ACQUfiTES EMESSEON AND RELAYED QESF‘LACEMEE‘V‘T'S EN LETHWM FLU-QEE§E SINGLE CRYSFALS TEEGSKS for ”we Degree cg p51. D. WCBLGAN STATE UNEYERSETY Robert B. Eng’ée 1966 SITY LIBRAR RIE IHIWUIH‘llIHHlHlllllllWllHtllHlHll Hill!“ 3 1293 016992 THESIS 6.3L This is to certify that the thesis entitled ACOUSTIC EMISSION AND RELATED DISPIACEMENTS IN LITHIUM FLUORIDE SINGLE CRYSTALS presented by Robert B. Engle has been accepted towards fulfillment of the requirements for Mdegree in Applied Mechanics #9 z - vw‘x/sL/L-f’f Major biofessor Date October 5. 1966 0-169 VJ LIBRARYU‘ Michigan 9 .ate i .1 University ; '! i ACOUSTIC EMISSION AND RELATED DISPLACEMENTS IN LITHIUM FLUORIDE SINGLE CRYSTALS By R obert B. Engle AN ABSTRACT Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOC TOR OF PHILOSOPHY Department of Metallurgy, Mechanics and Materials Science 1966 ABSTRACT ACOUSTIC EMISSION AND RELATED DISPLACEMENTS IN LITHIUM FLUORIDE SINGLE CRYSTALS by Robert B. Engle Acoustic emission is the result of lattice vibrations caused by mechanisms that govern deformation of crystalline materials when they are stressed. , Observations were made using single crystals of lithium fluoride oriented for easy-glide deformation when loaded in direct shear. Acoustic emission pulses and displacement pulses derived from rapid step displacements were observed. The largest displacements (10 x 10"6 inches) were found to occur in coincidence with large acoustic emission pulses. Smaller displacements (20 x 10"8 inches) and acoustic emission pulses were observed to occur together, though not in coinci- dence, throughout the tests. Some large acoustic emission pulses had no displacements associated with them. On the basis of these results a mechanism for the acoustic emission process is proposed, and is shown to be consistent with the observations of others who have studied acoustic emission in crystalline materials. Estimates of dislocation group velocities are made which agree with previous dislocation velocity measurements in lithium fluoride. The proposed mechanism involves the interaction between piled- up groups of dislocations and the obstacles that cause the pile-up. The pinning interaction causes an increase in local strain energy stored in the region of the obstacle. When the driving stress on the leading dislocation, composed of the applied stress and additional stress due ROBERT B. ENGLE to the pile -‘up itself, is large enough to cause breakaway and acceleration of part of the group, the local strain energy is available to excite lattice vibrations that appear as acoustic emission. A secondary emission process arises during collisions between coherent groups of moving dislocations and obstacles in their slip planes. Finally, an alternate type emission occurs at high stress levels after large plastic deformation when the stress concentrations at the leading edge of dislocation groups causes crack nucleation and propagation. ACOUSTIC EMISSION AND RELATED DISPLACEMENTS IN LITHIUM FLUOR IDE SINGLE CRYSTALS BY Ve/ Ll Robert Bf Engle A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Metallurgy, Mechanics and Materials Science 1966 (’ / t’ I /)<.-/ _. .~' ‘ g ' ‘_ 1-. .. / /, , /.I j “1’ c/ a" r" ~ ' / r“- //// fl ”5/ p7 / ACKNOWLEDGMENT I wish to thank Dr. Clement A. Tatro, who interested me in this project, and Dr. Terry Triffet, who assumed the role of major professor when Dr. Tatro went to Tulane University, for their encouragement and guidance. I gratefully acknowledge many informative discussions with faculty and graduate students of the Department of Metallurgy, Mechanics, and Materials Science and additional support from the department. The work accomplished for this thesis was supported by a National Science Foundation grant (0-14650) awarded to Michigan State University. ii Acknowledgments List of Figures . . . List of Tables . . . List of Appendices . I Introduction. . 11 Experimental Procedure TABLE OF CONTENTS III Presentation of Data . . IV Discussion V Conclusions. . . . . . . VI Suggestions for Further Research Bibliography . O O O O O O I O O O O O O Appendices......... iii Page ii iv vii viii 13 47 72 93 97 99 102 Figure 11. 12. l3. 14. 15. l6. 17. 18. 19. 20. 21. 22. 23'. LIST OF FIGURES Loadingsystem.................... Loading system supported on chains . . . . . . . . . Loading system -floating. . . . . . . . . . . . . . . Detail of crystal clamp and load bars . . . . . . . . . Data electronics, block diagram . . . . . . . . . . . Dataelectronics.................... Differential capacitor and load bars in assembly jig Differential capacitor installed in loader . . . . . . Sketch of differential capacitor. . . . . . . . . . . . Typical displacement calibration curve . . . . . . . Acoustic emission and displacement pulse frequency _ responseelectronics . Load transducer gage locations . . . . . . . . . . . . Load transducer bridge connection. . . . . . . . . . . Load calibration system . . . . . . . . . . . . . . . Acoustictransducer. . . . . . . . . . . . . . . . . . Acoustic transducer location . . . . . . . . . . . . . Relative location of acoustic and displacement transducers-pendulum removed from loader. . . . . Relative location of acoustic and displacement transducers with specimen revealed . . . . . . . . . Acoustic emission preamplifier installed . . . . . . LiF crystal orientations . . . . . . . . . . . . . Specimen geometry and state of applied stress . . . Easy-glide planes, TypeI crystal . . . . . . . . . . Easy-glide planes, Type II crystal. . . . . . . . . . iv Page 14 15 16 17 19 20 21 22 23 24 26 28 28 29 29 30 30 31 31 35 37 38 38 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. Easy-glide planes, Type III crystal. . . . . . . . Easy-glide planes, Type IV crystal. . . . . . . . Pulse counting and time interval system . . . . . RMS signal read-out ..... . . . . . . . . . . . Damped signal resulting from a single impulse . . Demodulatorcircuit . . . . . . . . . . . . . . . . Demodulated signal containing several impulses . Representative load-displacement curves. . . . . Acoustic emission pulse distribution - Type 1, Run 7. Displacement pulse distribution - Type I, Run 7 . . . Oscilloscope traces of demodulated acoustic emission and displacement pulses - Type I, Run? . . . . . . . Load-displacement and coincidences - Type 1, Run 7. Oscillograph record - event at T = 186 seconds . Oscilloscope traces - event at T = 186 seconds - Type I, Run 7. . . Micrograph of tested Type I crystal, Run 7 . . . . . Load-displacement and coincidences - Type 1, Run 3. Load-displacement and coincidences - Type 1, Run 4. Load-displacement and coincidences - Type 1, Run 9. Load-displacement and acoustic emission - Type 1' Run 13 O O O O O O O O O O O O O O O O O O O O O Load-displacement and RMS activity - Type 11, Run 12. Acoustic emission and displacement pulse distributions - Type III, Run 14 . . . . . . . . . . . . Load-displacement and acoustic emission - Type III, Rm 14 O O O O O O O O O O O O O O O O O O O O Oscillosc0pe traces - event at T = 436 seconds - Type I, Run 7O O O O O O O O O O O O O O O O O O O O O O 3 9 3 9 4O 42 44 44 45 49 52 54 55 56 59 62 64 65 67 68 69 70 83 47. 48. 49. 50. 51. 52. 53. 54. 55O 56O 57. 58. Dislocation velocity versus applied shear stress Acoustic emission pulse distribution - Type I, Run 3 O O O O O O O O O O O O O O O O O O O Displacement pulse distribution - Type 1, Run 3. . Acoustic emission pulse distribution - TypeI, Run4. . . . Displacement pulse distribution - Type I, Run 4. Acoustic emission pulse distribution - Type 1, Run 9 O O O O O O O O O O O O O O O O O O O Displacement pulse distribution - Type I, Run 9. Load-displacement and acoustic activity - Type I, Run 10 Oscilloscope traces Oscilloscope traces Oscillosc0pe traces Oscilloscope trac e 5 event at T event at T event at T event at T vi 292 seconds . 381 seconds . 465-7 seconds 478 seconds . 89 116 117 118 119 120 121 122 123 124 125 126 LIST OF TABLES Page List of runs and summary of results . . . ..... . 48 Pulse data from oscillograph and oscilloscope traces for coincident events, Type 1, easy- glide, Rm 7 O O O O O O O O O O O O O O O O O O O O O O 61 Dislocation velocities for coincident events - TypeI,easy-glide,Run7 ............. . 80 vii LIST OF APPENDICES Appendix Page A. Equipmentlist ..... 102 B. Acoustic channel gains and displacement pulse sensitivities...................... 104 C. Load-displacement data. . . . . . . . . . . . . . . . 105 D. Pulseheightdata................... 109 E. Additionaldataplots 116 F. Oscilloscope traces - Type 1, Run? . . . . . . . . . 123 viii INTR ODUCTION The fact that tin emits sounds when it is deformed has been known for many years? {Around 1930 Orowanl and Klassen -Nekludowa‘2 investigated this phenomenon in tin, that is now associated with the formation of mechanical twins. Other researchers have reported 3, 4, 5 ; however, no noises occurring with deformation in metals great significance was attributed to them. In 1950 the late Joseph Kaiseré’ 7 started experiments to determine if other materials produced acoustic phenomena when loaded. Tests of steel, aluminum, copper, lead, zinc, and wood revealed sub -audible acoustic pulses accompanying deformation. Each material produced a characteristic noise spectrum with amplitude and frequency distributions that correlated with the different regions of the stress-strain curve. He observed that the onset of emission occurred before the macroscopic yield point was reached, emission continued for a period of time when the load was held constant, and the emission process was irreversible. These observations led him to conclude that acoustic emission was related to plastic deformation in favorably-oriented crystals. Friction due to relative motion at crystal boundaries was suggested as the source mechanism. Subsequent investigations have confirmed the occurrence of acoustic .emission from all of the crystalline materials, that Kaiser tested, and various others as well, but have associated acoustic emission with more basic deformation mechanisms. They have not confirmed the rather detailed relationship between stress and the distributions of amplitude and frequency that Kaiser reported, but have exhibited the other characteristics that he observed. Interest in acoustic emission as a possible tool for non- 8, 9,10,11 destructive testing led Professor C. A. Tatro to start investigations while he was at Michigan State University from 1956 to 1962. B. H. Schofieldlz"16 at Lessells and Associates, Inc. , also started research into the acoustic emission phenomenon late in 1954. The fact that Kaiser, Tatro, and Schofield have all observed acoustic emission from every material they have tested suggests acoustic emission may itself become a research tool for deformation mechanics, if its sources can be identified and techniques for observation are refined. Both Tatro and Schofield believe the acoustic emission process is intimately related to microscopic mechanisms governing deformation of crystalline structures. Acoustic emission observed by Schofield characteristically consists of two components: a relatively high amplitude, low frequency, burst-type which occurs randomly with long quiescent periods between bursts; and a high frequency component with smaller amplitude that is continuous in nature, appearing much as broad- band noise in electronic equipment. Both components appear in elastic and plastic regions of the stress-strain curve. Schofield's results indicate there is a minimum strain rate below which the high frequency emission is notably absent. The equipment used at Michigan State University wascapable of only very low strain rates, and the fact that no high frequency emission has been observed there is in agreement with Schofield's observations. If the strain rate is high enough, high frequency emission generally starts before the nominal elastic limit is reached, increases in amplitude as the elastic limit is reached, and then diminishes in amplitude as plastic deformation continues. The emission pulse rate is observed to increase with increasing strain rate, and the frequency spectrum apparently shifts to higher frequencies as plastic strain increases. The burst-type emission also appears in most of the materials tested. However, the relative contribution of the burst-type is quite small when compared to the high frequency contribution above the threshold strain rate; nor can the total plastic deformation be explained on the basis of total acoustic emission for those tests made at low strain rates when the high frequency emission is not observed. The largest group of experiments has been run on aluminum alloys and pure aluminum. Schofieldlz’ l3 , reporting on tensile tests of specimens machined from 24ST-4 aluminum, characterized the emission as being primarily of the high frequency type; initial pulses were of lower frequency with higher frequency pulses starting at higher stresses with increasing rate. Burst-type emission was present with much higher amplitude,but with much less contribution to the total pulse count. The alloy polycrystalline specimens exhibited more of the burst-type emission than single crystal specimens of pure aluminum. Tests with pure aluminum single crystals produced much the same general behavior for high frequency emission, with some differences in character that could be related to the orientation of the crystals. The burst-type was notably absent during the early stages of deformation, but did occur during later stages of plastic deformation. In another series of tests1 5 single crystals of aluminum were loaded while immersed in an etchant bath so that the surface oxide film would not be a contributing factor. Crystals with the tensile axis in the (100) direction still exhibited high frequency emission in about the same amounts, though its initiation stress was shifted to a higher value. Very little burst-type was seen. On the other hand, single crystals oriented with a Schmid factor of 0. 5 gave no emission to strains as high as 20%. Polycrystalline specimens with grain densities ranging from 25 to 160 grains per cm2 were tested and gave no noticeable difference in emission from the oriented single crystals, thereby ruling out the importance of grain boundaries to the emission process as proposed by Kaiser. 1o, 11 and Shoemakerl7 on 2024 T-4 aluminum Reports by Tatro specimens tested at a low strain rate give more information on the burst-type emission because the high frequency component was not present to mask its occurrence. They observed burst-type emission in the elastic range that shifted to higher frequencies with lower amplitude and faster pulse rate in the plastic range, much as the high frequency emission of Schofield; however, the peak emission was recorded before the yield point was reached. 18,19 Tests by Liptai and Tatro11 on specimens of 2011 T-3 aluminum were nearly the same, except emission peaks were recorded in the plastic range. The only surface treatment that restored initial emission activity was electropolishing that removed 0. 05 inches from the surface. Electropolishing combined with electron bombardment partially restored initial emission activity, but shifted its onset to higher stress levels. R. G. Liptai18’ 19, testing single crystals of aluminum with a (100) tensile axis, reported burst activity much the same as for the 2011 T-3 specimens; except the activity seemed to be enhanced in the plastic region. He found thick anodized and reacted coatings increased the acoustic activity, while electropolishing afterwards would reestablish the as-received activity of the specimen. When testing single crystals of zinc, Schofield13 observed burst-type emission as the salient feature of the emission spectrum. Bursts occurred immediately upon loading at low load rates. Pulse rate and amplitude increased with stress, and emission continued for a period of time if the load was held constant. High frequency emission did not become apparent until gross deformation set in and stopped immediately when the load rate was reduced to zero. The amplitude of high frequency emission increased with deformation. A crystal coated with photoelastic adhesive gave several extremely energetic bursts associated with the formation of twins, though most of the burst-type emission occurred at stresses below that commonly thought of as being required for twin formation. One bi -crystal was observed to give an energetic burst associated with a grain boundary reorientation. 8-13, 17 Probing experiments have been run on carbon steel 3, magnesium and 40/60 solderlo, copper and lead13, .9 tinlz, brass and pure ironl6. All were observed to produce acoustic emission. R. T. Sedgwick20 tested single crystals of lithium fluoride in (100) compression in the elastic stress-strain region and observed what may be another burst-type emission. Crystals with thick surface films, that tended to raise the elastic stress-strain curve, produced burst-type emission when the films were etched from the loaded crystal. Burst-type emission was also observed in the elastic region during loading for both coated and uncoated crystals. This emission exhibited a delay character by continuing for a while after each increment of load. This is characteristic of burst emission observed by others in metals. Since the introduction of the concept of the moving dislocation in 1930 by Taylor“, Orowanzz, and Polanyi23, dislocation motion and interactions have played a major role in explaining the behavior of materials under load. The importance of dislocation theory in explaining the yield strength, ductility, work hardening, fatigue, and fracture of materials is recognized and supported by many experimental results. There are, however, few techniques that allow direct observation of dynamic behavior of dislocations. The sources of the acoustic emissions observed to date in crystalline materials are still subject to debate, though all of the present interpretations have one point in common. In each case, the results are most easily interpreted as being due to the motion of groups of dislocations, possibly to the motion of single dislocations. Previously, dislocation behavior has been inferred from metallographic, X-ray, and electron microscope studies that are inherently limited to observation of events near the surface, or sound attenuation measurements that can measure gross changes but not the details of dislocation behavior. Acoustic emission may be a phenomenon that will allow direct observation of dislocation motion. The earlier work of Schofieldlz’ 13' 14 , and the experiments of Tatro, Shoemaker, and Liptai all seem to indicate that acoustic emission is a surface phenomenon, or is controlled by the nature of the surface. Schofield's measurements on aluminum single crystalslz, for example, lead to a remarkably close, though tentative, correlation between the number of high frequency emission pulses and the number of slip lines formed at the surface. Liptai's results with coatings on aluminum1 8’ 19 also strongly indicate that burst-type emission is due to surface effects in crystals. Both, at the time, interpreted acoustic emission to be the result of energy released when piled-up dislocations broke through the surface barrier and left the crystal, causing the formation of new surface accompanied by heat and an elastic wave. This explanation is strengthened by those experiments that investigated the effects surface treatments and films had on the character of the emission. The burst-type in aluminum was observed to increase with coating thickness1 8’ 19. Shot-peened 1018 steel specimens produced more activity shifted to higher stress levelsll’ 17. Aluminum single crystals tested in an etchant bath produced either no emission (0. 5 orientation) or high frequency emission that started at higher stress levels for (100) orientation14’ 15. Later tests by Schofield15 on gold single crystals .indicated that some emission is not a surface phenomenon. Gold was chosen because it does not have an effective surface film. In these tests burst-type emission was much more prevalent, appeared to be partly reversible, and increased in activity with continued plastic strain. The high frequency type started immediately upon loading, peaked at about 0. 08% strain, and decreased thereafter. The burst-type emission was much more energetic when specimens were annealed after a test and were then retested. In this condition the crystals had coarse-grained annealing twins in the test section. Etching away a surface layer did not restore or alter the character of the emission as it did for aluminum. Cold work of screw faces or edge faces of the 0. 5 crystal did enhance the high frequency emission; and, in this case, surface etching returned the emission to normal. These results have led Schofield to identify the burst-type emission with the formation of stacking faults in gold and the high frequency emission with the movement of dislocations in the interior of the sample. Zincl3 also exhibits burst-type emission which Schofield has tentatively related to the formation of stacking faults (more probably micro-twins) and, in the more energetic cases, to twin formation and grain boundary reorientations. Schofield explained the relative lack of burst-type emission from aluminum by the fact that the energy required for the formation of stacking faults in aluminum is much higher than the energy required for the formation of stacking faults in gold or for the formation of micro -twins in zinc . Schofield now feels the acoustic emission process is related to internal deformation processes instead of surface effects. However, in so far as surface condition effects the motion of dislocations in the interior regions, surface condition can alter the acoustic emission process. The emission that R. Sedgwick20 observed in lithium fluoride appeared to come from two different sources. The bursts, observed when the surface films were etched away under load, can definitely be attributed to the egress of groups of dislocations that were piled up on a. slip plane due to the surface barrier. The fact that the surface coatings tended to raise the stress-strain curve led him to the conclusion that dislocations have some role to play in elastic deformation. He formulated a model depending upon the dislocation network present in the unstressed crystal to provide dislocation segments in favorably oriented slip planes that are pinned at points where they leave the planes. These segments act as Frank-Read sources which are activated by stresses too small to cause dislocations to leave the crystal, but which can contribute to deformation. Yield occurs when the force on the leading dislocation on a slip plane is large enough to cause it to leave the crystal. Back stresses would cause the sources to operate in reverse, so that the process can be thought of as a quasi-elastic one. According to Sedgwick's model the initial acoustic emission is due to the activation of these Frank-Read sources. Sources with the longest loop length will be activated at the lowest value of stress and will continue to Operate until dislocations emanating from them 10 become blocked and pile up in the slip plane. As stress increases the shorter sources activate. The emission due to the initial dislocation distribution should be symmetric about the activation stress required for the average loop length. New dislocations produced can also block the operation of the original source distribution, having the effect of shifting the mean loop length to lower values. This secondary effect causes the expected distribution of effective sources to be skewed towards the shorter length, and should, therefore, produce an emission distri- bution that is skewed towards higher stress as was observed. This mechanism, if verified, may explain the elastic region emission observed by others. In reviewing the data reported above, it becomes increasingly obvious that there are a multiplicity of mechanisms which can act as sources of acoustic emission. Any process involving a rapid transition between different dislocation configurations within the crystal lattice must also produce an accompanying acoustic emission. Whether or not the emission is detected would depend upon two factors. The first is the rate at which the process proceeds, and the second is the amount of energy dissipated in vibrational modes as the transition occurs. Acoustic emission pulses have been variously identified with slip line formation, formation of stacking faults, cracking of surface films, release of piled up dislocations, action of Frank-Read sources, twinning, grain boundary reorientation, and motion of individual dislocations in the interior of the specimens. The only estimate of 11 displacement has been made by Schofieldl3, who found that there was nearly a one to one relation between slip line formation in aluminum and the total high frequency pulse count. His estimate was based upon average values of slip line spacing, which are subject to question; these indicated that the average emission pulse corresponds to a displacement of from 50 to 150 x 10.8 cm and involves the motion of from 20 to 50 dislocations in aluminum. Similar estimates were made for zinc with the result that the displacement range was from 15 to 50 x 10"8 cm, corresponding to the motion of from 5 to 20 dislocations. A direct measurement of the displacements actually involved would greatly assist the positive identification of the acoustic emission mechanisms. R. L. Sproull32 has successfully used a capacitance-type transducer to measure displacements on the order of 1 Angstrom. Sproull's device is awkward for measurement of dynamic displace- ments, but his success indicated a capacitance transducer might provide the sensitivity required to detect small displacements related to acoustic emission. References3 to an ionization transducer utilizing a capacitance sensor looked promising, and efforts were directed towards developing a displacement transducer that would allow the measurement of such small dynamic displace- ments. If the necessary sensitivity could be achieved, acoustic emission measurements were planned to try to correlate the emission pulses with discrete displacements in order to learn more about the phenomenon. To identify the source or sources of acoustic emission, it is 12 imperative that the specimen's deformation modes be severely limited, and that as much as possible be known about dislocation motion in the material. Lithium fluoride is ideal for this purpose, since so much is already known from the experiments of Gilman and Johnston24-3l. It is well established from their work that lithium fluoride deforms plastically by {110} < 110> slip at room temperature. Slip is possible on other planes only at elevated temperature. Appropriate load geometry can limit slip to one or, at most, two slip planes. By utilizing appropriate orientation and load geometry to limit the possible deformation mechanisms, it was felt that it should be easier to relate acoustic emission to displacement. When it became apparent that the necessary displacement sensitivity was possible, work was started on making a direct- shear loader that would be quiet enough for acoustic emission studies. The objectives of the research were: first, to determine whether or not discrete slip displacements accompany acoustic emission; second, to determine the magnitude of the displacements if they exist; and third, to determine whether or not acoustic pulse height can be related to the size of such displacements. EXPERIMENTAL PR OC EDUR E The loading system shown in Figures 1, 2, and 3 was designed to provide noise-free, direct-shear loading to bar-shaped specimens. The system consists of a platform supported by two floats riding in independent water tanks. A pendulum is suspended from a frame that is mounted on the platform. The specimen under test is clamped between the pendulum and the platform. When the water level in one tank is lowered, the component of the pendulum weight down the resulting incline is applied to the crystal. The pendulum is restrained by the specimen under load, the four support wires, and four lateral wires that prevent sideways translation and rotation about the longitudinal axis of the specimen. The load system can be tilted to an angle of 30 degrees in ten minutes which gives a load rate of 4. 5 pounds per minute. This rate is constant, due to the combined effects of the lateral restraints and the variation in water flow as the tank water level decreases. Theload is transmitted to the crystal by square steel bars (1/2 in x 1/2 in) that also hold the displacement, load, and acoustic emission transducers. Figure 4 shows the detail of the crystal clamps and load bars. Simple beam theory predicts that bending moments in the specimen will produce a flexural stress of not more than 3. 2% of the maximum shear stress or 10% of the average shear stress applied by the loader. The load on the pendulum support wires decreases as the platform tilts. This places an axial load on the specimen 13 14 L-- O J \h¥ .: _—— 1’ K (Figures 4 and 16) u u \ l I ‘ I I /D‘,'l ‘ 1‘ I'll ' . / ,"\ / \ / \\\ / \ \ , \\\ / \\ / \\\ ‘\ // —“' \ -—’ X)‘\ A t E ) A \\’, muse? Water tanks. Flotation tanks. Load valve. Pendulum support wires. Pendulum weight. Bottom clamp assembly. Initial water level. Crystal load bar. Stabilizing counterweight. Platform. K. Specimen location. Figure 1. Loading system. 15 Figure 2. Loading system supported on chains. l6 Loader level - no load Figure 3. Loading system - floating. Dir ion of o L__ V Top load bar Top clamp % Pendulum Ir/ lower plate % Strain gages 'of load transducer -— . Displacement transducer moveable plate holder Specimen e35 ' 0. 003.,- 0. 005 inch . test' gap Displacement transducer side plate holder _. .i y . Platform / \ a la Well for acoustic transducer Bottom clamp B ottom load bar Figure 4. Detail of crystal clamp and load bars. 18 such that the ratio of average tensile stress to average shear stress increases nearly linearly with the angle of inclination of the loader from a value of zero at 0 degrees to 0. 28 at 30 degrees tilt. These extraneous loads caused slip on slip planes oriented at an angle to the load plane of the specimens. The load geometry, described later, severely restricted such undesirable slip in all cases. Furthermore, these additional stresses had no effect on the magnitude of the shear stress on the load plane of easy-glide specimens which were of primary interest. Acoustic emission, load, displacement, dynamic displacement, and time data were taken with the systems shown in Figures 5 and 6. During the first few runs, a consistent time reference was provided by recording the output of a time-mark generator on one channel of the tape recorder, and monitoring the recorded time marks with a counter in order to record remarks in the operating log based on test time. In later runs, and for all read-outs, the time marks were also displayed on the oscillograph to allow timewise correlations. These time marks were used to establish time intervals for the electronic pulse height analysis that was run for most of the tests. The displacement transducer was of the differential capacitance type utilizing a modified Decker Delta unit. Because the Delta unit injected too much noise into the system, it was modified by using batteries to supply filament and B+ voltages to the oscillator. In order to extend the frequency response of the unit and to facilitate its mounting on the loader, the Delta unit follower circuit was duplicated in a custom chassis mounted on the loader. The R. F. 19 ' TIME ACOUSTIC DISPLACEMENT LOAD REFERENCE EMISSION TIME MARK ADP DIFFERENTIAL LOAD GENERATOR CRYSTAL CAPACITOR BRIDGE PREAMPLIFIBR DECKER BRIDGE PROBE AMPLIFIER. FILTER CATHODE PROBE FOLLOWER EXCITER OSCILLOSCOPE Low PASS POWER AMPLIFIER FILTER SUPPLY I BUFFER AMPLIFIER AND _ SIGNAL I— CONDITIONER—l , __ - ' r PREAMPLIFIER PREAMPLIFIER X-Y , i RECORDER FILTER FILTER 'OSCILLOSCOPE ZERO AMPLIFIER SUPPRESSION SEVEN CHANNEL TAPE RECORDER T w I F DRIVER FILTER FILTER AMPLIFIER l DRIVER DRIVER CALVANOMETER DRIVER AMPLIFIER AMPLIFIER. MATCH AMPLIFIER MULTI -CHANNEL LIGHT BEAM OSCILLOGRAPH Figure 5. Data electronics, block diagram. 20 Right side Left side Figure 6. Data electronics. 21 excitation was supplied by the oscillator in the original Delta unit. The Decker probe was then connected directly to the cathode follower, thus reducing stray capacitance loading to a minimum with a resulting increase in frequency response. Driven shields were utilized to further reduce the effects of stray capacitance on the frequency response and linearity of the system. The differential capacitor is shown in Figures 7 and 8. The large central plate operated at ground potential and was attached to the lop load bar so that it would be the moving element. The smaller side plates were mounted on the lower stationary bar. Figure 7. Differential capacitor and load bars in assembly jig. 22 Figure 8. Differential capacitor installed in loader. The capacitor plates and the spacer used to mount the side plates were made of Invar. This eliminated effects of temperature change on the capacitor spacing and transducer sensitivity. The transducer responds to rotation about a vertical axis through the specimen and lateral displacement due to bending of the load bars (estimated at less than 10 x 10-6 in/lb of load) as well as to relative displacements of the crystal. The pendulum was restrained by lateral wires to prevent rotation. No method was found to eliminate contributions to the displacement due to the bending of the load bars. 23 The differential capacitor spacing ((11 + d2, Figure 9) was typically less than 0. 030 inches. This spacing gave sensitivities on the order of 2 millivolts per microinch displacement. The edge effects due to the close spacing, slight misalignments, and unavoidable amounts of stray capacitance made the output of the transducer non- linear with displacement. Proper selection of initial settings provided a sufficiently linear output for the range required. 0 Direction of motion of top load bar Adjustment 3 c rews Center plate (Imer) l \ Invar \ Spacer V? A ‘ — j . , , __ ////// "IL—1;- ‘7 ”— Epoxy laminated fiber glas s Side plates (Invar) Specimen position Center plate holder attaches to top load bar. Side plate holders attach to bottom load bar. Figure 9. Sketch of differential capacitor. In-place displacement calibrations were taken after each run in order to measure the transducer sensitivity at every point in the useful range. Figure 10 shows the calibration curve for Run 7. These were made by removing the specimen and translating the pendulum with 24 a micrometer screw in increments of 0. 001 inches over the full range of the transducer. It was possible to locate the initial position of the unloaded crystal on this curve, making the displacement sensitivity known throughout each test. +12 T +9 -- +6 .. +3 ‘- 0.. -3. I. -15 : e : - : : : . : 4 : - - -7 -6 —5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6 Displacement - 0. 001 inches Figure 10'. Typical displacement calibration curve. Transducer output - volts Initial adjustments of the transducer before installation in the loader were limited to aligning the central plate to be as parallel as possible with the side plates. After the transducer and test specimen were installed in the loader, a capacitance meter was used to adjust the initial capacities to values which would give the necessary sensitivity and linearity. Sensitivities were determined by calibration checks immediately after each test was completed. 25 The basic transducer output was a differential D. C. voltage. This output was fed to parallel cathode follower amplifiers which provided several outputs for a variety of signal conditioning and data recording purposes. One output was coupled to a preamplifier through a pulse transformer in an attempt to decouple the D. C. level and allow the amplification of dynamic components. This output was called the displacement pulse output. An alternate output was coupled to a preamplifier with a capacitor. This capacitor was chosen to cause differentiation of the displacement signal which gave a pulse for each step-like change in the displacement. This output, called the step displacement, and the displacement pulse output were amplified, filtered as necessary, and recorded on separate tape recorder channels. A third coupling amplifier was used to apply the D. C. displace- ment signal to the X-axis of an X-Y recorder to obtain a load-displace- ment curve. During later tests this same signal was applied simultaneously to a channel of the oscillograph in order to obtain a load-time record. When the oscillograph was obtained, the step displacement signal was recorded simultaneously on magnetic tape and the oscillo- graph. The acoustic emission and displacement pulse signals were recorded on the oscillograph through the tape recorder. Signal substitution was used to determine the frequency response of the dynamic displacement and acoustic emission channels. The calibration electronics are shown in Figure 11. 26 VARIABLE OS CILLATOR l ATTENUATOR 1 PULSE TRANSFORMER "“1-“I H |‘—_——‘| DISPLACEMENT l INPUT | ACOUSTIC PROBE MONITOR EMISSION l CATHODE I OSCILLOSCOPE I PREAMPLIFIER FOLLOW'ER I I ' I | I l Igl i ' | TAPE | OUTPUT I TAPE l RECORDER MONITOR I RECORDER I OSCILLOSCOPE I | I I DATA | DATA | ELECTRONICS _J L_ -ELECTRONICS _J Figure 11. Acoustic emission and displacement pulse frequency response electronic 3. 27 The load transducer was a four -arm strain gage bridge applied to the load bars (Figures 12 and 13). Gages 1 and 3 gave positive outputs for tensile strain, while gages 2 and 4 gave positive outputs for compressive strain. The configuration only responds to flexural strains produced by shear loading at the test section. Bridge power, bridge balancing, amplification, scale adjustment, and electrical calibration were provided by the bridge amplifier. The bridge amplifier output was then applied to a galvanometer amplifier for recording on the oscillograph and to the Y-axis of the X-Y recorder to complete the load-displacement curve. A load Calibration was performed using a split specimen and weights (Figure 14). The load bars were positioned and clamped in the loader as they would be during an actual test. The effect of a calibration resistor in the bridge amplifier was noted during the direct calibration, and this standard indication was used to interpret the load transducer output in later tests. The acoustic transducer, shown in Figure 15, was the same ADP crystal stack used by Liptai18 and Sedgwickzo. It was placed directly against the specimen in the well in the specimen clamp heads opposite the displacement transducer. The transducer location is illustrated in Figures 16, 17, and 18. The transducer was held in place with silicone vacuum grease, which also provided acoustic coupling between the transducer and the crystal under test. Figure 19 shows the acoustic amplifier, which was mounted on the loading frame, as close as possible to the acoustic crystal, to reduce attenuation of the acoustic output by capacitance loading. The 28 A Load direction Gagel [ J Gage 2 TOP load bar / Split specimen for / load calibration w Btt 1 db oom oa ar Gage 4 L J Gage 3 a Figure 12. Load transducer gage locations. Gage 4 \Gage 1 / Bridge supply voltage Gage 2 o———-B ridge output Figure 13. Load transducer bridge connection. Load direction Pendulum Split restraint specimen Pendulum \Kfl—l / ‘\ Platform . Figure 14. Load calibration system. . Figure 15. Acoustic transducer. 30 Top view Side view Figure 16. Acoustic transducer location. Figure 17. Relative location of acoustic and displacement transducers - pendulum removed from loader. 31 Figure 18. Relative location of acoustic and displacement transducers with specimen revealed. Figure 19. Acoustic emission preamplifier installed. 32 signal was then filtered and further amplified with the oscilloscope amplifier before being recorded on the tape recorder. In the earlier tests the signal recorded on tape was recorded on the oscillograph by simultaneous playback of the tape recording. In later tests, both the tape recording and the oscillograph recording were obtained simultaneously from the same amplifier output. An assembly jig was used to hold the load bars in alignment. while initial adjustments were made on the displacement transducer. This allowed adjustment of the test section gap and facilitated the installation and clamping of the test specimen. After the specimen was installed, the jig maintained alignment while the load bars with the specimen and displacement transducer were installed in the loader. The load transducer was monitored throughout the assembly and installation process in order to determine the size of any preload the specimen might be subjected to before the test began. A typical run required twelve hours and was timed so the actual loading test occurred after midnight when the laboratory was most quiet. Test preparations were started by supporting the loader with chains in order to make adjustments in pendulum alignment and to install the specimen. The specimen load bars were positioned in the assembly jig, and the crystal to be tested was clamped in place. Then the displacement capacitor was installed and initial adjustments were made. Next, the assembly jig with the clamped crystal and transducer was lowered into position in the loader. After the upper clamp 33 assembly was secured to the pendulum, the upper and lower grips were tightened; and the assembly jig was removed. During this installation procedure, the pendulum was restrained from movement by a pressure-operated piston mounted on the platform. This piston remained in place until the load test was started to prevent accidental loads on the specimen while adjustments were made. Final spacing and position adjustments were made on the differential capacitor using a capacitance meter to determine the correct values. Then the Decker probe, its cathode follower, and the acoustic preamplifier were mounted on the loader. At this point the loader was floated and the chains were removed. The data electronics were then checked, and calibration ‘ signals were recorded using the complete recording system. Just prior to the actual run, all gain and frequency settings were rechecked; and all parts of the system were turned on. Then the restraining post was removed to free the pendulum, a load . calibration was placed on the X-Y recorder and oscillograph, and the load valve was opened. In order to keep a test time record, the time mark signal (one pulse per second) was applied directly to the tape recorder. The simultaneous playback signal was then applied to an electronic counter and to the oscillograph. The oscillograph was turned on first, and test time started when the tape recorder was placed in its record mode. Event times were noted in an experiment log or on a supplemental tape rec order. 34 The seven channel tape recorder was used only for recording data during loading of the lithium fluoride crystals. When full load was reached, the signals being recorded on the oscillograph by using the playback from the tape recorder were switched directly to the oscillograph, so events occurring during unloading could be observed. When the loader was again level (no load), the specimen was removed from the load bars; and a steel comparison specimen was tested. Finally the steel comparison specimen was removed, and a direct displacement calibration was placed on the oscillograph and the X-Y recorder. The position of the displacement transducer was continuously monitored while specimens were being changed so that the initial position for both the lithium fluoride crystal and the steel comparison specimen could be located on the displacement calibration chart. The crystal orientations chosen were those illustrated in Figure 20. Type I was a {110} < 110> , easy-glide cyrstal. The first bracket refers to the plane on which the direct-shear load was applied. The second bracket specifies the direction of the stress on the loaded plane. Type II, III, and IV crystals had {100} < 110> , {110} < 100> , and { 100} < 100> orientations respectively. Of these four orientations, type III and IV crystals were expected to exhibit brittle fracture, since no easy-glide plane had a shear stress component in the easy-glide direction. The type II crystals were subjected to shear stress in the easy-glide direction on four of the six possible easy-glide planes. The amount of glide is limited, due to the geometry of the crystal clamps; and limited slip followed by 35 a b c / '{I‘lylpg}l<110> Type IV ’\a {loo}<100> _ / / a \b Type III {110} <1oo> [110] 100] I \ Six easy-glide planes 1'1 0] [001] c l I Type II I {loo}<1lo> I I I A Figure 20. LiF crystal orientations. 36 crack formation and fracture was expected. The loading placed shear stresses on all six easy-glide planes in the easy-glide direction in the Type I, easy-glide crystal. However, the primary slip system had twice the shear stress of the four other slip planes. Since one plane of the primary slip system was the plane of loading and was not restricted, this crystal was expected to show large plastic deformation before fracture. Figure 21 illustrates the specimen geometry and the state of applied stress. Figures 22 through 25 show the particular orientations of easy-glide planes relative to the plane of loading for all four orientations. The magnetic tape recordings of each run were subjected to extensive read-out by three different methods; electronic pulse counting, visual pulse counting from transcripts obtained with a light-writing oscillograph, and RMS signal measurements using a true RMS voltmeter in conjunction with an X-Y recorder. Events of particular interest were photographed using a storage oscillo- scope. Electronic pulse counting yielded the most reliable pulse height distributions as a function of time for both acoustic emission and displacement pulses. The counting system is shown schematically in Figure 26. The input band-pass filter was necessary to obtain a favorable. signal to noise ratio for read-out. Investigation proved the most favorable band-pass settings to be 380-420 hz for the displacement channel and 3. 6-3. 8 Khz for the acoustic emission channel. 37 /- - A-L \ Load plane 1 inc 1-...-- .._JL_ ll . lol/8 incha‘ L l" "'I Tabyr'o llll II __d 'J);“’ SIT--- V*"" Figure 21. Specimen geometry and state of applied stress. 38 Jeanine 3 2:5 .moddfim Ovfimlramdm .MN Ohdmwrm So: a L \ -IA \ x _. $52 A VIIM: ll 0 Q SQ," 3S L. \\ \\\ .IV 2:: 8:78 Sod. u floor- v x . ohms" SHZB\ \ Ade EH2, . n\ / . Hmoam H-ooL So: a III adores" 38:. map, 388 ear. I as: w , odd," :Hao g x \ \\10b 3:: \ 8:: Hauam HH8H AmHOV do Got ao 4.3930 H 2:8 .monda onfimlkmmm .NN ouummh o I HoHHH . _ . o moaam A33 do L. I o I“ :23 no r. I own.“ SH: a IV \. SHE l I \\ .. I. .J . . a. \, N\Ob Mu HHHOH—b A N p— ~ N\oa. n EELS 4%qu Z a Sr: 52: 8H1... 2 Iv \ will \ \Jofl o So: WV IVA—n ~\ A. H .. b \\x / 30» n Somme \\\\\\\\z . ammxxoax .. 0..." :2 o 39 .dmumlwno >H 09mm. 439:0 HHH 2th Jenna onfimlamdm .mN ouamwh £053 onfimlhmmm .vN ouswwh o» u 38:. .o 82V :0 "Io 33 oo o I. a. I p \Sofl :2: a IV \. Home \ ..mm.~, 38.8 I x _ N\Ob .lll mOHOH—b ‘/ we _ \ so: .5 \ . .. «has n Soda AM. \ 5o: SH: :2: a Aoaavao Qoru ”ooze. Niche" I Amomv :0 40 PULSE INPUT TIME MARK INPUT FROM FROM TAPE RECORDER TAPE RECORDER BAND-PASS FILTER OSCILLOSCOPE OSCILLOSCOPE t ELECTRONIC ELECTRONIC COUNTER COUNTER (PULSE) (TIME) Figure 26. Pulse counting and time interval system. 41 The trigger level adjustment, in conjunction with a square wave calibration signal, was used to set the oscilloscope to trigger only on pulses larger than a selected value. The positive gate output of the oscillosocope, which provided a pulse each time the sc0pe sweep was triggered, was applied to the counter. The dead time of the system, which consisted of the sweep time plus retrace time, could be varied by changing the sweep speed. Sweep speed was adjusted in each instance to minimize both the number of pulses missed, due to too long a sweep time, and multiple triggering on the same pulse, due to too short a sweep time. The appropriate speed was relatively easy to determine by visual observation of the oscilloscope. The total number of triggered sweeps was recorded at the end of each 10 second interval. The time intervals were determined by observing a second counter totaling the time reference pulses (l per second) that were recorded on the tape. Each data channel was scanned in this way at various trigger levels ranging down to' just above the steady noise level. Each signal channel was also played back through the system shown in Figure 27. This displayed the RMS signal strength versus time. The recorder pen was lifted every 10 or 20 seconds to indicate the end of each time interval used in the electronic pulse height analysis. The RMS voltmeter attenuator was set to give a small deflection for background noise, and the resulting traces show regions of acoustic emission or displacement pulse activity as spikes above noise. It was felt that this record might indicate regions of activity due to many small pulses that might not be apparent in the 42 INPUT FROM TIME MARKER TAPE INPUT FROM RECORDER TAPE RECORDER TRUE RMS _ . OSCILLOSCOPE VOLTMETER ,- 1"), :3, 3’. I; ‘ RMS MONITOR GATE OUTPUT A OUTPUT X-Y RECORDER COUNTER (x = TIME) (TIME) Figure 27. RMS signal read-out. 43 electronic pulse height analysis. It was possible to confirm coincident events by comparing the acoustic emission and displace- ment pulse RMS traces. The three primary data channels (time reference, acoustic emission, and displacement pulse) were also read-out with a light- writing oscillograph with frequency response to 3 Khz. Various signal filtering and tape playback speeds were used. These records, in addition to the similar record obtained during the tests, assisted in confirming pulse counts and coincident events. Demodulators were used to study the structure of both the acoustic emission and displacement signals in more detail. A mechanical impulse excites both the acoustic emission and displace- ment transducers to damped oscillations at their natural frequencies. In the absence of reflections these vibrations produce voltage signals that appear as damped sine waves. A single impulse will produce a signal that looks like the sketch in Figure 28. If this signal is demodulated with the circuit shown in Figure 29, the resulting signal will be an exponentially decreasingipulse. The rate of decay is governed by the time constant of the smoothing capacitor and the resistance R, usually the load resistance presented by the measuring instrument at the output of the demodulator. If the demodulator time constant is chosen to give just the enve10pe of the input signal, a single impulse will produce a single smooth demodulator output pulse (dashed line in Figure 28). If, however, other impulses are applied to the transducer before the signal from the first impulse has disappeared, the resulting signal 44 Figure 28. Damped signal resulting from a single impulse. PM I I T 4— Figure 29. Demodulator circuit. 0 w 45 will appear to have additional spikes as shown in Figure 30. Additional impulses may cause either an increase or a decrease in the log decrement of the input signal, depending upon whether the impulse is out of phase or in phase with the natural vibration of the transducer element. Only those impulses that produce a decrease in the log decrement would appear on the demodulated output. because half of the impulses, on the average, will be out of phase with the vibrating transducer. The number of events determined from the number of spikes in the demodulated signal will, therefore, only represent approximately one half of the events occurring. Figure 30. Demodulated signal containing several impulses. All crystals were investigated with a polarizing microscope to determine the mode of deformation. Only Type I and Type IV crystals were oriented to make dislocations visible in polarized light, since the Burgers vector must be in the direction of the analyzer or polarizer transmission axis. The other two crystal orientations would be expected to show the presence of stress 46 concentrations due to dislocations only if many are introduced into the specimen. PRESENTATION OF DATA Tests were run on fourteen crystals: nine Type I (easy-glide), two Type II, two Type III, and one Type IV. Of the tests of the Type I crystals: two were exploratory, one crystal was broken while being installed in the loader, and two were found to be misoriented. The remaining four Type I crystals all behaved consistently. Run 7 was the best documented test and data from it will be presented in detail. The other types were tested primarily as control crystals, since little or no emission was expected from any of them. Table 1 lists the tests and a brief summary of the results of each. The test numbers were assigned in the order in which the tests were run. Figure 31 shows a curve of shear load versus shear displace- ment for a representative of each type of crystal tested; the curves show both loading and unloading behavior. The flat portion at the top of three of the plots represents continued deformation at the maximum load for a few minutes before unloading began. All crystals were tested in as-received condition several months after they arrived from Harshaw Chemical Company. The only specimens that exhibited large instantaneous displacements, similar to Portevin-LeChatelier discontinuous slip, were the Type I crystals. Figure 32 presents the acoustic emission behavior observed during Run 7, a typical test of a Type I, easy-glide crystal. This plot shows the number of countable acoustic emission pulses that occurred in each ten second time interval and confirmed coincidences 47 48 Table 1. List of runs and summary of results Test number Summary of results Type I, {110} <110 >, easy-glide crystal 1 Exploratory - no data 2 Exploratory - deformation by uniform single slip 3 Acoustic emission, displacement pulses, and coincidences 4 Acoustic emission, displacement pulses, and coincidences 7 Acoustic emission, displacement pulses, and coincidences 8 Crystal broken while installing in loader 9 Acoustic emission, displacement pulses, and coincidences 10 No emission or displacement pulses above noise - misoriented 13 Limited emission - no displacement pulses above noise - misoriented Type II, {100}<110> crystal 5 No emission above noise - displacement pulse circuit inoperative 12 No emission or disphcement pulses above noise Type III, {110}<1oo> crystal 6 No emission or displacement pulses above noise 14 Some low level emission - no displacement pulses above noise Type IV, {100} <100> cyrstal ll Fracture on cleavage plane Load - pounds 50 .. ‘ 40y 35" 30 an 25 .4. 20 0 150 51b' 49 Ty e III Run 7 /7 Type I ’ / l 1 l 1 L 1 J I ‘l 1 f 0.5 1:0 1T5 2.0 215 3.0 -3 Displacement - 10 inches \ Preload Figure 31. Representative load-displacement curves. 50 .h Gum .H 0&8 .. nomudfluumwp 035m Gommmwgo 033504 .Nm ouamfirm mpsooo m u can. P D ©ODQD D P D P D b D DODOM D D D D D D FDOIOPVD b D D D D I D 'o-om CON 0°“ 0 11.fi.........q........._...4.........+4::.”vyyy..4+qnuupa“:+m moosvpfiofiou mafia unogoumamwp was oflmfloom pvauflnou q— l dfi . q.. —1.--«—.- -1 .—o A: omwos moan— ov :mfi noumonw momfism Id - I a! I q----....- - AC 09.8: moat. om dog .33on momfidm m .21. Z . 22......21212 . Ho omwon mean. A: Gdfi Hoodonm momfism m q fi.—-—-—- W ..«-—¢—-————.———-u‘u . - _o _ ofi mmwon most» a: was 4» fission, momadnm -¢ ¢ «.c—q——dd ¢ «-H« «a— —-—_¢-——-a .— q 4 1_—- o ommm A: memos @053 v was N smegma. momasnm _:__ _.:_A:_1_ 1.:_ 1.. Z. . _. :._ o :n: . ON : my: mv a m hm ¢ lwon «mm mos Lfiov SIBAIQIHI‘; puooes us; at siusAa }o .IaqtnnN 51 between emission and displacement pulses. Low Level emission appeared early in the test, at or before the yield stress. High level emission appeared only at higher stresses, with a lower occurrence rate. The tests that produced coincident acoustic emission and displacement pulses were tests 3, 4, 7, and 9 - all Type I, easy- glide crystals. Again, Run 7 was typical. The confirmed acoustic emission and displacement pulse coincidences shown at the bottom of the acoustic emission plot were those that appeared to occur simultaneously on the record obtained with the light-writing oscillograph. The recording speed used was 1 inch per second, which made resolution of coincident pulses possible to within 0. 05 seconds. Further verification of coincident events was obtained using a dual trace oscillosc0pe. Many pictures taken of the oscilloscope traces show that the coincidences are simultaneous to within a few milliseconds. Figure 33 presents the displacement pulse behavior observed for Run 7. The noise level observed during the pulse height analysis of this run was equivalent to the amplitude of pulses that would correspond to displacements of 2 x 10.7 inches at the transducer sensitivity and amplifier gain used. This displacement is equivalent to the displacement that would be produced by 9 unit dislocations when they pass out of the crystal. Larger displacement pulses observed during Run 7 involve groups of from 200 to 3200 unit dislocations. A complete tabulation of acoustic emission channel gains and displace- ment pulse sensitivities for all 14 tests is included in Appendix B. 52 000 com .N. gm .H REAR. .. non—dawnumfip mafia unmaoomamwa mpaooom 1 05a. cow com com 0“: .3. 223m moonogofioo @399 unmaoomamwp was oflmsoom pognflsou moosopwofioo Adams... mzm ‘1 1'. “l.‘fid‘ . ‘ 030: moan. ow Gaga ~36on osmium some: moan. cm swap Hoummum momgnw — u . _ u soon mocha v :93 Houmouw momaflm _ _ u a . — q a 1- . . vmwon mocha N =93 Houdmum momfifim - - q «d a u a q _ q.-— 1.. :5: ._ 4. sI'exueiu; puooes us; at siueAs ;o .quuan 53 This plot (Figure 33) shows the number of countable displace- ment pulses that occurred in each ten second interval as well as the confirmed pulse coincidences in corresponding intervals. In addition, coincidences confirmed with RMS signal traces are shown. The largest displacement pulses almost always had large acoustic emission pulses in coincidence with them. The pulse distribution plots (Figures 32 and 33) indicated that the small displace- ment pulses make a large contributionto the confirmed coincidences while the low level acoustic emission pulses make very little contribution. In fact, many large acoustic emission pulses were observed with no indication of displacement activity. This is illustrated by Figure 34 which presents a series of oscilloscope pictures showing acoustic emission signals above and displacement pulse signals below. These pictures were obtained using a dual trace amplifier in a storage oscillosc0pe. The signals were conditioned by very restrictive bandpass filters and a demodulator that severely attenuated the recorded signals. Only the largest signals appear in each channel. The numbers at the lower edge of the pictures identify many of the coincident events plotted with the pulse distributions. The three displacement pulses appearing in the third through fifth centimeter of the trace are not coincident with any acoustic emission pulse. There appeared to be a delay on the order of three seconds between three of the four acoustic emission pulses that appear above them and these three displacement pulses. These are the only large displacement pulses that were not coincident with emission pulses in Run 7. This apparent anomaly is explained in a later section. 54 Acoustic emission - 2 volts/cm Displacement - 0. 01 volts/cm —->- Increasing time - seconds 186190 245 249 256 268 274 289 299 358 368 381 391 406 406 436 465 478 481 501 556 Figure 34. Oscilloscope traces of demodulated acoustic emission and displacement pulses - Type 1, Run 7. 55 The Type I, easy-glide crystals were the only specimens to deform by large, instantaneous slip. This behavior is indicated by the horizontal segments and discontinuities in slope at several points on the load-displacement curve presented in Figure 35. Because load was proportional to time, the two vertical scales correspond identically. The early start of low level acoustic emission indicated that plastic deformation started immediately upon loading of Type I crystals. Furthermore, no well defined yield point was observed for any of the four typical Type I tests. Minor differences in initial behavior of the typical Type I tests may be attributed to differences in clamping which, in turn, will cause differing amounts of stress concentration at the boundary between the clamped region and the unrestricted test section and may introduce different numbers of fresh dislocations into the crystals due to indentation. The instantaneous slip events and pulse coincidences shown on the load-displacement curve for Run 7 agree quite well with each other. Figure 36 shows a segment of the light-writing oscillograph recordnthat was obtained during Run 7. The event shown occurred at T = 186 seconds. One second timing marks are at the top. The second trace down is a static reference trace. Then, in order, the displacement pulse, step displacement signal (explained on page 25), and acoustic emission signal appear. The D. C. displacement signal that was recorded .on the X-Y plotter is shown at the bottom along with another static reference trace. The instantaneous displace- ment shown on the D. C. displacement signal at the bottom coincides with the displacement pulse appearing on the step displacement channel. Time - sec onds 56 600.. 50" 45.» 500$ ' 404. 35" 4000 _ ‘b L— . 30v 0! E .— o 3" ad — - 300ib :25" ‘_ I, 'U --o n 8 —o u A — ’— 2018 mmpnfiocH .3. mafia “SoEoomamfip was ofimdoom apogee. inflow pnooomfifia m .0 303.3% * .E 93x WNN n _m_ ONNH m.mH woo 57¢ NN ommm om . Nb ow mew mMNH $.m u vmd .. was mH : I New mmmfi veg” coo «.md N6 3% mg OH om mow $3. $.N who Nw.m o.m a: mJN 2 mm .3. ¢>.m 37m vmd «In file ma 2 mm .4 oo: o o o n ... m .o 0 mm. 0.3» oz: 0N6 vmw wmo o6 mno odfi ma «m 5m Eli oo.m omo mm.: wé mom. 3 0 mm mom :3 23w $.m ww.m mg ovmm N.Nm Om mp om: 1mm omm H m a: 00m H omm H: com .o \caoz \ .NS \cz. 93 .93 z e 951.34 Hes/nu» 3}” > _m_zm\< 524 H 4 30:2: mmouum u > omgm Hmflwwm Madam >fioofio> 5.330.? oswm mcoflmuofimwp £03350 «seasoning £03350 «Soc/m poflmmm fig: >fioo~o> #82 Qsoum Ho Ho m0 m0 mo owwumefiw nmmmb omsnmsuq. .3304 mwdhm>< songs: H.308 QOSMHSQ mpsfinwmz Hongz oQfiH .h csm .opfiwsamdo .H “0&th I mason/0 unopfiofioo HOW moflfioofiocr Soflmoofimwfl .m 3an 81 causing an effective increase in the value of m in the relation v = n/8m. The value of Fi- is too small to be representative of the group size that makes the main contribution to the displacement and acoustic emission. The displacement signals indicate that the displacement proceeds by the motion of a few, possibly one or two, large groups and many small ones, rather than the movement of groups uniformly distributed about the mean value. The large groups will have the largest effect; therefore, the proper value for m to determine the group velocity lies between 53' and A/ 'th . That the leading portion of the displacement pulses have amplitudes much larger that the noise level equivalent, and the average group size ranges from 4 to 22 dislocations, which is on the same order of magnitude as the smallest detectable pulse, supports the above statements. Appendix B, listing the acoustic channel gains and displacement pulse sensitivities, shows that the smallest countable displacement pulse is related to the egress of at least nine unit dislocations for Run 7. If several groups move at the same time, or if the predominant group size is actually larger, the dislocation velocity will approach the lower limit. The velocities calculated above for instantaneous slip events are based on the assumption that the only mechanism operating is the breakaway of piled-up dislocation groups. The proposed model allows for emission accompanying crack formation and propagation, as well as collision processes. There is no reason for excluding these additional sources and they probably act simultaneously with the break- away process. If either one makes a contribution to the acoustic emission signal used to determine the number of acoustic 82 emission events occuring during the step displacement, the effect is to cause a high value of N (the number of spikes recorded in the emission signal). A reduction in N reduces the average velocity, given by V = 2N/8T, (6) though it has no effect on the lower limit velocity. Figure 46, especially, suggests the possibility of simultaneous slip and crack formation. The initial acoustic emission may be associated with the formation and propagation of a crack, with the remaining burst being due to the breakaway process. The collision process is equally probable but its occurrence is unlikely without prior breakaway emission. Though the process of dislocation pile -up can be considered as a collision process, it will progress slowly, with only one dislocation at a time coming to rest at an equilibrium position. Once a group collects at an obstacle, the subsequent release of a nearly coherent group will produce conditions conducive to collision emission. Any obstacle in the slip plane of the moving group is expected to interact with the group to impede its prOgress and produce collision emission which, in turn, will cause a high spike count and a correspondingly high estimate of the average dislocation velocity. In all, three types of acoustic emission are observed: the two types of high level emission pulses just discussed, associated with crack formation and breakaway of large dislocation groups that reach average velocities of about 500 to 600 inches per second, and low level 83 Emission 2 v/cm 5 milliseconds/cm Displacement 0. 2 v/cm Emission 0. 5 v/cm 5 millis ec onds/ cm Displacement 0. 01 v/cm Figure 46. Oscilloscope traces - event at T = 436 seconds - Type I, Run 7. 84 acoustic emission with no coincident displacement pulses. Smaller displacement pulses also are observed with no apparent acoustic emission (see Figure 33, Run 7). Three of the largest non- coincident displacement pulses are just detectable in Figure 34 between T = 190 and T = 245 seconds. Such non-coincident acoustic emission and displacement pulses are expected, according to the model of the emission process, when small groups of dislocations piled-up behind weak obstacles are finally forced to break away and move through the crystal. Since these groups must be much smaller than the smallest group associated with instantaneous slip, the driving force on the leading dislocations is smaller and lower velocities result. The combined effect of low velocity and unknown position of origin produces much longer and variable delays between the acoustic emission associated with the breakaway and the corresponding displace- ment pulse, making it difficult to determine correlations between them except on a statistical basis. A close look at the acoustic emission and displacement signals between T = 190 and 245 seconds (Figure 34) shows that there are emission pulses preceding each isolated displace- ment pulse by about three seconds. If these emission pulses are associated correctly with the delayed displacement pulses (the absence of other activity in the region makes this likely), and if the group of dislocations is assumed to have traveled the entire width of the crystal, the related dislocation velocity becomes v = 1/8x3 or about 0. 04 inches per second. 85 As a result of the low velocity, secondary emission by the collision process is improbable. For such small groups of dislocations, the terminal velocity of the group is probably similar to the bulk of the dislocations causing plastic flow. This is especially true for these three delayed events because the load angle is still small and, as discussed earlier, the ability of the loader to maintain the dislocation velocity is limited. The low level emission without coincident displace- ments therefore seems quite reasonable since the delay between the emission and corresponding displacement is variable; and, furthermore, many of these breakaway groups, being small and moving with low velocities, will become pinned again by other obstacles before they can leave the crystal. With a low velocity the individual dislocations of the group are notas apttncontinue moving coherently through the crysta1,which increases the possibility that they pile -up on other obstacles in their slip planes without producing collision emission and without resulting displacement pulses. The general appearance of the load displacement curves for all runs indicates that the bulk of the plastic deformation is smooth. This suggests that the deformation results from random motion of small groups of less than 9 dislocations (Run 7). An estimate of the upper limit dislocation velocity in this case is obtained from the load-displacement curve from Run 7, typical of the easy-glide tests. The average loading slope, away from the regions of instantaneous slip, is very nearly 248 x 103 seconds per inch of displacement. The unloading slope is 1090 x 103 seconds per inch of displacement. The corresponding relative velocities (velocity 86 of the top part of the test crystal relative to the bottom part) are 4. 03 x10.6 inches per second and 0. 92 x10.6 inches per second. The difference between these velocities represents the relative velocity due to plastic flow, since the unloading process is essentially elastic. This plastic flow velocity is 3.11 x 10-6 inches per second and may be attributed to the motion of groups of less than nine unit dislocations, because few detectable displacement pulses are observed in these regions of the load-displacement curve. The plastic flow velocity divided by the length of the unit Burgers vector for an easy-slip dislocation (22. 3 x 10-9 inches) indicates the number of dislocations that must leave the crystal in one second to accommodate the velocity: about 140 unit dislocations per second for Run 7. Let v = d/t inches per second represent the dislocation velocity, where t represents the time in seconds required for an edge dislocation to travel entirely through the crystal, a distance of 1/8 inch. If n represents the rate with which dislocations leave the crystal (dislocations per second) and m represents the number of dislocations moving at any given time, t = m/n seconds and v becomes n/8m inches per second. For any given value of n, the upper limit velocity occurs when m = 1 which represents the case when only one dislocation is moving at any given time. The upper limit dislocation velocity in the case of the smooth plastic flow observed during Run 7 (n = 140 per second) is, therefore, 17. 5 inches per second. This approximation, as before, ignores any finite accelerations the dislocations expe ri enc e . 87 In this case, no estimate of a lower limit velocity is made, because no measurement to determine m is possible. However, it is probable that many small groups move simultaneously and independently of each other. For example, if 100 groups of 5 move together, m would become 500 and the resulting dislocation velocity is v = n/8m : O. 035 inches per second. This is consistent with the velocity related to the identifiable delayed events discussed earlier. 25’ 31 indicate Reports on dislocation velocities in lithium fluoride that, for any given degree of hardness, dislocation velocity increases with applied stress over an extremely large range. For example, dislocation velocities may range from 2 x 10-5 to 2 x 10"1 inches per second as the applied shear stress varies from 900 to 1400 psi. The dependence appears to be linear in a log log plot. Dislocation velocities as high as 2 x 105 inches per second are reached with shear stress impulses of 40, 000 psi. The dislocation velocity appears to approach the velocity of (110)[110] shear waves asymptotically. Tests on a single crystal do not reflect the effects of work hardening. How- ever, additional tests on softer and harder crystals indicate that hardening causes a constant increase in the stress for all velocities up to about 0. 4 inches per second. Figure 48 qualitatively represents the above results. The upper limit dislocation velocity of 17. 5 inches per second for the smooth plastic flow of Run 7 appears to be high in comparison with the above data. The large number of dislocations available to accommodate the deformation, the velocity for the three delayed 88 displacement pulses, and the more reasonable calculation that considers the motion of many small groups, are in essential agreement with the above results. The dislocation velocities for the instantaneous slip events also appear to be high. Neither the lower limit nor the upper limit shows any regular stress dependence and the lower limit is higher than would normally be expected. The average dislocation velocity for these events appears to be constant for all stress levels. There are several reasons for these apparent discrepancies. The results quotedzs’ 31 are from measurements involving single dislocations moving in essentially virgin crystal, while the dislocation velocities for the instantaneous slip events reported here are for large groups of dislocations initially piled-up at obstacles. The effect of the pile-up is to magnify the driving stress on the leading dislocation by approximately the number of dislocations piled up behind it, and the effective driving stress is much higher than the applied stress appearing in Gilman and Johnston's reports. Consequently, generally higher velocities are to be expected. This is illustrated in Figure 47 by the two points on line c which indicate the expected velocity, va, based on the applied shear stress, Ta’ and the velocity, vC, resulting from the actual driving stress, TC. Furthermore, if the loader supplies energy to the system as fast or faster than drag forces on dislocations dissipate it, the dislocations can be expected to maintain their velocity, or even continue to accelerate. This feature of the loader is described earlier. Lack of a definite relation between stress and velocity is partially explained by the actual value of the driving stress being unknown. An- 2:: 10 ‘r _§hear wave velocity 10 a. 10 .- 10 10 ‘r' 10 ‘P 10.1". 10'21' Fast load rate Slow load rate Dislocation velocity - 4 in/sec y—a o i 100 11.000 10,:000 Ioofooo 'T 'T a C Io Applied shear stress - psi» Plot is intended to represent qualitative features only. Figure 47. Dislocation velocity versus applied shear stress. 90 other factor is the extremely low load rate and essentially unrestricted deformation of the present test. For example, the total plastic deformation for Run 7 is 2. 9 x 10.3 inches at a load rate of five pounds per minute (approximately 100% shear strain at a stress rate of 160 psi per minute). Such plastic deformation must cause a great amount of work hardening which reduces the stress dependence of dislocation velocity. This effect is shown by the dashed lines super- posed on Figure 47. Under these conditions, the higher velocities seem not unreasonable. No direct velocity comparisons are possible, since the relative purities, as-grown dislocation densities, and hard- ness of the crystals are not known. Signal amplitude comparisons between acoustic emission and displacement signals, both pulse and RMS, give no definite correlations. Large displacements canbe expected to have large acoustic emission pulses associated with them, but large acoustic emissions due to crack formation are not accompanied by displacement pulses. Further- more, the collision emission process that is very likely in conjunction with breakaway emission during instantaneous slip events adds an amount to the emission signal that depends more upon the number and type of collisions than upon the actual magnitude of the displacement. However, lack of pulse height correlation is due primarily to the irregularity of the acoustic emission pulses,because the RMS displace- ment signals are found to be related to the instantaneous slip displace- ments in the following way: (RMS amplitude)l/2 = KA . (7) 91 Measurement of such small changes in displacement presents several difficulties. The displacement sensitivity of the basic transducer is somewhat selectable, ranging between 900 and 2500 volts per inch, depending upon the initial settings of the displacement capacitor spacing. These sensitivities are achieved with the sacrifice of linearity, though the range of displacement is small and resulting changes in sensitivity are either negligible or can be corrected by use of calibration data. The biggest difficulty is to separate the small, rapid changes in displacement from the D. C. component of the signal that represents the total displacement. Three problems appear here. They are signal- to-noise ratio, frequency response of the transducer, and true repro- duction of the rapid step displacements. A basic sensitivity of 1000 volts per inch produces a change of 22.3 microvolts for the egress of one unit dislocation. Therefore, a noise level of 0. 4 millivolts peak- to-peak, such as is typical for this experiment, effectively obliterates these low level signals and, in Run 7, limits resolution of dynamic events to those involving more than 10 unit dislocations. The moving element of the displacement transducer is a relatively massive plate mounted at the end of a plastic beam. Its low natural frequency, about 400 Hz, limits its ability to follow rapid step displacements and smooths out details of the events. In the present experiment the dynamic sensitivity can be inferred only from the quasi-static sensitivity measured after each run and the gains and attenuations in the signal conditioning and playback equipment. This information is listed in Appendix B. 92 The third problem is evident in the relationship between displace- ment pulse RMS amplitude and measured displacements (equation 7), (RMS amplitude)l/Z 2 KA , which is believed to be partly the result of inductive coupling in the displacement pulse channel and partly the result of the mechanical response of the transducer. Since dynamic displacements are of primary interest, long term stability of the transducer is not important as long as it does not change appreciably during each run. Direct calibrations of the transducer taken several days apart indicated that the long term stability was sufficient for the present experiment. Additional precautions in construction and control of environment will be necessary if a transducer of this type is to be used to measure such small displacements in a creep-type experiment. It also should be mentioned in conclusion that some of the control specimens also give evidence of low level acoustic emission. In view of the additional load components (tension and flexure) applied to the crystals, the presence of low level acoustic emission from these control specimens is to be expected. The lack of displacement pulses during these control tests is due to the restrictive load geometry which makes easy-glide difficult unless the easy-glide plane is also the load plane. C ONC LUSIONS Dynamic displacements with magnitudes on the order of 2 x 10-7 inches and greater can be detected with the equipment and techniques reported here. More exact measurement of such small dynamic displacements and resolution of displacements on the order of 2 x 10-8 inches and smaller are believed practical with improvements suggested in the next section. Acoustic emission is due to lattice vibrations set up when excess elastic strain energy, stored in the volume around a blocking obstacle. as a result of dislocation pile -up on the obstacle, is released or allowed to relax rapidly. In lithium fluoride the method of release may be the breakaway of the dislocation group or the formation of a crack. Secondary emission can be produced as a result of collision processes between coherent moving groups of dislocations or high velocity single dislocations and blocking obstacles. The former is most likely, since single dislocations would rarely attain sufficient velocity to cause an emission before they become pinned by obstacles in their glide plane. The pinning process that causes groups of dislocations to collect is instrumental to the secondary emission process: it is the resulting stress concentration which causes group velocities sufficient to produce secondary collision emission. The acoustic emission behavior observed for the easy glide specimens can be qualitatively characterized by using the proposed model. The easy-glide crystals exhibit plastic behavior immediately upon loading. This is evidenced by the immediate, rapid occurrence of low level 93 94 acoustic emission from the breakaway of small groups of less than nine dislocations, or from the activation of Frank-R ead sources in the crystal. Displacement pulses are expected and observed with no particular value of time delay between emission and displacement pulses, because groups large enough to produce a displacement pulse are moving with low velocity, and the time delay is dependent upon the location of the blocking obstacle in the particular slip plane. Furthermore, due to the low velocity, many moving groups may be blocked again by other obstacles, producing collision emission without accompanying displacement pulses. After slip bands entirely fill the test section, line defect trails resulting from cross-glide dislocation multiplication provide increasing numbers of strong blocking obstacles for easy-glide, and cause the collection of larger piled-up dislocation groups. No high- level emission is expected before slip bands are completely developed; since before this point is reached, there is virgin crystal in which dislocation multiplication and easy-glide can take place. The stress concentration at the leading edge of such a group is sufficient to produce breakaway emission and accelerate much of the group to high velocity, creating an instantaneous slip event with coincident displacement pulses. Due to the high velocity attained and the increased number of blocking obstacles in the slip plane, collision emission is expected to be common under breakaway conditions. The additional mechanism of crack nucleation and propagation adds to the signals observed during the instantaneous slip events and also makes its own contribution, as evidenced by the presence of large emission pulses with no associated 95 displacement pulses - either coincident or delayed. Low level emission and non-coincident displacement pulses, related to the breakaway of small groups with low velocities, continue throughout the test, producing the bulk of the smooth plastic deformation. The data and the associated emission process model suggested above are in complete agreement with independent conclusions by Schofield15 and Tatroz'7 that the emission process is due to an internal, volumetric, rather than an external, surface mechanism. The model also agrees with Schofield's results indicating that high frequency emission requires a strain rate above a certain minimum value. The low level emission observed in this experiment corresponds to the high frequency emission observed by Schofield. Since the emission mechanism pr0posed here is essentially a dynamic mechanism, and the rate of dislocation multiplication and pile -up, as well as collision processes, are quite strain-rate dependent, low velocity dislocations associated with low strain rates would not produce detectable emission pulses at the noise levels presently achieved. It is possible to relate the burst-type emissions observed by Schofield in some materials to the breakaway process apparent in lithium fluoride. As in lithium fluoride, which shows burst-type emis- sions due both to cracking and dislocation breakaway, other materials have been shown to produce emission by other mechanisms, such as micro-twiming in zinc and stacking fault formation in gold. Such emission is to be expected, since these processes are characterized by a rapid change in structure, relieving local elastic strain energy concentrations. 96 However, the emission identified as low level in the present experi- ment, and high frequency in the work of others, is most probably due to the internal breakaway and collision processes proposed here. Each material may be expected to exhibit burst-type emission dependent upon crystallographic structure. The model satisfactorily explains the early results of other experimenters,notably Liptail 8, who related emission to surface effects. The emission he observed in aluminum with thick anodized and reacted coatings was no different than that proposed in this case. The effect of the coatings was to cause a preponderance of dislocation .r pile-ups at the interface between the aluminum and the coating layer; s and since the resulting release of dislocation groups was concentrated near the surface, other internal sources would be masked. The early experiments by Schofield and Tatro on aluminum indicated a surface mechanism because the oxide coating, normally present on aluminum, enhances the breakaway contribution at the surface. The same would be true of any surface treatment that tends to harden a surface layer of the specimen. SUGGESTIONS FOR FURTHER RESEARCH The conclusion that low level emission is associated with discrete displacements needs to be reinforced with additional data amenable to statistical analysis. In order to do this with a series of experiments similar to the present one, it will be necessary to improve signal-to- noise ratios in both the emission and displacement-pulse channels, and to improve the upper frequency response of the displacement transducer. The region where strain is less than 1. 5% is of primary interest, since this is where the basic breakaway process predominates: crack formation is absent because stress concentrations are small, and secondary collision emission is less apt to play an important role be- cause velocities are low. Pulse height correlation will be much more meaningful in this region with other mechanisms suppressed. With good correlations between acoustic emission pulse heights and displacement pulses, reasonable measures of the energy involved in the basic breakaway emission process should be made. Theoretical consideration of the additional energy stored in the crystal lattice due to the interaction of piled-up groups and blocking obstacles will be necessary to allow a better understanding of the breakaway emission process. Secondary collision emission in the region between 1. 5% strain and 10% strain or higher should be interesting from the viewpoint of studying the phenomenom of work hardening in crystalline materials. The present loader is not capable of controlled strain rates or of load rates higher than five pomds per minute. Further research on the relations between dislocation velocity, stress, and work hardening 97 98 should be pursued with a programmable loader capable of a wide range of load and strain rates. Continued improvement of signal-to-noise ratio of the displace- ment transducer will improve the dynamic displacement resolution and allow detection of single dislocation egress. A reduction in noise by a factor of ten will be sufficient, and is believed practical. Better load geometry will allow redesign of the moving element of the displacement transducer to allow stiffer supports and a less massive moving plate. The resulting shift to higher natural frequency will allow detection of more displacement details. Circuitry should be designed to allow more direct measurement of the rapid step displacements. One possibility that should be investi- gated is formation of a second displacement signal with high frequency components filtered out. A difference amplifier could then be used to compare the direct signal with the smoothed one and pick off rapid displacement changes without using capacitative or transformer c oupling. gr- 10. ll. 12. l3. l4. BIBLIOGRAPHY Orowan, E., and Becker, R., Zeit. Physik, Volume 79, (1932). Klassen-Nekludowa, M., Zeit. Physik, Volume 55, (1929). Crussuard, C., Leon, J. B., Plateau, J., and Blacket, C., "Sur la formation d'ondes sonores, au cours d'essai de traction, dans des eprouvettes metalliques", Comptes rendus des seances g3 l'Acadamie des Sciences, Volume 246, seance du 19 mai, (1958). Blewitt, T. H., Coltman, R. R., and Redman, J. K., "Low Temperature Deformation of Copper Single Crystals", Journal of Applied Physics, Volume 28, Number 6, June, (1957). Schmid, E., and Valanck, M. A., Zeit. Physik, Volume 75, (1932). Kaiser, J. , ”Untersuchungen uber das Auftreten von Gerauschen _ beim Zugversuch", Doctoral thesis, Munich Technische Hochschule, t (1950). . H H Kaiser, J. , Arkiv Fur das Eisenhuttenwessen, Volume 24, pg. 43- 44, (1953). Tatro, C. A. , "Sonic Techniques in the Detection of Crystal Slip in Metals", Engineering Research, Volume 1, Progress Report, The Engineering Experiment Station, College of Engineering, Michigan State University, (1957). Tatro, C. A. , "Sonic Techniques in the Detection of Crystal Slip in Metals", Division of Engineering Research, Progress Report, Michigan State University, January, (1959). Tatro, C. A. , "Acoustic Emission from Crystalline Materials Subjected to External Loads", Division of Engineering Research, Progress Report, Michigan State University, April, (1960). Tatro, C. A., and Liptai, R. G., "Acoustic Emission from Crystalline Substances", Paper presented at the Symposium on Physics and Nondestructive Testing, Southwest Research Institute, October, (1962). Schofield, B. H., Bareiss, R. A., and Kyrala, A. A., "Acoustic Emission Under Applied Stress", WADC Technical Report 58-194, April 30, (1958). Schofield, B. H. , "Acoustic Emission Under Applied Stress", ARL 150, December, (1961). Schofield, B. H. , "Acoustic Emission Under Applied Stress", ASD-TDR-53-509, Part 1, April, (1963). 99 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 100 Schofield, B. H. , “Acoustic Emission Under Applied Stress", ASD-TDR-63-509, Part 11, May, (1964). Schofield, B. H. , "Investigation of Applicability of Acoustic Emission", AFML-TR -65-106, May, (1965). Shoemaker, P. 5., "Acoustic Emission, An Experimental Method", M. S. thesis, Michigan State University, (1961). Liptai, R. G. , "An Investigation of the Acoustic Emission Phenomenon", Ph. D. thesis, Michigan State University, (1963). Liptai, R. G. , and Tatro, C. A. , "Acoustic Emission - A Surface Phenomenon", Paper presented at the Symposium on Nondestructive Testing of Aircraft and Missile Components, Southwest Research Institute, February, (1963). Sedgwick, R. T. , "An Investigation of Acoustic Emission from Coated and Uncoated Ionic Crystals", Ph. D. thesis, Michigan State University, (1965). Taylor, G. 1., Proceedings gfthe Royal Society, A, Volume 145, pg. 362, (1934). Orowan, E., Zeit. Physik, Volume 89, pg. 605, 614, 634, (1934). Polanyi, M., Zeit. Physik, Volume 89, pg. 660, (1934). Gilman, J. J., and Johnston, W. G., "Dislocations, Point-Defect Clusters, and Cavities in Neutron Irradiated LiF Crystals", Journal pprplied Physics, Volume 29, Number 6, June, (1958). Johnston, W. G., and Gilman, J. J., "Dislocation Velocities, Dislocation Densities, and Plastic Flow in Lithium Fluoride Crystals", Journal pf Applied Physics, Volume 30, Number 2, February, (1959). Gilman,J. J. , “Plastic Anisotropy of LiF and Other Rocksalt- Type Crystals", Acta Metallurgipa, Volume 7, September, (1959). Gilman, J. J. , “Dislocation Sources in Crystals", Journal o_f Applied Physics, Volume 30, Number 10, October, (1959). Gilman, J. J. , and Johnston, W. G. , "Behavior of Individual Dislocations in Strain-hardened LiF Crystals", Journal o_f Applied Physics, Volume 31, Number 4, April, (1960). Johnston, W. G. , and Gilman, J. J. , "Dislocation Multiplication in Lithium Fluoride Crystals", Journal pf Apglied Physics, Volume 31, Number 4, April, (1960). w’uamw. 30. 31. 32. 33. 34. 35. 36. 37. 101 Gilman, J. J. , "Direct Measurements of the Surface Energies of Crystals", Journal gfgpplied Physics, Volume 31, Number 12, December, (1960). Gilman, J. J., and Johnston, W. G. , "Dislocations in Lithium Fluoride Crystals", Solid State Physics, Volume 13, Academic Press, (1962). Sproull, R. L. , "Charged Dislocations in Lithium Fluoride", Phil. Mag., Volume 55, Number 56, August, (1960). Lion, K. S., Instrumentationi_n_Scientific Research, pg. 86, MCGraW-Hill, 1959. Lion, K. S., “Mechanic-Electric Transducer", The Reviewpf Scientific Instruments, Volume 27, Number 4, April, (1956). Koehler, J. S. , ”The Nature of Work-Hardening", Physical Review, Volume 86, Number 1, January, (1952). Orowan, E., in Dislocations i_nMetals, pg. 103, American Institute of Mining and Metallurgical Engineers, New York, (1954). Tatro, C. A., private communication. 100 ll. 12. 13. Appendix A. Equipment List It em Decker 02B-l -4D general purpose probe. Decker 904 delta unit. Tektronix type 122 low-level preamplifier. Tektronix type 532 oscillo- sc0pe with type 53/ 54D plug in unit. Variable electronic filters; Spencer Kennedy model 302, Spencer Kennedy model 308, Krohn-Hite model 335R. Ampex FR-l 100 tape recorder- reproducer. Sorensen ACR 2000 voltage regulator. CEC type 5-119 recording oscillograph. Galvanometer drivers: CEC type 1-162A galvanometer- driver amplifier, Brush Instruments universal amplifier model RD-5612-00. Baldwin SNB3 -06 ~1286 semiconductor strain gages. Ellis BAM-l bridge amplifier. Varian model F-80 X-Y recorder. Tektronix type 180 A time mark generator. 102 Used For Sensor for differential capacitor (displacement). Sens or power supply. Preamplifier for emission and displacement pulse signals. Amplifier and monitor for emission and displacement signals. Gate signal generator for pulse height analysis and time counter. Filters for recording and readout. Data storage. Line voltage regulator and filter to remove fast transients. Power for tape recorder. Multi-channel data recording. Galvanometer drivers for 5-119 oscillograph. Load bridge. Load bridge amplifier. L oad—displac ement plot. Time reference signal. 14. 15. l6. 17. 18. 19. 20. 21. 22. 103 Nuclear-Chicago model 151A electronic counter. Ballantine Laboratories, Inc. model 320 true root-mean- square voltmeter. Hewlett Packard model 522B electronic counter. Tektronix type 130 L-C meter. Tektronix type 564 storage oscilloscope with 3A1 dual trace amplifier and 2B67 time base plug in units. Type C-12 oscilloscope camera. Hewlett Packard 200CD oscillator, Arenberg 93 ohm attenuator, Hewlett Packard 400A RMS voltmeter, and Tektronix type 551 dual beam oscilloscope with 53/54D and 53 E/ 54E plug in units. Cohu 510 digital voltmeter. Harrison Lab 6204A D. C. power supply. ADP crystal stack. Time reference c ounter. Provide RMS signal for readout. Counter for pulse height analysis. Monitor for measuring initial capacitance of differential capacitor. C oincident puls e picture 3 . Calibration and setup signal and monitor. Meter to set initial balance of displacement transducer cathode followers. Zero suppression for displace- ment signal recorded on 5-119 oscillograph. Acoustic emission transducer. Appendix B Acoustic channel gains and displacement pulse sensitivities. Run Acoustic gain Displacement sensitivity Displacement Number True Relative True Relative Noise level 4 to Run 7 6 to Run 7 leoquiivalelfit 10 10 volt/in I ** {110}<110> Type I 30420,,< 2.42 1.11 9 43.7 3120-230 2.42 1.11 18 87.4 3230-420 2.42 1.11 45 218 2.78 12.5 34‘20 on 1.21 0.55 45 218 4 1.21 0.55 22.1 107 0.184 0.84 7 2.2 1 0.206 1 1.95 8.75 9 0.11 0.05 0.129 0.63 1.94 8.7 10 0.55 0.25 0.059 0.29 10.9 48.9 13 0.55 0.25 0.396 1.93 3.15 14.1 {100} <110> Type II 5 0.55 0.25 0.758 *** 12 0.55 0.25 0.595 2.88 2.1 9.4 {110} <1oo> Type 111 6 2.2 1 0.218 1.06 10.9 48.9 14 0.22 0.1 0.396 1.93 1.57 7.04 {100} <1oo> Type IV 11 0.22 0.1 0.396 1.93 * Subscripts indicate time interval during run if not constant for entire run. ** Number of unit dislocations equivalent to smallest countable pulse. 1|=6| 22. 3x10' 9inches. *** Displacement pulse data not recorded due to use of 2 Khz high-pass filter. 104 Appendix C. Load-displacement data. Run 3, Type I Run 4, Type I 105 Loading , Loading Displacement Load Displacement Load 10'3 in 1b ; 10"3 in lb -0.073 -6.66 5 -0.457 -16.79 -0.028 -3.06 1 -0.420 -14.77 0 0 ; -O.289 -9.51 +0.050 +3.89 ! -0.149 -2.22 0.105 7.94 +0.038 +2.63 0.194 11.99 i +0.029 +1.82 0.250 13.84 g 0 0 0.316 15.49 +0.047 +1.82 0.361 17.14 i 0.159 3.44 0.416 18.74 i 0.252 4.25 0.488 20.74 ‘ 0.457 5.47 0.538 20.99 0.625 6.28 0.572 22.29 1.110 8.30 0.616 23.94 : 1.465 9.92 0.716 26.19 : 1.539 10.32 0.733 26.49 i 2.062 15.18 0.783 26.64 I 2.345 18.08 0.883 28.29 ‘ 2.435 18.08 0.994 30.14 , 2.509 18.48 1.050 30.94 ' 2.677 18.48 1.094 31.59 1 2.789 20.78 1.194 33.14 ‘ 3.218 23.28 1.272 34.19 3.536 26.78 1.338 35.34 3.703 27.68 1.438 37.09 4.207 29.98 1.516 38.44 4.748 34.08 1.627 39.44 5.047 35.53 1.694 40.34 5.271 37.08 Unloading Unloading 0 0 9 0 0 0.042 2.45 ' 0.15 2.23 0.109 6.42 0.45 9.51 0.146 8.75 0.63 15.58 0.220 13.56 0.86 23.80 0.298 19.05 1.01 29.05 0.387 26.00 1.12 34.15 0.476 32.20 1.16 35.55 0.598 42.80 106 Run 5, Type 11 Run 6 continued Displacement Load Displacement Load 10'31n Lb 10'31n Lb Loading Unloading -0.167 -13.29 0 0 -0.145 -6.28 0.087 3.51 -0.102 -l.61 0.188 8.77 -0.058 +0.68 0.218 11.28 -0.044 +1.17 0.239 12.03 -0.015 +0.23 . 0.259 13.53 0 0 1 0.299 16.29 +0.015 +0.44 1 0.360 20.80 0.044 2.19 ‘ 0.441 27.56 0.174 4.82 0.542 35.58 0.319 7.74 0.603 40.09 0.464 10.66 0.643 43.35 0.595 13.87 l O. 682 16.79 1 Run 7, Type I 0. 798 20. 59 1 Displacement Load 0.900 23.51 -3. 0.987 26.41 1 10 1“ ‘Lb 1.103 28.81 0 0 1.205 31.26 0.045 1.30 1.393 35.41 0.125 3.64 1.466 37.11 . 0.250 5.78 1.582 39.71 . 0.348 6.85 1.756 41.81 g 0.530 8.14 0.635 8.83 Unloading ’ 0. 745 9. 68 -3. 0.930 11.45 Mean 310pe 71.23 lb/lO 1n 9 1.110 13.34 Run 6, ije 111 . 1.285 15.37 Displacement Load ' 1'330 15'37 _3 1.370 15.70 10 in Lb 1.410 16.41 . 1.552 19.71 LOadmg 1. 630 20. 77 0 0 1.730 22.19 0.071 1.50 1.865 23.96 0.324 4.51 1.880 24.29 0.607 9.52 1.936 25.30 0.688 11.53 2.045 26.08 0.768 13.78 2.175 28.21 0.870 16.79 2.265 30.10 1.011 21.80 . 2.330 31.68 . 1.112 25.56 ; 2.370 31.98 1.213 29.57 5 2.545 35.29 1.314 33.33 i 2.585 36.23 1.395 36.58 I 2.602 36.23 1.517 41.10 g 2.690 38.00 1.618 44.35 , 2.730 38.62 1.674 46.11 2.765 38.66 107 Run 7 continued Run 9 continued Displacement Load Displacement Load 10'31n Lb i 10'31n Lb 2. 810 39.63 I Unloading 2. 888 39.63 i 0 0 2.950 40.72 g 0.16 2.54 3.055 43.32 g 0.40 7.50 3.120 44.59 ‘ 0.57 11.95 3.180 45.44 0.68 15.34 0. 80 19.13 Unloading 0. 96 25. 80 0 0 1.11 33.64 0.026 1.30 1'20 36'91 1.25 39.39 0.111 6.49 . ° ' 0.175 11.09 Run10, TypeI 0. 240 1 5. 93 1 Displacement Load 0.282 19.71 -3 . 0.335 24.20 Loaggl 1“ Lb 0.410 31.04 ' g 0.452 34. 82 0 0 0.516 40.48 0.076 6.94 0.559 44.50 0.177 14.69 0.278 20. 29 Run 9, Type I 0. 369 23. 23 Displacement Load 0. 480 26. 97 -3 . 0. 581 30.71 1° 1“ Lb 0.682 34.44 0 0 0.783 38.72 0.034 1.49 0.884 42.72 0.116 3.06 , 0.955 45.39 . 2 . ' . g. :42 8129 ‘ Unloading 0.776 10.90 i 0 0 0.973 13.51 1 0.051 2.40 1.181 16.39 ‘ 0.152 9.21 1.357 18.58 0.201 12.68 1.425 18.69 0.354 25.37 1.506 19.53 0.480 39.38 1.656 21.62 0.495 45.39 1.761 22.66 1. 981 25. 28 ; Run 11, Type IV 2.268 28.67 1 Displacement Load 2.448 29.59 g -3. 2.828 29.59 1 10 1“ Lb 2. 888 30.37 ; 0 0 3.018 31.81 0.050 4.03 3.258 34.16 0.101 7.66 4.048 40.44 0.151 11.18 4.268 41.87 0.202 14.29 0. 252 17.30 0.303 19.88 108 Run 11 continued Run 13 continued Displacement Load Displacement Load 10'31n Lb ‘ 10'31n Lb 0.353 22.11 0. 569 22.69 0.404 23. 96 g 0.700 26.73 0.454 25.36 1 0.761 28.62 0. 505 26.33 “ 0. 811 30.23 0.565 27.14 0.872 32.39 0.656 27. 94 1.124 41.27 0.777 28. 48 1.185 43.37 0.787 28. 50 Fracture Unloading , O 0 12, , Dfiplacegii 11 Load 0' 045 3' 85 _3 0.121 11.93 10 in Lb 0.212 23.23 Loading 0.313 34. 81 0.387 43.37 0 0 0. 035 3. 82 Run 14, Type 111 0.166 13.78 Displacement Load 0.207 16.74 -3 . 0.278 21.59 10 1” ‘Lb 0.338 25.63 Loading 0.378 28.32 0.409 30.21 0 0 0.459 33.17 0.030 2.68 0.590 40.98 0.091 5.95 0.656 44.34 0.162 9.12 . 0.384 17.70 Unloading 0. 424 19. 04 0 0 0. 464 20. 43 0.052 3.55 0.585 23.46 0.133 9.48 0.626 23.92 0.229 16.21 0.686 24. 53 0.396 29.94 0.787 26.14 0.466 36.40 0.898 28.15 0. 517 40.71 0. 979 29.76 0. 560 44.34 1.110 32.44 1.200 34.32 Run 13, Type I 1.342 37. 27 Displacement Load 1. 382 3 8. 07 10-3 in Lb 1.398 38.34 Loading ; Unloading 0 0 0 0 0.105 10.85 0'010 0'27 0.135 15.70 0°378 9'65 0.226 17.31 0. 550 14.21 0 266 17 58 0.651 17.96 ' ' .792 24.13 0.327 18.12 0 0 367 18 52 0.974 34.05 ' ' 1. 034 38 34 0. 509 20. 95 ° Appendix D - Pulse height data. Run 3, Type I Key to columns Column 1 - ten second time interval ending at time indicated in seconds. Column 2 - displacement pulses in ten second time interval with pulse height between 0. 5 and 1. 0 volts. Column 3 - displacement pulses with pulse heights between 0. 3 and 0. 5 volts. Column 4 - displacement pulses with pulse heights between 0. 25 and O. 3 volts. Column 5 - total of columns 2, 3, and 4. Column 6 - emission pulses with pulse heights between 5 and 10 volts. Column 7 - emission pulses with pulse heights between 2. 5 and 5 volts. Column 8 - emission pulses with pulse heights between 1 and 2. 5 volts. Column 9 - emission pulses with pulse heights between 0. 5 and 1 volt. Column 10 - emission pulses with pulse heights between 0. 25 and 0. 5 volts. Column 11 - total of columns 6 through 10. Column 12 - positively identified RMS coincidences. Column 13 - RMS coincidences that were not enough above background to be considered for column 12. Column 14 - confirmed coincidences between displacement and acoustic emission pulses. Time Displacement Acoustic emission Coincidences Interval pulses pulses RMS Pulse 1 2 3 4 5 6 7 8 9 10 11 12 13 14 5 15 25 35 45 1 55 65 75 19 19 85 2 2. 95 105 2 115 Z 2 125 l 135 1 1 145 2 155 2 165 1 1 175 2 185 1 195 2 2 205 1 215 2 pa $HH hag—a I—‘D—‘ND—l .pquwmmv— N F‘N NHNNNrFCD 109 Run 3 continued Time Interval 1 225 235 245 255 265 275 285 295 305 315 325 335 345 355 365 375 385 395 405 415 425 435 445 455 465 475 485 495 505 515 525 535 545 555 565 575 Displacement pulses 2 3 4 5 9 10 19 10 10 1 8 9 12 12 15 15 5 6 2 ll 15 1 1 2 1 1 Nm¢¢mNHmm¢¢©deH n “83.03.3020 0015 0G0E00mamwfl .mv. 0nsmwh 330000 .. 05TH. coo com oov oom CON 2 on: o 000:0vw0fi00 00:5 E0800030€ .93 0305000 U0EHSG00 I . . . _ . . - _ _ . q ‘ . Ho 003830500 H0030 mzm m C I q C C d d U U G - C & C bi. O O H 030G m0§fl v pad N 6802503 m0m~5n~ I q u — I F > C O F" SIEAJmu; puoaas ‘9; u; smaAe g0 xaqumN 030a 0053 N J :08— H0ud0nw m0mfi9nm . 0...“? I d H 1 — I q I I Q . OH ON 003: 00:20 N J 3303 0025a 000.32: I H.308 . I . . . ‘ ‘ . ‘ . q o ___ .~_——. __da__ ___ _ d __ _ :OH .1. ON om n flaw 0&00m OIII. 0N n dew 0m00m .Ll' OH H Gflmm 0&00m 118 com 00£0 000 a 085. 00* . oom CON .0 90% .H 0&8 n £03££30€ 003m £30350 0309004 cog .om 0££mfim o iii-411%: 000£0E0£300 0035 u£0¢£0003030 U£0 030.9000 005.3300 Q q q - 0030£ 0053 CV £033 £0300£m 000H£nm a 0030£ 0053 ON £03 £0300£m 0003mm dq‘l‘ ‘-‘-q. — “ 0030£ 0053 A: £03 £0300£m 0003mm 4.14 ._ .1.._. a a 0030£ 00E3 v £033 £0300Hm 003.9% 0.. _____.___._______.._ .. . UmMOG mmgwu ¢ mun“ N AHOQEGQ. mmmfidm TA: .ON _______.______.._ . .og .ON SI‘EAISZNI'; puoaas 1:03, u; 31119100 ;0 .IaqmnN 119 .v ££m .H 0Q>H u £030333030 003m “£0E0003m03Q Am 00.0th 00£0000 .. 0838 00m OOmV OOm OON OOH fiTTIlllIIJllllTTfIITTIIIIIIIlITTTTIllITTTIITTIITTT. 000£030£300 003m «£0E0003m030 p£0 0300000 p0£££3£00 000£030£300 30£w30 mEm . a 4.0 q 0030£ 00££3 ON p£0 N £00503 0003mm _ 0030£ 0083 N 30303 000353 1 ... ._.- -_ ..¢.. ___—1 H1~4——-——___~ a O I O .—a C N srezuazu; puooas 1101 u; s1u0Aa }0 .quuan 120 .o ££m .H 0Q>H a £035333030 003m £30250 030500< .Nm 0££wfih 00£0 000 .. 0838 O \O O O 1.0 O O V‘ O O M O O N O «L I. 1. db Ii III II- P P 1 1+1. l-I -IO 1- *- 1v .4. 0 q» 0 0 0£0Um0£30 0 0035 39.0an 330 00 q .._.... .. .Ho 0030£ 0053 o¢ U£0 ON £00303 0003£nm 0030£ 0053 03 £033 £0300£w 000353 . . _. -0 __g q. .0 1. 0030£ 0053 v £033 £0300£m 0003£nm ..:....1......_._..__.... .1 :. 0030£ 0083 N £033 £0300£m 003.9% 2:12.21. 4... AT... '03 L 0030£ 00c£3 N 30303 0003mm smzuaqu; puoaes 1.10:1 u; s1u0A0 go .IaqumN 121 OOO OOm 095000 I 085. OO¢ OOm OON 002830300 003% sawhmoflwomo .0 gm .H 0Q>H u G033£30% 003E 3n0800030fifl .mm 0.3.th OOH . ._. .. . . . 003803230 0 H0530 mzm U-CW ‘ ‘ “' 0030: 0083 ON find OH G00E0n 003.9% q 003C 0053 m .N :03 H0300Hw 003.9% 003G 80.3 H0300Hm 0003a” . _-E u a. q: ¢ OA sremzenr; stun puooas us; u: square ;0 .IaqumN Time - seconds 122 45‘- 500" 40" 35.. 400» i- 3o~- " 3OO"'§254. o m I vs .. “20 o A 200“ 15'- lO~ 100» 5 L__ RMS acoustic activi O.L 0 . . o.'5 1:0 Displacement - 10.3 inches Figure 54. Load-displacement and acoustic activity - Type 1, Run 10. Appendix F. Oscilloscope traces - Type I, Run 7. Emission 2 volts/cm 5 milliseconds/cm Displatement 0.1 volts/cm Emission 1 volt/cm 5 milliseconds/cm Displacement 0.01 volts/ cm Figure 55. Oscilloscope traces - event at T = 292 seconds. 123 124 Emission Z volts/cm 5 millisec onds/cm DisplaCement 0. 2 volts/cm Emission 1 volt/cm 5 milliseconds/cm Displacement 0. 01 volts/cm Figure 56. Oscilloscope traces - event at T = 381 seconds. 125 Emission l volt/cm 140 milliseconds between large pulses Displacement 0. 05 volts/cm Emission l volt/cm 5 milliseconds/cm Leading pulses Displacement 0. 05 volts/cm Emission l volt/cm 5 milliseconds/cm Double trace showing all three pulses Displacement 0. 05 volts/cm Figure 57. Oscilloscope traces - events at T = 465-7 seconds. 126 Emission 1 volt/cm 5 milliseconds/cm Displacement 0. 2 volts/cm Figure 58. Oscilloscope traces - event at T = 478 seconds. ICHIcaN STATE UNIV. LIBRARIES IIHIIIIIIIWIIJHIHIMIIIIHMIIWIHIIHIIHHIHIIIHHI 31293016992541