.c .3 .5. 03...: '9 V... . mmwmkunn ., n... gt 3 I: a} 1 .. 4.9.5 fit .1 x. gfihlwk. . thJrBP x v Ax" ' ‘5": ; 5L5: .. an. «I; f... a. 5.1.“).W) 3%.. x '7} 1.9.4.13». . . ; 3" 1 lhro ; 1 av. . . :I : fit 1 . ’ gnaw» 1.1x h» I. .1» :13? v... 1‘ . 15‘ .r C... an“?! ..v 9F “V: . :23»: .1352}. I u r . .‘ 1.9.1:. r. :agixgiut, .S a. .‘1‘ O 3 ‘ . v .4 a .. l .. '15:: 0.!!- 0 Hm}! ‘,..I. “in. . L I!“ 94. $1,..mmw lat... Ll. . i a iris 3., 5‘ mg- ”8.. 1...... P. u‘ L . u E iizl 11:3; lfiik l3,“ . \h 17!. .I V ‘3 {Flv‘abz-s}. 11.1.4 (I! .:an......: 4‘ (153.17): :4 D7 . xy9c'. . ‘ ; . x '2‘ 11¢. 2.5.15. ) 355:8!- ‘l\,2!0 ‘f‘|’s|‘ vi £5v§55q ’34...er lllllllllllllllllIIIIHIHHIIHIIIUIIWIIIIHHIIJIIIIHHII 31293 01410 3224 This is to certify that the thesis entitled A QUALITATIVE EXAMINATION OF STRAIN USING ELECTRON BACKSCATTERING PATTERNS EMPLOYING A 35MM CAMERA BODY IN A SCANNING ELECTRON MICROSCOPE presented by ALAN W. GIBSON has been accepted towards fulfillment ‘ of the requirements for MAS- degree in WCIENCE /Mn Major professor Datem/ / 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution ‘A LIBRARY Michigan State University PLACE In RETURN BOX :3 rainw- w- chockout «on your record. TO AVOID FINES Mom on or baton dot. duo. DATE DUE DATE DUE DATE DUE II ? MSU In An Affirmative Action/Equal Opportunity Institution ‘ mm: A QUAIJTATIVE EXAMINATION OF STRAIN USING ELECTRON BACKSCATTERING PATTERNS EMPLOYING A 35NIM CANIERA BODY IN A SCANNING ELECTRON MICROSCOPE By Alan W. Gibson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Materials Science and Mechanics 1995 ABSTRACT A QUALITATIVE EXAMINATION OF STRAIN usmc ELECTRON BACKSCATTERING PATTERNS EMPLOYING A 35MM CAMERA BODY IN A SCANNING ELECTRON MICROSCOPE By Alan W. Gibson A comparison was made between two electron backseattering pattern (EBSP) recording systems. A system utilizing a 35mm camera body in an Hitachi $2500 scanning electron microscope (SEM), incorporating film transfer and exposure control, proved far superior in providing high quality EBSP images to a commercial LINK system incorporating a phosphor screen, low light television camera, and SEMPER image processing software. Also, EBSPs have been used to determine the appropriate amount of electropolishing required to obtain strain free material in commercially pure aluminum. EBSP quality, known to deteriorate with increasing strain, was shown to visually improve with increasing electropolishing time. Next, EBSPs have been used to qualitatively examine strain in commercially pure aluminum as a function of distance away from a surface that has been damaged by mechanical grinding. Visual comparisons showed an improvement in pattern quality away from the damaged surface. Additionally, EBSPs have been used to qualitatively investigate strain in a (A1203)p/6061 aluminum alloy metal matrix composite as a function of distance away from an (A1203) particulate. The .SP technique was found to be inappropriate due to the material’s small grain size. To Barbara... my loving wife, best friend, and eternal companion iii ACKNOWLEDGEMENTS This study was made possible by funding from the National Science Foundation, grant DMR-9257826. I would like to thank Dr. Martin A. Crimp for his guidance, friendship, and many interesting, informative discussions on countless subjects. His hard work and dedication have been an inspiration. I would like to thank Leo Szafranski, Machine Shop Technician, for aiding me with his expert craftsmanship. Thanks, also, are extended to the MSM office staff including his Taylor, Joann Peterson, Debbie Conway, and Lorna Coulter. In addition, I owe gratitude to Pat, Dean, Julian, Ojars and my other lab cohorts, as well as to all international students with whom I have had many interesting and informative discussions. Foremost, however, I would like to thank God for His forgiveness. iv TABLE OF CONTENTS List of Tables List of Figures Introduction 1. Literature Review 1.1 What are EBSPs & how are they formed 1.2 How and why are EBSPs used 1.3 EBSP recording systems 1.4 Sample Preparation 1.5 Objectives 2. Experimental Procedure 2.1 Material 2.2 Sample Preparation 2.2.1 GaAs 2.2.2 Commercially pure Al and (A1203)p/6O61 aluminum 2.3 Camera Systems 2.3.1 35mm Camera/Sample Set-up 2.3.2 Link Merlin EBSP Camera System/Sample Set-up 2.4 Acquiring EBSPs 2.5 EBSP Experiments 2.5.1 Recording .SPs using different films in the 35mm camera body vii viii 21 25 3O 39 41 41 41 41 41 45 45 48 48 50 50 2.5.2 LINK Merlin EBSP Camera System vs. 35mm camera using GaAs and commercially pure Al specimens 2.5.3 EBSPs recorded after 5 minutes and after 25 minutes of electropolishing using a commercially pure Al specimen 2.5.4 EBSPs recorded as a function of distance away from a surface that has been damaged by mechanical grinding with a 180 grit sandbelt using a commercially pure Al specimen 2.5.5 EBSPs recorded as a function of distance away from an (A1203) particulate in a (A1203)p/ 6061 aluminum alloy specimen 3. Experimental Results and Discussion 3.1 EBSPs recorded on 4 different films 3.2 LINK Merlin EBSP camera system vs. 35mm camera using GaAs and commercially pure Al specimens 3.3 EBSPs recorded after 5 minutes and after 25 minutes of electropolishing using a commercially pure Al specimen 3.4 EBSPs recorded as a function of distance away from a surface that has been damaged by mechanical grinding with a 180 grit sandbelt using a commercially pure Al specimen 3.5 EBSPs recorded as a function of distance away from an (A1203) particulate in an (A1203)P/6061 aluminum alloy specimen 4. Conclusions List of References vi 50 51 51 53 55 56 66 68 77 83 85 LIST OF TABLES Table 1 Backseatter coefficients, 1;, in percentages, as a function of atomic number and electron energy, E, for normal incidence. vii figure 10 LIST OF FIGURES Backseatter coefficient, 1;, as a function of atomic number, at 20 keV electron energy. Backseatter coefficient, ’7 , as a function of atomic number, plotted for a range of electron beam energies. Backseatter coefficient, 11, as a function of tilt, 0, for several elements as calculated by Monte Carlo simulation at 20 keV electron beam energy. Schematic illustrating the increase in backscattered electrons due to tilting the specimen with respect to the incident electron beam. Diagram illustrating backscattered electron trajectories produced when an electron beam is incident normal to a specimen and to a specimen tilted at 70°. Schematic illustrating the inelastic collisions, Bragg diffraction, and resultant formation and imaging of Kikuchi lines as a result of an incident electron beam. Schematic illustrating the inelastic collisions, Bragg diffraction, and resultant formation and imaging of EBSPs as a result of an incident electron beam. The formation of EBSP bands, on a macroscopic level, resulting from an electron beam incident upon a tilted specimen, and their intersection with an imaging medium. The formation and imaging of ISPs as a result of backscattered electrons diffracting from a (1 -1 0) plane and the correspondence of the [O O 1] and [l l 2] directions with locations on the EBSP. Cylindrical chamber used, in one of the fist published instances, to record electron backscattering patterns. viii ME? 11 12 15 16 18 20 26 11 12 13 14 15 16 17 18 19 20 Current as a function of voltage representative of the relationship seen when electmpolishing aluminum in perchloric acid based electrolytes. Current as a function of voltage representative of the more complex relationship seen when the potential exists for etching, polishing, and pitting. Diagram of electropolishing cell. Specimen/camera configuration, shown in the retracted position, mounted on the SEM goniometer stage. Commercially pure Al sample showing the mechanically ground surface and the electropolishing surface of interest upon which EBSPs were performed as a function of distance away from the damaged surface. (A1203)p/ 6061 aluminum matrix sample showing the (A1203) particulate away from which EBSPs were performed as a function of distance. EBSP images of GaAs recorded with the 35mm camera body system, using Kodak’s (a) TMaxTM 3200, (b) 'IMaxTM 400, (c) Technical Pan”, and ((1) Fine Grain Release PositiveTM film. EBSP images of GaAs recorded using the (a) commercial LINK Merlin EBSP camera system with no image processing, (b) commercial LINK Merlin EBSP camera system coupled with Semper 6.4 image processing of 15 seconds of frame averaging, and (c) 35mm camera body system. EBSP images of commercially pure Al recorded using the (a) commercial LINK Merlin EBSP camera system with no image processing, (b) commercial LINK Merlin EBSP camera system coupled with Semper 6.4 image processing of 15 seconds of frame averaging, and (c) 35mm camera body system. EBSP images of commercially pure Al recorded using the 35mm camera body system and Fine Grain Release PositiveTM film after (a) 5 minutes of electropolishing (b) 25 minutes of electropolishing. 34 35 46 52 54 57-58 61-62 64-65 67 21 22 EBSP images of commercially pure Al recorded, using the 70—75 35mm camera body system and Fine Grain Release PositiveTM film, at (a) the damaged edge, (b) 1.0 micron, (c) 2.0 microns, (d) 4.0 microns, (e) 5.0 microns, (t) 8.0 microns, (g) 10.0 microns, (h) 12.0 microns, (i) 15.0 microns, (j) 20.0 microns, (k) 25.0 microns, and (1) 30.0 microns from the damaged edge. EBSP images of (A1203)p/6061 aluminum alloy recorded, 78-80 using the 35mm camera body system and Fine Grain Release PositiveTM film, at (a) the (A1203)P/ 6061 matrix interface, and at (b) 1.0 micron, (C) 2.0 microns, (d) 5 .0 microns, and (e) 8.0 microns away from the (A1203)P/6O61 matrix interface. INTRODUCTION Diffraction studies often are carried out using transmission electron microscopy, X-ray diffraction, electron channelling pattern, and electron backscattering pattern (EBSP) techniques. Of these techniques, EBSPs are becoming increasingly popular due to their advantages. ISPs have exceptional spatial and angular resolution. Furthermore, the relatively easy sample preparation necessary to obtain patterns combined with the ease of experimentally obtaining patterns and crystallographic information has led to the increasing use of the EBSP technique in materials studies. Furthermore, unlike TEM related techniques, EBSPs may be obtained from bulk specimens. The EBSP method is capable of submicron resolution and is used for local crystallographic measurements, phase identification, local texture measurements, and strain quantification/ qualification. The formation of SIPs is directly dependent upon the material’s crystallography, and therefore, various interpretations of the ISPs can reveal extensive information related to the crystallography of the sample. The distribution, orientation, and delineation of the EBSP reveals information about the distribution, orientation, and crystallinity of the atomic planes in the sample. The delineation of the EBSP bands is dependent upon crystal perfection and therefore can yield localized information on dislocation density, stacking fault densities, or point defect concentrations. The lattice defects are responsible for causing a local bending of the lattice planes, thereby causing a diffuseness in the 2 EBSP band edges. An increasing number of lattice defects causes an increasing diffuseness in the band edges. The current study takes advantage of the increased band edge diffuseness by using EBSPs to study different deformation scenarios. EBSPs are used to qualitatively examine deformation in commercially pure aluminum as a function of distance away from a surface that has been damaged by mechanical grinding. In another experiment, EBSPs are used to study deformation as a function of distance away from an A1203 particulate in an (A1203)P/6061 aluminum alloy matrix. Additionally, EBSPs are used to study the effects of electropolishing on EBSP pattern quality using a commercially pure aluminum specimen. Other preliminary studies are made to find a superior, electron sensitive 35mm film to use in the current examinations. Furthermore, a comparison is made between two EBSP recording systems. One system incorporates a commercial LINK Merlin EBSP lowlight TV camera coupled to a phosphor screen. The other recording system uses a 35mm camera, mounted within the SEM vacuum chamber, and records EBSPs directly on film. 1. LITERATURE REVIEW 1.1 What are EBSPs & how are they formed Signals in the SEM In a scanning electron microscope (SEM), electrons are accelerated down the column through a series of electromagnetic lenses and strike a sample [1, 2]. After the primary incident electron beam strikes the sample in the SEM, the interactions of these electrons with the sample produce different types of signals. These signals may be used to form images and/or perform some type of analysis on the sample. The signals are produced by electron interaction events which may be categorized as being either elastic or inelastic. As described in more detail by Goldstein et al. [3], in an inelastic collision, there is some transfer of energy from an incident electron to an electron in the sample. This transfer of energy may be extremely small, or may include a total transfer of the incident electron’s energy. Secondary electrons, Brehmsstrahlung X- rays, and inner shell ionization are examples of the signals produced by inelastic events. Backseattered Electrons In an elastic collision, there is no transfer of energy from an incident electron to an electron in the sample. These types of collisions give rise to backscattered electrons. Backseattered electrons are strictly defined as being single, elastically scattered electrons whose trajectory is changed by more than 90° from the forward trajectory of the incident electrons and whose exit surface is the same as it’s entry surface [3]. 4 The fraction of incident electrons that are backscattered can be described by the backscatter coefficient, 1; , which is defined in Equation 1 as: ___ “3513:: i333 (1) 113 in n where 1). is the number of electrons incident on the sample surface, m is the number of backscattered electrons, in is the current of electrons impinging onto the sample, and in]; is the current of electrons being backscattered out of the sample. Though beyond the sc0pe of this research, a solid state physics theoretical approach to the understanding of backscattered electrons as been attempted by Dudarev et al. [4]. Goldstein et al. [3] and Reirner [5] explain in great detail the dependence Of m, the number of backscattered electrons produced, on many experimental and material parameters. The atomic number, Z, of the sample under investigation strongly effects rpm, thus effecting n. A plot of backscattered coefficient versus atomic number reveals a strong monotonic increase in r; with increasing Z, as shown in Figure 1. Figure 1 assumes an electron energy of 20 keV. The curve of 1; vs. Z can be fit with Equation 2 as 11=-0.0254+0.016Z- [1.86x10"] 22+ [8,3X10‘7) ] Z3 (2) where r; and Z are as noted above. It should be noted, however, that a closer 0.60 0.50 ~~ A A ‘ ‘A 0.40 + “ ‘9 ‘ 0.30 -~ ‘A‘ 0.20 1 T Backseatter Coefficient, TI 0 .L O 0.00 i i t . O 20 4O 60 80 100 Atomic Number, Z Figure 1 Backseatter coefficient, 17, as a function of atomic number, at 20 keV electron energy. Adapted from [3]. 6 examination of this curve would show that small increases in 2 would not necessarily correspond to a higher 1). When the material is a homogeneous mixture of elements on the atomic level, 1) follows a simple rule of mixtures described by Equation 3 “=21Ci'11‘ (3) where C, is the mass concentration of the individual constituents and n, is the pure elemental backscatter coefficient. Figure 2 (plotted for a range of beam energies) demonstrates that the backscatter coefficient depends on incident electron beam energy to only a slight degree. This appears contrary to what would be expected purely from a Monte-Carlo trajectory prediction [3]. Interestingly, some elements such as Al actually show a decrease in ’1 with increasing electron energy, as shown in Table 1. An important feature, utilized to great advantage in this study, is the dependence of 1] upon sample tilt [3]. Figure 3 demonstrates the reliance of ’1 on the specimen tilt for several elements. This dependence may be fitted with the expression in Equation 4 n(9)=1/(1+COSO)P (4) where the tilt angle, 0, on the X-axis is the complement of the smaller angle between 0.60 a " t f 0.50 «- ‘ A r: m ‘c‘ t x g 0.40 -_ a! 2 “A “f, a 8 030 f b A O i A5kev {as 0.20 .. I; now (0 a D40keV m A49keV 0.10 «L A 5 0.00 : t : i 0 20 40 60 80 100 Atomic Number, 2 Figure 2 Backseatter coefficient, 11, as a function of atomic number, plotted for a range of electron beam energies. Adapted from [3]. Table l Backseatter coefficients, 1;, in percentages, as a function of atomic number and electron energy, E, for normal incidence. Adapted from [5]. E[keV] Be Al Cu Au 9.3 5 17.5 31.3 47.8 11.0 5 17.1 31.0 48.0 13.4 5 16.4 31.4 49.1 17.3 5 15.9 31.0 49.2 25.2 4.5 15.1 30.7 50.1 41.5 4 14.5 30.1 50.7 62.1 4 13.7 29.9 51.3 81.8 4 13.5 29.4 51.0 102.0 3.5 13.3 29.1 51.3 1.00 0.90 1- ‘:080 «- ‘ 3 9 0:) 0.70 “- ‘ ‘ o O A E 0601- ° 0 A a: 5 ° 0 ‘ . A I o 0.50 ‘ o ’ A h I d) O O A § 0.40 ° A n m . . I AM: ‘5 0.30 A I 0A9 a“; A OCu 0.20 A I AA! IC 0.10 . I 0.00 1 1 1 1 1 . 0 15 30 45 60 75 90 Tilt, 9 (degrees) Figure 3 Backseatter coefficient, 1;, as a function of tilt, 0, for several elements as calculated by Monte Carlo simulation at 20 keV electron beam energy. Adapted from [3]. 10 the electron beam and the specimen plane, and P = 9/2‘”. A purely geometrical argument reveals that changing the angle of beam incidence by sample tilt is the same as shifting the trajectory of the backscatter electron cone. Figure 4 schematically illustrates this shifting of the electron cone trajectory when altering the sample tilt from 0° to near 70°, where an electron has an equal probability of landing anywhere on the darkened base of the cone. The effect of tilting the sample is now obvious in that more backscatter electrons are able to escape the surface due to simple geometric advantage. The conical shape, in this instance, is indicative of the general shape of the electron trajectories as given by Monte Carlo simulations. The Monte Carlo Simulation in Turbo Pascal program [6] demonstrates that at a beam energy of 25 keV, using aluminum’s atomic number, weight, and density, the backscatter coefficient increases with increasing sample tilt. At 0° tilt, i.e. beam perpendicular to sample surface, 1; = 0.160; at 30° tilt, 1) = 0.170; at 60° tilt, 11 = 0.400; and at 70° tilt, q = 0.470. Tilting of the sample not only increases the number of backscattered electrons, but also increases their tendency for forward scattering and changes their angular distribution, as shown in Figure 5 . Forward scattering occurs when the electron trajectories continue in approximately the same direction as the electron beam. Thus, as the sample is tilted, the backscatter electron signal is stronger in directions away from the incident beam and a detector placed in this appropriate position will register a much higher amount of backscattered electrons, resulting in a higher 11 than if it were placed elsewhere. A beam incident on a specimen tilted at 70° produces backscattered electrons that follow a more forward trajectory than if the specimen had 11 .5 88. Bag... .883 8885 58235 85 9 609.2 .23 538% 05 wean“ 3 26 88.629 Bounce—own E 88.6.: 05 wfiagmaa 058828 4 25mm m0 1(3). (3) Lastly, each plane, positioned nearly parallel with the primary beam, is irradiated from both sides and from all directions as explained above and shown in Figure 6. These 3 facts, taken together, explain the formation of Kikuchi lines, and for the most part, the formation of EBSPs. Figure 7, more accurately depicting the formation of EBSPs, shows the backscattered electrons exiting the same surface as 15 .3 3 Sec 8.53. .543 558—0 .522: 5 Lo :82 a ma 8:: 2085— me wfimfli e5 ecu—WEE .535.— 93 iguana—O wan—m £5558 Oman—Os 95 men—Ema: Ora—55m c 25E m2... .IUDvSv. “2... 3.031.! ¥C(O no PIG-CD to 20:.me.x 20....UW01X €09.00 OOIE. A. x 5 CCU—n- Gav—aald 20..P0(Cn-l_0 00(Im A A :On. .2 w... D Z mzoaaoc f 055%. 0— «Sumo 0— .PZM0.02- 16 .553 5.585 35285 5 we :82 a 8 “Emma Lo mfiwfifi e5 52558 “5:32 98 .couoEEe mun—m .meommaoo 3:305 05 95863: cum—55m ‘11 322:8 n 2522. <5” \ 4 o. .. 25m Emeoz. mofimsm m4n2>Ozw m._n_2 75 °) also produce an extremely large fraction of secondary electrons contributing to background signal, thus reducing band-to-background contrast. Goldstein et al. [3] report asymptotic behavior near 80° with the secondary electron coefficient going to infinity. Additionally, Dingley [37] reports that higher angles of incidence near 85 ° result in excessive probe elongation in the interaction volume causing diffuse patterns. Furthermore, since the interaction volume is effectively enlarged in the surface plane, unwanted EBSP diffraction information may be obtained from other grains, dislocation subcells, and phases. These considerations have lead to an incident angle of 70° as the standard for most EBSP studies. When using the EBSP technique, the SEM is used in spot mode as opposed to raster mode [37]. It is also necessary to carefully focus the beam while in secondary electron imaging mode prior to switching to spot mode [33]. These practices, like the above EBSP related sub-procedures, are necessary to improve spatial resolution by decreasing the interaction volume, thus decreasing residual electron noise and unwanted diffraction information while, at the same time, increasing the pattern-to- noise ratio. It is ultimately necessary for the beam diameter and interaction volume to be smaller than the crystal from which the EBSPs are obtained. In Dingley’s 1984 publication [37], achieving distinguishable patterns from 2 pm grains was possible, though inconvenient. Further reductions in beam current, corresponding to a probe 30 diameter of 10 nm, resulted in distinguishable patterns, but only after 15 minutes of exposure. In 1992, Dingley and Randle [13] reported a spatial resolution of 200 nm. 1.4 Sample Preparation Dingley and Randle [13] state that there are two essential prerequisites for obtaining EBSPs. The first, ensuring a sample tilt angle of 70° from horizontal, has already been discussed. The second requirement relates to obtaining a relatively strain free, clean surface suitable for the EBSP technique. Strain/deformation ”Strain free" does not mean free of all and any strain in the material. Obviously, as mentioned earlier, studying strain/deformation in a material is one important aspect for which EBSPs are very well suited. The strain that needs to be avoided, in this case, is the artifact strain induced on and near the surface of the material as a result of mishandling or sample preparation procedures. In softer materials, mechanical polishing, while removing deformed material, also introduces damage, or plastic deformation. For example, an annealed polycrystalline 70:30 brass sample abraded on 220—grit SiO2 paper can exhibit induced deformation as deep as 10 pm [44]. Indeed, Harland et al. [45] report that EBSP diffraction information is obtained from within 10 nm of the surface (albeit using a field emission SEM). Another analysis shows that at 30 keV, backscattering patterns disappear after applying a 10 nm coating of aluminum while at 20 keV, the patterns become invisible after applying a 5 run coating [37]. This implies the patterns originate from the top 10 nm of material [37]. Thus, it becomes obvious that the near-surface volume of the material needs to be free of artifact plastic deformation. The removal of this 31 deformed layer may be facilitated by several means. Electropolishing (discussed below), etching, acid string saw cutting, EDM, and ion milling may all be used to this end. However, electropolishing is more commonly used when preparing EBSP samples due to its relative ease, low expense, and wide expanse of reference material [11, 46, 47]. Additionally, archaic yet effective electropolishing cells are easily fabricated with inexpensive materials and components. Surface cleanliness and smoothness Surface cleanliness may be achieved by several methods. The production of a fresh fracture surface by methods such as cleaving a brittle material, ultrasonic cleaning in either water or a solvent, electropolishing, ion milling, and acid string cutting have all been used to produce a clean surface. Surface contamination occurring within the SEM chamber can have detrimental effects on EBSP analysis by reducing the quality of the pattern, thus possibly giving erroneous indications of strains. Wilkinson and Dingley [19] state that in order to avoid these surface contamination problems, the sample should only be observed for a few seconds at high magnification before switching to spot mode to obtain an EBSP. Not only does the secondary electron image darken due to surface hydrocarbon contamination, but the resulting EBSP pattern quality deteriorates with increasing beam exposure time. Wilkinson and Dingley [33] state that a 1 minute scanning mode beam exposure over a 40 um x 40 um area at SOOOX will produce the equivalent pattern degradation of a 1% strain. Yet another important prerequisite for producing surfaces that yield high quality EBSPs lies in the understanding of surface smoothness. The sample surface needs to 32 be smooth on the order of the beam diameter so that chaotic surface diffraction does not occur. Fortunately, locating areas that are flat on the order of the beam diameter is relatively easy even in macroscopically rough specimens [31]. Alternately, macroscopically smooth surfaces can often look impressive, but be deceivingly rough on the microscopic scale. Surface smoothness can best be achieved by proper electropolishing techniques. Electropolishing While the exact mechanisms of electropolishing are still not completely understood, this approach still has the ability of removing all traces of deformation induced by mechanical polishing operations in sample preparation. Electropolishing is also used to remove surface imperfections and oxides. Most metal alloys have a fine, thin, surface oxide layer present under normal equilibrium. Davidson [36] states that under good conditions, these oxides are easily penetrated by electrons and may normally be considered unimportant. In order to achieve these " good" conditions, metals are often electropolished, thus ensuring a thin, uniform oxide layer. Electropolishing is used to prepare samples for observation in optical, scanning electron, and transmission electron microscopy. As explained in Metals Handbook [47] , electropolishing is thought to include both a smoothing action and a brightening action. The smoothing action is accomplished by preferential attack of hills and ridges on the surface commonly produced by mechanical polishing. When the surface is made the anode in an electropolishing cell, a viscous polishing film forrus on the surface of the sample and acts as a resistance layer. The hills lying closer to the edge of this resistance layer, 33 have lower resistance than the valleys, and therefore have a higher current. This higher current causes the hills to dissolve much faster than the valleys resulting in a smoothing action. The brightening action is a result of the removal of surface irregularities on the order of 0.01 p.111. Etching, a common occurrence in stages of electropolishing, must also be avoided in order to achieve this bright surface if so desired. Electropolishing results from a simple relationship between current and voltage that is sometimes difficult to control. It is important to note, however, that the voltage is the quantity primarily affecting the polishing conditions, while the current is a variable depending on sample size [11]. Figure 11 schematically represents this relationship for the electropolishing of aluminum in a perchloric acid (I-IClO4) electrolyte. At lower voltages, a passivating film forms on the surface preventing the passage of current. This scenario is similar to placing the sample in a chemical etchant. Indeed, etching does occur under these conditions of low current! voltage. However, it is seen that electropolishing occurs over a continuous range above some critical voltage level. Above this level, the passivating film is broken allowing current to pass through to the sample affecting an ionic exchange between metal and electrolyte. A much more complex current-voltage relationship is illustrated in Figure 12. Here, etching occurs in the region between AB where no current passes. Electropolishing occurs in the region BC characterized by a constant current density with increasing voltage. Region CD often reveals its character by causing pitting on the sample. This pitting is due to gas bubbles breaking the polishing film surface causing momentary, localized increases in current. 34 POLISHING AZFHZDZUCO ETCHING VO LTAG E Figure 11 Current as a function of voltage representative of the relati ' . . . . onshrp seen when electropohshrng alumrnum 1n perchloric acid based electrolytes. 35 D C U PITTING R R B C E N POLISHING T ETCHING A VOLTAGE Figure 12 Current as a function of voltage representative of the more complex relationship seen when the potential exists for etching, polishing, and pitting. 36 In order to properly electropolish a specimen, only the surface for which electropolishing is desired should be in contact with the electrolyte. Many times, this can be accomplished by mounting the specimen in a mounting compound. However, some mounting compounds may react violently with the electrolyte. For example, Bakelite should not be used when electropolishing in perchloric acid compounds due to explosion hazards. Perchloric acid has a brief but violent history of explosion hazard [48]. A chemically inert, insulating lacquer is often used to paint the areas where electropolishing is unwanted. After Metals Handbook [47], electropolishing cells usually consist of a current source, usually DC, with some means of varying the potential. The cell also consists of a container for the electrolyte often surrounded by some type of cooling mechanism. Lower temperatures result in a slower, more controlled polishing process [11] as well as present safer conditions for electrolytes to function [48]. Another advantage of using lower temperatures lies in the expanding of the current-voltage plateau, discussed later [46]. A stainless steel beaker is often used as the container as it can act as the cathode as well as being impervious and inert to most electrolytes. Also, the cell often contains a means for stirring either the electrolyte or the specimen during polishing to help remove any chemical by-products from the specimen surface that might interfere with the electropolishing process. Ideally, the stirring should be nondirectional to prevent furrowing on the surface. But these macroscopic irregularities are often insignificant in EBSP analysis, as stated earlier. Prior to placing the specimen in the electropolishing cell, it is a good idea to pre-clean the sample. Aluminum alloys can be precleaned in a 5 % NaOH solution for 37 30 seconds in order to remove any surface contamination, dirt, or finger oil that might detrimentally affect the local electropolishing rates [49]. The specimen is made the anode and placed in the cell so that the surface to be polished faces the cathode. Depending on the type of polishing occurring, under passivating film or by gas evolution, the surface is placed either horizontally or vertically. Vertical placement, if at all possible, is best used as it allows bubbles better opportunity to escape from the surface thereby decreasing the chance for pitting. Experimental settings rarely produce the ideal current-voltage relationship shown above in Figure 12. Electropolishing curves rarely show this perfect plateau. However, there is usually an area on the curve where the slope falls off somewhat giving the researcher an indication of where to begin electropolishing. Lower temperatures, as already stated, aid in expanding this plateau. After an initial test, some minor adjustments may bring the cell to its optimum setting. These electropolishing current-voltage curves may be determined experimentally. After the cell is constructed, the voltage is adjusted on the potentiometer and corresponding currents are recorded. Plotting these data points may reveal the necessary plateau needed to obtain optimum electropolishing although some researches have had little success with this technique [50, 51]. As stated above, electropolishing is used to prepare samples for EBSP analysis in order to remove irregularities, oxides, and surface deformation artifacts often introduced during mechanical polishing. Indeed, most published research briefly states the manner in which the sample was prepared. 38 An aluminum MMC consisting of 140 um diameter silicon carbide fibers can be successfully prepared by mechanical polishing, followed by electropolishing with a 10% nitric acid in ethanol solution [19]. This electropolishing should be done, at the very least, the same day as the EBSP analysis to reduce excessive and uneven oxide build-up. A commercially pure 1100 aluminum alloy can be prepared for EBSP analysis by mechanically polishing down to 5 pm aluminum suspension followed by a 30% nitric acid in methanol electropolish and an etchant to reveal grain boundaries [21]. Bottcher et al. [16] sputtered the prepared surfaces of cold rolled electrical steel with neutral argon atoms in order to provide an accurately flat specimen. Dingley and Baba-Kishi [31] electropolished B-tin with Struers A2-l electrolyte and kept the sample in solvent until just prior to analysis. Dingley [37] electropolished IN718 superalloy in order to remove the surface work induced by mechanical polishing. Using a technique developed by Davies et al. [52], a series of papers on the penetration depth of ions in tungsten [53], silicon [54], and aluminum [55] describes the conversion of a known thickness of the metal to an oxide. The oxide layer is then dissolved in some aqueous solution to reveal an extremely uniform surface. This technique has been called electrochemical stripping or anodizing-stripping. Unfortunately, most sample preparation procedures, deemed so critical for .SP analysis, are unsatisfactorily explained in EBSP publications. This leaves the current researcher repeating studies to determine electropolishing procedures for selected materials. 39 1.5 Objectives A prior EBSP related investigation [38] in the MSM department at MSU used a commercial LINK system employing a phosphor screen imaging medium coupled with a low light TV camera. Also, an FeAl alloy of a sufficiently high atomic number was used resulting in average pattern quality. Even though hard prints of the patterns were not required for the texture analysis, a single hard print was made for illustrative purposes, using a piece of electron sensitive film placed in the SEM chamber. Sufficient for this previous study, the LINK Merlin EBSP camera system may not be adequate for other studies using low atomic number elements or studies requiring high quality .SPs for strain comparisons. Therefore, an alternate recording system, one using a 35mm camera body, was investigated in the present study. Initially, it was important to see if the 35mm camera body system had the ability to record high quality EBSP images with some appropriate, electron sensitive film. Additionally, it has been determined if the commercial LINK Merlin EBSP camera system had the ability to successfully yield high quality EBSP images, from relatively low atomic number aluminum alloys. This study describes the development of this 35mm camera body EBSP recording device and compares the results with the output of a commercial LINK Merlin EBSP system. Furthermore, because of the importance of sample preparation on obtaining high quality EBSP images, the effects of electropolishing on EBSP image quality has been examined. Once the feasibility of using the 35mm camera body system for recording high quality EBSPs had been determined, examinations into its usefulness were conducted. 40 Therefore, another goal of this study was to discover whether the system could be used to qualify strain in certain aluminum alloys. 2. EXPERIMENTAL PROCEDURE 2.1 Material The GaAs single crystal material used in this study was supplied by Professor M. Aslam, Department of Electrical Engineering, Michigan State University, in thin wafer, semiconductor, as-polished form, 0.62 millimeters thick. The commercially pure aluminum in this study was supplied by Dr. D. A. Grange, ALCOA Technical Center, for a previous study. The material was ingot cast and vacuum remelted and contained equiaxed grains averaging 2mm diameter in its as-received form. The (A1203)P/606l aluminum material used in this study was supplied by Prof. K. Subramanian, Department of Materials Science and Mechanics, Michigan State University. The (A1203),,/ 6061 aluminum material was extruded, solution annealed at 560° C for 1 hour, room aged at 24° C for 65 hours, then artificially aged at 170° C for 14 hours in its as-received form [56]. 2.2 Sample Preparation 2.2.1 GaAs In its supplied form, the GaAs semiconductor wafers were single crystal and polished to a mirror finish. The only additional preparation required prior to acquiring EBSPs was the rinsing of the wafer in methanol prior to analysis. 2.2.2 Commercially pure Al and (Ale,)p/6061 aluminum Mechanical Polishing The aluminum based materials were cut to an approximate 1.0 cubic centimeter shape using a Buehler Iso-cutTM diamond wafering blade mounted in a Struers 41 42 Accutom-STM high speed cutoff machine. Grinding on a 180 grit sandbelt was sometimes necessary to ensure that the surface to be studied was parallel to its opposing surface for mounting purposes. At this point, a 0.080 in. diameter hole was drilled and tapped into the sample face opposite the one to be used in the EBSP analysis. This was so a screw, which would serve as anode lead attachment, could be screwed into the sample without disturbing the material near the analysis surface. Heat Treatment The commercially pure Al was given a full anneal at 450° C for 1 hour and then furnace cooled. This was done to ensure a reasonably strain free, uniform material. The (A1203)p/6061 aluminum alloy was solution annealed at 560° C for 8 hours and then air cooled to eliminate all precipitate phases that might interfere with the EBSP analysis. Metallographic Preparation The surface of interest was then mechanically ground on 240, 320, 400, and 600 grit SiC metallographic papers, respectively. Following which they were polished using 5.0 pm, 3.0 pm, and 0.5 pm alumina lapping solutions, respectively, on billiard polishing cloths. Electropolishing An electrolytic cell was constructed in order to facilitate electropolishing of bulk samples. The potentiometric circuit cell consisted of a DC Struers PolipowerTM power source potentiostat containing both voltage and current readout meters. The positive (+) lead from the power source was connected to the screw in the sample, 43 thus making the sample the anode in the cell. All surfaces of the anode, except for that surface to be subjected to -SP analysis, as well as any parts of the positive lead that would be submersed in the electrolyte, were painted with MicroshieldrM acid resistant lacquer. A stainless steel beaker served as both the electropolishing solution container and cell cathode once the negative (-) lead from the potentiostat had been connected directly to the beaker. The stainless steel beaker was placed in an insulated dry ice/methanol bath maintained at 0° C. This bath was then placed on a Thermolyne Nuova IITM magnetic stirring plate so that the electrolyte could be continuously stirred, with a 1" magnetic stirring rod placed directly in the stainless steel beaker (Figure 13). Prior to being introduced to the electropolishing cell, the samples were subjected to 30 seconds of cleaning in a 5 % NaOH solution in order to remove any surface contamination that might cause uneven electropolishing. The optimum electropolishing conditions for the commercially pure Al were found to be an electrolyte consisting of 70 ml distilled H20, 350 ml ethanol, 50 ml 2- butoxyethanol, and 30 ml H2ClO4 (perchloric acid) maintained at 0° C. The approximate 0.60 sq. cm. surface area was electropolished for times ranging from 5 to 25 minutes (depending on the particular experiment) at 20.0 V resulting in a current of 0.66 mA (i.e. a current density of 1.1 mA/cm’). After electropolishing, the commercially pure aluminum sample was etched in a 10% HF(48%)/distilled H20 solution for 1 minute, rinsed in running hot tap water, methanol, and blow—dried in hot air. A final 3 minute ultrasonic methanol rinse followed and immediately preceded placing the sample in the SEM. The optimum electropolishing conditions for the (A1203)p/ 6061 Al were found to -----—--q .=8 ”Eamon—938.0 he Enhan— 28:»E whim Dmmam ozEmEm . 0_.Pw20<_2 m 0.52.922 " _ m._n:2JJ<>>< omEmOamma mam; «menu 19:? 20a: mo