International Tables FOR Crystallography: Volume B Reciprocal Space

Download as pdf or txt
Download as pdf or txt
You are on page 1of 42

INTERNATIONAL TABLES

FOR
CRYSTALLOGRAPHY

Volume B
RECIPROCAL SPACE

Edited by
U. SHMUELI
Contributing authors

E. Arnold: CABM & Rutgers University, 679 Hoes P. Goodman†: School of Physics, University of
Lane, Piscataway, New Jersey 08854-5638, USA. Melbourne, Parkville, Australia. [5.2]
[2.3] R. W. Grosse-Kunstleve: Lawrence Berkeley
M. I. Aroyo: Departamento de Fisı́ca de la Materia National Laboratory, 1 Cyclotron Road, Mailstop
Condensada, Facultad de Cienca y Technologı́a, 4-230, Berkeley, CA 94720, USA. [1.4]
Universidad del Paı́s Vasco, Apartado 644, 48080 J.-P. Guigay: European Synchrotron Radiation
Bilbao, Spain. [1.5] Facility, BP 220, F-38043 Grenoble, France. [5.3]
A. Authier: Institut de Minéralogie et de la Physique T. Haibach: Laboratory of Crystallography, Depart-
des Milieux Condensés, Bâtiment 7, 140 rue de ment of Materials, ETH Hönggerberg, HCI
Lourmel, 75015 Paris, France. [5.1] G 511, Wolfgang-Pauli-Strasse 10, CH-8093 Zurich,
H. Boysen: Department für Geo- und Umweltwis- Switzerland. [4.6]
senschaften, Sektion Kristallographie, Ludwig- S. R. Hall: Crystallography Centre, University of
Maximilians Universität, Theresienstrasse 41, 80333 Western Australia, Nedlands 6907, WA, Australia.
München, Germany. [4.2] [1.4]
G. Bricogne: Global Phasing Ltd, Sheraton House, H. Jagodzinski: Department für Geo- und Umwelt-
Suites 14–16, Castle Park, Cambridge CB3 0AX, wissenschaften, Sektion Kristallographie, Ludwig-
England, and LURE, Bâtiment 209D, Université Maximilians Universität, Theresienstrasse 41, 80333
Paris-Sud, 91405 Orsay, France. [1.3] München, Germany. [4.2]
P. Coppens: Department of Chemistry, Natural R. E. Marsh: The Beckman Institute–139–74, Cali-
Sciences & Mathematics Complex, State University fornia Institute of Technology, 1201 East California
of New York at Buffalo, Buffalo, New York 14260- Blvd, Pasadena, California 91125, USA. [3.2]
3000, USA. [1.2] R. P. Millane: Department of Electrical and
J. M. Cowley:† Arizona State University, Box 871504, Computer Engineering, University of Canterbury,
Department of Physics and Astronomy, Tempe, AZ Private Bag 4800, Christchurch, New Zealand.
85287-1504, USA. [2.5.1, 2.5.2, 4.3, 5.2] [4.5.1, 4.5.2]
L. M. D. Cranswick: Neutron Program for Materials A. F. Moodie: Department of Applied Physics, Royal
Research, National Research Council Canada, Melbourne Institute of Technology, 124 La Trobe
Building 459, Chalk River Laboratories, Chalk Street, Melbourne, Victoria 3000, Australia. [5.2]
River, Ontario, Canada K0J 1J0. [3.3.4] P. A. Penczek: The University of Texas – Houston
T. A. Darden: Laboratory of Structural Biology, Medical School, Department of Biochemistry and
National Institute of Environmental Health Molecular Biology, 6431 Fannin, MSB 6.218,
Sciences, 111 T. W. Alexander Drive, Research Houston, TX 77030, USA. [2.5.6, 2.5.7]
Triangle Park, NC 27709, USA. [3.5] P. S. Pershan: Division of Engineering and Applied
R. Diamond: MRC Laboratory of Molecular Biology, Science and The Physics Department, Harvard
Hills Road, Cambridge CB2 2QH, England. [3.3.1, University, Cambridge, MA 02138, USA. [4.4]
3.3.2, 3.3.3] S. Ramaseshan†: Raman Research Institute, Banga-
D. L. Dorset: ExxonMobil Research and Engineering lore 560 080, India. [2.4]
Co., 1545 Route 22 East, Clinton Township, M. G. Rossmann: Department of Biological Sciences,
Annandale, New Jersey 08801, USA. [2.5.8, 4.5.1, Purdue University, West Lafayette, Indiana 47907,
4.5.3] USA. [2.3]
F. Frey: Department für Geo- und Umweltwis- D. E. Sands: Department of Chemistry, University of
senschaften, Sektion Kristallographie, Ludwig- Kentucky, Chemistry–Physics Building, Lexington,
Maximilians Universität, Theresienstrasse 41, 80333 Kentucky 40506-0055, USA. [3.1]
München, Germany. [4.2] M. Schlenker: Laboratoire Louis Néel du CNRS, BP
C. Giacovazzo: Dipartimento Geomineralogico, 166, F-38042 Grenoble Cedex 9, France. [5.3]
Campus Universitario, 70125 Bari, Italy, and Insti- V. Schomaker†: Department of Chemistry, University
tute of Crystallography, Via G. Amendola, 122/O, of Washington, Seattle, Washington 98195, USA.
70125 Bari, Italy. [2.2] [3.2]
J. K. Gjønnes: Institute of Physics, University of Oslo, U. Shmueli: School of Chemistry, Tel Aviv University,
PO Box 1048, N-0316 Oslo 3, Norway. [4.3] 69 978 Tel Aviv, Israel. [1.1, 1.4, 2.1]
† Deceased. † Deceased.

v
CONTRIBUTING AUTHORS
J. C. H. Spence: Department of Physics, Arizona State M. Vijayan: Molecular Biophysics Unit, Indian Insti-
University, Tempe, AZ 95287-1504, USA. [2.5.1] tute of Science, Bangalore 560 012, India. [2.4]
W. Steurer: Laboratory of Crystallography, Depart- D. E. Williams†: Department of Chemistry, Univer-
ment of Materials, ETH Hönggerberg, HCI sity of Louisville, Louisville, Kentucky 40292, USA.
G 511, Wolfgang-Pauli-Strasse 10, CH-8093 Zurich, [3.4]
Switzerland. [4.6] B. T. M. Willis: Department of Chemistry, Chemistry
Research Laboratory, University of Oxford, Mans-
M. Tanaka: Institute of Multidisciplinary Research for field Road, Oxford OX1 3TA, England. [4.1]
Advanced Materials, Tohoku University, Japan.
[2.5.3] A. J. C. Wilson†: St John’s College, Cambridge,
England. [2.1]
L. Tong: Department of Biological Sciences, Columbia H. Wondratschek: Institut für Kristallographie,
University, New York 10027, USA. [2.3] Universität, D-76128 Karlsruhe, Germany. [1.5]
B. K. Vainshtein†: Institute of Crystallography, B. B. Zvyagin†: Institute of Ore Mineralogy (IGEM),
Academy of Sciences of Russia, Leninsky prospekt Academy of Sciences of Russia, Staromonetny 35,
59, Moscow B-117333, Russia. [2.5.4, 2.5.5, 2.5.6] 109017 Moscow, Russia. [2.5.4]

† Deceased. † Deceased.

vi
Contents
PAGE

Preface (U. Shmueli) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. xiii

Preface to the second edition (U. Shmueli) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. xiii

Preface to the third edition (U. Shmueli) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. xiv

PART 1. GENERAL RELATIONSHIPS AND TECHNIQUES


1.1. Reciprocal space in crystallography (U. Shmueli) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 2
1.1.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 2
1.1.2. Reciprocal lattice in crystallography .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 2
1.1.3. Fundamental relationships .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 3
1.1.4. Tensor-algebraic formulation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 5
1.1.5. Transformations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 7
1.1.6. Some analytical aspects of the reciprocal space .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 8

1.2. The structure factor (P. Coppens) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 10


1.2.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 10
1.2.2. General scattering expression for X-rays .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 10
1.2.3. Scattering by a crystal: definition of a structure factor .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 10
1.2.4. The isolated-atom approximation in X-ray diffraction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 10
1.2.5. Scattering of thermal neutrons .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 11
1.2.6. Effect of bonding on the atomic electron density within the spherical-atom approximation: the kappa formalism .. .. .. 11
1.2.7. Beyond the spherical-atom description: the atom-centred spherical harmonic expansion .. .. .. .. .. .. .. .. .. .. .. 12
1.2.8. Fourier transform of orbital products .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 15
1.2.9. The atomic temperature factor .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 17
1.2.10. The vibrational probability distribution and its Fourier transform in the harmonic approximation .. .. .. .. .. .. .. .. 17
1.2.11. Rigid-body analysis .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 18
1.2.12. Treatment of anharmonicity .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 20
1.2.13. The generalized structure factor .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 22
1.2.14. Conclusion .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 22

1.3. Fourier transforms in crystallography: theory, algorithms and applications (G. Bricogne) .. .. .. .. .. .. .. .. .. .. .. .. .. 24
1.3.1. General introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 24
1.3.2. The mathematical theory of the Fourier transformation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 24
1.3.3. Numerical computation of the discrete Fourier transform .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 52
1.3.4. Crystallographic applications of Fourier transforms .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 62

1.4. Symmetry in reciprocal space (U. Shmueli) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 114


1.4.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 114
1.4.2. Effects of symmetry on the Fourier image of the crystal .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 114
1.4.3. Structure-factor tables .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 117
1.4.4. Symmetry in reciprocal space: space-group tables .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 119
Appendix A1.4.1. Comments on the preparation and usage of the tables .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 122
Appendix A1.4.2. Space-group symbols for numeric and symbolic computations (U. Shmueli, S. R. Hall and
R. W. Grosse-Kunstleve) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 122
Appendix A1.4.3. Structure-factor tables .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 135
Appendix A1.4.4. Crystallographic space groups in reciprocal space .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 162

1.5. Crystallographic viewpoints in the classification of space-group representations (M. I. Aroyo and H. Wondratschek) .. .. .. .. 175
1.5.1. List of abbreviations and symbols .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 175
1.5.2. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 175
1.5.3. Basic concepts .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 176

vii
CONTENTS
1.5.4. Conventions in the classification of space-group irreps .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 179
1.5.5. Examples and discussion .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 182
1.5.6. Conclusions .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 191
Appendix A1.5.1. Reciprocal-space groups ðGÞ .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 192

PART 2. RECIPROCAL SPACE IN CRYSTAL-STRUCTURE DETERMINATION


2.1. Statistical properties of the weighted reciprocal lattice (U. Shmueli and A. J. C. Wilson) .. .. .. .. .. .. .. .. .. .. .. .. 195
2.1.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 195
2.1.2. The average intensity of general reflections .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 195
2.1.3. The average intensity of zones and rows .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 196
2.1.4. Probability density distributions – mathematical preliminaries .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 197
2.1.5. Ideal probability density distributions .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 200
2.1.6. Distributions of sums, averages and ratios .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 202
2.1.7. Non-ideal distributions: the correction-factor approach .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 203
2.1.8. Non-ideal distributions: the Fourier method .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 207

2.2. Direct methods (C. Giacovazzo) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 215


2.2.1. List of symbols and abbreviations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 215
2.2.2. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 215
2.2.3. Origin specification .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 215
2.2.4. Normalized structure factors .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 216
2.2.5. Phase-determining formulae .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 221
2.2.6. Direct methods in real and reciprocal space: Sayre’s equation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 230
2.2.7. Scheme of procedure for phase determination: the small-molecule case .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 231
2.2.8. Other multisolution methods applied to small molecules .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 232
2.2.9. Some references to direct-methods packages: the small-molecule case .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 234
2.2.10. Direct methods in macromolecular crystallography .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 235

2.3. Patterson and molecular replacement techniques, and the use of noncrystallographic symmetry in phasing (L. Tong,
M. G. Rossmann and E. Arnold) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 244
2.3.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 244
2.3.2. Interpretation of Patterson maps .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 247
2.3.3. Isomorphous replacement difference Pattersons .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 251
2.3.4. Anomalous dispersion .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 255
2.3.5. Noncrystallographic symmetry .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 258
2.3.6. Rotation functions .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 260
2.3.7. Translation functions .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 269
2.3.8. Molecular replacement .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 272
2.3.9. Conclusions .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 275

2.4. Isomorphous replacement and anomalous scattering (M. Vijayan and S. Ramaseshan) .. .. .. .. .. .. .. .. .. .. .. .. .. 282
2.4.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 282
2.4.2. Isomorphous replacement method .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 282
2.4.3. Anomalous-scattering method .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 284
2.4.4. Isomorphous replacement and anomalous scattering in protein crystallography .. .. .. .. .. .. .. .. .. .. .. .. .. .. 287
2.4.5. Anomalous scattering of neutrons and synchrotron radiation. The multiwavelength method .. .. .. .. .. .. .. .. .. .. 293

2.5. Electron diffraction and electron microscopy in structure determination (J. M. Cowley, J. C. H. Spence, M. Tanaka,
B. K. Vainshtein, B. B. Zvyagin, P. A. Penczek and D. L. Dorset) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 297
2.5.1. Foreword (J. M. Cowley and J. C. H. Spence) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 297
2.5.2. Electron diffraction and electron microscopy (J. M. Cowley) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 299
2.5.3. Point-group and space-group determination by convergent-beam electron diffraction (M. Tanaka) .. .. .. .. .. .. .. .. 307

viii
CONTENTS
2.5.4. Electron-diffraction structure analysis (EDSA) (B. K. Vainshtein and B. B. Zvyagin) .. .. .. .. .. .. .. .. .. .. .. 356
2.5.5. Image reconstruction (B. K. Vainshtein) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 361
2.5.6. Three-dimensional reconstruction (B. K. Vainshtein and P. A. Penczek) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 366
2.5.7. Single-particle reconstruction (P. A. Penczek) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 375
2.5.8. Direct phase determination in electron crystallography (D. L. Dorset) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 388

PART 3. DUAL BASES IN CRYSTALLOGRAPHIC COMPUTING


3.1. Distances, angles, and their standard uncertainties (D. E. Sands) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 404
3.1.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 404
3.1.2. Scalar product .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 404
3.1.3. Length of a vector .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 404
3.1.4. Angle between two vectors .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 404
3.1.5. Vector product .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 405
3.1.6. Permutation tensors .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 405
3.1.7. Components of vector product .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 405
3.1.8. Some vector relationships .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 405
3.1.9. Planes .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 406
3.1.10. Variance–covariance matrices .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 406
3.1.11. Mean values .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 408
3.1.12. Computation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 408

3.2. The least-squares plane (R. E. Marsh and V. Schomaker) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 410


3.2.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 410
3.2.2. Least-squares plane based on uncorrelated, isotropic weights .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 410
3.2.3. The proper least-squares plane, with Gaussian weights .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 413
Appendix A3.2.1 .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 416

3.3. Molecular modelling and graphics (R. Diamond and L. M. D. Cranswick) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 418
3.3.1. Graphics (R. Diamond) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 418
3.3.2. Molecular modelling, problems and approaches (R. Diamond) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 434
3.3.3. Implementations (R. Diamond) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 438
3.3.4. Graphics software for the display of small and medium-sized molecules (L. M. D. Cranswick) .. .. .. .. .. .. .. .. .. 442
n
3.4. Accelerated convergence treatment of R lattice sums (D. E. Williams) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 449
3.4.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 449
3.4.2. Definition and behaviour of the direct-space sum .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 449
3.4.3. Preliminary description of the method .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 449
3.4.4. Preliminary derivation to obtain a formula which accelerates the convergence of an Rn sum over lattice points X(d) .. .. 450
3.4.5. Extension of the method to a composite lattice .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 452
3.4.6. The case of n ¼ 1 (Coulombic lattice energy) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 453
3.4.7. The cases of n ¼ 2 and n ¼ 3 .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 454
3.4.8. Derivation of the accelerated convergence formula via the Patterson function .. .. .. .. .. .. .. .. .. .. .. .. .. .. 454
3.4.9. Evaluation of the incomplete gamma function .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 454
3.4.10. Summation over the asymmetric unit and elimination of intramolecular energy terms .. .. .. .. .. .. .. .. .. .. .. 455
3.4.11. Reference formulae for particular values of n .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 455
3.4.12. Numerical illustrations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 456

3.5. Extensions of the Ewald method for Coulomb interactions in crystals (T. A. Darden) .. .. .. .. .. .. .. .. .. .. .. .. .. 458
3.5.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 458
3.5.2. Lattice sums of point charges .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 460
3.5.3. Generalization to Gaussian- and Hermite-based continuous charge distributions .. .. .. .. .. .. .. .. .. .. .. .. .. 471
3.5.4. Computational efficiency .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 474

ix
CONTENTS
PART 4. DIFFUSE SCATTERING AND RELATED TOPICS
4.1. Thermal diffuse scattering of X-rays and neutrons (B. T. M. Willis) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 484
4.1.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 484
4.1.2. Dynamics of three-dimensional crystals .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 484
4.1.3. Scattering of X-rays by thermal vibrations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 487
4.1.4. Scattering of neutrons by thermal vibrations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 488
4.1.5. Phonon dispersion relations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 489
4.1.6. Measurement of elastic constants .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 490

4.2. Disorder diffuse scattering of X-rays and neutrons (F. Frey, H. Boysen and H. Jagodzinski) .. .. .. .. .. .. .. .. .. .. .. 492
4.2.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 492
4.2.2. Basic scattering theory .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 493
4.2.3. Qualitative treatment of structural disorder .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 495
4.2.4. General guidelines for analysing a disorder problem .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 507
4.2.5. Quantitative interpretation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 509
4.2.6. Disorder diffuse scattering from aperiodic crystals .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 526
4.2.7. Computer simulations and modelling .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 528
4.2.8. Experimental techniques and data evaluation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 530

4.3. Diffuse scattering in electron diffraction (J. M. Cowley and J. K. Gjønnes) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 540
4.3.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 540
4.3.2. Inelastic scattering .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 541
4.3.3. Kinematical and pseudo-kinematical scattering .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 542
4.3.4. Dynamical scattering: Bragg scattering effects .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 542
4.3.5. Multislice calculations for diffraction and imaging .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 544
4.3.6. Qualitative interpretation of diffuse scattering of electrons .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 544

4.4. Scattering from mesomorphic structures (P. S. Pershan) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 547


4.4.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 547
4.4.2. The nematic phase .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 549
4.4.3. Smectic-A and smectic-C phases .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 551
4.4.4. Phases with in-plane order .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 554
4.4.5. Discotic phases .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 561
4.4.6. Other phases .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 561

4.5. Polymer crystallography (R. P. Millane and D. L. Dorset) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 567


4.5.1. Overview (R. P. Millane and D. L. Dorset) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 567
4.5.2. X-ray fibre diffraction analysis (R. P. Millane) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 568
4.5.3. Electron crystallography of polymers (D. L. Dorset) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 583

4.6. Reciprocal-space images of aperiodic crystals (W. Steurer and T. Haibach) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 590
4.6.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 590
4.6.2. The n-dimensional description of aperiodic crystals .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 591
4.6.3. Reciprocal-space images .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 598
4.6.4. Experimental aspects of the reciprocal-space analysis of aperiodic crystals .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 621

PART 5. DYNAMICAL THEORY AND ITS APPLICATIONS


5.1. Dynamical theory of X-ray diffraction (A. Authier) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 626
5.1.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 626
5.1.2. Fundamentals of plane-wave dynamical theory .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 626
5.1.3. Solutions of plane-wave dynamical theory .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 630
5.1.4. Standing waves .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 633
5.1.5. Anomalous absorption .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 633

x
CONTENTS
5.1.6. Intensities of plane waves in transmission geometry .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 634
5.1.7. Intensity of plane waves in reflection geometry .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 638
5.1.8. Real waves .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 640
Appendix A5.1.1. Basic equations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 642

5.2. Dynamical theory of electron diffraction (A. F. Moodie, J. M. Cowley and P. Goodman) .. .. .. .. .. .. .. .. .. .. .. .. .. 647
5.2.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 647
5.2.2. The defining equations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 647
5.2.3. Forward scattering .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 647
5.2.4. Evolution operator .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 648
5.2.5. Projection approximation – real-space solution .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 648
5.2.6. Semi-reciprocal space .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 648
5.2.7. Two-beam approximation .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 649
5.2.8. Eigenvalue approach .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 649
5.2.9. Translational invariance .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 650
5.2.10. Bloch-wave formulations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 650
5.2.11. Dispersion surfaces .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 650
5.2.12. Multislice .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 651
5.2.13. Born series .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 651
5.2.14. Approximations .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 652

5.3. Dynamical theory of neutron diffraction (M. Schlenker and J.-P. Guigay) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 654
5.3.1. Introduction .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 654
5.3.2. Comparison between X-rays and neutrons with spin neglected .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 654
5.3.3. Neutron spin, and diffraction by perfect magnetic crystals .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 655
5.3.4. Extinction in neutron diffraction (nonmagnetic case) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 658
5.3.5. Effect of external fields on neutron scattering by perfect crystals .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 659
5.3.6. Experimental tests of the dynamical theory of neutron scattering .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 659
5.3.7. Applications of the dynamical theory of neutron scattering .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 660

Author index .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 665

Subject index .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 675

xi
Preface
By Uri Shmueli

The purpose of Volume B of International Tables for Crystal- The obviously delayed publication of Volume B is due to
lography is to provide the user or reader with accounts of some several reasons. Some minor delays were caused by a require-
well established topics, of importance to the science of crystal- ment that potential contributors should be approved by the
lography, which are related in one way or another to the concepts Executive Committee prior to issuing relevant invitations. Much
of reciprocal lattice and, more generally, reciprocal space. Efforts more serious delays were caused by authors who failed to deliver
have been made to extend the treatment of the various topics to their contributions. In fact, some invited contributions had to be
include X-ray, electron, and neutron diffraction techniques, and excluded from this first edition of Volume B. Some of the topics
thereby do some justice to the inclusion of the present Volume in here treated are greatly extended, considerably updated or
the new series of International Tables for Crystallography. modern versions of similar topics previously treated in the old
An important crystallographic aspect of symmetry in reci- Volumes I, II, and IV. Most of the subjects treated in Volume B
procal space, space-group-dependent expressions of trigono- are new to International Tables.
metric structure factors, already appears in Volume I of I gratefully thank Professor A. J. C. Wilson, for suggesting that
International Tables for X-ray Crystallography, and preliminary I edit this Volume and for sharing with me his rich editorial
plans for incorporating this and other crystallographic aspects of experience. I am indebted to those authors of Volume B who
reciprocal space in the new edition of International Tables date took my requests and deadlines seriously, and to the Computing
back to 1972. However, work on a volume of International Tables Center of Tel Aviv University for computing facilities and time.
for Crystallography, largely dedicated to the subject of reciprocal Special thanks are due to Mrs Z. Stein (Tel Aviv University) for
space, began over ten years later. The present structure of skilful assistance in numeric and symbolic programming,
Volume B, as determined in the years preceding the 1984 involved in my contributions to this Volume.
Hamburg congress of the International Union of Crystallography I am most grateful to many colleagues–crystallographers for
(IUCr), is due to (i) computer-controlled production of concise encouragement, advice, and suggestions. In particular, thanks
structure-factor tables, (ii) the ability to introduce many more are due to Professors J. M. Cowley, P. Goodman and C. J.
aspects of reciprocal space – as a result of reducing the effort of Humphreys, who served as Chairmen of the Commission on
producing the above tables, as well as their volume, and (iii) Electron Diffraction during the preparation of this Volume, for
suggestions by the National Committees and individual crystal- prompt and expert help at all stages of the editing. The kind
lographers of some additional interesting topics. It should be assistance of Dr J. N. King, the Executive Secretary of the IUCr,
pointed out that the initial plans for the present Volume and is also gratefully acknowledged. Last, but certainly not least, I
Volume C (Mathematical, Physical and Chemical Tables, edited wish to thank Mr M. H. Dacombe, the Technical Editor of the
by Professor A. J. C. Wilson), were formulated and approved IUCr, and his staff for the skilful and competent treatment of the
during the same period. variety of drafts and proofs out of which this Volume arose.

Preface to the second edition


By Uri Shmueli

The first edition of Volume B appeared in 1993, and was microscopy and diffraction in crystal structure determination.
followed by a corrected reprint in 1996. Although practically all The latter topic is here enhanced by applications of direct
the material for the second edition was available in early 1997, methods to electron crystallography.
its publication was delayed by the decision to translate all of Part 3, Dual Bases in Crystallographic Computing, deals with
Volume B, and indeed all the other volumes of International applications of reciprocal space to molecular geometry and
Tables for Crystallography, to Standard Generalized ‘best’-plane calculations, and contains a treatment of
Markup Language (SGML) and thus make them available the principles of molecular graphics and modelling and their
also in an electronic form suitable for modern publishing applications; it concludes with the presentation of a
procedures. convergence-acceleration method, of importance in the compu-
During the preparation of the second edition, most chapters tation of approximate lattice sums.
that appeared in the first edition have been corrected and/or Part 4 contains treatments of various diffuse-scattering
revised, some were rather extensively updated, and five new phenomena arising from crystal dynamics, disorder and low
chapters were added. The overall structure of the second edition dimensionality (liquid crystals), and an exposition of the
is outlined below. underlying theories and/or experimental evidence. The new
After an introductory chapter, Part 1 presents the reader with additions to this part are treatments of polymer crystallography
an account of structure-factor formalisms, an extensive treat- and of reciprocal-space images of aperiodic crystals.
ment of the theory, algorithms and crystallographic applications Part 5 contains introductory treatments of the theory of the
of Fourier methods, and treatments of symmetry in reciprocal interaction of radiation with matter, the so-called dynamical
space. These are here enriched with more advanced aspects of theory, as applied to X-ray, electron and neutron diffraction
representations of space groups in reciprocal space. techniques. The chapter on the dynamical theory of neutron
In Part 2, these general accounts are followed by detailed diffraction is new.
expositions of crystallographic statistics, the theory of direct I am deeply grateful to the authors of the new contributions
methods, Patterson techniques, isomorphous replacement and for making their expertise available to Volume B and for their
anomalous scattering, and treatments of the role of electron excellent collaboration. I also take special pleasure in thanking

xiii
PREFACE
those authors of the first edition who revised and updated their This editorial work was carried out at the School of Chemistry
contributions in view of recent developments. Last but not least, and the Computing Center of Tel Aviv University. The facilities
I wish to thank all the authors for their contributions and their they put at my disposal are gratefully acknowledged on my behalf
patience, and am grateful to those authors who took my requests and on behalf of the IUCr. I wish to thank many colleagues for
seriously. I hope that the updating and revision of future editions interesting conversations and advice, and in particular Professor
will be much easier and more expedient, mainly because of the Theo Hahn with whom I discussed at length problems regarding
new format of International Tables. Volume B and International Tables in general.
Four friends and greatly respected colleagues who Given all these expert contributions, the publication of this
contributed to the second edition of Volume B are no longer volume would not have been possible without the expertise and
with us. These are Professors Arthur J. C. Wilson, Peter devotion of the Technical Editors of the IUCr. My thanks go to
Goodman, Verner Schomaker and Boris K. Vainshtein. I asked Mrs Sue King, for her cooperation during the early stages of the
Professors Michiyoshi Tanaka, John Cowley and Douglas Dorset work on the second edition of Volume B, while the material was
if they were prepared to answer queries related to the being collected, and to Dr Nicola Ashcroft, for her collaboration
contributions of the late Peter Goodman and Boris K. during the final stages of the production of the volume, for her
Vainshtein to Chapter 2.5. I am most grateful for their prompt most careful and competent treatment of the proofs, and last but
agreement. not least for her tactful and friendly attitude.

Preface to the third edition


By Uri Shmueli

The second edition of Volume B appeared in 2001. Plans for the My gratitude is extended to the authors of new contributions
third edition included the addition of new chapters and sections, and to the authors of the first and second editions of the volume
the substantial revision of several chapters that existed in the for significant revisions of their chapters and sections in view of
second edition and minor revisions and updating of existing new developments. I wish to thank all the authors for their
chapters. The overall structure of Volume B remained excellent collaboration and for sharing with the International
unchanged. Tables for Crystallography their expertise. I hope that the
In Part 1, Chapter 1.5 on classifications of space-group tradition of keeping the contributions up to date will also persist
representations in reciprocal space has been extensively revised. in future editions of Volume B. This will be aided by significant
In Part 2, Chapter 2.2 on direct methods has been improvements in various aspects of technical editing which were
considerably extended to include applications of these methods already apparent in the preparation of this edition.
to macromolecular crystallography. Chapter 2.3 on Patterson Three greatly respected friends and colleagues who
and molecular replacement techniques has been updated and contributed to this and previous editions of Volume B
extended. Section 2.5.3 on convergent-beam electron diffraction passed away after the second edition of Volume B was
has been completely rewritten by a newly invited author, and published. These are Professors John Cowley, Boris
Section 2.5.6 on three-dimensional reconstruction has been Zvyagin and Donald Williams. I asked Professors John
updated and extended by a newly invited author, who has also Spence, Douglas Dorset and Pawel Penczek to take care of
added Section 2.5.7 on single-particle reconstruction. The any questions about the articles of the late John Cowley, Boris
Foreword to Chapter 2.5 on electron diffraction and microscopy Zvyagin and Boris Vainshtein in Chapter 2.5, and Dr Bill Smith
has also been revised. to answer any questions about Chapter 3.4 by the late Donald
In Part 3, Chapter 3.3 on computer graphics and molecular Williams. They all agreed promptly and I am most grateful for
modelling has been enriched by Section 3.3.4 on the implemen- this.
tation of molecular graphics to small and medium-sized My editorial work was carried out at the School of Chemistry
molecules, and a comprehensive Chapter 3.5 on modern of Tel Aviv University and I wish to acknowledge gratefully the
extensions of Ewald methods has been added, dealing with (i) facilities that were put at my disposal. I am grateful to many
inclusion of fast Fourier transforms in the computation of sums friends and colleagues for interesting conversations and
and (ii) departure from the point-charge model in Ewald exchanges related to this volume. Thanks are also due to my
summations. friends from the IUCr office in Chester for their helpful
In Part 4, Chapter 4.1 on thermal diffuse scattering of X-rays interest.
and neutrons has been updated, and Chapter 4.2 on disorder Finally, I think that the publication of the third edition of
diffuse scattering of X-rays and neutrons has been extensively Volume B would not have been possible without the competent,
revised and updated. tactful and friendly collaboration of Dr Nicola Ashcroft, the
Minor updates and corrections have also been made to several Technical Editor of this project during all the stages of the
existing chapters and sections in all the parts of the volume. preparation of this edition.

xiv
SAMPLE PAGES
1.1. Reciprocal space in crystallography

By U. Shmueli

1.1.1. Introduction try to give a unified demonstration of the usefulness of mutually


The purpose of this chapter is to provide an introduction to reciprocal bases as an interpretive tool.
several aspects of reciprocal space, which are of general impor- Consider the equation of a lattice plane in the direct lattice. It
tance in crystallography and which appear in the various chapters is shown in standard textbooks (e.g. Buerger, 1941) that this
and sections to follow. We first summarize the basic definitions equation is given by
and briefly inspect some fundamental aspects of crystallography, hx þ ky þ lz ¼ n; ð1:1:2:3Þ
while recalling that they can be usefully and simply discussed in
terms of the concept of the reciprocal lattice. This introductory where h, k and l are relatively prime integers (i.e. not having a
section is followed by a summary of the basic relationships common factor other than þ1 or 1), known as Miller indices of
between the direct and associated reciprocal lattices. We then the lattice plane, x, y and z are the coordinates of any point lying
introduce the elements of tensor-algebraic formulation of such in the plane and are expressed as fractions of the magnitudes of
dual relationships, with emphasis on those that are important in the basis vectors a, b and c of the direct lattice, respectively, and n
many applications of reciprocal space to crystallographic algo- is an integer denoting the serial number of the lattice plane within
rithms. We proceed with a section that demonstrates the role of the family of parallel and equidistant ðhklÞ planes, the interplanar
mutually reciprocal bases in transformations of coordinates and spacing being denoted by dhkl ; the value n ¼ 0 corresponds to the
conclude with a brief outline of some important analytical aspects ðhklÞ plane passing through the origin.
of reciprocal space, most of which are further developed in other Let r ¼ xa þ yb þ zc and rL ¼ ua þ vb þ wc, where u, v, w are
parts of this volume. any integers, denote the position vectors of the point xyz and a
lattice point uvw lying in the plane (1.1.2.3), respectively, and
assume that r and rL are different vectors. If the plane normal is
1.1.2. Reciprocal lattice in crystallography denoted by N, where N is proportional to the vector product of
The notion of mutually reciprocal triads of vectors dates back to two in-plane lattice vectors, the vector form of the equation of the
the introduction of vector calculus by J. Willard Gibbs in the lattice plane becomes
1880s (e.g. Wilson, 1901). This concept appeared to be useful in
the early interpretations of diffraction from single crystals N  ðr  rL Þ ¼ 0 or N  r ¼ N  rL : ð1:1:2:4Þ
(Ewald, 1913; Laue, 1914) and its first detailed exposition and the
recognition of its importance in crystallography can be found in For equations (1.1.2.3) and (1.1.2.4) to be identical, the plane
Ewald’s (1921) article. The following free translation of Ewald’s normal N must satisfy the requirement that N  rL ¼ n, where n is
(1921) introduction, presented in a somewhat different notation, an (unrestricted) integer.
may serve the purpose of this section: Let us now consider the basic diffraction relations (e.g. Lipson
& Cochran, 1966). Suppose a parallel beam of monochromatic
To the set of ai, there corresponds in the vector calculus a set of radiation, of wavelength , falls on a lattice of identical point
‘reciprocal vectors’ bi , which are defined (by Gibbs) by the scatterers. If it is assumed that the scattering is elastic, i.e. there is
following properties: no change of the wavelength during this process, the wavevectors
of the incident and scattered radiation have the same magnitude,
ai  bk ¼ 0 ðfor i 6¼ kÞ ð1:1:2:1Þ which can conveniently be taken as 1=. A consideration of path
and phase differences between the waves outgoing from two
point scatterers separated by the lattice vector rL (defined as
and above) shows that the condition for their maximum constructive
interference is given by
ai  bi ¼ 1; ð1:1:2:2Þ ðs  s0 Þ  rL ¼ n; ð1:1:2:5Þ

where s0 and s are the wavevectors of the incident and scattered


where i and k may each equal 1, 2 or 3. The first equation, (1.1.2.1), beams, respectively, and n is an arbitrary integer.
says that each vector bk is perpendicular to two vectors ai, as follows Since rL ¼ ua þ vb þ wc, where u, v and w are unrestricted
from the vanishing scalar products. Equation (1.1.2.2) provides the integers, equation (1.1.2.5) is equivalent to the equations of Laue:
norm of the vector bi : the length of this vector must be chosen such
h  a ¼ h; h  b ¼ k; h  c ¼ l; ð1:1:2:6Þ
that the projection of bi on the direction of ai has the length 1=ai ,
where ai is the magnitude of the vector ai . . . .
where h ¼ s  s0 is the diffraction vector, and h, k and l are
The consequences of equations (1.1.2.1) and (1.1.2.2) were integers corresponding to orders of diffraction from the three-
elaborated by Ewald (1921) and are very well documented in the dimensional lattice (Lipson & Cochran, 1966). The diffraction
subsequent literature, crystallographic as well as other. vector thus has to satisfy a condition that is analogous to that
As is well known, the reciprocal lattice occupies a rather imposed on the normal to a lattice plane.
prominent position in crystallography and there are nearly as The next relevant aspect to be commented on is the Fourier
many accounts of its importance as there are crystallographic expansion of a function having the periodicity of the crystal
texts. It is not intended to review its applications, in any detail, in lattice. Such functions are e.g. the electron density, the density of
the present section; this is done in the remaining chapters and nuclear matter and the electrostatic potential in the crystal, which
sections of the present volume. It seems desirable, however, to are the operative definitions of crystal structure in X-ray, neutron
mention by way of an introduction some fundamental geome- and electron-diffraction methods of crystal structure determina-
trical, physical and mathematical aspects of crystallography, and tion. A Fourier expansion of such a periodic function may be
2
1.2. The structure factor

By P. Coppens

1.2.1. Introduction
AðSÞ ¼ F^ fðrÞg; ð1:2:2:4bÞ
The structure factor is the central concept in structure analysis by
diffraction methods. Its modulus is called the structure amplitude. where F^ is the Fourier transform operator.
The structure amplitude is a function of the indices of the set of
scattering planes h, k and l, and is defined as the amplitude of
scattering by the contents of the crystallographic unit cell, 1.2.3. Scattering by a crystal: definition of a structure factor
expressed in units of scattering. For X-ray scattering, that unit is In a crystal of infinite size, ðrÞ is a three-dimensional periodic
the scattering by a single electron (2.82  1015 m), while for function, as expressed by the convolution
neutron scattering by atomic nuclei, the unit of scattering length PPP
of 1014 m is commonly used. The complex form of the structure crystal ðrÞ ¼ unit cell ðrÞ  ðr  na  mb  pcÞ; ð1:2:3:1Þ
n m p
factor means that the phase of the scattered wave is not simply
related to that of the incident wave. However, the observable,
which is the scattered intensity, must be real. It is proportional to where n, m and p are integers, and  is the Dirac delta function.
the square of the scattering amplitude (see, e.g., Lipson & Thus, according to the Fourier convolution theorem,
P PP
Cochran, 1966). AðSÞ ¼ F^ fðrÞg ¼ F^ funit cell ðrÞgF^ fðr  na  mb  pcÞg;
The structure factor is directly related to the distribution of n m p
scattering matter in the unit cell which, in the X-ray case, is the ð1:2:3:2Þ
electron distribution, time-averaged over the vibrational modes
of the solid. which gives
In this chapter we will discuss structure-factor expressions for PPP
X-ray and neutron scattering, and, in particular, the modelling AðSÞ ¼ F^ funit cell ðrÞg ðS  ha  kb  lc Þ: ð1:2:3:3Þ
h k l
that is required to obtain an analytical description in terms of the
features of the electron distribution and the vibrational displa-
Expression (1.2.3.3) is valid for a crystal with a very large
cement parameters of individual atoms. We concentrate on the
number of unit cells, in which particle-size broadening is negli-
most basic developments; for further details the reader is referred
gible. Furthermore, it does not account for multiple scattering of
to the cited literature.
the beam within the crystal. Because of the appearance of the
delta function, (1.2.3.3) implies that S = H with H ¼ ha þ
1.2.2. General scattering expression for X-rays kb þ lc .
The first factor in (1.2.3.3), the scattering amplitude of one unit
The total scattering of X-rays contains both elastic and inelastic cell, is defined as the structure factor F:
components. Within the first-order Born approximation (Born, R
1926) it has been treated by several authors (e.g. Waller & FðHÞ ¼ F^ funit cell ðrÞg ¼ unit cell ðrÞ expð2iH  rÞ dr: ð1:2:3:4Þ
Hartree, 1929; Feil, 1977) and is given by the expression
P R 2
Itotal ðSÞ ¼ Iclassical  n expð2iS  rj Þ 0 dr ; ð1:2:2:1Þ
n
1.2.4. The isolated-atom approximation in X-ray diffraction
where Iclassical is the classical Thomson scattering of an To a reasonable approximation, the unit-cell density can be
X-ray beam by a free electron, which is equal to described as a superposition of isolated, spherical atoms located
ðe2 =mc2 Þ2 ð1 þ cos2 2Þ=2 for an unpolarized beam of unit inten- at rj .
sity, is the n-electron space-wavefunction expressed in the 3n P
coordinates of the electrons located at rj and the integration is unit cell ðrÞ ¼ atom; j ðrÞ  ðr  rj Þ: ð1:2:4:1Þ
j
over the coordinates of all electrons. S is the scattering vector of
length 2 sin =.
Substitution in (1.2.3.4) gives
The coherent elastic component of the scattering, in units of P P
the scattering of a free electron, is given by FðHÞ ¼ F^ fatom; j gF^ fðr  rj Þg ¼ fj expð2iH  rj Þ
R P j j
Icoherent; elastic ðSÞ ¼  0  expð2iS  rj Þj 0 drj2 : ð1:2:2:2Þ ð1:2:4:2aÞ
j

or
If integration is performed over all coordinates but those of the P
jth electron, one obtains after summation over all electrons Fðh; k; lÞ ¼ fj exp 2iðhxj þ kyj þ lzj Þ
j
R P
Icoherent; elastic ðSÞ ¼ j ðrÞ expð2iS  rÞ drj2 ; ð1:2:2:3Þ ¼ fj fcos 2ðhxj þ kyj þ lzj Þ
j

where ðrÞ is the electron distribution. The scattering amplitude þ i sin 2ðhxj þ kyj þ lzj Þg: ð1:2:4:2bÞ
AðSÞ is then given by
R fj ðSÞ, the spherical atomic scattering factor, or form factor, is the
AðSÞ ¼ ðrÞ expð2iS  rÞ dr ð1:2:2:4aÞ Fourier transform of the spherically averaged atomic density
j ðrÞ, in which the polar coordinate r is relative to the nuclear
or position. fj ðSÞ can be written as (James, 1982)
10
1. GENERAL RELATIONSHIPS AND TECHNIQUES
Table 1.2.7.4. Closed-form expressions for Fourier transform of Slater-type functions (Avery & Watson, 1977; Su & Coppens, 1990)
R1 N
hjk i  0 r expðZrÞjk ðKrÞ dr; K ¼ 4 sin =:

N
k 1 2 3 4 5 6 7 8
0 1 2Z 2ð3Z2  K2 Þ 24ZðZ2  K2 Þ 24ð5Z2  10K2 Z2 þ K4 Þ 240ZðK2  3Z2 Þð3K2  Z2 Þ 720ð7Z6  35K2 Z4 þ 21K4 Z2  K6 Þ 40320ðZ7  7K2 Z5 þ 7K4 Z3  K6 ZÞ
K2 þ Z2 ðK2 þ Z2 Þ2 ðK2 þ Z2 Þ3 ðK2 þ Z2 Þ4 ðK2 þ Z2 Þ5 ðK2 þ Z2 Þ6 ðK2 þ Z2 Þ7 ðK2 þ Z2 Þ8

1 2K 8KZ 8Kð5Z2  K2 Þ 48KZð5Z2  3K2 Þ 48Kð35Z4  42K2 Z2 þ 3K4 Þ 1920KZð7Z4  14K2 Z2 þ 3K4 Þ 5760Kð21Z6  63K2 Z4 þ 27K4 Z2  K6 Þ
ðK2 þ Z2 Þ2 ðK2 þ Z2 Þ3 ðK2 þ Z2 Þ4 ðK2 þ Z2 Þ5 ðK2 þ Z2 Þ6 ðK2 þ Z2 Þ7 ðK2 þ Z2 Þ8

2 8K2 48K2 Z 48K2 ð7Z2  K2 Þ 384K2 Zð7Z2  3K2 Þ 1152K2 ð21Z4  18K2 Z2 þ K4 Þ 11520K2 Zð21Z4  30K2 Z2 þ 5K4 Þ
ðK2 þ Z2 Þ3 ðK2 þ Z2 Þ4 ðK2 þ Z2 Þ5 ðK2 þ Z2 Þ6 ðK2 þ Z2 Þ7 ðK2 þ Z2 Þ8

3 48K3 384K3 Z 384K3 ð9Z2  K2 Þ 11520K3 Zð3Z2  K2 Þ 11520K3 ð33Z4  22K2 Z2 þ K4 Þ


ðK2 þ Z2 Þ4 ðK2 þ Z2 Þ5 ðK2 þ Z2 Þ6 ðK2 þ Z2 Þ7 ðK2 þ Z2 Þ8

4 384K4 3840K4 Z 3840K4 ð11Z2  K2 Þ 46080K4 Zð11Z2  3K2 Þ


ðK2 þ Z2 Þ5 ðK2 þ Z2 Þ6 ðK2 þ Z2 Þ7 ðK2 þ Z2 Þ8

5 3840K5 46080K5 Z 40680K5 ð13Z2  K2 Þ


ðK2 þ Z2 Þ6 ðK2 þ Z2 Þ7 ðK2 þ Z2 Þ8

6 46080K6 645120K6 Z
ðK2 þ Z2 Þ7 ðK2 þ Z2 Þ8

7 645120K7
ðK2 þ Z2 Þ8

PðuÞ ¼ ð2hu2 iÞ3=2 expfjuj2 =2hu2 ig; ð1:2:10:1Þ around a vector k ð1 ; 2 ; 3 Þ, with length corresponding to the
magnitude of the rotation, results in a displacement r, such that
where hu2 i is the mean-square displacement in any direction.
r ¼ ðk  rÞ ¼ Dr ð1:2:11:1Þ
The corresponding trivariate normal distribution to be used for
anisotropic harmonic motion is, in tensor notation,
with
jr1 j1=2 2 3
PðuÞ ¼ expf 12 r1 j k
jk ðu u Þg: ð1:2:10:2aÞ 0 3 2
ð2Þ3=2 D ¼ 4 3 0 1 5; ð1:2:11:2Þ
2 1 0
Here r is the variance–covariance matrix, with covariant
components, and jr1 j is the determinant of the inverse of r.
Summation over repeated indices has been assumed. The or in Cartesian tensor notation, assuming summation over
corresponding equation in matrix notation is repeated indices,

jr1 j1=2 ri ¼ Dij rj ¼ "ijk k rj ð1:2:11:3Þ


PðuÞ ¼ expf 12 ðuÞT r1 ðuÞg; ð1:2:10:2bÞ
ð2Þ3=2
where the permutation operator "ijk equals +1 for i, j, k a cyclic
where the superscript T indicates the transpose. permutation of the indices 1, 2, 3, or 1 for a noncyclic permu-
The characteristic function, or Fourier transform, of PðuÞ is tation, and zero if two or more indices are equal. For i = 1, for
example, only the "123 and "132 terms occur. Addition of a
TðHÞ ¼ expf22
jk hj hk g ð1:2:10:3aÞ translational displacement gives
ri ¼ Dij rj þ ti : ð1:2:11:4Þ
or
TðHÞ ¼ expf22 HT rHg: ð1:2:10:3bÞ When a rigid body undergoes vibrations the displacements
vary with time, so suitable averages must be taken to derive the
With the change of variable b jk ¼ 22
jk, (1.2.10.3a) becomes mean-square displacements. If the librational and translational
TðHÞ ¼ expfb jk hj hk g: motions are independent, the cross products between the two
terms in (1.2.11.4) average to zero and the elements of the mean-
square displacement tensor of atom n, Uijn , are given by
n
U11 ¼ þL22 r23 þ L33 r22  2L23 r2 r3 þ T11
n
1.2.11. Rigid-body analysis U22 ¼ þL33 r21 þ L11 r23  2L13 r1 r3 þ T22
n
The treatment of rigid-body motion of molecules or molecular U33 ¼ þL11 r22 þ L22 r21  2L12 r1 r2 þ T33
n
ð1:2:11:5Þ
fragments was developed by Cruickshank (1956) and expanded U12 ¼ L33 r1 r2  L12 r23 þ L13 r2 r3 þ L23 r1 r3 þ T12
into a general theory by Schomaker & Trueblood (1968). The U13 ¼ L22 r1 r3 þ L12 r2 r3  L13 r22 þ L23 r1 r2 þ T13
n
theory has been described by Johnson (1970b) and by Dunitz n
(1979). The latter reference forms the basis for the following U23 ¼ L11 r2 r3 þ L12 r1 r3  L13 r1 r2  L23 r21 þ T23 ;
treatment.
The most general motions of a rigid body consist of rotations where the coefficients Lij ¼ hi j i and Tij ¼ hti tj i are the
about three axes, coupled with translations parallel to each of the elements of the 3  3 libration tensor L and the 3  3 translation
axes. Such motions correspond to screw rotations. A libration tensor T, respectively. Since pairs of terms such as hti tj i and htj ti i
18
1. GENERAL RELATIONSHIPS AND TECHNIQUES
(5) Parseval/Plancherel property. If u, w, U, W are as above, theory of certain Lie groups and coding theory – to list only a few.
then The interested reader may consult Auslander & Tolimieri (1979);
Auslander, Feig & Winograd (1982, 1984); Auslander & Tolimieri
1 (1985); Tolimieri (1985).
ðF ðNÞ½U; F ðNÞ½WÞW ¼ ðU; WÞW  One-dimensional algorithms are examined first. The Sande
jdet Nj
mixed-radix version of the Cooley–Tukey algorithm only calls
1
ðF ðNÞ½u; F ðNÞ½wÞW ¼ ðu; wÞW : upon the additive structure of congruence classes of integers. The
jdet Nj prime factor algorithm of Good begins to exploit some of their
multiplicative structure, and the use of relatively prime factors
leads to a stronger factorization than that of Sande. Fuller use of
(6) Period 4. When N is symmetric, so that the ranges of
the multiplicative structure, via the group of units, leads to the
indices k and k  can be identified, it makes sense to speak of
Rader algorithm; and the factorization of short convolutions then
powers of F ðNÞ and F ðNÞ. Then the ‘standardized’ matrices
yields the Winograd algorithms.
ð1=jdet Nj1=2 ÞF ðNÞ and ð1=jdet Nj1=2 ÞF ðNÞ are unitary matrices
Multidimensional algorithms are at first built as tensor
whose fourth power is the identity matrix (Section 1.3.2.4.3.4);
products of one-dimensional elements. The problem of factoring
their eigenvalues are therefore 1 and i.
the DFT in several dimensions simultaneously is then examined.
The section ends with a survey of attempts at formalizing the
1.3.3. Numerical computation of the discrete Fourier transform interplay between algorithm structure and computer architecture
for the purpose of automating the design of optimal DFT code.
1.3.3.1. Introduction It was originally intended to incorporate into this section a
The Fourier transformation’s most remarkable property is survey of all the basic notions and results of abstract algebra
undoubtedly that of turning convolution into multiplication. As which are called upon in the course of these developments, but
distribution theory has shown, other valuable properties – such as time limitations have made this impossible. This material,
the shift property, the conversion of differentiation into multi- however, is adequately covered by the first chapter of Tolimieri et
plication by monomials, and the duality between periodicity and al. (1989) in a form tailored for the same purposes. Similarly, the
sampling – are special instances of the convolution theorem. inclusion of numerous detailed examples of the algorithms
This property is exploited in many areas of applied mathe- described here has had to be postponed to a later edition, but an
matics and engineering (Campbell & Foster, 1948; Sneddon, abundant supply of such examples may be found in the signal
1951; Champeney, 1973; Bracewell, 1986). For example, the processing literature, for instance in the books by McClellan &
passing of a signal through a linear filter, which results in its being Rader (1979), Blahut (1985), and Tolimieri et al. (1989).
convolved with the response of the filter to a -function ‘impulse’,
may be modelled as a multiplication of the signal’s transform by 1.3.3.2. One-dimensional algorithms
the transform of the impulse response (also called transfer Throughout this section we will denote by eðtÞ the expression
function). Similarly, the solution of systems of partial differential expð2itÞ, t 2 R. The mapping t 7 ! eðtÞ has the following prop-
equations may be turned by Fourier transformation into a divi- erties:
sion problem for distributions. In both cases, the formulations
obtained after Fourier transformation are considerably simpler
than the initial ones, and lend themselves to constructive solution eðt1 þ t2 Þ ¼ eðt1 Þeðt2 Þ
techniques.
eðtÞ ¼ eðtÞ ¼ ½eðtÞ1
Whenever the functions to which the Fourier transform is
applied are band-limited, or can be well approximated by band- eðtÞ ¼ 1 , t 2 Z:
limited functions, the discrete Fourier transform (DFT) provides a
means of constructing explicit numerical solutions to the
problems at hand. A great variety of investigations in physics, Thus e defines an isomorphism between the additive group R=Z
engineering and applied mathematics thus lead to DFT calcula- (the reals modulo the integers) and the multiplicative group of
tions, to such a degree that, at the time of writing, about 50% of all complex numbers of modulus 1. It follows that the mapping
supercomputer CPU time is alleged to be spent calculating DFTs. ‘ 7 ! eð‘=NÞ, where ‘ 2 Z and N is a positive integer, defines an
The straightforward use of the defining formulae for the DFT isomorphism between the one-dimensional residual lattice Z=N Z
leads to calculations of size N 2 for N sample points, which and the multiplicative group of Nth roots of unity.
become unfeasible for any but the smallest problems. Much The DFT on N points then relates vectors X and X in W and
ingenuity has therefore been exerted on the design and imple- 
W through the linear transformations:
mentation of faster algorithms for calculating the DFT
(McClellan & Rader, 1979; Nussbaumer, 1981; Blahut, 1985;
Brigham, 1988). The most famous is that of Cooley & Tukey 1 X  
FðNÞ : XðkÞ ¼ X ðk Þeðk k=NÞ
(1965) which heralded the age of digital signal processing. N k 2Z=N Z
However, it had been preceded by the prime factor algorithm of X
Good (1958, 1960), which has lately been the basis of many new F ðNÞ : X  ðk Þ ¼ XðkÞeðk k=NÞ:
k2Z=N Z
developments. Recent historical research (Goldstine, 1977, pp.
249–253; Heideman et al., 1984) has shown that Gauss essentially
knew the Cooley–Tukey algorithm as early as 1805 (before
Fourier’s 1807 work on harmonic analysis!); while it has long
been clear that Dirichlet knew of the basis of the prime factor 1.3.3.2.1. The Cooley–Tukey algorithm
algorithm and used it extensively in his theory of multiplicative The presentation of Gentleman & Sande (1966) will be
characters [see e.g. Chapter I of Ayoub (1963), and Chapters 6 followed first [see also Cochran et al. (1967)]. It will then be
and 8 of Apostol (1976)]. Thus the computation of the DFT, far reinterpreted in geometric terms which will prepare the way for
from being a purely technical and rather narrow piece of the treatment of multidimensional transforms in Section 1.3.3.3.
specialized numerical analysis, turns out to have very rich Suppose that the number of sample points N is composite, say
connections with such central areas of pure mathematics as N ¼ N1 N2 . We may write k to the base N1 and k to the base N2
number theory (algebraic and analytic), the representation as follows:
52
1. GENERAL RELATIONSHIPS AND TECHNIQUES
Table A1.4.3.3. Monoclinic space groups
Each expression for A or B in the monoclinic system and for the space-group settings chosen in IT A is represented in terms of one of the following symbols:

cðhlÞcðkyÞ ¼ cos½2ðhx þ lzÞ cosð2kyÞ; cðhkÞcðlzÞ ¼ cos½2ðhx þ kyÞ cosð2lzÞ;


cðhlÞsðkyÞ ¼ cos½2ðhx þ lzÞ sinð2kyÞ; cðhkÞsðlzÞ ¼ cos½2ðhx þ kyÞ sinð2lzÞ;
sðhlÞcðkyÞ ¼ sin½2ðhx þ lzÞ cosð2kyÞ; sðhkÞcðlzÞ ¼ sin½2ðhx þ kyÞ cosð2lzÞ;
sðhlÞsðkyÞ ¼ sin½2ðhx þ lzÞ sinð2kyÞ; sðhkÞsðlzÞ ¼ sin½2ðhx þ kyÞ sinð2lzÞ; ðA1:4:3:1Þ

where the left-hand column of expressions corresponds to space-group representations in the second setting, with b taken as the unique axis, and the right-hand column
corresponds to representations in the first setting, with c taken as the unique axis.
The lattice types in this table are P, A, B, C and I, and are all explicit in the full space-group symbol only (see below). Note that s(hl), s(hk), s(ky) and s(lz) are zero for
h = l = 0, h = k = 0, k = 0 and l = 0, respectively.

Group symbol
Unique
No. Short Full Parity A B axis
3 P2 P121 2c(hl)c(ky) 2c(hl)s(ky) b
3 P2 P112 2c(hk)c(lz) 2c(hk)s(lz) c
4 P21 P121 1 k ¼ 2n 2c(hl)c(ky) 2c(hl)s(ky) b
k ¼ 2n þ 1 2s(hl)s(ky) 2s(hl)c(ky)
4 P21 P1121 l ¼ 2n 2c(hk)c(lz) 2c(hk)s(lz) c
l ¼ 2n þ 1 2s(hk)s(lz) 2s(hk)c(lz)
5 C2 C121 4c(hl)c(ky) 4c(hl)s(ky) b
5 C2 A121 4c(hl)c(ky) 4c(hl)s(ky) b
5 C2 I121 4c(hl)c(ky) 4c(hl)s(ky) b
5 C2 A112 4c(hk)c(lz) 4c(hk)s(lz) c
5 C2 B112 4c(hk)c(lz) 4c(hk)s(lz) c
5 C2 I112 4c(hk)c(lz) 4c(hk)s(lz) c
6 Pm P1m1 2c(hl)c(ky) 2s(hl)c(ky) b
6 Pm P11m 2c(hk)c(lz) 2s(hk)c(lz) c
7 Pc P1c1 l ¼ 2n 2c(hl)c(ky) 2s(hl)c(ky) b
l ¼ 2n þ 1 2s(hl)s(ky) 2c(hl)s(ky)
7 Pc P1n1 h þ l ¼ 2n 2c(hl)c(ky) 2s(hl)c(ky) b
h þ l ¼ 2n þ 1 2s(hl)s(ky) 2c(hl)s(ky)
7 Pc P1a1 h ¼ 2n 2c(hl)c(ky) 2s(hl)c(ky) b
h ¼ 2n þ 1 2s(hl)s(ky) 2c(hl)s(ky)
7 Pc P11a h ¼ 2n 2c(hk)c(lz) 2s(hk)c(lz) c
h ¼ 2n þ 1 2s(hk)s(lz) 2c(hk)s(lz)
7 Pc P11n h þ k ¼ 2n 2c(hk)c(lz) 2s(hk)c(lz) c
h þ k ¼ 2n þ 1 2s(hk)s(lz) 2c(hk)s(lz)
7 Pc P11b k ¼ 2n 2c(hk)c(lz) 2s(hk)c(lz) c
k ¼ 2n þ 1 2s(hk)s(lz) 2c(hk)s(lz)
8 Cm C1m1 4c(hl)c(ky) 4s(hl)c(ky) b
8 Cm A1m1 4c(hl)c(ky) 4s(hl)c(ky) b
8 Cm I1m1 4c(hl)c(ky) 4s(hl)c(ky) b
8 Cm A11m 4c(hk)c(lz) 4s(hk)c(lz) c
8 Cm B11m 4c(hk)c(lz) 4s(hk)c(lz) c
8 Cm I11m 4c(hk)c(lz) 4s(hk)c(lz) c
9 Cc C1c1 l ¼ 2n 4c(hl)c(ky) 4s(hl)c(ky) b
l ¼ 2n þ 1 4s(hl)s(ky) 4c(hl)s(ky)
9 Cc A1n1 h þ l ¼ 2n 4c(hl)c(ky) 4s(hl)c(ky) b
h þ l ¼ 2n þ 1 4s(hl)s(ky) 4c(hl)s(ky)
9 Cc I1a1 h ¼ 2n 4c(hl)c(ky) 4s(hl)c(ky) b
h ¼ 2n þ 1 4s(hl)s(ky) 4c(hl)s(ky)
9 Cc A11a h ¼ 2n 4c(hk)c(lz) 4s(hk)c(lz) c
h ¼ 2n þ 1 4s(hk)s(lz) 4c(hk)s(lz)
9 Cc B11n h þ k ¼ 2n 4c(hk)c(lz) 4s(hk)c(lz) c
h þ k ¼ 2n þ 1 4s(hk)s(lz) 4c(hk)s(lz)
9 Cc I11b k ¼ 2n 4c(hk)c(lz) 4s(hk)c(lz) c
k ¼ 2n þ 1 4s(hk)s(lz) 4c(hk)s(lz)
10 P2=m P12=m1 4c(hl)c(ky) 0 b
10 P2=m P112=m 4c(hk)c(lz) 0 c
11 P21 =m P121 =m1 k ¼ 2n 4c(hl)c(ky) 0 b
k ¼ 2n þ 1 4s(hl)s(ky) 0

136
1. GENERAL RELATIONSHIPS AND TECHNIQUES
Table A1.4.4.1 (cont.)

P421 2 No. 90 (143) P42 mc No. 105 (158)


(1) hkl: (2) hkl: (3) khl: 110/2 (4) khl: 110/2 (1) hkl: (2) hkl: (3) khl: 001/2 (4) khl: 001/2
(5) hkl: 110/2 (6) hkl: 110/2 (7) khl: (8) khl: (5) hkl: (6) hkl: (7) khl: 001/2 (8) khl: 001/2

P41 22 No. 91 (144) P42 bc No. 106 (159)


(1) hkl: (2) hkl: 001/2 (3) khl: 001/4 (4) khl: 003/4 (1) hkl: (2) hkl: (3) khl: 001/2 (4) khl: 001/2
(5) hkl: (6) hkl: 001/2 (7) khl: 003/4 (8) khl: 001/4 (5) hkl: 110/2 (6) hkl: 110/2 (7) khl: 111/2 (8) khl: 111/2

P41 21 2 No. 92 (145) I4mm No. 107 (160)


(1) hkl: (2) hkl: 001/2 (3) khl: 221/4 (4) khl: 223/4 (1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: 221/4 (6) hkl: 223/4 (7) khl: (8) khl: 001/2 (5) hkl: (6) hkl: (7) khl: (8) khl:

P42 22 No. 93 (146) I4cm No. 108 (161)


(1) hkl: (2) hkl: (3) khl: 001/2 (4) khl: 001/2 (1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: (6) hkl: (7) khl: 001/2 (8) khl: 001/2 (5) hkl: 001/2 (6) hkl: 001/2 (7) khl: 001/2 (8) khl: 001/2

P42 21 2 No. 94 (147) I41 md No. 109 (162)


(1) hkl: (2) hkl: (3) khl: 111/2 (4) khl: 111/2 (1) hkl: (2) hkl: 111/2 (3) khl: 021/4 (4) khl: 203/4
(5) hkl: 111/2 (6) hkl: 111/2 (7) khl: (8) khl: (5) hkl: (6) hkl: 111/2 (7) khl: 203/4 (8) khl: 021/4

P43 22 No. 95 (148) I41 cd No. 110 (163)


(1) hkl: (2) hkl: 001/2 (3) khl: 003/4 (4) khl: 001/4 (1) hkl: (2) hkl: 111/2 (3) khl: 021/4 (4) khl: 203/4
(5) hkl: (6) hkl: 001/2 (7) khl: 001/4 (8) khl: 003/4 (5) hkl: 001/2 (6) hkl: 110/2 (7) khl: 201/4 (8) khl: 023/4

P43 21 2 No. 96 (149)


(1) hkl: (2) hkl: 001/2 (3) khl: 223/4 (4) khl: 221/4 Point group: 42m Tetragonal Laue group: 4/mmm
(5) hkl: 223/4 (6) hkl: 221/4 (7) khl: (8) khl: 001/2 P42m No. 111 (164)
(1) hkl: (2) hkl: (3) khl: (4) khl:
I422 No. 97 (150) (5) hkl: (6) hkl: (7) khl: (8) khl:
(1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: (6) hkl: (7) khl: (8) khl: P42c No. 112 (165)
(1) hkl: (2) hkl: (3) khl: (4) khl:
I41 22 No. 98 (151) (5) hkl: 001/2 (6) hkl: 001/2 (7) khl: 001/2 (8) khl: 001/2
(1) hkl: (2) hkl: 111/2 (3) khl: 021/4 (4) khl: 203/4
(5) hkl: 203/4 (6) hkl: 021/4 (7) khl: 111/2 (8) khl: P421 m No. 113 (166)
(1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: 110/2 (6) hkl: 110/2 (7) khl: 110/2 (8) khl: 110/2
Point group: 4mm Tetragonal Laue group: 4/mmm
P4mm No. 99 (152) P421 c No. 114 (167)
(1) hkl: (2) hkl: (3) khl: (4) khl: (1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: (6) hkl: (7) khl: (8) khl: (5) hkl: 111/2 (6) hkl: 111/2 (7) khl: 111/2 (8) khl: 111/2

P4bm No. 100 (153) P4m2 No. 115 (168)


(1) hkl: (2) hkl: (3) khl: (4) khl: (1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: 110/2 (6) hkl: 110/2 (7) khl: 110/2 (8) khl: 110/2 (5) hkl: (6) hkl: (7) khl: (8) khl:

P42 cm No. 101 (154) P4c2 No. 116 (169)


(1) hkl: (2) hkl: (3) khl: 001/2 (4) khl: 001/2 (1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: 001/2 (6) hkl: 001/2 (7) khl: (8) khl: (5) hkl: 001/2 (6) hkl: 001/2 (7) khl: 001/2 (8) khl: 001/2

P42 nm No. 102 (155) P4b2 No. 117 (170)


(1) hkl: (2) hkl: (3) khl: 111/2 (4) khl: 111/2 (1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: 111/2 (6) hkl: 111/2 (7) khl: (8) khl: (5) hkl: 110/2 (6) hkl: 110/2 (7) khl: 110/2 (8) khl: 110/2

P4cc No. 103 (156) P4n2 No. 118 (171)


(1) hkl: (2) hkl: (3) khl: (4) khl: (1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: 001/2 (6) hkl: 001/2 (7) khl: 001/2 (8) khl: 001/2 (5) hkl: 111/2 (6) hkl: 111/2 (7) khl: 111/2 (8) khl: 111/2

P4nc No. 104 (157) I4m2 No. 119 (172)


(1) hkl: (2) hkl: (3) khl: (4) khl: (1) hkl: (2) hkl: (3) khl: (4) khl:
(5) hkl: 111/2 (6) hkl: 111/2 (7) khl: 111/2 (8) khl: 111/2 (5) hkl: (6) hkl: (7) khl: (8) khl:

166
1.5. CLASSIFICATION OF SPACE-GROUP REPRESENTATIONS

Fig. 1.5.5.5. Brillouin zone with asymmetric unit and representation domain of CDML for arithmetic crystal class mm2F: a2 < b2 þ c2 , b2 < c2 þ a2 and
c2 < a2 þ b2 . Space groups Fmm2  C2v 18 19
(42), Fdd2  C2v (43). Reciprocal-space group (Imm2)*, No. 44: a2 < b2 þ c2 , b2 < c2 þ a2 and c2 < a2 þ b2
(see Table 1.5.5.5). The representation domain of CDML is different from the asymmetric unit. Auxiliary points: T4: 0; 12 ;  12; Y2: 12 ; 0; 12; Y4: 12 ; 0;  12; Z2: 0; 0;  12.
Flagpoles: 0; 0; z:  12 < z < 0; 0; 12 ; z:  12 < z < 0. Wings: x; 0; z: 0 < x < 12 ;  12 < z < 0; 0; y; z: 0 < y < 12 ;  12 < z < 0.

Table 1.5.5.5. List of k-vector types for arithmetic crystal class mm2F: a2 < b2 þ c2 , b2 < c2 þ a2 and c2 < a2 þ b2
See Fig. 1.5.5.5. Parameter relations: x ¼  12  þ 12  þ 12 , y ¼ 12   12  þ 12 , z ¼ 12  þ 12   12 .
Wyckoff position of IT A,
k-vector label, CDML cf. Section 1.5.4.3 Parameters
 0; 0; 0 ex 2 a mm2 0; 0; 0
Z 12 ; 12 ; 0 ex 2 a mm2 0; 0; 12
 ; ; 0 ex 2 a mm2 0; 0; z: 0 < z < 12
LE ; ; 0 ex 2 a mm2 0; 0; z:  12 < z < 0
 [  [ Z [ LE 2 a mm2 0; 0; z:  12 < z  12
T 0; 12 ; 12 ex 2 b mm2 1
2 ; 0; 0
T  T2 0; 12 ; 12
Y 12 ; 0; 12 ex 2 b mm2 0; 12 ; 0
G ; 12 þ ; 12 ex 2 b mm2 1
2 ; 0; z: 0 < z  g0
G  H3 ¼ ½H2 T4  0; 12 ; z:  12 < z   12 þ g0 ¼ h2
GA ; 12  ; 12 ex 2 b mm2 1
2 ; 0; z: g2 ¼ g0 < z < 0
GA  H1 ¼ ½H0 T2  0; 12 ; z: 12  g0 ¼ h0 < z < 12
H 12 þ ; ; 12 ex 2 b mm2 0; 12 ; z: 0 < z  h0
HA 12  ; ; 12 ex 2 b mm2 0; 12 ; z: h2 ¼ h0 < z < 0
T2 [ H1 [ H [ Y [ HA [ H3 2 b mm2 0; 12 ; z:  12 < z  12
 0; ;  ex 4 c :m: x; 0; 0: 0 < x < 12
A 12 ; 12 þ ;  ex 4 c :m: x; 0; 12: 0 < x  a0
C 12 ; ; 12 þ  ex 4 c :m: x; 12 ; 0: 0 < x < c0 ¼ 12  a0
C  A1 x; 0; 12: 12  a0 ¼ c0 < x < 12
J ;  þ ;  ex 4 c :m: x; 0; z: ½ Z A0 G0 T
JA ;  þ ;  ex 4 c :m: x; 0; z: ½ T G2 A2 Z2 
K 12 þ ;  þ ; 12 þ  ex 4 c :m: x; 12 ; z: ½Y H0 C0 
K  J1 x; 0; z: ½Y4 G2 A2 
KA 12  ;  þ ; 12 þ  ex 4 c :m: x; 12 ; z: ½Y C0 H2 
KA  J3 x; 0; z: ½Y2 G0 A0 
A [ A1 [ J [ J3 [  [ JA [ J1 4 c :m: x; 0; z: 0 < x < 12; 0 < z  12
 ; 0;  ex 4 d m:: 0; y; 0: 0 < y < 12
B 12 þ ; 12 ;  ex 4 d m:: 0; y; 12: 0 < y < b0
D ; 12 ; 12 þ  ex 4 d m:: 1
2 ; y; 0: 0 < y  d0
D  B1 0; y; 12: 12  d0 ¼ b0  y < 12
E  þ ; ;  ex 4 d m:: 0; y; z: ½ Y H0 B0 Z
EA  þ ; ;  ex 4 d m:: 0; y; z: ½ Z2 B2 H2 Y
F  þ ; 12 þ ; 12 þ  ex 4 d m:: 1
2 ; y; z: ½T D0 G0 
F  E3 0; y; z: ½B2 T4 H2 
FA  þ ; 12  ; 12 þ  ex 4 d m:: 1
2 ; y; z: ½T G2 D0 
FA  E1 0; y; z: ½T2 B0 H0 
 [ B [ B1 [ E [ E1 [ EA [ E3 4 d m:: 0; y; z: 0 < y < 12;  12 < z  12
GP ; ;  8 e 1 x; y; z: 0 < x; y < 12; 0 < z  12

187
2. RECIPROCAL SPACE IN CRYSTAL-STRUCTURE DETERMINATION
tion to the structure factor as a random variable. This is of course
a necessary requirement for any statistical treatment. If, however,
the atomic composition of the asymmetric unit is widely
heterogeneous, the structure factor is then a sum of unequally
distributed random variables and the Lindeberg–Lévy version of
the central-limit theorem (cf. Section 2.1.4.4) cannot be expected
to apply. Other versions of this theorem might still predict a
normal p.d.f. of the sum, but at the expense of a correspondingly
large number of terms/atoms. It is well known that atomic
heterogeneity gives rise to severe deviations from ideal beha-
viour (e.g. Howells et al., 1950) and one of the aims of crystal-
lographic statistics has been the introduction of a correct
dependence on the atomic composition into the non-ideal p.d.f.’s
[for a review of the early work on non-ideal distributions see
Srinivasan & Parthasarathy (1976)]. A somewhat less well known
fact is that the dependence of the p.d.f.’s of jEj on space-group
symmetry becomes more conspicuous as the composition
becomes more heterogeneous (e.g. Shmueli, 1979; Shmueli &
Fig. 2.1.7.1. Atomic heterogeneity and intensity statistics. The histogram
Wilson, 1981). Hence both the composition and the symmetry appearing in this figure was constructed from jEj values which were
dependence of the intensity statistics are of interest. Other recalculated from atomic parameters published for the centrosymmetric
problems, which likewise give rise to non-ideal p.d.f.’s, are the structure of C6H18Cl2N4O4Pt (Faggiani et al., 1980). The space group of the
presence of heavy atoms in (variable) special positions, hetero- crystal is P1, Z ¼ 2, i.e. all the atoms are located in general positions. The
geneous structures with complete or partial noncrystallographic figure shows a comparison of the recalculated distribution of jEj with the
symmetry, and the presence of outstandingly heavy dispersive ideal centric [equation (2.1.5.11)] and acentric [equation (2.1.5.8)] p.d.f.’s,
denoted by 1 and 1, respectively.
scatterers.
The need for theoretical representations of non-ideal p.d.f.’s is 
Rb 1; if k¼m
exemplified in Fig. 2.1.7.1, which shows the ideal centric and fk ðxÞfm ðxÞpð0Þ ðxÞ dx ¼ km ¼ ; ð2:1:7:2Þ
acentric p.d.f.’s together with a frequency histogram of jEj values, a 0; if k 6¼ m
recalculated for a centrosymmetric structure containing a
platinum atom in the asymmetric unit of P1 (Faggiani et al., 1980). where ½a; b is the range of existence of all the functions involved.
Clearly, the deviation from the Gaussian p.d.f., predicted by the It can be readily shown that the coefficients dk are given by
central-limit theorem, is here very large and a comparison with
the possible ideal distributions can (in this case) lead to wrong Rb P
k
conclusions. dk ¼ fk ðxÞpðxÞ dx ¼ hfk ðxÞi ¼ cðkÞ n
n hx i; ð2:1:7:3Þ
Two general approaches have so far been employed in deri- a n¼0
vations of non-ideal p.d.f.’s which account for the above-
mentioned problems: the correction-factor approach, to be dealt where the brackets h i in equation (2.1.7.3) denote averaging with
with in the following sections, and the more recently introduced respect to the unknown p.d.f. pðxÞ and cðkÞ
n is the coefficient of the
Fourier method, to which Section 2.1.8 is dedicated. In what nth power of x in the polynomial fk ðxÞ. The coefficients dk are
follows, we introduce briefly the mathematical background of the thus directly related to the moments of the non-ideal distribution
correction-factor approach, apply this formalism to centric and and the coefficients of the powers of x in the orthogonal poly-
acentric non-ideal p.d.f.’s, and present the numerical values of the nomials. The latter coefficients can be obtained by the Gram–
moments of the trigonometric structure factor which permit an Schmidt procedure (e.g. Spiegel, 1974), or by direct use of the
approximate evaluation of such p.d.f.’s for all the three- Szegö determinants (e.g. Cramér, 1951), for any weight function
dimensional space groups. that has finite moments. However, the feasibility of the present
approach depends on our ability to obtain the moments hxn i
2.1.7.2. Mathematical background without the knowledge of the non-ideal p.d.f., pðxÞ.
Suppose that pðxÞ is a p.d.f. which accurately describes the
2.1.7.3. Application to centric and acentric distributions
experimental distribution of the random variable x, where x is
related to a sum of random variables and can be assumed to obey We shall summarize here the non-ideal centric and acentric
(to some approximation) an ideal p.d.f., say pð0Þ ðxÞ, based on the distributions of the magnitude of the normalized structure factor
central-limit theorem. In the correction-factor approach we seek E (e.g. Shmueli & Wilson, 1981; Shmueli, 1982). We assume that
to represent pðxÞ as (i) all the atoms are located in general positions and have
rationally independent coordinates, (ii) all the scatterers are
P
pðxÞ ¼ pð0Þ ðxÞ dk fk ðxÞ; ð2:1:7:1Þ dispersionless, and (iii) there is no noncrystallographic symmetry.
k Arbitrary atomic composition and space-group symmetry are
admitted. The appropriate weight functions and the corre-
where dk are coefficients which depend on the cause of the sponding orthogonal polynomials are
deviation of pðxÞ from the central-limit theorem approximation
and fk ðxÞ are suitably chosen functions of x. A choice of the set Non-ideal
ffk g is deemed suitable, if only from a practical point of view, if it pð0Þ ðjEjÞ fk ðxÞ distribution
allows the convenient introduction of the cause of the above
deviation of pðxÞ into the expansion coefficients dk. This ð2=Þ1=2 expðjEj2 =2Þ He2k ðjEjÞ=½ð2kÞ!1=2 Centric
requirement is satisfied – also from a theoretical point of view –
by taking fk ðxÞ as a set of polynomials which are orthogonal with
respect to the ideal p.d.f., taken as their weight function (e.g. 2jEj expðjEj2 Þ Lk ðjEj2 Þ Acentric
Cramér, 1951). That is, the functions fk ðxÞ so chosen have to obey ð2:1:7:4Þ
the relationship
204
2.2. Direct methods
By C. Giacovazzo

2.2.1. List of symbols and abbreviations 2.2.3. Origin specification


fj atomic scattering factor of jth atom (a) Once the origin has been chosen, the symmetry operators
Zj atomic number of jth atom Cs  ðRs ; Ts Þ and, through them, the algebraic form of the s.f.
N number of atoms in the unit cell remain fixed.
m order of the point group A shift of the origin through a vector with coordinates X0
P
p P
q P
N transforms ’h into
½
r p ; ½
r q ; ½
r N ; . . . ¼ Zjr ; Zjr ; Zjr ; . . .
j¼1 j¼1 j¼1 ’0h ¼ ’h  2h  X0 ð2:2:3:1Þ

½
r N is always abbreviated to
r when N is the number of atoms in and the symmetry operators Cs into C0s ¼ ðR0s ; T0s Þ, where
the cell
P P P P
p P
q P
N R0s ¼ Rs ; T0s ¼ Ts þ ðRs  IÞX0 s ¼ 1; 2; . . . ; m: ð2:2:3:2Þ
p; q; N; . . . ¼ fj2 ; fj2 ; fj2 ; . . .
j¼1 j¼1 j¼1
(b) Allowed or permissible origins (Hauptman & Karle, 1953,
s.f. structure factor 1959) for a given algebraic form of the s.f. are all those points in
n.s.f. normalized structure factor direct space which, when taken as origin, maintain the same
cs. centrosymmetric symmetry operators Cs. The allowed origins will therefore
ncs. noncentrosymmetric correspond to those points having the same symmetry environ-
s.i. structure invariant ment in the sense that they are related to the symmetry elements
s.s. structure seminvariant in the same way. For instance, if Ts ¼ 0 for s ¼ 1; . . . ; 8, then the
C ¼ ðR; TÞ symmetry operator; R is the rotational part, allowed origins in Pmmm are the eight inversion centres.
T the translational part To each functional form of the s.f. a set of permissible origins
’h phase of the structure factor Fh ¼ jFh j expði’h Þ will correspond.
(c) A translation between permissible origins will be called a
permissible or allowed translation. Trivial allowed translations
correspond to the lattice periods or to their multiples. A change
of origin by an allowed translation does not change the algebraic
2.2.2. Introduction
form of the s.f. Thus, according to (2.2.3.2), all origins allowed by
Direct methods are today the most widely used tool for solving a fixed functional form of the s.f. will be connected by transla-
small crystal structures. They work well both for equal-atom tional vectors Xp such that
molecules and when a few heavy atoms exist in the structure. In
recent years the theoretical background of direct methods has ðRs  IÞXp ¼ V; s ¼ 1; 2; . . . ; m; ð2:2:3:3Þ
been improved to take into account a large variety of prior
information (the form of the molecule, its orientation, a partial where V is a vector with zero or integer components.
structure, the presence of pseudosymmetry or of a superstructure, In centred space groups, an origin translation corresponding to
the availability of isomorphous data or of data affected by a centring vector Bv does not change the functional form of the
anomalous-dispersion effects, . . . ). Owing to this progress and to s.f. Therefore all vectors Bv represent permissible translations. Xp
the increasing availability of powerful computers, the phase will then be an allowed translation (Giacovazzo, 1974) not only
problem for small molecules has been solved in practice: a when, as imposed by (2.2.3.3), the difference T0s  Ts is equal to
number of effective, highly automated packages are today one or more lattice units, but also when, for any s, the condition
available to the scientific community.
The combination of direct methods with so-called direct-space ðRs  IÞXp ¼ V þ Bv ; s ¼ 1; 2; . . . ; m;  ¼ 0; 1 ð2:2:3:4Þ
methods have recently allowed the ab initio crystal structure
solution of proteins. The present limit of complexity is about 2500 is satisfied.
non-hydrogen atoms in the asymmetric unit, but diffraction data We will call any set of cs. or ncs. space groups having the same
at atomic resolution (~1 Å) are required. Trials are under way to allowed origin translations a Hauptman–Karle group (H–K
bring this limit to 1.5 Å and have shown some success. group). The 94 ncs. primitive space groups, the 62 primitive cs.
The theoretical background and tables useful for origin groups, the 44 ncs. centred space groups and the 30 cs. centred
specification are given in Section 2.2.3; in Section 2.2.4 the space groups can be collected into 13, 4, 14 and 5 H–K groups,
procedures for normalizing structure factors are summarized. respectively (Hauptman & Karle, 1953, 1956; Karle & Hauptman,
Phase-determining formulae (inequalities, probabilistic formulae 1961; Lessinger & Wondratschek, 1975). In Tables 2.2.3.1–2.2.3.4
for triplet, quartet and quintet invariants, and for one- and two- the H–K groups are given together with the allowed origin
phase s.s.’s, determinantal formulae) are given in Section 2.2.5. In translations.
Section 2.2.6 the connection between direct methods and related (d) Let us consider a product of structure factors
techniques in real space is discussed. Practical procedures for Q
n
A
solving small-molecule crystal structures are described in FhA11  FhA22  . . .  FhAnn ¼ Fhj j
Sections 2.2.7 and 2.2.8, and references to the most extensively j¼1
!
used packages are given in Section 2.2.9. The integration of direct P
n Q
n
methods, isomorphous replacement and anomalous-dispersion ¼ exp i Aj ’hj jFhj jAj ; ð2:2:3:5Þ
techniques is briefly discussed in Section 2.2.10. j¼1 j¼1

The reader interested in a more detailed description of the


topic is referred to a recent textbook (Giacovazzo, 1998). Aj being integer numbers.
215
2.2. DIRECT METHODS
SAPI: Fan, H.-F. (1999). Crystallographic software: (2) low-resolution data create additional problems for direct
teXsan for Windows. http://www.rigaku.com/downloads/journal/ methods since the number of available phase relationships per
Vol15.1.1998/texsan.pdf. reflection is small.
SnB: Weeks, C. M. & Miller, R. (1999). The design and Sheldrick (1990) suggested that direct methods are not
implementation of SnB version 2.0. J. Appl. Cryst. 32, 120–124. expected to succeed if fewer than half of the reflections in the
SHELX97 and SHELXS: Sheldrick, G. M. (2000). The range 1.1–1.2 Å are observed with jFj > 4
ðjFjÞ (a condition
SHELX home page. http://shelx.uni-ac.gwdg.de/SHELX/. seldom satisfied by protein data).
SHELXD: Sheldrick, G. M. (1998). SHELX: applications to The most complete analysis of the problem has been made by
macromolecules. In Direct methods for solving macromolecular Giacovazzo, Guagliardi et al. (1994). They observed that the
structures, edited by S. Fortier, pp. 401–411. Dordrecht: Kluwer expected value of  (see Section 2.2.7) suggested by the tangent
Academic Publishers. formula for proteins is comparable with the variance of the 
SIR97: Altomare, A., Burla, M. C., Camalli, M., Cascarano, parameter. In other words, for proteins the signal determining the
G. L., Giacovazzo, C., Guagliardi, A., Moliterni, A. G. G., phase is comparable with the noise, and therefore the phase
Polidori, G. & Spagna, R. (1999). SIR97: a new tool for crystal indication is expected to be unreliable.
structure determination and refinement. J. Appl. Cryst. 32, 115– Quite relevant results have recently been obtained by inte-
119. grating direct methods with some additional experimental
SIR2004: Burla, M. C., Caliandro, R., Camalli, M., Carrozzini, information. In particular, we will describe the combination of
B., Cascarano, G. L., De Caro, L., Giacovazzo, C., Polidori, G. & direct methods with:
Spagna, R. (2005). SIR2004: an improved tool for crystal structure (a) direct-space techniques for the ab initio crystal structure
determination and refinement. J. Appl. Cryst. 38, 381–388. solution of proteins;
XTAL3.6.1: Hall, S. R., du Boulay, D. J. & Olthof-Hazekamp, (b) isomorphous-replacement (SIR–MIR) techniques;
R. (1999). Xtal3.6 crystallographic software. http://xtal. (c) anomalous-dispersion (SAD–MAD) techniques;
sourceforge.net/. (d) molecular replacement.
Point (d) will not be treated here, as it is described extensively in
IT F, Part 13.

2.2.10. Direct methods in macromolecular crystallography 2.2.10.2. Ab initio crystal structure solution of proteins
2.2.10.1. Introduction Ab initio techniques do not require prior information of any
The smallest protein molecules contain about 400 non- atomic positions. The recent tremendous increase in computing
hydrogen atoms, so they cannot be solved ab initio by the algo- speed led to direct methods evolving towards the rapid devel-
rithms specified in Sections 2.2.7 and 2.2.8. However, traditional opment of multisolution techniques. The new algorithms of the
direct methods are applied for: program Shake-and-Bake (Weeks et al., 1994; Weeks & Miller,
(a) improvement of the accuracy of the available phases 1999; Hauptman et al., 1999) allowed an impressive extension of
(refinement process); the structural complexity amenable to direct phasing. In parti-
(b) extension of phases from lower to higher resolution (phase- cular we mention: (a) the minimal principle (De Titta et al., 1994),
extension process). according to which the phase problem is considered as a
The application of standard tangent techniques to (a) and (b) constrained global optimization problem; (b) the refinement
has not been found to be very satisfactory (Coulter & Dewar, procedure, which alternately uses direct- and reciprocal-space
1971; Hendrickson et al., 1973; Weinzierl et al., 1969). Tangent techniques; and (c) the parameter-shift optimization technique
methods, in fact, require atomicity and non-negativity of the (Bhuiya & Stanley, 1963), which aims at reducing the value of the
electron density. Both these properties are not satisfied if data do minimal function (Hauptman, 1991; De Titta et al., 1994). An
not extend to atomic resolution (d > 1.2 Å). Because of series effective variant of Shake-and Bake is SHELXD (Sheldrick,
termination and other errors the electron-density map at d > 1998) which cyclically alternates tangent refinement in reciprocal
1.2 Å presents large negative regions which will appear as false space with peak-list optimisation procedures in real space
peaks in the squared structure. However, tangent methods use (Sheldrick & Gould, 1995). Detailed information on these
only a part of the information given by the Sayre equation programs is available in IT F (2001), Part 16.
(2.2.6.5). In fact, (2.2.6.5) express two equations relating the A different approach is used by ACORN (Foadi et al., 2000),
radial and angular parts of the two sides, so obtaining a large which first locates a small fragment of the molecule (eventually
degree of overdetermination of the phases. To achieve this Sayre by molecular-replacement techniques) to obtain a useful
(1972) [see also Sayre & Toupin (1975)] suggested minimizing nonrandom starting set of phases, and then refines them by
(2.2.10.1) by least squares as a function of the phases: means of solvent-flattening techniques.
The program SIR2004 (Burla et al., 2005) uses the tangent
 2
P P 

formula as well as automatic Patterson techniques to obtain a first
ah Fh  Fk Fhk  : ð2:2:10:1Þ imperfect structural model; then direct-space techniques are used
h k
to refine the model. The Patterson approach is based on the use
of the superposition minimum function (Buerger, 1959;
Even if tests on rubredoxin (extensions of phases from 2.5 to Richardson & Jacobson, 1987; Sheldrick, 1992; Pavelcı́k, 1988;
1.5 Å resolution) and insulin (Cutfield et al., 1975) (from 1.9 to Pavelcı́k et al., 1992; Burla et al., 2004). It may be worth noting
1.5 Å resolution) were successful, the limitations of the method that even this approach is of multisolution type: up to 20 trial
are its high cost and, especially, the higher efficiency of the least- solutions are provided by using as pivots the highest maxima in
squares method. Equivalent considerations hold for the appli- the superposition minimum function.
cation of determinantal methods to proteins [see Podjarny et al. It is today possible to solve structures up to 2500 non-hydrogen
(1981); de Rango et al. (1985) and literature cited therein]. atoms in the asymmetric unit provided data at atomic (about 1 Å)
A question now arises: why is the tangent formula unable to resolution are available. Proteins with data at quasi-atomic
solve protein structures? Fan et al. (1991) considered the question resolution (say up to 1.5–1.6 Å) can also be solved, but with
from a first-principle approach and concluded that: greater difficulties (Burla et al., 2005). A simple evaluation of the
(1) the triplet phase probability distribution is very flat for potential of the ab initio techniques suggests that the structural
proteins (N is very large) and close to the uniform distribution; complexity range and the resolution limits amenable to the ab
235
2.3. Patterson and molecular replacement techniques, and the use of noncrystallographic
symmetry in phasing

By L. Tong, M. G. Rossmann and E. Arnold

2.3.1. Introduction space group to the Laue symmetry is produced by the translation
2.3.1.1. Background of all vectors to the Patterson origin and the introduction of a
centre of symmetry. The latter is a consequence of the relation-
Historically, the Patterson has been used in a variety of ways to ship between the vectors AB and BA. The Patterson symmetries
effect the solutions of crystal structures. While some simple for all 230 space groups are tabulated in IT A (2005).
structures (Ketelaar & de Vries, 1939; Hughes, 1940; Speakman, An analysis of Patterson peaks can be obtained by considering
1949; Shoemaker et al., 1950) were solved by direct analysis of N atoms with form factors fi in the unit cell. Then
Patterson syntheses, alternative methods have largely superseded
this procedure. An early innovation was the heavy-atom method P
N
which depends on the location of a small number of relatively Fh ¼ fi expð2ih  xi Þ:
strong scatterers (Harker, 1936). Image-seeking methods and i¼1
Patterson superposition techniques were first contemplated in
the late 1930s (Wrinch, 1939) and applied sometime later Using Friedel’s law,
(Beevers & Robertson, 1950; Clastre & Gay, 1950; Garrido,
1950a; Buerger, 1959). This experience provided the encourage-
ment for computerized vector-search methods to locate indivi- jFh j2 ¼ Fh  Fh
dual atoms automatically (Mighell & Jacobson, 1963; Kraut, 1961; N " N #
P P
Hamilton, 1965; Simpson et al., 1965) or to position known ¼ fi expð2ih  xi Þ fj expð2ih  xj Þ ;
molecular fragments in unknown crystal structures (Nordman & i¼1 j¼1

Nakatsu, 1963; Huber, 1965). The Patterson function has been


used extensively in conjunction with the isomorphous replace-
ment method (Rossmann, 1960; Blow, 1958) or anomalous which can be decomposed to
dispersion (Rossmann, 1961a) to determine the position of
heavy-atom substitution. Pattersons have been used to detect the P
N P
N P
N
jFh j2 ¼ fi2 þ fi fj exp½2ih  ðxi  xj Þ: ð2:3:1:3Þ
presence and relative orientation of multiple copies of a given i¼1 i6¼j
chemical motif in the crystallographic asymmetric unit in the
same or different crystals (Rossmann & Blow, 1962). Finally, the
orientation and placement of known molecular structures On substituting (2.3.1.3) in (2.3.1.2), we see that the Patterson
(‘molecular replacement’) into unknown crystal structures can be consists of the sum of N 2 total interactions of which N are of
accomplished via Patterson techniques. weight fi2 at the origin and NðN  1Þ are of weight fi fj at xi  xj .
The function, introduced by Patterson in 1934 (Patterson, The weight of a peak in a real cell is given by
1934a,b), is a convolution of electron density with itself and may R
be defined as wi ¼ i ðxÞ dx ¼ Zi ðthe atomic numberÞ;
U
R
PðuÞ ¼ ðxÞ  ðu þ xÞ dx; ð2:3:1:1Þ
V
where U is the volume of the atom i. By analogy, the weight of a
peak in a Patterson (form factor fi fj ) will be given by
where PðuÞ is the ‘Patterson’ function at u, ðxÞ is the crystal’s R
periodic electron density and V is the volume of the unit cell. The wij ¼ Pij ðuÞ du ¼ Zi Zj :
Patterson function, or F 2 series, can be calculated directly from U
the experimentally derived X-ray intensities as
Although the maximum height of a peak will depend on the
2 X
hemisphere
spread of the peak, it is reasonable to assume that heights of
2
PðuÞ ¼ 2 jFh j cos 2h  u: ð2:3:1:2Þ
V h
peaks in a Patterson are proportional to the products of the
atomic numbers of the interacting atoms.
There are a total of N 2 interactions in a Patterson due to N
The derivation of (2.3.1.2) from (2.3.1.1) can be found in this atoms in the crystal cell. These can be represented as an N  N
volume (see Section 1.3.4.2.1.6) along with a discussion of the square matrix whose elements uij, wij indicate the position and
physical significance and symmetry of the Patterson function, weight of the peak produced between atoms i and j (Table
although the principal properties will be restated here. 2.3.1.1). The N vectors corresponding to the diagonal of this
The Patterson can be considered to be a vector map of all the matrix are located at the Patterson origin and arise from the
pairwise interactions between the atoms in a unit cell. The vectors convolution of each atom with itself. This leaves NðN  1Þ
in a Patterson correspond to vectors in the real (direct) crystal vectors whose locations depend on the relative positions of all of
cell but translated to the Patterson origin. Their weights are the atoms in the crystal cell and whose weights depend on the
proportional to the product of densities at the tips of the vectors atom types related by the vector. Complete specification of the
in the real cell. The Patterson unit cell has the same size as the unique non-origin Patterson vectors requires description of only
real crystal cell. The symmetry of the Patterson comprises the the NðN  1Þ=2 elements in either the upper or the lower triangle
Laue point group of the crystal cell plus any additional lattice of this matrix, since the two sets of vectors represented by
symmetry due to Bravais centring. The reduction of the real the two triangles are related by a centre of symmetry
244
2.4. Isomorphous replacement and anomalous scattering

By M. Vijayan and S. Ramaseshan†

2.4.1. Introduction interesting applications (Koetzle & Hamilton, 1975; Sikka &
Isomorphous replacement is among the earliest methods to be Rajagopal, 1975). More recently there has been a further revival
employed for crystal structure determination (Cork, 1927). The in the development of anomalous-scattering methods with the
power of this method was amply demonstrated in the classical advent of synchrotron radiation, particularly in view of
X-ray work of J. M. Robertson on phthalocyanine in the 1930s the possibility of choosing any desired wavelength from a
using centric data (Robertson, 1936; Robertson & Woodward, synchrotron-radiation source (Helliwell, 1984).
1937). The structure determination of strychnine sulfate penta- It is clear from the foregoing that the isomorphous replace-
hydrate by Bijvoet and others provides an early example of the ment and the anomalous-scattering methods have a long and
application of this method to acentric reflections (Bokhoven et distinguished history. It is therefore impossible to do full justice
al., 1951). The usefulness of isomorphous replacement in the to them in a comparatively short presentation like the present
analysis of complex protein structures was demonstrated by one. Several procedures for the application of these methods
Perutz and colleagues (Green et al., 1954). This was closely have been developed at different times. Many, although of
followed by developments in the methodology for the application considerable historical importance, are not extensively used at
of isomorphous replacement to protein work (Harker, 1956; present for a variety of reasons. No attempt has been made to
Blow & Crick, 1959) and rapidly led to the first ever structure discuss them in detail here; the emphasis is primarily on the state
solution of two related protein crystals, namely, those of of the art as it exists now. The available literature on isomorphous
myoglobin and haemoglobin (Kendrew et al., 1960; Cullis et al., replacement and anomalous scattering is extensive. The refer-
1961b). Since then isomorphous replacement has been the ence list given at the end of this part is representative rather than
method of choice in macromolecular crystallography and most of exhaustive.
the subsequent developments in and applications of this method During the past few years, rapid developments have taken
have been concerned with biological macromolecules, mainly place in the isomorphous replacement and anomalous-scattering
proteins (Blundell & Johnson, 1976; McPherson, 1982). methods, particularly in the latter, as applied to macromolecular
The application of anomalous-scattering effects has often crystallography. These developments are described in detail in
developed in parallel with that of isomorphous replacement. International Tables for Crystallography, Volume F (2001).
Indeed, the two methods are complementary to a substantial Therefore, they have not been dealt with in this chapter. Signif-
extent and they are often treated together, as in this article. icant developments in applications of direct methods to macro-
Although the most important effect of anomalous scattering, molecular crystallography have also occurred in recent years. A
namely, the violation of Friedel’s law, was experimentally summary of these developments as well as the traditional direct
observed as early as 1930 (Coster et al., 1930), two decades methods on which the recent progress is based are presented in
elapsed before this effect was made use of for the first time by Chapter 2.2.
Bijvoet and his associates for the determination of the absolute
configuration of asymmetric molecules as well as for phase
2.4.2. Isomorphous replacement method
evaluation (Bijvoet, 1949, 1954; Bijvoet et al., 1951). Since then
there has been a phenomenal spurt in the application of 2.4.2.1. Isomorphous replacement and isomorphous addition
anomalous-scattering effects (Srinivasan, 1972; Ramaseshan & Two crystals are said to be isomorphous if (a) both have the
Abrahams, 1975; Vijayan, 1987). A quantitative formulation for same space group and unit-cell dimensions and (b) the types and
the determination of phase angles using intensity differences the positions of atoms in both are the same except for a repla-
between Friedel equivalents was derived by Ramachandran & cement of one or more atoms in one structure with different types
Raman (1956), while Okaya & Pepinsky (1956) successfully of atoms in the other (isomorphous replacement) or the presence
developed a Patterson approach involving anomalous effects. of one or more additional atoms in one of them (isomorphous
The anomalous-scattering method of phase determination has addition). Consider two crystal structures with identical space
since been used in the structure analysis of several structures, groups and unit-cell dimensions, one containing N atoms and the
including those of a complex derivative of vitamin B12 (Dale et al., other M atoms. The N atoms in the first structure contain subsets
1963) and a small protein (Hendrickson & Teeter, 1981). In the P and Q whereas the M atoms in the second structure contain
meantime, the effect of changes in the real component of the subsets P, Q0 and R. The subset P is common to both structures in
dispersion correction as a function of the wavelength of the terms of atomic positions and atom types. The atomic positions
radiation used, first demonstrated by Mark & Szillard (1925), also are identical in subsets Q and Q0 , but at any given atomic position
received considerable attention. This effect, which is formally the atom type is different in Q and Q0 . The subset R exists only in
equivalent to that of isomorphous replacement, was demon- the second structure. If FN and FM denote the structure factors of
strated to be useful in structure determination (Ramaseshan et the two structures for a given reflection,
al., 1957; Ramaseshan, 1963). Protein crystallographers have
been quick to exploit anomalous-scattering effects (Rossmann, FN ¼ FP þ FQ ð2:4:2:1Þ
1961; Kartha & Parthasarathy, 1965; North, 1965; Matthews, 1966;
Hendrickson, 1979) and, as in the case of the isomorphous
replacement method, the most useful applications of anomalous and
scattering during the last two decades have been perhaps in the
field of macromolecular crystallography (Kartha, 1975; Waten- FM ¼ FP þ FQ 0 þ FR ; ð2:4:2:2Þ
paugh et al., 1975; Vijayan, 1981). In addition to anomalous
scattering of X-rays, that of neutrons was also found to have
where the quantities on the right-hand side represent contribu-
† Deceased. tions from different subsets. From (2.4.2.1) and (2.4.2.2) we have
282
2.5. Electron diffraction and electron microscopy in structure determination

By J. M. Cowley,† J. C. H. Spence, M. Tanaka, B. K. Vainshtein,† B. B. Zvyagin,† P. A. Penczek


and D. L. Dorset

2.5.1. Foreword dynamical diffraction effects as the basis for obtaining crystal
By J. M. Cowley and J. C. H. Spence structure information. The fact that dynamical diffraction is
dependent on the relative phases of the diffracted waves then
Given that electrons have wave properties and the wavelengths implies that relative phase information can be deduced from the
lie in a suitable range, the diffraction of electrons by matter is diffraction intensities and the limitations of kinematical diffrac-
completely analogous to the diffraction of X-rays. While for tion, such as Friedel’s law, do not apply. The most immediately
X-rays the scattering function is the electron-density distribution, practicable method for making use of this possibility is conver-
for electrons it is the potential distribution which is similarly gent-beam electron diffraction (CBED) as described in Section
peaked at the atomic sites. Hence, in principle, electron diffrac- 2.5.3.
tion may be used as the basis for crystal structure determination. A further important factor, determining the methods for
In practice it is used much less widely than X-ray diffraction for observing electron diffraction, is that, being charged particles,
the determination of crystal structures but is receiving increasing electrons can be focused by electromagnetic lenses. Many of the
attention as a means for obtaining structural information not resolution-limiting aberrations of cylindrical magnetic lenses
readily accessible with X-ray- or neutron-diffraction techniques. have now been eliminated through the use of aberration-
Electrons having wavelengths comparable with those of the correction devices, so that for weakly scattering samples the
X-rays commonly used in diffraction experiments have energies resolution is limited to about 1 Å by electronic and mechanical
of the order of 100 eV. For such electrons, the interactions with instabilities. This is more than sufficient to distinguish the indi-
matter are so strong that they can penetrate only a few layers of vidual rows of atoms, parallel to the incident beam, in the prin-
atoms on the surfaces of solids. They are used extensively for the cipal orientations of most crystalline phases. Thus ‘structure
study of surface structures by low-energy electron diffraction images’ can be obtained, sometimes showing direct representa-
(LEED) and associated techniques. These techniques are not tion of projections of crystal structures [see IT C (2004), Section
covered in this series of volumes, which include the principles and 4.3.8]. However, the complications of dynamical scattering and of
practice of only those diffraction and imaging techniques making the coherent imaging processes are such that the image inten-
use of high-energy electrons, having energies in the range of sities vary strongly with crystal thickness and tilt, and with the
20 keV to 1 MeV or more, in transmission through thin speci- defocus or other parameters of the imaging system, making the
mens. interpretation of images difficult except in special circumstances.
For the most commonly used energy ranges of high-energy Fortunately, computer programs are readily available whereby
electrons, 100 to 400 keV, the wavelengths are about 50 times image intensities can be calculated for model structures [see IT C
smaller than for X-rays. Hence the scattering angles are much (2004), Section 4.3.6]. Hence the means exist for deriving the
smaller, of the order of 102 rad, the recording geometry is projection of the structure if only by a process of trial and error
relatively simple and the diffraction pattern represents, to a and not, as would be desirable, from a direct interpretation of the
useful first approximation, a planar section of reciprocal space. observations.
Extinction distances are hundreds of ångstroms, which, when The accuracy with which the projection of a structure can be
combined with typical lattice spacings, produces rocking-curve deduced from an image, or series of images, improves as the
widths which are, unlike the X-ray case, a significant fraction of resolution of the microscope improves but is not at all compar-
the Bragg angle. able with the accuracy attainable with X-ray diffraction methods.
The elastic scattering of electrons by atoms is several orders of A particular virtue of high-resolution electron microscopy as a
magnitude greater than for X-rays. This fact has profound structural tool is that it may give information on individual small
consequences, which in some cases are highly favourable and in regions of the sample. Structures can be determined of ‘phases’
other cases are serious hindrances to structure analysis work. On existing over distances of only a few unit cells and the defects and
the one hand it implies that electron-diffraction patterns can be local disorders can be examined, one by one.
obtained from very small single-crystal regions having thick- The observation of electron-diffraction patterns forms an
nesses equal to only a few layers of atoms and, with recently essential part of the technique of structure imaging in high-
developed techniques, having diameters equivalent to only a few resolution electron microscopy, because the diffraction patterns
interatomic distances. Hence single-crystal patterns can be are used to align the crystals to appropriate axial orientations.
obtained from microcrystalline phases. More generally, for all electron microscopy of crystalline mate-
However, the strong scattering of electrons implies that the rials the image interpretation depends on knowledge of the
simple kinematical single-scattering approximation, on which diffraction conditions. Fortunately, the diffraction pattern and
most X-ray diffraction structure analysis is based, fails for elec- image of any specimen region can be obtained in rapid succession
trons except for very thin crystals composed of light-atom by a simple switching of lens currents. The ready comparison of
materials. Strong dynamical diffraction effects occur for crystals the image and diffraction data has become an essential compo-
which may be 100 Å thick, or less for heavy-atom materials. As nent of the electron microscopy of crystalline materials but has
also been of fundamental importance for the development of
a consequence, the theory of dynamical diffraction for electrons
electron-diffraction theory and techniques. The development of
has been well developed, particularly for the particular special the nanodiffraction method in the field-emission scanning
diffracting conditions relevant to the transmission of fast elec- transmission electron microscope (STEM) has allowed micro-
trons (see Chapter 5.2), and observations of dynamical diffrac- diffraction patterns to be obtained from subnanometre-sized
tion effects are commonly made and quantitatively interpreted. regions, and so has become the ideal tool for the structural
The possibility has thus arisen of using the observation of analysis of the new microcrystalline phases important to
nanoscience. The direct phasing of these coherent nanodiffrac-
† Deceased. tion patterns is an active field of research.
297
2. RECIPROCAL SPACE IN CRYSTAL-STRUCTURE DETERMINATION
Table 2.5.3.9. Dynamical extinction lines appearing in ZOLZ reflections for all crystal space groups except Nos. 1 and 2
Point groups 2, m, 2/m (second setting, unique axis b)
Incident-beam
direction
Space group [h0l]
3 P2
4 P21 0k0 A2 B2
21 B3
5 C2
6 Pm
7 Pc h0lo A2 B2
c A3
8 Cm
9 Cc he0lo A2 B2
c A3
10 P2/m
11 P21/m 0k0 A2 B2
21 B3
12 C2/m
13 P2/c h0lo A2 B2
c A3
14 P21/c 0k0 A2 B2
21 B3
h0lo A2 B2
c A3
15 C2/c he0lo A2 B2
c A3

Point group 222


Incident-beam direction
Space group [100] [010] [001] [hk0] [0kl] [h0l]
16 P222
17 P2221 00l A2 B2 00l A2 B2 00l A2 B2
21 B3 21 B3 21 B3
18 P21212 0k0 A2 B2 h00 A2 B2 h00 A2 B2 h00 A2 B2 0k0 A2 B2
212 B3 211 B3 211 B3 211 B3 212 B3
0k0
212
19 P212121 0k0 A2 B2 h00 A2 B2 h00 A2 B2 00l A2 B2 h00 A2 B2 0k0 A2 B2
212 B3 211 B3 211 B3 213 B3 211 B3 212 B3
00l 00l 0k0
213 213 212
20 C2221 00l A2 B2 00l A2 B2 00l A2 B2
21 B3 21 B3 21 B3
21 C222
22 F222
23 I222
24 I212121

Point group mm2


Incident-beam direction
Space group [100] [010] [001] [hk0] [0kl] [h0l]
25 Pmm2
26 Pmc21 00l A2 B2 00l 00l A2 B2 h0lo A2 B2
c, 21 A3 B3 21 B3 21 B3 c A3
27 Pcc2 00l 00l 0klo A2 B2 h0lo A2 B2
c2 A3 c1 A3 c1 A3 c2 A3
28 Pma2 h00 A2 B2 ho0l A2 B2
a A3 a A3
29 Pca21 00l 00l A2 B2 h00 A2 B2 00l A2 B2 0klo A2 B2 ho0l A2 B2
21 B3 c, 21 A3 B3 a A3 21 B3 c A3 a A3

324
2. RECIPROCAL SPACE IN CRYSTAL-STRUCTURE DETERMINATION

Fig. 2.5.3.16. CBED patterns of Sm3Se4. The procedures for identifying the symmetry are also shown. (a, b) [111] incidence at 80 kV: the WP symmetry is 3m (a)
and the projection (proj.) WP symmetry is 3m (b). (c, d) [100] incidence at 80 kV: the WP symmetry is 2mm (c) and the projection WP symmetry is 4mm (d).
Dynamical extinction lines A2 and A3 are seen (d). The point group is determined to be 4 3m. (e) Spot diffraction pattern taken with the [100] incidence at 80 kV
shows the absence of 0kl reflections. The lattice type is determined to be I. ( f ) [100] incidence at 100 kV: dynamical extinction lines A in HOLZ
reflections confirm the existence of a glide plane. The space group is determined to be I 4 3d.

346
2.5. ELECTRON DIFFRACTION AND ELECTRON MICROSCOPY IN STRUCTURE DETERMINATION
cation second’ approach. It is possible to reverse this order by
using invariants with the supporting rationale that once
approximately homogeneous classes of images were found, it
should be easy to align them subsequently as within each class
they will share the same motif.
A practical approach to reference-free alignment known as
alignment by classification (Dube et al., 1993) is based on the
observation that for a very large data set and centred particles
one can expect that although the in-plane rotation is arbitrary,
there is a high chance that at least some of the similar images will
be in the same rotational orientation. Therefore, in this approach
the images are first (approximately) centred, then subjected to
classification, and subsequently aligned.
In its simplest form, the multireference alignment belongs to
the class of supervised classification methods: given a set of
templates (i.e., reference images; these can be selected unpro-
cessed particle projections, or class averages that resulted from
preceding analysis, or projections of a previously determined EM
structure, or projections of an X-ray crystallographic structure),
each of the images from the available data sets is compared
(using a selected discrepancy measure) with all templates and
assigned to the class represented by the most similar one. Equally
often multireference alignment is understood as a form of
unsupervised classification, more precisely K-means classification,
even if the description is not formalized in terms of the latter.
Given a number of initial 2D templates, the images are compared
with all templates and assigned to the most similar one. New
templates are calculated by averaging images assigned to their
predecessors and the whole procedure is repeated until a stable Fig. 2.5.7.3. Principle of random conical tilt reconstruction. A tilt pair of
images of the same grid area is collected. By aligning the particle images in
solution is reached. the untilted micrograph (left), the Eulerian angles of their counterparts in
the tilted micrograph (right) are established. The particle images from the
tilted micrograph are used for 3D reconstruction of the molecule (bottom).
2.5.7.7. Initial determination of 3D structure using tilt experiments The set of projections form a cone in Fourier space; information within the
The 2D analysis of projection images provides insight into the cone remains undetermined.
behaviour of the protein on the grid in terms of the structural
consistency and the number and shape of projection images. In the problem is difficult as the tilt data tend to be very noisy and
order to obtain 3D information, it is necessary to find geometrical have very low contrast.
relations between different observed 2D images. The most robust Given three Eulerian angles and centred tilted projections, a
and historically the earliest approach is based on tilt experiments. 3D reconstruction is calculated. There are numerous advantages
By tilting the stage in the microscope and acquiring additional of the RCT method. (i) Assuming the sign of the tilt angle is read
pictures of the same area of the grid it is possible to collect correctly (it can be confirmed by analysing the defocus gradient
projection images of the same molecule with some of the in the tilted micrographs), the method yields a correct hand of the
required Eulerian angles determined accurately by the setting of structure. (ii) With the exception of the in-plane rotation of
the goniometer of the microscope. untilted projections, which can be found relatively easily using
In random conical tilt (RCT) reconstruction (Radermacher et alignment procedures, the remaining parameters are determined
al., 1987), two micrographs of the same specimen area are by the experimental settings. Even if they are not extremely
collected: the first one is recorded at a tilt angle of ~50 while the accurate, the possibility of a gross error is eliminated, which
second one is recorded at 0 (Fig. 2.5.7.3). If particles have positively distinguishes the method from the ab initio computa-
preferred orientation on the support carbon film (or within the tional approaches that use only untilted data. (iii) The compu-
amorphous ice layer, if no carbon support is used), the projec- tational analysis is entirely done using the untilted data, which
tions of particles in the tilted micrographs form a conical tilt have high contrast. (iv) The RCT method is often the only
series. Since in-plane rotations of particles are random, the method of obtaining 3D information if the molecule has strongly
azimuthal angles of the projections of tilted particles are also preferential orientation and only one view is observed in untilted
randomly distributed; hence the name of the method. The micrographs. The main disadvantage is that the conical projection
untilted image is required for two reasons: (i) the particle series leaves a significant portion of the Fourier space unde-
projections from the untilted image are classified, thus a subset termined. This follows from the central section theorem [equa-
corresponding to possibly identical images can be selected tion (2.5.6.8) of Section 2.5.6]: as the tilt angle is less than 90 , the
ensuring that the projections originated from similar and simi- undetermined region can be thought to form a cone in three
larly oriented structures; and (ii) the in-plane rotation angle dimensions and is referred to as the missing cone. The problem
found during alignment corresponds to the azimuthal angles in can be overcome if the molecule has more than one preferred
three dimensions (one of the three Eulerian angles needed). The orientation. Subsets of particles that have similar untilted
second Eulerian angle, the tilt, is either taken from the micro- appearance (as determined by clustering) are processed inde-
scope setting of the goniometer or calculated based on geome- pendently and for each a separate 3D structure is calculated. If
trical relations between tilted and untilted micrographs. The third the preferred orientations are sufficiently different, i.e., the
Eulerian angle corresponds to the angle of the tilt axis of the orientations of the original particles in three dimensions are
microscope stage and is also calculated using the geometrical sufficiently different in terms of their angles with respect to the z
relations between two micrographs. In addition, it is necessary to axis, the 3D structures can be aligned and merged, all but elim-
centre the particle projections selected from tilted micrographs; inating the problem of the missing cone and yielding a robust, if
although various correlation-based schemes have been proposed, resolution-limited, initial model of the molecule (Penczek et al.,
381
3.1. Distances, angles, and their standard uncertainties

By D. E. Sands

3.1.1. Introduction Subscripts are used for quantities that transform the same way as
A crystal structure analysis provides information from which it is the basis vectors ai ; such quantities are said to transform covar-
possible to compute distances between atoms, angles between iantly. Superscripts denote quantities that transform the same
interatomic vectors, and the uncertainties in these quantities. In way as coordinates xi ; these quantities are said to transform
Cartesian coordinate systems, these geometric computations contravariantly (Sands, 1982a).
require the Pythagorean theorem and elementary trigonometry. Equation (3.1.2.4) is in a form convenient for computer
The natural coordinate systems of crystals, though, are deter- evaluation, with indices i and j taking successively all values from
mined by symmetry, and only in special cases are the basis vectors 1 to 3. The matrix form of (3.1.2.4) is useful both for symbolic
(or coordinate axes) of these systems constrained to be of equal manipulation and for computation,
lengths or mutually perpendicular.
It is possible, of course, to transform the positional parameters u  v ¼ uT gv; ð3:1:2:6Þ
of the atoms to a Cartesian system and perform the subsequent
calculations with the transformed coordinates. Along with the
where the superscript italic T following a matrix symbol indicates
coordinates, the transformations must be applied to anisotropic
a transpose. Written out in full, (3.1.2.6) is
thermal factors, variance–covariance matrices and other impor-
tant quantities. Moreover, leaving the natural coordinate system 0 10 1 1
of the crystal sacrifices the simplified relationships imposed by g11 g12 g13 v
translational and point symmetry; for example, if an atom has u  v ¼ ðu1 u2 u3 Þ@ g21 g22 g23 A@ v2 A: ð3:1:2:7Þ
fractional coordinates x1, x2 , x3 , an equivalent atom will be at g31 g32 g33 v3
1 þ x1 , x2 , x3 , etc.
Fortunately, formulation of the calculations in generalized
rectilinear coordinate systems is straightforward, and readily If u is the column vector with components u1 ; u2 ; u3, uT is the
adapted to computer languages (Section 3.1.12 illustrates the use corresponding row vector shown in (3.1.2.7).
of Fortran for such calculations). The techniques for these
computations are those of tensor analysis, which provides a
compact and elegant notation. While an effort will be made to be 3.1.3. Length of a vector
self-sufficient in this chapter, some proficiency in vector algebra is By (3.1.2.1), the scalar product of a vector with itself is
assumed, and the reader not familiar with the basics of tensor
analysis should refer to Chapter 1.1 and Sands (1982a). v  v ¼ ðvÞ2 : ð3:1:3:1Þ

3.1.2. Scalar product The length of v is, therefore, given by


The scalar product of vectors u and v is defined as
v ¼ ðvi v j gij Þ1=2 : ð3:1:3:2Þ
u  v ¼ uv cos ’; ð3:1:2:1Þ
Computation of lengths in a generalized rectilinear coordinate
system is thus simply a matter of evaluating the double summa-
where u and v are the lengths of the vectors and ’ is the angle tion vi v j gij and taking the square root.
between them. In terms of components,

u  v ¼ ðui ai Þ  ðv j aj Þ ð3:1:2:2Þ 3.1.4. Angle between two vectors


i j
u  v ¼ u v ai  aj ð3:1:2:3Þ By (3.1.2.1) and (3.1.2.4), the angle ’ between vectors u and v is
i j given by
u  v ¼ u v gij : ð3:1:2:4Þ
’ ¼ cos1 ½ui v j gij =ðuvÞ: ð3:1:4:1Þ
In all equations in this chapter, the convention is followed that
summation is implied over an index that is repeated once as a
An even more concise expression of equations such as (3.1.4.1) is
subscript and once as a superscript in an expression; thus, the
possible by making use of the ability of the metric tensor g to
right-hand side of (3.1.2.4) implies the sum of nine terms
convert components from contravariant to covariant (Sands,
1982a). Thus,
u1 v1 g11 þ u1 v2 g12 þ . . . þ u3 v3 g33 :
vi ¼ gij v j ; uj ¼ gij ui ; ð3:1:4:2Þ
The gij in (3.1.2.4) are the components of the metric tensor [see
Chapter 1.1 and Sands (1982a)] and (3.1.2.4) may be written succinctly as

gij ¼ ai  aj : ð3:1:2:5Þ u  v ¼ ui vi ð3:1:4:3Þ

404
3.2. The least-squares plane

By R. E. Marsh and V. Schomaker†

3.2.1. Introduction times go astray. As to the propagation of errors, numerous


By way of introduction, we remark that in earlier days of crystal treatments have been given, but none that we have seen is
structure analysis, before the advent of high-speed computers and altogether satisfactory.
routine three-dimensional analyses, molecular planarity was We refer all vectors and matrices to Cartesian axes, because
often assumed so that atom coordinates along the direction of that is the most convenient in calculation. However, a more
projection could be estimated from two-dimensional data [see, elegant formulation can be written in terms of general axes [e.g.,
e.g., Robertson (1948)]. Today, the usual aim in deriving the as in Shmueli (1981)].
coefficients of a plane is to investigate the degree of planarity of a The notation is troublesome. Indices are needed for atom
group of atoms as found in a full, three-dimensional structure number and Cartesian direction, and the exponent 2 is needed as
determination. We further note that, for such purposes, a crys- well, which is difficult if there are superscript indices. The best
tallographer will often be served just as well by establishing the way seems to be to write all the indices as subscripts and
plane in an almost arbitrary fashion as by resorting to the most distinguish among them by context – i, j, 1, 2, 3 for directions; k, l,
elaborate, nit-picking and pretentious least-squares treatment. p (and sometimes K, . . . ) for atoms. In any case, atom first then
The approximate plane and the associated perpendicular direction if there are two subscripts; direction, if only one index
distances of the atoms from it will be all he needs as scaffolding for a vector component, but atom (in this section at least) if for a
for his geometrical and structural imagination; reasonable weight or a vector. And
d1 , e.g., for the standard uncertainty of
common sense will take the place of explicit attention to error the distance of atom 1 from a plane. For simplicity in practice, we
estimates. use Cartesian coordinates throughout.
Nevertheless, we think it appropriate to lay out in some detail The first task is to find the plane, which we write as
the derivation of the ‘best’ plane, in a least-squares sense,
through a group of atoms and of the standard uncertainties 0 ¼ m  r  d  mT r  d;
associated with this plane. We see two cases: (1) The weights of
the atoms in question are considered to be isotropic and uncor-
related (i.e. the weight matrix for the positions of all the atoms is where r is here the vector from the origin to any point on the
diagonal, when written in terms of Cartesian axes, and for each plane (but usually represents the measured position of an atom),
atom the three diagonal elements are equal). In such cases the m is a unit vector parallel to the normal from the origin to the
weights may have little or nothing to do with estimates of random plane, d is the length of the normal, and m and r are the column
error in the atom positions (they may have been assigned merely representations of m and r. The least-squares condition is to find
for convenience or convention), and, therefore, no one should the stationary values of S  ½wk ðmT rk  dÞ2  subject to mT m ¼ 1,
feel that the treatment is proper in respect to the theory of errors. with rk , k ¼ 1; . . . ; n, the vector from the origin to atom k and
Nevertheless, it may be desired to incorporate the error estimates with weights, wk , isotropic and without interatomic correlations
(variances) of the atom positions into the results of such calcu- for the n atoms of the plane. We also write S as S  ½wðmT r  dÞ2 ,
lations, whereupon these variances (which may be anisotropic, the subscript for atom number being implicit in the Gaussian
with correlation between atoms) need to be propagated. In this summations ð½. . .Þ over all atoms, as it is also in the angle-bracket
case the distinction between weights (or their inverses) and notation for the weighted average over all atoms, for example in
variances must be kept very clear. (2) The weights are anisotropic hri – the weighted centroid of the groups of atoms – just below.
and are presumably derived from a variance–covariance matrix, First solve for d, the origin-to-plane distance.
which may include correlation terms between different atoms;
the objective is to achieve a truly proper Gaussian least-squares 1 @S
result. 0¼ ¼ ½wðmT r  dÞ ¼ 0;
2 @d
d ¼ ½wmT r=½w  mT hri:

3.2.2. Least-squares plane based on uncorrelated, isotropic


weights Then
This is surely the most common situation; it is not often that one
will wish to take the trouble, or be presumptive enough, to assign S  ½wðmT r  dÞ2  ¼ ½wfmT ðr  hriÞg2 
anisotropic or correlated weights to the various atoms. And one
 ½wðmT sÞ2   mT ½wssT m  mT Am:
will sometimes, perhaps even often, not be genuinely interested
in the hypothesis that the atoms actually are rigorously coplanar;
for instance, one might be interested in examining the best plane
Here sk  rk  hri is the vector from the centroid to atom k. Then
through such a patently nonplanar molecule as cyclohexane.
solve for m. This is the eigenvalue problem – to diagonalize A
Moreover, the calculation is simple enough, given the availability
(bear in mind that Aij is just ½wsi sj ) by rotating the coordinate
of computers and programs, as to be a practical realization of the
axes, i.e., to find the 3  3 arrays M and L, L diagonal, to satisfy
off-the-cuff treatment suggested in our opening paragraph. The
problem of deriving the plane’s coefficients is intrinsically
nonlinear in the way first discussed by Schomaker et al. (1959; MT AM ¼ L; MT M ¼ I:
SWMB). Any formulation other than as an eigenvalue–
eigenvector problem (SWMB), as far as we can tell, will some-
A and M are symmetric; the columns m of M are the direction
† Deceased. cosines of, and the diagonal elements of L are the sums of
410
3.3. Molecular modelling and graphics

By R. Diamond and L. M. D. Cranswick

3.3.1. Graphics A third form, suitable only for rhombohedral cells, is


By R. Diamond 0 1
p þ 2q p  q p  q
aB C
M ¼ @ p  q p þ 2q p  q A
3
3.3.1.1. Coordinate systems, notation and standards p  q p  q p þ 2q
01 2 1 1 1 11
3.3.1.1.1. Cartesian and crystallographic coordinates þ  
It is usual, for purposes of molecular modelling and of Bp q p q p qC
B C
computer graphics, to adopt a Cartesian coordinate system using 1 B1 1 1 2 1 1C
M 1 ¼ B  þ  C
mutually perpendicular axes in a right-handed system using the 3a B
Bp q p q p qC
C
ångström unit or the nanometre as the unit of distance along such @1 1 1 1 1 2A
axes, and largely to ignore the existence of crystallographic   þ
coordinates expressed as fractions of unit-cell edges. Transfor- p q p q p q
mations between the two are thus associated, usually, with the
input and output stages of any software concerned with model- in which
ling and graphics, and it will be assumed after this section that all pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
coordinates are Cartesian using the chosen unit of distance as the p ¼  1 þ 2 cos  q ¼  1  cos ;
unit of coordinates. For a discussion of coordinate transforma-
tions and rotations without making this assumption see Chapter which preserves the equivalence of axes. Here the chiralities of
1.1 in which formulations using co- and contravariant forms are the Cartesian and crystallographic axes are the same if p is chosen
presented. positive, and different otherwise, and the two sets of axes coin-
The relationship between these systems may be written cide in projection along the triad if q is chosen positive and are 
X ¼ Mx x ¼ M 1 X out of phase otherwise.

in which X and x are position vectors in direct space, written as 3.3.1.1.2. Homogeneous coordinates
column vectors, with x expressed in crystallographic fractional Homogeneous coordinates have found wide application in
coordinates (dimensionless) and X in Cartesian coordinates computer graphics. For some equipment their use is essential, and
(dimension of length). they are of value analytically even if the available hardware does
There are two forms of M in common use. The first of these sets not require their use.
the first component of X parallel to a and the third parallel to c Homogeneous coordinates employ four quantities, X, Y, Z and
and is W, to define the position of a point, rather than three. The fourth
0 1 coordinate has a scaling function so that it is the quantity X/W
a’= sin  0 0 (as delivered to the display hardware) which controls the left–
B C
M ¼ @ aðcos   cos  cos Þ= sin  b sin  0 A right positioning of the point within the picture. A point with
a cos  b cos  c jX=Wj < 1 is in the picture, normally, and those with jX=Wj > 1
0 1 are outside it, but see Section 3.3.1.3.5.
sin =a’ 0 0 There are many reasons why homogeneous coordinates may
B C
M 1 ¼ @ ðcos  cos   cos Þ=b’ sin  1=b sin  0 A be adopted, among them the following:
ðcos  cos   cos Þ=c’ sin  1=c tan  1=c (i) X, Y, Z and W may be held as integers, thus enabling fast
arithmetic whilst offering much of the flexibility of floating-point
working. A single W value may be common to a whole array of
in which
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X, Y, Z values.
’¼ 1  cos2   cos2   cos2  þ 2 cos  cos  cos  (ii) Perspective transformations can be implemented without
the need for any division. Only high-speed matrix multiplication
¼ sin  sin  sin   :
using integer arithmetic is necessary, provided only that the
drawing hardware can provide displacements proportional to the
’ is equal to the volume of the unit cell divided by abc, and is ratio of two signals, X and W or Y and W. Rotation, translation,
unchanged by cyclic permutation of ,  and  and of ,  and scaling and the application of perspective are all effected by
  . The Cartesian and crystallographic axes have the same chir- operations of the same form, namely multiplication of a four-
ality if the positive square root is taken. vector by a 4  4 matrix. The hardware may thus be kept rela-
The second form sets the first component of X parallel to a and tively simple since only one type of operation needs to be
the third component of X parallel to c and is provided for.
0 1
a b cos  c cos  (iii) Since kX, kY, kZ, kW represents the same point as X, Y, Z,
B C W, the hardware may be arranged to maximize resolution without
M ¼ @ 0 b sin  cðcos   cos  cos Þ= sin  A
risk of integer overflow.
0 0 c’= sin  For analytical purposes it is convenient to regard homo-
0 1
1=a 1=a tan  ðcos  cos   cos Þ=a’ sin  geneous transformations in terms of partitioned matrices
B C   
M 1 ¼ @ 0 1=b sin  ðcos  cos   cos Þ=b’ sin  A: M V X
;
0 0 sin =c’ U N W

418
3.3. MOLECULAR MODELLING AND GRAPHICS
Table 3.3.4.1. Functionality of software for crystal structure display
Comparison/ Extended
Ball overlay structures/
and Polyhedral Cartesian of multiple topology Magnetic Incommensurate
Program stick ADPs MSDA display coordinates structures analysis structures structures
ATOMS Yes Yes Yes Yes Yes
Balls&Sticks Yes Yes
BALSAC Yes Yes
Cameron Yes Yes
CaRIne Yes Yes Yes
Crystallographica Yes Yes Yes
CrystalMaker Yes Yes Yes Yes
Crystal Studio Yes Yes Yes Yes
CrystMol Yes Yes Yes Yes
Diamond Yes Yes Yes Yes Yes
DrawXTL Yes Yes Yes Yes
FpStudio Yes Yes Yes Yes
GRETEP Yes Yes Yes
Mercury Yes Yes
MolXtl Yes Yes Yes
OLEX Yes Yes Yes
ORTEP-III Yes Yes
ORTEP-3 for Windows Yes Yes Yes Yes
ORTEX/Oscail X Yes Yes
PEANUT Yes Yes Yes
Platon/Pluton Yes Yes Yes
PowderCell Yes
PRJMS Yes Yes
SCHAKAL Yes Yes
STRUPLO Yes Yes
STRUPLO for Windows Yes Yes
STRUVIR Yes
VENUS Yes Yes Yes Yes Yes
XmLmctep Yes Yes
X-Seed Yes Yes Yes
Xtal-3D Yes Yes Yes Yes
XtalDraw Yes Yes Yes

Crystallography Volume G (2005)] is slowly becoming a standard 3.3.4.2.2. Anisotropic displacement parameters
in this regard, displacing the single-crystal SHELX INS format, A subset of the programs that display ball-and-stick structures
which has been a de-facto standard file format for much crystal- can also display surfaces related to anisotropic displacement
lographic data exchange. Entering crystallographic data by hand parameters (ADPs) (also known colloquially as ‘thermals’,
is slow and often introduces errors via typographical mistakes. ‘anisotropic thermal ellipsoids’ or ‘ORTEPS’). By default, most
Such mistakes can be minimized by importing structures using a programs display the ellipsoid surfaces at a probability of 50%
known file type, or reformatting using a text editor or spreadsheet and normally allow this value to be changed to values between 1
program into a known file type. A variety of software programs and 99%. Programs that can draw ADPs include ATOMS,
can be used for translating crystallographic structure files; Cameron, Crystallographica, CrystalMaker, Crystal Studio,
however, the output (especially the handling of the symmetry CrystMol, Diamond, DrawXTL, FpStudio, GRETEP, MolXtl,
operators and the space group) should be carefully checked. The ORTEP-III, ORTEP-3 for Windows, ORTEX, PEANUT,
CCP14 website (Cockroft & Stephenson, 2005) lists a variety of Platon, VENUS, XmLmctep, X-Seed and XtalDraw.
programs that can be used for this, of which a specialist program
is Cryscon (Dowty, 2005).
3.3.4.2.3. Mean-square displacement amplitude
3.3.4.2. Types of crystal structure display and functionality When a more thorough investigation of the ADPs would be
informative (Hummel, Raselli & Bürgi, 1990), PEANUT can be
The following information was current at the time of writing,
used for plotting the mean-square displacement amplitude
but most software is continually changing with the insertion of
(MSDA), root-mean-square displacements (RMSDs) and differ-
new features. Thus occasional checks for updated functionality
ence surfaces. MSDA ‘peanuts’ can be displayed where the ADPs
can be useful. Most software distributions include an ‘updates’
are non-positive-definite and the ellipsoids cannot be drawn.
file containing new features and bug fixes. Detailed information
ORTEP-3 for Windows also has an option for plotting MSDAs.
on the software referred to in this section, including functionality,
Care should be taken to ensure the resulting display is
authorship, source and availability, is given in Tables 3.3.4.1 and
correct.
3.3.4.2.

3.3.4.2.1. Ball and stick 3.3.4.2.4. Polyhedral display


This is one of the most fundamental methods of displaying a A method for understanding inorganic and intermetallic
crystal structure and almost all software supports this. The structures is the use of coordination polyhedra. The faces defined
exceptions are STRUVIR and STRUPLO for Windows [a port of by the outer coordinated atoms generate a polyhedral object that
STRUVIR incorporating a graphical user interface (GUI)], which is displayed instead of the individual atoms. This can aid in
are both optimized for the polyhedral display of crystal struc- understanding the structures of polymeric inorganic materials
tures. involving both simple and complex tilt systems, and distorted
443
3. DUAL BASES IN CRYSTALLOGRAPHIC COMPUTING
Table 3.4.2.2. Untreated lattice-sum results for the dispersion energy (n = 6) of
crystalline benzene (kJ mol1, Å) XðxÞ ¼ x1 a1 þ x2 a2 þ x3 a3 ;
Truncation Number Number Calculated
limit of molecules of terms energy where x1 ; x2 ; x3 are the fractional cell coordinates of X. A lattice
6.0 26 524 69.227 vector in direct space is defined as
8.0 51 1313 76.007
10.0 77 2631 78.179 XðdÞ ¼ d1 a1 þ d2 a2 þ d3 a3 ;
12.0 126 4718 79.241
14.0 177 7531 79.726
16.0 265 11274 80.013
where d1 ; d2 ; d3 are integers (specifying particular values of
x1 ; x2 ; x3 ) designating a lattice point. Vd is the direct-cell volume
18.0 344 15904 80.178
which is equal to a1  a2  a3. A general point in the direct lattice
20.0 439 22049 80.295
is X(x); the contents of the lattice are by definition identical as
Converged value 80.589 the components of x are increased or decreased by integer
amounts.
the original sum, which contains the difference terms, is not The reciprocal-lattice vectors are defined by the relations
increased.
aj  bk ¼ 1 j¼k
Pcell allP
one cells ¼0 j 6¼ k:
Vn ¼ ð1=2Þ 0
Qjk Rn
jk WðRÞ
j k

Pcell allP
one cells
A general vector in reciprocal space H(r) is defined as
þ ð1=2Þ 0
Qjk Rn
jk ½1  WðRÞ:
j k
HðrÞ ¼ r1 b1 þ r2 b2 þ r3 b3 :

In the accelerated-convergence method the difference terms


are expressed as an integral of the product of two functions. A reciprocal-lattice vector H(h) is defined by the integer triplet
According to Parseval’s theorem (described below) this integral h1 ; h2 ; h3 (specifying particular values of r1 ; r2 ; r3 ) so that
is equal to an integral of the product of the two Fourier trans-
forms of the functions. Finally, the integral over the Fourier HðhÞ ¼ h1 b1 þ h2 b2 þ h3 b3 :
transforms of the functions is converted to a sum in reciprocal (or
Fourier-transform) space. The choice of the convergence function
W(R) is not unique; an obvious requirement is that the relevant In other sections of this volume a shortened notation h is used for
Fourier transforms must exist and have correct limiting beha- the reciprocal-lattice vector. In this section the symbol H(h) is
viour. Nijboer and DeWette suggested using the incomplete used to indicate that it is a particular value of H(r).
gamma function for W(R). More recently, Fortuin (1977) showed The three-dimensional Fourier transform gðtÞ of a function f ðxÞ
that this choice of convergence function leads to optimal is defined by
convergence of the sums in both direct and reciprocal space: R
gðtÞ ¼ FT3 ½ f ðxÞ ¼ f ðxÞ expð2ix  tÞ dx:
2 2
WðRÞ ¼ ðn=2; w R Þ=ðn=2Þ;
The Fourier transform of the set of points defining the direct
2 2 lattice is the set of points defining the reciprocal lattice, scaled by
where ðn=2Þ and ðn=2; w R Þ are the gamma function and the
the direct-cell volume. It is useful for our purpose to express the
incomplete gamma function, respectively:
lattice transform in terms of the Dirac delta function ðx  xo Þ
which is defined so that for any function f ðxÞ
R1
ðn=2; w2 R2 Þ ¼ tðn=2Þ1 expðtÞ dt R
w2 R2 f ðxo Þ ¼ ðx  xo Þf ðxÞ dx:

and We then write


P P
ðn=2Þ ¼ ðn=2; 0Þ: FT3 f ½XðxÞ  XðdÞg ¼ Vd1 ½HðrÞ  HðhÞ:
d h

The complement of the incomplete gamma function is


First consider the lattice sum over the direct-lattice points X(d),
relative to a particular point XðxÞ ¼ R, with omission of the
ðn=2; w2 R2 Þ ¼ ðn=2Þ  ðn=2; w2 R2 Þ: origin lattice point.
P
S0 ðn; RÞ ¼ jXðdÞ  Rjn :
d6¼0

3.4.4. Preliminary derivation to obtain a formula which


accelerates the convergence of an Rn sum over lattice points The special case with R ¼ 0 will also be needed:
X(d)
P
The three-dimensional direct-space crystal lattice is specified by S0 ðn; 0Þ ¼ jXðdÞjn :
the origin vectors a1, a2 and a3 . A general vector in direct space is d6¼0

defined as
450
3.5. Extensions of the Ewald method for Coulomb interactions in crystals

By T. A. Darden

3.5.1. Introduction crystallization, the lowest-free-energy polymorph at given


High-precision single-crystal X-ray structural analysis of small temperature and pressure may not be the likeliest to form. The
organic molecules, yielding the space group, the unit-cell para- kinetics of growth of microcrystals may largely determine which
meters and the fractional coordinates of the atoms making up the low-energy polymorph appears (Dunitz, 2003). However, it is
molecule(s) in the asymmetric unit, has become a routine matter generally agreed that accurate calculation of the relative free
as long as crystals of sufficient quality can be obtained. The energy of polymorphs is a prerequisite for predicting crystal
thermodynamic stability of the crystal, as described by the structures.
enthalpy of sublimation Hsub , can also be determined experi- To assess progress towards solving this latter problem, a series
mentally (although not always to high precision). Theoretical of blind tests of crystal-structure prediction has been undertaken
models for calculating intermolecular interaction energies can be (Day, Motherwell, Ammon et al., 2005). The results of these tests
used to connect the crystal structure to the molar enthalpy of have highlighted the need for continued improvements in
sublimation using the relationship sampling methods and intermolecular energy potentials. Since
extensive sampling of the crystal-structure parameters is neces-
sary [between 104 and 105 starting structures, each followed by
Hsub ð0 KÞ ¼ Elattice ; parameter minimizations (Price & Price, 2005)], there is a trade-
off in the computational cost versus accuracy of the inter-
where the lattice or packing energy Elattice is the total (molar) molecular energy functions used. Calculating the work of trans-
intermolecular interaction energy between all the molecules in forming between polymorphs is yet more ambitious in terms of
the crystal, which are treated as rigid entities with zero-point sampling. Consequently empirical force fields are likely to be
energies of intra- and intermolecular vibrations neglected. needed for the near term at least.
Connection to experimentally accessible heats of sublimation at In the remainder of the introduction we outline some of the
higher temperatures involves thermodynamic corrections. approaches to empirical potentials used in the calculation of the
Methods for calculating thermodynamic quantities of solids are lattice energy, and then, motivated by these developments,
discussed in Gavezzotti (2002a) and (in more detail) in Frenkel & discuss techniques for efficient summation of the electrostatic
Smit (2002). and other slow-decaying interaction terms that occur in these
Thus, given a parameterized intermolecular potential-energy potential functions.
function, or if computationally affordable a first-principles Methodological developments in the intermolecular force
approach such as density-functional theory (or preferably, when fields used in crystal-structure prediction from early times to the
it becomes feasible for crystals, a good-quality post-Hartree–Fock present state of the art have been reviewed (Price & Price, 2005).
potential-energy surface that describes dispersion interactions), Until recently, these force fields were made up of atom–atom
one can sum the intermolecular energies to obtain the lattice interactions. The earliest involved only repulsion and dispersion,
energy as a function of the above parameters defining the crystal usually in the ‘exp-6’ form
structure. If such an energy function is used together with a
P
method for systematic search of the crystal-structure parameters, U MN ¼ Urep
MN MN
þ Udisp ¼ Aij expðBij rij Þ  Ci;j r6
ij ;
one could in principle predict the minimum-lattice-energy crystal i2M;j2N
structure for a rigid organic molecule. To extend this approach to
flexible molecules one would need to minimize the sum of the
intramolecular energy plus the lattice energy. If the experimental where UMN denotes the intermolecular potential energy between
crystal structure corresponds to the thermodynamic minimum- molecules M and N and rij is the distance between atoms i 2 M
energy structure (i.e. it is not a metastable state determined by and j 2 N. Sometimes the exponential form in the above equa-
crystal-growth kinetics), one could in principle predict the tion is replaced by a simpler power law, as in the Lennard–Jones
experimental crystal structure of an organic compound through potential. As was pointed out by Dunitz (2003), in comparison
this minimization protocol. Moreover, one could ideally predict with more sophisticated force fields, this repulsion–dispersion
the additional metastable forms of the crystal. form readily allows analysis of the significance of particular
Prediction of the structure of crystals of an organic molecule atom–atom interactions, since the interactions are short-ranged
from its molecular structure is a difficult problem that has been and thus can be localized. That is, the r6 form of the attractive
compared to the protein folding problem (Dunitz, 2003; Dunitz & dispersion energy means that interaction energies are halved for
Scheraga, 2004). Like the protein folding problem, a solution of every 12% increase in distance. In contrast, introduction of long-
the crystal prediction problem has significant practical ramifica- range Coulombic interactions not only entails subtleties in lattice
tions. A compound is often polymorphic, that is it has more than summation (the subject of this contribution), but greatly
one crystal structure, and it may be difficult to characterize the complicates the assignment of ‘key’ atom–atom interactions.
conditions under which a particular crystal structure is formed Gavezzotti and Fillipini systematically explored the use of the
(Dunitz & Bernstein, 1995). Polymorphs may have very different exp-6 potential in fitting organic crystal structures with and
physical properties. An obvious example is diamond versus without hydrogen-bond interactions (Gavezzotti & Fillipini,
graphite, but other commercially important examples include 1994). They were surprisingly successful in accounting for weak
food additives, various solid forms of explosives and the bio- hydrogen bonding in this way, but selective use of point charges
availability of various forms of a drug such as ritonavir (Chem- improved the directionality of the potential. Earlier, Williams
burkar et al., 2000). A method for predicting the possible crystal derived exp-6 parameters for the atoms C, H, N, O, Cl, F and
structures of the compound, and ideally for predicting the polar H for use in organic crystal structures, but found it neces-
dominant crystal structure given the experimental conditions, sary (Williams & Cox, 1984) to supplement these with selected
would thus be very valuable. Note that due to the subtleties of point charges, both atomic and at off-atom sites. Price and
458
4.1. Thermal diffuse scattering of X-rays and neutrons

By B. T. M. Willis

4.1.1. Introduction vibrational energies are many orders of magnitude less than
Thermal motion of the atoms in a crystal gives rise to a reduction X-ray energies. The situation is much more favourable with
in the intensities of the Bragg reflections and to a diffuse distri- thermal neutrons because their wavelength is comparable with
bution of non-Bragg scattering in the rest of reciprocal space. interatomic spacings and their energy is comparable with a
This distribution is known as thermal diffuse scattering (TDS). quantum of vibrational energy (or phonon). The neutron beam is
Measurement and analysis of TDS give information about the scattered inelastically by the lattice vibrations, exchanging energy
lattice dynamics of the crystal, i.e. about the small oscillatory with the phonons. By measuring the energy change for different
displacements of the atoms from their equilibrium positions directions of the scattered beam, the dispersion relations !ðqÞ can
which arise from thermal excitations. Lattice-dynamical models be determined. Brockhouse & Stewart (1958) reported the first
form the basis for interpreting many physical properties – for dispersion curves to be derived in this way; since then the neutron
example, specific heat and thermal conductivity – which cannot technique has become the principal experimental method for
be explained by a static model of the crystal. obtaining detailed information about lattice vibrations.
Reference to a lattice-dynamical model is found in Newton’s In this chapter we shall describe briefly the standard treatment
Principia, which contains a discussion of the vibrations of a linear of the lattice dynamics of crystals. There follows a section on the
chain of equidistant mass points connected by springs. The model theory of the scattering of X-rays by lattice vibrations, and a
was used to estimate the speed of sound in air. The vibrational similar section on the scattering of thermal neutrons. We then
properties of a one-dimensional crystal treated as a linear chain refer briefly to experimental work with X-rays and neutrons. The
of atoms provide the starting point for several modern treatises final section is concerned with the measurement of elastic
on the lattice dynamics of crystals. constants: these constants are required in calculating the TDS
The classical theory of the dynamics of three-dimensional correction to measured Bragg intensities (see Section 7.4.2 of IT
crystals is based on the treatment of Born & von Kármán (1912, C, 2004).
1913). In this theory, the restoring force on an atom is determined
not by the displacement of the atom from its equilibrium posi-
tion, but by its displacement relative to its neighbours. The 4.1.2. Dynamics of three-dimensional crystals
atomic motion is then considered in terms of travelling waves, or For modes of vibration of very long wavelength, the crystal can
‘lattice vibrations’, extending throughout the whole crystal. These be treated as a homogeneous elastic continuum without referring
waves are the normal modes of vibration, in which each mode is to its crystal or molecular structure. The theory of the propaga-
characterized by a wavevector q, an angular frequency !(q) and tion of these elastic waves is based on Hooke’s law of force and
certain polarization properties. on Newton’s equations of motion. As the wavelength of the
For twenty years after its publication the Born–von Kármán vibrations becomes shorter and shorter and approaches the
treatment was eclipsed by the theory of Debye (1912). In the separation of adjacent atoms, the calculation of the vibrational
Debye theory the crystal is treated as a continuous medium properties requires a knowledge of the crystal structure and of
instead of a discrete array of atoms. The theory gives a reasonable the nature of the forces between adjacent atoms. The three-
fit to the integral vibrational properties (for example, the specific dimensional treatment is based on the formulation of Born and
heat or the atomic temperature factor) of simple monatomic von Kármán, which is discussed in detail in the book by Born &
crystals. It fails to account for the form of the frequency distri- Huang (1954) and in more elementary terms in the books by
bution function which relates the number of modes and their Cochran (1973) and by Willis & Pryor (1975).
frequency. Before setting up the equations of motion, it is necessary to
An even simpler model than Debye’s is due to Einstein (1907), introduce three approximations:
who considered the atoms in the crystal to be vibrating inde- (i) The harmonic approximation. When an atom is displaced
pendently of each other and with the same frequency !E. By from its equilibrium position, the restoring force is assumed to be
quantizing the energy of each atom in units of h- !E, Einstein proportional to the displacement, measured relative to the
showed that the specific heat falls to zero at T = 0 K and rises neighbouring atoms. The approximation implies no thermal
asymptotically to the Dulong and Petit value for T much larger expansion and other properties not possessed by real crystals; it is
than h- !E =kB . (h- is Planck’s constant divided by 2 and k is
B a reasonable assumption in the lattice-dynamical theory provided
Boltzmann’s constant.) His theory accounts satisfactorily for the the displacements are not too large.
breakdown of equipartition of energy at low temperatures, but it (ii) The adiabatic approximation. We wish to set up a potential
predicts a more rapid fall-off of specific heat with decreasing function for the crystal describing the binding between the atoms.
temperature than is observed. However, the binding involves electronic motions whereas the
Deficiencies in the Debye theory were noted by Blackman dynamics involve nuclear motions. The adiabatic approximation,
(1937), who showed that they are overcome satisfactorily using known as the Born–Oppenheimer approximation in the context
the more rigorous Born–von Kármán theory. Extensive X-ray of molecular vibrations, provides the justification for adopting the
studies of Laval (1939) on simple structures such as sylvine, same potential function to describe both the binding and the
aluminium and diamond showed that the detailed features of the dynamics. Its essence is that the electronic and nuclear motions
TDS could only be explained in terms of the Born–von Kármán may be considered separately. This is possible if the nuclei move
theory. The X-ray work on aluminium was developed further by very slowly compared with the electrons: the electrons can then
Olmer (1948) and by Walker (1956) to derive the phonon instantaneously take up a configuration appropriate to that of the
dispersion relations (see Section 4.1.5) along various symmetry displaced nuclei without changing their quantum state. The
directions in the crystal. approximation holds well for insulators, where electronic tran-
It is possible to measure the vibrational frequencies directly sition energies are high owing to the large energy gap between
with X-rays, but such measurements are very difficult as lattice filled and unfilled electron states. Surprisingly, it even works for
484
4.2. Disorder diffuse scattering of X-rays and neutrons

By F. Frey, H. Boysen and H. Jagodzinski

4.2.1. Introduction higher temperatures. Hence a thermodynamical interpretation of


diffraction data needs a careful study of diffuse scattering as a
Diffuse scattering of X-rays, neutrons and other particles is an function of temperature or some other thermodynamical para-
accompanying effect in all diffraction experiments aimed at meters. This can be done in very rare cases only, so the omission
structure analysis with the aid of elastic scattering. In this case, of this subject seems justified.
the momentum exchange of the scattered photon or particle As shown in this chapter, electron-density fluctuations and
includes the crystal as a whole; the energy transfer involved distribution functions of defects play an important role in the
becomes negligibly small and need not be considered in diffrac- complete interpretation of diffraction patterns. Both quantities
tion theory. Static distortions as a consequence of structural may best be studied in the low-angle scattering range. Hence
changes cause typical elastic diffuse scattering. Many structural many problems cannot be solved without a detailed interpreta-
phenomena and processes contribute to diffuse scattering, and a tion of low-angle diffraction (also called small-angle scattering).
general theory has to include all of them. Hence the exact Disorder phenomena in magnetic structures are also not
treatment of diffuse scattering becomes very complex. specifically discussed here. Magnetic diffuse neutron scattering
Inelastic scattering is due to dynamical fluctuations or ioniza- and special experimental techniques constitute a large subject by
tion processes and may become observable as a ‘diffuse’ contri- themselves. Many aspects, however, may be analysed along
bution in a diffraction pattern. A separation of elastic from similar lines to those given here.
inelastic diffuse scattering is generally possible, but difficulties Glasses, liquids or liquid crystals show typical diffuse diffrac-
may result from small energy exchanges that cannot be resolved tion phenomena. Particle-size effects and strains have an
for experimental reasons. The latter is true for scattering of important influence on the diffuse scattering. The same is true for
X-rays by phonons, which have energies of the order of dislocations and point defects such as interstitials or vacancies.
102–103 eV, values which are considerably smaller than 10 keV, These defects are mainly described by their strain field, which
a typical value for X-ray quanta. Another equivalent explanation, influences the intensities of sharp reflections like an artificial
frequently forwarded in the literature, is the high speed of X-ray temperature factor: the Bragg peaks diminish in intensity while
photons, such that the rather slow motion of atoms cannot be the diffuse scattering increases predominantly close to them.
‘observed’ by them during diffraction. Hence, all movements These phenomena are less important from a structural point of
appear as static displacement waves of atoms, and temperature view, at least in the case of metals or other simple structures. This
diffuse scattering is pseudo-elastic for X-rays. This is not true in statement is true as long as the structure of the ‘kernel’ of defects
the case of thermal neutrons, which have energies comparable to may be neglected when compared with the influence of the strain
those of phonons. Phonon-related or thermal diffuse scattering is field. Whether dislocations in more complicated structures meet
discussed separately in Chapter 4.1, i.e. the present chapter is this condition is not yet known.
mainly concerned with the elastic (or pseudo-elastic other than Commensurate and incommensurate modulated structures
thermal) part of diffuse scattering. A particularly important and quasicrystals frequently show a typical diffuse scattering, a
aspect concerns diffuse scattering related to phase transitions, in satisfactory explanation of which demands extensive experi-
particular the critical diffuse scattering observed at or close to the mental and theoretical study. A reliable structure determination
transition temperature. In simple cases, a satisfactory description becomes very difficult in cases where the interpretation of diffuse
may be given with the aid of a ‘soft phonon’, which freezes at scattering has not been incorporated. Many erroneous structural
the critical temperature, thus generating typical temperature- conclusions have been published in the past. The solution of
dependent diffuse scattering. If the geometry of the lattice is problems of this kind needs careful thermodynamical consid-
maintained during the transformation (i.e. there is no breakdown eration as to whether a plausible explanation of the structural
into crystallites of different cell geometry), the diffuse scattering data can be given.
is very similar to diffraction phenomena described in this chapter. For all of the reasons mentioned above, this article cannot be
Sometimes, however, very complicated interim stages (ordered or complete. It is hoped, however, that it will provide a useful guide
disordered) are observed, demanding a complicated theory for for those who need a full understanding of the crystal chemistry
their full explanation (see, e.g., Dorner & Comes, 1977). of a given structure.
Obviously, there is a close relationship between thermo- The study of disorder in crystals by diffuse-scattering techni-
dynamics and diffuse scattering in disordered systems repre- ques can be performed with X-rays, neutrons or electrons. Each
senting a stable or metastable thermal equilibrium. From the of these methods has its own advantages (and disadvantages) and
thermodynamical point of view, the system is then characterized they often can (or have to) be used in a complementary way (cf.
by its grand partition function, which is intimately related to the Chapter 4.3 of this volume). Electron diffraction and microscopy
correlation functions used in the interpretation of diffuse scat- are usually restricted to relatively small regions in space and thus
tering. The latter is nothing other than a kind of ‘partial partition supply information on a local scale, i.e. local defect structures.
function’ where two atoms, or two cell occupancies, are fixed such Moreover, electron-microscopy investigations are carried out on
that the sum of all partial partition functions represents the grand thin samples (films), where the disorder could be different from
partition function. This fact yields the useful correlation between the bulk, and, in addition, could be affected by the high heat load
thermodynamics and diffuse scattering mentioned above, which deposited by the impinging electron beam. X-rays and neutrons
may well be used for a determination of thermodynamical sample larger crystal volumes and thus provide thermo-
properties of the crystal. This important subject shall not be dynamically more important information on averages of the
included here for the following reason: real three-dimensional disorder. These methods are also better suited to the analysis of
crystals generally exhibit diffuse scattering by defects and/or long-range correlated cooperative disorder phenomena. On the
disordering effects that are not in thermal equilibrium. They are other hand, electron microscopy and diffraction often allow more
created during crystal growth, or are frozen-in defects formed at direct access to disorder and can therefore provide valuable

492
4.3. Diffuse scattering in electron diffraction

By J. M. Cowley† and J. K. Gjønnes

4.3.1. Introduction The kinematical description can be used for electron scattering
The origins of diffuse scattering in electron-diffraction patterns only when the crystal is very thin (10 nm or less) and composed of
are the same as in the X-ray case: inelastic scattering due to light atoms. For heavy atoms such as Au or Pb, crystals of
electronic excitations, thermal diffuse scattering (TDS) from thickness 1 nm or more in principal orientations show strong
atomic motions, scattering from crystal defects or disorder. For deviations from kinematical behaviour. With increasing thick-
diffraction by crystals, the diffuse scattering can formally be ness, dynamical scattering effects first modify the sharp Bragg
described in terms of a nonperiodic deviation ’ from the reflections and then have increasingly significant effects on the
periodic, average crystal potential, ’ : diffuse scattering. Bragg scattering of the diffuse scattering
produces Kikuchi lines and other effects. Multiple diffuse scat-
tering broadens the distribution and smears out detail. As the
’ðr; tÞ ¼ ’ ðrÞ þ ’ðr; tÞ; ð4:3:1:1Þ
thickness increases further, the diffuse scattering increases and
the Bragg beams are reduced in intensity until there is only a
where ’ may have a static component from disorder in addition diffuse ‘channelling pattern’ where the features depend in only a
to time-dependent fluctuations of the electron distribution or very indirect way on the incident-beam direction or on the
atomic positions. sources of the diffuse scattering (Uyeda & Nonoyama, 1968).
In the kinematical case, the diffuse scattering can be treated The multiple-scattering effects make the quantitative inter-
separately. The intensity Id as a function of the scattering variable pretation of diffuse scattering more difficult and complicate the
u ðjuj ¼ 2 sin =Þ and energy transfer h is then given by the extraction of particular components, e.g. disorder scattering.
Fourier transform F of ’ Much of the multiple scattering involves inelastic scattering
processes. However, electrons that have lost energy of the order
of 1 eV or more can be subtracted experimentally by use of
Iðu; Þ ¼ jðu Þj2 ¼ jF f’ðr; tÞgj2 ¼ F fPd ðr; Þg ð4:3:1:2Þ
electron energy filters (Krahl et al., 1990; Krivanek et al., 1992)
which are commercially available. Measurement can be made
and may also be written as the Fourier transform of a correlation also of the complete scattering function Iðu; Þ, but such studies
function Pd representing fluctuations in space and time (see have been rare. Another significant improvement to quantitative
Cowley, 1981). When the energy transfers are small – as with TDS measurement of diffuse electron scattering is offered by new
– and hence not measured, the observed intensity corresponds to recording devices: slow-scan charge-couple-device cameras
an integral over : (Krivanek & Mooney, 1993) and imaging plates (Mori et al.,
1990).
IðuÞ ¼ Id ðuÞ þ Iav ðuÞ There are some advantages in the use of electrons which make
R it uniquely valuable for particular applications.
Id ðuÞ ¼ Id ðu; Þ d ¼ F fPd ðr; 0Þg

and also

Id ðuÞ ¼ hjðuÞj2 i  jhðuÞij2 ; ð4:3:1:3Þ

where the brackets may indicate a time average, an expectation


value, or a spatial average over the periodicity of the lattice in the
case of static deviations from a periodic structure.
The considerations of TDS and static defects and disorder of
Chapters 4.1 and 4.2 thus may be applied directly to electron
diffraction in the kinematical approximation when the differ-
ences in experimental conditions and diffraction geometry are
taken into account.
The most prominent contribution to the diffuse background in
electron diffraction, however, is the inelastic scattering at low
angles arising mainly from the excitation of outer electrons. This
is quite different from the X-ray case where the inelastic
(‘incoherent’) scattering, SðuÞ, goes to zero at small angles and
increases to a value proportional to Z for high values of juj. The
difference is due to the Coulomb nature of electron scattering,
which leads to the kinematical intensity expression S=u4 ,
emphasizing the small-angle region. At high angles, the inelastic
scattering from an atom is then proportional to Z=u4, which is
considerably less than the corresponding elastic scattering
ðZ  f Þ2 =u4 which approaches Z2 =u4 (Section 2.5.2) (see Fig.
4.3.1.1). Fig. 4.3.1.1. Comparison between the kinematical inelastic scattering (full
line) and elastic scattering (broken) for electrons and X-rays. Values for
† Deceased. silicon [Freeman (1960) and IT C (2004)].

540
4.4. Scattering from mesomorphic structures

By P. S. Pershan

4.4.1. Introduction most thoroughly documented mesomorphic phases, there are


The term mesomorphic is derived from the prefix ‘meso-’, which others not included in the table which we will discuss below.
is defined in the dictionary as ‘a word element meaning middle’, The progression from the completely symmetric isotropic
and the term ‘-morphic’, which is defined as ‘an adjective liquid through the mesomorphic phases into the crystalline
termination corresponding to morph or form’. Thus, meso- phases can be described in terms of three separate types of order.
morphic order implies some ‘form’, or order, that is ‘in the The first, or the molecular orientational order, describes the fact
middle’, or intermediate between that of liquids and crystals. The that the molecules have some preferential orientation analogous
name liquid crystalline was coined by researchers who found it to to the spin orientational order of ferromagnetic materials. In
be more descriptive, and the two are used synonymously. It the present case, the molecular quantity that is oriented is a
follows that a mesomorphic, or liquid-crystalline, phase must symmetric second-rank tensor, like the moment of inertia or the
have more symmetry than any one of the 230 space groups that electric polarizability, rather than a magnetic moment. This is the
characterize crystals. only type of long-range order in the nematic phase and as a
A major source of confusion in the early liquid-crystal litera- consequence its physical properties are those of an anisotropic
ture was concerned with the fact that many of the molecules that fluid; this is the origin of the name liquid crystal. Fig. 4.4.1.2(a) is
form liquid crystals also form true three-dimensional crystals with a schematic illustration of the nematic order if it is assumed that
diffraction patterns that are only subtly different from those of the molecules can be represented by oblong ellipses. The average
other liquid-crystalline phases. Since most of the original meso- orientation of the ellipses is aligned; however, there is no long-
morphic phase identifications were performed using a ‘misci- range order in the relative positions of the ellipses. Nematic
bility’ procedure, which depends on optically observed changes in phases are also observed for disc-shaped molecules and for
textures accompanying variation in the sample’s chemical clusters of molecules that form micelles. These all share the
composition, it is not surprising that some three-dimensional common properties of being optically anisotropic and fluid-like,
crystalline phases were mistakenly identified as mesomorphic. without any long-range positional order.
Phases were identified as being either the same as, or different The second type of order is referred to as bond orientational
from, phases that were previously observed (Liebert, 1978; Gray order. Consider, for example, the fact that for dense packing of
& Goodby, 1984), and although many of the workers were very spheres on a flat surface most of the spheres will have six
clever in deducing the microscopic structure responsible for the neighbouring spheres distributed approximately hexagonally
microscopic textures, the phases were labelled in the order of around it. If a perfect two-dimensional triangular lattice of
discovery as smectic-A, smectic-B etc. without any attempt to indefinite size were constructed of these spheres, each hexagon
develop a systematic nomenclature that would reflect the on the lattice would be oriented in the same way. Within the last
underlying order. Although different groups did not always few years, we have come to recognize that this type of order, in
assign the same letters to the same phases, the problem is now which the hexagons are everywhere parallel to one another, is
resolved and the assignments used in this article are commonly possible even when there is no lattice. This type of order is
accepted (Gray & Goodby, 1984). referred to as bond orientational order, and bond orientational
Fig. 4.4.1.1 illustrates the way in which increasing order can be order in the absence of a lattice is the essential property defining
assigned to the series of mesomorphic phases in three dimensions the hexatic phases (Halperin & Nelson, 1978; Nelson & Halperin,
listed in Table 4.4.1.1. Although the phases in this series are the 1979; Young, 1979; Birgeneau & Litster, 1978).
The third type of order is the positional order of an indefinite
lattice of the type that defines the 230 space groups of conven-

Table 4.4.1.1. Some of the symmetry properties of the series of three-


dimensional phases described in Fig. 4.4.1.1
The terms LRO and SRO imply long-range or short-range order, respectively, and
QLRO refers to ‘quasi-long-range order’ as explained in the text.
Molecular Positional order
orientation Bond
order within orientation Normal Within
Phase layer order to layer layer
Smectic-A (SmA) SRO SRO SRO SRO
Smectic-C (SmC) LRO LRO† SRO SRO
Hexatic-B LRO† LRO QLRO SRO
Smectic-F (SmF) LRO LRO QLRO SRO
Smectic-I (SmI) LRO LRO QLRO SRO
Crystalline-B (CrB) LRO LRO LRO LRO
Crystalline-G (CrG) LRO LRO LRO LRO
Crystalline-J (CrJ) LRO LRO LRO LRO
Crystalline-E (CrE) LRO LRO LRO LRO
Crystalline-H (CrH) LRO LRO LRO LRO
Fig. 4.4.1.1. Illustration of the progression of order throughout the sequence Crystalline-K (CrK) LRO LRO LRO LRO
of mesomorphic phases that are based on ‘rod-like’ molecules. The shaded
section indicates phases in which the molecules are tilted with respect to the † Theoretically, the existence of LRO in the molecular orientation, or tilt, implies that there
smectic layers. must be some LRO in the bond orientation and vice versa.

547
4. DIFFUSE SCATTERING AND RELATED TOPICS
for the intensities Il ðRhk Þ from the data IðR; ZÞ on the diffraction structure can then be determined by an exhaustive exploration of
pattern. The parameters llat, laxial and 0 , as well as the cell the parameter space using molecular model building. The first
constants and possibly other parameters, can also be refined example above would belong to this class. The second class of
as part of the profile-fitting procedure using nonlinear methods is appropriate when the size of the helical repeating unit
optimization. is such that its structure is described by too many variable
A suite of programs for processing fibre diffraction data parameters for the parameter space to be explored a priori. It is
is distributed (and often developed) by the Collaborative then necessary to phase the fibre diffraction data and construct
Computational Project for Fibre and Polymer Diffraction an electron-density map into which the molecular structure can
(CCP13) in the UK (http://www.ccp13.ac.uk/) (Shotton et al., be fitted and then refined. The second example above would
1998). belong to this class. The second class of methods therefore mimics
conventional protein crystallography quite closely. The third class
of problems applies when the structure is large, but there are too
4.5.2.6. Structure determination few diffraction data to attempt phasing and the usual determi-
nation of atomic coordinates. The solution to such problems
4.5.2.6.1. Overview varies from case to case and usually involves modelling and
Structure determination in fibre diffraction is concerned with optimization of some kind.
determining atomic coordinates or some other structural para- An important parameter in structure determination by fibre
meters, from the measured cylindrically averaged diffraction diffraction is the degree of overlap (that results from the
data. Fibre diffraction analysis suffers from the phase problem cylindrical averaging) in the data. This parameter is equal to the
and low resolution (diffraction data rarely extend beyond 3 Å number of significant terms in equation (4.5.2.17) or the number
resolution), but this is no worse than in protein crystallography of independent terms in equation (4.5.2.24), and depends on the
where phases derived from, say, isomorphous replacement or position in reciprocal space and, for a polycrystalline fibre, the
molecular replacement, coupled with the considerable stereo- space-group symmetry. The number of degrees of freedom in a
chemical information usually available on the molecule under particular datum is equal to twice this number (since each
study, together contribute enough information to lead to precise structure factor generally has real and imaginary parts), and is
structures. What makes structure determination by fibre diffrac- denoted in this section by m. Determination of the Gnl ðRÞ from
tion more difficult is the loss of information owing to the the cylindrically averaged data Il ðRÞ therefore involves separ-
cylindrical averaging of the diffraction data. However, in spite of ating the m=2 amplitudes jGnl ðRÞj and assigning phases to each.
these difficulties, fibre diffraction has been used to determine, The electron density can be calculated from the Gnl ðRÞ using
with high precision, the structures of a wide variety of biological equations (4.5.2.7) and (4.5.2.11).
and synthetic polymers, and other macromolecular assemblies.
Because of the size of the repeating unit and the resolution of the
diffraction data, methods for structure determination in fibre 4.5.2.6.2. Helix symmetry, cell constants and space-group
diffraction tend to mimic those of macromolecular (protein) symmetry
crystallography, rather than small-molecule crystallography The first step in analysis of any fibre diffraction pattern is
(direct methods). determination of the molecular helix symmetry uv . Only the zero-
For a noncrystalline fibre one can determine only the mole- order Bessel term contributes diffracted intensity on the meri-
cular structure from the continuous diffraction data, whereas for dian, and referring to equation (4.5.2.6) shows that the zero-order
a polycrystalline fibre one can determine crystal structures from term occurs only on layer lines for which l is a multiple of u.
the Bragg diffraction data. However, there is little fundamental Therefore, inspection of the distribution of diffraction along the
difference between methods used for structure determination meridian allows the value of u to be inferred. This procedure is
with noncrystalline and polycrystalline fibres. For partially crys- usually effective, but can be difficult if u is large, because the first
talline fibres, little has so far been attempted with regard to meridional maximum may be on a layer line that is difficult to
rigorous structure determination. measure. This difficulty was overcome in one case by Franklin &
As is the case with protein crystallography, the precise Holmes (1958) by noting that the second Bessel term on the
methods used for structure determination by fibre diffraction equator is n ¼ u, estimating G00 ðRÞ using data from a heavy-atom
depend on the particular problem at hand. A variety of tools are derivative (see Section 4.5.2.6.6), subtracting this from I0 ðRÞ, and
available and one selects from these those that are appropriate using the behaviour of the remaining intensity for small R to infer
given the data available in a particular case. For example, the the order of the next Bessel term [using equation (4.5.2.14)] and
structure of a polycrystalline polynucleotide might be determined thence u.
by using Patterson functions to determine possible packing Referring to equations (4.5.2.6) and (4.5.2.14) shows that the
arrangements, molecular model building to define, refine and distribution of Rmin for 0 < l < u depends on the value of v.
arbitrate between structures, difference Fourier synthesis to Therefore, inspection of the intensity distribution close to the
locate ions or solvent molecules, and finally assessment of the meridian often allows v to be inferred. Note, however, that the
reliability of the structure. As a second example, to determine the distribution of Rmin does not distinguish between the helix
structure of a helical virus, one might use isomorphous replace- symmetries uv and uuv . Any remaining ambiguities in the helix
ment to obtain phase estimates, calculate an electron-density symmetry need to be resolved by steric considerations, or by
map, fit a preliminary model and refine it using simulated detailed testing of models with the different symmetries against
annealing alternating with difference Fourier analysis, and assess the available data.
the results. The various tools available, together with indications For a polycrystalline system, the cell constants are determined
of where and how they are used, are described in the following from the ðR; ZÞ coordinates of the spots on the diffraction pattern
sections. as described in Section 4.5.2.6.4. Space-group assignment is based
Although a variety of techniques are used to solve structures on analysis of systematic absences, as in conventional crystal-
using fibre diffraction, most of the methods do fall broadly into lography. However, in some cases, because of possible overlap of
one of three classes that depend primarily on the size of the systematic absences with other reflections, there may be some
helical repeat unit. The first class applies to molecules whose ambiguity in space-group assignment. However, the space group
repeating units are small, i.e. are represented by a relatively small can always be limited to one of a few possibilities, and ambiguities
number of independent parameters or degrees of freedom (after can usually be resolved during structure determination (Section
all stereochemical constraints have been incorporated). The 4.5.2.6.4).
576
4.6. Reciprocal-space images of aperiodic crystals

By W. Steurer‡ and T. Haibach

4.6.1. Introduction nD lattice symmetry. The intersection of the nD hypercrystal with


The discovery of materials with icosahedral diffraction symmetry the dD physical space is equivalent to  a projection
Pn of the
(Shechtman et al., 1984) was the main reason for the reassessment weighted nD reciprocal lattice  ¼ H ¼ i¼1 hi di jhi 2 Z
of the definition of crystallinity and for the introduction of the onto the dD physical
Pn space. The
 resulting set (Fourier module)
concept of aperiodic crystals. The first aperiodic crystal, i.e. a M ¼ Hk ¼ i¼1 hi ai jhi 2 Z is countably dense. Countably
material with Bragg reflections not located only at reciprocal- dense means that the dense set of Bragg peaks can be mapped
lattice nodes, was identified long before (Dehlinger, 1927). In the one-to-one onto the set of natural numbers. Hence, the Bragg
following decades a wealth of incommensurately modulated reflections can be indexed with integer indices on an appropriate
phases and composite crystals were discovered. Nevertheless, basis. The Fourier module of the projected reciprocal-lattice
only a few attempts have been made to develop a crystallography vectors Hk has the structure of a Z module of rank n. A Z module
of aperiodic crystals; the most powerful of these was the higher- is a free Abelian group, its rank n is given by the number of free
dimensional approach (see de Wolff, 1974, 1977; Janner & generators (rationally independent vectors). The dimension of a
Janssen, 1979, 1980a,b; de Wolff et al., 1981). In fact, incom- Z module is that of the vector space spanned by it. The vectors ai
mensurate structures can be easily described using the higher- are the images of the vectors di projected onto the physical space
dimensional approach and also, fully equivalently, in a dual way: Vk . Thus, by definition, the 3D reciprocal space of an ideal
as a three-dimensional (3D) combination of one or more periodic aperiodic crystal consists of a countably dense set of Bragg
basic structures and one or several modulation waves (de Wolff, reflections only. Contrary to an ideal crystal, a minimum distance
1984). However, with the discovery of quasicrystals and their between Bragg reflections does not exist in an aperiodic one. In
noncrystalline symmetries, the latter approach failed and summary, it may be stressed that the terms aperiodic and periodic
geometrical crystallography including the higher-dimensional refer to properties of crystal structures in dD space. In nD space,
approach received new attention. For more recent reviews of the as considered here, lattice symmetry is always present and,
crystallography of all three types of aperiodic crystals see van therefore, the term crystal is used.
Smaalen (1995), of incommensurately modulated structures see Besides the aperiodic crystals mentioned above, other classes
Cummins (1990), of quasicrystals see Steurer (1990, 1996), of of aperiodic structures with strictly defined construction rules
quasicrystals and their crystalline approximants see Goldman & exist (see Axel & Gratias, 1995). Contrary to the kind of
Kelton (1993) and Kelton (1995). Textbooks on quasicrystals aperiodic crystals dealt with in this chapter, the Fourier spectra of
have been written by Janot (1994) and Senechal (1995). aperiodic structures considered in the latter reference are
According to the traditional crystallographic definition, an continuous and contain only in a few cases additional sharp
ideal crystal corresponds to an infinite 3D periodic arrangement Bragg reflections ( peaks).
of identical structure motifs. Its symmetry can be described by Experimentally, the borderline between aperiodic crystals and
one of the 230 3D space groups. Mathematically, a periodic their periodic approximations (crystalline approximants) is not
structure can be generated by the convolution of a function sharply defined. Finite crystal size, static and dynamic disorder,
representing the structure motif with a lattice function. The chemical impurities and defects broaden Bragg peaks and cause
structure motif can be given, for instance, by the electron-density diffuse diffraction phenomena. Furthermore, the resolution
distribution ðrÞ of one primitive unit cell of the structure. The function of the diffraction equipment is limited.
However, the concept of describing an aperiodic structure as a
P3 gðrÞ is represented by a set of  functions at the
lattice function
dD physical-space section of an nD crystal (see Section 4.6.2) is
nodes r ¼ i¼1 ki ai of a 3D lattice  with basis ai, i ¼ 1; . . . ; 3,
and ki 2 Z (Z is the set of integer numbers). In reciprocal space, only useful if it significantly simplifies the description of its
this convolution corresponds to the product of the Fourier structural order. Thus, depending on the shape of the atomic
transform GðHÞ ofR the lattice function gðrÞ and the Fourier surfaces, which gives information on the atomic ordering,
transform FðHÞ ¼ v ðrÞ expð2iH  rÞ dr of the structure motif incommensurately modulated structures (IMSs, Sections 4.6.2.2
ðrÞ. GðHÞ is represented by the reciprocal lattice  decorated and 4.6.3.1), composite structures (CSs, Sections 4.6.2.3 and
P3 4.6.3.2), or quasiperiodic structures (QSs, Sections 4.6.2.4 and
with  functions on the reciprocal-lattice nodes H ¼ i¼1 hi ai ,
with the reciprocal-basis vectors ai , i ¼ 1; . . . ; 3, defined by 4.6.3.3) can be obtained from irrational cuts. The atomic surfaces
ai  aj ¼ ij and hi 2 Z. The product GðHÞ  FðHÞ is called the are continuous ðn  dÞ-dimensional objects for IMSs and CSs,
weighted reciprocal lattice; the weights are given by the structure and discrete ðn  dÞ-dimensional objects for QSs. A class of
factors FðHÞ. Thus, the characteristic feature of an ideal crystal in aperiodic crystals with discrete fractal atomic surfaces also exists
direct and reciprocal space is the existence of a lattice. In direct (Section 4.6.2.5). In this case the Hausdorff dimension (Haus-
space, this lattice is decorated with identical structure motifs dorff, 1919) of the atomic surface is not an integer number and
preserving translational and point symmetry in the framework of smaller than n  d. The most outstanding characteristic feature
space-group symmetry. In reciprocal space, only the point of a fractal is its scale invariance: the object appears similar to
symmetry between structure factors is maintained. The Fourier itself ‘from near as from far, that is, whatever the scale’ (Gouyet,
spectrum (or Fourier image, i.e. the Fourier transform) of the 1996).
electron-density distribution of an ideal crystal consists of a To overcome the problems connected with experimental
countably infinite set of discrete Bragg peaks with a strictly resolution, the translational symmetry of periodic crystals is used
defined minimum distance. as a hard constraint in the course of the determination of their
This crystal definition can be generalized to n > 3 dimensions. structures. Hence, space-group symmetry is taken for granted and
A d-dimensional (dD) ideal aperiodic crystal can be defined as a only the local atomic configuration in a unit cell (actually,
dD irrational section of an n-dimensional (nD, n > d) crystal with asymmetric unit) remains to be determined. In reciprocal space,
this assumption corresponds to a condensation of Bragg reflec-
‡ To whom correspondence should be addressed. tions with finite full width at half maximum (FWHM) to  peaks
590
5.1. Dynamical theory of X-ray diffraction

By A. Authier

5.1.1. Introduction (Bragg et al., 1926). For perfect or nearly perfect crystals, dyna-
The first experiment on X-ray diffraction by a crystal was mical theory should be used. Geometrical theory presents
performed by W. Friedrich, P. Knipping and M. von Laue in 1912 another drawback: it gives no indication as to the phase of the
and Bragg’s law was derived in 1913 (Bragg, 1913). Geometrical reflected wave. This is due to the fact that it is based on the
and dynamical theories for the intensities of the diffracted X-rays Fourier transform of the electron density limited by the external
were developed by Darwin (1914a,b). His dynamical theory took shape of the crystal. This is not important when one is only
into account the interaction of X-rays with matter by solving interested in measuring the reflected intensities. For any problem
recurrence equations that describe the balance of partially where the phase is important, as is the case for multiple reflec-
transmitted and partially reflected amplitudes at each lattice tions, interference between coherent blocks, standing waves etc.,
plane. This is the first form of the dynamical theory of X-ray dynamical theory should be used, even for thin or imperfect
diffraction. It gives correct expressions for the reflected inten- crystals.
sities and was extended to the absorbing-crystal case by Prins Until the 1940s, the applications of dynamical theory were
(1930). A second form of dynamical theory was introduced by essentially intensity measurements. From the 1950s to the 1970s,
Ewald (1917) as a continuation of his previous work on the applications were related to the properties (absorption, inter-
diffraction of optical waves by crystals. He took into account the ference, propagation) of wavefields in perfect or nearly perfect
interaction of X-rays with matter by considering the crystal to be crystals: anomalous transmission, diffraction of spherical waves,
a periodic distribution of dipoles which were excited by the interpretation of images on X-ray topographs, accurate
incident wave. This theory also gives the correct expressions for measurement of form factors, lattice-parameter mapping. In
the reflected and transmitted intensities, and it introduces the recent years, they have been concerned mainly with crystal optics,
fundamental notion of a wavefield, which is necessary to under- focusing and the design of monochromators for synchrotron
stand the propagation of X-rays in perfect or deformed crystals. radiation [see, for instance, Batterman & Bilderback (1991)], the
Ewald’s theory was later modified by von Laue (1931), who location of atoms at crystal surfaces and interfaces using the
showed that the interaction could be described by solving standing-waves method, determination of phases using multiple
Maxwell’s equations in a medium with a continuous, triply reflections [for reviews of n-beam diffraction, see Weckert &
periodic distribution of dielectric susceptibility. It is this form Hümmer (1997) and Chang (2004); for recent determinations of
which is most widely used today and which will be presented in phases, see Chang et al. (2002), Mo et al. (2002), Weckert et al.
this chapter. (2002), Shen & Wang (2003)], characterization of the crystal
The geometrical (or kinematical) theory, on the other hand, perfection of epilayers and superlattices by high-resolution
considers that each photon is scattered only once and that the diffractometry [see, for instance, Tanner (1990) and Fewster
interaction of X-rays with matter is so small it can be neglected. It (1993)], etc.
can therefore be assumed that the amplitude incident on every Modern developments include the extension of dynamical
diffraction centre inside the crystal is the same. The total theory to time-dependent phenomena (Chukhovskii & Förster,
diffracted amplitude is then simply obtained by adding the 1995; Shastri et al., 2001; Graeff, 2002a,b, 2004; Malgrange &
individual amplitudes diffracted by each diffracting centre, taking Graeff, 2003; Sondhauss & Wark, 2003; Adams, 2004) and the
into account only the geometrical phase differences between study of the influence of the coherence of the source (Yamazaki
them and neglecting the interaction of the radiation with matter. & Ishikawa, 2002, 2004).
The result is that the distribution of diffracted amplitudes in For reviews of dynamical theory, see Zachariasen (1945), von
reciprocal space is the Fourier transform of the distribution of Laue (1960), James (1963), Batterman & Cole (1964), Authier
diffracting centres in physical space. Following von Laue (1960), (1970), Kato (1974), Brümmer & Stephanik (1976), Pinsker
the expression geometrical theory will be used throughout this (1978), Authier et al. (1996), Authier & Malgrange (1998), and
chapter when referring to these geometrical phase differences. Authier (2005). Topography is described in Chapter 2.7 of IT C
The first experimentally measured reflected intensities were (2004), in Tanner (1976) and in Tanner & Bowen (1992). For the
not in agreement with the theoretical values obtained with the use of Bragg-angle measurements for accurate lattice-parameter
more rigorous dynamical theory, but rather with the simpler mapping, see Hart (1981). For online calculations in the case of
geometrical theory. The integrated reflected intensities calculated multiple diffraction, grazing incidence or for strained crystals, see
using geometrical theory are proportional to the square of the http://sergey.gmca.aps.anl.gov.
structure factor, while the corresponding expressions calculated A reminder of some basic concepts in electrodynamics is given
using dynamical theory for an infinite perfect crystal are in Section A5.1.1.1 of the Appendix.
proportional to the modulus of the structure factor. The inte-
grated intensity calculated by geometrical theory is also propor-
5.1.2. Fundamentals of plane-wave dynamical theory
tional to the volume of the crystal bathed in the incident beam.
This is due to the fact that one neglects the decrease of the 5.1.2.1. Propagation equation
incident amplitude as it progresses through the crystal and a The wavefunction
associated with an electron or a neutron
fraction of it is scattered away. According to geometrical theory, beam is scalar while an electromagnetic wave is a vector wave.
the diffracted intensity would therefore increase to infinity if the When propagating in a medium, these waves are solutions of a
volume of the crystal was increased to infinity, which is of course propagation equation. For electrons and neutrons, this is Schrö-
absurd. The theory only works because the strength of the dinger’s equation, which can be rewritten as
interaction is very weak and if it is applied to very small crystals.
How small will be shown quantitatively in Sections 5.1.6.5 and
5.1.7.2. Darwin (1922) showed that it can also be applied to large 
þ 42 k2 ð1 þ Þ
¼ 0; ð5:1:2:1Þ
imperfect crystals. This is done using the model of mosaic crystals
626
5.2. Dynamical theory of electron diffraction

By A. F. Moodie, J. M. Cowley† and P. Goodman†

5.2.1. Introduction Here k ¼ jkj is the scalar wavenumber of magnitude 2=, and
Since electrons are charged, they interact strongly with matter, so the interaction constant
¼ 2me=h2 . This constant is
that the single scattering approximation has a validity restricted approximately 103 for 100 kV electrons.
to thin crystals composed of atoms of low atomic number. For fast electrons, ’=W is a slowly varying function on a scale
Further, at energies of above a few tens of keV, the wavelength of of wavelength, and is small compared with unity. The scattering
the electron is so short that the geometry of two-beam diffraction will therefore be peaked about the direction defined by the
can be approximated in only small unit cells. incident beam, and further simplification is possible, leading to a
It is therefore necessary to develop a scattering theory specific forward-scattering solution appropriate to HEED (high-energy
to electrons and, preferably, applicable to imaging as well as to electron diffraction).
diffraction. The development, started by Born (1926) and Bethe
(1928), and continuing into the present time, is the subject of an
5.2.3. Forward scattering
extensive literature, which includes reviews [for instance: Howie
(1978), Humphreys (1979)] and historical accounts (Goodman, A great deal of geometric detail can arise at this point and,
1981), and is incorporated in Chapter 5.1. Here, an attempt will further, there is no generally accepted method for approximation,
be made to present only that outline of the main formulations the various procedures leading to numerically negligible differ-
which, it is hoped, will help the nonspecialist in the use of the ences and to expressions of precisely the same form. Detailed
tables. No attempt will be made to follow the historical devel- descriptions of the geometry are given in the references.
opment, which has been tortuous and not always logical, but The entrance surface of the specimen, in the form of a plate, is
rather to seek the simplest and most transparent approach that is chosen as the x, y plane, and the direction of the incident beam is
consistent with brevity. Only key points in proofs will be sketched taken to be close to the z axis. Components of the wavevector are
in an attempt to display the nature, rather than the rigorous labelled with suffixes in the conventional way; K0 ¼ kx þ ky is the
foundations of the arguments. transverse wavevector, which will be very small compared to kz.
In this notation, the excitation error for the reflection is given by

5.2.2. The defining equations K02  jK0 þ 2hj2


h ¼ :
No many-body effects have yet been detected in the diffraction of 4jkz j
fast electrons, but the velocities lie well within the relativistic
region. The one-body Dirac equation would therefore appear to
be the appropriate starting point. Fujiwara (1962), using the An intuitive method argues that, since ’=W  1, then the
scattering matrix, carried through the analysis for forward scat- component of the motion along z is little changed by scattering.
tering, and found that, to a very good approximation, the effects Hence, making the substitution b ¼ expfikz zg and neglecting
of spin are negligible, and that the solution is the same as that @2 =@z2 , equation (5.2.2.1) becomes
obtained from the Schrödinger equation provided that the rela-
 
tivistic values for wavelength and mass are used. In effect a @ 1 2 2
Klein–Gordon equation (Messiah, 1965) can be used in electron ¼i ðrx; y þ K0 Þ þ
’ ; ð5:2:3:1Þ
@z 2kz
diffraction (Buxton, 1978) in the form
  where
2 82 mjej’ 82 m0 jejW jejW
r b þ b þ 1þ b ¼ 0:
h2 h2 2m0 c2
2 @2 @2
rx; y  2
þ 2;
@x @y
Here, W is the accelerating voltage and ’, the potential in the
crystal, is defined as being positive. The relativistic values for
mass and wavelength are given by m ¼ m0 ð1  v2 =c2 Þ1=2, and and ðx; y; 0Þ ¼ expfiðkx x þ ky yÞg.
taking ‘e’ now to represent the modulus of the electronic charge, Equation (5.2.3.1) is of the form of a two-dimensional time-
jej, dependent Schrödinger equation, with the z coordinate replacing
time. This form has been extensively discussed. For instance,
Howie (1966) derived what is essentially this equation using an
 ¼ h½2m0 eWð1 þ eW=2m0 c2 Þ1=2 ; expansion in Bloch waves, Berry (1971) used a Green function in
a detailed and rigorous derivation, and Goodman & Moodie
(1974), using methods due to Feynman, derived the equation as
and the wavefunction is labelled with the subscript b in order to the limit of the multislice recurrence relation. A method due to
indicate that it still includes back scattering, of central impor- Corones et al. (1982) brings out the relationship between the
tance to LEED (low-energy electron diffraction). HEED and LEED equations. Equation (5.2.2.1) is cast in the
In more compact notation, form of a first-order system,
!   !
½r2 þ k2 ð1 þ ’=WÞ b ¼ ðr2 þ k2 þ 2k
’Þ b ¼ 0: ð5:2:2:1Þ @ b 0 1 b
@ b ¼ 2 2 @ b :
@z ðrx; y þ k þ 2k
’Þ 0
@z @z
† Deceased.

647
5.3. Dynamical theory of neutron diffraction

By M. Schlenker and J.-P. Guigay

5.3.1. Introduction energy, hence of velocity, close to the Maxwell distribution


Neutron and X-ray scattering are quite similar both in the characteristic of the temperature T of the moderator. Frequently
geometry of scattering and in the orders of magnitude of the basic used moderators are liquid deuterium (D2, i.e. 2H2) at 25 K, heavy
quantities. When the neutron spin is neglected, i.e. when dealing water (D2O) at room temperature and graphite allowed to heat
with scattering by perfect nonmagnetic crystals, the formalism up to 2400 K; the corresponding neutron distributions are termed
and the results of the dynamical theory of X-ray scattering can be cold, thermal and hot, respectively.
very simply transferred to the case of neutrons (Section 5.3.2). The interaction of a neutron with an atom is usually described
Additional features of the neutron case are related to the neutron in terms of scattering lengths or of scattering cross sections. The
spin and appear in diffraction by magnetic crystals (Section main contribution corresponding to the nuclear interaction is
5.3.3). The low intensities available, coupled with the low related to the strong force. The interaction with the magnetic
absorption of neutrons by most materials, make it both necessary field created by atoms with electronic magnetic moments is
and possible to use large samples in standard diffraction work. comparable in magnitude to the nuclear term.
The effect of extinction in crystals that are neither small nor bad
enough to be amenable to the kinematical approximation is
therefore very important in the neutron case, and will be 5.3.2.2. Scattering lengths and refractive index
discussed in Section 5.3.4 together with the effect of crystal The elastic scattering amplitude for scattering vector s, f ðsÞ, is
distortion. Additional possibilities arise in the neutron case defined by the wave scattered by an object placed at the origin
because the neutrons can be manipulated from outside through when the incident plane wave is
i ¼ A exp½iðk0  r  !tÞ,
applied fields (Section 5.3.5). Reasonably extensive tests of the written as
s ¼ A½ f ðsÞ=r exp½iðkr  !tÞ with k ¼ jk0 j ¼
predictions of the dynamical theory of neutron diffraction have jk0 þ sj ¼ 2=. In the case of the strong-force interaction with
been performed, with the handicap of the very low intensities of nuclei, the latter can be considered as point scatterers because
neutron beams as compared with X-rays: these are described in the interaction range is very small, hence the scattering amplitude
Section 5.3.6. Finally, the applications of the dynamical theory in is isotropic (independent of the direction of s). It is also inde-
the neutron case, and in particular neutron interferometry, are pendent of  except in the vicinity of resonances. It is conven-
reviewed in Section 5.3.7. tionally written as b so that most values of b, called the
scattering length, are positive. A table of experimentally
measured values of the scattering lengths b is given in IT C for
the elements in their natural form as well as for many individual
5.3.2. Comparison between X-rays and neutrons with spin isotopes. It is apparent that the typical order of magnitude is the
neglected fm (femtometre, i.e. 1015 m, or fermi), that there is no systematic
variation with atomic number and that different isotopes have
5.3.2.1. The neutron and its interactions very different scattering lengths, including different signs. The
An excellent introductory presentation of the production, first remark implies that scattering amplitudes of X-rays and of
properties and scattering properties of neutrons is available neutrons have comparable magnitudes, because the characteristic
(Scherm & Fåk, 1993, and other papers in the same book). A length for X-ray scattering (the scattering amplitude for forward
stimulating review on neutron optics, including diffraction by scattering by one free electron) is R ¼ 2:8 fm, the classical
perfect crystals, has been written by Klein & Werner (1983). electron radius. The second and third points explain the impor-
X-rays and neutrons are compared in terms of the basic quan- tance of neutrons in structural crystallography, in diffuse scat-
tities in Table 4.1.3.1 of IT C (2004), where Chapter 4.4 is devoted tering and in small-angle scattering. Scattering of neutrons by
to neutron techniques. condensed matter implies the use of the bound scattering lengths,
The neutron is a massive particle for which the values relevant as tabulated in IT C. The ‘free’ scattering length, used in some
to diffraction are: no electric charge, rest mass m ¼ presentations, is obtained by multiplying the bound scattering
1:675  1027 kg, angular momentum eigenvalues along a given lengths by A=ðA þ 1Þ, where A is the mass of the nucleus in
direction h- =2 (spin 12) and a magnetic moment of 1.913 atomic units.
nuclear magneton, meaning that its component along a A description in terms of an interaction potential is possible
quantization direction z can take eigenvalues z ¼ using the Fermi pseudo-potential, which in the case of the nuclear
0:996  1026 A m2 . The de Broglie wavelength is  ¼ h=p interaction with a nucleus at r0 can be written as VðrÞ ¼
where h is Planck’s constant (h ¼ 2h- ¼ 6:625  1034 J s) and p ðh2 =2mÞbðr  r0 Þ, where  denotes the three-dimensional Dirac
is the linear momentum; p ¼ mv in the nonrelativistic approx- distribution.
imation, which always applies in the context of this chapter, v Refraction of neutrons at an interface can be conveniently
being the neutron’s velocity. The neutron’s wavelength, , and described by assigning a refractive index to the material, such
kinetic energy, Ec , are thus related by  ¼ h=ð2mEc Þ1=2, or, that the wavenumber in the material, k, is related to that in a
in practical units,  ½Å ¼ 9:05=ðEc ½meVÞ1=2 . Thus, to be of vacuum, k0 , by k ¼ nk0. Here
interest for diffraction by materials, neutrons should have kinetic
energies in the range 100 to 102 meV. In terms of the velocity, !1=2
 ½Å ¼ 3:956=ðv ½km s1 Þ. 2 X
n¼ 1 b ;
Neutron beams are produced by nuclear reactors or by spal- V i i
lation sources, usually pulsed. In either case they initially have an
energy in the MeV range, and have to lose most of it before they
can be used. The moderation process involves inelastic inter- where the sum is over the nuclei contained in volume V. PWith
actions with materials. It results in statistical distributions of typical values, n is very close to 1 and 1  n ¼ ð2 =2VÞ i bi is
654

You might also like