Search Documents

By Frank G. Miller

The lower cost of producing oil from naturally flowing wells compared with pro-luction costs accruing from artificial lifting methods has stimulated much research, rith the joint purpose of extending the natural flowing life of wells and of increas-ng the productivity of small wells. The iboratory investigation described herein was made to determine the effect of a lotted liner on the pressure gradient in the icinity of the wall of a flowing well. Flow estrictions caused by the liner induce nergy losses between the oil-bearing for-lation and the inside of the liner, which lay be reduced by improved liner design, nd thereby the flow capacity of wells may e increased. These losses have been evalu-ted for rectangular slots of different sizes, nd means are suggested for reducing them. The dearth of information concerning the ature of the flow of fluids from uncon-slidated sands through small openings in he walls of. pipes—as represented by the ow of oil from an oil sand into a slotted ner—is acknowledged. Although the bene-ts to be realized by operators through the se of optimum liners is just beginning to e recognized generally, Coberlyl in 1937 ointed out that the functioning of screen is is an important factor in the pro-uction of oil. From his study on the :lection of size of oil strings for wells in California, Parks2 concluded, in part, that the perforated section of the oil string deserves continued study, especially with respect to the selection of perforations. The growing use of gravel packing resulting from certain research findings lends even more significance to the problem of selecting optimum liners, as the application of these methods usually requires slotted liners or other kinds of perforated liners. The present study was made to obtain further information for the guidance of engineers and operators in selecting slotted liners and to develop a laboratory method that may be used in analyzing possible pressure conditions surrounding most types of perforated casings used in producing oil wells, so that the present inadequate fund of knowledge on this important subject may be expanded. Discussion of ProbleM At the beginning of the investigation the following assumptions were made regarding bottom-hole conditions of a producing oil well and the oil-sand reservoir it penetrates: (I) Conditions of steady-state flow exist, (') a a homogeneous liquid is being produced) (3) no contaminating material (such as drilling fluid) is present between the liner and the face of the oil-bearing sand, (4) the well is vertical and completely penetrates a horizontal sand stratum of uniform thickness occurring between two impermeable strata, and (5) the flow is wholly viscous. If the foregoing assumptions are made, with the added requirements that the sand

Jan 1, 1941

By Frank G. Miller

The lower cost of producing oil from naturally flowing wells compared with pro-luction costs accruing from artificial lifting methods has stimulated much research, rith the joint purpose of extending the natural flowing life of wells and of increas-ng the productivity of small wells. The iboratory investigation described herein was made to determine the effect of a lotted liner on the pressure gradient in the icinity of the wall of a flowing well. Flow estrictions caused by the liner induce nergy losses between the oil-bearing for-lation and the inside of the liner, which lay be reduced by improved liner design, nd thereby the flow capacity of wells may e increased. These losses have been evalu-ted for rectangular slots of different sizes, nd means are suggested for reducing them. The dearth of information concerning the ature of the flow of fluids from uncon-slidated sands through small openings in he walls of. pipes—as represented by the ow of oil from an oil sand into a slotted ner—is acknowledged. Although the bene-ts to be realized by operators through the se of optimum liners is just beginning to e recognized generally, Coberlyl in 1937 ointed out that the functioning of screen is is an important factor in the pro-uction of oil. From his study on the :lection of size of oil strings for wells in California, Parks2 concluded, in part, that the perforated section of the oil string deserves continued study, especially with respect to the selection of perforations. The growing use of gravel packing resulting from certain research findings lends even more significance to the problem of selecting optimum liners, as the application of these methods usually requires slotted liners or other kinds of perforated liners. The present study was made to obtain further information for the guidance of engineers and operators in selecting slotted liners and to develop a laboratory method that may be used in analyzing possible pressure conditions surrounding most types of perforated casings used in producing oil wells, so that the present inadequate fund of knowledge on this important subject may be expanded. Discussion of ProbleM At the beginning of the investigation the following assumptions were made regarding bottom-hole conditions of a producing oil well and the oil-sand reservoir it penetrates: (I) Conditions of steady-state flow exist, (') a a homogeneous liquid is being produced) (3) no contaminating material (such as drilling fluid) is present between the liner and the face of the oil-bearing sand, (4) the well is vertical and completely penetrates a horizontal sand stratum of uniform thickness occurring between two impermeable strata, and (5) the flow is wholly viscous. If the foregoing assumptions are made, with the added requirements that the sand

Jan 1, 1941

By Aaron J. Miles, Eugene A. Stepenson

The pressure distribution in a producing oil or gas reservoir has been obtained mathematically in a limited number of special cases where the boundary of the reservoirs are simple geometric figures with the wells conveniently located within them. Using some of the previous work as a basis, it can be proved that a number of additional problems of pressure distribution can be solved by use of soap film. This method is particularly applicable to a reservoir drained by a number of wells, or to a reservoir of irregular shape, or to both. Since this method obtains the pressure distribution, it also establishes the drainage boundaries of the wells. This paper derives the analogy between the equation representing the pressure distribution in a reservoir and the equation of the deflected membrane. It also describes the apparatus and the method of operation. The experimental method is employed to solve a problem whose mathematical solution is known and the results compared. These are found to be in good agreement. The method is applicable to any homogeneous liquid, compressible or incompressible, or to a gas, flowing in a reservoir of uniform thickness, permeability and porosity. The flow state may be either the steady state or the transient state wherein the density or pressure of the fluid in the reservoir declines at a constant rate. PRESSURE-DISTEIBUTION PBOBLEM Fluids encountered in the oil reservoir are compressible and expand and move toward the drilled hole where the Dressure is lowest. The loss of fluid from the reservoir lowers the pressure in the more distant parts of the reservoir, permitting the fluids in these parts to expand and move toward the well. The area around the well affected by drainage is increasing continuously, and the density, pressure, and velocity of flow

Jan 1, 1938

By Charles E. Reed

With the increasing importance of clay in drilling operations which demand more precise and exacting control over its behavior, there has come the realization that most of our present methods of control over many of its important properties still represent more of an art than a science. Such a situation with respect to a material of such industrial significance is unfortunate, and naturally many investigations have been conducted to improve our knowledge. In general, these investigations have revealed the complex nature of clay behavior and while sometimes they may not have seemed to contribute directly to a solution of the field operator's problems they have been of inestimable importance in bringing to light the broad fundamental scientific principles upon which must be based any effective working theory correlating the known facts of clay behavior and predicting profitable new lines of attack for its improvement. Major Facts Concerning Behavior of Clays In the study of clays, as in that of most other colloidal materials, it has been impossible to make progress by the study of isolated facts. On the other hand, analysis of large groups of facts, many of which seem at first completely unrelated, can lead to a deeper insight into the structure of clays and the mechanism of their behavior. It will be well to list the major facts on which the analysis here presented is based, starting with those that have emerged repeatedly from past investigations and which thereby constitute an impregnable foundation upon which must rest any comprehensive theory of clay behavior: 1. Clays consist of hydrated aluminosilicates of extremely fine and variable particle size. 2. Suspended in water, clays exhibit variable degrees of swelling, depending upon their origin and upon the chemical and colloidal nature of any other substances present. The magnitude of the swelling is a direct indication of the extent to which the clay has become dispersed.

Jan 1, 1938

By D. T. MacRoberts

This paper embodies the results of theoretical studies concerning gas reservoirs, especially the effect of drilling programs of various intensities upon pressure depletion and ultimate recoveries. The primary objective has been to determine the net quantity of recoverable gas that could be obtained as a result of each additional well, in comparison with the gas . that might be secured from a single well ideally placed with respect to both acreage and geological structure. No method is yet known whereby the relative merits of intensive and sparse drilling can be compared experimentally; a reservoir must either be drilled closely or developed with wide spacing, but both programs cannot be conducted simultaneously. Hence, the analytical method seems to be the only approach to the problem. While the present state of our theoretical knowledge does not offer a precise solution to the problem of gas flow in sand bodies, it takes such appropriate cognizance of the limitations of the methods used as to restrict any answer obtained to the realm of engineering accuracy. In solving the problem for the average gas field, it was found that practically the same answers resulted as to pressure distribution, regardless of the size of the drainage radius, even though the latter was assumed to vary from a few feet as a minimum to that of the entire reservoir as a maximum. Such limits represent extreme possibilities; therefore, the practical conditions that are associated with normal development programs, in which wells are placed in the center of tracts varying in size from 40 to 640 acres, will of necessity result in less extreme variations than those indicated in the tables accompanying this paper. Stated more concisely, the amount of gas unrecovered at abandonment (of the particular field used for illustrative purposes) will lie somewhere between 4.6 and 6 per cent of the total original content, no matter how intensively the field is drilled. Gas reservoirs may be divided into two general types, those in which the energy is intrinsic, and those of which the energy is extrinsic. † The first type contains energy in the form of a compressed gas, which expands

Jan 1, 1938

By Ben Gumpertz

The important factors in any study of the screening of sand with a gravel envelope, as applied to use in oil wells, are: (I) sand size and shape, (z) gravel size and shape, (3) ratio of gravel size to sand size, (4) flow velocity, (5) viscosity of the fluid. Perhaps the first extensive investigation of the function of gravel in screening sand was performed by Coberly and Wagner,' from which the following conclusion was drawn: "The diameter of the largest grains in a single classified sand suitable for gravel packing is approximately 13 times the grain size at the ten-percentile analysis of the formation sample." ?;his applies to sands classified by Tyler screens and is for gravel grains of the maximum size that could be used to form a stable bridge. Field practice has shown that the screening effect of a given gravel for a given sand apparently varies with the flow velocity." Also, it has been assumed by some operators that the thickness of the walls of the gravel envelope surrounding the well has little effect upon the screening process. It is thought by others that a narrow envelope of gravel might form an ineffective screen. Thus the purposes of this work are: 1. To investigate the effect of gravel-sand size ratio upon screening action under the condition of varying flow velocities. 2. To investigate the effect upon screening action of varying the thickness of the gravel screen. 3. To investigate sand migration in the gravel screen and the effect of flow velocity and gravel size upon sand migration. Description of Apparatus In an attempt to reproduce field conditions for the purposes of this investigation, a flow tube was constructed, according to the manner of Urenj3 to represent a radial drainage segment of an oil sand. To arrive at the dimensions of the flow tube, a radial drainage segment of rectangular cross section was assumed to define an area of z sq. in. on the face of a well 6 in. in diameter. From this the dimensions of the segment were calculated out to a distance of 5 ft. from the face of the well. By computing cross-sectional areas at a number of sections along the rectangular segment and transforming these areas into the circular form, a drainage cone was obtained having everywhere the same cross-sectional area as the drainage element of rectangular cross section. To construct the apparatus, two sections of pipe of 7%-in. inside diameter, one 13 in. and the other 4 ft. long, were joined by a bolted flange. A mandrel having the dimensions of the drainage cone was centered in the pipe and a neat cement mixture was cast around it. A protective gasket was placed between the two sections of pipe so that the cement body could be separated at that point. When the mandrel was withdrawn a drainage cone was obtained (Fig. I).

Jan 1, 1941

By Ben Gumpertz

The important factors in any study of the screening of sand with a gravel envelope, as applied to use in oil wells, are: (I) sand size and shape, (z) gravel size and shape, (3) ratio of gravel size to sand size, (4) flow velocity, (5) viscosity of the fluid. Perhaps the first extensive investigation of the function of gravel in screening sand was performed by Coberly and Wagner,' from which the following conclusion was drawn: "The diameter of the largest grains in a single classified sand suitable for gravel packing is approximately 13 times the grain size at the ten-percentile analysis of the formation sample." ?;his applies to sands classified by Tyler screens and is for gravel grains of the maximum size that could be used to form a stable bridge. Field practice has shown that the screening effect of a given gravel for a given sand apparently varies with the flow velocity." Also, it has been assumed by some operators that the thickness of the walls of the gravel envelope surrounding the well has little effect upon the screening process. It is thought by others that a narrow envelope of gravel might form an ineffective screen. Thus the purposes of this work are: 1. To investigate the effect of gravel-sand size ratio upon screening action under the condition of varying flow velocities. 2. To investigate the effect upon screening action of varying the thickness of the gravel screen. 3. To investigate sand migration in the gravel screen and the effect of flow velocity and gravel size upon sand migration. Description of Apparatus In an attempt to reproduce field conditions for the purposes of this investigation, a flow tube was constructed, according to the manner of Urenj3 to represent a radial drainage segment of an oil sand. To arrive at the dimensions of the flow tube, a radial drainage segment of rectangular cross section was assumed to define an area of z sq. in. on the face of a well 6 in. in diameter. From this the dimensions of the segment were calculated out to a distance of 5 ft. from the face of the well. By computing cross-sectional areas at a number of sections along the rectangular segment and transforming these areas into the circular form, a drainage cone was obtained having everywhere the same cross-sectional area as the drainage element of rectangular cross section. To construct the apparatus, two sections of pipe of 7%-in. inside diameter, one 13 in. and the other 4 ft. long, were joined by a bolted flange. A mandrel having the dimensions of the drainage cone was centered in the pipe and a neat cement mixture was cast around it. A protective gasket was placed between the two sections of pipe so that the cement body could be separated at that point. When the mandrel was withdrawn a drainage cone was obtained (Fig. I).

Jan 1, 1941

By H. H. Evinger, M. Muskat

Although the shooting of oil wells for the purpose of increasing production has been practiced since 1866, present-day shooting technique has been arrived at almost wholly by a process of trial and error. The difficulties attending either a theoretical analysis of the problem or an experimental attack become apparent when one actually embarks on a theoretical investigation or attempts to formulate an intelligent experimental program. The literature on the subject consists mainly of descriptions of current practice, scattered eye-witness accounts of events following the shooting of a well, and vague speculations on the mechanical processes involved. In contrast to quarry blasting, the results of oil-well shooting cannot be determined by means of visual inspection, and it is perhaps for this reason that only one attempt at what approaches a direct experimental attack has been reported in the literature.' In these experiments, charges of liquid nitroglycerin were exploded in holes drilled in a sandstone outcrop. In order to observe the character and extent of the resulting fractures, the rock was blasted away so as to expose the original shot holes. After all this had been done it was found impossible to distinguish the fracturing due to the original shots from that due to the subsequent blasting. The enlargement of the holes due to the pulverizing action of the shots was measured and an attempt was made to determine the effectiveness if sand tamping. Only five shots were made and few definite conclusions could be drawn. Some small-scale experiments to deter-mine the effect of tamping were performed by Snelling and IIalL2 Small charges (zo grams) of dynamite were detonated at the bottom of a cylindrical hole in a lead block and the resulting enlargement of the hole was taken as a measure of the effectiveness of the explosion. It was found that when the shot was tamped with a sufficient amount of sand or clay, the volume enlargement was approximately double that found when no stemming had been used. The properties of the explosives have been the subject of much study by the manufacturers, the U. S. Bureau of Mines, and various other agencies. For example, velocities of detonation for many commercial explosives are to be found in the Blasters' Handbook, issued by the Du Pont company. Many other properties are given in a treatise on the subject by Naoum.~ The shock wave that accompanies the detonation of high explosives has been studied photographically by Payman4 and others. The hydrodynamical theory of the shock wave has received much attention, and an extended theoretical treatment is given by Bateman.6 Two aspects of the well-shooting problem will be considered in this paper; namely, the fracturing of subsurface strata, and possible damage to casing. With respect to the former of these, several calculations have been made giving the stresses in the rocks surrounding the shot. In order to arrive at numerical results at all, however, it is necessary to introduce a great many simplifying

Jan 1, 1941

By M. Muskat, H. H. Evinger

Although the shooting of oil wells for the purpose of increasing production has been practiced since 1866, present-day shooting technique has been arrived at almost wholly by a process of trial and error. The difficulties attending either a theoretical analysis of the problem or an experimental attack become apparent when one actually embarks on a theoretical investigation or attempts to formulate an intelligent experimental program. The literature on the subject consists mainly of descriptions of current practice, scattered eye-witness accounts of events following the shooting of a well, and vague speculations on the mechanical processes involved. In contrast to quarry blasting, the results of oil-well shooting cannot be determined by means of visual inspection, and it is perhaps for this reason that only one attempt at what approaches a direct experimental attack has been reported in the literature.' In these experiments, charges of liquid nitroglycerin were exploded in holes drilled in a sandstone outcrop. In order to observe the character and extent of the resulting fractures, the rock was blasted away so as to expose the original shot holes. After all this had been done it was found impossible to distinguish the fracturing due to the original shots from that due to the subsequent blasting. The enlargement of the holes due to the pulverizing action of the shots was measured and an attempt was made to determine the effectiveness if sand tamping. Only five shots were made and few definite conclusions could be drawn. Some small-scale experiments to deter-mine the effect of tamping were performed by Snelling and IIalL2 Small charges (zo grams) of dynamite were detonated at the bottom of a cylindrical hole in a lead block and the resulting enlargement of the hole was taken as a measure of the effectiveness of the explosion. It was found that when the shot was tamped with a sufficient amount of sand or clay, the volume enlargement was approximately double that found when no stemming had been used. The properties of the explosives have been the subject of much study by the manufacturers, the U. S. Bureau of Mines, and various other agencies. For example, velocities of detonation for many commercial explosives are to be found in the Blasters' Handbook, issued by the Du Pont company. Many other properties are given in a treatise on the subject by Naoum.~ The shock wave that accompanies the detonation of high explosives has been studied photographically by Payman4 and others. The hydrodynamical theory of the shock wave has received much attention, and an extended theoretical treatment is given by Bateman.6 Two aspects of the well-shooting problem will be considered in this paper; namely, the fracturing of subsurface strata, and possible damage to casing. With respect to the former of these, several calculations have been made giving the stresses in the rocks surrounding the shot. In order to arrive at numerical results at all, however, it is necessary to introduce a great many simplifying

Jan 1, 1941

By M. C. Leverett, W. B. Lewis

The dynamic behavior of a multiple fluid system is completely describable in terms of driving forces and resistances to flow. The latter are proportional to the viscosity of the fluid under consideration and inversely proportional to a property of the system termed the " effective permeability " to that fluid, a property therefore defined by this relation. Since three phases are usually present in oil reservoirs, it is essential to know the effective permeabilities to the several fluid phases, and the relationship between the composition of the fluid in the pore space and that of the flowing stream, in order to handle satisfactorily certain problems of oil production. In this regard, there are many unknown factors and parameters, and many that can be evaluated only indirectly. However, it is expected that work along lines similar to this present undertaking will help materially in solving these problems. .. . . Previous workers in this field have investigated the two-phase systems, water-gas,' water-oil2 in unconsolidated sands, and water-gasg in consolidated sands. The extension of the study of the flow of heterogeneous fluid systems from two to three phases involved only slight changes in technique. The apparatus used in the present work, except for slight changes to be described, was that used by Leverett in his water-oil experiments.~ Correlation and interpretation of data were made by methods analogous to those used in two- phase work. The relative permeability to a phase—the effective permeability divided by the permeability of the system to a homogeneous fluid—was employed throughout in correlating the rate of flow of that phase with the composition of the pore space. Quantitative measurements of two independent physical properties of the system were required to determine the composition of the pore space. As in the two-phase work, the fraction of water was determined electrically, The gas fraction was determined by measuring changes in volume accompanying measured pressure changes. Suitable account was taken of the solubility of the gas in the liquid phases, and of the vapor pressure of the water phase. Since this work was exploratory in nature, no attempt was made to obtain highly accurate numerical results. Reasonable precision and reproducibility were sought. Experimental Procedure variables Considered.—Of the properties of the fluids, the oil viscosity is most likely in practice to vary widely. Variations in oil viscosity were therefore included in this work, but variations in other properties of the fluids are as yet unexplored. It appears unlikely that the normally encountered variations will produce significant changes. ~h~ observations were made at fairly high pressure gradients, no attempt being made to detect a variation of the relative permeabilities with pressure gradient. It should be pointed out also that since most of the data were taken on one sand size only, it call Only be assumed at present that changes in

Jan 1, 1941

By W. B. Lewis, M. C. Leverett

The dynamic behavior of a multiple fluid system is completely describable in terms of driving forces and resistances to flow. The latter are proportional to the viscosity of the fluid under consideration and inversely proportional to a property of the system termed the " effective permeability " to that fluid, a property therefore defined by this relation. Since three phases are usually present in oil reservoirs, it is essential to know the effective permeabilities to the several fluid phases, and the relationship between the composition of the fluid in the pore space and that of the flowing stream, in order to handle satisfactorily certain problems of oil production. In this regard, there are many unknown factors and parameters, and many that can be evaluated only indirectly. However, it is expected that work along lines similar to this present undertaking will help materially in solving these problems. .. . . Previous workers in this field have investigated the two-phase systems, water-gas,' water-oil2 in unconsolidated sands, and water-gasg in consolidated sands. The extension of the study of the flow of heterogeneous fluid systems from two to three phases involved only slight changes in technique. The apparatus used in the present work, except for slight changes to be described, was that used by Leverett in his water-oil experiments.~ Correlation and interpretation of data were made by methods analogous to those used in two- phase work. The relative permeability to a phase—the effective permeability divided by the permeability of the system to a homogeneous fluid—was employed throughout in correlating the rate of flow of that phase with the composition of the pore space. Quantitative measurements of two independent physical properties of the system were required to determine the composition of the pore space. As in the two-phase work, the fraction of water was determined electrically, The gas fraction was determined by measuring changes in volume accompanying measured pressure changes. Suitable account was taken of the solubility of the gas in the liquid phases, and of the vapor pressure of the water phase. Since this work was exploratory in nature, no attempt was made to obtain highly accurate numerical results. Reasonable precision and reproducibility were sought. Experimental Procedure variables Considered.—Of the properties of the fluids, the oil viscosity is most likely in practice to vary widely. Variations in oil viscosity were therefore included in this work, but variations in other properties of the fluids are as yet unexplored. It appears unlikely that the normally encountered variations will produce significant changes. ~h~ observations were made at fairly high pressure gradients, no attempt being made to detect a variation of the relative permeabilities with pressure gradient. It should be pointed out also that since most of the data were taken on one sand size only, it call Only be assumed at present that changes in

Jan 1, 1941

By H. H. Reamer, B. H. Sage

Attempts have been made to predict the volumetric and phase behavior of naturally occurring hydrocarbon mix-turesl-6 but these methods have not been extended to the higher pressures and tem-peratures that are currently encountered in some producing formations. Furthermore, the methods of prediction at present avail-able have not been tested in a sufficient number of cases to ensure their applica-bility to all naturally occurring hydrocar-bon mixtures, even at the lower pressures and temperatures. For these reasons the behavior of a number of hydrocarbon mix-tures from a field in which the subsurface pressures are above 4500 lb. per sq. in.t and the temperatures are in the neighbor-hood of 250°F. was investigated. Method The objective of this program was to determine the volumetric behavior of ma-terial as it existed in the reservoir as a function of pressure, temperature and com-position. Strictly speaking, this cannot be accomplished in any simple fashion. How- ever, the variations in composition that are encountered within a producing reservoir may be approximated roughly by means of mixtures made up of two arbitrarily chosen constituents. Experience indicates that the liquid and gas phases from a high-pressure gas trap afford a simple source of material for this purpose. Throughout this discus-sion the samples of the gas phase and liquid phase from the production trap will be called the "trap-gas sample" and the "trap-liquid sample," respectively. The volumetric behavior of both of the trap samples was determined throughout the temperature interval between 100" and 250°F. at pressures up to 5000 lb. per sq. in. In addition, the behavior of eight different mixtures containing varying proportions of the two trap samples was studied throughout the same temperature and pressure intervals. The experimental results yielded information concerning the specific volume of the material as a function of pressure, temperature and the relative quantity of the trap samples. The relative amount of each of the paraffin hydrocarbons of lower molecular weight than n-hexane was determined for each of the trap samples. In addition, the specific weight, viscosity and average molecular weight of the hexanes-and-heavier fraction of the trap-liquid sample were determined as well as similar information for each of five fractions separated from this material by distillation. From these data the weight fractions of each of the components from methane through pentane were determined for the trap samples and eight systematically chosen mixtures. Materials The gas and liquid samples were obtained from the production trap of a well in the Rio Bravo field in California during the fall of 1939. Information concerning the condi- 70

Jan 1, 1941

By H. H. Reamer, B. H. Sage

Attempts have been made to predict the volumetric and phase behavior of naturally occurring hydrocarbon mix-turesl-6 but these methods have not been extended to the higher pressures and tem-peratures that are currently encountered in some producing formations. Furthermore, the methods of prediction at present avail-able have not been tested in a sufficient number of cases to ensure their applica-bility to all naturally occurring hydrocar-bon mixtures, even at the lower pressures and temperatures. For these reasons the behavior of a number of hydrocarbon mix-tures from a field in which the subsurface pressures are above 4500 lb. per sq. in.t and the temperatures are in the neighbor-hood of 250°F. was investigated. Method The objective of this program was to determine the volumetric behavior of ma-terial as it existed in the reservoir as a function of pressure, temperature and com-position. Strictly speaking, this cannot be accomplished in any simple fashion. How- ever, the variations in composition that are encountered within a producing reservoir may be approximated roughly by means of mixtures made up of two arbitrarily chosen constituents. Experience indicates that the liquid and gas phases from a high-pressure gas trap afford a simple source of material for this purpose. Throughout this discus-sion the samples of the gas phase and liquid phase from the production trap will be called the "trap-gas sample" and the "trap-liquid sample," respectively. The volumetric behavior of both of the trap samples was determined throughout the temperature interval between 100" and 250°F. at pressures up to 5000 lb. per sq. in. In addition, the behavior of eight different mixtures containing varying proportions of the two trap samples was studied throughout the same temperature and pressure intervals. The experimental results yielded information concerning the specific volume of the material as a function of pressure, temperature and the relative quantity of the trap samples. The relative amount of each of the paraffin hydrocarbons of lower molecular weight than n-hexane was determined for each of the trap samples. In addition, the specific weight, viscosity and average molecular weight of the hexanes-and-heavier fraction of the trap-liquid sample were determined as well as similar information for each of five fractions separated from this material by distillation. From these data the weight fractions of each of the components from methane through pentane were determined for the trap samples and eight systematically chosen mixtures. Materials The gas and liquid samples were obtained from the production trap of a well in the Rio Bravo field in California during the fall of 1939. Information concerning the condi- 70

Jan 1, 1941

By L. S. Darken

In the study of reactions occurring in liquid iron, alone or in contact with a liquid oxide or slag phase, it has been found that the experimental data over a limited range of concentration can in some cases be represented satisfactorily by the simple equilibrium expression where (P) denotes the concentration of a product and (R) the concentration of a reactant in the reaction r1R1 + r2R2 + • • • D P lP1 + P 2P2 + • • • Some of the reactions for which it holds are listed in Table 1. For others, however, the simple equation apparently does not hold. This has been attributed to an improper method of expressing concentration (for instance, weight per cent instead of mol fraction); or it has been supposed that the solutions are not ideal in the sense that the activity of a component is not proportional to its true mol fraction. But it may be that the problem has been oversimplified by neglecting some molecular species actually present; that is, by considering only the ostensible simple components and thus leaving out of account possible compound formation. It is the purpose of the present paper to point out how, in a group of related cases, a reasonable consistent assumption relative to the presence of a compound makes it possible to express the experimental data satisfactorily by the simple equilibrium expression when this compound is taken into account, and to bring into harmony data referring to different reactions in liquid iron. There are several a priori reasons for believing that solutions in liquid iron should not deviate far from ideality, provided that the concentration of solute is not too great and is expressed in proper terms. The factors that commonly lead to nonideality in systems at ordinary temperature— as ionization or the occurrence of polar and nonpolar molecules—are presumably less effective in inetal and in slag systems at high temperature. Again, the type of equation commonly used to express the value of

Jan 1, 1940

By L. S. Darken

In the study of reactions occurring in liquid iron, alone or in contact with a liquid oxide or slag phase, it has been found that the experimental data over a limited range of concentration can in some cases be represented satisfactorily by the simple equilibrium expression where (P) denotes the concentration of a product and (R) the concentration of a reactant in the reaction r1R1 + r2R2 + • • • D P lP1 + P 2P2 + • • • Some of the reactions for which it holds are listed in Table 1. For others, however, the simple equation apparently does not hold. This has been attributed to an improper method of expressing concentration (for instance, weight per cent instead of mol fraction); or it has been supposed that the solutions are not ideal in the sense that the activity of a component is not proportional to its true mol fraction. But it may be that the problem has been oversimplified by neglecting some molecular species actually present; that is, by considering only the ostensible simple components and thus leaving out of account possible compound formation. It is the purpose of the present paper to point out how, in a group of related cases, a reasonable consistent assumption relative to the presence of a compound makes it possible to express the experimental data satisfactorily by the simple equilibrium expression when this compound is taken into account, and to bring into harmony data referring to different reactions in liquid iron. There are several a priori reasons for believing that solutions in liquid iron should not deviate far from ideality, provided that the concentration of solute is not too great and is expressed in proper terms. The factors that commonly lead to nonideality in systems at ordinary temperature— as ionization or the occurrence of polar and nonpolar molecules—are presumably less effective in inetal and in slag systems at high temperature. Again, the type of equation commonly used to express the value of

Jan 1, 1940

By Karl Fetters, John Chipman

The relationship between the composition of the slag and that of the underlying metal during the refining of a heat of liquid steel may best be studied in the light of the two broad physicochemical concepts of equilibrium and reaction rate. Actual steel-making processes, typified by basic open-hearth practice, involve such complexities of operating variables that it is impossible to distinguish with any degree of certainty between the general limiting conditions imposed by chemical equilibrium and the effects associated with rates of diffusion or of chemical reaction. Further advances in the physical chemistry of steelmaking will depend upon our ability to distinguish clearly between these two kinds of effects, and the experimental work that must be done falls naturally into two classifications. The first is the study of chemical equilibrium that cannot be carried out with assurance under operating conditions but must be done in the laboratory. The second is the -subject of rates of reaction and diffusion and the mechanism of the process as a whole, which can best be studied under operating conditions. The latter is not entirely independent of the former, since its interpretation often requires a knowledge of the equilibrium state. It is the purpose of this and of several papers that will follow to present the results of labora- tory investigations of equilibria involving liquid steel and slag. Open-hearth slag is a complex mixture containing at least eight major components and perhaps twice as many whose influence may be classified as probably minor. It is not surprising that progress toward a complete understanding of such a material at high temperatures has been slow. Attempts to apply the established principles of physical chemistry to the behavior of the individual components of this mixture have been beset by an almost complete lack of data on its chemical properties in the liquid state. This has necessitated the adoption of various methods of approximation based upon such assumptions as, for example, that for a given component the ratio of its mol fraction in the slag to that in the metal is a constant, or that the molecular composition of the liquid slag is the same as its crystalline constitution after solidification. From the thermo-dynamic viewpoint, such assumptions are merely useful dodges, which enable one to hazard a guess as to the magnitude of the more fundamental quantities, the chemical potentials and activities of the components of slag. The problem of determining the activity of each substance in an eight-component svstem seems too complex to be undertaken without the aid of generalizations derived from experimental data on similar systems containing a smaller number of components. Accordingly, it was planned in the present investigation to study the equilibrium of molten iron with the simplest

Jan 1, 1941

By Karl Fetters, John Chipman

The relationship between the composition of the slag and that of the underlying metal during the refining of a heat of liquid steel may best be studied in the light of the two broad physicochemical concepts of equilibrium and reaction rate. Actual steel-making processes, typified by basic open-hearth practice, involve such complexities of operating variables that it is impossible to distinguish with any degree of certainty between the general limiting conditions imposed by chemical equilibrium and the effects associated with rates of diffusion or of chemical reaction. Further advances in the physical chemistry of steelmaking will depend upon our ability to distinguish clearly between these two kinds of effects, and the experimental work that must be done falls naturally into two classifications. The first is the study of chemical equilibrium that cannot be carried out with assurance under operating conditions but must be done in the laboratory. The second is the -subject of rates of reaction and diffusion and the mechanism of the process as a whole, which can best be studied under operating conditions. The latter is not entirely independent of the former, since its interpretation often requires a knowledge of the equilibrium state. It is the purpose of this and of several papers that will follow to present the results of labora- tory investigations of equilibria involving liquid steel and slag. Open-hearth slag is a complex mixture containing at least eight major components and perhaps twice as many whose influence may be classified as probably minor. It is not surprising that progress toward a complete understanding of such a material at high temperatures has been slow. Attempts to apply the established principles of physical chemistry to the behavior of the individual components of this mixture have been beset by an almost complete lack of data on its chemical properties in the liquid state. This has necessitated the adoption of various methods of approximation based upon such assumptions as, for example, that for a given component the ratio of its mol fraction in the slag to that in the metal is a constant, or that the molecular composition of the liquid slag is the same as its crystalline constitution after solidification. From the thermo-dynamic viewpoint, such assumptions are merely useful dodges, which enable one to hazard a guess as to the magnitude of the more fundamental quantities, the chemical potentials and activities of the components of slag. The problem of determining the activity of each substance in an eight-component svstem seems too complex to be undertaken without the aid of generalizations derived from experimental data on similar systems containing a smaller number of components. Accordingly, it was planned in the present investigation to study the equilibrium of molten iron with the simplest

Jan 1, 1941

By R. D. Springer, R. L. Wilcox, P. A. PALMARTAN, R. K. Waring, E. A. Anderson

A knowledge of the mutual liquid solubility of zinc and lead is of importance in various phases of zinc metallurgy. The determination of this solubility has been the subject of numerous investigations, but unfortunately the results vary widely, particularly at higher temperatures, and it has been impossible to decide what data are most reliable. It was decided to investigate the matter and in view of the disagreement in the published results, it was felt necessary to take rather unusual precautions to insure the reliability of the data. For this reason determinations were made by four different methods, and the procedures of several of the previous investigators were duplicated. The previously used methods can be divided into two classes: methods involving freezing the entire melt and taking samples of the lead-rich and zine-rich phases in the solid state,l,2 and methods in which the samples are taken while the bath is liquid and at the desired temperature.3,4 Data obtained by methods involving freezing the entire melt show much lower solubility of lead in zinc than methods in which the liquid is sampled at temperature. The latter method was used by Spring and Romanoff5 and by Hass and Jellinek.6 The results of these two investigations are in agreement at temperatures near the melting point but at higher temperatures Hass and Jellinek find a higher mutual solubility. The low-temperature section of the zinc-lead equilibrium diagram is given by Peirce and Anderson.7 The diagram is based on the data of Hodge and Heyer,8 Arnemann,9 and Peirce.10

Jan 1, 1934

By F. Keeler, C. M. Craighead

Aluminum alloys containing magnesium and silicon are susceptible to strengthening and hardening by suitable heat-treatments, and they constitute a class of alloys of considerable commercial importance. Generally, they contain these two elements in the proportions required for the formation of the intermetallic compound Mg2Si, or an excess of silicon; no commercial wrought alloys are made with an appreciable excess of magnesium. The reason for not having excess magnesium is obvious from the results given in this paper, since it is shown that excess magnesium results in a proportional decrease in the solubility of Mg2Si. This is the eighteenthl of a series of papers from the Aluminum Research Laboratories dealing with the equilibrium relations in aluminum-alloy systems; most of the previous publications concerned binary aluminum-alloy systems; this is the fourth of a group pertaining to ternary aluminum-alloy systems. The work on this system was completed in 1930, but the results were not published because of the pressure of other work. Hanson and Gaylor2, in 1921, presented the results of an extensive investigation of the aluminum-magnesium-silicon system. Their results indicated that an excess of magnesium over the amount required to form the compound MgzSi caused a noticeable decrease in the solid solubility of MgzSi in aluminum. A more recent investigation3 in these laboratories showed that the solid solubility of MgzSi in aluminum was 1.85 per cent at 595' C. and decreased to 0.30 per cent at 300' C. The results of the latter investigation were used as the basis for the present work. The Present Investigation In the present investigation, a study was made of the effect of various amounts of excess magnesium upon the solid solubility of MgzSi in

Jan 1, 1936

By W. L. Fink, L. A. Willey

This paper is the nineteenth of a series from the Aluminum Research Laboratories, presenting the results of the investigations of equilibrium relations in aluminum-base alloys made from electrolytically refined aluminum. In previous papers equilibrium relations have been determined and correlated for 12 binary systems. Equilibrium relations have also been determined for portions of the aluminum corners of three ternary systems (aluminum-iron-silicon, aluminum-magnesium-silicon and aluminum-copper-magnesium). The object of the present paper and those to follow is to increase our knowledge of the aluminum corner of the more important ternary systems. In 1926 Dix and Keith † suggested a new method for designating the phases encountered in alloy systems. The system has been used to a very limited extentl,2,3 The system has since been slightly modified and will be used in this and future papers. In the system as modified, a phase is designated by placing in parentheses the chemical symbols of the elements essential to the formation of that phase. The symbols in the parentheses are separated by dashes in order to clearly distinguish them from a chemical formula. Elements that are not necessary to the formation of the phase in question, but which may be present in solid solution, are not indicated in the designation. If two or more phases would have the same designation according to these rules, they are distinguished by prefixing Greek letters, and are definitely identified by X-ray diffraction patterns. Previous Investigations The solid solubility of magnesium in aluminum has been well established by the work of Dix and Keller4, Schmid and Siebe15, and Saldau and Sergeev6. There is no agreement as to the composition of the magnesium constituents at the aluminum end of the binary aluminum-magnesium system. Hanson and Gayler7 report Mg2Al3; Boyer8 found that the alloy

Jan 1, 1937