Microscopy: Science, Technology, Applications and Education A. Mndez-Vilas and J. Daz (Eds.

) ______________________________________________

Basic Principles and Application of Electron Channeling in a Scanning Electron Microscope for Dislocation Analysis
R. J. Kamaladasa and Y.N. Picard
Materials Science and Engineering Department, Carnegie Mellon University, 5000 Forbes Ave., Pittsburgh, PA, USA 15213-3890 Dislocations are extended defects within crystalline solids that often influence the mechanical, electrical, magnetic and optical properties of the material. Establishing more quantitative relationships between dislocations and material properties relies on effective microscopy or spectroscopy methods that can resolve both dislocation type and position within the solid. Diffraction has historically been an effective mechanism in electron microscopy for creating image contrast necessary to locate and identify various dislocations within materials. Transmission electron microscopy (TEM) has long been the most conventional method employing diffraction contrast for dislocation analysis. However, scanning electron microscopes (SEM) can also access diffraction contrast through the phenomenon of electron channeling. This article highlights recent research progress in the imaging and identification of dislocations using SEM-based electron channeling. A brief introduction to electron channeling is provided and various example microscopy images of dislocations are presented. Practical details are provided on how to access electron channeling in a conventional SEM, optimizing channeling contrast, and intepreting the nature of channeling contrast features. Keywords Scanning Electron Microscopy; SEM; electron channeling; ECCI; Dislocations; Defects; Diffraction; ECP

1. Introduction
Dislocations are line defects in crystalline solids that can strongly influence the material properties. Determining the density and types of dislocations within a solid is a crucial step towards explaining how specific material properties are affected by dislocations. The most common method for observing dislocations is by the use of diffraction contrast in transmission electron microscopy (TEM). However this technique suffers a number of limitations related to the destructive sample preparation necessary to obtain electron transparent samples (< 300 nm thickness), the limited viewing area accessible for each TEM sample (order of few microns), and the deviation from bulk-like conditions for analyzing dislocation behavior in such thin specimens. Electron channeling offers the ability to execute diffraction contrast imaging inside a scanning electron microscope (SEM), removing all of these limitations imposed by TEM analysis with still sufficient imaging resolution to analyze individual dislocations . This chapter will cover experimental and theoretical developments in electron channeling, describe the experimental approach for conducting electron channeling in the SEM, and present example images of various dislocations.

2. Background of Electron Channeling

Kikuchi-like bands were observed by Coates [1] in 1967 while imaging single-crystals using a conventional SEM. Coates observed that these Kikuchi-like bands were readily produced in single-crystals at low magnification, which implied an angular dependence by the scanning mechanism was giving rise to these patterns. The mechanism for generating these bands was defined as electron channeling and was immediately highlighted as a technique which could aid determination of crystal orientation using an SEM. Hence Coates predicted that with a stage-rocking configuration, where the specimen was tilted in two orthogonal directions with respect to a fixed incoming electron beam, it would be possible to obtain electron channeling patterns (ECPs) from bulk crystals and even from individual grains in polycrystalline materials. Booker et al [2] suggested that these Kikuchi-like bands could be explained by the super-position of two Blochwaves. Booker went on to describe the analogous nature of bend contours observed at high-magnifications in TEM and ECPs observed at low magnification in scanning mode microscopy. The group went on to predict the possibility of observing sub-grain boundaries and dislocations at high magnification using electron channeling contrast imaging (ECCI). Not long afterwards, Clarke [3] and Stern [4] published experimental ECCI micrographs of dislocation networks in thin-foils by virtue of both transmission and backscatter SEM configurations. Meanwhile, a theoretical explanation for channeling contrast, which seemed more intricate than conventional diffraction contrast observed in TEM, was slowly developed. In 1971 Clarke and Howie [5] used a many-beam Bloch wave approach to analyze the attenuation of anomalous absorption and the contrast effects of strain fields such as screw dislocations and stacking faults. The theory approximated that for maximum dislocation contrast the depth should be 0.2 x g from the surface, the probe size should be less than (g b) g /5 and that the divergence angle should be less than 1/(|g| g) (where g is the extinction distance).

FORMATEX 2010

1583

Microscopy: Science, Technology, Applications and Education A. Mndez-Vilas and J. Daz (Eds.) ______________________________________________

Spencer et al [6] in 1971 developed a dynamical many beam Bloch wave approach to take into account multiple scattering between forward and backscattered intensities. The theory predicted that the probe size should be approximately less than (g b) g /4 and that the divergence angle should be less than 1/(|g| g) for observing crystal defects. The versatility of the dynamical approach was that it predicted the disappearance of channeling bands analogous to vanishing of diffraction spots in TEM diffraction patterns due to double diffraction. Another dynamical many beam approach was proposed by Sandstrom et al [7] in 1973 which took into account the energy losses due to single-electron excitations and plasmons. This addition led to a better explanation of why the contrast-to-noise ratio of ECCI depends on the magnitude of the energy window. They theorized that it would be possible to enhance the contrast of ECPs by the use of a small negative bias grid, a cascade of retarding grids or an electron velocity analyzer. A high-energy filter mounted scintillator detector was used by Pitaval and Morin et al, in 1977 [8] and 1979 [9], allowing the first demonstration of invisibility criterion for dislocations in bulk specimens. They also used the forescatter (high-tilt) geometry which decreased absorption effects and increased contrast. This geometry became the preferred choice for dislocation imaging until Simkin and Crimp [10] in 1999 showed that with sufficient probe current and modern high detector collection efficiencies, backscatter (low-tilt) geometries were also a pragmatic approach to dislocation imaging. Several Monte Carlo formulations of a two-beam dynamical diffraction theory were published in the early 80s, as well as a transport equation approach by Spencer and Humphreys [11] that included cumulative small angle interactions by the use of a Rutherford elastic scattering cross-section approximation. The theory was able to fit the experimental information depth of an ECP arising from thin films with better accuracy than the forward-backscattering approximation made by Spencer et al [6]. A few practical uses for ECPs were discussed by Joy et al [12] in 1982 where they discussed in detail how crystallographic orientation mapping, measuring lattice bending and observing contrast between grains can be achieved by selected area channeling patterns and rocking beam systems. They went on to describe means of calculating lattice parameters by measuring the angular width of channeling bands, calculating dislocation densities and extent of mechanical deformation or radiation damage by resolving the distortion of ECPs. However, orientation and phase determination became more accessible for SEM users via electron backscatter diffraction (EBSD). With no need for beam or specimen rocking, EBSD merely relied on acquiring Kikuchi patterns using 2-D detectors (phosphor screen). A simplified experimental set-up combined with automated indexing schemes rendered EBSD as an attractive and commercially viable alternative to channeling for many of these applications. As EBSD continued to supplant ECP for orientation analysis, researchers continued to explore theoretical and experimental aspects of electron channeling. In 1995, Dudarev [13] et al was able to take into account dynamical elastic scattering, quasi-elastic phonon scattering by means of a Rutherford cross-section approximation, and low-angle inelastic scattering by means of Bethe energy loss law, which were all possible by describing electron channeling inside a crystal with a kinetic equation for the one-particle density matrix and reducing it to an inhomogeneous transport equation. While ECP analysis had been less utilized, a number of ECCI studies have been published during the past couple of decades. In 1990, Czernuszka et al [14] imaged near surface dislocations of Si, Ni and Ga thin films and were able to show g b = 0 invisibility criterion for screw dislocations. Misfit dislocation clusters were observed by Wilkinson et al [15] in Si1-xGex epitaxial thin films grown on a Si substrate up to depths greater than 1 micron. The study showed that the contrast was at a maximum when the incident beam was perpendicular to the line direction. Researchers studying intermetallics over the years have employed ECCI and published several papers that show the imaging of deformation twins, twin-grain boundary interaction, persistent slip bands (PSBs) and dislocation cells. Ahmed et al [16], observed the interesting effect of constant channeling contrast emanating from high density dislocation structures, irrespective of the diffraction condition or deviation from Bragg. This was explained later by Dudarev et al [17] claiming that dislocations which run parallel to the incident beam have randomly distributed kinks that scatter electrons in a range of phases that lead to channeling contrast at almost all tilt angles. Ahmed et al also showed the formation of dislocation cells before PSB nucleation, and hinted the role of dislocation cells in the formation of PSBs. With bulk sampling capabilities of the SEM, these ECCI studies have a vast potential in opening up a new approach to non-destructive examination (NDE) for other material systems. Recent studies of ECCI have been centered on electronic materials including GaN and SiC. Trager-Cowan et al [18], and Picard et al [19-20], have used a combination of electron backscatter diffraction (EBSD) and ECCI to image dislocations and atomic steps. These studies have focused on improving detector and specimen geometry, while also understanding the contributions of diffraction vector, deviation parameter and the dislocation Burgers vector to the channeling contrast for individual dislocations in ECCI. Due to advances in computation power, 2-dimensional simulations are more accessible for visualizing the accuracy of theoretical work. Twigg and Picard in 2009 [21] simulated ECCI micrographs for screw dislocations in 4H-SiC with a Bloch wave single inelastic scattering theory for fast electrons devised by Rossouw et al, that accounted for the depths of normal absorption and anomalous absorption, and by using the conventional HowieWhelan equations used in diffraction contrast in TEM. The simulated ECCI micrographs for screw dislocations have close agreement with the

1584

FORMATEX 2010

Microscopy: Science, Technology, Applications and Education A. Mndez-Vilas and J. Daz (Eds.) ______________________________________________

observed experimental images and directly account for the influence of dislocation Burgers vector. Winkelmann et al [22] in 2007, in an effort to simulate ECPs in GaN, used the reciprocity principle to relate the incident beam dependence of an ECP to that of an outgoing diffracted beam by an EBSD process. This work included a quasi-elastic backscattering event from atomic nuclei and assumed it was sufficient to account for all the energy losses. By using further approximations, the group was successful in simulating qualitative EBSD patterns with HOLZ effects and fine zone axis structure.

3. Principles of Electron Channeling

Although the term channeling contrast is used to describe the contrast of an ECP, the maximum contrast occurs due to the same electron-electron inelastic scattering processes that give rise to Kikuchi bands in TEM. However, to understand the subtle contrast features of an ECP, the effects of other inelastic absorption processes needs to be accounted for as we have discussed in the historical progression of a better theory for electron channeling. At the time electron channeling was discovered, the term Kikuchi-line was almost exclusively used for TEM. Since the ECPs were obtained under a SEM, Booker et al [2] called them inverse channeling patterns because ECPs were more comparable to proton channeling patterns observed at the time. The term channeling is not entirely misleading since it is used to describe the motion of charged particles inside a crystalline lattice. Although the wave formalism is preferred for quantitative analysis, the particle approach is simple in its ability to explain orientation contrast and defect contrast qualitatively. Rather than the lattice being a set of atomic points, channeling assumes that the effect of a crystal lattice can be described in terms of channels or paths where the particle can preferably penetrate to a higher depth before scattering (Figure 1a, 1b). Hence, certain orientations of the crystal will backscatter more electrons than others, giving rise to orientation contrast.

Fig. 1 Illustration of an atomic lattice relative to incoming electron trajectories so that a relatively more (a) closed channel or (b) open channel condition is obtained. The presence of an edge dislocation (c) can locally convert an open channel to a closed channel condition.

At low magnifications the scanning motion of the electron beam ensures that many channels are accessed over a wide angular range, leading to the wide variations in backscattered electrons and the formation of an ECP in the resulting image. At high enough magnifications a single channel corresponding to a specific orientation can be isolated. Hence the image for a perfect crystal at high magnifications should show no contrast or rather a constant signal. But the presence of a local crystal defect like a dislocation or a stacking fault may block the channel and will preferentially scatter more electrons back towards the detector (Figure 1c). Or the case might be that a dislocation may open a channel and let the electrons penetrate to a higher depth, reducing the number of electrons that scatter back towards the detector. In either case, the effect gives rise to dislocation contrast via electron channeling. The reader should be aware that the basic channeling description outlined here is a greatly simplified approach and is not the preferred method to describe contrast effects for ECPs and ECCI. In modern literature, the reciprocity principle is used to relate ECPs to the widely known phenomenon of EBSD. Although they are not exact reciprocal versions of each other, the wide angular collection of electrons by a 2-dimensional detector (phosphor screen) in EBSD can be corresponded with the large angular width of electron beam scanning via low-magnification imaging for ECP, as illustrated in Figure 2. The important distinction between EBSD patterns and ECP is that the former directly informs us

FORMATEX 2010

1585

Microscopy: Science, Technology, Applications and Education A. Mndez-Vilas and J. Daz (Eds.) ______________________________________________

of the specimen orientation relative to the detector and the latter directly informs us of the specimen orientation relative to the incoming electron beam. Knowledge of the incoming beam trajectory relative to the crystal specimen is critical for obtaining and isolating specific diffraction conditions so that channeling contrast features will be stronger in intensity and more easily related to specific extended defect types.

Fig. 2 Illustration depicting the reciprocal nature of EBSD and ECP/ECCI.

4. Experimental Method and Results

Electron channeling can be carried out both at high-tilt and low-tilt configurations using any suitable backscatter electron detection apparatus, such as forescatter diode detectors mounted on commercial EBSD systems or polepiecemounted backscatter detectors. A high-tilt configuration will increase surface sensitivity and increase the contrast arising from atomic steps and surface topography. The low-tilt configuration will increase surface penetration. For commonly grown crystal orientations, the surface normal will correspond to a zone axis that should be easily visible at the low-tilt configuration. Selecting either configuration will limit the accessible diffraction conditions, and therefore g b visibility criterion for a given crystal system should be considered before choosing the appropriate configuration. For the purpose of simplicity, we will consider only the low-tilt configuration of ECCI. Employing a conventional backscatter diode detector, optimal channeling conditions include operating the detector in a Z-contrast mode or COMPO Mode where signal from all hemispeheres (A+B) or quadrants (A+B+C+D) of the backscatter detector are summed together. Surface preparation is not needed for specimens with sufficiently smooth surfaces. Adequate conductive paths need to be created since channeling requires a higher current than usual secondary electron imaging. The user must balance between a higher spot-size for sufficient backscatter electron yield and smaller spot size for adequate imaging resolution. For general purposes, a voltage of 20kV and an aperture of 30-100 microns can be used. The voltage can be decreased in order to decrease surface penetration into specimen as in shown in Figure 3, where lower voltage (5 kV) allows direct imaging of atomic steps (Fig. 3a) in a backscatter mode while increasing the voltage leads to higher resolution imaging and sharper contrast for individual dislocations (Fig. 3c).

Fig. 3 ECCI micrographs showing diffraction contrast of screw dislocations in GaN (0002) at (a) 5kV (b) 10kV (c) 20kV.

The specimen can be mounted on a conventional flat or angled SEM stubs for obtaining backscatter or forescatter geometries under typical SEM working distances (5-10 mm). After the specimen is brought to an appropriate working distance (WD), final adjustments can be made to the WD so that the backscatter electron yield captured by the detector is maximized. This can be achieved by imaging at a low magnification and observing when the ECP contrast and brightness are highest as WD is varied.

1586

FORMATEX 2010

Microscopy: Science, Technology, Applications and Education A. Mndez-Vilas and J. Daz (Eds.) ______________________________________________

To capture a well defined ECP, imaging is performed at the lowest possible magnification at slow scan rates (0.5-3 minutes/frame). The beam can be focused at the sample surface or defocused with the crossover point far below the specimen in order to decrease the beam convergence and sharpen ECP line features. Although this provides a feature rich ECP, there might be a slight rotation of the pattern relative to that the ECP obtained when focused to the sample surface. Hence for orientation purposes, such as isolating diffraction conditions, ECPs should be generated when focused to the specimen surface. A comprehensive review on mapping ECPs has been published by Joy et al [12]. If certain ECPs are difficult to index, a corresponding EBSD pattern simulated and indexed by modern commercial software can be related back to the ECP. Any reliable kinematic computation of EBSD/ECP line positions should allow identification of ECP lines in the experimental pattern. Generally, the optimal contrast for defect imaging occurs at areas where the ECP shows a strong dark-light contrast (eg major channeling band). After such a position is centered along the SEM optic axis by virtue of tilt and rotation, obtaining a high magnification image will allow the imaging of dislocations that are visible in the obtained diffraction condition. In GaN, surface penetrating threading dislocations appear as dark-light spots. As predicted by the dynamical theory of channeling contrast by Spencer et al [6] in 1972, it is observed that lower index diffraction conditions give sharper contrast than higher index diffraction conditions (Figure 4). Screw dislocations in GaN generally have a Burgers vector of c<0001>. But since these dislocations penetrate the surface, surface relaxation effects allow these dislocations to be visible even under diffraction conditions that would satisfy the g b = 0 invisibility condition. Since there are two types of Burgers vectors that are 180 with respect to each other, c[0001] and c[000-1], it is not surprising that there are only two kinds of spot features [19-20] (with an 180 rotation of bright-dark line of contrast) in each ECCI micrograph in Figure 4. The figure also shows that by selecting diffraction vectors 90 with respect to each other, the bright-dark line of transition rotates by 90. Thus, the acting diffraction vector, g, determines the bright-dark contrast directionality of each individual screw dislocation [19], consistent with diffraction contrast behavior for viewing screw dislocations imaged end-on by plan-view TEM [23].

Fig. 4 ECCI micrographs of GaN (0002) under different diffraction conditions: (a) g = [11-20] (b) g = [10-10].

The strength of the strain field induced by the defect will depend on the magnitude of the Burgers vector. GaN threading edge dislocations are known to have smaller magnitude Burgers vector than threading screw dislocations, and are thus expected to appear spatially smaller [20] than screw dislocations by ECCI (Figure 5). Note that the directionality of dark-light contrast occurs at a perpendicular angle when comparing the stronger intensity fluctuations (screw) and the weaker intensity fluctuations (edge), a consequence of their respective Burgers vectors being perpendicular to each other. (Threading edge dislocations in GaN are known to have a Burgers vector of a/3<11-20>).

Fig. 5: Arrays of edge dislocations in GaN (0002) appear as smaller dark/bright spots than screw dislocations due to smaller Burgers vector.

FORMATEX 2010

1587

Microscopy: Science, Technology, Applications and Education A. Mndez-Vilas and J. Daz (Eds.) ______________________________________________

In a SrTiO3 (001) substrate sample, g b = 0 invisibility criterion can be demonstrated since there are non-surface penetrating dislocation with no surface relaxation effects. By accessing sufficient diffraction conditions, it becomes possible to determine Burgers vectors of sub-surface dislocations. In Figure 6, dislocation line segments visible under g = 100 condition are denoted by black rectangles which show no contrast under g = 010. Dislocation segments visible in g = 010 condition are denoted by white rectangles and show no contrast under g = 100.

Fig. 6: ECCI micrographs of SrTiO3 (001) recorded in a backscatter geometry with (a) g = 100 and (b) g = 010. White rectangles denote dislocation lines visible in (a) and invisible in (b). Black rectangles denote dislocation lines visible in (b) and invisible in (a).

Apart from individual dislocations, we can use ECCI to observe interacting pairs of planar and threading dislocations that have given rise to Frank-Read type dislocation structures, dislocation arrays associated with prismatic slip, and arrays of sub-surface basal plane dislocations that stretch over a large areas in Figure 7 (b), (c) and (a) respectively.

Fig. 7: ECCI micrographs of (a) an array of basal plane dislocations in a 300nm GaN (0002) film grown on a GaN substrate, (b) Frank-read source imaged at 15kV in a HVPE grown GaN (0002) substrate, and (c) dislocation arrays associated with prismatic slip induced by mechanical stress in a SrTiO3 (001) substrate.

5. Discussion and Summary

Although not widely applied over the decades as a means for directly imaging crystal defects, modern SEMs with increased source brightness and higher BSE collection efficiency detectors have made ECCI more accessible. ECCI can be used to predict the directionality of the Burgers vector by the angle of dark-light contrast seen in dislocations and relating it to the diffraction condition being used. While all surface penetrating dislocations can be imaged under all diffraction condition due to surface relaxation, the Burgers vector of sub-surface dislocations can readily be resolved

1588

FORMATEX 2010

Microscopy: Science, Technology, Applications and Education A. Mndez-Vilas and J. Daz (Eds.) ______________________________________________

through application of the invisibility criterion. If the magnitude of the Burgers vector is different for various dislocations types (eg screw vs. edge dislocations in GaN) they can be distinguished by spatial size of the contrast features observed by ECCI. This not only makes the observation of fundamental dislocation interaction phenomena (i.e. Frank-Read sources, dislocation arrays) possible through ECCI, but also gives the research community an easier capacity to coordinate diffraction based analysis of ECCI to other methods for more robust analysis of materials and devices than offered by TEM. Compared to the limitations of executing diffraction contrast in TEM, ECCI provides a versatile approach to carry out dislocation analysis in the SEM. The non-destructive nature of ECCI will greatly benefit failure analysis since dislocations can be imaged without destroying or modifying the specimen. Particular instances where ductile failure leaves a smooth fracture surface for observation, ECCI can potentially be applied to distinguish the nature of dislocations that contributed to the failure. For optoelectronic materials such as thin films of GaN or solar cell materials such as multicrystalline silicon, where dislocation density plays a crucial role on influencing device performance, ECCI can be used to dynamically characterize dislocations between growth processes or post device processing. Novel memory storage materials that exhibit a reliance on dislocation presence or interactions would undoubtedly benefit from ECCI, such as resistive switching materials like SrTiO3 or other ferroelectric and piezoelectric materials. ECCI under in-situ conditions with applied stress, electric and/or thermal fields are a probable and highly unique opportunity for SEM-based investigations of dislocation behavior in bulk materials. By using specialized SEMs with rocking beam capability or stage rocking capabilities, it will be possible to carry out the same studies in polycrystalline materials. Future advances in SEM resolution, detection, and beam rocking will only further advance opportunities for applying ECCI for non-destructive and comprehensive dislocation analysis.
Acknowledgements We would like to acknowledge Fang Liu, Wenkan Jiang, Prof. Marek Skowronski, Prof. Paul Salvador, Prof. Lisa Porter, and Prof. Robert Davis of Carnegie Mellon University for providing GaN and SrTiO3 specimens. We would also like to thank Daniel Koleske of Sandia National Labs, Greg Mulholland and Bob Metzger of Kyma Technologies, and Kenneth Jones of the Army Research Lab for providing GaN specimens used in this study.

References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] Coates DG. Kikuchi-like reflection patterns obtained with scanning electron microscope. Philosophical Magazine, 1967; 16(144): 1179-1185. Booker GR, et al. Some comments on interpretation of Kikuchi-like reflection patterns observed by scanning electron microscopy. Philosophical Magazine, 1967;16(144) 1185-1191 Clarke DR. Observation of crystal defects using scanning electron microscope. Philosophical Magazine, 1971;24(190): 973979. Stern RM, et al. Dislocation images in high-resolution scanning electron-microscope. Philosophical Magazine, 1972;26(6):1495-1499. Clarke DR and Howie A. Calculations of lattice defect images for scanning electron microscopy. Philosophical Magazine, 1971;24(190): 959-971. Spencer JP, Hirsch PB, and Humphrey CJ. Dynamical theory for contrast of perfect and imperfect crystals in scanning electronmicroscope using backscattered electrons. Philosophical Magazine, 1972; 26(1):193-213. Sandstro R, Spencer JF, and Humphrey CJ, Theoretical model for energy-dependence of electron channeling patterns in scanning electron-microscopy. Journal of Physics D-Applied Physics, 1974;7(7):1030-1046. Pitaval M, et al. Energy filtering of backscattered electrons for SEM. Journal De Microscopie Et De Spectroscopie Electroniques, 1977; 2(2):185-196. Morin P, et al. Electron-channeling imaging in scanning electron-microscopy. Philosophical Magazine A- 1979; 40(4):511524. Simkin BA and Crimp MA. An experimentally convenient configuration for electron channeling contrast imaging. Ultramicroscopy, 1999;77(1-2):65-75. Spencer JP and Humphreys CJ, A multiple-scattering transport-theory for electron channeling patterns. Philosophical Magazine a 1980; 42(4):433-451. Joy DC, Newbury DE, and Davidson DL, Electron channeling patterns in the scanning electron-microscope. Journal of Applied Physics, 1982;53(8):R81-R122. Dudarev SL, Rez P, and Whelan MJ, Theory of electron backscattering from crystals. Physical Review B, 1995;51:3397-3405. Czernuszka JT, et al. Imaging of dislocations using backscattered electrons in a scanning electron-microscope. Philosophical Magazine Letters, 1990;62(4):227-232. Wilkinson AJ, et al. Electron channeling contrast imaging of interfacial defects in strained silicon-germanium layers on silicon. Philosophical Magazine A, 1993;68(1):59-80. Ahmed J, Wilkinson AJ, and Roberts SG, Characterizing dislocation structures in bulk fatigued copper single crystals using electron channelling contrast imaging (ECCI). Philosophical Magazine Letters, 1997;76(4):237-245. Dudarev SL, et al. Decoherence in electron backscattering by kinked dislocations. Acta Crystallographica Section A, 1999; 55(1):234-245. Trager-Cowan C, et al. Electron backscatter diffraction and electron channeling contrast imaging of tilt and dislocations in nitride thin films. Physical Review B, 2007;75(8).

FORMATEX 2010

1589

Microscopy: Science, Technology, Applications and Education A. Mndez-Vilas and J. Daz (Eds.) ______________________________________________

[19] Picard YN, et al. Diffraction contrast and Bragg reflection determination in forescattered electron channeling contrast images of threading screw dislocations in 4H-SiC. Journal of Applied Physics, 2008;104:124906. [20] Picard YN, et al. Resolving the Burgers vector for individual GaN dislocations by electron channeling contrast imaging. Scripta Materialia, 2009;61:773-776. [21] Twigg ME and Picard YN. Simulation and analysis of electron channeling contrast images of threading screw dislocations in 4H-SiC. Journal of Applied Physics, 2009;105:093520. [22] Winkelmann A, et al. Many-beam dynamical simulation of electron backscatter diffraction patterns. Ultramicroscopy, 2007;107(4-5):414-421. [23] Tunstall WJ, Hirsch PB, Steeds J. Effect of surface stress relaxation on the electron microscope images of dislocations normal to thin metal foils. Philosophical Magazine, 1964; 9:99.

1590

FORMATEX 2010