Search Documents

Jan 1, 1940

Jan 1, 1940

Jan 1, 1940

Jan 1, 1940

Jan 1, 1940

By Jr. E. H. Dix.

In selecting the subject, "Acceleration of the Rate of Corrosion by High Constant Stresses,'' for the 1940 Institute of Metals Division Lecture, I have been influenced by its highly theoretical and speculative nature as well as by the great practical importance of the correct interpretation and appraisal of the engineering significance of the phenomenon. I lay no claim to originality in talking on this topic, since it has been the subject of discussion for over a quarter of a century. However, as our technical thinking advances, old phenomena take on new and more significant aspects. In our studies at the Aluminum Research Laboratories, we have encountered many perplexing and seemingly contradictory results and have been fascinated by the problem of dovetailing all of the many pieces of information to make a complete picture. I hope that in the brief period of this lecture I may share with you some of the enjoyment we have had in attempting to meet this challenge to metallurgical science. The emphasis generally has been placed on the spontaneous failure by cracking of a metal under the combined action of high stress and corrosion, hence the present-day use of the terms "stress corrosion" or "stress corrosion cracking." The older term "season cracking," applied more specifically to copper-base alloys, includes, by inference, the element of time. For the purpose of the discussion today I will use "stress corrosion" to cover broadly the effect of stress in accelerating the rate of corrosion of metals. To stay within the limits of allotted time the disrussiorl will cover the effect of high static stresses rather than repeated or dynamic stresses, which would take us into the related field of corrosion-fatigue. The subject is particularly appropriate for an Institute of Metals Division Lecture because much the same phenomenon has been observed in almost all metal systems. Thus, in addition to the season cracking of brass may also be mentioned the "caustic embrittlement" of steel boiler

Jan 1, 1940

By Clarence Zener

Theoretical considerations by one of the authors have ledl to the prediction that the dynamic internal friction of annealed metals has a broad maximum at a certain grain size. This prediction they have verified2 for alpha brass. In the present paper the theory is extended to include the precise manner in which the internal friction varies with grain size on either side of the maximum. New experimental data are presented to test this extended theory. The physical basis for the type of internal friction that is being investigated lies in the interplay between stress and temperature. Just as local fluctuations in temperature give rise to local fluctuations in stress, so likewise fluctuations in stress give rise to fluctuations in temperature. But temperature gradients give rise to necessarily irreversible thermal currents, and hence to internal friction. The theory for this type of internal friction has been developed quantitatively,³ and has been experimentally verified,4,5 for the particularly simple case of the transverse thermal currents accompanying transverse vibrations. Of more fundamental importance are the intercrystalline thermal currents accompanying all types of vibration. The variation in stress across grain boundaries arises from the elastic anisotropy and at least partial random orientation of the individual crystallites. An exact quantitative treatment of the internal friction due to these thermal currents is of course out of the question. However, much information can be obtained from general arguments.1,6 Thus when the frequency of vibration f is low, or when the grain size d is small, adjacent grains remain nearly in thermal equilibrium, and vibration proceeds isothermally with little internal friction. Again, when the frequency is high, or the grain size large, the heat flow per cycle is small, the vibration proceeds adiabatically with little internal friction. The internal friction will become a maximum in the transition region. The degree of adiabatic-

Jan 1, 1940

By John T. Norton

In recent years, the damping capacity of metals has come to be recognized as an important index of certain types of structural change. In a number of instances, this property has proved to be particularly sensitive, but insufficient work has been done to permit a general interpretation. There is considerable question at the present time as to just what factors are responsible for the internal dissipation of energy when a metal is carried through a cycle of stress, and these factors must be ascertained if the value of the damping method as a metallurgical tool is to be determined. Since internal stresses are among the chief causes of dissipation of internal energy, or damping capacity, it was thought that a study of the changes in damping during the recovery and recrystalliza-tion of a cold-worked metal might throw some light upon the subject, particularly if the damping changes were correlated with other properties that are better understood. Probably more work has been done on the recrystallization phenomena of alpha brass than on any other metal, so this was chosen as the material for the present investigation. Damping measurements on brass have been made by Förster and Köster.1 They showed that the annealing of a brass that had been cold-drawn to 50 per cent reduction produced a rapid fall in the damping, a flat minimum occurring at approximately the temperature of recrystallization. The damping remained at a low value until grain growth was well under way and then it increased rapidly with increasing grain size. Köster and Rosenthal2 continued the work in greater detail. They pointed out that the damping decrease does not correspond to the hardness drop but occurs at a lower temperature, suggesting that the relief or redistribution of internal stresses has a large effect while the actual recrystallization effect is small. They also show the increase in damping at the higher annealing temperatures. Their curves are reproduced in Fig. 1.

Jan 1, 1940

By R. A. Harrington, H. R. Nelson

The: structure of anodic films formed on aluminum and other metals under varied conditions is not well known despite the wide use of such films both as protective coatings and as the dielectric medium in electrolytic condensers. The present work was undertaken as a survey of the possibilities of electron diffraction in the investigation of anodic films, especially in comparing the properties of anodic films formed under different sets of conditions and subjected to various subsequent treatments. In addition, an effort has been made to compare surface films formed on aluminum by anodic treatment with surface films formed by heating, immersion in chemicals, etc. The elcctron diffraction method in many cases has been successfully applied to the study of thin surface films. Its principal advantage is that it can be used without removing the films from the base material upon which they have been formed. The low penetrating power of electrons, in the "reflection" diffraction method which has been used in this work, prevents the appearance of diffraction patterns from substances at depths greater than about 100 Angstrom units (lop6 cm.). Different investigations of the structure of anodic films have showll widely varying results, depending on the methods used in preparing the films. Apparently under some conditions these films are crystalline and under other conditions they are amorphous, and it has been difficult to say which type of structure will be found in any particular case. Schmid and Wassermann's1 X-ray diffraction patterns of eloxal‡ films show two broad, diffuse bands, which these investigators attribute to finely divided cubic y-Al2O3.§ Burgers, Claassen, and Zernicke³ obtained sharp X-ray

Jan 1, 1940

By L. H Levenson, Charles S. Barrett

The orientation of metal crystals and grains, both large and small, can be detcrmined conveniently and accurately without the use of X-ray apparatus. This seems to be appreciated by so few metallurgists that it may be useful to explain in detail the method we have found most satisfactory in the laboratory—the reflection of light from crystallographie etch pits. This note includes a description of a suitable inexpensive instrument and instructions for etching various metals and for plotting the data. As the stereographie projection is used to facilitate handling the data, this note may he considered as an appendix to an earlier review of the metallurgical uses of the stercographie projection.' Various methods involving etch pits have been used by other investigators. Bridgman's method2 consists of attaching the specimen to a sphere (a transparent sphere is convenient4 , holding it at arm's length, and marking a spot on it when it has been turned so that the eteh pits reflect light into the eye of the observer from a lamp standing behind him. It has an accuracy of about 2" and is extrernely rapid. A greater accuracy can be obtained by having the reflected light fall on a screen, as has been done by Czochralski,5 Chalmers,6 and Schubnikov.7 All of these methods have been intended for undeformed single crystals. Tammann and his coworkers3,8 developed a method for estimating the number of grains in a polyerystalline aggregate having orientations in a given region of the stereographic triangle and thus obtained statistical information about deformation and reerystallization textures. While the method has had some use in recent years,9 it is not applicable to the majority of problems in physical metallurgy. Smith and Mehl10 determined orientations by plotting the directions of the sides of individual etch pits on a stereographic projection. The accuracy is limited to about 3" and the method requires careful polishing and high magnification.

Jan 1, 1940

By Carl H. Samans

In their recent complete review of the subject of age-hardening,' R. F. Mehl and L. K. Jetter classify the main types of precipitation-hardening alloys under two headings, depending upon the nature of the precipitate and the character of the precipitation. In the alloys falling under the first heading, which display "continuous precipitation" and are typified by the aluminum-copper alloys, "precipitation seems to occur without much regard to grain boundaries and apparently at much the same rate at every point," although an acceleration in regions plastically deformed has definitely been established.2 In the second type of alloys the precipitation is "discontinuous," as in the copper-silver system for example, and "seems to start in restricted regions, primarily those contiguous with the grain boundaries, and to spread inward in an advancing wave." The reasons for the occurrence of these two greatly different types of precipitation are unknown, as are those for many other reactions that take place in the solid state. In the present paper only the first or "continuous" type of precipitation will be dealt with, for which a complete crystallographic mechanism has been worked out for the process of precipitation from a supersaturated solid solution of copper in aluminum. Based upon this mechanism, an explanation of the various property changes found experimentally will be developed in the second paper of this series. In later papers ('discontinuous precipitation" in copper-silver alloys will be explained, and an effort will be made to rationalize the two processes and types of precipitation. Crystallography of the Precipitation Process Through the extensive investigation by R. F. Mehl, C. S. Barrett and F. N. Rhines3 on the Widmanstätten structure in aluminum-rich copper

Jan 1, 1940

By William L. Fink and

Although age-hardening in an aluminum-base alloy containing magnesium was observed by Alfred Wilml as early as 1911, it was not until 1919 that a theory of the mechanism of age-hardening was proposed. Merica, Waltenberg and Scott12 studying the age-hardening characteristics of Duralumin at the National Bureau of Standards, came to the conclusion that age-hardening of this alloy is caused by the precipitation of CuAl2, which is less soluble at the aging temperature than at the solution heat-treating temperature. It was postulated that the initially precipitated particles were of fine colloidal dimensions but that they gradually grew in size. Maximum hardness was assumed to be attained when the particles reached a critical size. This theory was at first received skeptically, since it was a widely accepted view at that time that the solid solution state should represent the hardest state of an alloy. Shortly thereafter, Jeffries and Archer proposed their well-known slip interference theory.'13 They suggested that age-hardening was accomplished by the precipitated particles acting as keys interfering with slip along crystallographic glide planes. They showed that this mechanical theory led direct,ly to the assumption that maximum hardening would be produced by particles of a critical size, which they suggested would have a diameter approximately 10 times that of the aluminum atom. This contribution amplified and explained so rationally the age-hardening theory of Merica, Waltenberg and Scott that wide acceptance resulted. Objections to Simple Precipitation Theory From time to time certain experimental facts have been cited against the simple precipitation theory. The first of these was the discovery by Fraenkel4 that, although the electrical resistivity of Duralumin decreased during aging at higher temperatures (from 100' to 250' C.), it increased during aging at ordinary temperatures. This was not in accordance

Jan 1, 1940

By L. H. Levenson, Charles S. Barrett

Since 1925, when the preferred orientations in compressed aluminum were first determined1,2 the orientations have been described as a fiber texture in which a face diagonal, [110], of the face-centered cubic lattice is parallel to the compression axis. Subsequent experiments have seemed to confirm this texture,3-5 and a number of theories (recently summarized in references 6 and 7) have been proposed to explain it and to relate it to the behavior of single crystals during compression.3,8-11 The experiments reported in this paper prove that the orientations are not adequately described as quoted above—i.e., as a [110] fiber texture—although this is a first approximation to the actual texture. They also show that the texture cannot be thought of as a stable orientati011 that is reached and maintained after a sufficient amount of deformation. The actual texture, when studied in detail at successive stages of compression, indicates that a hitherto unsuspected range of orientations is present, both in the polycrystalline metal as a whole and in various individual grains in its interior, and that a state of dynamic equilibrium exists within this range. The reasons for the failure of previous theories to predict the actual texture become obvious when the heterogeneous nature of the flow is considered. In addition to the deformation texture, the manner of its development, and its theory, this paper presents data on the following related subjects: the identity of strain markings in aluminum; the orientation changes in aluminum single crystals during compression and a comparison of these with the conditions in individual grains of a polycrystalliile specimen; the origin of asterism ill deformed single crystals, which heretofore has been regarded as a distortion localized at the surfaces or slip planes; and the relation of plastic strain in individual grains ill all aggregate to the strain of the whole.

Jan 1, 1940

By Charles S. Barrett

Recrystallization textures—thc orientations of grains after recrys-tallization—have been studied extensively not only because of their metallurgical importance but also because of the information they yield regarding the atomic movements during reerystallization. Recrystallized grains must possess the same orientations as the submicroscopic nuclei from which they have grown, and thus they offer a direet road by which experimenters can penetrate the submicroscopic realm. Many perplexing results have come of these studies, but one of the most striking has been the apparent difference in behavior between single crystals and polyerystalline grains. For example, Burgers and Louwerse have shown1 that when single crystals of aluminum are deformed by compression and are reerystallized (at 600' C.) the new grains appear with orientations that are different from the orientations of the deformed crystal in which they grew; but polycrystalline specimens, on the other hand, appear to have a [110] fiber texture both before and after reerystallization and thus to retain their texture during recrystallization. This apparent difference between the habits of single crystals and small grains is the subject of the first part of the present research and has been explained by findings reported herein. A similar situation appears to exist for aluminum when elongated or drawn into wires, for single crystals alter their orientation 2,3 and polycrystalline wires retain their texture"5*" upon recrystallization after elongation (at least with high-purity aluminum), and it is possible that the explanation found in this paper for compression can be extended to cover elongatioil and other types of deformation as well. The second part of this paper deals with the theory of reerystallization textures. In the most successful theory for aluminum reerystallized after compression,1 much weight is given to the orientation of highly stressed nuclei in the deformed material. Taylor's "local distortions" arc assumed to exist along the slip planes where displacement has occurred, consisting of fragments rotated in a specific way with reference to the slip plane and slip direction; namely, about an axis lying in the slip plane perpendicular to the slip direction.1,7,89

Jan 1, 1940

By C. H. Mathewson, J. A. Collins

The process of plastic extension of face-centered cubic single crystalline rods is so well understood, since the notable studies of Taylor and Elam,1 that the change in orientation of any crystal by slip and rotation may be predicted from the measured elongation. It is customary to assume a very homogeneous deformation, wherein gliding lamellar arc free to rotate so that their slip direction continually approaches the axis of the rod, until a second slip system comes into operation; whereupon conjugate slip proceeds quantitatively to an end adjustment with the [121] direction in the axis of the rod. This process has been checked in malty face.-centered cubie materialsand and readily permits the preparation of specimens that have been subjtvted to various determinable amounts of shearing action. It is elearly of interest to study in such measurably strained speeimens the, development of a recrystallized structure on annealing and if possilble detect the source of nuclei from which the new crystals grow. Strain-:~nneali~ig experiments of this charactcr have been performed by others but not in detail sufficient to permit a study, for example, of the number of recrystallized grains vs. the amount of shear. In the experiments now reported, this was done with a wide variety of orientations covering nearly the entire range of slip on a single system and corresponding to amounts of shear (ratio of shearing distance to distance between lamellae at the conclusion of a test of from 0.08 to 0.57. A particular point wa-made of determining the crystallite rotation, or small angular range. assumed by the octahedral slip plane after these various amounts of shear. The, mechanism of reerystallization and the magnitude and effeet of recovery on the recrystallization process was investigated.

Jan 1, 1940

By A. I. Blank, S. E. Maddigan

Many experimental and theoretical investigations have been made on the processes involved in annealing and recrystallization of metals. Most of these, however, have treated the time element as constant and heat-treatments used have been for relatively short times. Aside from the early work by Carpenter and Elam,' the few that have investigated the effect of time on the processes have usually done so for fairly short time periods. During recent years the growing interest in long-time creep properties of metals and alloys under stress has made increasingly important a better knowledge of the recrystallization during prolonged heating. The recrystallization temperature of 70-30 brass given by Clark and White" in 1932 was somewhat above 400°F. (204°C.) but under 500" (260°C.), whereas in our laboratory recrystallization has been observed in creep specimens at 204°C., the lowest temperature of Clark and White's range. Since the creep properties are quite different above and below the recrystallization temperature, erroneous conclusions can readily be drawn by the application of short-time recrystallization data to creep problems. It was thus considered of both practical and theoretical value to make long-time annealing studies near the minimum recrystallization temperature. In this way useful additions could be made to data on recrystallization temperatures and, since the reactions in this range are relatively slow, additional insight could be obtained into the factors involved in annealing. Several comprehensive reviews of recrystallization processes have been written,3-6 therefore only a brief review of the generally accepted facts will be given here. Though the effects of heat-treatment usually are expressed as related to observations with varying temperatures, it is logical to assume that the same order of occurrence will exist with time as a variable. Thus, we may postulate that with increasing time we mill observe the following

Jan 1, 1940

By R. M. Brick

The etched mierostructures of cold-worked alpha brasses, after reductions in excess of about 20 per cent, exhibit dark lines or markings, which have been termed "deformation bands," "etch bands," and "strain markings." The terms will be used synonymously in this report. The markings are similar in their general appearance to etch figures found in cold-worked iron, aluminum and other cubic metals but in certain respects may differ noticeably. At reductions of from about 20 to 50 per cent in brass, they appear as straight dark lines of irregular spacing and at random directions in different grains, but apparently intimately associated with definite crystallographic planes, usually {ill), in each crystal. At higher reductions, the markings on the rolling plane tend to appear in a direction transverse to the rolling direction in all crystals but may bend away at considerable angles, particularly near grain boundaries. Samans,1 on the basis of X-ray and micrographic studies of two single crystals of brass, rolled to a 50 per cent reduction, showed that the lines of deformation were the edges of thin mechanical twins parallel to octahedral planes. Barrett and Levenson2-4 have made comprehensive studies of the deformation of iron and of aluminum. They show that in these two metals the deformation markings originate by division of the original homogeneous crystal into banded areas differing in orientation by an amount that increases with the degree of deformation. The markings thus cannot be twin bands, and Barrett, in a review of this subject,2 attributed Samans' results on brass to an accidental choice of deformation that left the bands in approximately a twinned relationship. On the other hand, Mathewson6 has suggested that in some cases mechanical twinning offers a better rationalization of the coverage of some (111) pole figures of worked brass than the assumption of band rotation into symmetry positions. Considering next a deformed and annealed alpha brass, the most striking aspect of its microstructure is the multiplicity of twin bands.

Jan 1, 1940

By H. l. Burghoff

During the course of preparation of crystals of alpha brass for an investigation of their creep characteristics, a number of critically oriented crystals were produced. In each of these specimens, Po, the specimen axis, lay near some important crystallographic direction or on or near the border of two possible slip systems. Because of the unusual nature of their orientations, it was considered of interest to deform these specimcns in simple tension and to study orientation changes and any unusual occurrences, different from what would be expected for specimens ill which only one slip system had distinctly the highest maximum resolved shear stress. Choice of actual operating slip system in a crystal offering an immediate possibility of two slip systems would be expected to be due to chance., defeet, or some inhomogeneity. Even in an ideally perfect crystal so oriented, the natural oscillations of the atoms about their lattice positions would give rise to a condition wherein one slip system would be farored over the other at the critical shear stress. In the large, nonperfeet crystals of the present investigation, slip on both systems, but in different portions of the specimen, might occasionally be expected. Determination of Orientation and Operating Slip System Orientations of specimens before and after deformation were determined by Greninger&apos;s back-reflection Laue method.&apos; Elam2 has shown that deformation of alpha brass crystals proeceds as with crystals of aluminum and of copper; that is, the slip plane is a {111) plane and the slip direction is a <110> direction. There are, however, four possible slip planes of {111) form and each plane has three possible slip directions, so that there is a, total of 12 possible slip systems in a face-centered cubic lattjce. It is evident that the system for which the tangential component of shear is maximum will function

Jan 1, 1940

By Paul A. Beck

While the recrystallization properties of most of the practically important metals are known in considerable detail, those of lead are still relatively little known in spite of some valuable contributions in recent years. This condition probably is due to the difficulties involved in applying the usual methods of investigation. The microscopic method of discriminating between the elongated deformed grains of the familiar lenticular shape and the equiaxed recrys-tallized grains is not successful with fully refined lead: lenticular grains cannot be seen under the microscope with any amount of deformation whatever. The probable cause of this phenomenon was recognized long ago. With large deformation the time of recrystallization at room temperaturc is shorter than the period of preparation of the specimen for microscopic examination; with small deformation the deformed grains arc not noticeably elongated in the first place. However, with small cleformation the migration of the grain boundaries can be observed. The only systematic investigations yet published on the recrystallization of lead, those by R. S. Russell, 1-3 make use of the conditions just described. Russell considered recrystallization completed when no further migration of boundaries could be observed. A serious objection to this method arises from the fact that there is no evidence to support the view that grain-boundary migration results only from recrystallization. If such migration does result from other sources after recrystallization is completed—after the deformed grains have been replaced by new unde-formed grains—the "time of recrystallization" as determined by Russell may be too long. Furthermore, the method is subject to the above.-mentioned limitation because of the unavoidable period of preparation of the sample. According to Russell, even with the fastest technique 5 min. clapsed before the deformed samples were ready for examination, and consequently no reearystallization time shorter than 5 min. could be measured. The lack of success in detecting the recrystallization of lead by the usual hardness or tensile tests is again caused by the time necessary for

Jan 1, 1940

By E. R. Parker, R. H. Thielemann

The conventional short-time tensile test provides a reliable means of predicting the sustained load-carrying capacity of steels only when the temperature is such that continuous plastic flow does not occur. At elevated temperatures, stresses considerably less than the short-time rupture value may produce continuous flow, with fracture occurring only after long periods of time. The amount of flow or creep accompanying failure varies for different steels, and depends, to a large extent, on the temperature and the duration of the test. Service records of cracking stills, steam superheaters and high-temperature boilers have shown that brittle fractures sometimes occur with little or no warning. In installations of this type, localized stresses are often encountered, and steels must be capable of withstanding a certain amount of deformation without fracture. The sustained-load rupture test determines the expected life and corresponding ductility of steels at various stresses and temperatures. It also yields additional information regarding the effect of micro-structure and metallurgical stability on the high-temperature properties. Present Status of Sustained-load Rupture Test The sustained-load rupture test conducted at elevated temperatures is not entirely new. Dr. Zay Jeffriesl presented in 1919 the results of a large number of rupture tests made on copper, iron and tungsten. These tests were made with various strain rates and at temperatures ranging from the boiling point of liquid air to 1000° C. At that time he observed the interesting phenomenon characteristic of metals, that a temperature exists below which a metal is ductile and breaks with a transcrystalline fracture (through the grain) and above which the metal is brittle and breaks with an intergranular fracture (through the grain boundaries). He observed that this temperature varied with the testing speed, and that a minimum temperature, slightly above the recrystallization temperature, existed at which the cohesion of the grains themselves was

Jan 1, 1939