Search Documents

By Holbrook G. Botset

Experiments performed in this laboratory on the flow of gas-liquid mixtures through unconsolidated sands have been described and discussed in an earlier paper.4 In these earlier experiments a definite relationship was found between the liquid saturation of a sand and its permeability to the liquid or gas phase. Furthermore, the nature of the sand, the permeabilities of the sands over a range between 17 and 260 darcies, appeared to have a negligible effect upon the saturation-permeability relation. Likewise experiments indicated that the viscosity of the liquid, or the substitution of an immiscible liquid for the gas phase, had only a very slight effect on the permeability-saturation relation. This conclusion has since been definitely confirmed by the experiments of M. C. Leverett3 on the flow of oil and water through unconsolidated sands. The sands used by Leverett were of very low permeability (between 6.8 and 1.04 darcies), while the oil-water viscosity ratio varied between 90 and 0.057. Since the sands that constitute petroleum reservoirs are largely consolidated, although of widely varying degrees of consolidation, it was desirable to extend the experiments on gas-liquid mixtures to consolidated sand. In the study of consolidated sands additional variables are involved, but for the initial experiments the only new variable to be introduced was the cementing material. Experimental Procedure A comparatively limited assortment of sandstones was available for the selection of the experimental material. Of these, the Nichols buff sandstone appeared to be most suitable. It gives no visual evidence of any bedding planes, it has a firm structure, permitting easy handling, and its permeability is about 0.5 darcy. Its only disadvantage is the fact that it contains an appreciable amount of ferric oxide which, for experiments involving water, will hydrate and reduce the permeability. This

Jan 1, 1940

By R. A. David, D. J. Vink, D. L. Katz

A phase diagram showing boundary curve and quantity of liquid in the two-phase region was determined for a mixture of natural gas and natural gasoline in the region of its critical conditions. The temperatures and pressures of phase measurements were in the range of 85 to 212 F. and 1300 to 2600 Ib. per sq. in., respectively, with the critical conditions at 169.5 F. and 2615 Ib. per sq. in. abs. Color phenomena were observed in the region of the boundary curve from 102' to 192 F. Approximate densities of the single-phase and two-phase regions and analysis of the system are included. Boundary curves between the single-phase and the two-phase regions have been determined for several binary mixtures containing hydrocarbons. 2,3,4.6,7,10,11,12,14 These investigators did not report the relative amounts of liquid and vapor within the two-phase region. The equilibrium values obtained from the dew-point and bubble-point data can be used to compute the lines showing percentage of liquid on the pressure-temperature phase diagram8 but the positions of the lines near the critical temperature and pressure are uncertain. Relative amounts of vapor and liquid were obtained for a gasoline and a naphtha1 within the two-phase area but not in the region of the critical temperature and pressure. The bubble-point and dew-point lines, as well as the percentage liquid lines within the two-phase region, were determined by visual observations in glass apparatus. Apparatus An arrangement of the apparatus is given in Fig. 1, showing the Jerguson gauge A with glass windows B. Pressure was supplied by the

Jan 1, 1940

By John E. Sherborne

Much valuable scientific research has been performed in recent years on the subject of phase behavior of hydrocarbons.l-11 Engineers employed in petroleum production are interesting themselves in this work as well as in methods of applying the fundamental data available to the solution of their various problems. Recently a number of papers have been published in which applications of phase behavior have been made to specific cases pertaining to critical phenomena.l2-16 Little effort, however, has been made in the literature to show the relation between changes occurring in the critical region and the more common phase behavior, therefore it is believed that a presentation of the fundamentals of phase behavior with reference to hydrocarbons is timely. Study of phase behavior is not new. In the metallurgical field, knowledge of heterogeneous equilibria has advanced tremendously, particularly with reference to solid-solid and solid-liquid behavior. Much is known about vapor-liquid equilibria too, but few engineers are familiar with this subject. Definition of Terms In a discussion of this sort, a definition of terms used is most important. Such terms as "pressure," "temperature" and "volume" need little definition other than mention of the units in which they are considered. Pressure is expressed in pounds per square inch absolute. Temperature is usually expressed as degrees Fahrenheit or degrees Rankine. In considering thermodynamic and phase behavior, the use of the absolute, Rankine, scale is desirable. Volume is expressed as specific volume, such as cubic feet per pound. This will be recognized as the reciprocal of the specific weight. In considering systems composed of more than one component, it is sometimes desirable to consider molal

Jan 1, 1940

By B. H. Sage, W. N. Lacey

An apparatus is described for the measurement of the pressure-volume-temperature relations of pure substances, simple mixtures and complex mixtures with an over-all absolute uncertainty, which is usually not more than 0.2 per cent. The equipment is suitable for studies in the gaseous, two-phase, and liquid regions. The behavior of such systems may be investigated at pressures as high as 10,000 lb. per sq. in. for temperatures between 70 and 600 F. The methods of measurement and the construction of the apparatus are described in some detail. The measurement of the volumetric behavior of multicomponent systems requires the simultaneous measurement of the pressure, volume, temperature and weight of each component of the system. In the present apparatus a sample having constant weight is investigated at a number of independently controlled pressures and temperatures, therefore it is necessary only to establish the weight of each component added before a given set of measurements. This method is in distinction to that in which the total volume remains constant and the pressure is varied by addition or removal of material, the system having variable weight. An objective in the design of the apparatus to be described was to develop equipment that would yield results involving no uncertainties larger than 0.25 per cent throughout the range of pressures from 150 to 10,000 Ib. per sq. in. at temperatures between 70" and 600" F. Although this over-all absolute uncertainty of measurement may not appear to be small, it necessitated care in the plans for control and measurement of each of the fundamental variables. In general, any uncertainty greater than 0.05 per cent in the pressure, volume, absolute temperature or weight of each component was considered undesirable; therefore the attempt was made to obtain a precision of measurement in each case such that the absolute uncertainty would be only half this amount except in certain regions near the ends of the experimental range of the variable under consideration.

Jan 1, 1940

By B. H. Sage, W. N. Lacey, W. M. Morris

The volumetric behavior of isobutane at temperatures below its critical temperature has been studied by several investigators. Seibert and Burrelll measured the vapor pressure of isobutane from the ice point to the critical temperature. Dana and co-workers2 determined among other thermodynamic quantities the density of the saturated liquid and the saturated gas and the vapor pressure from approximately 0 to 130 F. Sage and Lacey3 studied a number of the thermodynamic properties of this hydrocarbon from 70" to 250" F. The specific volume in the gas and liquid region was determined at pressures up to 3000 Ib. per sq. in.? throughout this temperature interval. In addition, the vapor pressures were measured at 30" intervals. These data taken together serve to establish the thermodynamic behavior of isobutane at temperatures between 0" and 250" F. with reasonable accuracy. However, apparently there are no data available at the higher temperatures. For this reason an investigation of the volumetric behavior of isobutane at temperatures between 100" and 460" F. was carried out with apparatus that has been described in another paper (p. 136, this volume). Materials The isobutane used in this investigation was obtained from the Philgas Division of the Phillips Petroleum Co. That company's special analysis upon this sample indicated it to contain less than 0.03 mol per cent of impurities. In order to avoid traces of noncondensable gases, this hydrocarbon material was subjected to two successive fractionations at a high reflux ratio in a column approximately 4 ft. long and % in. in diameter, which was packed with small glass rings. The middle portion of distillate from the first fractionation was used in the second fraction-ation and a portion of the overhead from this second distillation process was repeatedly condensed at liquid-air temperatures at a pressure only slightly above the sublimation pressure of isobutane at this temperature.

Jan 1, 1940

By H. T. Byck

The plastering properties of six representative California drilling muds were studied over a temperature range of 70 to 175 F. at several mud weights, using a high-pressure circulating filter press with full size consolidated cores. In all tests, with untreated as well as with chemically treated muds, the filtration rates at elevated temperature were higher than at low temperatures and in nearly every case this increased filtration rate was notably larger than could be attributed alone to the decreased viscosity of water—the flowing medium—at elevated temperatures. It is concluded that no existing method will permit even an approximate determination of the filtration rate at high temperature from data at room temperature. It is necessary to measure filtration at the temperature actually anticipated in the well, or to make a sufficient number of tests at various lower temperatures so that a small extrapolation of these data to the anticipated well temperature may be applied. Such tests should be carried out on the mud as actually used in the well. The viscosity of each of the muds was determined over the same temperature range. While no relationship was observed between mud viscosity and plastering properties, it was found that chemical treatment of the muds for viscosity reduction shifted the temperature of minimum viscosity to appreciably higher values than for untreated muds. Importance of Plastering Properties A survey of the literature on rotary drilling muds clearly shows the growing realization of the importance of plastering properties,'-5 not only from the point of view of the successful completion of the mechanical operation of drilling the well but also from the point of view of minimum impairment of the natural flow possibilities of a given well. Poor plastering properties with attendant high filtration rate (large water loss to the formation) and thick mud sheath, lead to stuck drill

Jan 1, 1940

By K. C. Ten Brink, Allen D. Garrison

A rrevious paper1 reviewed some of the properties of clays and shales and presented some data on the nature of the gelling phenomenon. It included a brief discussion of origin of clays and shales, their composition, structure, retention of water, ionization, base exchange, and the nature of the gelling of water suspensions. Other publications of the Institute2-8 have discussed the properties of clays and their relations to drilling-fluid behavior and control. It is the purpose of this article to present some new data obtained and arranged in a different manner and making possible a better understanding of the chemical reactions involved in the control of drilling-fluid characteristics during the normal drilling operations. Experimental Procedure The difficulties involved in the measurements of the flow characteristics of plastic fluids are too familiar to recite. The possible applications of some of the well-known instruments to this problem have been adequately discussed by Hobson9 and he has given an excellent bibliography. Although there are many viscosimeters of varied design, none of them appear to fulfill the most obvious demands of accuracy in the measurement of clay or shale suspensions. Since the fluid characteristics are dependent on the rate of flow of the fluid, and the viscosity usually falls off very rapidly as the rate is increased, it is clear that the first requirement is that the instrument shall be capable of a wide variety of speeds. The viscosity of the fluid at any fixed and constant rate of flow is also a function of time, so that the second demand is that the instrument shall be capable of a constant and continuous motion for a prolonged period of time.

Jan 1, 1940

By P. B. Elkin, K. C. Ten Brink, Allen D. Garrison

It is the purpose of this article to present the results of an investigation of certain clay and shale suspensions having viscosities that are materially influenced by fluid motion, and to discuss the bearing of these results upon the choice of policy in the control of drilling fluids. A previous article described the design and the characteristics of an instrument adapted to measure the viscosity of plastic suspensions during continuous uniform motion and at a variety of speeds.' It has long been recognized that the viscosity of a clay suspension depends on the past history of the fluid motion. Since an abundance of literature is available on this phenomenon,2 it is not necessary to describe it in more than general terms. It is usually observed that rapid motion reduces the viscosity of a clay suspension, and that some time is required for the viscosity to fall to its equilibrium value after the steady motion is started. On reducing the speed of motion to some lower value, the viscosity again rises, but it also requires some time to attain a steady viscosity value at the slower rate of motion. Thus clay suspensions are usually observed to be increasing in viscosity as time passes when the measurement has been immediately preceded by fluid motion that was more rapid than that employed in the measurement. And it is usually observed that the viscosity falls as time passes when the measurement has been preceded by more quiescent conditions. Reed,3 and Broughton and Hand4 have reported a condition of observation of clay suspensions in which the viscosity was observed to fall during the first few moments of steady motion, and rise thereafter. The characteristics of the new instrument are particularly well adapted to studies of this kind. With it we have been able to observe all of the phenomena described above, and it is important to understand their nature because they all occur at some time during normal drill-

Jan 1, 1940

By John D. Gill

The purpose of this paper is to invite attention away from the obvious. direct monetary costs of oil-field overdevelopment (as distinguished from "social costs") to a consideration of the role played by the exercise of freedom to create even allegedly excessive1 plant facilities in contributing to the maintenance of equilibrium in the wholc economy and in promoting material progress. The obvious and most commonly discussed effect of overdevelopment of oil fields is illustrated by the allegation that already the excessive development of the East Texas field has cost upward of $162,000,000;2 in other words, the drilling of 12,500 wells more than the number estimated as the requirement for economical recovery of the reserve has unnecessarily raised the total cost of development that amount. Many other illustrations of so-called excessive development of oil pools could be presented. Some of them, like Cushing, Powell and Long Beach, might be drawn from the era antedating widespread discussion in the trade of such matters as optimum well spacing, reservoir energy, the conservation rate of depletion of reserves, etc. Numerous other cases can he found in the records of the current era, which began with the introduction of rationalized restriction of production. Besides East Texas, there may be mentioned Oklahoma City, Centralia and. more recently, the town-lot development of Bloomingdale, Michigan.

Jan 1, 1940

By G. A. Beiswenger

The application of the law of supply and demand to the sale of crude oil is generally conceded, but the motives underlying the buyer's (refiner's) demands are not always obvious to the seller (producer). In some cases this obscurity has led to the conclusion that the refiner may be unnecessarily particular when he selects one crude and rejects another. This preference, however, is very largely the reflection of the refiner's product demands and the quality of the stocks obtainable from the selected crude. The purpose of this paper is to point out critical characteristies of the major classes of refined products and to show how crudes vary in their suitability for such products. With relatively few exceptions, available processing methods and addition agents make it theoretically possible to produce any type of refined product from any crude source, although it may not be economically sound to do so. The widespread application of thermal cracking for converting higher boiling stocks into gasoline is well known. More recently cracking and refining in the presence of catalysts9 have been developed, as well as various processes for polymerizing lighter hydrocarbons into gasolines having desirable characteristics.6 For several years, selective solvents have been applied commercially to refining inferior quality kerosene, Diesel fuel and lubricating oil distillates.3 Improved methods for refining lubricating oils from crude residual4 formerly suitable only for fuel, are now in use and a number of special solvent processes are available for dewaxing lubricating oils to low pour point.5 These improved refining methods have been developed not only to meet the demand for improved product quality but also to enable the refiner to obtain quality products from crude sources that would not yield such products by the older refining methods. In many cases, the utilization of these modern refining tools necessitates an appreciable expenditure for equipment as well as added manufacturing costs, and their application therefore must be justified economically. Despite the availability of such tools, the added processing costs involved and yields obtainable may render their application entirely impractical. To illustrate this point, inspections of typical cuts from various crudes

Jan 1, 1940

By Norman D. Fitzgerald

The growth of domestic requirements for motor fuel has been phenomenal, rising year after year in a fashion almost unique among commodities, resisting depressions and forging rapidly ahead in times of prosperity. Consumption advanced from 226 millions of barrels in 1925 to 522 millions of barrels in 1938; or 130 per cent. The economic influences generating and sustaining this demand are of vital concern to the petroleum industry. The general factors affecting utilization of motor fuel are well known but it is the purpose of this paper to employ the recently developed technique of economic dynamics to describe them mathematically. Such problems in economic analysis are complicated by the freedom of all the variables to fluctuate, their mutual interdependence, changing incidence upon one another, and the nonquantitative character of some of them. Economic quantities and relationships, unlike those of the physical world, are continuously changing and this dynamic characteristic is both a distinguishing and confusing feature in economic research. The complexities of the problems require application of the analytical devices of mathematical statistics to uncover the probable relationships between factors. The procedure employed is to formulate a logical relationship linking the dependent to the independent variables, and then simultaneously, rather than sequentially, determine their separate effects by multiple correlation. This transformation of common knowledge to quantitative terms makes it possible to measure the probable effects of changing conditions upon requirements, assisting in the sound interpretation of economic policies. The utility of the results of such dynamic analysis is in sharp contrast to the sterility of the trends and other simple empiricisms of descriptive statistics. It is not intended in this study to throw light on the reasons why any motor fuel at all is consumed; but to discover relationships explaining why more was consumed at one time than at another. Specifically the problem is confined to the measurement of the influences of statistically

Jan 1, 1940

By V. R. Garfias, R. V. Whetsel, J. W. Ristori

World consumption of petroleum and related fuels for civil purposes in 1939 was close to 2,024,000,000 bbl., or approximately 110,000,000 bbl. more than in 1938, the United States accounting for 96,000,000 bbl. of the increased amount. With the annexation of Czechoslovakia and part of Poland, civil consurnption in Germany during 1939, notwithstanding sevre rationing in the last four months of the year, was 54,000,000 bbl., or only 1,600,000 bbl. less than in 1938, and domestic production of pctroleum and related fuels was some 34,000,000 bbl. short of civil consumption. The mounting petroleum requirements of the military forces of greater Germany are, of course, not included in these figures of civil consumption. To supply this deficiency Germany now must rely almost exclusively on the Rumanian fields, which in 1939 produced about 35,000,000 bbl in ewes, of the Rumanian civil consumption. However, these fields arc supplying and have supplied for years the oil requirements of the neighboring

Jan 1, 1940

By James Terry Duce

In order to facilitate interpretation of the data in this chapter, we print the following excerpts from the Circular to Authors. The field is the unit in this tabulation. In cases of fields extending across State boundaries, such as Rodessa, it is suggested that each State author treat the section of the field in his State as a unit, and by a footnote indicate that the field extends into an adjacent State. Each space in Table 1 may represent one of four possibilities; either it is not applicable to the particular field, or the proper entry is not determinable, or the proper entry may be determinable but is not determinable from data available to the author, or the proper entry is determinable. When spaces are not applicable, leave blank; in spaces where the proper entries are determinable from data available to the author, please insert y; in spaces where the proper entries are determinable by the author, make such entries. y implies a hope that in some future year a definite figure will be available. The entry of an 0 is a positive declaration, and an important declaration where it is in order. Inability to determine precisely the correct entry for a particular space should not lead the author to insert merely y. Contributions of great value may be made by the author in many cases where entries are not subject to precise determination. In such cases the author should use his good judgment and make the best entry possible under the circumstances. For many spaces the correct entries represent the opinion of the author (for example, "Area Proved") and in such cases the entry need not be hedged to such extent as when the quantities are definite yet can be ascertained only approximately by the author. In cases under definite headings but where figures are only approximate, the author may use x. It is thought that the nearest whole numbers are sufficiently accurate for our purposes except as to percentage of sulphur in oil. If an author desires to report any other figures to tenths, he may do so. The quantity of gas produced should include, where possible, gas sold or otherwise marketed, and gas blown into the air, burned as flares or otherwise wasted. Segregation of these figures would be interesting if the authors can make such segregation. In any event, the figures should represent as nearly as possible the total quantity of gas removed from the reservoir. Under the columns on "Depth," the average depth to the top of the productive zones and to the bottom of the productive wells, when subtracted, should give the approximate thickness of the productive zone. For fields where this is not true

Jan 1, 1940

By Warren B. Weeks

Continued development in the deeper fields discovered during 1937 and 1938 was largely responsible for the 16 per cent (2,913,400-bbl.) increase in production, bringing the 1939 output to 21,376,230 bbl., the largest for South Arkansas since 1929. The older settled fields continued to decline and Rodessa was added to this classification. Magnolia field accounted for much of the increase, with Schuler field showing only a slight increase. There were 248 wells drilled during 1939, five more than during the previous year. Of 198 wells drilled in proven fields, 21 were dry; 46 of the 50 wildcats drilled were dry. The old depth record of 8332 ft. was increased to the present record of 9028. Trend of Prospecting Production from the porous " Reynolds " oolite zone of the Smackover limestone continued to be the objective of prospecting. During the year 20 seismograph, 3 gravity meter, 1 magnetometer, 1 soil analysis and 1 electromagnetic crews were reported in the state. The geophysical work was largely concentrated in three southern counties: Columbia, Lafayette and Union. As indicated in Table 2, the same three counties received most of the important drilling tests. At the end of the year only three seismograph crews remained in the areas. The area for prospecting Smackover limestone was greatly narrowed during the year by the failure of two "fault zone" wells to find satisfactory permeability in the oolite zone, and by the relatively thin sections of porosity found in wells drilled in townships 18 and 19 South. Geology Most of the wells listed in Table 2 afforded additional control on structure of the lower beds, but gave little new information on stratigraphy. Of the 248 wells drilled, 3 ended in Tertiary, 34 in Gulf, 1 in Paluxy, 29 in Glen Rose, 15 in Travis Peak, 15 in Cotton Valley, 139 in Smackover, 2 in Eagle Mills and 10 in Carboniferous. The age of the beds between the Travis Peak formation and Carboniferous is still questionable. Most

Jan 1, 1940

By V. H. Wilhelm

Drilling activity in California during 1939 decreased approximately 15 per cent from that of the previous year. New reserve discoveries were the lowest in many years, but extensions, development of deeper sands, and reappraisals, indicated enough new potential to more than equal 1939 withdrawals of about 223,644,000 barrels. On June 1 a new uniform method of allocation was adopted, which had the immediate effect of reducing production, so that approximately 6,000,000 bbl. of crude were withdrawn from storage during the year. Market demand averaged 628,845 bbl. per day as compared with a daily production of 612,945 barrels. The rate of drilling was slow during the first five months because of the unfavorable storage situation. Subsequent to the adoption of the new uniform method of allocation on June 1, drilling was speeded up and continued at a rapid rate throughout the remainder of the year. As a result of prospecting, four new fields were discovered and seven fields were measurably extended. All of the new discoveries were made in the San Joaquin Valley. Statistics of production and development arc given in Fig. 1 and Tables 1 to 3. New Discoveries Coalinga-Northeast.—The Amerada Petroleum Corporation discovered a second Eocene shore-line field north of the Coalinga Nose when it brought in well S.P.L. 7-17 in sec. 17-19-16, flowing 4230 bbl. of 32.4" gravity oil and 14,200,000 cu. ft. of gas at a depth of 8237 ft., from the "Gatchell" sand of Eocene age. This field, 2 miles north of the Coalinga Nose field, is considered to be a separate structure because it has a gas cap and a lower water table. The structure is a broad, gentle anticlinal nose probably separated from the southerly Coalinga Nose field by a shallow syncline. Production is sealed updip by the pinching out of oil sands. The average sand thickness is approximately 200 ft., the top 75 ft. containing mainly gas. The sand has an average permeability of 1000 millidarcys and an average porosity of 25 per cent. The productivc area of the field is estimated to be between 1200 and 1500 acres.

Jan 1, 1940

By Alfred H. Bell, George V. Cohee

In 1939 Illinois produced 94,302,000 bbl. of oil—almost three times the amount of oil produced at the peak in 1908, when development in the southeastern Illinois field was at its height (Fig. I). It represents nearly a fourfold increase over the 24,075,000 bbl. produced in 1938 (Table 6). This rapid increase in oil production in Illinois is largely due to development in the Salem and Louden fields (Table I). During 1939 daily production for the state increased from 135,000 to 332,572 bbl. A notable increase in production for 1940 is expected as a result of the drilling of the Devonian limestone in the Salem and Centralia fields,

Jan 1, 1940

By G. F. Fix, R. E. Esarey

During 1939, Indiana experienced a grcat increase in prospecting and drilling for oil and gas. Major activity, as during the preceding year, was in the southwestern part of the state, the Indiana portion of the Eastern Interior Coal Basin. An increasing interest was evident during the year, however, in the entire state, chiefly in northern Indiana (on the south flanks of the Michigan Basin) and in the old Trenton field of east central Indiana. In the latter area several tests to deeper horizons in the Trenton and to the underlying St. Peter sandstone were completed. Results in this area have been discouraging, however, since these deeper horizons are very erratic. The regional structure in the Indiana portion of the Michigan Basin is generally favorable to oil accumulation, but the picture is complicated by a uniform mantle of glacial drift and a lack of subsurface information. Several wells were started during the last half of 1939 to test the Devonian and Trenton formations in this area. Production from these formations was found at several localities many years ago, hut, no sustained effort has been made to locate pools in this area, which covers perhaps 20 counties. There were 377 wells completed in Indiana during 1939, and 77 others were in various stages of drilling at the close of the year. Of the 377 completed, 255 were classed as field locations and 122 as wildcats. The field wells were divided as follows: 156 oil wells, 39 gas wells, and 60 dry holes; the wildcats: 98 dry holes, 19 oil wells and 5 gas wells (wildcat oil and gas wells include deeper production in oil fields and field extensions as well as new field discoveries). During the year 218 more wells were completed than during 1938. Outstanding developments for 1939 include the discovery of several new oil and gas fields. In the Griffin field, which was discovered during the closing days of the previous year, 103 wells were drilled, including field extensions. Of these, 97 were productive oil wells. This field now covers more than 2000 acres of productive area and has not as yet been defined. There are five producing horizons in the field, including the Waltersburg, Tar Springs, Elwren, and Aux Vases sandstones of Chester age and the McClosky horizon of the Ste. Genevieve limestone. Produe-

Jan 1, 1940

By Frederick N. Rhines

Pure copper that has been allowed to oxidize at an elevated temperature in the air is found to be covered with two distinguishable layers of oxide scale. The outer of these, which is very thin, is composed of cuprie oxide, while the inner layer, comprising the bulk of the scale, is made up chiefly of cuprous oxide (Fig, la). If a small quantity of an alloying element, such as silicon, is present in the copper, silica particles are found disseminated through the cuprous oxide, deposited most abundantly near the metal surface, and, in addition, silica particles occur to a limited depth within the copper itself (Fig. 1b). It is this inner layer, composetl of oxide particles precipitated in a matrix of metallic copper, that is known as the "subscale" or the zone of internal oxidation. Under a low oxygen pressure it is possible to produce the subscale without the formation of external layers of copper oxides, thus leaving the metal surface bright and apparently clean while oxidation has occurred, in fact, to a considerable depth within the metal (Fig. 2). Although little is yet known in a detailed way of the mechanism of internal oxidation, the general nature of the process is evident. By virtue of the considerable solubility of oxygen in copper,' it is possible to bring the alloying element in contact with oxygen within the metallic solid solution. If the oxide of the alloying element is more stable than cuprous oxide, the foreign oxide may form and, since most oxides are relatively insoluble in copper, it will not long remain in solution, but mill separate as a precipitate. As the metallic solute is exhausted in the zone of precipitation, oxygen that is constantly being supplied at the surface diffuses farther into the alloy and thus produces an ever-thickening shell of subscale. The distribution of the reacting elements in the subscale and adjoining zones must be somewhat as indicated in Fig. 3. The conditions requisite to internal oxidation to be anticipated are thus: (I) a solubility of oxygen in the alloy, (2) the ability of oxygen to diffuse at an appreciable rate into the alloy, (3) the possibility of the

Jan 1, 1940

By L. L. Wyman

The resultant embrittlement caused by the exposure of oxygen-bearing copper when hot and exposed to reducing gases has been the subject of many studies.' Little attention, however, has been given to the possible embrittling effects due to atmospheres or conditions that under some circumstances might be the source of a reducing gas. At rather infrequent intervals there occur reports of copper becoming embrittled during normal annealing cycles in commercial steam-atmos-phered furnaces such as customarily are used for this operation. As might be expected, several other factors exist—such as oil, dirt on the wire, or in the furnace, or metallic surface contacts—that might produce reducing gases and be the source of the trouble. However, in consideration of the statement made by Ruder,2 it would seem that under some conditions steam itself might prove to be the source of trouble. A consideration of the amount of the dissociation of water vapor up to temperatures of the melting point of copper shows that3 the amount of hydrogen that could come from this source is far too minute to account for any embrittlement. What such information does not tell is what will happen at the surface of copper when subjected to steam at elevated temperatures, for entirely different conditions may persist, which would account for Ruder's observation. Furthermore, there are no data as to what might happen at the surfaces of other materials present, the resultant of which might affect the condition of the copper present. For this reason, it was decided to make a series of experiments comparable to those previously conducted by the writer,' which would clarify the effect of the steam on copper, and in copper in the presence of steel. Material The material used in these experiments was tough-pitch copper wire of rectangular cross section, about % in. thick, picked at random from the regular factory production of material used for windings, etc., in the electrical industry. Chemical analysis of this material gave the follow-

Jan 1, 1940

By Frederick N. Rhines, Robert F. Mehl

Relatively few data have been collected on the rates of diffusion in ternary solid solution systems. In general it does not seem worth while to gather extensive data for such systems until the factors that determine the rates of diffusion in binary systems are better understood. But the rates of diffusion in a ternary system of limited solid solubility, in which the terminal solid solution is in equilibrium with an intermediate phase and may thus be represented as a quasibinary system, represents a case of special interest. In this case, where by reason of the association of the two solute metals in an intermediate phase a special attraction of the two solute atoms may be safely assumed, some interaction of these atoms during simultaneous diffusion in the solvent lattice might be expected. It has frequently been proposed that molecular solid solutions exist in quasibinary and in other ternary (and binary) solid solutions. There is no direct evidence for such solid solutions; arguments based upon the supposed anomaly in certain physical properties1 are hardly important; unexpected density relationships have proved deceptive;2 suggestions of molecular solid solutions have been made on the basis of solution laws in their application to certain systems, but without result;3 attempts to find lattice parameters in the solid solution Al-MgZnz, which cannot be calculated from the separate effects of magnesium and zinc and might therefore be taken as evidence for a molecular aggregation, have given negative results.* Fundamentally, methods for the determination of the state of molecular aggregation in solution depend upon the kinetic behavior of the particles constituting the system. If it could be shown that the two solutes in a quasibinary ternary system diffuse in molecular proportions, despite a radical difference in the separate rates of diffusion of each solute when diffusing alone in the solvent lattice, some basis for the concept of molecular solid solution would be afforded. Freche5 examined the rates of

Jan 1, 1940