsing well. Since the pulse-test instruments have a sensitivity of about 0.001 psi, pulses of several hours or less in duration will generate a measurable response in most reservoirs. For this reason, many well pairs can be tested in a short period of time with little interference in field operations. Comparison of pulse-test results to conventional testing methods shows that the pulses obey unsteady-state, compressible-flow theory and thus provide a measure of both transmissibility (kh/µ) and storage (Fch). In additian, the method can be used qualitatively to describe communication across faults and between zones, and direction and tnagnitude of fracture trends. INTRODUCTION To obtain maximum usefulness of new reservoir analysis methods as well as to aid the field engineer in understanding local reservoir anomalies, improved methods are needed to describe heterogeneities in specific reservoirs. Optimum well spacing, equipment specifications, well completion design and especially the economics of pressure maintenance and secondary recovery programs depend on the extent and location of the heterogeneities that affect flow behavior in the field. One method for obtaining a measure of areal reservoir heterogeneity is an interference test. In the simplest interference test. constant-rate production or injection is initiated at one well and the effect of this flow is measured as pressure vs time at another well. Thus, during interference tests the flow characteristics of the formation are determined in situ. Despite their obvious usefulness in reservoir delineation. interference tests have not enjoyed frequent usage. Reasons for this have been the weeks or months required in many fields to obtain a pressure response measurable with conventional gauges, and the interruption of routine field operation during the field-wide shut-in normally required during these long times. Also, only average properties of the relatively large volume of the reservoir disturbed during these long times can normally be obtained — variations in properties between individual wells cannot be delineated. To eliminate these drawbacks of conventional tests, a special reservoir description technique called pulse testing' has been developed. Pulse-testing has several advantages that make it more convenient and practical for field use than interference tests. Valid tests are more certain because use of a series of flow disturbances gives rise to a diagnostic pressure response that can more readily be distinguished from unknown trends in reservoir pressure and other "noise." The time required is only a fraction of that for an interference test because a special, high-sensitivity differential pressure gauge can detect a much smaller pressure change. Variations in properties between individual wells can be delineated because of the relatively small volume of the reservoir disturbed during these short times. Finally, wells other than the pulsing and responding wells do not have to be shut in or carefully regulated.

Jan 1, 1967

By P. A. Beck, S. Kohara, M. N. Parthasarathi

IT has been shown1,3 that many observed phenomena related to annealing textures in f.c.c. metals can be adequately accounted for, at least in a qualitative way, by considering the orientation dependence of the boundary mobility, without any hypotheses concerning preferred orientation of nuclei. It was noted1 that this approach is particularly successful in the common cases of primary recrystallization in highly cold-worked polycrys-talline metals, where nuclei occur in great abundance, and that in the other extreme case, where the number of nuclei is very limited, or even only a single crystal is grown, the orientation dependence of boundary mobility has only a small effect, or none at all, on the orientations resulting from annealing. Recently Dunn described examples3-6 where observed secondary recrystallization textures apparently cannot be accounted for on the basis of oriented growth alone, but also require oriented nuclei, or at least the absence of nuclei in certain orientation ranges. In general, the great difficulty in correlating the preferred orientation of recrystallized grains with that of nuclei is inherent in the fact that nuclei are by nature very small, and that they are usually identifiable as "nuclei" only after growth, i.e., after they have already been subject to growth selectivity. An investigation was carried out by Liebmann and Lücke,' and also a very similar one by Graham and Cahn,' to obtain quantitative information on the orientation depend- ence of boundary migration rates in the recrystallization of slightly deformed aluminum single-crystal rods or plates, in which artificially nucleated recrystallized grains of predetermined orientation were grown. Liebmann and Lücke foupd maximum boundary mobility for boundaries across which the misorientation corresponded to a rotation of 40 deg around a common [Ill] axis, in accordance with earlier conclusions.9,1 Also, they confirmed that such an orientation relationship predominates among crystals growing into slightly extended single crystals from presumably random artificial nuclei. On the other hand, Graham and Cahn found no preferred orientation among crystals grown in a similar manner. It is not as yet known how these observations may be reconciled with each other. The view has been expressed in various papers5'10,21 that selective growth based on the orientation dependence of boundary mobility is not in itself sufficient to account for the formation of sharp re-crystallization textures, such as those often observed in highly deformed polycrystalline metals after annealing. The present work was undertaken to ascertain whether or not it is possible to obtain a sharp re-crystallization texture in a case where random nuclei are assuredly present in large numbers. It was previously found1' that high-purity aluminum crystals rolled 80 pct on the (110) plane in the [1T2] direction, when suitable precautions are taken, may be annealed for periods of about 1000 sec at 350°C without recrystallization. It was also found that the reluctance to recrystallize in this case is due to the scarcity of material in sufficiently deviating orientations to serve as nuclei. By using such a matrix,

Jan 1, 1959

By Josiah Royce

To evaluate chemically the sample of rock obtained by diamond drilling, it has long been recognized that the analyses of the two components of the sample, core and sludge, must be given appropriate influence in computing the average analysis of any unit of depth. The purpose of this investigation is to set forth what means are available for apportioning the effect of core and sludge on the final analysis, what variables affect the problem, and what combination of applied mathematics will closest approximate the truth under each condition as these variables proceed within their limitations. A drill hole is bored in iron ore exploration principally to test variations in rock composition with depth and is usually directed as nearly normal to the bedding of a horizon to be tested as possible. This practice has a tendency to minimize variation in composition laterally which in any event is not likely to be great. It is obvious that the opportunity for change in analysis of a particular rock is not statistically as great radially in a diamond drill hole where the distance in which such a change may occur is from 0.719 in. (EX bit) to 1.469 in. (NX bit) as there would be longitudinally even in a' run as short as 5 ft. Variations in composition of bedded or layered rocks are usually greater normal to the bedding than parallel thereto. Even in massive rocks, like porphyries, variations are functions of distance. Hence. in either case, variations along the hole are of greater effect than across it. Let us examine Fig 1 briefly to observe a cross section of a unit of depth of a typical diamond drill hole. If we neglect radial change in chemical composition, which we have seen is small compared to lengthwise variation, we can see that if core recovery was 100 pct, the core and its surrounding area, which would be recovered as cuttings (assuming 100 pct sludge recovery), must analyze the same if the sampling is perfect. This is the foundation of this paper, namely, that if core recovery is complete we may assume that the core constitutes as nearly as possible a perfect sample of the ground drilled. Now let us pass to the methods of weighting core and sludge in the final analysis and then develop each in turn. One might employ core analyses only, sludge analyses only, or apportion the influence of each by one of several methods. The most common apportionment is by direct proportion to relative theoretical vol- ' ume occupied by each component in the cylinder hollowed out by the drill, known as the Longyear formula. Another logical treatment would be to consider core and sludge in the final analysis according to the weight of each recovered. This procedure was elaborately set forth and refined by R. S. Moeh1man.l Then to be sure that we have embraced all the possibilities, we must admit that some other method or methods may be devised empirically or by mathematical maneuvers. Should a drill machine be capable of recovering 100 pct of the core in all types of rock, certainly core alone need be analyzed and no complicated mathematics are required; but although some companies have had fair success in obtaining high percentages of core in certain homogeneous ores on the Mar-quette Range, their same methods have proved disappointing on the Menominee, Cuyuna, and Gogebic Ranges where the iron formation often varies widely from flinty chert beds interstratified with soft hematite to wholly leached ore material composed of an unpredictable mixture of hematite and limonite with occasional zones of sugary, recrystallized chert. In spite of numerous mechanical improvements in core barrels, few drilling programs can count on complete core recovery. The sludge alone might be analyzed but for practical reasons this is inadvisable when core is obtained due to the numerous opportunities for contaminating or losing portions of the cuttings. The Longyear formula involves a weighting of the core and sludge in proportion to the theoretical volume of each, that is, where the volume of the core recovered is one-third of the total cubic volume of the cylindrical hole made by the drill, the analysis of the core is given one-third of the weight in the final analysis. This approach to

Jan 1, 1950

By T. F. Mueller, P. A. Witherspoon

For the case of an infinite radial system operating at constant terminal rate, the reservoir engineer often uses the "point source" solution of the diffiusivity equation to study pressure interference effects. At early times and at short distances from the inner boundary these solutions are invalid. The amount of error is often not precisely defined because of mathematical difficulties. Work presented here shows that, if dimensionless time is defined appropriately, previous solutions of the pressure equation can be displayed .as a family of curves on one chart. These solutions include the point source solution (referred to in the field of hydrology as the Theis solution) and other solutions obtained with digital computer methods. With these curves, an exact evaluation of the pressure drop within a reservoir or an aquifer can be made by the engineer. Examples of field problem solutions are presented. In most reservoirs the error involved when the Theis solution is employed is often negligible; whereas, in the calculation of interference effects in an aquifer, a substantial error can occur through such an approach. INTRODUCTION Flow equations are used in petroleum engineering to study the behavior of individual wells and reservoirs. In the case of wells, the pressure response at the wellbore face is the major point of interest; whereas, in the case of reservoirs, the pressure response at the interface of the aquifer boundary is sought. To aid in such studies, the flow equations have been solved in terms of the behavior at these two inner boundaries. Only limited work has been published in regard to the pressure conditions away from these points, i.e., within the reservoir or aquifer. Theis and Mortada are among the few who have reported on this problem. The Theis approach employs the exponential integral and is valid for pressure conditions that occur some distance away from the flow disturbance. It is derived from the concept of a point source, as opposed to a flow across some finite area. The Mortada results, on the other hand, are valid at all points within the reservoir or aquifer. They are presented in terms of dimension less ratios of the radius where the pressure is desired to the radius where the flow rate is measured. Their main use, in the past, has been in aquifer studies. The published results are presented in the form of graphs that are limited to a maximum radius ratio of 64. These graphical results are cumbersome to interpolate at non-integral radius ratios, so that one may be forced to utilize a rather involved analytical expression presented by Mortada. BASIC EQUATIONS The solutions of Mortada and Theis are both based on the diffusivity equation as applied to the case of an infinite radial system subject to a constant terminal rate. The equation is obtained by combining the material balance equation with Darcy's flow equation. The assumptions implicit in the use of this equation are as follows: (1) a single fluid is present that occupies the entire pore volume; (2) the reservoir is horizontal, homogeneous, uniform in thickness, and of infinite radial extent; (3) compressibility and viscosity of the fluid remain constant at all pressures; and (4) fluid density obeys the equation Using the diffusivity equation in situations where the above conditions do not hold will result in errors. These errors (not discussed here) but only the errors which arise in the solution of the equation itself. The diffusivity equation for the homogeneous reservoir conditions cited above can be written in cylindrical coordinates, as To obtain a dimensionless equation, so a single solution can be used for applications of different porosity, permeability and fluid properties, the following transformations are usually made: After these transformations are made, Eq. 3 .can be written in dimensionless form, as MORTADA SOLUTION One solution of Eq. 6 has been given by Mortada.

Jan 1, 1966

By I. F. Jackson, H. L. Hartman

Hard-metal inserts have received considerable attention in the mining industry recently for a variety of excavating and penetrating purposes. Their use in percussive drilling and in coal cutting machines is well known, but attempts to utilize them in cutting materials such as dimension stone have been discouraging because of excessive tool wear and failure. Particular attention is given to those properties of hard-metal inserts that affect their resistance to wear. A consideration of the actual wear mechanism shows the importance of correct mounting of the inserts and of the selection of suitably powered driving motor so as to achieve minimum tool wear and maximum tool life. Comparisons are made between the performances of types of hard-metal inserts, relating their inherent properties with results obtained in laboratory tests. As part of a broad program to improve the quarrying and processing of industrial minerals, The Pennsylvania State University has been investigating the machinability of slate under laboratory and field conditions. It should be emphasized that the results of the research presented in this paper are also applicable to other natural structural materials, and particularly to the more workable ones such as limestone and sandstone. Improvements which accrue in cutting and sawing should prove of benefit to all the dimension stone industry. Further, it may be pointed out that the results reported here are applicable to rotary drilling (with drag bits) as well as to cutting and sawing of rock. INVESTIGATIONS IN RELATED FIELDS The investigation was initiated with an extensive survey of available literature related to metal cutting and to rotary drilling. The following similar features between metal cutting and slate cutting became evident: 1) Hard-metal inserts are widely used as a cutting tool in both cases. 2) The forces acting on the tool may be resolved into normal and tangential components. 3) Rate of penetration and tool wear are important criteria in each case. On the other hand, important differences also exist: 1) Most metals, exhibit plastic behavior, while slate (and most rock) is a brittle material under normal conditions. 2) In metal cutting the component of force acting parallel to the surface of the workpiece is considerably greater than the force holding the tool against the workpiece. 111 slate cutting, this situation is reversed. Similarities between drag-bit rotary drilling and slate cutting are as follows: 1) In both casts, a tool, usually of the hard-metal insert type, is used to penetrate a brittle material. 2) The drilling forces and related variables, such as rotation speed, are similar. 3) Chip formation in each case is essentially the same. 4) The shape of the tool is similar in rotary drilling and slate cutting. Important differences are also apparent: 1) The objective of drilling is to produce a directed hole, while in cutting it is to form a kerf. In core drilling, however, a kerf is also produced. 2) The horizonta1 component of force in rotary drilling decreases from the center to the outside of the tool, whereas in slate cutting the component is almost uniformly distributed across the width of the tool. 3) The power absorbed in the drill rod is greater in rotary drilling. 4) The clogging action of chips, if excessive power is applied to the tool and if flushing is insufficient, is more serious in rotary drilling. Because of the striking similarities between metal cutting and slate cutting, and between rotary drilling and slate cutting, it was considered that the findings of investigators in those fields would serve as a guide for experimental studies of slate cutting. The following is thus devoted to a discussion of those aspects of metal cutting and rotary drilling pertinent to the present investigations.

Jan 1, 1962

By C. H. Brehaut, W. N. Taylor, R. S. Douglas, H. A. Shannon

THE name for this new method of mining is derived from a composition of Horizontal, Radial, Diamond, and the drilling is from raises. This method, worked out at Copper Mountain, B.C., is believed to be distinct in many features from other known methods. GEOLOGY The ore deposits of Copper Mountain are related to the Copper Mountain stock, which is an ideal example of the magmatic differentiation in that it grades from a basic gabbro at the border to an almost pure orthoclase-pegmatite core. The "contact" ore bodies lie along the contact of the gabbro and an older complex of volcanics and sediment termed the "Wolfe Creek" formation. These older rocks, collectively referred to as greenstone, had been fractured prior to the intrusion of the stock and were highly metamorphosed during the intrusion. The "outlying" ore bodies are somewhat different in character, though a similar structural pattern prevails. It is believed that a deep-seated magma of which the Copper Mountain stock is a part has given off solutions that caused a partial granitization of some of the volcanics. Mining problems are caused further by the presence of a series of soft barren dikes intersecting a great many of the ore bodies, which are about 30 ft. wide, strike north-south, and dip at a high angle to the east; and are further increased by the presence of four major faults, with a great many minor faults running at acute angles from them. Fault zones average 8 ft. in width, and consist of gouge and breccia. They strike northeast and dip steeply to the northwest. The faults cut through the ore zones and the dikes, and continue on into the gabbro. The general mine structure with two types of ore bodies, bounded by weak walls and intersected by several series of soft barren dikes and major faults, indicated that a method of mining should be devised that would remove the ore at low cost with the least disturbance of the waste walls and dikes. THEORETICAL ESTIMATE OF STOPING WITH THE DIAMOND DRILL After consideration of various methods, the most suitable one appeared to be horizontal drilling to break the ore in a manner similar to shrinkage stoping. Drilling vertical holes or rings from a bench and breaking to an open slot was eliminated because the weak walls of ore, dike or waste would not permit this type of opening. It was then decided to drill either horizontal radial holes from a raise or parallel holes from a slot cut across the ore body. After discussion, the latter method was believed unsuitable, for the following reasons:

Jan 1, 1945

By Craig S. Hartley

Expressions are obtained for the energy changes associated with the reaction of (a& (111) slip dislocations on intersecting (110)planes in anisotropic bcc metals. An energy criterion for assessing the likelihood of dissociation of the products of such reactions is also presented. It is found that the "burrier reactions" which form a(100) dislocations at the intersection of two active {110) slip planes are more energetically favorable in metals which exhibit a high value of Zener's anisotropy factor, A, than those which have a low value. The results are presented in a form which permits the stacking fault energy to be obtained from a measurement of the separation between par-tials in a dissociated configuration. However, until accurate calculations or measurements of the stacking fault energies involved are available, it is not possible to assess the physical importance of dissociated dislocations. In a recent paper,' the energy changes associated with several types of reactions between two slip dislocations, (a/2)(111){110), in bcc structures were calculated.* Isotropic elasticity and the approxima- tion v = -3- were employed. The purpose of this work is to present calculations of the energy changes for many of the same reactions using anisotropic elasticity. The problem of dissociation of a(100) and a(110) dislocations is also considered, and maximum fault energies for which dissociation will be energetically favorable are calculated for several bcc metals. Two general types of reactions are considered; those for which the reactant (a/2)(111) dislocations have long-range attractive forces and those for which the reverse is true. An example of the former is: (a/2)[lll] + (a/2)[lll]-a[l00] while the latter are typified by: (a/2)[lll] + (a/2)[111] -a[011] Only reactants lying in different slip planes are considered; therefore, the products must lie along (111) or (100) directions, which are the intersection of two {llO} planes. It will be assumed that the reactants and products are infinitely long parallel dislocations, since in this case the energy change associated with the reactions is a maximum.' THEORY The self-energy per unit length of a straight mixed dislocation in an anisotropic medium can be written? where b is the Burgers vector, K is an appropriate combination of the single-crystal elastic constants, and R and ro are, respectively, outer and inner cut-off radii of the elastic solution. The energy given by Eq. [I] does not account for any variation of the core energy with orientation. This could be manifested by an orientation dependence of the core radius or, equivalently, the Peierls width, of the dislocation. However, the energy contribution due to this source is expected to be small, and current models of the dislocation core are not sufficiently accurate to justify such a refinement. It has already been shown that for the isotropic case the energy contributions due to nonzero tractions across the cores of the reactants and products exactly cancel one another in the reaction.' Accordingly, it will be assumed that this contribution to the total energy change in the anisotropic case is small. In the subsequent discussion it is also assumed that the core radii of the reactant and product dislocation are the same and that, where stacking faults are formed, the faulted region is bounded by the centers of the partials. Consequently only changes in elastic energy due to the reactions will be considered. When the dislocation is parallel to either the (111) or the (100) directions, K may be written:375 K = (Ke sin2 a + Ks cos2 a) [2] where K, and Ks are the combination of elastic constants corresponding to an edge and screw dislocation lying along the same direction as the mixed dislocation, and a is the angle between the direction tangent to the dislocation line and the Burgers vector. Eq. [2] should not be confused with the isotropic approximation to the variation in energy with line Orientation.6 It should be noted that the essentially isotropic expression for K is a result of the characteristic symmetry of the (111) and (100) directions and is not, in general, valid for other dislocation directions in anisotropic cubic metals. The energy* change for a reaction in which the re- actant and product dislocations are parallel perfect dislocations can be written: where Ep and E, refer to the self-energies of the products and reactants, respectively. For dislocations parallel to (100) and (111) directions, Eq. [3] becomes:

Jan 1, 1969

By Jr. F. M. Perkins.

Capillary forces play a controlling role in water-drive displacement processes both in laboratory experiments and in actual reservoirs, but their quantitative importance may be quite different in the two cases. Because of the importance of conducting laboratory experiments which are representative of field conditions, it is necessary to understand exactly the role of capillary forces in the displacement process. Though a number of experimental investigations related to this subject are contained in the literature, there appears to be a lack of information pertaining to unsteady-state experiments in water-wet media. This experimental study was conducted to obtain additional laboratory data to clarify further the role of capillary forces in both the macroscopic and microscopic flow of oil and water in porous materials. THEORETICAL CONSlDERATlONS The capillary pressure is defined as the difference in pressure between a continuous oil phase and a continuous water phase in a porous material.' The magnitude of this pressure difference depends on the interfacial curvature and the interfacial tension. The interfacial curvature is determined by the geometry of the pore spaces, the wettability of the rock surfaces, and the quantity of cach phase present. Capillary forces are involved in a water-drive displacement process in that they exert a controlling influence on the microscopic fluid distribution which in turn is reflected in the saturation or macroscopic flow behavior. Microscopic Fluid Distribution Because of the microscopic nature of the displacement of oil by water, it is necessary to consider the flow and the fluid distribution in individual pores. On this microscopic scale the capillary Forces, which act over a distance of one or two sand grain diameters, control the distribution of oil and water under static equilibrium conditions. When an external force is applied to the fluids, such as in a water-injection experiment, the applied forces tend to distort the oil-water interfaces. However, in most fine-grained, water-wet sands, the applied pressure difference across one or two grain diameters is usually several orders of magnitude less than the capillary pressure difference. These con-siderations lead to the theory' that even during flow the capillary forces continue to control the microscopic distribution of oil and water within the pores of a porous material for all practical reservoir and laboratory flow rates. This concept of capillary forces controlling the microscopic distribution of fluids has been substantially verified by other investigators3-7 who have found a lack of dependence of relative permeability and residual oil saturation on rate of fluid injection. Macroscopic Distribution The microscopic influence of capillary forces cannot be observed easily and only the effect on the macroscopic or average saturation can be detected. The saturation, of course, is really the point of interest. During a water flood, large differences in saturation at the flood front cause large capillary pressure gradients. This, in turn, causes water to advance ahead of the flood front thereby reducing the capillary pressure gradient in this region. The result is that in homogeneous porous media capillary pressure gradients tend to cause a diffuse displacement front. At low rates in laboratory columns, the front may extend over the entire column length. When the advancing water first reaches the outflow face of a core, the water, which is the wetting phase, cannot be produced because the pressure in the water just inside the core is lower than the pressure in the oil-filled space around the outflow face. This difference in pressure is equal to the capillary pressure for the water saturation existing at the outflow face. Water, therefore, accumulates at the outflow end of the core which causes a reduction in the capillary pressure. Because the capillary pressure does not vanish except at the residual oil saturation,7,8,9 water will not be produced until the residual oil saturation exists at the outflow face. This entire effect,"' which is called the "boundary effect", results in a region of relatively high water saturation near the outflow face. At low rates of injection in a short column, this region of high water saturation may extend over a considerable portion of the column. The influence of capillary forces on the macroscopic flow of oil and water have been described by Leverett.10 For unidirectional, viscous flow in the absence of gravity segregation, the expression in dimensionless form for the fraction of water in the flowing stream, f, is

Jan 1, 1958

By R. E. Hudrlik, G. W. Preckshot

The diffusion coefficients of silver from solid silver in molten lead were measured to within ± 0.8 pet in a columnar type diffusion cell ower, the temperature range of 326° to 530°C. Fick's law describes the process up to 530°C where the laminar mechanism appareltly breaks down. These is negligible resistance at the interface as shown by mathematical analyses. The diffusion coefficients are found concentration independent. IT would seem that diffusion in liquid metals would be free of such effects as molecular structure, dissociation. polarization. and compound formation. This view was taken by Gorman and preckshot in their study of diffusion of copper from solid copper into molten lead. They reported diffusion coefficients which were independent of the concentration over the range of 478° to 750°C. They found that the Stokes-Einstein equation with constant radius of the diffusing specie represented the diffusion data better than Eyring's rate theory equation and Sheibel's correlation. The radius of diffusion was found to be that of the doubly charged copper. There appeared to be no resistance across the solid-liquid boundary. In the present work the diffusion coefficients for silver in liquid lead were measured over a range of temperatures of 350° to 505°C. The solubility of silver in lead over the range of 303° to 630°C was also obtained. These results are compared with calculated or correlated values or with data in the literature. EXPERIMENTAL Procedure—The experimental equipment techniques and procedures were those reported in detail by Gorman and preckshot9 and will not be repeated here. Measured values of WT, Co, A. L were obtained for various diffusion times and the diffusion coefficient was computed for the case of no resistance at the interface9, 11 by: WT/CoAL = 1- 8/p2 n=1 1/(2n - 1)2 exp[-(2n - 1)2p2 Dt/4L2] [1] or where there was resistance at the interface by: WT = 1- ?n=1 2h2/ap2L [sxp [-Dan2t]/[(h2 + an2) L + h] The roots an are those of the transcendental equation3 tan (an L) = Iz/cun. The diffusion coefficient is that defined by Hartley and Crank.7 The total silver in the lead cylinder and equilibrium slug was determined by a cupellation technique' with proper correction for losses. Analysis of known samples showed that this method is surprisingly accurate. The amount of silver in the lead adhering to the silver cylinder was obtained in the same fashion as shown by Gorman and preckshot.9 The small errors involved in this determination are not critical since the silver in this adhering lead layer is only 3 to 15 pet of the total diffused. Materials—Electrolytic silver containing 99.9+ pet Ag obtained from General Refineries of Minneapolis, Minn. was used for all but runs 7 and 8. For the balance of the runs this silver was reduced with hydrogen at 1100°C and its oxygen content was found to be about 0.017 pet. For the runs. 7 and 8, phosphorous-reduced silver of the same purity was obtained from Handy and Harman Co. of Chicago, Ill. The densities of the phosphorus-reduced silver and the hydrogen-reduced electrolytic silver were 10.484 and 10.487 g per cm3, respectively. These values agree with those reported for pure silver. Silver which was reduced at 900 C had an average density of 9.998 g per cm3, indicating porosity. This silver was used for a number of runs which were not tabulated in Table I. These are indicated by crosses on Fig. 2. The 99.999 pet Pb was obtained from the National Lead Co. Research Laboratory of Brooklyn, New York. DISCUSSION OF RESULTS The diffusion and solubility results are reported in Table I for eleven runs using either phosphorus-reduced electrolytic silver or hydrogen-reduced silver at 1100° C. The solubility data shown in Fig. 1 show the excellent agreement with that reported by Heycock and Neville.8 The data of Friedrichs5 apparently are in error. The experimental solubility data of this work are reported to 0.3 pet. The experimental diffusion coefficients computed from Eq. [1] are reported within 1.2 pet of the mean and are plotted in Fig. 2. These are expressed within +0.8 pet of the experimental values over the entire temperature range by: D= 8.26 x 10 -5 e-1925/RT . [3] There appears to be little difference due to the

Jan 1, 1961

By Morris Muskat, James M. McDowell

A report is given of electrical analog experiments on the effect of casing perforation completions on well productivity for ideal uniform reservoirs under steady state homogeneous fluid flow conditions. Graphical results are presented for the effect of perforations of different densities and various degrees of penetration into the formation surrounding the casing or cement sheath. The latter range from perforations which terminate flush with the casing to those which extend into the formation a distance equal to the casing diameter. For the special case where the perforations terminate immediately beyond the casing with semi-spherical protrusions the electrical analog results correct and supersede previously reported theoretical calculations. It is found that the penetration of the perforations into the surrounding productive section may so increase the resultant productivity as to approach or even exceed that for open hole completions. INTRODUCTION A theoretical analysis of the effect of casing perforation on well completion, without additional penetration of the formation, was reported several years ago.' In this treatment the perforations were represented by mathematical sinks and no explicit account was taken of the imperviousness to flow of the casing in which the perforations were made. This procedure was justified by consideration of the mutual interference between the sinks, which seemed to indicate that the analytical method automatically eliminated flow across the casing radius between the perforations. However, both further theoretical investigation and the electrical analog measurements described below showed that this assumption in the previously developed theory was not valid and that the calculated well capacities would considerably exceed the correct values. § Even if this simplifying assumption were retained the generalization of the theory to situations where the perforations extend into the producing formation would be extremely difficult. On the other hand, the electrolytic model analog of this type of steady-state homogenous fluid system is, in principle, so simple that it has not been worthwhile to attempt to develop the exact analysis even for the more limited case of strict casing perforation. The whole problem has therefore been studied anew by electrical models to cover the range from holes in the casing which terminate flush with the casing surface to such as extend into the surrounding rock to a depth equal to the casing diameter. It is to be expected that the productivity of a well completed with casing radius* rw, perforated with holes of radius Tp, and extending to depths d into the surrounding rock, would depend primarily on the dimensionless geometrical parameters Te Tw TP d of the system, such as —, —, — , and —, where re is the Tw, a a rw external boundary radius and a is the mean vertical separation between the perforations. It will also depend, of course, on the number of perforation lines or total perforation density. While such parameters automatically control analytical treatments, it would require a very extensive series of model experiments to fully disclose their functional effect on the well productivity. For practical purposes, however, a rather limited range of the variables suffices to give the effects of interest. Accordingly, in the experimental study reported here the casing diameter has been kept fixed? at six in., the perforation diameter has been taken as either one-fourth in. or one-half in., their density has been varied from one to eight per ft, and their penetration into the formation has ranged from zero to six in., with a primary scaling factor of one to six. EXPERIMENTAL METHOD AND RESEARCH Although electrolytic model experiments have been made for years, there still seems to be some question as to what set of conditions will give the best results. For this reason a number of electrode materials and electrolytes were tried. Those chosen for the final experiments were copper electrodes and a weak copper sulphate solution as the electrolyte. This 'Assuming that the casing has been cemented with a uniform and impermeable cement sheath, the effective casing radius is to be interpreted as the external radius of the cement annulus. It is also assumed here that the casing is set through the complete producing section that it is perforated throughout with a uniform density and that tie formation has a a and isotropic permeability. As noted later, however, a uniform scale change permits reinterpreting the primary measurements to correspond to a casing diameter of 12 in.

Jan 1, 1950

By R. Taggart, D. H. Polonis, W. C. Gallaugher

In the Ti-Cu system, the a' phase can be produced over a wide range of alloy composition witJwut the retention of measurable amounts of the ß or ? phases. This paper reports on the decomposition of this hexagonal martensite phase front the standpoint of solute corzcentratiorz, tempering temperature, and coherency strain condition. The mechanism of precipitation during tempering comprises localized precipitation which is followed by discontinuous precipitation if sufficient coherency strains are present. The localized precipitation process in hypoeutectoid alloys is described by the generalized Johnson-hiehl equation and an activation energy of 46,000 cal per mole. In hypereutectoid alloys the corresponcEing activation energy is 51,000 cal per mole. The rate-controlling process has been proposed as the difision of copper along a' platelet interfnces. SEVERAL investigations have been reported concerning the kinetics of decomposition of the mar-tensitic a' phase in titanium binary alloys.' In the Ti-Mo system,' two modes of a' decomposition have been observed in the presence of /3 phase; one of these involves the precipitation of fine a from a', and the other involves diffusion across the a'-ß interface. In studies of the Ti-Cr system, Rostoker2 did not detect the formation of TiCr2 during tempering. In Ti-Ni alloys,3 the intermediate phase Ti2Ni has been found to precipitate directly from a' along the interfaces between plates. With the exception of the study of Ti-Ni alloys the previous investigations of tempering phenomena in substitutional martensites are mainly qualitative and do not present a detailed description of the precipitation processes. The following limitations restrict the correlation of the microstruc-tural processes and reaction kinetics during tempering in most binary-alloy systems. 1) A mixture of at least two phases characterizes the constitution of quenched alloys. 2) Difficulties have been encountered in obtaining uniform structures throughout quenched samples. 3) The reaction products, and in particular their morphology. have not been clearly resolved. 4) The martensitic a' phase forms over only a limited composition range for most titanium-base binary alloys. Gomez and polonis4 showed that the Ti-Cu system provides an excellent basis for investigating the tempering of a' over a range of composition without the complication of retained P or the transition w phase. In the present work the effects of solute concentration, tempering temperature, and coherency strain conditions are considered with reference to the over-all precipitation process in quenched alloys of the Ti-Cu system. Both microstructural observations and kinetic data are correlated to define the rate-controlling processes which govern the observed localized and discontinuous precipitation reactions. An earlier paper5 was devoted to a discussion of the modes of heterogeneous nucleation of Ti2Cu from the martensitic a' structure. The present paper will therefore emphasize the progress of the tempering reaction beyond the initial stages. EXPERIMENTAL METHODS The Ti-Cu alloys for this study were prepared by conventional arc-melting procedures. Chemical analysis revealed that the alloy buttons contained approximately 0.02 wt pct 0, 0.04 wt pct N, and within 0.1 wt pct of the intended copper concentration. The progress of the tempering reaction was followed by means of electrical-resistance measurements utilizing a Kelvin double bridge, microhard-ness readings with a 400-g load, and X-ray dif-fractometry patterns to reveal line-broadening effects. Metallographic specimens examined at magnifications greater than X1000 were shadowed with germanium to reveal fine structural details. Direct carbon replicas were prepared for the electron-microscopy studies. EXPERIMENTAL RESULTS Property Changes. The changes of microhardness that accompany the precipitation of Ti2Cu from a' are shown in Figs. 1 and 2. As the solute concentration increases the peak hardness, for a given tempering temperature, increases. In alloys of given composition the time to reach the peak hardness agreed with the time to attain maximum X-ray diffraction peak breadth, as shown in Fig. 3. The maximum hardness and the maximum line breadth increased with lower tempering temperatures, and the time to reach the maximum also increased. The a' peaks broadened during the initial stages of tem-

Jan 1, 1965

By W. E. Newell

The basic principles and assumptions involved in D. L. White's solution5 for ultrasonic wave amplification in piezoelectric semiconductors are summarized. If the gain per unit length is maximized at each frequency without regard to the drift-field power density, the gain increases linearly with frequency. Therefore it would seem that very high gain is possible for frequencies in the hundreds of megacycles. However, with present materials the resulting drift-field power density is so large that cooling would be a major problem. For this reason it is more realistic to judge the capabilities of an ultrasonic amplifier in terms of the gain that is possible with a specified power density. In this case the gain at high frequencies does not increase but decreases more rapidly than l/f. The calculation of the gain us frequency at a fixed power level involves the repeated solution of a cubic equation and is therefore tedious. For the purpose of comparing different materials, sufficient accuracy can be obtained from an approximate solution which may easily be plotted as three straight lines on logarithmic graph paper. This method is used to compare the Performances of CdS, CdSe, ZnO, and GaAs. At a power density of 10 watts per cu cm, the first three give maximum gains of 60 to 80 db per cm in the vicinity of 60 megacycles. At the same power density, GaAs has a maximum gain of about 7 db per cm near 150 megacycles. Although an increase in mobility usually increases the gain for a specified power density, it is shown that there is an optimum mobility above which the gain decreases. An approximate expression for the optimum mobility is established. It is well-known that the interaction between charged particles and an electromagnetic wave can be used to amplify the electromagnetic wave. In the conventional traveling-wave tube, electrons in a vacuum are caused to move at a velocity slightly greater than the wave velocity, thereby coupling energy into the wave.' In principle, the same effect would be evident in a solid except that repeated collisions prevent the electrons from attaining a velocity greater than that of an electromagnetic wave. However, the drift velocity of electrons can be made to exceed the velocity of an acoustic wave in a solid. In 1956, weinreich2 noted that the deformation potential provided one means of coupling energy from drifting electrons into an acoustic wave, but the effect was too small to be of any practical importance. It was not until 1961 that White and Hutson demonstrated that the piezoelectric effect provided a strong enough coupling to give appreciable amplification of an ultrasonic wave and awakened wide interest in electron-phonon interactions. Qualitatively the effect may be described as follows. An acoustic wave propagating through a piezoelectric material is accompanied by an electric-field wave. If the vibrations are properly oriented with respect to the crystal axes of a piezoelectric semiconductor, the local electric fields are axially directed and cause bunching of the charged carriers. If the drift velocity of these carriers is less than the velocity of sound, they tend to be dragged along by the acoustic wave. This electroacoustic effect results in a net potential appearing across the material and in attenuation of the acoustic wave. On the other hand, if the drift velocity exceeds the velocity of sound, energy is coupled from the carriers to the acoustic wave and amplification results. Related effects which are being investigated include ultrasonic amplification in semimetals6-' and semi-conductor, and current saturation in piezoelectric semiconductors.'4-17 There are various problems which hinder the immediate application of ultrasonic amplification in a practical device. First, there are the traditional problems encountered in the design of acoustic delay lines: attenuation in the electromechanical transducers, reflected waves (which can lead to self-oscillation in an amplifying device), mechanical tolerances necessary to maintain phase coherence, and so forth. Second, there is the problem of selecting the best material from among the various piezoelectric semiconductors as they become available in single-crystal form with a wide choice of characteristics. This paper deals with the second problem and shows that the most realistic way of comparing materials is in terms of the gain which can be obtained without exceeding a

Jan 1, 1964

By R. H. McLean

Jet impingement produces two mechanisms to clean the bottom of a borehole during jet-bit drilling operations. One is an impact-pressure wave in the immediate area of jet impingement. The other is crossflow, which spreads across the bottom away from the pressure wave. Measurements of the vertical distribution of the cross-flow at several positions beneath a 4 3/4-in. tricone jet bit in a laboratory model show that the maximum velocity in the crossflow is directly proportional to the square root of the product of the rate of flow and the velocity of the jets through the nozzles (QV). Other measurements in the laboratory model show that gradients in the impact-pressure wave are directly proportional to the jet QV if the diameters of the nozzles are held constant. Consistency of the basic relationships through changes in the Reynolds numbers of the jets from 31,000 to 93,000 suggests that the laboratory results should be representative of typical flow in field operations having higher Reynolds numbers. Estimates of the impact-pressure waves under conditions other than those in the jet-bit model may be made by comparison with a less complex model—impingement of a jet against an unrestricted flat surface normal to the jet. Equations derived from current jet technology describe the wave in this simple system for a wide range of conditions. INTRODUCTION The efficiency of rotary drilling operations is strongly linked to the efficiency with which cuttings are removed. Cuttings remaining on the bottom of the borehole will impede further penetration by the bit until they are removed. Wasteful regrinding is prevented if the fluid circulated past the bit removes cuttings as rapidly as they are made. Most investigations of bottom-hole cleaning under bits have been concerned with the rate-of-penetration effects of bit type, nozzle configuration, fluid properties, pressure gradients and utilization of hydraulic horsepower. These investigations have shown that drilling muds, weighted to keep the pressure in the borehole greater than in adjacent formations and carrying material which forms a filter cake, apparently plaster cuttings against the bottom, making them difficult to remove. In addition to a demonstration of the factors causing poor cleaning, the most significant result of these previous investigations has been the increased use of jet bits accompanied by better hydraulic programs and better selection of drilling fluids. Undoubtedly, these practices have improved bottom-hole cleaning, but rates of penetration when drilling with muds are still often lower than would be expected with perfect cleaning." Although there is general agreement with the concept that increasing hydraulic energy at a jet bit often increases the rate of penetration by improving cleaning, research workers have differed as to the best criterion for evalyating hydraulics. Some lean toward maximizing the product of the rate of flow and the velocity through the nozzles (QV) others prefer to maximize the hydraulic horsepower at the bit;'"-' "till another says that neither criterion is universally suitable."' This controversy shows that conclusive field evidence has not established the superiority of either of these criteria. Analysis of the fluid flow which cleans the bottom should contribute to resolving the controversy. Impinging jets clean the bottom by creating two mechanisms capable of dislodging cuttings. One is the impact from the momentum of the jet and the other is the flaw parallel to the bottom away from the area of impingement. The impact force appears in the form of a pressure wave on the bottom and, in this paper, is called the impact-pressure wave. The flow parallel to the bottom is called crossflow. Fig. 1 illustrates these two mechanisms. This paper describes crossflow and the impact-pressure wave beneath jet bits and relates them to controllable parameters. Through these relations, criteria for use in designing hydraulics to achieve maximum use of the mechanisms of cleaning are derived. PROPERTIES OF CROSSFLOW CONCEPTS DEVELOPED IN PREVIOUS INVESTIGATIONS Concepts of scavenging the bottom by crossflow have been well discussed by Bobo, et al." They examine the means by which crossflow may create forces on cuttings and conclude that the cleaning action of crossflow can be increased by increasing the velocity of the flow stream close to the bottom. Using boundary layer theory, they indicate that a reduction in the viscosity of the drilling

Jan 1, 1965

By R. Wachtell

IN order to study better the phenomena at work in various phases of diffusion of the Fe/Si system when compounded and alloyed by powder metallurgy methods, several attacks have been planned. Electrical, hardness and other measurements have been discussed e1sewhere.l,2 The metallographic phenomena will be considered here. Concerning the metallography of the ferrosilicon alloys, some initial discussion is in order. As this laboratory discovered to its chagrin, there are many traps that lie in wait for even the experienced metal-lographer when he deals with alloys of high silicon content. Chief of these is probably the anomalous behavior of these materials under conventional etching techniques. We have before us, for instance, Corson's paper" of 1928, wherein is mentioned and illustrated (probably for the first time) the peculiar "barley shell" structure sometimes found in these alloys. Again in 19412 he same author reports the structure in 5-14 pct silicon irons. Houghton and Becker mention its presence' but make only tentative effort to explain it. In 1943, Hurst and Riley6 discussed the "barley shell" and declared it to be a false structure, resulting from the peculiar and characteristic film-forming propensities of the alloy. At about the same time, Wrazej7 proved correct the contentions of Hurst and Riley, establishing pretty well the mechanisms of formation and identifying the constituents involved. Hurst and Riley6 a1so describe in their paper (1943) a "cracked film" structure, also a pseudo-morph, the precise nature of which they do not suggest. Careful examination of such a film reveals that it not only cracks but begins to curl away from the surface of the metal and peel, much in the manner of flaking paint. The careful microscopist will be able to focus on the topmost edge of the flake, and, by optical sectioning, follow it down to the metal surface. We have been able, by oblique illumination of such specimens, to observe the raised segment of the film, the shadow that it casts, and the mating segment on the metal into which it fits. As late as 1946, Hurst and Riley8 found occasion to correct misinterpretation of this "cracked film" structure in a work by others on a 12 pct Si/Fe alloy. The "barley shell" and "cracked film" pseudo-morphs are not the only ones encountered in studies of these materials. A third pseudomorph, which may, however, have some structural significance is a fine striation of the surface. Closely spaced and parallel striae extend entirely across single grains. Each grain shows a different direction for the striae, and different closeness of the spacing. The regularity of this structure, and the fact that the striae do not cross from one grain to another, suggest that the structure is a film cracking controlled in some way by the crystal plane orientation of the surface being etched. In general it is wise for the metallographer dealing with these alloys to cultivate a feeling of uncertainty not only of the more complex questions of interpretation, but also of the basic question of whether that which he sees is indeed a true representation of the structure. We have taken, as a first test of the validity of a structure, its reproducibility "in situ" after repolishing and re-etching of the samples. Thus we knew the "barley shell" to be false (fig. 1) even before encountering the definitive works on the subject. The Metallography of Silicon Diffusion in Laminate Bars: As a first step in the study of the metallography of the diffusion process, we have pressed a series of laminate bars. These bars, or "sandwiches" as we have termed them, were pressed with covering layers of iron powder, and a "filler" of the Fe/Si master alloy under study. Bar dimensions were approximately %, x 3 x 1/4 in. The total

Jan 1, 1951

By T. A. Rickard

The interior of West Australia is an arid table-land, elevated 1400 feet above the sea. This plateau is flanked to the south by the Tertiary limestones which fringe the Great Australian Bight. It is bordered northward by the Carboniferous beds of the Fitzroy river and westward by the granite of the Darling hills, while to the cast this wide area, about 900 miles square, slopes downward imperceptibly into an undulating plain of sand, which stretches with dismal persistence across the boundary of South Australia. The waters of the ocean receded from this tract of land long ago; it is probably the oldest landsurface on the globe, and represents the basal wreck of a much larger continent. Fig. 1 is a map of this region. The Coolgardie and Kalgoorlie gold-fields are situated in the southwestern part of the region. The rock-formation consists of granite penetrated by diorites and andesites. The latter are occasionally associated with tuff's, which have been readily mistaken for sedimentaries. There are no fossil-bearing rocks, such as would afford a datum-line from which to measure the relative geological age of the prevailing formation. On the extreme edges of the mining territory there are, it is true, remnants of sand-rock which are considered identical with the " Desert Sandstone " of Queensland, determined by Daintree to be of Mesozoic age. But even this formation has evidently been laid down so long subsequent to the underlying rocks that it serves merely to emphasize their much greater antiquity. In their characteristics and in their relations to each other, the granite and the diorite of the Coolgardie region appear to me much to resemble the Laurentian granite and the Huronian

Jan 1, 1899

By J. H. Reynolds

The possibility of using a semiconductor, metal, semiconductor structure as a light detector is discussed. A brief theoretical argument is presented which predicts that this structure should have pho-toresponse characteristics similur to those of a metal-semiconductor diode, but with the magnitude of the current increased by a transistor gain factor. Measurements made at low frequencies on a particulur S-M-S device agree u~th this prediction aizd shoul that significant gain factors are attainable. The current gain is substantially higher than that predicted for the metal base transistor. It is proposed that the operation of these units is more closely approxinzated by the metal-gate model which predicts higher current gains. A qualitative discussion of its high-frequency performance indicates that detection of light modulated at microwave frequencies by a S-M-S deuice is possible. ADVANCES in the technology of chemical vapor deposition'-4 have made it possible to form a S-M-S structure consisting of a thin metallic layer sandwiched between layers of single-crystal semiconductor. It has been proposed5-7 that, if the metal layer is made sufficiently thin, this structure could be used in an amplifier having several advantages over presently available transistors including higher-frequency performance. Calculations8 predict that the maximum gain-bandwidth product is nearly an order of magnitude higher than that of a bipolar transistor having the same geometry. Maximum oscillating frequencies in the microwave range are predicted. Electrical measurements made on S-M-S devices have shown that, although the theoretical frequency characteristics are not presently attainable, current and power gain is obtained. Another possible application of the S-M-S structure is its use in a light detector. This structure is a three-layer relative of the metal-semiconductor diode. Light detection using M-S diodes is the subject of considerable present interest. They are known to have a short response time and the ability to detect light having a photon energy below the semiconductor bandgap. In this paper, the mechanism by which S-M-S photocurrent is generated is discussed. An expression is derived which shows that the response characteristics are similar to those of a M-S diode, but that the magnitude of the current is increased due to the transistorlike behavior of the device. The results of low-frequency response measurements on a particular S-M-S structure are presented. These agree with the theoretical predictions and show that significant gain factors are attainable. A qualitative discussion of the high-frequency characteristics is presented which predicts that detection of light at microwave frequencies by a S-M-S device is possible. THEORY In this discussion, the photocurrent, iT, measured across terminals attached to the semiconductor emitter and collector layers is related to the current pho-toemitted from the metal base layer. The latter will be referred to as the primary photocurrent. The behavior of the primary photocurrent is determined by the characteristics of the two metal-semiconductor diodes. The diode photoemission is illustrated in Fig. 1, which shows the emitter and collector potential barriers. The metal thickness has been greatly exaggerated. As will be discussed later, the actual barrier shape can be much more complex, but this diagram illustrates the essential features of the photoemission process. An absorbed photon having an energy greater than the barrier height can emit an electron into the emitter or collector. Since this process is discussed

Jan 1, 1968

By W. A. Tiller, W. C. Johnston

The solute distribution ahead of an advancing solid-liquid interface is controlled by varying the momentum boundary layer thickness in the liquid adjacent to the interface. Single pass zone-melting experiments on Pb-Sn alloys are described in which the liquid is caused to rotate in the plane of the solid-liquid interface due to the influence of a magnetic .field rotating in this plane. The variation of the diffusion boundary layer thickness with field strength is determined. DURING the freezing of an alloy a partitioning of solute occurs at the solid-liquid interface, the magnitude depending upon the difference in the solid and liquid solubilities at the interface temperature. Because the diffusion coefficient is so small in the liquid relative to normal freezing rates, either a pileup or a depletion of solute occurs at the interface when the partition coefficient, kO, is either less than or greater than unity, respectively. This solute distribution is completely characterized by i) the solute concentration in the liquid far from the interface, C) the ratio of freezing velocityf, to diffusion coefficient, D, iii) the partition coefficient, ko, and iv) the thickness of the diffusion boundary layer, 6,, at the interface (this assumes that an electric field is not applied across the interface). For a particular system, C ko and D are fixed so that the character of the solute distribution can be changed only by a variation of either f or 6c. As is well known, the freezing velocity can be changed by varying the thermal environment of the alloy. Whereas, the thickness of the solute rich layer can be changed by varying the fluid flow conditions in the liquid. It is the purpose of this paper to describe a technique for controlling the fluid flow conditions in the liquid. During most solidification experiments the driving force for fluid flow arises from either mechanical mixing, thermal convection or electromagnetic mixing. The momentum boundary layer thickness, 6m at the solid-liquid interface, within which negligible movement of the liquid occurs, has been calculated for the special case of mechanical mixing where crystals are solidified by the Czochralski technique.l,2 The calculation of 6m for the case of thermal convection has also been made.3 In some calculations it has been assumed that the momentum boundary layer thickness, 6,, may be equated to the thickness of the diffusion boundary layer 6c, at the interface. However, as may be seen from Appendix I this is only true in very special cases and in general where a is a constant of magnitude determined by the particular liquid alloy. Neither calculations nor experimental data exist for the dependence of upon electromagnetic mixing. In this paper experiments are described in which the liquid is caused to rotate in the plane of the interface due to the torque developed on it by a magnetic field, H, rotating in this plane. The variation of is controlled by the variation of H. A magnetic field rotating in a plane perpendicular to the interface has been used for the purification of aluminum. Experimentally, we determined the variation of 5C with H by using the zone-melting technique of pfann5and a theoretical relationship of Burton et a1.2 pfann5 has shown that, after the passage of one molten zone through a long cylindrical charge of uniform concentration, CO, the solute distribution, C,, in the solid is as shown in Fig. 1. The equation describing this distribution is

Jan 1, 1962

By A. E. Worthington, R. F. Meldau

An analog computer study was made of the reduction of the magnitude of the SP by formation geometry and resistivity effects. Practical field .situations were studied. The analog data were converted to the form of departure charts. These charts may he used to determine the reduction in the SP due to the above eflects pith good precision Quantitative interpretation of the SP log in terms of formation water resistivity is presently based on the premise that the observed SP is equal to the electrochemical potential developed across a shale between the mud filtrate and the formation water. This procedure inherently assumes that, first, no electro kinetic (streaming) potential is present, and, second, that the SP is not reduced by resistivity or geometry effects in the formation. When, however, such a reduction of the observed SP does occur, one may write, SP = b,E, b, E, (1) where SP is the observed self-potential opposite the center of the bed under consideration. In this case, E, is the electrochemical potential, E, is the electrokinetic potential, b, is the electrochemical reduction factor, and h, is the electrokinetic reduction factor. The first assumption above is equivalent to assuming E, = O in Eq. 1 and the second assumption is equivalent to h, = b, = 1. Examination of electric logs shows that for many formations of interest the observed SP is reduced below its maximum value. The object of this work is to aid quantitative SP interpretation in at least some of these situations by obtaining values for b, and b, in Eq. 1 as functions of realistic combinations of bed thickness, depth of invasion, and formation resistivity. The only previously published work in this field known to the authors is by H. G. Doll' in 1948. He evaluated b, under conditions permitting exact analytical solution. His work clearly established the importance and qualitative nature of formation geometry and resistivity effects upon the SP, but the necessity of obtaining analytical solutions restricted his study to situations of limited practical interest. In order to extend such studies to practical combinations of formation parameters which do not lend themselves to feasible analytical solutions, the results reported in this paper were obtained by use of a special analog computer. THE ANALOG MODEL The analog model used for this study was the Guyed Analog Computer which originally was designed and built for resistivity log departure curve studies. It consists of a two-dimensional mesh of resistors, with the vertical mesh corresponding linearly to vertical distance along the borehole, and the horizontal mesh corresponding to radial distance from the borehole axis. By proper choice of horizontal resistor values, a three-dimensional, axially symmetric, borehole-formation problem can be modeled with this two-dimensional array. These and other details of the model are contained in a paper by H. Guyod'. The geometry used for the SP studies with the analog is depicted in Fig. 1. This configuration has a horizontal plane of symmetry, and, hence, the model, and the following discussion, need only to be concerned with one quadrant such as shown in Fig. 2. A further simplification arises by considering that both the natural potential sources and those used in the model have fixed internal impedance. Therefore, the current flows produced by these sources can be separated and studied independently ( i.e., the superposition theorem can be applied). SP = SP, i SP, = h,E, + b,E, ,(7)

By E. T. Turkdogan, P. Grieveson, J. F. Beisler

Experimental results are given for the rate of reduction of silica from silicate melts by paphite-saturated iron in the presence of carbon monoxide. It is shown that, when gas bubbles are present at the slag-metal interface, the interfacial chemical reaction is the rate-controlling process in the transfer of silicon from slag to metal. From the silicon-concentration profiles in the metal sample, the diffusivity of silicon in paphite-saturated iron is found to be (5.4 * 0.4) x 10 sq cm per sec at 1600°C. Some consideration is given to the transport of oxygen from the slug-metal interface to the walls of the gvaphite metal container as becoming the rate-controlling process in the absence of gas bubbles at the slag-metal interface. ATTEMPTS were made by Fulton and chipman1 to investigate the kinetics of reduction of silica from calcium alumino-silicate melts by graphite-saturated iron in an atmosphere of carbon monoxide. In these experiments with 400 to 500 g of metal and slag, both phases were stirred in such a manner that the interface was essentially undisturbed. Initially, the graphite-saturated iron contained about 15 pct Si and, after about 8 or 10 hr of reaction time at 1600°C, the silicon content of the metal in- creased by about 0.2 pct. In a written discussion on this paper of Fulton and Chipman, schuhmann2 pointed out the possibility that this observed slowness of the reduction of silica could be due to the transport of oxygen across the liquid metal boundary layer at the slag-metal interface. Although this explanation of Schuhmann is quite feasible for the experimental data of Fulton and Chipman on silica reduction, such a reasoning should not be extended to other rate data on slag-metal reactions unless sufficient evidence is available satisfying the boundary conditions for such a transport process. The hydrodynamics of stirred liquids are very complex and even for a simple method of stirring there is no exact mathematical solution which will predict the mass transfer coefficient accurately. The best available solution of this problem is that given by Lewis' and later modified by mayers~ for a two-liquid system where liquids are stirred by counter-rotating stirrer blades without disturbing the interface. In such a system the mass transfer coefficient is estimated from an empirical relationship using the data on viscosity and density of the two liquids, rate of revolution, the length of stirrer blades, and the diffusivity of the element which is transferred from one phase to the other. It was found in these investigations' that the estimated and observed mass transfer coefficients, in systems containing two immiscible aqueous solutions, agreed within *40 pct. Although this is not considered to be an unreasonable error for the study of industrial processes, uncertainties of this magnitude cannot be tolerated in any theoretical study of reaction kinetics. In view of the complexity of the

Jan 1, 1963

By David Turnbull

IN isothermal recrystallization processes, new crystals generally grow into the matrix until they impinge upon other new crystals or an external surface, at constant linear rates G. Before impingement the perceptible course of growth can be described by the equation: 1 = G(t-7) C1I where, G = dl/dt, 1 is a crystal dimension measured in a constant direction, t is the time, and 7, the nucleation period, is a positive intercept on the time axis. Fig. 1 is a schematic representation of I as a function of time for a recrystallizing grain. G is dependent upon temperature, driving energy (strain or surface energy), relative grain and boundary orientations, but is generally independent of time. The frequency of nucleation, fi, (time" volume") can be defined by the equation: N = 1/fV [2] where ? is the mean nucleation period and V is the volume of the specimen that has not recrystallized. The kinetics of primary and secondary recrystallization generally can be described satisfactorily in terms of the parameters N and G only.'-" After recrystallization is complete the average grain size 7 increases with time by "normal grain growth;" didt, the average rate of grain growth, is strongly time dependent and has not yet been precisely related to G for the motion of the individual grain boundaries constituting the system. It has been suggested4* " that the elementary act in grain boundary migration is closely related to the elementary act in grain boundary self-diffusion. Although the distance of atom movement in the two processes may be somewhat different, there is reason to expect that the activated states may be very similar, so that the free energy of activation for grain boundary migration should be of the same order of magnitude as for grain boundary self-diffusion. Therefore, it is desirable to develop a satisfactory basis for comparing data on self-diffusion and grain boundary migration and to make such comparisons where possible. Theory The formal theory of grain boundary migration rates is analogous to the theory for the rate of growth of crystals into supercooled liquids reviewed elsewhere 6-8. Boreliuss has shown that the latter theory describes, within the theoretical uncertainty, the growth of selenium crystals into supercooled liquid selenium. Motto and more recently Smolu-chowski" have derived expressions for the rate of boundary migration in recrystallization. The treatment to be presented is similar to Mott's excepting that the formalism of the absolute reaction rate theory will be used. The atomic mobility, M, in grain boundary migration is defined by: G = -M6p/6x where p is the chemical potential per atom and x is the coordinate measured in the direction of grain boundary movement. Let AF be the free energy difference per gram atom on the two sides of the boundary and k the thickness of the boundary. For RT>>AF the potential gradient across the boundary (6p/6x) is essentially linear and it follows that: SF/8x = - aF/N\ [4] where N is Avogadro's number. According to the Nernst-Einstein equation, M is related to a diffusion coefficient, Do, for matter transport in grain boundary migration by the equation: M = Da/kT [5] Substituting eqs 4 and 5 into eq 3 gives the basic relation between Do and G: G = DoaF/\RT [6] Do values may be calculated from experimental values of G from eq 6 and directly compared with the coefficient of self-diffusion within the crystal, DL, or the grain boundary self-diffusion coefficient D,. However, a more convenient, though equivalent, basis for comparing atomic mobility in grain boundary migration and self-diffusion is through the constants of the absolute reaction rate theory. According to this theory diffusion coefficients may be written:" D = k2(kT/h) exp [-AF,/RT] 171 aFa, the free energy of activation, is related to the measured energy of activation, Q, by the equation: AFA = Q - T aSx - RT [8] where aSa is the entropy of activation. Substituting eqs 8 and 7 into eq 6 gives: G = ek(kT/h) (aF/RT) exp [(AS,,)C/R] exp C-Qc/RTI C91 where the subscript G refers to boundary migration. The relationship between the driving free energy and the free energy of activation in boundary migration is indicated schematically in Fig. 2. Experience indicates that the variation of G with temperature can be described by: G= Go exp [- Qc/RT] [10] where Go and Qc are generally temperature independent over wide ranges of temperature. Comparison of eq 9 with eq 10 gives:

Jan 1, 1952