Search Documents

By C. W. Phillips, D. N. Frey

A commercial Ti-Fe-Cr alloy, Ti-150, exhibits a martensitic transformation on cooling and two nucleation and growth reactions, one above and one below the Mg-Mf region, on isothermal holding below the single-phase ß temperature range. The reactions were followed by metallographic, X-ray diffraction, and microhardness methods. THE rapidly expanding engineering application of titanium and its alloys makes increasingly urgent an adequate knowledge of the effect of heat treatment on this material. The work herein reported is an attempt to determine certain aspects of the behavior of a commercially available alloy, Ti-150, during heat treatment. The alloy, with a nominal composition of 2.7 pct Cr, 1.3 pct Fe, 0.25 pct 0, 0.02 pct N, and 0.02 pct C, was received in the as-rolled condition as a 3/4 in. round. Slugs 0.11 to 0.12 in. thick were cut under a coolant using a silicon carbide wheel. Preliminary work established that the alloy consists of two phases, a (hexagonal close-packed) and /3 (body-centered cubic), at temperatures below 1750" to 1775"F, and of only /3 phase at higher temperatures. The slugs were heated for 1 hr in air at 1800°F to form a single-phase /3 matrix, then transferred to molten tin (1250° to 1650°F) or lead (1150°F and below) for isothermal holding. Holding temperatures ranged from 700° to 1650°F, times from 0.001 to 23 hr. The treated slugs were then sectioned, one part for metallographic examination and diamond pyramid hardness, the other, after deep etching to remove the high oxygen surface,' for comparative lattice parameter determination and X-ray phase identification. Hardness measurements were taken on the lightly etched surfaces of the metallographic specimens; a 1000 g load on a 136° diamond pyramid was used. Five impressions on each sample were made and the average of the hardness calculated. The use of microhardness measurements to follow the course of an isothermal transformation reaction is undesirable because of the possibility of wide differences in microconstituent hardness.' They were used in this case, however, because of the small amount of material available. At the higher transformation temperatures, where the above-mentioned objection was apparent, the readings were proportioned to correspond to the estimated ratio of transformed and untransformed constituents present. At lower temperatures, resulting in much finer microconstituents, the microhardness measurements were more uniform. The standard deviation of the mean of five impressions calculated from these latter values was 4.2 DPH. Employing 2a limits (probability 0.95) the experimental error of the hardness measurements averages was 8.4 DPH. The etchant used was composed of one part glyc-erol and one part 48 pct hydrofluoric acid." Of a number of etches tried, this one appeared most satisfactory. All micrographs were taken at 500X magnification and with oblique lighting to permit easier identification of the acicular structure resulting from rapid cooling. No advantage was gained by using polarized light for metallographic examination. Phase identification and lattice parameter measurements were made with a North American Philips high angle spectrometer in conjunction with a recording potentiometer. Considerable difficulty was encountered with some of the lattice parameter measurements because of the extreme broadness of the few diffraction lines obtained from samples consisting mainly of strained, untempered, quenched structures.

Jan 1, 1953

By C. L. Sollenberger, Will Mitchell, Ford F. Miskell

Heat treatment of ore prior to comminution reduces power requirements for grinding, reduces grinding media wear, and improves recovery of iron values from a typical Minnesota magnetic taconite. Test data demonstrating this, as well as an analysis of the economics of commercial application of the technique, are presented. THE taconites in general are hard, tough ores, difficult to grind. Liberation of iron mineral constituents usually is accomplished by grinding the ore through at least 100 mesh, and often it has been found necessary to grind substantially through 325 mesh to achieve satisfactory recovery and grade in the concentration process. Because of the fineness of grind required and the enormous tonnages of material contemplated for treatment in the future, costs resulting from grinding media wear and power consumption, together with capital investment required for comminution, approach astronomical figures. Economy in any one of these elements per ton of material ground could very well reflect a considerable saving to the ferrous industry in yearly costs. With this in mind, the Research Laboratories of Allis-Chalmers Manufacturing Co. have launched a program to discover means of effecting this economy. The initial phase of the work as described here deals with a heat treatment of crushed raw ore, followed by thermal shock in cooling, for the preparation of rod mill feed. Several investigators1-" ave noted an improvement in the grindabilities of ores treated in this manner. One investigator2 subjected low grade iron ores to heat treatment in an electric furnace. Basing his conclusions on screen analyses of crushed products, he observed that treated ore was more easily crushed than the untreated. He reported that observation of the treated and untreated ore through a microscope revealed cracks following the grain boundaries in treated ore, whereas no cracks were present in untreated ore. However, few if any have quantified the improvement in terms of total hp hours saved, reduction of wear of grinding media, or reduction of capital costs of grinding equipment involved. By means of Bond rod mill and ball mill grindabilitiesh nd by comparative wear tests, the conditions maintained during this investigation have been definitely evaluated. Heat treatment experiments have been made on a batch scale in a muffle furnace, followed by continuous scale experiments in a rotary kiln under various conditions of temperature, atmosphere, retention time, and quenching to determine the combination that would give the greatest improvement in the grindability of a taconite without affecting adversely the magnetic susceptibility of the magnetite. From these data, relative costs have been calculated for grinding both untreated and heat- treated ore on a basis of plant capacity of 120 tons per hour. Heat requirements for a rotary kiln, as well as the kiln capacity required for the treatment, have been estimated. For these tests 50 tons of ore were obtained from the Aurora mine of the Erie Mining Co., Hibbing, Minn. The ore was reported7 to have been selected from the magnetic portion of the lower chert horizons. An average chemical analysis of' this portion is shown in Table I and an approximate mineralogi-cal analysis is shown in Table 11. No attempt was made to differentiate between the various silicates because the compositions vary widely on the different horizons and depend on the degree of oxidation. Minnesotite, however, is the predominant silicate. The siderite grains were dispersed throughout the ore but were essentially associated with the rock-forming minerals. In a series of preliminary tests in which batches of —6 mesh ore were heat-treated in a muffle fur-

Jan 1, 1953

By Will Mitchell, Ford F. Miskell, C. L. Sollenberger

Heat treatment of ore prior to comminution reduces power requirements for grinding, reduces grinding media wear, and improves recovery of iron values from a typical Minnesota magnetic taconite. Test data demonstrating this, as well as an analysis of the economics of commercial application of the technique, are presented. THE taconites in general are hard, tough ores, difficult to grind. Liberation of iron mineral constituents usually is accomplished by grinding the ore through at least 100 mesh, and often it has been found necessary to grind substantially through 325 mesh to achieve satisfactory recovery and grade in the concentration process. Because of the fineness of grind required and the enormous tonnages of material contemplated for treatment in the future, costs resulting from grinding media wear and power consumption, together with capital investment required for comminution, approach astronomical figures. Economy in any one of these elements per ton of material ground could very well reflect a considerable saving to the ferrous industry in yearly costs. With this in mind, the Research Laboratories of Allis-Chalmers Manufacturing Co. have launched a program to discover means of effecting this economy. The initial phase of the work as described here deals with a heat treatment of crushed raw ore, followed by thermal shock in cooling, for the preparation of rod mill feed. Several investigators1-" ave noted an improvement in the grindabilities of ores treated in this manner. One investigator2 subjected low grade iron ores to heat treatment in an electric furnace. Basing his conclusions on screen analyses of crushed products, he observed that treated ore was more easily crushed than the untreated. He reported that observation of the treated and untreated ore through a microscope revealed cracks following the grain boundaries in treated ore, whereas no cracks were present in untreated ore. However, few if any have quantified the improvement in terms of total hp hours saved, reduction of wear of grinding media, or reduction of capital costs of grinding equipment involved. By means of Bond rod mill and ball mill grindabilitiesh nd by comparative wear tests, the conditions maintained during this investigation have been definitely evaluated. Heat treatment experiments have been made on a batch scale in a muffle furnace, followed by continuous scale experiments in a rotary kiln under various conditions of temperature, atmosphere, retention time, and quenching to determine the combination that would give the greatest improvement in the grindability of a taconite without affecting adversely the magnetic susceptibility of the magnetite. From these data, relative costs have been calculated for grinding both untreated and heat- treated ore on a basis of plant capacity of 120 tons per hour. Heat requirements for a rotary kiln, as well as the kiln capacity required for the treatment, have been estimated. For these tests 50 tons of ore were obtained from the Aurora mine of the Erie Mining Co., Hibbing, Minn. The ore was reported7 to have been selected from the magnetic portion of the lower chert horizons. An average chemical analysis of' this portion is shown in Table I and an approximate mineralogi-cal analysis is shown in Table 11. No attempt was made to differentiate between the various silicates because the compositions vary widely on the different horizons and depend on the degree of oxidation. Minnesotite, however, is the predominant silicate. The siderite grains were dispersed throughout the ore but were essentially associated with the rock-forming minerals. In a series of preliminary tests in which batches of —6 mesh ore were heat-treated in a muffle fur-

Jan 1, 1953

By C. A. Siebert, S. H. Bush

Isothermal temper-embrittlement studies were conducted on a 5140 steel at various temperatures for times as long as 3000 hr. Specimens from the embrittled steel were subjected to impact tests, metallographic examination, microhardness, and lattice-parameter measurements. SEVERAL of the steels in the low alloy group are subject to a pronounced decrease in impact energy if held within, or cooled slowly through, a range of temperatures below the lower critical. This phenomenon, known as temper embrittlement, is quite pronounced in the chromium steels of which AISI 5140 is typical. This study considered the response of one heat of 5140 steel when embrittled isothermally at various temperatures for times as long as 3000 hr. Excellent reviews of the work done in the field of temper embrittlement have been written by Hollomonl and Woodfine.' Both authors list the various theories which have been advanced to explain temper embrittlement. Some of the mechanisms considered were: 1—a transformation below 700°C; 2—an allotropic modification of iron; 3—precipitation from the ferrite of such compounds as carbides, nitrides, phosphides, or chromium oxides; 4—decomposition of retained austenite; 5—modification of special carbides; 6—grain-boundary segregation; and 7—selective distribution of carbides. The present study investigated the changes in hardness, Meyer's number, impact properties, micro-structure, and lattice parameter which occurred on embrittlement. Previous studies considered the influence of microstructure and of prior austenitic grain size3 on the response of 5140 steel to temper embrittlement. Equipment and Procedure All samples were taken from one heat of 5140 steel; the chemical composition of this steel was: 0.43 pct C, 0.86 pct Mn, 0.020 pct P, 0.019 pct S, 0.27 pct Si, 0.84 pct Cr, 0.10 pct Ni, 0.06 pct Mo, 0.10 pct Cu, and the remainder Fe. The material was received as % in. rounds. This stock was cut down to blanks measuring approximately 0.425~0.425~2.196 in. Heat Treatment: The rough blanks were normalized at 1600°F (870°C) for 1 hr, then austenitized at 1600°F (870°C) for 1 1/2 hr, and quenched into an agitated oil bath. The specimens were tempered at 1275°F (690°C) for 5 hr and water quenched. These tempered specimens were designated as the toughened state. Embrittling Treatments: The tempered bars were divided into sets and subjected to the various embrittling treatments listed in Table I. Temperatures from 750" to 1050°F were investigated for times of 1 to 3000 hr. Special tests to determine the effect of heating a short time at a higher temperature after long-time embrittlement at some lower temperature were also conducted. All specimens were water quenched at the end of the embrittling cycle, then ground and notched to obtain standard Charpy V-notch specimens. Impact Tests: The bars were fractured in a 260 ft-lb Charpy impact unit; specimens were fractured at temperatures ranging from —315°F to room temperature. Fixed temperatures used consisted of —315°F with liquid nitrogen, —115°F with dry ice and acetone, and 32°F with ice and water. Intermediate temperatures were obtained by suspending the bars in a jig over the liquid nitrogen, dry ice and acetone, or ice and water.

Jan 1, 1955

By G. F. Sager, B. J. Nelson

Magnesium alloy forgings offer higher and more uniform mechanical properties than heat treated magnesium alloy castings and are used principally for light weight parts that may be stressed in fatigue or subject to impact loads, or where space limitations do not permit the use of the necessarily bulkier magnesium alloy castings. Compositions and Properties of Magnesium Forging Alloys The nominal compositions and properties of the principal commercial magnesium forging alloys are given in Table 1. The two that are most commonly used are AMC58S and AM65S. The former has the highest strength and is, therefore, employed for heavy duty applications. However, the forging of this alloy and of AMC57S ordinarily requires the use of slow speed hydraulic presses. The forging alloy AM65S combines good, although somewhat lower, mechanical properties with excellent forging characteristics. This alloy can be forged on hammers, upsetters, rolls, and mechanical presses, as well as on hydraulic presses. These characteristics make it particularly useful for many applications. Manganese Segregation in AM65S The alloy AM65S contains 5 pct tin, 3 1/2 pct aluminum, and a small amount of manganese, the manganese being added to insure satisfactory resistance to corrosion. Some manganese segregation was occasionally encountered in this alloy. In many applications, such segregation is of no particular consequence, but when forgings are acid dipped prior to the application of various chemical coatings, the segregated particles of manganese constituent may stand in relief. The resultant surface roughening may impair appearance and is undesirable in forgings employed in the textile industry, where very smooth surfaces are essential to avoid the catching and breaking of fine threads. An effort was therefore made to determine the cause of manganese segregation in this alloy and find a method for its elimination. Radiographic examination of a series of cast ingot sections and extruded samples of the AM65S type having manganese contents ranging from zero to about 0.7 pct revealed no appreciable manganese segregation when the manganese was less than about 0.4 pct. A small amount was observed with a manganese content of 0.46 pct and considerable with manganese contents of 0.6 pct or higher. Attempts to eliminate the manganese segregation by variations in the melting and casting conditions were unsuccessful and it soon became apparent that the trouble was attributable to the fact that the manganese contents that caused segregation were in excess of the amounts that would remain in solution at temperatures encountered in the melting and/or casting operations prior to the complete solidification of the ingots. A knowledge of the amounts of manganese that would be soluble in molten alloys of the AM65S type at various temperatures was obviously desirable. The liquid solubilities of manganese in binary Mg-Mn alloys and in Mg-Al-Mn and Mg-Zn-Mn alloys have been determined by Tiner and others,* but no data on the solubility of manganese in molten magnesium alloys containing tin and aluminum are given in the literature. For this reason, an investigation to determine the solubility of manganese in a magnesium alloy containing 5 pct tin and 3.5 pct aluminum (AM65S type) was undertaken.

Jan 1, 1950

By John B. Myers

Vermiculite is the commercial name used to describe a micaceous mineral that expands appreciably when heated. This process, called exfoliation, results in a light weight material. The light weight material is also called vermiculite, or more exactly, expanded vermiculite. Practically all uses of vermiculite are for the expanded material. Composition and Properties Vermiculite, in its natural state, has the characteristic micaceous habit and a perfect basal cleavage and readily splits into thin laminae, which are soft, pliable, and inelastic. The monoclinic crystal faces are often marked by triangular lines at 60° and 120°. The hardness varies from 1.5 to 2 or more, depending on the composition. The specific gravity varies from 2.2 to 2.7; specific heat 0.2; fusion temperature 2,000° to 2,500°F; luster pearly to greasy; color amber, bronze, brown, dark green, and black. Feel is like talc and sometimes soapy when wet. When heated quickly at elevated temperatures, vermiculite exfoliates by expanding at right angles to the cleavage into long wormlike pieces (the name vermiculite is from the Latin vermiculari, to breed worms). This important characteristic of expansion is believed to be a mechanical separation of the layers when the contained water is converted to steam. The increase in bulk volume of commercially prepared material is 8 to 12 times, but individual flakes may expand as much as 30 times. There is no expansion parallel to the cleavage. Vermiculite may also be expanded by several chemical processes such as soaking in hydrogen peroxide, weak acids and other electrolytes. The color changes during expansion. Heat-treated material changes to a light golden bronze or cream color depending on composition. The conditions of heating affect color. Heating in an oxidizing atmosphere produces reddish colors and heating in a reducing atmosphere produces dull gray colors. The iron is generally not oxidized during chemical expansion. Vermiculite is hydrated magnesium-aluminum silicate. Although there has been much research on the chemical and structural composition, there is not yet complete agreement on the exact formula and composition. This is to be expected since vermiculite is not a mineral of fixed composition and different workers have examined different varieties. It has been proposed that vermiculite is not a single species but a family of related minerals. This has also resulted in many names for the particular vermiculite found in a certain deposit. All of this difficulty arises from the very nature of its crystal chemistry. Unlike most minerals, certain crystal structure properties and the chemical composition can be changed by the ordinary forces of nature or in the laboratory. This would suggest that vermiculite be thought of more as a variable mineral rather than a group of related species. It was stated on page 261 of this volume in the chapter on Clay Minerals by H. H. Murray of the Georgia Clay Co., that vermiculite is an equidimensional, expandable clay mineral. Gruner (ref. 14, Chap. 12) showed that the structure consists of a layer made up of two silica tetrahedral sheets with a central magnesium, iron octahedral sheet separated by two layers of water molecules and adsorbed magnesium and calcium ions. Mathieson and Walker,2a in a recent publication, agree that vermiculite must be regarded as a true clay mineral. The characteristic properties of the mineral, such as high cation ex-

Jan 1, 1960

By T. L. Berger, M. C. lnman, M. F. Rose

AFTER shock loading, mechanical twinning has been observed by several authors1-4 for a variety of fcc metals and alloys. It has been shown2,3 that low stacking fault energy materials deform primarily by the formation of mechanical twins and stacking faults during dynamic loading. Further, Inman, Murr, and Rose2 have shown that a correlation exists between the average hardness and the volume percent of mechanically twinned material in stainless steel. It has been observed in nickel' that mechanical twins are produced during shock deformation if the peak pressure associated with the shock wave is greater than 350kbar. It is the purpose of this work to discuss shock-hardening mechanisms in nickel and to determine to what extent the initial condition influences the final response. The "flying plate technique" was used to generate planar shock waves with peak pressures as high as 710 kbar. The details of the technique5" have been discussed elsewhere and will not be repeated here. The material used in this study was polycrystalline nickel of 99.98 pct purity and had an initial grain size of 0.18 mm. Prior to any deformation, all samples were given a stress-relief anneal of 6 hr at 600°C in a vacuum furnace. After annealing, the initial condition of the samples was changed by cold rolling individual samples from 0 pct to a maximum of 80 pct reduction in thickness. This provided samples with controlled initial histories for shock deformation. Immediately after shock loading, the specimens were slowed and quenched by suitable arrangement of buffer material and water bath. This technique minimized specimen damage and the effects of shock heating on the samples. Since deformation twins in high stacking fault energy fcc metals are only produced at high stress levels, the initial defect configuration prior to mechanical twinning should greatly influence the number of twins formed. In fact, Zukas and Fowler7 have found that prestraining in iron and steel greatly suppresses the formation of mechanical twins during the shock deformation process. Following the methods of Warren and warekois8 and Cohen and Wagner,9 X-ray diffraction profile analysis was performed to determine the incidence of stacking faults and mechanical twins in the deformed nickel. Analysis was made on samples in the annealed, cold-rolled, shock-loaded, and cold-rolled—shock-loaded states. The estimated accuracy of the peak position measurements is *0.01 deg in 2?. Therefore the accuracy in the measurement of the relative peak shifts between the (400) and (222) diffraction line profiles is on the order of *0.04 deg in 2?. None of our measurements indicated peak shifts greater than 0.04 in 28. This would correspond to a fault density of -1 pct and sets the limit on the sensitivity of our instrument. Similarly, analysis of the shift in the center of gravity of the diffraction line profiles yields the minimum measurable value of the twin fault probabilitya to be -0.002. Fig. 1 shows mechanical twinning in nickel shock-loaded to 400 kbar. In this photograph, fault intersections are clearly visible. The thickness of the individual fault lamellae was determined from the image widths of their lines of intersections by the

Jan 1, 1968

By P. Dayal, L. S. Darken

THE migration of carbon in austenite under the influence of an electric current has been observed qualitatively several times.14 Lebedev and Guterman5 recently have reported quantitative measurements; their method consisted of welding a low-carbon to a high-carbon wire, passing current, and later estimating the carbon distribution metal-lographically. We believe the method reported in this paper is superior in principle since electrolytic migration and ordinary diffusion are not necessarily coupled as in the weld method; also for the first time the extent of migration is determined by combustion analysis for carbon. A uniform tubular specimen (S.A.E. 1045 or 1095) was mounted as illustrated in the figures. the whole assembly being in an atmosphere of purified nitrogen. Direct current (129 to 226 amp at 0.65 to 1.0 v) served as source both of heat and of potential gradient for carbon migration. The duration of experiment was 7.5 to 12 hr, during which time the temperature was controlled (±10°C) manually. The experimental conditions are indicated in the first five columns of table I. With this experimental arrangement, the temperature initially has a maximum value at the central plane. Also, in this vicinity (by virtue of the symmetry), the carbon entering any volume element is equal in quantity to that leaving; the loss or accumulation of carbon occurs near the electrodes since the carbon moves uniformly near the center and is "frozen" at the cold ends. Hence, providing the time of experiment is not sufficiently long to seriously influence the conditions at the central plane, carbon crosses this plane only under the influence of the electrical potential gradient and not under the influence of temperature or concentration gradient. The amount of carbon which has crossed this plane was determined by combustion analysis of the two halves of the specimen. The results of these analyses are shown in the sixth and seventh columns of table I. It will be noted that substantial amounts of carbon migrated from the anode portion to the cathode portion (except at 920°C, which temperature was too low to exhibit the effect in the allotted time). In spite of precau-tions taken, considerable decarburization occurred as may be seen by comparing cols. 8 and 5. In three cases (indicated by asterisks in table I) the specimen was ground on one side, polished, and examined metallographically before analysis. One specimen was normalized and re-examined. Micrographs of selected areas in specimens 4 and 6 are shown in figs. 1 and 2, respectively. Evidence of carbon migration from the anode toward the cathode is quite apparent. Fig. 2 shows abundant graphite as well as cementite in a region near the cathode. The temperature of this zone is estimated at about 800°C. In fig. 1, particularly in the micrograph on the extreme right (the right side of which is of unaltered carbon content), there is evidence of slight depletion of carbon in the immediate vicinity of the cathode as contrasted to the general increase in carbon of the cathode half of the specimen. Although this may be attributed in part to decarbur-ization, it is believed that it is occasioned also by thermal diffusion—migration (of carbon) from a cool to a hot zone4—a phenomenon superimposed upon the electrical migration. Evidence of carbon migration was visible particu-larly during the course of the high-temperature,

Jan 1, 1951

By B. H. Kear, H. G. F. Wilsdorf

After a few percent strain, dislocations in disordered Cu3Au are arranged in groups in ulell-defined slip planes, which contrasts with the more or less random distribution of dislocations in the corresponding ordered alloy. Unusually long and straigkt dislocation segments observed in the ordered alloy are interpreted in terms of cross slip of screw dislocations from their slip planes into cube planes. It is proposed that cross slip, and also dislocation reactions of the Lomer type, play an important role in work hardening of the ordered alloy. ACCORDING to to dislocations in the disordered lattice are only partial dislocations in the corresponding super lattice, so that in the super-lattice dislocations must occur as pairs, with each pair connected by a ribbon of antiphase boundary. Recently, striking confirmation of the theory was obtained by Marcinkowski, Brown, and Fisher3j4 in a study of dislocation configurations in thin foils of the Cu3Au alloy by transmission electron microscopy. In the present investigation, the observations made by these workers have largely been confirmed. Moreover, it has been found that dislocations in the plastically deformed ordered alloys are often unusually long and straight. This paper is mainly concerned with these new observations. A tentative explanation is offered in terms of cross slip of dislocations in the ordered alloys. PREPARATION OF SPECIMENS An alloy close to the stoichiometric composition Cu3Au was prepared by melting together high-purity copper and gold (99.995 pct) under vacuum in a graphite mold. The resulting ingot was rolled down and cut into strips 1 1/2 by 1/8 by 0.005 in. in thickness. All strips were annealed at 900°C for 3 hr and cooled to 500'~. A few of the strips were quenched from ~OO'C, while the remainder were slowly cooled in the temperature range 400" to 150°C at intervals of a few degrees every 24 hr to produce a high degree of order. After this annealing treatment the average grain size of the material was about 0.1 mm. Material in different states of order was obtained by reannealing fully ordered material at temperatures below the critical temperature: followed by quenching. The equilibrium degree of order attained at a particular reannealing temperature was obtained from the paper by Keating and warren.' A simple hand-operated tensile jig was used to give the strips the required amounts of plastic strain. Following straining, a number of specimens were cut from each strip by means of an electrolytic cutting device.= Finally, each specimen was thinned electrolytically using the polishing procedure described by Bakish and Robertson,7 and these thin sections were examined in the Philips EM-100B electron microscope operating at 100 kv. EXPERIMENTAL OBSERVATIONS Antiphase boundaries. The average domain size in the ordered alloys was found by direct observation to be about 0.25 p. Also, it was observed that the grown-in configuration of antiphase boundaries was changed little by plastic deformation, at least up to 6 pct strain. Fig. 1 shows an electron micrograph of a thin foil of partially ordered alloy (S - 0.8) after 3 pct strain. The micrograph shows a few dislocation pairs and a somewhat crystallo-graphic network of antiphase boundaries. The normal to this foil was found by diffraction analysis to be approximately 11121, with principal directions in

Jan 1, 1962

By Leon Merker

FLAME fusion growth of strontium titanate crystals was undertaken to obtain large transparent crystals on which physical data could be gathered. The fact that strontium titanate is a cubic crystal and probably has a high index of refraction indicated that it might be useful in optics. No large boules* of strontium titanate had been synthesized previously, and none have been found as natural occurrences. The flame fusion process, invented by Verneuil in 1904,1 was first applied to grow large crystals of ruby and has since been used extensively to grow crystals of corundum and spinel on a commercial scale. Only during the past nine years has this technique been applied to the growth of large single crystals of other high melting compounds such as rutile,2 mullite,3 and ferrites4 One of several significant advantages of the flame fusion technique over more conventional methods of growing crystals from a melt is that the crystals are grown without the use of melt containers. This fact alone excludes other methods of growing crystals from a melt whenever there are compounds with a melting point near 2000°C or above. Growth of the crystal, similar to that of a stalagmite, takes place inside a circular refractory furnace that does not contact the crystal, Fig. 1. This eliminates a serious source of impurities as well as the likelihood of multi-crystalline solidification along the container walls. The flame fusion technique is particularly advantageous for research because it produces large crystals quickly and at lower cost. A corundum crystal 25 mm long and 12 mm in diam can be grown in 2 to 3 hr. Operating conditions permit observation of the growing crystal during any stage of the process. The various flame patterns and a wide range of oxidation-reduction conditions can be investigated in a relatively short time. It should be pointed out that formation of the foot of the boule is the most critical step in growing an unfractured boule. When the tip of the sintered cone is fused to initiate crystal growth, the original area of fusion should be kept as small as possible so that one crystallite can assume dominance over all the others and propagate as a single crystal. If the boule start happens to consist of several crystals, usually the entire boule also contains several crystals. In many cases these tightly cohering crystals exhibit different crystallographic orientations. Although the basic technique of the original Verneuil process has been retained, refinements in method and burner design have been developed in recent years. For example, boule growth may be initiated on a seed crystal rather than by fusion of a sintered cone. The danger of multicrystalline or twinned boule starts is thus eliminated, and this procedure may also be used to grow boules of predetermined crystallographic orientation. Prior Art: Of the many attempts to produce artificial gem crystals, most early efforts% ere directed toward synthesis of .rubies and sapphires and diamonds. None of these efforts resulted in a process that was commercially practical. Without exception only tiny crystals of corundum were obtained, and those at excessive cost. Verneuil announced his process in 1904; Ravier6 in 1934 presented a detailed description of a commercial boule production plant. Descriptions of the German synthetic stone industry, the process details, and the equipment used have been published by Barnes7 and by Merker.'

Jan 1, 1956

By N. K. Chen, R. Maddin

In the extension of molybdenum single crystals at room temperature, the slip planes were found to be of the type (1101; the slip direction <111>. Theories of plasticity of body-centered cubic metals have been examined, and an explanation based upon planes of highest atomic density seems plausible to explain the plastic behavior of molybdenum single crystals. IN body-centered cubic metals, slip has been reported to occur on several planes; these are of the type {110), {112), and (123). Andrade has correlated&apos; the operative slip planes in body-centered cubic metals with the testing temperature relative to the melting point. In this correlation, if T is the absolute temperature at which slip takes place and Tm the absolute melting point, then the acting slip planes in body-centered cubic metals may be arranged as shown in Table I. However, deviations from this correlation have been reported in iron and iron-silicon alloys2 and also in molybdenum deformed at a temperature around 2400°C." In the case of iron and other body-centered cubic metals or alloys in which all three slip planes have been reported to be active at any one temperature, the crystallographic mechanism of slip is not well understood. Numerous investigators in this field have assumed that the resolved shear stress is a deciding factor. Such considerations were given by Gough in his experiments with repeated torsion of iron single crystals4 and recently by Opinsky and Smoluchowski in their study of iron-silicon alloys," and will be dealt with later in this report. A third possibility is that slip lines in body-centered cubic metals may be formed by alternate slip on nonparallel (110) planes and not by slip on any plane of higher indices. This was advanced by Elam in the case of ß brass." Since (1101 planes are planes of highest atomic density in the body-centered cubic structure, this hypothesis would be in accordance with the plastic behavior of metals in other systems, namely, face-centered cubic and hexagonal close-packed. Experimental evidence in support of this hypothesis is lacking. Nevertheless the work of Barrett, Ansel, and Meh12 has definitely demonstrated that the slip resistance on {110} planes in iron is very much smaller compared with that on (112) and (123) planes. In the present investigation, molybdenum single crystals were carefully documented as to the operative slip planes and lattice reorientations after plastic extension at room temperature. In three crystals studied, (110) planes were found to be active invariably rather than the (112) planes, which were previously reported for molybdenum at this temperature&apos; and predicted by Andrade&apos;s theory.&apos; In two other cases (112) and (123) planes seemed to operate together with (110) planes as active slip planes. However, it is believed, after careful analysis of crystallite and lattice rotations, that slip on (112) and (123) planes may be spurious. Particular reference will be made to the formation of asterism, lattice rotation, and crystallite rotation in an attempt to show that slip apparently observed on {1121} and (123) planes is actually slip on nonparallel {110} planes. Materials and Treatment Sintered molybdenum rods ? in. in diam, obtained from the Fansteel Metallurgical Co., were reported to have a purity of 99.9 pct.* One specimen, Mo-13, ¼ in. in diam, was supplied by Westinghouse Electric Corp. Single crystals were prepared, using the technique devised by Chen, Maddin, and Pond." Mo-13 was ¼ in. in diam x 7 in. long with a reduced gage section of 0.200 in. in diam x 3 in. long. The center of

Jan 1, 1952

By H. H. Rachford, D. S. Howard

In the withdrawal of fluids from tilted aquifers it is of value to be able to predict pressure patterns during the course of the pressure decline. As an example of this, in the displacement of fluids from oil reservoirs being produced by water-drive, prediction of the reservoir pressure is an important factor in fixing production rate. The prediction of pressure patterns in the depletion of an inclined aquifer has conventionally been made by solving the well known flow equations for the pressure Patterns in a horizontal aquifer, and correcting these patterm for the pressure difference due to the hydrostatic head associated with the vertical distance between a point in the horizontal aquifer and the corresponding point in the tilted aquifer. Inasmuch as the equation describing flow in tilted strata is different from that for horizontal flow. there was question as to the validity of the simple corrective procedure current1y in use. To examine the magnitude of error made using this procedure, an inclined aquifer was chosen for which the error was expected to he large, and the equation for flow in this tilted stratum was solved numerically. The results of the conventional procedure of correcting the solution of the equation for horizontal flow was compared with the solution of the correct equation by studying the difference between predicted pressures by the two methods. Three forms of the hydrostatic head correction were studied. The results for all three forms of the head correction are similar. The error in the approximate solution is quite dependent on the position of the horizontal aquifer with respect to the inclined one. The error is greatest at the producing end ad increases with amount of fluid withdrawn. The magnitude of the errors is not large if a good choice is made for the position of the horizontal datum. The most satisfactory position for the datum in all cases is near or slightly about the mid-point of the inclined stratum. Use of the best approximate method studied with the best datum tested yielded predicted pressures in error no greater than 7 psi. INTRODUCTION The movement of water by expansion from an aquifer to displace fluids from a contiguous oil reservoir was first recognized over 20 years ago.9 Since that time much work has been done to relate this mechanism to reservoir behavior mathematically. The result of this effort on the part of many workers has led to an understanding of underground fluid movement which has enabled us to follow and predict the behavior of the water-drive reservoir. One reason that good progress has been made in the application of mathematics to the study of pressure patterns and water movements in aquifers has been that the differential equation which describes this process is of a relatively simple form. The equation for the isothermal, horizontal, viscous flow of water through homogeneous sandstone is identical with that for the diffusion of heat through homogeneous solids, and this equation and means of obtaining solutions for it with many types of boundary conditions have been studied extensively. That this work has been highly successful is clearly emphasized by the large number of solutions developed for problems of interest both by methods of formal mathemat-

Jan 1, 1957

By I. Fatt, D. N. Saraf

A method is described for measuring two- and three-phase relative permeabilities in sandstones or sand packs using a nuclear magnetic resonance (NMR) technique to determine fluid saturations. Two- and three-phase relative permeabilities have been determined on Boise sandstone using the NMR technique of saturation measurement. Three-phase relative permeability to water was found to depend only on the water saturation, whereas three-phase permeability to oil depended on both the water and oil saturations. Relative permeability to gas in three-phase flow was found to depend only on the total liquid saturation. INTRODUCTION Three-phase relative permeabilities are extremely useful in calculating field performance for reservoirs being produced by simultaneous water and gas drives. Three-phase relative permeability data are also needed for analyzing solution gas-drive reservoirs which are partially depleted and are being produced by water drive. Some thermal recovery processes involve three-phase flow which require three-phase relative permeability data for predicting reservoir behavior. Unfortunately, three-phase relative permeability measurements have rarely been made. Also, because of the scarcity of three-phase data, it has not been possible to date to relate other measured rock characteristics to the relative permeabilities with a great certainty. Leverett and Lewis,l Reid2 and Snell3 have reported three-phase relative permeability data on unconsolidated sands. Leverett and Lewis used ring electrodes spaced along the length of the sand sample to measure the resistivity of the sample which was assumed to be monotonically related to brine saturation. Gas saturation was determined from pressure-volume measurements. Oil saturation was obtained by material balance on the cell containing the sand sample. This method is involved and time consuming. Another difficulty arises from the fact that the resistivity of the sand is a function not only of saturation of brine but also of the distribution and saturation history of the brine in the pore spaces. Reid2 used a gamma ray absorption technique for measuring liquid saturation. This method has the disadvantage that total liquid saturation rather than oil or brine saturation is all that can be measured and still another method is required to determine the saturations of individual components. snel13 used a neutron bombardment method which also required a separate determination of the individual component saturations. Caudle et al.4 measured three-phase relative permeability on consolidated sandstones using vacuum distillation for determining fluid saturations. Distillation after each reading makes this technique very lengthy and time consuming. Corey et a1. 5 and Naar and Wygal 6 measured three-phase relative permeability on sandstones by the capillary-pressure method. Semipermeable diaphragm assemblies were used at each end of the core specimen to keep the water phase in the core. Gravimetric methods were used to determine fluid saturations. Sarem7 recently reported an unsteady-state method for measuring three-phase relative permeability on sandstones. This method is an extension of Welge&apos;s method8 for measuring two-phase relative permeability. Although Sarem&apos;s method is simple and comparatively fast, the assumptions involved may oversimplify the problem. Sarem&apos;s assumption, that in all rocks relative permeability to each fluid will depend only on the saturation of that fluid, seems to be rather unrealistic. Neglecting capillary effects at the end of the core is also a weak assumption. Donaldson and dean9 measured three-phase relative permeability using a method similar to Sarem&apos;s.

By R. D. Satterley

The Sherwood mine is an iron-ore mine owned and operated by the Inland Steel CO. in the Iron River district of the Menominee Range in Michigan. The property consists of an 80-acre tract in the village of Mineral Hills, about a mile north of the city of Iron River. The Ore mined is a reddish brown, nonbessemer hematite and limonite. The ore bodies lie in a large syncline having an east-west strike and are associated with a rock structure that is folded in a complex manner. The two 40-acre tracts run north and south and the shaft is at about the center of the north forty. The shaft is 1286 ft. deep, with the principal operating level at a depth of 1000 ft. from surface. A new level at a depth of 1200 ft. is in process of development. The ore body at the 1000 ft. level is east and west at approximately the line between the north and south forties and at this elevation is from 50 to 250 ft. wide and extends above the level at varying heights of from 50 to 250 feet. The 1000-ft. haulage drift extends south from the shaft, through the east-west ore body to the slates on the south side of the body. The ore body was intercepted in the drifting at about goo ft. from the shaft and was about 200 ft. wide at that point. A crosscut was extended to the east and west property lines along the south boundary of the ore body. This east-west ore body is laid out in areas go ft. wide for stoping with pillars 40 ft. wide. With this plan a stope is approximately 80 by 200 ft. and has a pillar on both sides that is approximately 40 by 200 ft. The ore has a height of from 50 to 250 ft. above the elevation of the level, therefore the height of the stopes varies in accordance with the height of the ore body at each location. From the crosscut, two main raises, each 6 by 7 ft., are put up into each 80 ft. stope area (Fig. 1). One of the main raises serves as a manway and the other as an ore raise, through which the ore is transferred from the stope to the tramming equipment. The original layout as for scraping of the ore from under the stopes into the ore raises. The ore raise is put halfway (east and west) along the 80-ft. stope area and usually is in iron formation. The manway and ore raise are on approximately 15-ft. centers. These main raises are Put up to a height of from 22 to 30 ft. above the level floor, then from the toll of the ore raise a scraper drift (8 X 8 ft.) is driven across the width of the ore body, which is in the neighborhood of 200 ft. From the scraper drift, mill raises are put up at 12½-ft. centers, staggered on opposite sides of the drift, and these are extended 23 ft. above the elevation of the scraper-drift floor. The sublevel connecting the top of these mill raises is called a mill "sub." Other sublevels are developed parallel to the scraper and mill subs at 25-ft. vertical intervals above the mill sub. The mill sub and the other subs above the mill sub are known as sublevels or dog drifts, and are 4 ft. wide and 6 ft. high. These dog drifts

Jan 1, 1946

By V. Koump, T. F. Perzak, R. G. Olsson

The vate of dissolutioz of carbon in molten Fe-C alloys urns studied by votating- cylindrical g-vaphite sairlples in a stationary crucible containing the melt. The rate of dissolltion was determdined from the change in the bnth co)npositiorz. Alccss-tvansfev coeffi&apos;cients, calcilated with the assurlption that the inte)yacial liquid ulas saturated with carbon, were found to be in agreement with an existing Correlation fov mass transfer jroni votating cylinders. It is corzcllrded tlat, itz the wnge of- Reynolds&apos; numbers itzvestigated (790 to 18,000), the vate of dissolutio of carborz in Fe-C alloys is controlled by tile rate of carbon difirsion from the interface. The dissolution of carbon in molten Fe-C alloys has been the subject of a number of investigations. Most of these investigations have been inconclusive because of poorly defined experimental variables. Two of the more sophisticated studies are summarized below. Dahlke and Knacke&apos; studied the phenomenon by, in part, immersing graphite rods into 0.5 pct carbon-steel for different lengths of time. Since the experiments were conducted in air or in CO atmospheres, the melt was strongly agitated by gas evolution. It was concluded that the change in bath composition as a function of immersion time was in agreement with the assumption that the dissolution process was transport-controlled. The hydrodynamic conditions in these experiments, however, were not well-defined and gave no indication of the nature of the dissolution process at higher fluid velocities. In a recent study Shurygin and ruk investigated the dissolution of carbon in molten Fe-C, Fe-Si, Fe-P, and Fe-Ni alloys. Small graphite disks were rotated in a crucible containing molten alloys under such conditions that only the lower face of the disk was exposed to the melt. The results of this investigation indicate that, over the range of velocities (100 to 1000 rpm) and bulk liquid concentrations (0.7 to 4.5 wt pct C) considered, the dissolution of carbon in the iron alloys is limited by the diffusion of carbon from the interface to the bulk liquid (a transport mechanism). For the particular geometry used by these authors, exact analytical solutions of hydrodynamic and diffusion problems are aailable. Hence the experimental results obtained by Shurygin and Kryuk permit the evaluation of diffusion coefficients of carbon in the four systems investigated. In the present investigation a series of experiments were conducted in which the rate of dissolution of carbon in Fe-C alloys was measured under conditions which permitted wide variation of the fluid velocity and the carbon concentration of the bulk liquid. Graphite cylinders were rotated at different speeds in a stationary concentric crucible containing various molten Fe-C alloys. THEORETICAL DEVELOPMENT In general, the resistance to mass transfer from a solid interface to a turbulent liquid is assumed to reside in a laminar boundary layer adjacent to the interface. If it is also assumed that the diffusion coefficient is independent of concentration, the steady-state rate of dissolution for the case of arbitrary liquid and solid densities can be obtained from the integration of the equations of continuity. The derivation is similar to that used by Lommel and chalmers4 for the case of pl = ps. The solution, expressed in terms where N = rate of diffusion of the solute from the interface. Exact analytical solutions of the hydrodynamic equation for the boundary layer thickness are available for only a limited number of geometries. Consequently, a mass-transfer coefficient, k, is usually

Jan 1, 1967

By Charles C. Templeton

Any explanation of the nature of multiphase fluid flow through porous media must ultimately involve a knowledge of the mechanism of such flow through the individual pores of the medium. To provide quantitative information of the latter type, a study was made of gas-liquid and liquid-liquid displacements in single uniform Pyrex capillaries with diameters as small as 4 microns. Quantitative observation of air-liquid displacements with zero contact. angle indicates the adequacy of a model based on Poiseuille&apos;s Law and the independence of capillary pressure and interfacial velocity. Consequently, this work discounts previous claims that flow is abnormally retarded in very small capillaries. For the displacement of a very viscous oil by water in a capillary initially filled with water, flow is faster than predicted by the above model. This behavior is probably due to the existence of a thin annular film of water between the oil column and the capi1lary wall as postulated by Yuster. INTRODUCTION The need for experimental observation of the microscopic nature of multiphase fluid flow, to further the understanding of the results of macroscopic flow investigations, has been previously recognized. Chatenever and Calhoun1 in connection with the API Project 47B have extensively studied the flow of fluids through layers of glass or lucite spheres between glass or lucite plates. Lowman2 prepared an interesting film concerning flow in a single capillary, using an apparatus which made a moving interface appear stationary in the field of a microscope. The purpose of the present investigation was to deter- mine what contribution might be made to this field by comprehensive quantitative observation of gas-liquid and liquid-liquid displacements in single capillaries of uniform, very small diameter. The techniques used are similar to those of Lowman but incorporate improvements necessary for quantitative work. Quantitative observations must begin with fluids of simple chemical composition to provide a basis for analyzing the more complex behavior that may be expected with natural reservoir fluids. However, some qualitative observations on systems involving crude oils are presented. Experiments of this sort involve the observation of the velocity of the interface separating two fluids, and have been performed in horizontal or vertical capillaries by several investigators.3,4,5,6,7,8 The most general treatment of their mechanics, by Brittin,9 involves forces due to capillarity, viscous resistance, gravity, an "end-drag" effect, and the rate of change of momentum of the capillary contents. Since the range of pore diameters of interest in petroleum reservoirs is about 0.1 to 100 microns, it was desired to perform these experiments in the smallest capillaries possible, to ascertain whether capillary diameter had any effect on the existing theory based on data from larger capillaries.* It was also important to be able more systematically to vary the direction and rate of displacement than was done by previous investigators. Finally, to concentrate attention on viscous and capillary forces by eliminating gravity as a parameter, this study involves only horizontal capillaries. The work of Eley and Pepper&apos; with a "large" (0.766-mm diameter) horizontal capillary is the previous investigation which is most similar to the present method of quantitative observation. Here capillaries with diameters as small as 4 microns will be studied.

Jan 1, 1955

By R. A. Morse, W. W. Owens, D. R. Parrish, T. M. Geffen

Laboratory studies of several factors affecting measurements of relative permeability were made using the three-section plastic-covered core technique. Results show that the core assembly, properly constructed, will perform as a single unit, and that the testing technique will, under suitable conditions of pressure gradient, gas expansion, and migration of partial water saturation, permit measurement of flow characteristics not affected by technique. Wettability equilibrium is readily established in cores exhibiting strong wetting preference to water or oil when initially saturated with water. Laboratory tests must be conducted so that saturation changes represent those that occur in the reservoir. Immediate implications of saturation history are (1:) that the possibility exists of increasing the displacement efficiency of solution gas drive reservoirs over the natural process, and (2) residual gas saturations following water flooding in gas or gas condensate reservoirs will be 15 to 50 per cent pore space rather than 1 to 11 per cent as generally believed. INTRODUCTION Solutions of petroleum reservoir problems pertaining to production performance require the use of true relative permeability characteristics. &apos;This relationship of fluid conductivity and saturation has been obtained by reservoir engineers in four ways, namely: 1. From past gross reservoir performance and the extrapolation of this data based on experience,1,2 2. By using published fluid flow relationships obtained in laboratory studies on general type porous materials,3, 4, 5 3. By attempting a mathematical derivation of flow behavior," using some experimentally obtained characteristics of reservoir rocks, and 4. By laboratory flow tests using representative rock samples of a reservoir. The first three methods listed above have shortcomings which make their use limited or questionable. Production characteristics of only certain processes are obtained from field data and these are not available at the beginning of a reservoir&apos;s producing life, at which time they are desirable. It is fortuitous if general fluid flow characteristics obtained experimentally have accurate application to specific field problems. Also, it is felt that at this time there is not sufficient knowledge of the flow behavior of oil, water, and gas in porous materials to enable applicable analytical description of this to be made based on other measured rock characteristics. Measurement of relative permeability in the laboratory offers the only direct method subject to adequate checking for determination of flow characteristics applicable to field problems. Primarily, this paper deals with laboratory experiments to establish the effects of several factors on the measurement of relative permeability and the practical significance of this knowledge. Laboratory flow tests are essentially model studies; the gross reservoir being represented by a small rock sample. It is necessary to simulate in experiments all the factors of importance which influence flow in the reservoir, as well a.; to eliminate effects peculiar to the model studied to assure measurement of valid flow behavior. Factors which affect the measurement and behavior of fluid flow in porous materials have been studied by other investigators.3,7,8,9,10,11 This analysis is not intended to evaluate fully all factors involved, but to establish first the validity of measurements by the technique used and then study flow behavior under various conditions. The studies reported are as follows: 1. Effects of pressure gradient, 2. Effects of gas expansion in tests conducted at near atmospheric pressure, 3. Effects of migration of a partial water saturation in oil-gas-water systems, 4. Effects If wettability characteristics of the system. and 5. Effects of the fluid saturation history. The first three are important only in laboratory measure-

Jan 1, 1951

By D. R. Parrish, T. M. Geffen, R. A. Morse, W. W. Owens

Laboratory studies of several factors affecting measurements of relative permeability were made using the three-section plastic-covered core technique. Results show that the core assembly, properly constructed, will perform as a single unit, and that the testing technique will, under suitable conditions of pressure gradient, gas expansion, and migration of partial water saturation, permit measurement of flow characteristics not affected by technique. Wettability equilibrium is readily established in cores exhibiting strong wetting preference to water or oil when initially saturated with water. Laboratory tests must be conducted so that saturation changes represent those that occur in the reservoir. Immediate implications of saturation history are (1:) that the possibility exists of increasing the displacement efficiency of solution gas drive reservoirs over the natural process, and (2) residual gas saturations following water flooding in gas or gas condensate reservoirs will be 15 to 50 per cent pore space rather than 1 to 11 per cent as generally believed. INTRODUCTION Solutions of petroleum reservoir problems pertaining to production performance require the use of true relative permeability characteristics. &apos;This relationship of fluid conductivity and saturation has been obtained by reservoir engineers in four ways, namely: 1. From past gross reservoir performance and the extrapolation of this data based on experience,1,2 2. By using published fluid flow relationships obtained in laboratory studies on general type porous materials,3, 4, 5 3. By attempting a mathematical derivation of flow behavior," using some experimentally obtained characteristics of reservoir rocks, and 4. By laboratory flow tests using representative rock samples of a reservoir. The first three methods listed above have shortcomings which make their use limited or questionable. Production characteristics of only certain processes are obtained from field data and these are not available at the beginning of a reservoir&apos;s producing life, at which time they are desirable. It is fortuitous if general fluid flow characteristics obtained experimentally have accurate application to specific field problems. Also, it is felt that at this time there is not sufficient knowledge of the flow behavior of oil, water, and gas in porous materials to enable applicable analytical description of this to be made based on other measured rock characteristics. Measurement of relative permeability in the laboratory offers the only direct method subject to adequate checking for determination of flow characteristics applicable to field problems. Primarily, this paper deals with laboratory experiments to establish the effects of several factors on the measurement of relative permeability and the practical significance of this knowledge. Laboratory flow tests are essentially model studies; the gross reservoir being represented by a small rock sample. It is necessary to simulate in experiments all the factors of importance which influence flow in the reservoir, as well a.; to eliminate effects peculiar to the model studied to assure measurement of valid flow behavior. Factors which affect the measurement and behavior of fluid flow in porous materials have been studied by other investigators.3,7,8,9,10,11 This analysis is not intended to evaluate fully all factors involved, but to establish first the validity of measurements by the technique used and then study flow behavior under various conditions. The studies reported are as follows: 1. Effects of pressure gradient, 2. Effects of gas expansion in tests conducted at near atmospheric pressure, 3. Effects of migration of a partial water saturation in oil-gas-water systems, 4. Effects If wettability characteristics of the system. and 5. Effects of the fluid saturation history. The first three are important only in laboratory measure-

Jan 1, 1951

By S. Kado, G. Derge, L. Yang

Self-diffusion coefficients of silver in molten silver have been measured by means of the capillary-reservoir method in the temperature range 1002" to 1105°C. The results can be .represented by the equation D = (7.10 i 0.06) X 10-4 exp [(—8150 ± 1130)/RT]. The heat of activation agrees within the limit of experimental uncertainty with that for the viscous flow. The validity of the Eyring and the Stokes-Einstein equations are examined on the basis of the self-diffusion coefficients obtained and the known viscosity data of molten silver. MEASUREMENT of the self-diffusion coefficient of silver in molten silver was made in the temperature range 1275° to 1378°K, using the capillary-reservoir method and &"&apos; as the radioactive tracer. The experimental arrangement is shown in Fig. 1. Silver of 99.99 pct purity (fine silver), was melted and degassed in vacuum in a graphite crucible situated inside a McDanel vacuum jacket. The crucible (4 in. tall and 1 in. ID) was protected both at the top and at the bottom with graphite shields and contained about 220 g of silver. The sample holder was a piece of graphite rod (1 in. long and 1/2 in. diam) with 4 capillaries drilled longitudinally into it. The holder was screwed onto a long graphite rod (12 in. long and 74 in. diam) which was connected to a steel disk hanging on a nylon string through a glass pulley. The steel disk kept the sample holder at the center of the McDanel tube and was heavy enough to overcome the buoyancy when the holder was lowered into the molten silver. By turning a handle connected to the pulley through a tapered glass joint, it was possible to raise or lower the sample holder without breaking the vacuum. Two such arrangements were constructed, sitting side by side in 2 Kanthal furnaces. One with non-radioactive silver in its crucible was used for the diffusion run and the other with radioactive silver in its crucible was used for filling the capillaries. To fill the capillaries with radioactive silver, the sample holder was degassed first and then lowered into the radioactive molten silver in vacuum. Purified argon gas was then admitted into the system to push the liquid silver into the capillaries. After cooling in argon, the filled capillaries were polished on emery paper to expose the top of the rod and used immediately as samples for the diffusion run which was carried out at the same temperature as that used for filling the capillaries. During the diffusion run, the samples were degassed in vacuum and lowered slowly into the nonradioactive molten silver before purified argon was admitted into the system. After a predetermined period of time, the samples were raised out of the melt and cooled in argon. The temperature of the melt was measured before and after the run with a chromel-alumel thermocouple inside a McDanel protection tube. Another chromel-alumel thermocouple attached to the outside wall of the McDanel vacuum jacket was used to indicate any fluctuation in the temperature of the furnace during the run, which was found to be less than ± 2°C for the longest run made (93 min). The temperature gradient inside the molten silver was found to be negligible. No significant difference was observed in the results obtained whether the melt was stagnant during the run or was stirred by raising and lowering the sample holder slowly a few times a minute for a distance of about 1/4 in. inside the melt.

Jan 1, 1959

By D. F. Stoneburner

The thermoelectric power and electrical conductivity of a series of liquid alloys of thallium were determined in order to study the relation of chemical bonding to semiconduction in liquids. The results indicate metallic behavior for the alloys of thallium with arsenic, antimony, and bismuth, while the sulfides, selenides, and tellurides indicate semiconduction. The results are compared with the Properties of the same alloys in the solid state. This work supports the observation that long-range order is not a necessary requirement for semiconduction. PREVIOUS work1,5 has shown that some sulfides and oxides behave as semiconductors in the liquid state. Although the band theory has been of great help in understanding the properties of solid semiconductors, its quantitative application to liquids is difficult. The very existence of liquid electronic semiconductors shows that ascribing the discrete structure of the energy spectrum to strict lattice periodicity, or long-range order, is not always satisfactory. Rather, the short-range order determines the structure of energy levels and the width of the forbidden gap and, consequently, the concentration of free holes and electrons. This is borne out by the analysis of data accumulated by Regel and coworkers.6 The similarity of electronic structure of liquid and solid phases of certain liquid metals was pointed out by Knight et a1.,7 who examined nuclear resonance in solid and liquid Li, Na, Rb, Cs, Mg, Al, Ga, Sn, and Bi. They found noticeable changes in the electronic structure upon melting only in the cases of bismuth and gallium. They concluded that "a liquid metal possesses a band structure which is very like that of the solid, provided the short-range structure is not altered during melting." In semiconductors in which the near-neighbor arrangement remains the same on fusion, the material would be expected to retain its semiconducting properties. The quantum theory of semiconductors, with its band structure of energy levels, is linked exclusively to long-range order. Furthermore, it has limited usefulness in predicting new semiconductors. Many sulfide8-12 have made comparisons of properties on the basis of crystal structure or the position in the periodic table, both methods involving the same factors as chemical bonding. Mooser and Pearson8"10 have pointed out that crystal structure depends on chemical bonding, and that this bonding is a property of the atoms themselves. As the periodic arrangement of the atoms offers an easy comparison of such factors as atomic weight and number of valence electrons it provides clues to expected similarities of bonding for various elements or groups of elements. This approach has led Mooser and Pearson to formulate their valence-bond theory of semiconduction, in which they define a semiconducting bond. These authors state that "semiconductivity is the result of the presence in a solid of predominantly covalent bonds. These lead, through the process of electron sharing, to completely filled s and p orbi-tals in the valence shells of all atoms of elemental semiconductors while in semiconducting compounds only one atom of any two bonded together need acquire closed s and p orbitals. The presence of empty &apos;metallic&apos; orbitals in some atoms of a compound does not destroy the semiconductivity provided that these atoms are not bonded together; it may, however, give rise to pivoting resonance in

Jan 1, 1965