Electron Backscatter Diffractionin Materials Science

Second Edition

Adam J. Schwartz · Mukul Kumar ·Brent L. Adams · David P. FieldEditors

Electron BackscatterDiffraction inMaterials Science

Second Edition

123

EditorsAdam J. SchwartzLawrence Livermore National LaboratoryPhysical and Life Sciences Directorate7000 East AvenueLivermore CA 94550USAschwartz6@llnl.gov

Brent L. AdamsDepartment of Mechanical EngineeringBrigham Young UniversityProvo UT 84602455B Crabtree Technology BuildingUSAb l adams@byu.edu

Mukul KumarLawrence Livermore National LaboratoryPhysical and Life Sciences Directorate7000 East AvenueLivermore CA 94550USAkumar3@llnl.gov

David P. FieldSchool of Mechanical and MaterialsEngineeringWashington State UniversityPullman WA 99164-2920Dana 239EUSAdfield@wsu.edu

First hard cover printing 2000, Kluwer Academic / Plenum Publishers

ISBN 978-0-387-88135-5 e-ISBN 978-0-387-88136-2DOI 10.1007/978-0-387-88136-2

Library of Congress Control Number: 2009920955

c© Springer Science+Business Media, LLC 2009All rights reserved. This work may not be translated or copied in whole or in part without the writtenpermission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connectionwith any form of information storage and retrieval, electronic adaptation, computer software, or by similaror dissimilar methodology now known or hereafter developed is forbidden.The use in this publication of trade names, trademarks, service marks, and similar terms, even if they arenot identified as such, is not to be taken as an expression of opinion as to whether or not they are subjectto proprietary rights.

Printed on acid-free paper

springer.com

Contents

1 Present State of Electron Backscatter Diffractionand Prospective Developments . . . . . . . . . . . . . . . . . . . . . 1Robert A. Schwarzer, David P. Field, Brent L. Adams,Mukul Kumar, and Adam J. Schwartz1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Generation and Interpretation of Electron Backscatter

Diffraction Patterns . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Experimental Set-Up of an EBSD System . . . . . . . . . . . . . 31.4 The Components of an Automated EBSD System . . . . . . . . . 4

1.4.1 The Pattern Acquisition Device . . . . . . . . . . . . . . 41.4.2 Mechanical Stage and Digital

Beam Scanning . . . . . . . . . . . . . . . . . . . . . . . 51.5 Spatial Resolution . . . . . . . . . . . . . . . . . . . . . . . . . 71.6 SEM Specifications for Good EBSD Performance . . . . . . . . . 91.7 The Radon or Hough Transformation for Band Localization . . . 111.8 Indexing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121.9 Fast EBSD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131.10 Ion Blocking Patterns . . . . . . . . . . . . . . . . . . . . . . . . 151.11 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2 Dynamical Simulation of Electron BackscatterDiffraction Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Aimo Winkelmann2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.2 Model of Electron Backscatter Diffraction . . . . . . . . . . . . . 212.3 Dynamical Electron Diffraction in EBSD . . . . . . . . . . . . . 22

2.3.1 Dynamical Electron Diffraction in EBSD . . . . . . . . . 222.3.2 Dynamical Electron Diffraction in EBSD . . . . . . . . . 232.3.3 Dynamical Electron Diffraction in EBSD . . . . . . . . . 24

2.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252.4.1 A Real-Space View of EBSD . . . . . . . . . . . . . . . 252.4.2 Full Scale Simulation of EBSD Patterns . . . . . . . . . 272.4.3 The Influence of the Energy Spectrum of the

Backscattered Electrons . . . . . . . . . . . . . . . . . . 282.4.4 Dynamical Effects of Anisotropic Backscattering . . . . . 30

2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

v

vi Contents

3 Representations of Texture . . . . . . . . . . . . . . . . . . . . . . . 35Jeremy K. Mason and Christopher A. Schuh3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353.2 Rotations and Orientations . . . . . . . . . . . . . . . . . . . . . 36

3.2.1 Defining a Rotation . . . . . . . . . . . . . . . . . . . . 363.2.2 Defining an Orientation . . . . . . . . . . . . . . . . . . 37

3.3 Pole Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383.4 Discrete Orientations . . . . . . . . . . . . . . . . . . . . . . . . 40

3.4.1 Axis-Angle Parameters . . . . . . . . . . . . . . . . . . 413.4.2 Rodrigues Vectors . . . . . . . . . . . . . . . . . . . . . 423.4.3 Quaternions . . . . . . . . . . . . . . . . . . . . . . . . 423.4.4 Euler Angles . . . . . . . . . . . . . . . . . . . . . . . . 45

3.5 Orientation Distribution Functions . . . . . . . . . . . . . . . . . 463.5.1 Circular Harmonics . . . . . . . . . . . . . . . . . . . . 463.5.2 Spherical Harmonics . . . . . . . . . . . . . . . . . . . . 473.5.3 Hyperspherical Harmonics . . . . . . . . . . . . . . . . . 483.5.4 Generalized Spherical Harmonics . . . . . . . . . . . . . 493.5.5 Symmetrized Harmonics . . . . . . . . . . . . . . . . . . 49

3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

4 Energy Filtering in EBSD . . . . . . . . . . . . . . . . . . . . . . . 53Alwyn Eades, Andrew Deal, Abhishek Bhattacharyya,and Tejpal Hooghan4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534.3 Energy Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544.4 Operating the Filter . . . . . . . . . . . . . . . . . . . . . . . . . 564.5 Early Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 574.6 Patterns at Different Energies . . . . . . . . . . . . . . . . . . . 604.7 Localization of the Signal . . . . . . . . . . . . . . . . . . . . . 614.8 Future Energy Filters in EBSD . . . . . . . . . . . . . . . . . . . 624.9 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . 62

5 Spherical Kikuchi Maps and Other Rarities . . . . . . . . . . . . . 65Austin P. Day5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.2 Electron Backscatter Patterns . . . . . . . . . . . . . . . . . . . 655.3 Spherical Kikuchi Maps . . . . . . . . . . . . . . . . . . . . . . 655.4 EBSP Detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.5 EBSP Imaging and Uniformity . . . . . . . . . . . . . . . . . . . 685.6 EBSP Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . 685.7 Spherical Kikuchi Maps from EBSPs . . . . . . . . . . . . . . . 685.8 Kikuchi Band Profiles . . . . . . . . . . . . . . . . . . . . . . . 725.9 Spherical Kikuchi Map Inversion . . . . . . . . . . . . . . . . . 745.10 Uses for Spherical Kikuchi Maps . . . . . . . . . . . . . . . . . 755.11 Colour Orientation Contrast Images . . . . . . . . . . . . . . . . 765.12 STEM in the SEM . . . . . . . . . . . . . . . . . . . . . . . . . 765.13 Unusual Features in EBSPs . . . . . . . . . . . . . . . . . . . . 77

Contents vii

6 Application of Electron Backscatter Diffractionto Phase Identification . . . . . . . . . . . . . . . . . . . . . . . . . 81Bassem El-Dasher and Andrew Deal6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 816.2 Considerations for Phase ID with EBSD . . . . . . . . . . . . . . 826.3 Case Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

6.3.1 Simultaneous EBSD/EDS Phase Discrimination . . . . . 856.3.2 Distinguishing � and �′ in Ni Superalloys . . . . . . . . . 866.3.3 Volume Fraction Determination in a Multiphase Alloy . . 89

7 Phase Identification Through Symmetry Determinationin EBSD Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97David J. Dingley and S.I. Wright7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 977.2 Basis of the Phase Identification Method . . . . . . . . . . . . . . 977.3 Determination of the Crystal Unit Cell . . . . . . . . . . . . . . . 987.4 Discovering the Lattice Symmetry . . . . . . . . . . . . . . . . . 1007.5 Re-Indexing the Pattern According to the Discovered

Crystal Class . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1017.6 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

7.6.1 Case 1, A Cubic Crystal . . . . . . . . . . . . . . . . . . 1027.6.2 Case 2, A Hexagonal Crystal . . . . . . . . . . . . . . . 1047.6.3 Case 3, A Trigonal Crystal . . . . . . . . . . . . . . . . . 104

7.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

8 Three-Dimensional Orientation Microscopy by SerialSectioning and EBSD-Based Orientation Mappingin a FIB-SEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109Stefan Zaefferer and Stuart I. Wright8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1098.2 The Geometrical Set-Up for 3D Characterisation in a FIB-SEM . 1108.3 Automatic 3D Orientation Microscopy . . . . . . . . . . . . . . 1138.4 Software for 3D Data Analysis . . . . . . . . . . . . . . . . . . . 1138.5 Application Examples . . . . . . . . . . . . . . . . . . . . . . . 114

8.5.1 The 3D Microstructure and Crystallography ofPearlite Colonies . . . . . . . . . . . . . . . . . . . . . . 114

8.5.2 Microstructure of “Nanocrystalline” NiCo Deposits . . . 1158.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

8.6.1 Accuracy and Application Limits . . . . . . . . . . . . . 1198.6.2 Materials Issues . . . . . . . . . . . . . . . . . . . . . . 120

8.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

viii Contents

9 Collection, Processing, and Analysis of Three-DimensionalEBSD Data Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Michael A. Groeber, David J. Rowenhorst,and Michael D. Uchic9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1239.2 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . 1239.3 Processing Strategies . . . . . . . . . . . . . . . . . . . . . . . . 124

9.3.1 Registration and Alignment of Sections . . . . . . . . . . 1249.3.2 Segmentation of Grains . . . . . . . . . . . . . . . . . . 1269.3.3 Clean-Up Routines . . . . . . . . . . . . . . . . . . . . . 127

9.4 Analysis Capabilities . . . . . . . . . . . . . . . . . . . . . . . . 1299.4.1 Morphological Descriptors . . . . . . . . . . . . . . . . 1299.4.2 Crystallographic Descriptors . . . . . . . . . . . . . . . 133

9.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

10 3D Reconstruction of Digital Microstructures . . . . . . . . . . . . 139Stephen D. Sintay, Michael A. Groeber, and Anthony D. Rollett10.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13910.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

10.2.1 2D–3D Inference . . . . . . . . . . . . . . . . . . . . . . 13910.2.2 3D Polycrystal Microstructure Generation . . . . . . . . 140

10.3 Data Collection and Analysis . . . . . . . . . . . . . . . . . . . . 14010.3.1 Data Sources . . . . . . . . . . . . . . . . . . . . . . . . 14010.3.2 Identifying Features . . . . . . . . . . . . . . . . . . . . 14110.3.3 Statistical Description of Features . . . . . . . . . . . . . 141

10.4 Methods for 3D Structure Inference . . . . . . . . . . . . . . . . 14110.4.1 Monte Carlo-Based Histogram Fitting . . . . . . . . . . 14310.4.2 Observation-Based Domain Constraint . . . . . . . . . . 145

10.5 Generation of 3D Structure . . . . . . . . . . . . . . . . . . . . . 14710.5.1 Packing of Ellipsoids . . . . . . . . . . . . . . . . . . . 14710.5.2 Relaxation of Boundaries . . . . . . . . . . . . . . . . . 149

10.6 Quality Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 14910.6.1 Size Distribution Comparison . . . . . . . . . . . . . . . 14910.6.2 Shape Distribution Comparison . . . . . . . . . . . . . . 14910.6.3 Neighborhood Comparison . . . . . . . . . . . . . . . . 15110.6.4 Boundary Structure Comparison . . . . . . . . . . . . . . 151

10.7 Thoughts on Current Conditions and Future Work . . . . . . . . . 151

11 Direct 3D Simulation of Plastic Flow from EBSD Data . . . . . . . 155Nathan R. Barton, Joel V. Bernier, Ricardo A. Lebensohn,and Anthony D. Rollett11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15511.2 Material and Microstructural Model . . . . . . . . . . . . . . . . 156

11.2.1 Three-Dimensional Microstructure Generation . . . . . . 15711.2.2 Micromechanical Model . . . . . . . . . . . . . . . . . . 15811.2.3 Finite Element Model . . . . . . . . . . . . . . . . . . . 159

11.3 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . 15911.4 Directions for Further Computational Development . . . . . . . . 16211.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

Contents ix

12 First-Order Microstructure Sensitive Design Basedon Volume Fractions and Elementary Bounds . . . . . . . . . . . . 169Surya R. Kalidindi, David T. Fullwood, and Brent L. Adams12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16912.2 Quantification of Microstructure . . . . . . . . . . . . . . . . . . 17012.3 Microstructure Sensitive Design Framework . . . . . . . . . . . . 17012.4 Property Closures . . . . . . . . . . . . . . . . . . . . . . . . . . 172

13 Second-Order Microstructure Sensitive DesignUsing 2-Point Spatial Correlations . . . . . . . . . . . . . . . . . . . 177David T. Fullwood, Surya R. Kalidindi, and Brent L. Adams13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17713.2 Definition and Properties of the 2-Point Correlation Functions . . 178

13.2.1 Boundary Conditions . . . . . . . . . . . . . . . . . . . 17913.2.2 Properties of the 2-Point Functions . . . . . . . . . . . . 17913.2.3 Visualization of the 2-Point Functions . . . . . . . . . . . 17913.2.4 Metrics from 2-Point Correlations . . . . . . . . . . . . . 18013.2.5 Collecting 2-Point Correlations from Material Samples . 180

13.3 Structure Property Relations . . . . . . . . . . . . . . . . . . . . 18113.3.1 Localization Tensors . . . . . . . . . . . . . . . . . . . . 18213.3.2 Effective Tensors . . . . . . . . . . . . . . . . . . . . . . 184

13.4 Microstructure Design . . . . . . . . . . . . . . . . . . . . . . . 186

14 Combinatorial Materials Science and EBSD: A HighThroughput Experimentation Tool . . . . . . . . . . . . . . . . . . 189Krishna Rajan14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18914.2 Introduction to Combinatorial Methods . . . . . . . . . . . . . . 189

14.2.1 High Throughput EBSD Screening . . . . . . . . . . . . 19014.2.2 Informatics and Data . . . . . . . . . . . . . . . . . . . . 194

14.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

15 Grain Boundary Networks . . . . . . . . . . . . . . . . . . . . . . . 201Bryan W. Reed and Christopher A. Schuh15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20115.2 Measurement and Classification of Local Network Elements . . . 202

15.2.1 General Definitions for Single Boundaries . . . . . . . . 20215.2.2 Structures with More than One Boundary . . . . . . . . . 203

15.3 Geometry of the Network Structure . . . . . . . . . . . . . . . . 20415.3.1 Percolation Measures of the Grain Boundary Network . . 20515.3.2 Crystallographic Constraints . . . . . . . . . . . . . . . . 206

15.4 Microstructure-Property Connections . . . . . . . . . . . . . . . 20815.4.1 Composite Averaging vs. Percolation Theory . . . . . . . 20915.4.2 Crystallographic Correlations . . . . . . . . . . . . . . . 211

15.5 Conclusions and Future Outlook . . . . . . . . . . . . . . . . . . 212

x Contents

16 Measurement of the Five-Parameter Grain BoundaryDistribution from Planar Sections . . . . . . . . . . . . . . . . . . . 215Gregory S. Rohrer and Valerie Randle16.1 Introduction: Grain Boundary Planes and Properties . . . . . . . 21516.2 Serial Sectioning . . . . . . . . . . . . . . . . . . . . . . . . . . 21616.3 Single-Surface Trace Analysis . . . . . . . . . . . . . . . . . . . 21716.4 Five-Parameter Stereological Analysis . . . . . . . . . . . . . . . 218

16.4.1 Parameterization and Discretization of the Spaceof Grain Boundary Types . . . . . . . . . . . . . . . . . . 218

16.4.2 Measurement of the Grain BoundaryCharacterization Distribution . . . . . . . . . . . . . . . . 219

16.4.3 Performance of the Stereological Analysis . . . . . . . . 22116.4.4 Comparison GBCDs Measured Stereologically

and by Serial Sectioning in the Dual Beam FIB . . . . . . 22316.5 Examples of Five-Parameter Analyses . . . . . . . . . . . . . . . 224

17 Strain Mapping Using Electron Backscatter Diffraction . . . . . . . 231Angus J. Wilkinson, David J. Dingley, and Graham Meaden17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231

17.1.1 The Need for Local Strain Assessment . . . . . . . . . . 23117.1.2 Competing Strain Mapping Techniques . . . . . . . . . . 23117.1.3 Review of Applications of EBSD

to Analysis of Elastic Strains . . . . . . . . . . . . . . . . 23217.2 Cross-Correlation-Based Analysis

of EBSD Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . 23417.2.1 Geometry: Linking Pattern Shifts to Strain . . . . . . . . 23417.2.2 Pattern Shift Measurement . . . . . . . . . . . . . . . . . 23517.2.3 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . 23717.2.4 Illustrative Applications . . . . . . . . . . . . . . . . . . 239

17.3 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . 247

18 Mapping and Assessing Plastic Deformation Using EBSD . . . . . . 251Luke N. Brewer, David P. Field, and Colin C. Merriman18.1 Plastic Deformation Effects on the EBSD Pattern and

Orientation Map . . . . . . . . . . . . . . . . . . . . . . . . . . 25118.2 Pattern Rotation Approaches . . . . . . . . . . . . . . . . . . . . 253

18.2.1 Mapping Orientations and Misorientations . . . . . . . . 25318.2.2 Average Misorientation Approaches . . . . . . . . . . . . 25518.2.3 Measurement and Calculation

of GND Densities . . . . . . . . . . . . . . . . . . . . . . 258

19 Analysis of Deformation Structures in FCC MaterialsUsing EBSD and TEM Techniques . . . . . . . . . . . . . . . . . . . 263Oleg V. Mishin, Andrew Godfrey, and Dorte Juul Jensen19.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26319.2 Orientation Noise in EBSD Data . . . . . . . . . . . . . . . . . . 265

19.2.1 A Quantitative Description of Orientation Noise . . . . . 26519.2.2 Postprocessing Orientation Filtering Operations . . . . . 266

19.3 Quantitative TEM–EBSD Comparison . . . . . . . . . . . . . . . 26819.4 Heterogeneity in Microstructural Refinement . . . . . . . . . . . 271

Contents xi

19.4.1 Analysis of Local Heterogeneity . . . . . . . . . . . . . 27119.4.2 Potential for Analysis of Large-Scale Heterogeneities . . 272

19.5 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . 273

20 Application of EBSD Methods to Severe PlasticDeformation (SPD) and Related Processing Methods . . . . . . . . 277Terry R. McNelley, Alexandre P. Zhilyaev, SrinivasanSwaminathan, Jianqing Su, and E. Sarath Menon20.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27720.2 Microstructures During the Initial ECAP Pass . . . . . . . . . . . 27820.3 Microstructures Developed by Machining . . . . . . . . . . . . . 28220.4 Grain Refinement During FSP . . . . . . . . . . . . . . . . . . . 28420.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288

21 Applications of EBSD to Microstructural Control inFriction Stir Welding/Processing . . . . . . . . . . . . . . . . . . . . 291Sergey Mironov, Yutaka S. Sato, and Hiroyuki Kokawa21.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29121.2 Brief Explanations of FSW/P Terminology . . . . . . . . . . . . 29221.3 Microstructural Evolution . . . . . . . . . . . . . . . . . . . . . 29221.4 Material Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29621.5 Structure-Properties Relationship . . . . . . . . . . . . . . . . . 29821.6 Summary and Future Outlook . . . . . . . . . . . . . . . . . . . 299

22 Characterization of Shear Localization and Shock Damagewith EBSD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301John F. Bingert, Veronica Livescu, and Ellen K. Cerreta22.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30122.2 Shear Localization . . . . . . . . . . . . . . . . . . . . . . . . . 302

22.2.1 Constrained Shear in Pure Fe—Shear Zone Geometry . . 30222.2.2 Constrained Shear in Pure Fe—Texture Development . . 30622.2.3 Effect of Morphology on Grain Instability in Cu . . . . . 307

22.3 Shock Loading Damage in Tantalum . . . . . . . . . . . . . . . . 30922.3.1 Effect of Shock Duration on Incipient Spall Structure . . 31022.3.2 Effect of Pressure on Incipient Spall Structure . . . . . . 313

22.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

23 Texture Separation for �/� Titanium Alloys . . . . . . . . . . . . . 317Ayman A. Salem23.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31723.2 Microstructure of �/� Titanium Alloys . . . . . . . . . . . . . . 31723.3 Texture of Ti-6Al-4V . . . . . . . . . . . . . . . . . . . . . . . . 318

23.3.1 Separation of Primary and Secondary Alpha Texture . . . 31923.3.2 EBSD + BSE Imaging Technique . . . . . . . . . . . . . 31923.3.3 EBSD or XRD + Heat Treatment Technique . . . . . . . 320

23.4 Texture Separation Using EBSD + EDS Technique . . . . . . . . 32023.4.1 Texture Separation Using EBSD + EDS Technique . . . . 32023.4.2 Microstructure Observations . . . . . . . . . . . . . . . . 32123.4.3 Chemical Composition Maps (EDS) . . . . . . . . . . . 321

23.5 Industrial Application: Controlling Texture DuringHot-Rolling of Ti-6Al-4V . . . . . . . . . . . . . . . . . . . . . 322

xii Contents

23.5.1 Microstructure Evolution . . . . . . . . . . . . . . . . . 32323.5.2 Overall Texture Evolution . . . . . . . . . . . . . . . . . 32323.5.3 Primary-Alpha (�p) Textures . . . . . . . . . . . . . . . 32423.5.4 Secondary-Alpha (�s) Texture . . . . . . . . . . . . . . . 325

23.6 Industrial Application: Controlling Texture DuringHot-Rolling of Ti-6Al-4V . . . . . . . . . . . . . . . . . . . . . 326

24 A Review of In Situ EBSD Studies . . . . . . . . . . . . . . . . . . . 329Stuart I. Wright and Matthew M. Nowell24.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32924.2 In Situ Postmortem Experiments . . . . . . . . . . . . . . . . . . 33024.3 Deformation Stage Experiments . . . . . . . . . . . . . . . . . . 33124.4 Heating Stage Experiments . . . . . . . . . . . . . . . . . . . . . 332

24.4.1 Phase Transformation . . . . . . . . . . . . . . . . . . . 33224.4.2 Recrystallization and Grain Growth . . . . . . . . . . . . 333

24.5 Combined Heating and Tensile Stage Experiments . . . . . . . . 33524.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335

25 Electron Backscatter Diffraction in Low Vacuum Conditions . . . . 339Bassem S. El-Dasher and Sharon G. Torres25.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33925.2 Considerations for Low Vacuum EBSD . . . . . . . . . . . . . . 34025.3 Example Applications . . . . . . . . . . . . . . . . . . . . . . . 341

25.3.1 Microstructural Analysisof AlN-TiB2 Ceramic Composite . . . . . . . . . . . . . 341

25.3.2 Characterization of CaHPO4·2H2O Single Crystals . . . . 342

26 EBSD in the Earth Sciences: Applications, CommonPractice, and Challenges . . . . . . . . . . . . . . . . . . . . . . . . 345David J. Prior, Elisabetta Mariani, and John Wheeler26.1 Development of EBSD in Earth Sciences . . . . . . . . . . . . . 34526.2 Current Practice, Capabilities, and Limitations . . . . . . . . . . 346

26.2.1 Range of Materials and Preparation . . . . . . . . . . . . 34626.2.2 Speed of Data Collection . . . . . . . . . . . . . . . . . 34726.2.3 Spatial Resolution . . . . . . . . . . . . . . . . . . . . . 34726.2.4 Misindexing . . . . . . . . . . . . . . . . . . . . . . . . 34826.2.5 Polyphase Samples . . . . . . . . . . . . . . . . . . . . 350

26.3 Application of EBSD in Earth Sciences . . . . . . . . . . . . . . 35126.3.1 Rock Deformation and Solid Earth Geophysics . . . . . . 35226.3.2 Metamorphic Processes . . . . . . . . . . . . . . . . . . 35526.3.3 Meteorites . . . . . . . . . . . . . . . . . . . . . . . . . 35626.3.4 Other Areas . . . . . . . . . . . . . . . . . . . . . . . . 356

26.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357

Contents xiii

27 Orientation Imaging Microscopy in Researchon High Temperature Oxidation . . . . . . . . . . . . . . . . . . . . 361Bae-Kyun Kim and Jerzy A. Szpunar27.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36127.2 High Temperature Oxidation . . . . . . . . . . . . . . . . . . . . 36227.3 Experimental Procedure . . . . . . . . . . . . . . . . . . . . . . 363

27.3.1 Oxidation of Samples and Oxide Formation . . . . . . . 36327.3.2 Sample Preparation and Geometry in OIM . . . . . . . . 36427.3.3 Microstructure and Texture Measurement . . . . . . . . . 36527.3.4 Oxidation of Low Carbon Steel . . . . . . . . . . . . . . 365

27.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . 36827.4.1 Grain Growth in Iron Oxide . . . . . . . . . . . . . . . . 36827.4.2 Effect of the Oxidation Process on Microstructure . . . . 37127.4.3 Oxidation of Pure Iron . . . . . . . . . . . . . . . . . . . 373

27.5 Cracks and Defects . . . . . . . . . . . . . . . . . . . . . . . . . 38427.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395

Contributors

Brent L. Adams Department of Mechanical Engineering, Brigham YoungUniversity, 455B CTB, Provo, UT 84602-4201, USA, b l adams@byu.edu

Nathan R. Barton Lawrence Livermore National Laboratory, L-129, 7000 EastAvenue, Livermore, CA 94550, USA, barton22@llnl.gov

Joel V. Bernier Lawrence Livermore National Laboratory, L-129, 7000 EastAvenue, Livermore, CA 94550, USA, bernier2@llnl.gov

Abhishek Bhattacharyya 1 Becton Drive, Franklin Lakes, NJ 07417 USA,abhatt72@yahoo.com

John F. Bingert Materials Science and Technology Division, Los Alamos NationalLaboratory, MST-8, MS G755, Los Alamos, NM 87545, USA, bingert@lanl.gov

Luke N. Brewer Sandia National Laboratories, New Mexico, PO Box 5800 MS1411, Albuquerque, NM 87123-1411, USA, lnbrewe@sandia.gov

Ellen K. Cerreta Los Alamos National Laboratory, MST-8, MS G755, Los Alamos,NM 87545, USA, ecerreta@lanl.gov

Austin P. Day Aunt Daisy Scientific Ltd., Durlow, Dixton Road, Monmount,Gwent NP25 3PP; KE Developments, The Mount, Toft, Cambridge CB23 2RL,United Kingdom, auntdaisy@btinternet.com

Andrew Deal GE Global Research, One Research Circle, Niskayuna, NY 12309,USA, deal@research.ge.com

David J. Dingley H. H. Wills Physics Laboratory, Bristol University, TyndallAvenue, Bristol BS8 1TL, United Kingdom, djdingley@hotmail.com

Alwyn Eades Department of Materials Science and Engineering, Lehigh University,5 East Packer Avenue, Bethlehem, PA 18015-3195, USA, jae5@lehigh.edu

Bassem S. El-Dasher Lawrence Livermore National Laboratory, L-367, 7000 EastAvenue, Livermore, CA 94550, USA, eldasher2@llnl.gov

David P. Field School of Mechanical and Materials Engineering, Washington StateUniversity, Dana 239E, Pullman, WA 99164-2920, USA, dfield@wsu.edu

David T. Fullwood Department of Mechanical Engineering, Brigham YoungUniversity, 435 CTB, Provo, UT 84602-4201, USA, dfullwood@byu.edu

xv

xvi Contributors

Andy Godfrey Laboratory of Advanced Materials, Department of Materi-als Science and Engineering, Tsinghua University, Beijing 100084, China,awgodfrey@mail.tsinghua.edu.cn

Michael A. Groeber Wright Patterson Air Force Base, 2210 Tenth St. Building 655Room 110, WPAFB, OH 45433, USA, michael.groeber.ctr@wpafb.af.mil

Tejpal Hooghan Texas Instruments Incorporated, 13536 North Central Expressway,Dallas, TX 75243, USA, thooghan@ti.com

Dorte Juul Jensen Risø National Laboratory for Sustainable Energy, MaterialsResearch Department, Center for Fundamental Research: Metal Structures in FourDimensions, Technical University of Denmark, Building 228, P.O. Box 49, DK-4000Roskilde, Denmark, dorte.juul.jensen@risoe.dk, www.risoe.dk

Surya R. Kalidindi Department of Materials Science and Engineering, DrexelUniversity, LeBow 346, 3141 Chestnut Street, Philadelphia, PA 19104, USA,skalidin@coe.drexel.edu

Bae-Kyun Kim Department of Mining and Materials Engineering, McGillUniversity, Montreal, QC Canada; Corporate R&D Institute, Samsung Electro-Mechanics, Suwon, Kyungki-Do, Korea 443-743, baekyun.kim@yahoo.ca,baekyun.kim@samsung.com

Hiroyuki Kokawa Department of Materials Processing, Graduate School of Engi-neering, Tohoku University, 6-6-02 Aramaki-aza-Aoba, Aoba-ku Sendai 980-8579,Japan, kokawa@material.tohoku.ac.jp

Mukul Kumar Lawrence Livermore National Laboratory, L-356, 7000 EastAvenue, Livermore, CA 94550, USA, kumar3@llnl.gov

Ricardo Lebensohn Los Alamos National Laboratory, MST-8, MS G755, LosAlamos, NM 87545, USA, lebenso@lanl.gov

Veronica Livescu Los Alamos National Laboratory, MST-8, MS G755, Los Alamos,NM 87545, USA, vlivescu@lanl.gov

Elisabetta Mariani Department of Earth and Ocean Sciences, University ofLiverpool, Liverpool L69 3GP, United Kingdom, mariani@liverpool.ac.uk

Jeremy K. Mason Department of Materials Science and Engineering, MassachusettsInstitute of Technology, Cambridge, MA 02139, USA, jkmason@mit.edu

Terry R. McNelley Department of Mechanical and Astronautical Engineering,Naval Postgraduate School, 700 Dyer Road, Monterey, CA 93943-5146, USA, tmc-nelley@nps.edu

Graham Meaden BLG Productions Ltd., 3 Sydenham Road, Briston BS6 5SH,United Kingdom, grahammeaden@blueyonder.co.uk

E.S. Menon Department of Mechanical and Astronautical Engineering, NavalPostgraduate School, 700 Dyer Road, Monterey, CA 93943-5146, USA,skmeno1@nps.edu

Colin C. Merriman School of Mechanical and Materials Engineering,Washington State University, PO Box 642920, Pullman, WA 99164-2920, USA,merrimac@wsu.edu

Contributors xvii

Sergey Mironov Department of Materials Processing, Graduate School of Engi-neering, Tohoku University, 6-6-02 Aramaki-aza-Aoba, Aoba-ku Sendai 980-8579,Japan, smironov@material.tohoku.ac.jp

Oleg V. Mishin Risø National Laboratory for Sustainable Energy, MaterialsResearch Department, Center for Fundamental Research: Metal Structures in FourDimensions, Technical University of Denmark, DK-4000 Roskilde, Denmark,oleg.mishin@risoe.dk

Matthew M. Nowell EDAX-TSL, 392 East 12300 South, Suite H, Draper, UT84020, USA, Matt.Nowell@ametek.com

David J. Prior Department of Earth and Ocean Sciences, University of Liverpool,Liverpool L69 3GP, United Kingdom, davep@liv.ac.uk

Krishna Rajan Department of Materials Science and Engineering, Iowa State Uni-versity, 2220 Hoover, Ames, IA 50011-2300, USA, krajan@iastate.edu

Valerie Randle School of Engineering, Materials Research Centre, Swansea Univer-sity, Singleton Park, Swansea SA2 8PP, United Kingdom v.randle@swansea.ac.uk

Bryan W. Reed Lawrence Livermore National Laboratory, L-356, 7000 EastAvenue, Livermore, CA 94550, USA, reed12@llnl.gov

Gregory S. Rohrer Department of Materials Science and Engineering, CarnegieMellon University, 5000 Forbes Avenue, Weh 3325, Pittsburgh, PA 15213, USA,gr20@andrew.cmu.edu

Anthony D. Rollett Department of Materials Science and Engineering, CarnegieMellon University, 5000 Forbes Avenue, Reh 148, Pittsburgh, PA 15213, USA,rollett@andrew.cmu.edu

David J. Rowenhorst The United States Naval Research Laboratory, Code 6355,Washington, DC 20375, USA, david.rowenhorst@nrl.navy.mil

Ayman A. Salem Wright Patterson Air Force Base, 2210 Tenth St. Building 655Room 053, WPAFB, OH 45433, USA, ayman.salem@wpafb.af.mil

Yutaka S. Sato Department of Materials Processing, Graduate School of Engineer-ing, Tohoku University, 6-6-02 Aramaki-aza-Aoba, Aoba-ku Sendai 980-8579, Japan,ytksato@material.tohoku.ac.jp

Christopher A. Schuh Department of Materials Science and Engineering, Mas-sachusetts Institute of Technology, Cambridge, MA 02139, USA, schuh@mit.edu

Adam J. Schwartz Lawrence Livermore National Laboratory, L-352, 7000 EastAvenue, Livermore, CA 94550, USA, schwartz6@llnl.gov

Robert A. Schwarzer Institute of Physics, Clausthal University of Technology, D-38678 Clausthal-Zellerfeld, Germany, post@robert-schwarzer.de

Stephen D. Sintay Department of Materials Science and Engineering, CarnegieMellon University, 5000 Forbes Ave. Roberts Hall Room 148, Pittsburgh, PA 15213,USA, ssintay@andrew.cmu.edu

J.-Q Su Naval Postgraduate School, 700 Dyer Road, Monterey, CA 93943-5146,USA, jsu@nps.edu

xviii Contributors

S. Swaminathan General Electric Co, Bangalore, India,srinivasan.swaminathan@ge.com

Jerzy A. Szpunar Department of Mining and Materials Engineering, McGill Uni-versity, 3610 University St., Montreal, QC, Canada H3A 2B2,jerzy.szpunar@mcgill.ca

Sharon G. Torres Lawrence Livermore National Laboratory, L-344, 7000 EastAvenue, Livermore, CA 94550, USA, torres4@llnl.gov

Michael D. Uchic Wright Patterson Air Force Base, 2210 Tenth St. Building 655Room 076, WPAFB, OH 45433, USA, michael.uchic@wpafb.af.mil

John Wheeler Department of Earth and Ocean Sciences, University of Liverpool,Liverpool L69 3GP, United Kingdom, johnwh@liverpool.ac.uk

Angus J. Wilkinson Department of Materials, University of Oxford, Parks Road,Oxford OX1 3PH, United Kingdom, angus.wilkinson@materials.ox.ac.uk

Aimo Winkelmann Max-Planck-Institut fur Mikrostrukturphysik, Weinberg 2, D-06120 Halle, Germany, winkelm@mpi-halle.mpg.de

Stuart I. Wright EDAX-TSL, 392 East 12300 South, Suite H, Draper, UT 84020,USA, Stuart.Wright@ametek.com

Stefan Zaefferer Max-Planck-Institute for Iron Research, Max-Planck-Straße 1, D-40237 Dusseldorf, Germany, s.zaefferer@mpie.de

Alexander Zhilyaev Institute for Metals Superplasticity Problems, Ufa, 450001Russia and Centro Nacional de Investigaciones Metalurgicas, Madrid 28040, Spain,zhilyaev@cenim.csic.es, AlexZ@anrb.ru

Abbreviations

1D 1-dimensional2D 2-dimensional3D 3-dimensionalAD analog-to-digitalADC arbitrary defined cellADR adaptive mesh refinementAMIS average intragrain misorientationARB accumulative roll bondingAS advancing sideB/T basal/transverseBCC body centered cubicBEKD backscatter electron Kikuchi diffractionBEKP backscatter electron Kikuchi patternBF bright fieldBKD backscatter Kikuchi diffractionBKP backscatter Kikuchi patternBSE backscatter electronCA Cellular AutomataCA compression axisCBED convergent beam electron diffractionCCD charge coupled deviceCD crystal deformationCD crystal directionCDF cumulative distribution functionCFD chip flow directionCI confidence indexCIND crystallographic interface normal distributionCIP computer integrated polarizationCOCI color orientation contrast imageCPO crystallographic preferred orientationsCSL coincident site latticeDA digital-to-analogDF dark field

xix

xx Abbreviations

EBSD electron backscatter diffractionEBSP electron backscatter patternECAE equal channel angular extrusionECAP equal channel angular pressingECP electron channeling patternsED extrusion directionEDM electric discharge machiningEDS energy dispersive spectroscopyEDX energy dispersive x-ray spectroscopyEM edge millingEPMA electron probe microanalyzerESEM environmental SEMESR equivalent sphere radiusFCC face-centered cubicFE field emissionFE finite elementFEGSEM field emission gun SEMFEM finite element methodFESEM field emission SEMFFT fast Fourier transformFIB focused ion beamFSP friction stir processingFSW friction stir weldingFSW/P friction stir welding/processingFZ fundamental zoneGAM grain average misorientationGB grain boundaryGD growth directionGEM generalized effective mediumGHAB general high angle boundaryGND geometrically necessary dislocationsGOS grain orientation spreadGSH generalized spherical harmonicsHAB high angle boundaryHAGB high angle grain boundaryHCP hexagonal close packedHMR high misorientation regionHOLZ higher order Laue zonesHPT high-pressure torsionIBP ion blocking patternICDD International Centre of Diffraction DataICSD inorganic crystal structure databaseID identificationIMD integrated angular misorientation density

Abbreviations xxi

IMD intragranular misorientation deviationIND interface normal distributionIPF inverse pole figureIQ image qualityITO indium tin oxideKAM kernal average misorientationkV kilovoltsLAB low angle boundaryLAGB low angle grain boundaryLC low carbonLD long durationLEED low-energy electron diffractionLMR low misorientation regionLPO lattice preferred orientationLSEM large-strain extrusion machiningLV-EBSD low-vacuum EBSDMC Monte CarloMCD modified crystal deformationMIMS mesoscale interface mapping systemMODF misorientation distribution functionMRD multiples of random distributionMSD microstructure sensitive designMSMV maximum subgrid – minimum varianceND normal directionOCF orientation correlation functionODF orientation distribution functionOFHC oxygen free high conductivityOIM orientation imaging microscopyPC principal componentPC personal computerPCA principal component analysisPDF powder diffraction filePDF pair distribution functionPF pole figurePQ pattern qualityPSN particle-stimulated nucleationRD rolling directionRFN rake face normalRHEED reflection high-energy electron diffractionRMS root mean squareRS retreating sideRVE representative volume elementSACP selected area channeling patternSAECP selected area electron channeling patterns

xxii Abbreviations

SD shear directionSD short durationSE secondary electronSEM scanning electron microscopeSIT silicon intensified targetSKM spherical Kikuchi mapSM surface millingSOS scalar orientation spreadSPD severe plastic deformationSPN shear plane normalSSD statistically stored dislocationsSTEM scanning transmission electron microscopeSZ stir zoneSZ shear zoneT transverseTCP tetrahedrally close packedTD transverse directionTEM transmission electron microscopeTMAZ thermomechanically affected zoneTMP thermomechanical processingTRD twin related domainUHV ultra high vacuumWD welding directionWDS wavelength dispersive spectroscopyXRD x-ray diffractionYAG yttrium aluminum garnet

Chapter 1

Present State of Electron Backscatter Diffractionand Prospective Developments

Robert A. Schwarzer, David P. Field, Brent L. Adams, Mukul Kumar,and Adam J. Schwartz

1.1 Introduction

Electron backscatter diffraction (EBSD), whenemployed as an additional characterization techniqueto a scanning electron microscope (SEM), enablesindividual grain orientations, local texture, point-to-point orientation correlations, and phase identificationand distributions to be determined routinely onthe surfaces of bulk polycrystals. The applicationhas experienced rapid acceptance in metallurgical,materials, and geophysical laboratories within thepast decade (Schwartz et al. 2000) due to the wideavailability of SEMs, the ease of sample preparationfrom the bulk, the high speed of data acquisition, andthe access to complementary information about themicrostructure on a submicron scale. From the samespecimen area, surface structure and morphology ofthe microstructure are characterized in great detailby the relief and orientation contrast in secondaryand backscatter electron images, element distribu-tions are accessed by energy dispersive spectroscopy(EDS), wavelength dispersive spectroscopy (WDS), orcathodoluminescence analysis, and the orientations ofsingle grains and phases can now be determined, as acomplement, by EBSD.

The first observation of a diffraction patternin backscattering mode was reported in 1928 byNishikawa and Kikuchi in the same volume wheretransmission electron microscopy (TEM) Kikuchi pat-

R.A. Schwarzer (�)Institute of Physics, Clausthal University of Technology,D-38678 Clausthal-Zellerfeld, Germanye-mail: post@robert-schwarzer.de

terns were discussed (Nishikawa and Kikuchi 1928).The researchers placed a recording film to capturethe pattern in transmission, and then placed a film infront of the specimen so as to obtain an image frombackscattered electrons. This technique was discussedin detail by Alam, Blackman, and Pashley in 1954(Alam et al. 1954) and later investigated by Venablesand co-workers (Venables and Harland 1973; Venablesand Bin-Jaya 1977). The early literature dubbed thetechnique high-angle Kikuchi diffraction and it hasbeen referred to by several additional acronyms in thepast two decades. Those that are most notable, otherthan EBSD, include the more accurate nomenclature ofbackscatter Kikuchi diffraction (BKD) or backscatterelectron Kikuchi diffraction (BEKD). (Note: Acronymsof EBSP, BKP and BEKP are also common in litera-ture and these refer specifically to the image formedby the diffraction technique—i.e., electron backscatterdiffraction pattern.) The terms “electron backscatterdiffraction” and “backscatter Kikuchi diffraction” areoften used interchangeably in the literature.

Fully automated EBSD has developed into amature alternative to X-ray pole figure measure-ments in quantitative texture analysis without suchconstraints as ghost problems, defocusing effects, orinconsistent data as a consequence of specimen tiltsthrough large angles. Moreover, automated EBSD hasopened new horizons in quantitative texture analy-sis because of its outstanding high spatial resolu-tion, its access to orientation correlations and ori-entation stereology, its high speed, and its abilityto represent the texture and grain boundary char-acter distribution visually and quantitatively via anorientation map. Because SEMs and commercialEBSD systems are readily available, electron backscat-ter diffraction is no longer an academic technique

1A.J. Schwartz et al. (eds.), Electron Backscatter Diffraction in Materials Science,DOI 10.1007/978-0-387-88136-2 1, © Springer Science+Business Media, LLC 2009

2 R.A. Schwarzer et al.

reserved to only a few select research laboratories,but rather is well on the way to becoming a tool for pro-cess development and quality control. Additionally, thetechnique enables three-dimensional (3D) volumetricreconstruction of the microstructure from consecutivesurface sections that are created by mechanical serialsection, as described in Chapter 16 by Rohrer and Ran-dle, or focused ion beam (FIB) milling, as discussed inthe chapters by Zaefferer and Wright; Groeber, Rowen-hort, and Uchic; and Sintay, Groeber, and Rollett. Forthe purpose of 3D reconstruction and analysis, how-ever, the speed and ease of handling the EBSD system,as well as the capability to re-examine the results atany time, are decisive requirements.

Automated EBSD at present is limited to materi-als in which grain sizes larger than several tens ofnanometers in diameter and several square millimetersin area can be characterized. Surface strains must notbe excessive, and the specimens must be compatiblewith the general requirements of electron microscopy.In particular, the specimens should be conductive andshould not decompose in vacuum or under the electronbeam. The surface should be reasonably flat and freefrom foreign layers.

1.2 Generation and Interpretation ofElectron Backscatter DiffractionPatterns

EBSD patterns are generated on a phosphor screen bybackscatter diffraction of a stationary beam of high-energy electrons from a volume of crystal materialapproximately 20 nm deep in the specimen, times theprojected area of the incident beam. The characteris-tic feature of a backscatter Kikuchi pattern is the reg-ular arrangement of parallel bright bands on a steepcontinuous background (Fig. 1.1), rather than a reg-ular array of diffraction spots as is generated in theTEM in selected area diffraction from a single crys-tallite. The intersections of Kikuchi bands form promi-nent and distinct zone axes.

The geometry of a Kikuchi pattern can be inter-preted as a gnomonic projection of the crystal latticeon the flat phosphor screen. The point of impinge-ment of the primary beam on the specimen surfaceis the center of projection. The lattice planes can be

Fig. 1.1 Backscatter Kikuchi pattern from cadmium at 20 keV,acquired with an analog video camera

Fig. 1.2 Schematic of the diffracting cones with respect to thereflecting plane, the specimen, and the phosphor screen

imagined to be stretched out to intersect the screen inthe center of the lines of their related Kikuchi bands.Figure 1.2 contains a schematic showing the incidentbeam on the specimen with a given unit cell orien-tation and a specified diffracting plane giving rise tobackscattered “Kikuchi” diffraction. The two diffract-ing cones are the edges of the Kikuchi band, and theplane through the center of these cones is the geomet-ric projection of the diffracting plane onto the phos-phor screen.

When more than one such Kikuchi band is consid-ered, the angles between the projected plane normalorientations correspond to the interplanar angles, andthe angular width of a Kikuchi band {hkl} is twice

1 Present State of Electron Backscatter Diffraction and Prospective Developments 3

the Bragg angle ϑhkl. Thus, the width of the bands isrelated to the interplanar spacing, dhkl, according toBragg’s law:

2 · dhkl · sin ϑhkl = n · λ (1.1)

where n is the order of reflection and λ is thewavelength of the incident electron beam, which isdependent on the accelerating voltage of the SEM.The extinction rules for the expected reflections (i.e.,Kikuchi bands) of the specific crystal structure aredetermined by the structure factor of the crystal. Inaddition, higher order reflections may appear as a set ofstraight lines parallel to the band edges. A decrease inaccelerating voltage, U, causes an increase in electronwavelength and hence an increase in the width of theband, which is, to a first approximation, ϑhkl ∼ 1/U1/2.An appreciable increase in band width and a deviationfrom the usual straight-line approximation to the shapeof real conical sections is observed at low acceleratingvoltages, in particular for high-order Kikuchi lines.

This simple geometric model and the kinematicalapproximation do not explain the exact intensity dis-tribution in a Kikuchi pattern. To fully quantify theintensity distribution, the dynamical theory of elec-tron diffraction must be employed (Reimer 1985). Themechanisms that lead to the formation of the char-acteristic diffraction contrast features in EBSD pat-terns, including Kikuchi bands as well as the promi-nent circular Kikuchi envelopes around zone axes—byappearances like higher order Laue zone (HOLZ) linesfrom thin foils in convergent beam electron diffrac-tion (CBED)—have been described with the applica-tion of dynamic diffraction. Excellent agreement hasbeen obtained between experimental patterns and sim-ulations in extended many-beam dynamical calcula-tions using the Bloch wave approach (Winkelmannet al. 2007; Winkelmann 2008), as is discussed inChapter 2 by Winkelmann.

1.3 Experimental Set-Upof an EBSD System

Instrumentation for generating and capturing electronchanneling patterns (ECP) from selected small spec-imen regions is still available with some commercialSEMs, but the spatial resolution rarely exceeds 50 �m,

as a consequence of the large spherical aberration ofthe probe-forming lens and the pivoting beam. As aresult of the relatively poor resolution and the knowl-edge that many materials of interest have grain sizessmaller than 50 �m, the EBSD technique has largelytaken the place of ECPs in materials and earth sciencesinvestigations. In EBSD, a stationary beam is directedonto the grain of interest to form a Kikuchi pattern.The spot size, and hence the interaction volume of theprimary beam with the crystal contributing to the pat-tern, can be made more than two orders of magnitudesmaller than with ECP. Spatial resolution, as well asdepth resolution in EBSD, depends on specimen tilt,density of the specimen, and accelerating voltage. Thelowest practical beam voltage is about 3 keV, if a phos-phor screen without an aluminum top layer is used.

For quantitative texture analysis, a statistically sig-nificant number of individual grain orientations arerequired. The interactive, or manual, collection of sucha database by the operator is both inconvenient andtime consuming. Fully automated methods have beendeveloped for acquisition and indexing of Kikuchi pat-terns within the SEM (Adams et al. 1993) and withinthe TEM (Zaefferer and Schwarzer 1994; Schwarzerand Sukkau 1998). A number of commercial systemsare currently available, which can be added to new orexisting SEMs. Automated EBSD systems generallyrequire little operator input; after the initial set-up ofthe system, the only input required is the step size. TheEBSD software controls the SEM and rasters the beamacross the specimen on a user-specified pre-definedgrid, pausing at each point only long enough to acquirethe backscatter diffraction pattern, index the orienta-tion, and record the x, y coordinates and the orientationvectors. As discussed in detail below, scanning can beperformed either by translating the specimen in the xand y directions with respect to the stationary primarybeam with a high-precision computer-controlled spec-imen stage (Adams et al. 1993); or by stepping the pri-mary beam under digital computer-control across thestationary specimen surface in a similar way as in con-ventional scanning electron microscopy (Kunze et al.1994). The positions at which diffraction patterns aremeasured may constitute some clusters of individualpoints, a dotted line, or a raster field on the specimensurface. For digital beam scanning, a fast and high-resolution (>12 Bit) digital-to-analog (DA) converteris recommended. A fine raster grid allows for very pre-cise positioning of the measured spots on the inclined

4 R.A. Schwarzer et al.

specimen surface, so as to correct for distortions of thegrid due to the steep tilt of the specimen surface andthe image rotation during dynamical focusing of theprobe-forming lens.

1.4 The Components of an AutomatedEBSD System

An automated EBSD system consists of three mainparts: the SEM, the pattern acquisition device (or cam-era), and the software. To achieve the best possibleperformance, these parts must be considered simul-taneously when designing and setting up a system.In general, it is not recommended that you constructyour own system from scratch. Significant effort isnecessary to ensure the coupled system works syn-chronously and to develop the complex software forcontrolling the SEM functions, the pattern acquisition,and the data interpretation.

The following intrinsic difficulties of EBSD must beaddressed:

• steep specimen tilt, approximately 70◦ relative tothe incident beam (see Fig. 1.3);

• low contrast and intensity, and high backgroundnoise in the backscatter Kikuchi patterns;

• disposition to pattern degradation by contaminationand deformation layers;

Fig. 1.3 Schematic of the typical EBSD geometry, showing thepole piece of the SEM, the electron beam, the tilted specimen,and the phosphor screen

• decomposition and charging of low-conductingmaterials under the beam;

• requirements of high speed, high spatial resolution,and high accuracy of measurement.

1.4.1 The Pattern Acquisition Device

The backscatter Kikuchi pattern is commonly pro-jected onto a transparent phosphor screen (approxi-mately 5 cm in diameter), which is about 2 cm awayfrom the specimen. The screen preferably stands paral-lel with the primary beam and the tilt axis of the stage,but can be rotated about 20 degrees from that planein any direction. The pattern is either viewed with ahigh-sensitivity camera through a window from out-side the specimen chamber, or the phosphor screen isplaced on a fiber optic bundle, which is directly cou-pled to the camera sensor. The phosphor screen is gen-erally matched to the spectral response of the sensor foroptimum performance. Common phosphors employedfor EBSD applications include P20 and P43. P20 hasa short decay time at high current densities, whichoccurs in photon counting tubes, but exhibits a longdecay at low current densities. This latter property isa good match for direct view low light systems. It isyellow/green emitting at 540 nm and has a decay timeof about 1–10 ms with an efficiency (lumens/watt) of30. P43 phosphor is preferred for most applicationswith TV camera output, because of its efficiency andlinearity. It is also fast enough for most high frame-rate applications. It is green emitting (548 nm) witha 1.2 ms decay time and an efficiency of 50. A thin,reflective, aluminum coating is often deposited ontothe phosphor screen. This coating enhances the bright-ness of the phosphor by reflecting light back toward thecamera. It also acts as somewhat of a passive energyfilter in that it absorbs low energy electrons beforethey arrive at the phosphor screen. The most impor-tant function of the coating is to ground the phos-phor screen, as an electrically floating phosphor willcharge and degrade the performance of the SEM andwill interfere with orientation mapping by automatedEBSD. (Alternatively, an indium tin oxide (ITO) layer,or some other conductive and transparent coating, canbe deposited on the substrate window.)

In most EBSD systems, the acquisition device ismounted on a retractable stage. This enables a precise

1 Present State of Electron Backscatter Diffraction and Prospective Developments 5

translation of screen and camera at a fixed spacingfrom each other along the optic axis of the camera sys-tem so that the diffraction pattern projected onto thescreen is kept in focus. A travel of several centimetersis required to provide adequate space for bulky speci-mens when grain orientations need to be measured tothe edge of the material. The accurate displacementof the acquisition device can also be used for cali-brating the EBSD system with the “pattern magnifica-tion method” or “moving screen method” (Day 1993;Hjelen et al. 1993). When retracting the screen fromthe specimen, the pattern “zooms” out from the pat-tern center, which can thus be located quite easily. Thisfeature is used to provide an accurate calibration ofthe system. The pattern center and the specimen-to-screen distance can be accurately calibrated by mea-suring the locations of several corresponding zone axeson the non-displaced and the displaced patterns. Noinitial estimates of the calibration parameters and noknowledge of the crystallography of the sample arerequired. The displacement should be more or less dou-ble the initial specimen-to-screen distance for the ref-erence measurement. However, pattern intensity fallsoff with the square of the specimen-to-screen distance.Furthermore, the precise movement of the device mustbe done in situ under vacuum. In order to guarantee aclean vacuum, a bellows system is recommended overthe method of greased O-rings.

There are several types of camera systems that havebeen used for EBSD image detection. Historically,Peltier-cooled and intensified charge coupled device(CCD) cameras and silicon intensified target (SIT)cameras were used for automated work, and the moreexpensive slow scan CCD cameras were applied forhigh quality imaging and phase identification. Cur-rently, CCD cameras are used for both rapid scan rateimaging and for high quality EBSD image collection.CCD cameras can produce binned images on the orderof ∼100 × 100 pixels at the rate of near 1000 framesper second with sufficient intensity for reliable index-ing. The practical indexing limit is currently in therange of 600–800 images per second, but that numberis likely to continue to increase with more powerfulcomputers and optimized image-handling algorithms.To obtain high quality EBSD patterns for phase iden-tification or publication purposes, there is typically noon-chip binning performed and the full image is col-lected using time averaging techniques to obtain suffi-cient light intensity and contrast.

Some emphasis has to be placed on the light optics.A high-quality macro lens with a small f-stop (largeaperture, “fast lens”) is a good choice in the caseof a short distance between the phosphor screen andthe camera sensor chip. The sensitivity of the acqui-sition system can be almost doubled, at the expenseof high cost and practical inconvenience, by couplingthe CCD sensor with a (tapered) fiber optic bun-dle to the phosphor screen. The highest efficiency isexpected from on-chip deposition of the phosphor orfrom direct exposure of the sensor chip to the pattern-forming electrons. Such a sensor chip will presumablybe placed inside the specimen chamber, either on asmall retractable rod or directly on the specimen stage.

The digital image is the only source of informa-tion for pattern recognition. Software can correct forpoor image quality or distortions only to some extent.Hence the camera has to be chosen with care, mak-ing a trade-off between sensitivity, noise, number ofpixels, image quality, and cost. Almost all currentEBSD systems have moved to video or digital cam-eras with solid-state sensors, either to intensified or tointegrating CCD cameras. These cameras are economi-cal, and the sensor geometry is fixed without producingundue distortions nor “blooming” or burn-in of brightspots (Schwarzer 1989), as has been the case with for-mer vacuum tube sensors. It is worth mentioning thatPeltier-cooling of the sensor chip or the photocathodeof the image intensifier, in order to reduce noise, isineffective at short exposures of less than a second.

1.4.2 Mechanical Stage and DigitalBeam Scanning

Two computer-controlled sampling modes are usedfor automated EBSD: stage-scan mode, in which thespecimen is translated mechanically under the focusedstationary primary beam (Adams et al. 1993); and dig-ital beam scan mode, in which the focused primarybeam is moved across the stationary specimen surface(Kunze et al. 1994). The combination of both modesenables large area scans with high accuracy and speedwhereby small, slightly overlapping fields are sam-pled by digital beam scan and stitched together aftercoarse mechanical steps of the stage from one field tothe next.

6 R.A. Schwarzer et al.

Advantages of the mechanical stage scan include:the accommodation of large measured fields, only lim-ited in size by the range of travel of the specimen stage;step size calibration does not depend on SEM mag-nification; there are no geometric distortions resultingfrom the tilted specimen surface or beam deflection;and from one measured point to the next, diffrac-tion geometry is identical, i.e., pattern center posi-tion, specimen-to-screen distance, background inten-sity, and focus settings remain constant. Hence, muchless elaborate EBSD software is sufficient. A high-performance stage, however, is necessary to keep thespecimen surface exactly in the plane of translation.Step sizes less than 0.5 �m in x and y must be pos-sible. Such a computer-controlled, high precision x–ystage is a relatively slow and expensive alternative rel-ative to the digital beam scan. In addition, the mechan-ical stage has a higher uncertainty of the measurementposition due to play or lag in the system.

Digital beam scanning, on the other hand, enablesan extremely high speed as well as precision in beampositioning. It is cost-effective and not susceptible tobreakage. However, the diffraction geometry and pat-tern center change at each point on the specimen as aresult of the varying beam tilt as the beam is steppedacross and down the specimen (Fig. 1.4). Therefore,

Fig. 1.4 Raster grid on a tilted specimen surface with digitalbeam scan

the system has to be calibrated dynamically from spotto spot (“autocalibration” [Schwarzer 1997]). Other-wise, errors in calculated grain orientations may eas-ily exceed several degrees; or indexing of the patternmay fail in particular at low magnifications and withincreasing distance of the measured location from thepoint on which the system had been calibrated ini-tially. Because of the importance of proper calibration,dynamic calibration of the pattern center has been per-formed on commercial systems since the developmentof beam scanning in 1994. A simple test for accuracyof calibration can be made by measuring across a largefield on a single crystal and checking the uniformity oforientation data. This is also how most systems are cal-ibrated initially. The necessity of a precise calibrationof the pattern center as well as the diffraction length(specimen-to-screen distance) has been verified in acomputer simulation for transmission Kikuchi patternsby Morawiec (1999).

The EBSD software must address two additionaldifficulties with the digital beam scan. As a conse-quence of the steep forward scattering of electrons, thespecimen surface has to be steeply inclined, typicallyat about 20◦ to the primary beam (i.e., typically 70ºfrom horizontal) in backscatter diffraction in order togenerate a Kikuchi pattern of sufficient intensity. Whenscanning down the specimen line after line, the primarybeam spot runs out of focus—increasing in diameter—so that spatial resolution decreases severely. This holdstrue for low as well as high magnifications becausethe requirements of high resolution scale with usefulmagnification. SEM hardware capabilities for dynamicfocusing of most SEMs, however, do not accommo-date the steep specimen tilts as required for EBSD,nor do they accommodate specimen tilts about an axisat an oblique angle to the axis of the specimen stage.This latter limitation would demand a free port for thecamera on that side of the specimen chamber which ispointing at a right angle to the stage axis.

Defocusing has a detrimental effect on spatial res-olution and reliability of indexing, in particular withfine-grain materials; so that dynamic focusing is indis-pensable not only at low but also at medium magni-fications. Pattern quality is a very sensitive indicatorof incorrect focus settings due to the diffuseness ofKikuchi patterns. The superposition of two or morediffraction patterns has two adverse effects on index-ing. First, the small grains contribute less diffractedintensity to the signal. Their faint patterns are

1 Present State of Electron Backscatter Diffraction and Prospective Developments 7

overlooked at best, at the expense of the larger grains,which are covered by the majority of the primarybeam spot. Second, spurious bands from faint pat-terns may be picked up and enter the set of bands forindexing. Pattern interpretation may then fail becauseof the inconsistency of reflections; or even worse,a false orientation may be the result. The effect ofspurious reflections on the reliability of indexing hasbeen clearly demonstrated in a simulation for trans-mission Kikuchi patterns (Morawiec 1999). The beamaperture is significantly smaller in SEM with a fieldemission (FE) gun, so that the depth of focus is sub-stantially increased and the demand for dynamic focus-ing on steeply tilted surfaces is alleviated to the sameextent.

The steep specimen tilt of about 70◦ from horizon-tal causes a further complication of EBSD: the beamspot on the specimen is elongated and hence spatialresolution is reduced in vertical direction by about 3times. Orientation maps, as well as conventional SEMimages, when taken at the same specimen tilt, are fore-shortened to the same extent. Therefore, allowancemust be made for this kind of image distortion inquantitative stereology, either by using different lengthscales for x and y or by stretching out the foreshortenedimage.

The signal to noise ratio in EBSD patterns is quitepoor. The backscatter Kikuchi pattern is superimposedon a background, which is almost 25 times higher inintensity than the useful signal and depends on thegrain orientation, i.e., the actual diffraction pattern.Moreover, the intensity distribution of the backgroundchanges during digital beam scans with the position ofthe beam spot on the specimen surface, as well as withlocal specimen density (phase) and surface relief. Afurther fluctuation may be caused by variations of theprobe current due to instabilities of the emission cur-rent of the gun, drift in the alignment of the column,specimen charging, or build-up of carbon contamina-tion. The quality of diffraction patterns is improvedsignificantly by “flat fielding.” In this case, the raw pat-tern is normalized to a flat field image that containsthe background and image artifacts (e.g., scratches onthe screen, blind or bright dots on the camera chip),but no features of the Kikuchi pattern. There are sev-eral ways to obtain such a flat field image: The beamcan be scanned across an area large enough to containmany grains of different orientations. The Kikuchi pat-terns of these grains are integrated so that they level

out to form an even background. The primary beamcan also be defocused in spot mode to the extreme sothat the Kikuchi patterns fade away. Finally, the back-ground can be reconstructed from the actual diffrac-tion pattern by dedicated software filtering (Field 1997;Schwarzer and Sukkau 1998). In fast EBSD mode (seebelow) with off-line indexing a sequence of patterns, aflat image can be constructed a posteriori by summingup and averaging several patterns out of the sequencethat had been acquired at different positions of thebeam spot on the sample. Each of these techniqueshave specific advantages and limitations. Defocusedspot mode, for instance, adequately reduces the shad-ows due to surface relief. Background reconstructionby filtering is particularly useful in case of a coarsegrain microstructure or a strong texture.

Consequently, advanced EBSD software has notonly to control the digital beam scan or the mechanicalstage scan, but in addition has to control the modesof SEM operation (switching between imaging andspot mode) and pattern acquisition (Schwarzer 1997).Switching the SEM between imaging and spot mode isnecessary for automated experimental flat fielding. Thefinal lens currents (respectively, the working distancesand magnification) have to be read for autocalibrationand dynamic focusing. The final lens current must beset by the computer as a function of x–y beam positionfor software-controlled dynamic focusing. The accel-erating voltage is read as a measure of electron wave-length when the band widths are optionally used forindexing.

1.5 Spatial Resolution

A high spatial resolution in orientation measurementis required for the study of fine grain and heav-ily deformed materials, of recrystallization and graingrowth, of grain boundary characterization, and ofnanomaterials. But why does the spatial resolution inEBSD fall more than one order of magnitude behindspatial resolution in conventional SEM imaging, andstill further behind when compared to the spatial res-olution of a TEM? The inherent resolution of EBSDis governed not by the diameter of the beam spotat the point of impact on the surface, but primarilyby the excitation volume—that is, the fraction of theinteraction volume of the primary electrons within the

8 R.A. Schwarzer et al.

Fig. 1.5 Interaction volume, excitation volume and spatial res-olutions, �, with (a) backscatter Kikuchi patterns from a bulkspecimen in the SEM, (b) transmission Kikuchi patterns from a

thin foil in the TEM, and (c) ion blocking patterns from a bulkspecimen in the scanning ion microscope (schematical represen-tations)

sample from which the pattern forming electrons arebackdiffracted and leave the crystal without furtherscattering. The shading in Fig. 1.5 indicates this vol-ume fraction. This demonstrates why, for orientationmicroscopy in a SEM, it is not wise to reduce thespot size below the diameter of the excitation volume.The adverse effects would be a reduced beam cur-rent, hence less intense patterns, and possibly a strongincrease in contamination rate by polymerization ofhydrocarbons under the beam.

As a consequence of the steep sample tilt, the elon-gated projection of the beam spot, and the forwardscattering, the spatial resolution in EBSD along thebeam direction on the sample surface, �v, is aboutthree times worse than �x. The information depth, �z,is limited by the mean free depth of penetration of thebackscattered electrons in the sampled material at theactual beam voltage. The excitation volume increasesfor light materials and high beam voltages. The TEM,on the other hand, is operated at a significantly higheraccelerating voltage than the SEM. However, the spa-tial resolution, �, in microbeam TEM diffraction isstill approximately the diameter of the beam size,because the sample is thinned to the range of the meanfree path of the energetic electrons, so that only a

small interaction volume can develop (Fig. 1.5b). Inthis case, the information depth, dz, equals the foilthickness.

Spatial resolution in EBSD can be improved tosome extent by lowering the beam voltage from typ-ically 20 kV down to a few kV. However, beambrightness and the sensitivity of the phosphor screen,and hence the pattern intensity, are likewise reduced.While the resolution within a grain is of low sig-nificance, it becomes quite critical when the beamapproaches a grain boundary. An intelligent patternindexing software program can then improve reso-lution by taking account of the intensity levels ofsuperimposed patterns, rejecting less likely orientationsolutions, and comparing orientations in neighboringpixels.

Spatial resolution with copper is better than 0.05 �mat 20 kV using a tungsten filament, and currently lessthan about 0.02 �m with a field emission (FE) gun, asa result of the higher beam current in the small probe.This is roughly the same resolution as that predictedby Venables and Harland in 1973. Backscatter diffrac-tion patterns have been found to disappear when a thinforeign surface layer about twice the thickness of theRutherford elastic mean free path is present at a given

1 Present State of Electron Backscatter Diffraction and Prospective Developments 9

beam energy; i.e., a depth resolution of about 100 nmis assumed for Al, 20 nm for Ni and 10 nm for Au at40 kV accelerating voltage and 20◦ angle of incidenceto the surface (Michael and Goehner 1994). Theoret-ical and experimental values of mean free path relateto amorphous materials, but can be significantly largerand orientation dependent in crystals, as a consequenceof the channeling effect. Therefore, real informationdepths in EBSD are expected to be larger than theseestimates.

High spatial resolution requires an intense primarybeam spot as well as a small interaction volume of theprimary electrons beneath the specimen surface. Thelatter can be reached only by lowering the acceleratingvoltage significantly from about 20 kV, as is usual inpresent systems with a thermionic cathode, to less thanabout 5 kV. A high beam current in a small spot at lowaccelerating voltages is the domain of the field emis-sion SEM. The drawbacks of low accelerating voltagesare the susceptibility of the beam to magnetic strayfields (hence a small working distance is mandatory,which, however, may conflict with the design of cur-rent pattern acquisition systems), the low efficiency ofpresent phosphor screens, and the high susceptibilityof pattern quality to preparation artifacts or foreign sur-face layers.

Because spatial resolution depends on the size ofthe beam spot rather than on the actual magnifica-tion of the SEM, a high spatial resolution can beobtained by correct focus settings, irrespective of lowmagnification. Hence, a large specimen area may bestudied by coupling both a mechanical stage and dig-ital beam scan. The accessible specimen area is lim-ited only by the largest field of view of the SEM atthe lowest magnification and largest working distance.By slightly oversampling, i.e., by choosing a densityof the scanning grid high enough to characterize eachgrain only a few times on the average, the global tex-ture of a large area can be measured conveniently byorientation microscopy. The advantages over conven-tional X-ray pole figure measurement are numerous.The selected specimen area is scanned uniformly, andthe scanned area can be adjusted to irregular shapes.Inhomogeneities in microstructure and texture remainvisible in the orientation maps. Consistent data areobtained, whereas data from X-ray pole-figure mea-surement may be more or less biased due to largevariations of specimen tilt, variations of informationdepth, and variations of the pole-figure window. The

angular instrument resolution is usually higher (about0.5◦ with EBSD, whereas X-ray pole figures are mea-sured with typically 3◦ to 5◦ angular step width). TheX-ray count rate has to be checked for linearity. TheODF calculation from individual grain orientationsdoes not suffer from ghost artifacts. Because almostthe same maximum area can be measured with EBSDwith digital beam scan and in X-ray pole figure mea-surements with an oscillation stage, grain statistics aresimilar and depend on the ratio of average grain sizeto measured area. Automated EBSD competes well inspeed with X-ray diffraction, but is a more univer-sal instrument because of the additional capabilitiesof the SEM.

1.6 SEM Specifications for Good EBSDPerformance

A high beam current is required in spots from 0.02 �m(or slightly less) to 0.5 �m in diameter (to matchthe material-specific resolution limits of EBSD) at amedium working distance (to accommodate the attach-ment of the EBSD system and additional detectors),and at accelerating voltages between about 10 kV to30 kV. A further essential requirement is long-term sta-bility over several hours of the beam current as well asof the mechanics of the specimen stage. Field emis-sion guns have a brightness of about 3 orders of mag-nitude higher than thermionic emitters, but the cross-over—as the effective source of electrons—is less than10 nm in diameter, as compared to 10 �m for a LaB6

emitter. FE guns are superior to any thermionic gunin producing high beam current in small probes of0.02 �m and less, whereas single-crystal LaB6 emittersare superior when the beam spot exceeds about 0.5 �m.SEMs with a field emission source enable high cur-rents of several tens of nA in beam spots of a few nmdiameter, whereas current in small beam spots dropsdramatically when produced with a thermionic emittergun. Therefore, FE SEMs are the first choice for high-speed and high-resolution orientation microscopy. Adetector with low sensitivity can be offset by a highbeam current only to some extent. It is worth keep-ing in mind that contamination rate increases rapidlywith current density. Therefore, it is wise to focus thebeam only down into a useful spot diameter according

10 R.A. Schwarzer et al.

to the actual grain size and the physical resolution limitof backscatter Kikuchi diffraction that is in the rangeof some tens of nm, depending on the material andaccelerating voltage. Furthermore, a low beam currentis generally desirable for the production of patterns toreduce sample damage and charging of low-conductivematerials.

Another invaluable advantage of FE guns is theirmuch smaller beam aperture and hence their largerdepth of focus. In the range of spot sizes that are ofinterest for EBSD, performance depends greatly on thedesign of the lens system. FE-SEMs are usually opti-mized for high resolution at low accelerating voltagesand short working distances. Cold FE guns in particu-lar suffer from significant current fluctuations and needa regular reconditioning (flashing) after duty periods ofabout one hour. They are therefore not so well suitedfor automated EBSD. Schottky FE guns, on the otherhand, can reach an adequate long-term stability of thebeam current. The main drawback of a FE SEM, how-ever, is the higher costs. Beam currents of thermionicguns with a conventional tungsten hairpin filament areabout 4 times lower than currents with a LaB6 cathode.Tungsten filaments are still standard with medium per-formance SEM since they are fairly economical, needonly a moderately high vacuum in the gun chamber,and are known for their excellent beam current sta-bility. In addition, the lifetime of a tungsten filamentmay easily exceed 150 hours. When changed on a reg-ular basis and operated with some care, the lifetimeis a minor source for interruption of long-term scans.In conclusion, a single-crystal LaB6 gun is a goodeconomic compromise at present, but the trend goesdefinitively to Schottky FE guns.

A great challenge of automated EBSD is the studyof low-conductive surfaces such as minerals, oxides (asdiscussed in Chapter 27 by Kim and Szpunar), geologi-cal samples (as discussed in Chapter 26 by Prior, Mar-iani, and Wheeler), hard coatings, integrated circuitswith dielectric layers, specimens with non-metallicinclusions, or embedded samples. There are severalexperimental techniques available which intend eitherto reduce the resistance of the specimen, to reduce theprobe current density, to increase the secondary elec-tron emission coefficient, or to compensate for surfacecharging (Schwarzer 1994). Charging problems arealleviated to some extent by the steep inclination of thespecimen surface to the beam. A conductive coatingwith carbon—not to say gold or other heavy metals—

as in conventional SEM surface imaging, however, isprohibitive, since any foreign layer degrades patternquality as a consequence of the low information depthin backscatter Kikuchi diffraction. A low-vacuum inthe SEM specimen chamber is a convenient means bywhich to suppress specimen charging as described inChapter 25 by El-Dasher and Torres. If available, a“variable pressure” SEM working at a chamber pres-sure in the 1 mbar (100 Pa) range and a beam voltageof about 20 kV or higher is a good choice when insu-lating materials are in the scope of investigation.

Excessive scattering of the pattern-forming elec-trons on their path to the phosphor screen is an adverseside effect of low vacuum that results in a diffuse pat-tern. Hence, the shortest possible specimen-to-screendistance and a high accelerating voltage are manda-tory to reduce this unwanted scattering of the pattern-forming electrons when working at a low vacuumin the specimen chamber. With decreasing specimen-to-screen distance, a larger angular section of theKikuchi pattern is captured. The same translation ofthe beam spot on the sample with digital beam scanresults in the same travel of the pattern center on thescreen (cf. Fig. 1.4), but angular deviation of the ref-erence direction increases with decreasing specimen-to-screen distance. Hence, dynamic pattern center cali-bration becomes indispensable for correct orientationmeasurement the closer the screen is placed to thespecimen.

An essential requirement is a clean vacuum in thespecimen chamber in order to exclude excessive for-mation of carbon contamination. A turbomolecularpump backed by a dry roughing pump is thereforerecommended, while greased vacuum sealings shouldbe avoided. The specimen stage should accommo-date large specimens and a eucentric tilt from 0◦ toabout 75◦ from the horizontal plane. The x–y transla-tion should be made in the surface plane of the spec-imen. A free port of at least 5 cm wide is requiredat normal direction to the tilt axis of the stage about1 cm beneath the eucentric point, for mounting thecamera and the phosphor screen. Finally, a fast SEMcomputer interface is mandatory for high speed dig-ital beam scans, flat imaging, and dynamic focusing.Unfortunately, most high-performance SEMs today arenot optimized for automated EBSD. Therefore, a trade-off has to be made between the performance of thesystem, the intended applications, and the availablehardware.