Search Documents

By W. C. Ellis, E. S. Greiner

Metallic titanium has been prepared in small quantities since the beginning of the century. Hunter1 reported in 1910 that he obtained a malleable product of 99.9 pct purity by the reduction of the tetrachloride by sodium in a closed container. Later Van Arke12 and Fast prepared small quantities of malleable titanium by thermal decomposition of titanium iodide. Another method for producing malleable titanium from tetrachloride was reported by Kr0ll4 who used magnesium as the reducing agent. More recently Dean, Long, Wartman, and Anderson5 working at the U.S. Bureau of Mines have reported the preparation of the metal on a pilot plant scale by the reduction of titanium tetrachloride by magnesium in an atmosphere of helium. Their product, which is malleable, was used in the present investigation. Early in 1948 after the work of this paper was completed Campbell, Jaffee, Blocher, Gurland, and Gonser6 described the strength and working characteristics of pure titanium which was prepared by the iodide method on a scale producing 300 to 600 g per run, in the form of 0.3- to 0.4-in. diam rods. Their product is characterized by low hardness and large capacity for cold work which they attribute to a high degree of purity with respect to oxygen, nitrogen, and carbon. Titanium has a melting point of about 1725 C," it may be classed as a light metal since the specific gravity is 4.5 g per cc-I. The crystal structure is close-packed hexagonal from room temperature to 885°C. Above this temperature the structure is reported7 as body-centered cubic. The change in crystal structure has been associated with discontinuities in specific heat,8 electric resistanceg and theimoelec-tric force10 at temperatures near 890°C. It is reportedll that annealed titanium has a tensile strength of 78,700 psi which can be increased by cold-work to 123,000 psi. In the present investigation, the lattice constants at room temperature, the thermal expansivity between 30 and 800°C, as well as the electrical resistivity and thermoelectric force against platinum between — zoo and I000 C have been determined. Materials For the measurements, two series of specimens were prepared from titanium furnished by the U. S. Bureau of Mines. The first was made from an annealed sintered compact, 0.38 in. thick. Sections of the compact were cold-swaged with intermediate anneals in vacuum to the sizes used. Sizes over 0.25-in. diam were annealed at I000°C and smaller sizes at 800°C. Specimens originating from the sintered compact are designated in this paper as A. The second series of specimens was made from a rod of titanium, 0.19 in. in diam, which was cold-swaged with the necessary intermediate anneals to the required sizes. This series of specimens is designated B. The impurities in the specimens of titanium, determined by chemical analysis, are given in Table 1. Oxygen and nitrogen which Were not determined by analysis, are probably present although in small quantities in view of the ductility of the material.

Jan 1, 1949

By A. F. Lucas

Introduction BELLE ISLE, located in the low sea marshes near Atchafalaya Bay, is the southeasternmost of the famous Five Salt Islands of Louisiana. Rising about 80 ft. (24 m.) above the level of the surrounding marsh and being some 800 acres (323 ha.) in extent, it constitutes a striking landmark on the low Gulf coast, and since the eighteenth century, when it played an important part in the operation of the buccaneers, it has been intermittently frequented by man. In 1897, the writer discovered the great mass of salt that underlies the island and shortly afterward pub-lished a brief account of the geology of Belle Isle and the neighboring salt islands.' In that paper he described principally the salt, but pointed out also the occurrence of petroleum and of sulphur in the strata above the salt. Since that time the island has been extensively explored by the drill, and the developments up to 1908 have been described by Veatch2 and Harris," and Knapp.4 The writer's early conclusions as to the probable presence of sulphur in considerable quantity have been confirmed by the later drilling, some of the wells having found quantities of gas containing sulphur dioxide and a number of sulphur-bearing strata as well. Under the stimulus of the present keen demand for sulphur for war purposes, a syndicate of New York capitalists was recently formed to explore the sulphur resources of the island in more detail. Under the writer's direction, six wells were drilled for this syndicate during the winter of 1916-17. The present paper presents briefly the results of this work and sums up the present knowledge of the geology of this highly interesting locality.

Jan 9, 1917

By E. I. Larsen, F. R. Hensel

According to statements by Guertler1 Smith and Hamilton were the first to study the copper-titanium alloys, but owing to the presence of large amounts of impurities their data are inconclusive. M. A. Hunter and I. W. Bacon later investigated the electrical conductivity and the temperature coefficients at room temperature. The present writers reported their results2 in November, 1930. Recently Kroll published data on copper-titanium alloys3 which agree closely with those obtained by the authors. Material Tested The titanium used was in the form of square sintered rods, the analysis of which runs as follows: titanium, 95.4 per cent; silicon, 0.7; iron,l.8; aluminum, 1.4; copper, 0.1; manganese, 0.05. The ingots were prepared in a high-frequency induction furnace in air atmosphere. Their analyses are given in Table 1. Up to 5 per cent titanium the alloys could be forged. Above that amount they were red short and broke up easily. Owing to their rapid absorption of nitrogen and oxygen from the air, the alloys with the higher amounts of titanium tend to form surface films which greatly increase the casting difficulties, especially when small masses of metal are handled. It was, however, possible to prepare sound ingots up to 25 per cent titanium. Thermal and X-ray Investigation The cooling curves of some of the alloys are plotted in Fig. 1, and R tentative diagram constructed from these data is given in Fig. 2. From this it is evident that the solid solubility of copper for titanium is approxi-mately 4 to 5 weight per cent titanium at 878" C., which is the eutectic temperature. The melting point of copper is lowered gradually by adding

Jan 1, 1932

By Frank Alfred Fahrenwald

A complete list of the elements as given for 1915 includes 18 that melt above 1,700" C. There does not exist one complete thermal equilibrium diagram for any pair of these 18 elements. Several of these elements have been combined, in relatively small proportions, with metals of lower melting point: e.g., tungsten, molybdenum, and vanadium, in steels, or with nickel and cobalt; but of this refractory group, platinum is the only one that has been extensively studied in this connection. When one considers that, with few exceptions, the industries have made use of alloys to the exclusion of pure metals, it is reasonable to expect that alloys of this refractory group will invade fields in which the pure metals have found little application. The past neglect of this class of metals has, no doubt, been partly due to the rarity of some of them, for even though alloys of very valuable properties should be developed, their cost, unless in exceptional applications, would be prohibitive. There is also a well-founded reluctance to undertake research involving very high temperatures which, in the case of most of these elements, must be accompanied by an inert atmosphere as well. This combination of high temperatures*and control of atmosphere within the heating chamber of any furnace of present design is a difficult problem. It is the purpose of this paper to give results,' of an investigation of one binary series of this group, in which methods were employed that avoided the usual difficulties accompanying high-temperature alloy investigation. , Historical.—The alloys of tungsten and molybdenum have received very little consideration during the past. A review of available literature reveals very little regarding this system, and that which has been recorded seems to be at fault. Mennicke2 writes, "Die bekannteste Verbindung

Jan 1, 1917

By F. H. Vitovec, D. W. Hoeppner

Pusk-pull fatigue tests were conducted on copper tricrystals of 99.988 pct purity to ascertain the role of grain boundaries in the initiation and propagation of fatigue cracks. Significant differences in behavior were found for specimens which possessed different transverse-boundary misorientation. In speciwens with low boundary angles cracks initiated within the transverse boundary, while higher angles led to transcrystalline fatigue failure. It is suggested that at low angular misorientation moving dislocations may interact with dislocations of the boundary or dislocations present in adjacent pains on favorably oriented glide planes, thus initiating a fatigue crack. MANY fatigue studies have been concerned with fatigue-crack initiation within grains and the mechanism causing initiation and propagation1-3 Although the initiation of fatigue cracks in or near grain boundaries of pure metals has been observed and reported in the literature, the mechanism of this phenomenon has received little attention.4"11 EXPERIMENTAL PROGRAM Testing Procedure. To investigate the role of grain boundaries regarding initiation and propagation of fatigue cracks, copper tricrystals were tested in push-pull. Axial-stress tests were used to avoid the stress gradients introduced by stressing of some other nature. Copper was selected as the most suitable test material since extensive work has been done on the formation of fatigue-induced slip in copper. Tricrystal specimens were used to provide a grain boundary geometry with one boundary transverse to the principal stress axis and one or more boundaries nearly coincident with the direction of maximum resolved shear stress. The boundary energy and the slip characteristics in the vicinity of the transverse boundary depend on the relative orientations of grains across the boundary. Testing was done at room temperature, at a frequency of 700 cpm, in an atmosphere of either air or argon. It is known that the incidence of grain boundary cracking increases with increasing temperature and decreasing frequency.4 The fatigue machine applied a uniaxial tension-compression load to the specimen by a flat spring which was actuated by an adjustable eccentric. Use was made of an adjustable head and a load cell which had strain gages mounted at 120-deg intervals around its periphery, thus providing a means for eliminating any detectable superimposed bending moment. A clip gage was mounted between the gripping heads to record the cyclic strain amplitude applied to the specimen. Each specimen's hysteresis characteristics were recorded by supplying the load and strain signals to an oscilloscope. Microstructural changes were observed and recorded with an optical microscope which was mounted on the fatigue machine. Immediately prior to insertion in the machine, each specimen was chemically polished. Extreme care was exercised while inserting the specimen in the test machine to avoid either bending the specimen or introducing a mean load. Each specimen was stressed in the tensile direction first and subsequently the load was reversed. Specimen Preparation. Tricrystal fatigue specimens of 99.988 pct Cu were grown from the melt using a modified Bridgman technique. Ingots about 1 in. high were grown to the configuration shown in Fig. 1. Upon sectioning the resultant ingot, several specimens were provided with an identical shape and grain boundary orientation. Except for plane sectioning and final polishing, this method eliminated machining the specimens. The apparatus used for growing the tricrystals is shown in Fig. 2. A spectrographically pure graphite mold, Fig. 3, was inserted in a vycor tube which was mounted on a vertical zone-refining table. Prior to insertion in the tube the mold was assembled as follows. The insert and graphite rods were fixed in position in the mold. The copper was then placed in the mold and the entire mold assembly was positioned in the vycor tube, Fig. 2. At this point the tube was sealed, evacuated, purged with helium gas, and finally placed under a slightly positive helium pressure. The heating coil was then adjusted to melt the copper charge, after which it was raised at a speed of 4 in. per hr. It is also possible to use

Jan 1, 1964

By P. A. Witherspoon, S. P. Neuman

Evaluating the permeability of a caprock overlying a potential gas storage reservoir is a very critical problem. Pumping water from the reservoir can be used as an evaluation tool in analyzing this problem. Fluid level changes that occur in the aquifer us well as in the caprock can be measured with appropriately placed wells. If the leakage of water from the caprock into the aquifer is considerable, the effects will be apparent in the aquifer. If the leakage is slight, however, it will not be possible to detect it with certainty from observations in the aquifer alone. Fluid level measurements in the caprock must be relied upon. and improved methods of analyzing such effects have been developed which are based on a theoretical analysis of fluid flow through a caprock of infinite thickness. An example applying these methods to field data is discussed. INTRODUCTION One of the most critical problems in evaluating an aquifer gas storage project is determining the tightness of the caprock overlying the formation to be used as the storage reservoir. A formation that has previously held oil or gas obviously has a suitable caprock, but an aquifer that contains only water gives no such assurance. A number of aquifer projects in the United States have been troubled by gas leaking out of the intended storage zone, and the ensuing difficulties have led to the development of new evaluation methods. One of these new methods is pump testing wherein water is removed from the aquifer at some controlled rate prior to injection of gas. This fluid withdrawal causes a pressure drop to move out through the aquifer for considerable distances in a matter of days or weeks. Depending on the properties of the caprock, a pressure transient can also pass upward (as well as downward) through the caprock layers adjacent to the aquifer. Thus, if the operator has placed observation wells at appropriate distances from the pumping well, the rapidity with which the pressure transients reach different points in the system can be used to investigate the fluid transport properties of both the aquifer and its caprock. The usefulness of pump testing has been recognized by groundwater hydrologists for many years as a means of determining the potential yield and properties of aquifers used in water supply They have introduced the term "leaky aquifer" for a system in which an aquifer is overlain (or underlain) by semipermeable caprock layers. The ease with which water leaks into the aquifer during pumping can, of course, be very beneficial in bringing additional water to the pumped well. Hydrologists have therefore devoted considerable attention to this prob-lem. From the gas storage standpoint, however, the tighter the caprock layers that overlie the intended storage reservoir, the better are the conditions for minimizing or eliminating any vertical migration of gas. Thus, after a suitable geologic structure has been found, the emphasis in aquifer storage projects is in determining that the caprock is tight. Attention has recently been focused on the use of pump testing as one approach to solving this problem.23,24 This paper presents a further development on evaluating the permeability of a slightly leaky caprock when the caprock is of infinite thickness. From the practical standpoint, this means that the caprock layers are thick enough that pressure transients do not reach the outer boundaries of the system during the pumping test. In a subsequent paper, an analysis of the case where the caprock is of finite thickness will be presented. PREVIOUS WORK ON LEAKY AQUIFERS Jacob" developed a partial differential equation describing the flow of water in an aquifer of permeability k that is overlain by a leaky caprock of permeability k'. Fig. 1 shows a schematic cross-section of the system under consideration. One of his principle assumptions was that if k > > k', the direction of flow is essentially vertical in the caprock and horizontal in the aquifer. Neuman19 confirmed the validity of Jacob's assumption using a mathematical model. Another assumption was that a permeable source layer overlies the caprock (Fig. 1) and is able to maintain a constant hydraulic head at the upper boundary of the caprock. By neglecting the effects of compressibility within the caprock, Jacob1* developed a solution for a bounded circular aquifer. Later, Hantush and Jacobx5 used the same assumptions to solve the case of an infinite radial aquifer that is pumped at a constant rate. Their solution may be expressed in di-mensionless parameters by

By R. N. Thurston, E. A. Nesbitt, G. Y. Chin

A general relalionship between the amount of glide shear (due to slip) and the macroscopic shape change has been developed. Since the deformation can be large, finite strain analysis is employed. In this treatnzent, the shape change is expressed by a deformation gradient matrix, F - [axi/aXj1, where X and x refer to the initial and final positions of a particle. This matrix is readily evaluated in terms of the amounl of glide shear, a , and the direction cosines of the slip-plane normal and the slip direction for a single active slip system. In the case of deformation from slip on several systems, the product F(a) of the several deformation gradient matrices is first calculated. Then, by assuming that the final configuration is reached by a long series of small shears of magnitude a, occurring more or less alternately in the several slip sys- IN problems of crystal plasticity, it is often necessary to relate the amount of glide in the operating slip systems to the macroscopic strain components. This relationship is relatively simple when only small strains are considered. In this case the separate strain contributions from several slip systems are additive. Such a procedure is incorrect in the case of large plastic deformation. To the authors' knowledge, no general treatment of the latter problem has appeared in the literature. In specific instances, Mark, Polanyi, and schmid3 have derived the relationship between glide and axial elongation during tensile pulling of a single crystal when only one slip system is active. A similar relationship for duplex slip was worked out by v. Goler and Sachs.4 Taylor and Elam5 have also studied in detail problems of large plastic deformation. Due to uncertainties as to the slip systems at the time, terns, the final deformation gradient matrix can be obtained. Mathematically, the problem is reduced to finding the limit of F(dN as N — m while a — 0 in such a way that the Product Na = 0, a ftnite conslant designatzng the accunzulated amount of shear. It turns out that this limit is simply , where F, is the matrix whose elements are the coefficients of a in the F(a) matix. Application is made of the present treatment to a fcc crystal of Perlrzalloy compressed on the (110) plane and contrained to elongate zn the [112] directzon. In addition, it is shown that one rnay readily obtain from the general analyszs the well-known formulas relating elongation, amount of glide shear, and amount of lattice rotation for crystals deforming by single and double slip under tension. however, they were mainly concerned with proving that slip occurs on {111}(110) systems in fcc metals. The present general treatment arose out of recent investigations of magnetic anisotropy in cold-worked Fe-Ni alloys6 as well as some related problems in strength anisotropy of single crystals.' Detailed application of the general analysis will be provided in the case of a Permalloy single crystal compressed on the (110) plane and constrained to elongate only in the [112] direction. It will be seen that, as expected, the small strain approximation leads to significant errors after moderate straining. It will also be shown that the present general treatment yields the formulas of Mark et al. and of v. Goler and Sachs in those specific cases. GENERAL CONSIDERATIONS Consider a homogeneous deformation in which a material point initially at moves to both positions being referred to the same set of Cartesian axes. Such a deformation can be

Jan 1, 1967

By D. E. Thomas, C. E. Birchenall

ALTHOUGH Eoltzmann gave a mathematical solution for the diffusion equation (for planar diffusion in infinite 01. semi-infinite systems only) in 1894 allowing for variation of the diffusion coefficient with a change in concentration, it was not until 1933 that this solution was applied to an experimentally investigated metallic system. The calculation was carried out by Matano' on the data obtained by Grube and Jedele3 for the Cu-Ni system. Since that time concentration dependence of the diffusion coefficient has been demonstrated for many pairs of metals. However, the nature of this dependence has never been fully elucidated. Many investigators have suspected that these variations could be related to the thermodynamic properties of the solutions, one of the earliest explicit statements being contained in a discussion of irreversible transport processes by Onsager' in 1931. Development along these lines has been greatly retarded by the lack of reliable data on the variation of tliffusivity with concentration, the paucity of the thermodynamic data for the same systems at the same temperatures and compositions, and an incomplete understanding of the relation of the thermodynamic properties of the activated state for diffusion to the bulk thermodynamic properties. The last factor has been discussed by Fisher, Hollomon, and Turnbull.5 In many instances where data exist, it is difficult to know which are acceptable. This problem probably applies more strongly to diffusion data than to activity measurements. For instance, four sets of observers"-" have reported self-diffusion coefficients for copper. The average spread between extreme results is a factor of about four, though the individual sets of data are self-consistent to about 20 pct. Thus one or more factors are out of control, at least in these experiments, making estimates of internal error unreliable. The most reliable diffusion data in most systems have resulted from the use of welded couples with a plane interface from which layers for analysis are machined parallel to the interface after diffusion. The layers are analyzed, and the result is a graphical relation between distance and concentration, usually called the penetration curve. Given the same set of analytical data and distances and following the same procedure in computation, different observers will generally produce diffusion coefficients which vary appreciably, especially at the extremes of the concentration range. Experiments must be carefully designed so that the precision is good enough to answer a particular question unequivocally. In the first calculation of the dependence of the diffusion coefficient on concentration in the metallic solid solution Cu-Ni, Matano found that the coefficient was insensitive to concentration from 0 to 70 pct Cu, after which it rose more and more steeply to some undetermined value as pure copper was approached.' The same behavior was reported for Au-Ni, Au-Pd, and Au-Pt.* The data used were those of Grube and Jedele which were very good at the time, but are not considered particularly good by present standards. Furthermore, the method of calculation makes the ends of the diffusion coefficient-concentration curve unreliable. For better reliability, the high copper end of the curve has been checked by incremental couples, where the concentration spread is 67.7 to 100 atomic pct Cu. The implication of the curves calculated by Matano was that diffusion is very concentration sensitive in one dilute range of this completely isomorphous system and hardly at all in the other. Matano's result is confirmed. Later Wells and Mehll0 published data on diffusion in Fe-Ni at 1300°C, which represent a thorough test of the shape of the concentration dependence curve. They ran couples with the following ranges of nickel concentration: 0-25 pct, 1.9-20.1 pct, 0-20.1 pct, 20.1-41.8 pct, 0-99.4 pct, and 79.3-99.4 pct. Although the trend of the data indicates an S-shaped concentration dependence, their curve was drawn to the pattern set by Matano. Their original data have been recalculated for the 0-99.4 and 79.3-99.4 pct couples. Wells and Mehl's points and two independent recalculations from the raw data are plotted in Fig. 1. What appears to be the best curve is drawn through them. This curve shows little sensitivity to composition in both dilute ranges with a strong dependence at intermediate composi-tions.? Similar experiments on the Cu-Pd system are reported here at temperatures where solubility is unlimited. These lead to the same type of concentration dependence for the diffusion coefficients as was found upon recalculation of the data for the Fe-Ni system. Experimental Procedure Cu-Pd: The concentration dependence of the diffusion coefficient may be determined by the use of

Jan 1, 1953

By G. G. Luther, C.E. Jackson

The relationship between hardenability and weldability has been mentioned many times. The ease of making a hardness survey has led to its wide use as a criterion of weldability and with a given class or series of steels for which the ductility under weld heat-treatment is known the test is entirely suitable. However, in studying a new steel the hardness survey is of less value, since it is based on the assumption that there is a correlation between hardness and ductility. Several attempts have been made to derive indices for weldabilityl and hardenability from the chemical analysis. In most of these proposals the effects of the different elements were considered to be additive, but recently Grossmann2 proposed a novel method of calculating the hardenability of a steel from its chemical composition. In the method proposed, a steel is considered as having a base hardenability due to its car- bon content and grain size, and this base hardenability is multiplied by a factor for each chemical element present. The final product is the hartlenability. The purpose of the present study is to point out the possible relationship between hardenability and the effect of welding on hardness and ductility for a number of plain carbon and low-alloy steels. The composition of the 25 steels reported in this paper as determined by chemical analysis is given in Table I. To eliminate the effect of variations in finishing temperatures, the 1/2-in. plates were normalized after heating for one hour at 1650°F. in a con-trolled-atmosphere furnace. In making comparative tests, it is essential to hold conditions as Constant as possible; hence full automatic weltling control was used. The electrodes, all supplied by

Jan 1, 1943

By G. G. Luther, C. E. Jackson

The relationship between hardenability and weldability has been mentioned many times. The ease of making a hardness survey has led to its wide use as a criterion of weldability and with a given class or series of steels for which the ductility under weld heat-treatment is known the test is entirely suitable. However, in studying a new steel the hardness survey is of less value, since it is based on the assumption that there is a correlation between hardness and ductility. Several attempts have been made to derive indices for weldabilityl and hardenability from the chemical analysis. In most of these proposals the effects of the different elements were considered to be additive, but recently Grossmann2 proposed a novel method of calculating the hardenability of a steel from its chemical composition. In the method proposed, a steel is considered as having a base hardenability due to its car- bon content and grain size, and this base hardenability is multiplied by a factor for each chemical element present. The final product is the hartlenability. The purpose of the present study is to point out the possible relationship between hardenability and the effect of welding on hardness and ductility for a number of plain carbon and low-alloy steels. The composition of the 25 steels reported in this paper as determined by chemical analysis is given in Table I. To eliminate the effect of variations in finishing temperatures, the 1/2-in. plates were normalized after heating for one hour at 1650°F. in a con-trolled-atmosphere furnace. In making comparative tests, it is essential to hold conditions as Constant as possible; hence full automatic weltling control was used. The electrodes, all supplied by

Jan 1, 1943

By J. H. Hollomon, J. C. Fisher

THE fundamental problem concerning the fracture of both crystalline and noncrystalline solids is the divergence between the actual and the theoretically computed fracture stresses; the stress required for fracture, computed from the forces between atoms, is many times that actually observed with commercially available materials. Computations have been made1, 2 indicating that the theoretical fracture stress should be from 100 to 1000 times that actually observed. There are other incidental problems that have arisen, perhaps the most important of which is the problem of size effect often observed in the measurement of the fracture stress of non-metallic solids, such as glass. Another effect concerns the relation between deformation and the stress required for fracture. Experiments recently made with pearlitic steels indicate that the tensile fracture stress increases with increasing tensile deformation and decreases with increasing prior compressive deformation. This marked anisotropy of the effect of deformation on the fracture stress is in contrast with the essential isotropy of the flow stress. Early experiments on rock salt crystals by Joffé, et a13 have illustrated the effect of tensile strain in increasing the tensile stress required for fracture. A consistent theory must be constructed that will explain not only the divergence of the measured fracture stress from the theoretical value, the size effect, and the effect of deformation; but will also explain and predict the effects of combined stress, of temperature, of strain rate, and of metallurgical structure. With respect to this last variable, it has now been relatively well established that the carbide particle is the element of structure controlling the fracture of steels. In steels carbide particles vary in shape from plates to spheroids. It has been found that not only does the magnitude of the fracture stress depend upon the size and shape of these particles, but also its dependence upon strain is affected by changes in shape and size of the particles. The precipitate resulting from the age hardening of duralumin also modifies the fracture stress of this alloy. Nonmetallic inclusions greatly lower the fracture stress of all metals. The magnitude of the lowering depends on the size and shape of the included particles. In glass, the major source of the low fracture stress has been fairly conclusively demonstrated by Griffith4 to arise from real flaws or cracks present in the material. He showed that the stress concentration at the end of a defect was of sufficient magnitude to raise the stress at this location to a value comparable with the theoretical strength of the material. If the cracks were reduced in size, Griffith showed that the observed fracture strength increased by virtue of the decrease in stress concentration. He also pointed out that cracks disposed perpendicularly to the applied tensile stress would be more effective in lowering the fracture

Jan 1, 1947

By Junius D. Edwards

WHEN Charles Martin Hall invented the electrolytic process for the production of aluminum, one basic requirement was a supply of pure alumina. Now, more than 50 years later, the same requirement still holds, and there are no technical straws in the wind to indicate that this situation is likely to change. Long ago, Aluminum Company of America took a number of steps to see that this country should not face shortages of alumina and aluminum in time of emergency. As one step to this end, the Company developed and used a supply of bauxite in South America in order to conserve dwindling Arkansas reserves. Another, and more a program than a step, was to investigate every type of process that might bear the possibility of utilizing the low-grade ores of aluminum that are more plentiful than bauxite. It is a long story, but the investigation included electric-furnace processes, acid extraction processes and alkaline extraction processes, with or without preliminary furnace and sinter treatments. Some of the processes developed were operated in commercial production. This work has been continuous since the first World War. The turn of the wheel came sooner than was expected when, in 1942, bauxite shipments through the Caribbean became uncertain and hazardous through the operation of submarines. The tremendous requirements of aluminum for the war effort were consuming high-grade bauxite at an alarming rate and a national search for substitutes began. High-grade bauxites, for technical and economic reasons, have been defined as those containing about 7 per cent or less of silica. While bauxite of this grade was limited in amount in the United States, there was in sight a substantially greater volume of low-grade, high-silica bauxite not amenable to economic treatment by the Bayer process, but richer in alumina than the clays or other available low-grade ores. It was at this point, and after several years of intensive research on this particular phase of the problem, that Aluminum Research Laboratories brought to completion the invention of one of its research staff, which made practical the utilization of large reserves of low-grade bauxite in America. BAYER PROCESS The Bayer process has been almost universally used for the 'production of alumina from bauxite. This process, in outline, comprises the digestion of finely ground bauxite in a solution of caustic soda (NaOH) under pressure and at an elevated temperature. Alumina in the bauxite goes into solution as sodium aluminate and the impurities-oxides of iron, titanium and silicon-remain undissolved. The undissolved residue, known as "red mud," from its color, is separated and run to waste. Aluminum hydrate (A1202.3H20) is precipitated from the solution of sodium aluminate, and the liquor containing the caustic soda and some alumina is returned to the process for digesting more bauxite. The aluminum hydrate, after washing, is calcined at a high temperature to produce

Jan 1, 1945

By John C. Shyne, Eric R. Morgan

WHEN temper rolled low carbon sheet is stored at room temperature before use, changes take place in its mechanical properties. This phenomenon is known as strain aging. Normally these changes are observed by means of Olsen ductility and Rockwell B hardness tests but' unfortunately such tests are relatively insensitive. Strain aging can be readily observed in the laboratory by means of tensile tests. The tensile test method is even more sensitive if the aging is allowed to take place in specimens which have been prestrained in tension rather than by cold rolling. In fact one of the great benefits of commercial temper rolling is that it introduces a residual stress pattern which tends to mask the effects of strain aging when the sheet is finally used in a pressing operation. Curves B, C, and D of Fig. 1 illustrate the changes that take place during the aging of commercial rimmed steel at room temperature following straining in tension, curve A, Fig. 1. The first important change that can be observed is an increase in the stress required to produce plastic flow. This increase is, in this paper, called the aging index. As aging proceeds and the aging index increases, there is also an observable return of the yield point, an increase in the ultimate tensile strength, and a decrease in elongation to fracture. The results of strain aging in deep drawing steel manifest themselves as stretcher strains (L?ders bands) buckling, 2nd tearing. In deep drawn parts such effects are undesirable and strain aging is a serious problem. The rate of strain aging of a steel depends upon its composition and prior mechanical and thermal history? but for most temper rolled ope

Jan 1, 1958

By G. W. Josephson

SINCE the earliest years of recorded history the durability and protective qualities of asphalt and tar have been known and utilized. The mummies of early Egyptian kings were coated with asphaltic materials, and asphalt mastics were common in Babylon. However, according to various authorities, it was not until 1780 that the first forerunner of present-day composition roofing was discovered,1 and the large-scale expansion of the industry has taken place since 1900. HISTORICAL BACKGROUND Development was stimulated by the growing popularity of manufactured gas. The by-product tars from these operations accumulated in large waste piles that had no commercial value and were a source of annoyance and expense to the gas companies. These tars were offered at very attractive terms to anyone who would take them away, and the infant "composition" roofing industry benefited accordingly. At first the process consisted simply of dipping sheets of burlap or other fabric in a tank of molten tar or asphalt. The long strip of roofing was rolled up for convenient handling, and the early experimenters soon found that it was necessary to dust the surface of their new "roll roofing" with some kind of material that would prevent the tarred surfaces from sticking together. Any finely divided, low priced local material like sand or fine gravel served this purpose and from its application the roofing-granule industry grew. As time went on other characteristics of a properly applied mineral surfacing were recognized. As with wooden shingles, tar and asphalt can be ignited by flying embers from near-by fires. It was found that by increasing the largest particle size to the maximum that could be successfully embedded, and by properly controlling the amount of smaller fractions retained in the grade so as to cover the asphalt completely a Nigh degree of resistance to fire was attained. Needless to say, in this early period asphaltic roofing needed such selling points to gain acceptance in a field dominated by wood, slate, and tile. When asphalt is protected from the sun's rays it is extremely durable, but when exposed to sunlight its surface is comparatively rapidly decomposed. Completely covering asphalt roofing with mineral surfacing protects the asphalt and adds materially to the life of the roof-conservatively, 500 per cent. Slate Granules In the production of split slate shingles a large quantity of waste slate is accumulated, as 60 to go per cent of the rock quarried is discarded. This waste was a continuing burden on the industry. Some was marketed as ground slate flour for filler in paint and linoleum, and there is a record that as early as 1906 two carloads of coarse screenings from slate flour were marketed as granules for surfacing roofing. This new field appeared to the slate pro-

Jan 1, 1944

By Daniel N. Beshers, Ralph H. Condit

An expression for the diffusion coefficient of inlerstitial atoms in the bcc lattice is derived which takes account of jumps between tetrahedral and octahedral interstices. The results are applied to the case of carbon diffusion in iron to find values for jump frequencies of the types, octahedral-tetrahedral, tetra-hedral-tetrahedral, and tetrahedral-octahedral. ThERE has been renewed discussion1-5 concerning the mechanism of diffusion of interstitial atoms in the bcc lattice. This lattice has both octahedral and tetrahedral interstices, and it has been pointed out that the interstitial atoms may make jumps of the octahedral-octahedral type, hereafter called o-o, octahedral-tetrahedral, o-t, and tetrahedral-tetrahedral, t-t. Doubt has been cast on the applicability of snoek's8 model for internal friction to some interstitial migrations in bcc lattices wherein he proposed 0-0 jumps as the principal mechanism of migration. Also, the anomalous excess diffusion of carbon in iron at high temperatures4 has been ascribed to t-t jumps,2,3 even though the sites normally occupied are octahedral. In this paper an expression is derived for the diffusion coefficient, taking account of the three types of jump mentioned. The relationship of the coefficient to the corresponding internal friction relaxation time is found for the limiting case corresponding to Snoek's model. These results are applied to the case of carbon in iron. In Fig. 1, the structure is shown with the locations of the interstices. Each o site has four t site neighbors coplanar with it, and each t site lies halfway between two o sites. There are four t site first neighbors of a t site lying at the vertices of a tetrahedron. Let the average fraction of atoms occupying o sites and t sites be No and Nt, respectively. Since there are six o and twelve t interstices per unit cell having a lattice parameter, a, the concentration of interstitial atoms will It will be assumed that three types of jumps may be distinguished, t — t, t — o, o — t,for which the jump

Jan 1, 1968

By Thomas J. Menner

Jan 1, 1983

By H. A. Saller, F. A. Rough

Studies of the V-U system have been made to determine the constitutional diagram. The diagram is fairly simple, since no intermediate phases are formed. Additions of vanadium lower the uranium melting point and allotropic transformation temperatures. The maximum solid solubility of vanadium in uranium is 12 atomic pct and the maximum solid solubility of uranium in vanadium is about 4 atomic pct. WITH the advent of high purity iodide vanadium,' it became of interest to study the alloying characteristics of vanadium and uranium. Alloys were prepared covering the entire system and were studied using a combination of metallographic examination and thermal and X-ray analyses. As a consequence of this work, the V-U constitutional diagram was evolved. In order to prepare suitable alloys for equilibrium-phase studies, it is desirable to have the purest possible starting materials. For this investigation, calcium-reduced vanadium, de Boer process crystal-bar vanadium, and good Mallinkrodt Chemical Co. uranium were used. The first work was done using the best vanadium then available, that being calcium-reduced vanadium produced at Knolls Atomic Power Laboratory. Subsequently, the important features of the diagram were checked using alloys made with crystal-bar vanadium produced at Battelle. These checks are in agreement with the earlier work. Analyses of the starting materials are contained in Table I. Only minor amounts of any impurities were present in either the vanadium crystal bar or uranium. Nevertheless, care was exercised in selecting the uranium for the alloys. The procedure followed was to section the stock into suitable pieces for melting, and macroetch each piece. Only those pieces which appeared to be clean by macroetch examination were used for alloying. Experimental Procedure Melting: Alloys with 0 to 50 atomic pct V were induction melted in beryllia (BeO) crucibles under vacuum of 10 microns pressure or better. A number of remelts was frequently required to insure uniformity of the individual ingots. Although no visible reaction between the alloy and crucible was found in any of these operations, the alloys were checked by spectroscopic analysis. Only trace amounts of beryllium were detected in the alloys. Alloys containing 50 to 100 atomic pct V were arc melted in a water-cooled copper crucible under an atmosphere of helium. A tungsten electrode was used. Ingots of 10 to 15 g were produced readily by turning and remelting a few times. Such small ingots were necessary because high purity vanadium was not available in appreciable quantities at this time. After melting, the ingots were examined metal-lographically for segregation and were consolidated by remelting before fabrication was attempted. Spectroscopic analyses showed that only small amounts of tungsten were picked up during the arc-melting process. Fabrication of Alloys: Alloys containing up to 20 atomic pct V could be fabricated by rolling in air at 600°C with 5 min between passes. After each pass through the rolling mill the alloys were returned to

Jan 1, 1954

By J. Moore Samuel

Before the year 1909, no measurements of dust losses and flue gases had been made at the Copper Queen Reduction Works, at Douglas, Ariz. At that time the "unaccounted" loss of the smelter had reached a figure that drew attention to the possibility of large stack losses. The measurement of dust losses had not been studied much at that time, and estimates of dust losses, with a few exceptions, were the results of tests by what has been termed '(chemical methods," that is, the aspiration and filtration through cotton, wool or some similar medium of a few liters of gas, the dust being weighed on a chemical balance, and the volume of gas computed from the volume of aspirator water. The possible error of this method is enormous, and the probable error much too large for reliable results. Since the object of these tests was to discover not only the extent of the dust losses, but also the possibility of profitably reducing them, it was decided that the method adopted must have the following characteristics: 1. The sample of gas taken for filtration must be representative of the whole volume of gas, that is to say, it must have the same dust content. This condition will be fulfilled in all cases if the gas sample is drawn off without change of rate of flow. Then it will contain the same proportion of entrained or suspended particles of dust as the original gas. In the case of flue gases which contain only fume, at a temperature above or near the sublimation point, this condition need not be adhered to since the fume will not behave as entrained particles, but more nearly as a true gas. Consequently the gas sample in this case only may be drawn off at any rate of flow. 2. The gas sample taken must be large enough (a) to yield sufficient dust to permit of accurate chemical and physical examination; (b) so that the ratio of the total volume of gas passing through the flue to the volume of gas sample taken shall not be too large.

Jan 1, 1917

By S. G. Cupschalk, F. A. Dahlman, J. J. Wert

Studies have been undertaken to determine the indicidual effects of particle size, degree of long-range ovder, antiphase domain size, and root mean square stran on the microhardness and yield strength of ordered alloys. Dnta have been analyzed for Cu3Pt initzally ordered to a value of 0.82 and after deformations of 1 and 6 pct. It was observed that deformation fleatly reduced the degree of long-range order. Furtherrnore, wztkin this range of relatively small deforntntlons, the average particle size changed very little while the antiphase domain size was greatly reduced. Smultaneosly, the mcrohardness changed by a factor of two durzng the deforrtation process. PREVIOUS studies have reported some of the effects of cold work on the broadening of X-ray diffraction peaks. These investigations were performed on powder and wire samples representing both ordered and disordered states; i.e., the specimens were initially studied in a severly cold-worked condition. By comparing the difference in line shape between the annealed and cold-worked peaks, fundamental information was obtained concerning particle size, strain distribution in different crystallographic directions, degree of long-range order, and change in antiphase domain size. Considerable theoretical work has been done concerning the analysis of diffraction data obtained from cold-worked metals. Stokes' expressed the change in diffraction profiles in terms of Fourier coefficients. Much of the work in this area has been summarized by warren2 in an extensive review article concerning the analysis of plastic deformation by X-ray diffraction. Cohen and Bever3 applied these techniques in studying the effects of cold work on alloy systems exhibiting long-range order. They utilized the Fourier coefficients of fundamental peaks in conjunction with those of the superlattice peaks to determine the change in antiphase domain size. Little work of this nature has been reported for ordered systems that have undergone small degrees of plastic deformation. The purpose of this investiga-tion was to determine the effects of small deformations in such a material with respect to particle size, strain distribution in various crystallographic directions, antiphase domain size, degree of long-range order, and hardness. EXPERIMENTAL PROCEDURE CusPt was used for the initial investigation since the order-disorder transformation takes place with- out a change in crystal structure. The transformation is readily detectable via X-ray diffraction techniques due to the large difference in the scattering factors of copper and platinum. Additionally, the alloy is relatively low melting (approximately 1300°C) and is easily deformable in both the ordered and disordered states. 1) Specimen Preparation and Cold Working. A 100-g, 12-in. diam., cylindrical specimen of Cu3Pt was prepared by melting and casting 99.99 pct pure Cu and Pt i.n vacuo. Prior to any mechanical working, the material was homogenized in a vacuum for 60 hr at 100O0C, and surface defects were removed by machining to a depth of approximately 116 of an in. The material was then cold-rolled, with an intermediate anneal, into a strip approximately 12 in. wide by 14 in. thick. Straightening and flattening removed another 0.025 in. from the thickness. After a recrystallization treatment at 750°C for 30 min, the specimen was slow-cooled from 55OoC, at the rate of 6°C per hr, down to 150°C to induce superlattice formation. This treatment yielded an ASTM grain size of 7 and a degree of long-range order equal to 0.83 0.06. After obtaining X-ray and Knoop hardness data, the sample was cold-rolled approximately 0.75 pct in one pass through a hand-operated jewelers' mill. X-ray and hardness data were again obtained and the specimen was reduced an additional 5.41 pct in a single pass through the mill. 2) X-Ray Measurements. The specimen was examined in the ordered condition and after the two degrees of cold working previously mentioned using a General Electric XRD-5 unit equipped with a spectrometer and scintillation counter. Using Mo-Ka radiation with a zirconium filter, six orders of the 100 reflection were obtained. It was anticipated that point counting would be necessary for an accurate determination of the low-intensity peaks and tails: however, it was demonstrated that, by using a scanning speed of 0.2 deg per min and the appropriate time constant, the recorded data were sufficiently accurate. Thus, for ease of experimental procedure, all peaks were recorded on chart paper. Specimen position in the holder was considered to be insignificant after making a series of measurements of the same peak area in different positions with respect to the beam. Since peak overlapping did occur at high values of 20, it was necessary to separate the peaks graphically prior to analyzing the data in order to minimize this source of error. The peak tails were also carefully drawn to obtain the best possible data. Fourier coefficients of the line profiles were calculated on an IBM 7072 computer, and graphical meth-ods2j3 were employed in analyzing the results. For this type of calculation, in which the line profile is represented by intensities taken at set intervals, the intervals selected must be sufficiently small to give an accurate representation of the line profile. It was decided that for 20 = 0.02 deg the line profiles were

Jan 1, 1967

By E. V. Foran

Among the many newer disclosures that have accompanied the petroleum industry's progressively deeper exploratory drilling is the increased frequency with which the operators are encountering reservoirs of which the products at the time of discovery exist wholly or predominantly in the gas phase. The popular term "distillate pool" is most often used within the industry when referring to this type of reservoir, and in this paper the term is used in that sense. These pools have been observed in many different surroundings. In some places the entire reservoir is filled with the wet gas containing its water-white condensible products in the gas phase but apparently with no dark oil or other hydrocarbon liquids in the reservoir. The 8200-ft. horizon in the Big Lake field in West Texas is a good example of this type. At other places the wet, or distillate-bearing, gas may form a large gas cap directly overlying and in contact with an extensive reserve of dark oil. Oklahoma City, Kettleman Hills, the 8500-ft. horizon in the Big Lake field, and many others are examples of this type. There are also many places in which the wet distillate-bearing gas fills nearly the entire reservoir space, while the dark oil occupies only a narrow, bandlike zone, which marks the boundary of the reservoir along its lowest productive contour. The heavier components of these deep pool gases are generally water white in color, a large portion of which are condensible in a conventional oil and gas separator, when operated under pressures of the order of 400 to 800 Ib. per sq. in. Although the water-white products recovered in the separator sometimes have end points of 650" F. and higher on standard Engler distillation, they occur only in the gas phase in the reservoir. It is entirely possible that even some small amounts of dark colored reservoir components may occur in the gas phase when the temperature is 300" F. and pressures exceed 6000 Ib. per sq. in. Close approaches to these conditions have been encountered in the deepest producing wells. The condensible content of the deep reservoir gases of the many pools that have been tested or produced to date shows a wide variation in both condensible content, per 1000 cu. ft. of gas, and physical character.

Jan 1, 1939