ABSTRACT A STUDY OF THE RELATIONSHIP BETNEEN THE YIELD POINT, DELAY TIME, AND THE ERITTLE TO DUCTILE TRANSITION IN A LON CARBON ALLOY by John Hrinevich, Jr. In this investigation ingot iron samples were thermally treated in three different ways to vary the precipitate ‘morphology. Samples were annealed by heating to 13500 F for one and one half hours followed by furnace cooling. Samples were solution treated by heating to 13500 F for one and one half hours followed by quenching in iced brine. Samples were solution treated-aged by heating to 13500 F for one and one half hours, quenched in iced brine, and then heated for six hours at 158° F. Mbdified Izod impact samples, 0.505 inch diameter cylindrical compression samples, and delay time specimens were made from the ingot iron in the three conditions and tested at temperatures between 14 and 305° F to determine the relationship between the behavior of the yield point, delay time, and the transition temperature. The annealed samples showed a brittle to ductile transition at approximately 1600 F. All annealed samples tested in compression below 1600 F showed a yield point. Samples tested above 1600 F showed no yield point. The annealed samples also showed a delay time when tested below 1600 F. Above 1600 F, these samples showed no delay time. The yield point and delay time disappeared above the brittle to ductile transition. The solution treated and solution treated aged samples showed no brittle to ductile transition. The solution treated and solution treated aged samples showed no brittle to ductile transition, yield point, or delay time in these tests. Careful examination of all three types of samples under the Optical and electron microscopes revealed that there‘were definite differences between the samples. The annealed samples showed wide grain boundaries with definite precipitates present. The as-quenched samples showed very little grain boundary precipitates and narrow well define grain boundaries. The solution treated aged samples like the as-quenched samples did not show much grain boundary precipitate. However, the low aging temperature of 1580 F did cause a fine general precipitate which was reflected in the higher yield and tensile strengths of the solution treated-aged samples. Carbides, nitrides, and oxides made up the precipitates observed. The yield point arises when dislocations from the interior of the grain move outward until they meet the grain boundary. At the grain boundary, the dislocations pile-up causing a stress concentration until dislocations in the neighboring grain break away, fresh dislocations are generated in the neighboring grain, the dislocations penetrate the grain boundary and move through the neighboring grain, or fresh dislocations are generated from grain boundary sources. The important point is that the grain boundary can be a major obstacle. The delay time likewise arises because the dislocations are not able to move as soon as the stress is applied. The annealed samples were the only samples to show a brittle to ductile transition, yield point, delay time, and grain boundary precipitates. Below the brittle to ductile transition, the dislocations can move relatively easily in the interior of the grains. When the dislocations reach the grain boundary, a pile-up occurs at the precipitates which causes a stress concentration. At low temperatures with oxides, nitrides, and carbides in the grain boundaries, the dislocations cannot propagate the slip into the next grain before the grain boundary actually fractures. Above the transition temperature, the dislocations have enough mobility to move around the precipitates and thus reduce the stress concentration. In ingot iron the brittle to ductile behavior is related to the precipitates in the grain boundary and thus to the yield point and delay time. A STUDY OF THE RELATIONSHIP BETWEEN THE YIELD POINT, DELAY TIME AND THE BRITTLE TO DUCTILE TRANSITION IN A LOW CARBON ALLOY by John Hrinevich, Jr. A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY in Metallurgy Department of Metallurgy, Mechanics, and Materials Science 1966 ACKNOWLEDGMENTS I wish to express my sincere appreciation to Professors Austen J. Smith and Lawrence E. Malvern. Their guidance and counsel throughout this research were invaluable. Thanks are also given to Dr. William E. Taylor who suggested the problem and to Dr. Howard Womachel who gave many helpful suggestions during the project. Appreciation is also expressed to Dr. Robert Engle who helped with the instrumentation, to Dr. C. T. Wei for many helpful discussions, to Mr. Donald Childs and his staff for their excellent job of preparing samples, and to my fellow graduate students. This project was supported in part by the National Science Foundation under Grant No. 6-24898. To my dear wife Mary Alice, who spent many hours in typing the rough drafts and the final draft, I will always be thankful for her understanding, encouragement, and confidence. I also wish to thank my parents, for their encouragement all through college and this project. ii TABLE OF CONTENTS LIST OF FIGUIIESO o o o o o o o o O o O O O O O O 0 LIST OF TABLES. . . . . . . . . . . . . . . . . . CHAPTER I INTRODUCTIOI‘I O O O O O O O O O O O O 0 CHAPTER II HISTORY AND BACKGROUND. . . . . . . .. 2.1 2.2 2.3 2.4 CrystaIIOgraphic and Metallographic Features of Brittle and Ductile Fractures Brittle to Ductile Transition Theory. . . The Sharp Yield POinto o o o o o o o o o Delay-Time Phenomenia. o o o o o o o o 0 CHAPTER III 'EJFETIJIKTRL KETHCDS- . . . . . . . . 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 Material . . . . . . . . . . . . . . . Atmosphere for Heat-Treating. . . . . . Heat-Trcating Equioncnt. . . . . . . . Heat-Treating Procedure. . . . . . . . '. Aging Experiments. . . . . . . . . . . . Impact Tests. . . . . . . . . . . . . . Static Compression Tests . . . . . . . . Pressure Bar System for Delay—Time PleasurementSo o o o o o o o o o o o o 0 Loading Device for Delay—Time PleasurelnentSo o o o o o o o o o o o o o Instrumentation for Delay-Time NeaSllremerltSoooooooooooo o 0 Samples and Procedure for Delay—- Time Teasurements. . . . . . . . . . . . Optical and Electron Microsc0pe Investigation. . . . . . . . . . . . . . iii 17 22 28 28 28 34 36 39 42 46 49 57 61 65 69 CHAPTER IV EXPERIMENTAL RESULTS. . . . . . . . . 4.1 4.2 4.3 4.4 CHAPTER V 5.1 5.2 5.3 5.4 5.5 5.6 Aging Experiments. . . . . . . . . . . Impact Tests. . . . . . . . . . . . . Static Compression Tests. . . . . . . Delay—Time Measurements. . . . . . . DISCUSSION. . . . . . . . . . . . . . Observed Differences in the Brittle to Ductile Transition, Yield Point, and Delay Time in the Annealed, Solution Treated, and Solution Treated-Aged SampIeS................ Observed Chemical and Structural Differences Between the Different TypeSOfSamples........... Origin of the Precipitates Observed in the Annealed Samples. . . . . . . . . Relationship Between the Brittle to Ductile Transition, Yield Point, and Delay-Time.............. A Proposed Mechanism for the Brittle to Ductile Transition in Ingot Iron. . . Additional Observations. . . . . . . . CHAPTER VI CONCLUSIOP‘IS O O O O O O O O O O O O O BIBLIOGRAPIW O O O O O O O O O O O C O O O O O 0 iv Page 70 7O 73 8O 97 109 109 109 120 123 125 127 130 132 LIST OF TABLES Table Page 1 Average Chemical Composition of the Ingot Iron . . . . . . . . . . . . . . . 29 2 Hardness (Rk) Data For A Typical SpRCimen [1ng At 1580 F o a 0-0-0 0 o o 71 3 Impact Data For As-Received Samples . . 75 4 Impact Data For Annealed Samples . . . . 77 5 Impact Data For Solution Treated samples a o o oo o o o o o o o o o o o o 79 6 Impact Data For Solution Treated-Aged samples 0 O O O O O O O. O O C O O O O O 82 7 Proportional Limit And Yield Drop For Annealed samples 0 o o o o o o o o o 96 8 Proportional Limit And Yield Drop For Solution Treated Samples . . . . . . 97 9 PrOportional Limit And Yield Drop For Solution Treated-Aged Samples . . . 98 10 History And Data For Delay-Time Specj-mens O O O C O O O O O O O O O O O 100 11 Chemical Composition of the Ingot Iron Before Heat.Treamento o o o o o o o e o 110 12 Chemical Composition of the Ingot Iron After Heating for Twenty-four HOurs at 14800 F. O O O O O O C O O O I I O O O O 112 FIGURE U! \DCDVO 10 11 12 13 14 15 16 17 18 19 20 LIST OF FIGURES Stress Versus Temperature For Brittle Strength, Yield Strength and Three Times the Yield Strength . . . . . . . . . . . Brittle Strength, Fracture Stress, and Yield Stress Versus Temperature . . . . Stress Versus Temperature For Brittle FraCtllreoooooooooooooooo Cross-Section of As-received Ingot Iron. Longitudinal Section of As-received IHgOtIronooooooooooooooo Cross-Section of Drying Train . . . . . Assembled Drying Set-Up . . . . . . . . Cross-Section of Furnace. . . . . . . . Complete Furnace . . . . . . . . . . . Overall Heat-Treating Facility . . . . . Impact Sample . . . . . . . . . . . . . Completed Impact Samples . . . . . . . . Impact Testing Facility . . . . . . . . Close-Up Of Sample Ready For Testing . . Overall View Of Static Stress Strain TeStingoooooooooooooooo Close-Up of Sample and Heat Sinks . . . Striker And Pressure Bar . . . . . . . . Hyge Bed . . . . . . . . . . . . . . . . Pillow Blocks . . . . . . . . . . . . . Aluminum Funnel Used To Keep Striker And Spreader Bars Aligned During Impact . . vi PAGE 31 33 35 37 38 4O 43 44 45 47 50 51 53 55 56 58 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 LIST OF FIGURES (continued) Side View Of Mounted Hyge Tester . . . . Schematic Sketch Of Hyge Tester . . . . Strain Gage Bridge . . . . . . . . . . . Overall View of Instrumentation . . . . Cross-Section of Sample Holder . . . . . Sample and Sample Holder In Position . . Typical Hardness Versus Time For A 0 Solution Treated Sample Aged at 158 F . Impact Strength Versus Temperature For [ls-Received Samples 0 o o o o o o o o 0 Impact Strength Versus Temperature For AnnealedsampleSoooooooooooo Impact Strength Versus Temperature For 8°1Ution Treated Samples 0 o o o o o o 0 Impact Strength Versus Temperature For Solution Treated-Aged Samples . . . . . Stress Strain Curve For Annealed Samples TestedAtl4°F...-........ Stress Straig Curve For Annealed Samples TEStedAt7OFooeooooooooe Stress Strain Curve For Annealed Samples TeStedAt158oFoooocoo-coco Stress Strain Curve For Annealed Samples TeStedAt3OOOFooooooooooso Stress Strain Curve gar Solution Treated Samples Tested At 14 F . . . . . . . . Stress Strain Curve or Solution Treated Samples TeSted At 70 F o o o o o o o 0 Stress Strain Curve For Solution Treated Samples TeSted At 1580 F o o o o o o o 0 vii Page 59 6O 63 66 67 68 72 74 76 78 81 83 84 85 86 87 88 89 39 4O 41 42 43 44 45 46 47 48 49 50 51 52 53 54 LIST OF FIGURES (continued) Stress Strain Curve For Solution Treated Samples TeStEd At 3000 F o o o o o o o o 0 Stress Strain Curve For Solution Treated- Aged Samples Tested At 14° F . . . . . . . Stress Strain Curve For Solution Treated- Aged Samples Tested At 700 F 3 . . . . . . Stress Strain Curve For Solution Treated- Aged Samples Tested At 158° F . . . . . . Stress Strain Curve For Solution Treated- AgEd Samples TeStEd At 3000 F o o o o o o PrOportional Limit Versus Test Temperature For Annealed, Solution Treated, And Solution Treated-Aged Samples 0 o o o o o o o o o o o OscilloscOpe Trace For Annealed Samples TeStedAtl4oFoooo000000000o OscilloscOpe Trace For Annealed Samples TestedAt70°F.............. Oscilloscope Trace For Annealed Samples TeStedAt158°F............. Oscilloscope Trace For Annealed Samples TestedAt300°F............. Oscilloscope Trace For Solution Treated Samples Tested At 14° F . . . . . . . . . Oscilloscope Trace Far Solution Treated Samples 1.83th At 70 F o o o o o o o o o OscilloscOpe Trace For Solution Treated Samples Tested At 158° F . . . . . . . . Oscilloscope Trace For Solution Treated- Aged Samples Tested At 14° F . . . . . . OscilloscOpe Trace For Solution Treated- Aged Samples Tested At 70° F . . . . . . Oscilloscope Trace For Solution Treated- Aged Samples Tested At 158° F . . . . . . viii Page 90 91 92 93 94 99 103 103 104 104 105 105 106 107 107 108 55 57 58 59 6O 61 62 63 64 LIST OF FIGURES (continued) Fracture Surface of a Sample That Failed In A Brittle FaShiOn O O O O O O O O O O O Microstructure of the Annealed Ingot Iron 0 O C O O O O O O O O O O O O O O O O Enlargement of a Typical Precipitate Found in the Annealed Ingot Iron . . . . . . . . Enlargement of a Typical Precipitate Found in the Annealed Ingot Iron . . . . . . . . Typical Microstructure of the As-Quenched Samples . . . . . . . . . . . . . . . . . Typical Microstructure of the Solution Treated-Aged Samples. . . . . . . . . . . Phase Diagram For Iron-Oxygen o o o o o 0 Side View Of Brittle Fracture . . . . . . Cross Section Of A Typical Tough Fracture. Side View Of A Typical Ductile Sample . . ix Page 113 115 116 117 118 119 121 126 128 129 CHAPTER I INTRODUCTION Low—carbon steels at room temperature are normally considered impact resistant and ductile. Impact strengths as high as forty-five foot-pounds in a standard Izod test with considerable plastic deformation are commonly recorded. At very low temperatures, however, low—carbon steels often behave quite differently. Their impact strength in the standard Izod test may drop as low as two foot—pounds and the fracture shows practically no plastic deformation. This change from ductile behavior to brittle behavior occurs abruptly over a small temperature region. The average temperature of the transition zone is referred to as the Transition Temperature. Some structures made of low carbon steel are used at temperatures below the transition temperature of the material and may fail in a brittle manner. The failures, while not an everyday event, have happened frequently enough to be both a very costly and aggravating problem for nearly the last eighty years. One of the most famous and catastrophic brittle failures of a low carbon structure was the Boston molasses tank.1 One January day in 1919, when the tank contained 2,300,000 gallons of molasses, it burst open. Twelve persons were drowned in molasses or died of injuries, forty others were injured, and horses were drowned. Houses were damaged and a portion of the Boston Elevated Railway structure was knocked over. Bridges also have suffered from brittle failure.2 Just prior to World War II, about 50 bridges of a type known as a Vierendeel truss were built across the Albert Canal in Belgium. Some of these bridges were built of welded or rolled I-beam and plate, others entirely of plate. In March, 1938, when the weather was quite cold, the bridge at Hasselt, with a span of 245 feet, collapsed into the canal. Eyewitnesses heard a sound like a shot and saw a crack open in the lower chord. Six minutes later the bridge broke into three pieces, and fell into the canal. All the fractures were brittle, some through welds, others in solid plate away from the welds. The bridge was lightly loaded at the time. Within two years, two similar bridges failed in the same way. Brittle failure has claimed a number of ships.3 Between 1942 and 1952 about 250 welded ships suffered one or more brittle fractures of such severity that the vessels were lost or were in a dangerous condition. Nineteen of these 250 ships broke completely in two or were abandoned after their backs were broken. In the same ten-year period, 1200 welded ships suffered brittle cracks, generally less than 10 feet in length, which did not disable the ships but were potentially dangerous. For brittle fracture to occur, three conditions seem to be necessary. 1) Low temperature, such as exists in the winter months. 2) The presence of a notch, introducing triaxial stress. 3) High strain rate, or impact loading. This factor is not wholly necessary for the initiation of brittle failure. The temperature at which the fracture changes from brittle to ductile, the transition temperature, is affected by various metallurgical factors.5 It has been found that a fully-killed steel will have a lower transition temperature than a semikilled or rimmed steel. Small ferritic grain size, the use of a lower finishing temperature in hot rolling, increasing manganese content, nickel in amounts up to 1.80% and silicon in amounts up to 0.25% lower the transition temperature. Cold working and increasing contents of carbon, phOSphorous, molybdenum, and boron increase the transition temperature. The first small amounts of aluminum lowers the transition temperature, but increasing amounts cause no change. Chromium appears to have little effect. In this work, the beginnings of a more fundamental understanding of the brittle to ductile transition in low carbon alloys will be sought. CHAPTER II HISTORY AND BACKGROUND 2.1 Crystallographic and Metallographic Features pf Brittle and Ductile Fractures Brittle fracture may occur in ferritic steel and some other body-centered cubic materials and certain hexagonal close-packed metals. Face centered cubic metals like pure aluminum, copper, silver, and gold,even at temperatures near absolute zero and with sharp notches, do not fail in a brittle fashion or cleave. During ductile fracture, slip in iron occurs in a 6 on planes containing the (111) family of directions. The {110}, {112}, {123} , or combinations of pencil fashion these families are the active slip planes. The resulting fracture surface is wavy and irregular. A brittle or cleavage failure occurs by separation without macro- plastic deformation on the {100} or cube faces.7 The resulting fracture surface is relatively smooth and well defined. The highpspeed cracks occurring in a cleavage failure are not smooth but run in a discontinuous fashion.8 Since each crystal cleaves on a (100) face, the crack must change direction as it goes from one crystal to another. Cleavage probably starts independently in neighboring grains. The resulting crack segments are then connected by plastic deformation at the grain boundaries.9 One should be able to observe the crack segments adjacent to or ahead of the main fracture in steel that has failed in a brittle manner. Jaffe10 has observed such unconnected crack segments. The surface of a brittle fracture shows little plastic deformation.6 X-ray diffraction studies indicate that some plastic deformation does occur. The depth of the cold worked layer as revealed by etching the surface away and re-examining with x-ray techniques was approximately 0.05 mm. 2.2 WEWWW One of the earliest attempts to explain the transition from brittle to ductile behavior was carried out by Mesnager,12 who noted that to produce a brittle fracture, high speed loading was not necessary. Slow bending or slow tension can introduce brittle fracture if the specimen contains a sharp deep notch. His theory of triaxial tension in notch brittleness described.how brittle fracture could occur under very low speed loading when a notch or crack was present. When a uniform stress below the elastic limit is imposed on a plate in a direction perpendicular to a crack, a comparatively high stress in the same direction will exist just behind the root of the crack. The stress at the crack root is biaxial. If plastic flow is to occur at the crack root, there must be lateral contraction of the material at that place in order to preserve constant volume of material. This lateral contraction is Opposed by the large amount of material, stressed to a lower value behind the root of the crack. This will induce a state of triaxial stress, the third stress being perpendicular to the plane of the plate and tending to contract the material laterally at the root of the notch. As a result the axial stress at the root of the crack will build up beyond the uniaxial yield stress, Y, of the material to some value, Yh, before flow can occur. The ratio of Yn 1:O Y is known as the ”plastic constraint factor." Ludwikl3 in 1923 developed a similar theory. Until 1945, it had been believed that if the notch was of correct depth and sharp-ended, the stress at the root and the plastic constraint factor would rise to infinity. Orowan, Nye, and Cairns14 in 1945 showed that the maximum constraint factor is appr0ximately three. As a result if a sharp crack is present in a stressed plate, the stress at the crack root must rise to about three times the normal uniaxial yield stress of the material before plastic flow will occur. Davidinkov and Wittman further explained the phenomenon of transition in 1937. They drew a logical qualitative picture using the concept of ”brittle strength." The resulting picture is shown in Figure 1.15 The brittle strength B, yield strength, Y, and three times the yield sesameem mama» may means mummy oz< memameem maMHw .meozemem maeeaam mom meme mmmmem a ”$5on a... Nazism“. 2 up w a _ _ _ . _ _ 44; _ " Jib _ \NKE _ a. _ a0 _ u. 6. [Y O _ 0v _ so“ . é " e. . 6am. . 9‘ m k 4 k . w W masks: a“ .u , .9 fl 2:95qu was >oxcop4>ua nouns 0» If. -D $538.15 strength, 3Y, are sketched as a function of temperature. Experimentally, it is known that the yield strength rises with decreasing temperature. It was assumed that the brittle strength rose with decreasing temperature but not at so steep a gradient. Above temperature T2 where the curves of brittle strength and of 3Y intersect, the material will be fully ductile, with.or without the presence of a notch. Below temperature T1, where the brittle strength intersects the yield strength, the material will always fracture in a brittle manner. The region between these two temperatures is the ”notch brittle zone,“ wherein the material will behave in a ductile fashion in the presence of uni-axial stress, but if a notch is present, it will fail in a brittle fashion before the stress of 3Y with attendent plastic flow can be reached. In practice the transition from full ductile to fhll.brittleness occurs in a narrow temperature zone rather than at a particular temperature. 16 determined the fracture In 1951, Eldin and Collins and yield stress of A.I.S.l. 1020 steel at temperatures from 12°K.to 1850K. From 12 to 61.50K, all specimens exhibited typically brittle fracture, with no reduction in area, and the strength drapped steadily and linearly as the temperature was increased. At 61.5°K.a sharp transition occurred. Above 61.5°K the reduction of area increased rapidly as the testing temperature was increased, and a yield stress, as well as a fracture stress could be measured. A reproduction of some of their data is shown in Figure 2. High strain rates favor brittle failure. This effect may very well be related to the strain-rate dependence of the yield strength. The yield strength of most metals is not particularly sensitive to strain rate.17 In typical nonferrous metals, for instance, increasing the strain rate by several orders of magnitude will increase the yield strength by less than 20 percent. In low-carbon steels, however, a corresponding increase in strain rate will hincrease the yield stress by one hundred to two hundred percent. The brittle strength, on the other hand, shows a relatively small dependence on strain rate. The net result at higher strain rates is to displace the yield stress curve upward while the brittle strength curve remains nearly fixed. Brittle failure is thus more likely to occur. 18’19 furthered our understanding of the Robertson transition temperature with studies carried out on flat plates. In effect, a saw cut was placed at the edge of the plate and a brittle crack was initiated by a wedging impact force. The specimen had a temperature gradient across it, the notch end being cold, and the apposite end being warm. All the while the specimen was in tension in a direction perpendicular to the notch. The crack traveled across the plate, and was arrested at some particular point. Beyond this point and above the 10 A: meeeaammzme memam> mmmxem neeaw oz< .mmmmem maseoaae .meuemmem maeeumm N MMDU Hm no meahammezme com em. 02 on o _ _ . _ _ '1 Ofix I; 0.2 / .W mmhm Ntah0‘c . . ConSLdering that C; is not too large, that L0 ooeys the relationship in O; = ln B-fiT where B and ,3 are constants, , + K is a constant, and that 03 varies linearly with temperature over small temperature ranges, the transition temperature, Tc’ may be expressed as I L * -, erc=og*+C-(39-£E§-k)1.2 (5) where 6, C, k constants the triaxiality; 1/ 3 .0 II Thus the transition temperature is dependent on the grain size, friction of unlocked dislocations, the strength of dislocation locking, and the degree of triaxiality of the applied stress. 17 .Z-E’: mmxgmm Cottrell's32 theory for the sharp yield point in body-centered cubic materials is based on the strong affinity of dislocations for carbon, nitrogen, and other atoms. In the unyielded or strain-aged condition, the dislocations are anchored by their atmospheres. In the yielded state, the dislocations are free of any atmosphere. Considerably higher stress is required to move a dislocation pinned by an atmosphere than to move one that is free of any atmosphere. To cause yielding, a stress high enough to move the dislocations with their atmospheres is required. However, as soon as the movement has begun, the dislocations are freed from their atmospheres and can move with a reduced stress. Since the change from the pinned to the unpinned state happens relatively fast at the beginning of plastic deformation, the material suddenly becomes softer and yields markedly. One of the best examples of sharp yield behavior is soft steel with ‘10-3 to 10'1 weight percent of Carbon. Sharp yield points have been observed in a number of other materials.33 Cadmium, zinc, brass, and molybdenum are a few examples. Cottrell and Bilby34 have described the atmosphere in greater detail. An important feature of the theory is that the atmosphere condenses into a line of solute atoms lying parallel and close to the dislocation at the position of strongest bonding. Under stress small segments of the dislocation are bowed out. Thermal fluctuations help the 18 loOps grow to a stable size and then pull the rest of the dislocation away from its atmosphere. The authors calculated the ratio of the theoretical break-away stress at temperature T to the break-away stress at zero degrees Kelvin. The variation of the ratio with temperature agreed well with the variation of the yield strength with temperature. Face centered cubic materials do not in general show a strong yield point.35 Small atoms like carbon and nitrogen dissolved in a body-centered cubic metal distorts the lattice locally into a body-centered tetragonal structure. Because this distortion is not spherically symmetrical,these atoms interact with shear stresses, as well as with hydrostatic stresses, and so interact with. both the screw and edge components of dislocations. A solute atom with a spherically-symmetrical elastic field, on the other hand, can interact only with an edge dislocation, since to a first approximation, a screw has no hydrostatic component in its field. In the face-centered cubic lattice the arrangements of lattice positions around a substitutional site or the mainly used interstitial site at the center of the cube is too symmetrical to allow a solute atom in either site to produce a non-spherical distortion. Although there should be no direct elastic interaction between a pure screw dislocation and such soluteatoms, it may still be possible to anchor the screws indirectly. In the face- centered-cubic lattice ordinary unit dislocations in closepacked planes dissociate into pairs of half- dislocations, the lines of which run parallel to each other and are spaced a few atoms apart. The two dislocations in such a pair are joined to each other by a stacking fault and must glide together as a unit in the slip plane. Their Burgers vectors intersect at 600 so that there is no orientation of the line of the pair for which its members can both be pure screws; always, at least one of them must have a substantial edge component and therefore be able to be locked by solute atoms. If just one dislocation is locked, the other one cannot move either since they are coupled together by the stacking fault. While this locking is no doubt weaker, in general, than that in the body centered cubic lattice, nevertheless the possibility of producing yield points in face centered cubic crystals cannot be ruled out. Adair, HOOk, and McGaughey36 have shown that the initial yield characteristics of iron with 20 parts per million carbon, 50 parts per million nitrogen, and 80 parts per million oxygen are strongly influenced by the extent of the segregation of interstitials to grain boundaries. Specimens cooled from 12920 F showed grain boundary precipitates in thin film specimens at 12,000 X. Samples quenched from 12920 F showed no precipitates in the grain boundaries. Slowly cooled samples showed a marked yield point and the quenched samples showed no yield point. It is inferred that strong locking occurs when the grain boundaries have considerable segregation. This causes the marked yield drop. Stein, Low, and Seybolt37 develOped a technique to lower the carbon content to as low as 5 x 10"3 parts per million in single crystals of iron using a hydrogen atmosphere continuously purified with ziroconium hydride. Samples with carbon contents from 5 x 10"3 parts per million to 0.02 percent were pulled in tension. The yield strength of samples with more than one part per million of carbon showed a strong dependence on temperature. The yield strength of samples with less than one part per million of carbon showed a marked reduction in temperature dependence. A reasonable explanation for these observations is that the temperature dependence of the yield strength in alpha iron of usual purity is due primarily to the interaction of mobile dislocations with interstitial impurities in solid solution. Cottrell's theory of dislocation locking in the case of iron and related body-centered-cubic metals is well excepted. The dependence of the yield drop on interstitial impurities, the kinetics of strain aging, and direct observation by transmission electron microscopy of particles formed on dislocations provide a body of strong supporting evidence. However, the theory that unlocking dominates the yield drop is not entirely 21 satisfactory because the temperature, strain-rate, and grain-size dependence of the upper yield stress, lower yield stress, and of subsequent flow stress values are very similar. Yet the flow stress cannot be governed by an unlocking mechanism at low temperatures where the pinning points are essentially stationary. Gilman and Johnston38 accounted for the yield drop in lithium fluoride crystals quantitatively in terms of the rapid multiplication of dislocations and the stress- dependence of dislocation velocity. In lithium fluoride no evidence of unlocking has been found and grown-in dislocations remain locked. Dislocations responsible for the slip are heterogeneously nucleated and multiply rapidly. Hahn39 carried out the same type of analysis for alpha iron. According to him the abrupt yield is due to the presence of a small number of mobile dislocations initially present, a rapid dislocation multiplication and the stress dependence of the dislocation velocity. Considering the plastic strain rate, multiplication rate, and the dislocation velocity, the plastic strain is expressed as follows: 0.5 bf (,0. + Ca; ) (Z’Q)'n(6’-q€p)n (6) m '0 II where 6p plastic strain rate b = Burgers vector f 2410-1 C = experimental parameter related to the multiplication of dislocations. a = experimental parameter related to the multiplication of dislocations. ’l = the resolved shear stress corresponding to the unit velocity. :3 II experimental parameter related to the dislocation velocity. proportionality factor relating the change in stress to the plastic strain. The stress-strain curve calculated from this equation at constant strain-rate shows a pronounced yield drOp as experienced with real metals. This suggests that unpinning may not be the only cause of the sharp yield point. 2.3 legyegime Phenomenia Liu, Kramer, and Steinberg4O observed that close- packed hexagonal zinc and body-centered cubic beta-brass showed a definite increase in critical resolved shear stress at temperatures below room temperature when the strain durationwausdecreased to about 10"3 seconds. However, in face-centered cubic copper crystals no delay time was apparent. Clark and Wood41 found that rapid-load tests on type 302 stainless steel, SAE 4130 normalized steel, SAE 4130 quenched and tempered steel, 24S-T 23 aluminum alloy, and 753-T aluminum alloy showed that no delay occurs for the initiation of plastic deformation and 31x.of these materials exhibit a definite yield point in their static stress strain curve. Annealed low carbon steel showed a delay time and its static stress strain curve does have a definite yield point. '7 I Kramer and Maddin;4' and Lui, Kramer and Steinberg4O imply that the type of crystal structure determines whether the delay time phenomenon will occur whereas Clark and Wood41 feel that the appearance of a yield point on the static stress strain curve is the important criterion. This is al*ost the same criterion because usually body-centered cubic materials show a strong yield point. Clark and Wood43 compared the delay times for the initiation of yielding in a steel treated in four different ways: a) annealed; b) annealed and wet-hydrogen-treated to eliminate the static upper yield point; c) annealed, wet-hydrogen-treated and recarburized; d) annealed, wet-hydrogen-treated and renitrided. All four of the samples behaved in a qualitamnmly similar manner when subjected to rapidly applied constant stress, that is, a delay time for the initiation of plastic deformation exists for all four materials; the relations between this delay time and the stress are all quite similar in form; and the effect of temperature upon these relations is essentially the same in all cases. Furthermore, the Dehavior of these materials is similar to the behavior 24 of annealed low-carbon steels. However, definite quantitative differences in behavior are exhibited between the four materials. The test results showed that the principal effect of changing the carbon and nitrogen content of an annealed low carbon steel subjected to rapidly applied constant stress was to change the stress corresponding to a given value of the delay time. The magnitude of the stress for a given delay time was decreased by a reduction of the carbon and/or nitrogen concentration. This supports the theory that the delay time phenomenon is a result of the dislocations being unable to move as soon as the stress is applied. As the carbon and nitrogen content is decreased, there are interstitial atoms available for pinning, the g15L;;;:;_43 can move more easilfi and the delay time increases as a :rsult. 'Vreeland, Wood, and Clark44 carried out some interest- ing repeated stress-pulse experiments which also supported the idea that the delay time is a result of the inability of the dislocations present to move as soon as an external stress is applied. The tests employed showed the effect of stress-pulses and aging on the delay time. Stress- pulses of essentially constant magnitude and duration on the specimen were used. The pulse was as follows: a) Stress of 45,000 r 800 psi applied in approximately 0.007 seconds. b) Stress held essentially constant for 0.029 3 0.001 seconds. m m c) Stress removed in approximately 0.003 seconds. The delay time for the material when subjected to a stress of 45,000 pounds per square inch is approximately 0.050 seconds, which is greater than the duration of stress-pulse and less than the duration of two stress pulses. Thus, the material might be expected to yield during the second stress-pulse if there were no recovery between pulses. The specimens were aged for various intervals of ‘time between stress-pulses at temperatures of 70, 150, and 2000 F. The procedure for aging the specimens after a stress-pulse was as follows: 1) Specimen removed from rapid-load machine and placed in a bath at the desired temperature within five minutes after the stress pulse. 2) Specimen removed from the bath after the desired aging period and immediately cooled in powdered dry ice. 3) Specimen brought to approximately 700 F in an alcohol bath five minutes prior to the next stress-pulse. 4) Stress-pulse applied when the specimen reached 70° F. The results of the repeated stress-pulse tests showed that a definite recovery from the effects of the previous stress-pulse takes place when the time and temperature between stress-pulses is sufficient. It was found that approximately fifty percent of the microstrain is recovered after'removing the load- This indicates that some of the dislocations which were displaced by the applied stress return toward their original positions when the stress is removed, as might be expected. The microstrain rate, prior to the initiation of yielding, decreases with time. This might be accounted for by a depletion of the reservoir of dislocations which may be moved by the applied stress. This may also account for the decreasing amount of microstrain induced by successive stress-pulses when recovery does not take place. When the aging treatment between stress-pulses induces recovery, a greater decrease in the microstrain in successive stress-pulses if found. The recovery mechanism may be explained by the diffusion of carbon and nitrogen to the dislocations which have been displaced. The resulting array of dislocations, anchored by atmospheres of carbon and nitrogen, may be expected to be more stable than the original array under the particular stress condition. Successive stress-pulses and aging cycles then produce less microstrain, and yielding does not take place. The more stable configuration of dislocations for the particular stress condition is also indicated by the results of the rapid load tensile tests on the 27 material which had previously been subjected to stress- pulses and aging cycles in which recovery took place. The delay times for the initiation of yielding are increased by a factor of approximately four over the delay times for the original material for corresponding stresses. Fisher45 derived an expression for the delay time. Taking into account the energy necessary to unpin the dislocation, the energy necessary to extend the dislocation line, the rate of thermal nucleation of 100ps, and the idea that the delay time is associated with the generation of a large number of dislocation loops; the expression for the delay time is as follows: Log (D.T.) = A + BGZ/IT (7) where (D.T.) delay time G - the rigidity modulus ”E0 = resolved shear stress on the slip plane. T = absolute temperature The agreement with experimental data was good except at low temperatures. Thus the delay time for yielding under constant stress depends upon the single parameter Gz/TLT and not upon stress or temperature independently. CHAPTER III EXPERIMENTAL METHODS §._l_ Material Armco magnetic ingot iron was used for all specimens. The ingot iron was purchased in cold-drawn three quarter inch diameter bars twelve feet long. The chemical composi- tion was determined from a 0.400 inch by 0.400 inch by 2.0 inch sample machined from the middle of a twelve foot bar. Care was taken to remove equal amounts from all four sides to eliminate surface effects. The analysis was carried out by the United States Steel Research Laboratories. Table 1 shows the average chemical composition. The iron in the as-received condition was essentially all ferrite with some oxides and manganese sulfide present. Figure 4 shows a typical cross-section microstructure of the as-received ingot iron. As expected, the as-received ingot iron showed considerable cold working. Figure 5, a longitudinal section, illustrates this point. _3_.3_ Atmosphere g9; M-Treatking High purity dried argon gas (99.995) was used as a protective atmosphere for all high temperature treatments. The argon was purchased from the National Cylinder Gas Company in cylinders containing 330 cubic feet of the compressed gas. Only one cylinder of gas was required for the entire test series. 28 TABLE 1 Average Chemical Composition of the Ingot Iron. Element Percent by EELEEE Carbon . . . . . . . . . . . . . . . . . 0.015 Manganese . . . . . . . . . . . . . . . 0.063 Silicon . . . . . . . . . . . . . . . . 0.004 Phosphorous . . . . . . . . . . . . . . 0.004 Sulfur . . . . . . . . . . . . . . . . . 0.021 Nitrogen . . . . . . . . . . . . . . . . 0.004 oxygen 0 O O O O O O O O O O O O O O O O 0 g 092 30 Etch: 4 percent Nitric Acid in Amyl Alcohol 500 x Mount # 65-129 FIGURE 4 Cross-section of As-received Ingot Iron. The matrix is ferrite with some manganese sulfide and oxides present. Etch: 4 percent Nitric Acid in Amyl Alcohol 100 x Mount # 65-131 FIGURE 5 Longitudinal Section of As-received Ingot Iron. The elongated grains show that the material has been cold worked. 31 In use the gas was passed through a Linde Inert-Gas Double-Stage Regulator to reduce the tank pressure to a constant value of five pounds per square inch and then through a needle valve to regulate the flow before being passed into the drying train. From the drying train, it was piped into the heat-treating furnace in a 1.25 inch soft copper tube. The drying train was assembled especially for these experiments. A one-inch-diameter 36 inchslong heavy-wall black iron steam pipe was threaded on both ends for standard pipe fittings. The tube was then pickled in an aqueous 30-percent nitric-acid solution to remove the oxides left from.manufacturing. After rinsing in water and drying, a two-inch plug of fine clean copper turnings was pushed just past the center of the pipe. Next a tight-fitting 80 mesh austenitic stainless steel screen was placed over the copper turnings, and three inches of minus ten plus thirty mesh titanium powder were added. Over the titanium powder another tightly fitting 80-mesh austenitic stainless steel screen was placed followed by another two-inch plug of fine clean cOpper turnings. Figure 6 shows a cross section of the drying train. The assembled drying train was placed in a Burrell- High Temperature Furnace, Type CTA1-9, and standard pipe- reducers were screwed onto both ends, using teflon tape as a sealant. Brass half unions were used to connect the in- coming and outgoing c0pper tubes. A platinum platinum- 33 maize? mucsou ”3.35. 0735mm mo ZOHHOmmImmOmO o MMDUHHH $2.2K3F KUEQOU 235m «338 22235. Zumxom l\ A |l\\l\.l.\ o O O 0 Gal’\ v \I\|l\..l\ O O 0 )\ol\ (It 0 O O o 0 82’s l“)\_ O O O O .(llk \( \/\_ O o o 0 0 TI\ 1 \/\/\/\/ J T J F mntmmlk mm a DKmm4m mm mnHm HN HMDUHh 6O MMHmMH mowm ho EUHmMm UHHm3w 222.234. \\ a skiEiE .. «(m «.5 sq ‘m. maaz mmmzomHm0mmlm< mom mmBaMMEQH mDmfim> HHHmfimem Eng/NEE“ mm MMDUHm Gav mmaiumazm» 8m m5 omm Mmu com m: cm. Mm. 02 he om A a _ _ A _ fl _ _ fl 2 f, '2 HLSNBHlS .lOVdN/ K} (a 37:!) 0 fl) IMPACT DATA FOR AS-RECEIVED SANPLES TABLE 3 Test Temperature Sample Impact (0F) Number Strength 14 169 2.0 14 170 1.0 14 171 1.0 Average Impact Strength 1.3 72 166 3.0 72 16 2.0 72 16‘ 3.0 Average Impact Strength 2.67 158 163 4.0 158 164 5.0 158 165 23.0 Average Impact Strength 12.3 300 160 26.0 300 161 25.0 300 16 27.0 Average Impact Strength 26.0 76 nnqmaém amazemzzaa mom mMDHamcmuxo sou ouauoz mqmz MQHm do mMDth CHAPTER VI COIICLUSIOI‘QS 1. The annealed ingot iron used for these tests showed a brittle to ductile transition at approximately 1600 F. The solution treated and the solution treated and aged samples showed no brittle to ductile transition down to 140 F. 2. The annealed samples showed a marked yield point below the brittle to ductile transition. Above the transition, no yield point existed in the annealed samples. The solution-treated and the solution- treated and aged samples showed no yield point in these tests. 3. The annealed samples showed a marked delay time below the brittle to ductile transition. Above the transition, no delay time existed for the annealed samples. The solution-treated and the solution- treated and aged samples showed no delay time. 4. Low temperature aging, 1600 F for six hours, of solution treated samples did not markedly alter the samples as did annealing. For the test conditions used in these tests, no brittle to ductile transition, yield point, or delay time was 130 detected. The yield strength was considerably higher, however. 5. Grain boundary precipitates appear to be important in determining whether a strong yield point, delay time, and thus a high brittle to ductile transition will be present. 6. The transition temperature of as-received ingot iron can be markedly lowered by quenching from 13500 F rather than slow cooling. 10. 11. 12. 13. 14. 132 BIBLIOGRAPHY 9' '7. ngg Beggtd, Vol. 37, no. 9, Sept. 1, 1921, pp. 372. Shank, M. F., "Brittle Failure of Steel Structures- a Brief History,” Metal Ptgggggfi, Vol. 66, no. 3, September, 1954, pp. 83-88. Tipper. c. F-. The mm: Erastus Sim Cambridge University Press, Cambridge, 1962, pp. 1- 20. Shank, M. "E., "Brittle Failure of Nonship Steel- Plate, 'MgsnanisalEnginggring Vol- 76 l~0-1. 1954, pp. 23-28. Lorig, C. H., “Influence of Metallurgical Factors,” Enhaxinr 2f.Mgrgia at Les Ignngratnrgg. American Society for Metals, Cleveland, Ohio, 1954, pp. 71-105. Cottrell, A. H., Dislocations gnd Plastic Flo win Ctygtals, Clarendon Press, Oxford, 1958, pp. 2- Barrett, C. S., Sttucture gt Metals, McGraw Hill Book Company, New York, 19 2, pp. 385-387. Tipper, C. F., The B thgtgge Story, University Press, Cambridge, 1 62, pp. 21-24. Tipper, C. F., The Brittle Enacturg S or , University Press, Cambridge, 1962, pp. 25-28. Jaffe, L. D., Reed, E. L., Man, H. C., "Discontinuous Crack Propagation-Further Studies,“ Transactions of the American Institute of Mining and Metallurgical Engineers, Vol. 185, 1949, pp. 683-687. Smallman, R. E., Modern Physical Metallurgy, Butterworths, Washington, 1963, pp. 72-73. Mesnager, A., "Rapport Non-officiele,“ No. Abf, International Association for Testing Materials, Brussels Congress, 1906, pp. 1-16. Ludwik, P., "Ueber Kerbwirkungen bei Flusseisen,“ Stahl and Eisen, Vol. 43, 1923, p. 999. Orowan, E., Nye, J. F., Cairns, w. J., Theoretical Research Report No. 16/45, Armament Research Dept. MOS, London, 1945. 15. 16. 17. 18. 21. 22. 25. 26. 27. 28. 133 Orowan, E., ”Classical and Dislocation Theories of Brittle Fracture," ct , John Wiley, New York, 1959, p. 148. Eldin, A. 3., Collins, S. C., "Fracture and Yield Stress of 1020 Steel at Low Temperature," Journal of Applied Physics, Vol. 22, 1951, pp. 1296-1297. Low, J. R., "The Influence of Mechanical Variables,” Behavior of Metals at Low Temperatures, American Society for Metals, Cleveland, Ohio, 1953, pp. 48-56. Robertson, T. 3., "Propagation of Brittle Fracture in Steel," Journal of the Iron and Steel Institute, V01. 175, 1953, pp. 361-3740 Robertson, T. 8., "Brittle Fracture of Mild Steel,” Engineering, 1951, pp. 445-448. Weasel, E. T., "Abrupt Yielding and the Ductile-to- Brittle Transition in Body-Centered-Cubic Metals," Journal of Metals, Vol. 9, 1957, pp. 930-935. Stroh, A. N., ”Brittle Fracture and Yielding,” Philosophical Magazine, Vol. 46, pp. 968-972. Mott, N. F., ”Fracture in Metals," Journal of the Iron and Steel Institute, Vol. 183, 1956, pp. 233- 243. Koehler, J. 3., "Kature of Work Hardening,” Physical Review, Vol. 86, 1952, pp. 52-59. Cottrell, A. H., "Theory of Brittle Fracture in Steel and Similar Metals," Transactions of the American Institute of Mining and Metallurgical Engineers, V01. 212’ 1958, pp. 192-2030 Low, J. R., Symposium on the Relation of Properties to Microstructure, A.S.M., 1954, p. 163. Owen, W. S., Averbach, B. L., and Cohen, M., “Brittle Fracture of Mild Steel at -196° C,” Transactions of the American Society for Metals, Vol. 50, 1958, Pp 0 634-655 0 Petch, N. J., ”The Ductile-Cleavage Transition in Alpha Iron," F a , John Wiley and Sons, Inc., New York, 1959’ pp. 54-640 Griffith, A. A., "The Phenomena of Rupture and Flow in Solids," Philosophical Transaction of the Royal Society, VOI. A221, 1920, pp. 163-198. 29. 31. 32. 33. 34. 35. 36. 37. 38. 39. 134 Hall, E. O., "The Deformation and Aging of Mild Steel: II Characteristics of Luders Deformation," Physical Society of London, Proceedings, Section B, Vol. 64, pp. 742-752. Cracknell, A., and Petch,/N. J., "Frictional Forces, on Dislocation Arrays at the Lower Yield Point in Iron," Acta Metallurgica, Vol. 3, 1955, pp. 186-189. Heslop, J., and Petch, N. J., The Stress to Move a Free Dislocation in Alpha Iron," Phi1050phical Magazine, Vol. 1, Ser. B, 1956, pp. 866-873. Cottrell, A. H., Dislogatigns gnd Plastic Flow in ngstals, Clarendon Press, Oxford, 1953, pp. 139-145. Edwards, C. A., Phillips, D. L., Liu, Y. H., "The Yield Point in Steel,“ Journal of the Iron and Steel Institute, Vol. 147, pp. 145-172. Cottrell, A. R., and Bilby, B. A., "Dislocation Theory of Yielding and Strain Aging in Iron," Proceedings Royal Physical Society, London, 1949, Vol. A62, PP. 49-62 o Ardley, G. W., Cottrell, A. H., "Yield Points in Brass Crystals," Proceedings of the Royal Society of London, Vol. 219, pp. 328-341. Adair, A. M., Hook, R. E., and McGaughey, R. L., "The Effect of Thermal History on the Yield Behavior of the American Institute of Mining and Metallurgical Engineers, Vol. 236, No. 2, February, 1966, pp. 174-178. Stein, D. F., Low, J. R., Seybolt, A. U., ”The Mechanical Properties of Iro Single Crystals Containing Less Than 5 x 10' ppm Carbon," Acta "Metallurgica, Vol. 11, No. 11, November, 1963, pp. 1253-1262. Johnson, w. 8., and Gilman, J. J., Dislocation Velocities, Dislocation Densities, and Plastic Flow in Lithium Fluoride Crystals," Journal of Applied Physics, Vol. 30, No. 2, 1959, pp. 129-144. Hahn, G. T., "A Model for Yielding With Special Reference to the Yield-Point Phenomena of Iron and Related BCC Metals," Acta Metallurgica, Vol. 10, No. 8, August, 1962, pp. 727-738. 40 . 41. 42. 43. 44. 45. 46. 47. 49. 50. 135 Lur, T., Kramer, I. R., and Steinberg, M. A., ”The Delay-Time Phenomenon in Metal Single Crystals,” Acta Metallurgica, Vol. 4, No. 4, July, 1956, pp . 36(1-‘370 . Clark, D. 3., Wood, D. 3., ”The Time Delay for the Initiation of Plastic Deformation at Rapidly Applied Constant Stress,” Proceedings of the American Society for Testing Materials, Vol. 49, 1949, pp. 717-721. Kramer, I. R., and Maddin, R., “Delay Time for the Initiation of Slip in Metal Single Crystals,“ 'Journal of Metals, Vol. 4, No. 4, February, 1952, Pp. 197-203. Wood, D. S., Clark, D. 3., ”Delayed Yield in Annealed Steels of Very Low Carbon and Nitrogen Content,” Transactions of the American Society for Metals, V01. 44, 1952’ pp. 726-745. Ureeland, T., Wood, D. 3., Clark, D. 3., “A Study of the Mechanism of the Delayed Yield Phenomena,” Transactions of the American Society for Metals, V01. 45’ 1953, pp. 620-631. Fisher, J. C., "Application of Cottrell's Theory of Yielding to Delayed Yield in Steel,” Transactions of the American Society for Metals, Vol. 47, 1955, pp. 451-462. Smith, M. C., Alloy Series in Physical Metallurgy, Harpers and Brothers Publishers, New York, 1956, pp. 200-208. Seybolt, A. U., "Solubility of Oxygen in Alpha Iron," American Institute of Mining and Metallurgical Engineers, Vol. 200, 1954, pp. 641-644. Seybolt, A. U., “Precipitation of Iron Oxide from Alpha Fe-O Solid Solutions," Transactions of the American Institute of Mining and Metallurgical Engineers, Vol. 200, 1954, pp. 979-982. Wert, C. A., "Solid Solubility of Cementite in Alpha Iron," Journal of Metals, Vol. 2, 1950, pp. 1242-1244. Allen, N. P., "The Mechanical Properties of the Ferrite Crystals," Journal of the.Iron and Steel Institute, Vol. 191, 1959, pp. 1-18. 51. 52. 53. 54. 55. Kamber, K., Keefer, D., and Nert, C., ”Interactions of Interstitials with Dislocations in Iron,“ Acta Metallurgica, Vol. 9, 1961, pp. 403-414. Tsow, A. L., Nutting, J., Menter, J. w., "The Quench- Aging of Iron," Journal of the Iron and Steel Institute, Vol. 172, 1952, pp. 163-171. Leslie, W. C., and Keh, A. S., "An Electron Transmission Study Of The Strain Aging Of Iron," Journal of the Iron and Steel Institute, Vol. 200, 1962, pp. 722-728. Russell, To Lo, WOOd, D. So, Clark, D. So, ”The Influence of Grain Size On The Yield Phenomenon In Steel, Acta Metallurgica, Vol. 9, 1961, pp 0 1054‘1063 0 Phillips, V. A., "New Evidence For Segregation At Grain Boundaries, Subgrain Boundaries, And Dislocations In Dilute Iron-Carbon-Nitrogen Alloys," Acta Metallurgica, Vol. 11, 1963, pp. 1139-1150.