VESTA: a Three-Dimensional Visualization System

for Electronic and Structural Analysis

Koichi MOMMA
1
and Fujio IZUMI
2
National Institute for Materials Science,
1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan
April 3, 2011
1
Current aliation: National Museum of Nature and Science,Tokyo, 3-23-1 Hyakunincho, Shinjuku-
ku, Tokyo 169-0073. E-mail: vesta.dev@gmail.com
2
E-mail: IZUMI.Fujio@nims.go.jp
Contents
LICENSE AGREEMENT x
1 INTRODUCTION AND BACKGROUND 1
1.1 Understanding Crystal and Electronic Structures in Three Dimensions . . . . . . 1
1.2 Circumstances behind the Development of VESTA . . . . . . . . . . . . . . . . . 1
1.3 Overview of the Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3.1 General features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3.2 Structural models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3.3 A variety of structural information derived on selection of objects . . . . . 3
1.3.4 Visualization of volumetric data in two and three dimensions . . . . . . . 6
1.3.5 Cooperation with external programs . . . . . . . . . . . . . . . . . . . . . 8
1.3.6 Input and output les . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Supported Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.1 Minimum requirements of hardware . . . . . . . . . . . . . . . . . . . . . 9
1.4.2 Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.3 Mac OS X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.4 Linux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.5 Programming Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5.1 GUI parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5.2 Core parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2 GETTING STARTED 12
2.1 Execution of VESTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.1 Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.2 Mac OS X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.3 Linux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Trouble Shooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Notes on GUI Widgets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3 MAIN WINDOW 15
3.1 Components of the Main Window . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2 Menus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.1 File menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.2 Edit menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2.3 View menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2.4 Objects menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2.5 Utilities menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.6 Help menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
ii
3.3 Tools in Toolbar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3.1 Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3.2 Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3.3 Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3.4 Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4 REPRESENTATION OF STRUCTURE AND VOLUMETRIC DATA 21
4.1 Structural Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.1.1 Objects to be displayed . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.1.2 Styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.2 Volumetric Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5 OPEN AND IMPORT FILES 25
5.1 Open New Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.1.1 Four modes of opening les . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.1.2 A list output after getting new structural data . . . . . . . . . . . . . . . 25
5.2 Import Data Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.2.1 Structural data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.2.2 Volumetric data to draw isosurfaces . . . . . . . . . . . . . . . . . . . . . 26
5.2.3 Volumetric data for surface coloring . . . . . . . . . . . . . . . . . . . . . 27
6 CREATE AND EDIT CRYSTAL STRUCTURES 29
6.1 Structure Dialog Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6.1.1 Space-group symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6.1.2 Lattice parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6.1.3 Fundamental equations in structure analysis . . . . . . . . . . . . . . . . . 31
6.1.4 Parameters related to thermal motion . . . . . . . . . . . . . . . . . . . . 32
6.1.5 Structure parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6.1.6 A list of sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
6.1.7 Options and buttons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
6.2 Additional Lattice Settings Dialog . . . . . . . . . . . . . . . . . . . . . . . . . . 36
6.2.1 Non-conventional lattice settings . . . . . . . . . . . . . . . . . . . . . . . 36
6.2.2 Create super- and sub-lattices . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.2.3 Options and buttons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
7 CREATE BONDS AND POLYHEDRA 41
7.1 Three Search Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7.2 Operating Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7.2.1 Search Bonds and Atoms . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7.2.2 A list of bond specications . . . . . . . . . . . . . . . . . . . . . . . . . . 43
7.3 Hydrogen Bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
8 ADDITIONAL OBJECTS 45
8.1 Vectors on Atoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
8.1.1 A list of atoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
8.1.2 A list of vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
8.1.3 How to attach a vector to atoms . . . . . . . . . . . . . . . . . . . . . . . 46
8.2 Lattice Planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
iii
8.2.1 Appearance of lattice planes . . . . . . . . . . . . . . . . . . . . . . . . . . 46
8.2.2 Add lattice planes frame box . . . . . . . . . . . . . . . . . . . . . . . . . . 48
8.2.3 A list of lattice planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
9 INTERACTIVE MANIPULATIONS 49
9.1 Drawing Boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
9.1.1 Ranges of fractional coordinates . . . . . . . . . . . . . . . . . . . . . . . 49
9.1.2 Cuto planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
9.1.3 A list of cuto planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
9.2 Dene a View Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
9.2.1 Manner of specifying directions . . . . . . . . . . . . . . . . . . . . . . . . 50
9.2.2 Current orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
9.2.3 View direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
9.3 Rotate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
9.3.1 Drag mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
9.3.2 Animation mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
9.4 Magnify . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
9.5 Translate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
9.6 Select . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
9.6.1 Atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
9.6.2 Bond . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
9.6.3 Coordination polyhedron . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
9.7 Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
9.8 Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
9.8.1 Bond angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
9.8.2 Dihedral angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
10 PROPERTIES OF OBJECTS 60
10.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
10.1.1 Unit cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
10.1.2 Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
10.2 Atoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
10.2.1 Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
10.2.2 Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
10.2.3 Atom style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
10.2.4 Radius and color . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
10.3 Bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
10.3.1 Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
10.3.2 Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
10.3.3 Bond style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
10.3.4 Radius and color . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
10.4 Polyhedra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
10.4.1 Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
10.4.2 Polyhedral style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
10.4.3 Planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
10.4.4 Edges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
10.5 Isosurfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
iv
10.5.1 Isosurfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
10.5.2 Surface coloring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
10.6 Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
11 OVERALL APPEARANCE 71
11.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
11.2 Lighting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
11.3 Projection Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
11.4 Depth-Cueing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
12 UTILITIES 74
12.1 Equivalent Positions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
12.2 Geometrical Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
12.3 Standardization of Crystal Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
12.4 Niggli-Reduced Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
12.5 Powder Diraction Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
12.6 Site Potentials and Madelung Energy . . . . . . . . . . . . . . . . . . . . . . . . . 80
12.7 2D Data Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
12.8 Line Prole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
12.9 Peak Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
12.10Conversion of Electron Densities . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
13 TWO-DIMENSIONAL DATA DISPLAY 85
13.1 Components of the 2D Data Display Window . . . . . . . . . . . . . . . . . . . . 85
13.2 Menus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
13.3 Tools in the Toolbar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
13.3.1 Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
13.3.2 Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
13.3.3 Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
13.4 Create and Edit a 2D Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
13.4.1 (hkl) plane in the bounding box . . . . . . . . . . . . . . . . . . . . . . . 87
13.4.2 (hkl) plane dened by two vectors . . . . . . . . . . . . . . . . . . . . . . 89
13.4.3 Project along [hkl] axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
13.5 Controlling Properties of a 2D Image . . . . . . . . . . . . . . . . . . . . . . . . . 89
14 PREFERENCES 92
14.1 Settings for RIETAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
14.2 Font for Text Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
14.3 Open a New File in . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
14.4 Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
14.5 Start-up Search for Bonds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
14.6 A Setting for Raster Image Export . . . . . . . . . . . . . . . . . . . . . . . . . . 93
14.7 Default isosurface level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
15 INPUT AND OUTPUT FILES 95
15.1 File Formats of Volumetric Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
15.2 Directories for User Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
15.3 Files Used by VESTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
v
15.4 Input Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
15.4.1 Structural data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
15.4.2 Volumetric data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
15.4.3 Structural and volumetric data . . . . . . . . . . . . . . . . . . . . . . . . 104
15.5 Output Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
15.5.1 Data les . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
15.5.2 Raster images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
15.5.3 Vector images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
15.5.4 Output text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
APPENDIX 111
A Scalability of VESTA 111
B Keyboard Shortcuts 112
vi
List of Figures
1.1 Screenshots of VESTA running on three kinds of operating systems . . . . . . . . 3
1.2 Crystal structure of masutomilite represented as the ball-and-stick model . . . . 4
1.3 Crystal structure of a polymorph of vitamin B1 represented as the space-lling
model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Crystal structure of beryl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Crystal structure of sodalite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6 Crystal structure of the tetragonal form of melanophlogite . . . . . . . . . . . . . 6
1.7 A thermal-ellipsoid model of 17-(2H-indazol-2-yl)androsta-5,16-dien-3-ol having
an indazole substituent at the C17 position [1]. C: brown, N: green, O: red, H:
sky-blue. The probability for atoms to be included in the ellipsoids was set at 50
% except for the small and spherical H atoms. . . . . . . . . . . . . . . . . . . . . 6
1.8 Electron-density distribution in MgB
2
. . . . . . . . . . . . . . . . . . . . . . . . 7
1.9 A (001) slice illustrating electron-density distribution in MgB
2
. . . . . . . . . . 7
1.10 Nuclear-density distribution in the paraelectric phase of KH
2
PO
4
. . . . . . . . . 8
3.1 Main window of VESTA running on Windows XP . . . . . . . . . . . . . . . . . 15
4.1 Crystal structure of quartz represented as stick model with dot surfaces . . . . . 21
5.1 Import Data dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.2 A dialog box to choose operations for volumetric data . . . . . . . . . . . . . . . 27
5.3 Distributions of electron densities and eective spin densities in an O
2
molecule . 28
6.1 Structure dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6.2 Crystal structure of Cs
6
C
60
, with C
60
represented by translucent polyhedra . . . 34
6.3 Additional Lattice Settings dialog box. . . . . . . . . . . . . . . . . . . . . . . . . . 36
6.4 An example of lattice transformation between primitive and face-centered-cubic
lattices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.5 A dialog box asking whether additional sites are searched or not . . . . . . . . . 40
7.1 Bonds dialog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7.2 Search molecules mode with and without Beyond the boundary option enabled 43
7.3 Crystal structure of -AlOOH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
8.1 Vectors dialog box displaying a list of atoms in SrFeO
2
. . . . . . . . . . . . . . . 45
8.2 A dialog box to create or edit specications of a vector . . . . . . . . . . . . . . . 46
8.3 Lattice Plane dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
8.4 A section of a dierence Fourier map inserted in a ball-and-stick model of AlOOH 47
9.1 Boundary dialog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
vii
9.2 Electron density distribution in D-sorbitol . . . . . . . . . . . . . . . . . . . . . . 50
9.3 Orientation dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
9.4 Calculation of a dihedral angle for four carbon atoms in an aromatic ring in
3-[4-(dimethylamino)phenyl]-1-(2-hydroxyphenyl)-prop-2-en-1-one . . . . . . . . . 58
10.1 General page in the Properties dialog box . . . . . . . . . . . . . . . . . . . . . . 60
10.2 Atoms page in the Properties dialog box . . . . . . . . . . . . . . . . . . . . . . . 61
10.3 Thermal ellipsoids of an atom (A) with principal ellipses and (B) without them . 62
10.4 Bonds page in the Properties dialog . . . . . . . . . . . . . . . . . . . . . . . . . 63
10.5 Polyhedra page in the Properties dialog . . . . . . . . . . . . . . . . . . . . . . . 64
10.6 Isosurfaces page in the Properties dialog box . . . . . . . . . . . . . . . . . . . . . 65
10.7 Comparison between the two modes of rendering isosurfaces . . . . . . . . . . . . 66
10.8 The 64a
1g
orbital for the {Cd[S
4
Mo
3
(Hnta)
3
]
2
}
4
ion [2] with a ball-and-stick
model. The isosurface levels of the wave function were set at 0.01a
3/2
0
(yellow)
and 0.01a
3/2
0
(blue), where a
0
is the Bohr radius. . . . . . . . . . . . . . . . . 67
10.9 Composite images of isosurfaces and a ball-and-stick model for albatrossene . . . 67
10.10Relations between orientation of isosurface and the two parameters for opacity
setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
10.11Sections page in the Properties dialog box . . . . . . . . . . . . . . . . . . . . . . 69
10.12Distribution of nuclear densities obtained for KOD at 580 K by MEM from single-
crystal neutron diraction data [3]. Coordinate ranges from (0, 0, 0) to (1, 1, 1)
were drawn with a cuto plane of (111). D atoms are highly disordered around
O atoms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
10.13Relations among saturation levels, data values, and colors of sections . . . . . . . 70
11.1 Overall Appearance dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
11.2 Crystal structure of post-perovskite, MgSiO
3
, rendered by parallel and perspec-
tive projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
11.3 Crystal structure of mordenite rendered with and without depth-cueing . . . . . 73
12.1 Equivalent Positions dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
12.2 Geometrical Parameters dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
12.3 Simulation of an X-ray powder diraction pattern of YBa
2
Cu
4
O
8
. . . . . . . . . 79
13.1 2D Data Display window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
13.2 A dialog box for specications of a 2D image . . . . . . . . . . . . . . . . . . . . 87
13.3 Three modes of creating a 2D image . . . . . . . . . . . . . . . . . . . . . . . . . 88
13.4 Birds eye view of a (001) slice of rutile-type TiO
2
. . . . . . . . . . . . . . . . . 90
13.5 Electron-density distribution on the (001) plane in rutile-type TiO
2
with and
without grid edges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
13.6 Contours page in the Side Panel . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
14.1 Preferences dialog box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
15.1 Two types of the grids for volumetric data . . . . . . . . . . . . . . . . . . . . . . 95
viii
List of Tables
6.1 Axis choices in the orthorhombic space groups . . . . . . . . . . . . . . . . . . . . 31
6.2 Non-standard settings in two triclinic space groups . . . . . . . . . . . . . . . . . 31
15.1 Files used by VESTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
A.1 Maximum numbers of objects allowed in VESTA. . . . . . . . . . . . . . . . . . . 111
B.1 Keyboard shortcuts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
ix
LICENSE AGREEMENT
VESTA LICENSE
Version 2
Copyright 20072009, Koichi Momma and Fujio Izumi
VESTA is currently copyrighted with its source code not open to the public; we wish to
control the development and future of VESTA by ourselves. This software is distributed free
of charge for academic, scientic, educational, and non-commercial users. Users belonging to
commercial enterprises may also use this software at no cost until a license for business users is
established.
Permission to use this software is hereby granted under the following conditions:
1. Drawings produced by VESTA may be used in any publications provided that its use is
explicitly acknowledged. A suitable reference for VESTA is:
K. Momma and F. Izumi, VESTA: a three-dimensional visualization system for electronic
and structural analysis, J. Appl. Crystallogr., 41, 653658 (2008).
2. You should not redistribute any copy of the distributed les including MADEL and
STRUCTURE TIDY unless you have a written permission from us.
Part of these terms may be changed without any prior announcement. This software is
provided as is without any expressed or implied warranty.
x
Chapter 1
INTRODUCTION AND
BACKGROUND
1.1 Understanding Crystal and Electronic Structures in Three
Dimensions
Progress in modern structure-renement techniques of the maximum-entropy method (MEM)
[4, 5, 6, 7] and MEM-based pattern tting (MPF) [8, 9], has made it easier and more popular to
determine three-dimensional (3D) distribution of electron densities from X-ray diraction data
and densities of coherent-scattering lengths (nuclear densities), b
c
, from neutron diraction data.
On the other hand, rapid developments of computer hardware and software have accelerated
and facilitated electronic-structure calculations aording physical quantities including electron
densities, wave functions, and electrostatic potentials.
Such technological advances in recent years bring demands for integrated 3D visualization
systems to deal with both structural models and volumetric data such as electron and nuclear
densities. The crystal structures and spatial distribution of various physical quantities obtained
experimentally and by computer simulations should be understood three-dimensionally. Despite
the availability of many structure-drawing programs, cross-platform free software capable of
visualizing both crystal and electronic structures in three dimensions is very few; if any, they
are not very suitable for displaying those of inorganic and metallic compounds.
To improve such a situation, we have developed a new integrated system VESTA (Visualization
for Electronic and STructural Analysis) for 3D visualization of crystal structures and volumetric
data on personal computers.
1.2 Circumstances behind the Development of VESTA
VESTA is a successor to two visualization programs, VICS and VEND, in the VENUS (Visualiza-
tion of Electron/NUclear and Structures) software package
1
[10, 11]. VENUS, which was devel-
oped by Dilanian, Izumi, and Kawamura with help from Ohki and Momma during 20012006,
comprises the following ve programs:
1. VICS (VIsualization of Crystal Structures) for displaying and manipulating crystal struc-
tures [10, 11],
2. VEND (Visualization of Electron/Nuclear Densities) for displaying and manipulating
volumetric data [10, 11],
1
http://homepage.mac.com/fujioizumi/visualization/VENUS.html
1
3. PRIMA (PRactice Iterative MEM Analyses) for MEM analysis from X-ray and neutron
diraction data [8],
4. ALBA (After Le Bail Analysis) for the maximum-entropy Patterson method [12],
5. Alchemy: A le converter to make it possible to analyze observed structure factors and
their estimated standard deviation, which result from Rietveld analysis using GSAS [13]
and FullProf [14], by MEM with PRIMA [15].
VICS and VEND are programs for 3D visualization of crystal and electronic structures,
respectively. They were written in the ANSI C language with the full use of the OpenGL
2
technology. The graphical user interfaces (GUIs) of VICS and VEND were constructed by using
GLUT and GLUI libraries.
VICS and VEND saw the light of day at the end of 2002 and, since then, continued their
growth to be used widely in a variety of studies. However, we never get full satisfaction from their
usability and performance. First, the combined use of VICS and VEND to visualize both crystal
and electronic structures via text les is rather troublesome; on-the-y visualization of these two
kinds of images is highly desired. Second, their GUIs are not very user-friendly because they
are based on the old-fashioned toolkits, GLUT and GLUI, which have been no longer upgraded.
Above all things, they lack scalability and require large system resources owing to unrened
programming.
1.3 Overview of the Program
1.3.1 General features
VESTA [16, 17] is a 3D visualization system for structural models and volumetric data (voxel
data) such as electron and scattering-length (nuclear) densities, wave functions, and electro-
static potentials. Objects (atoms, bonds, coordination polyhedra, isosurfaces, etc.) are rotated,
expanded, shrunken, and translated fast in three dimensions, particularly in the presence of
video cards accelerating the OpenGL instruction set, e.g., those powered by the GeForce and
RADEON graphic processing units (GPUs). The scalability of VESTA is very high; it enables
us to deal with a practically unlimited numbers of objects such as atoms, bonds, polyhedra,
and polygons on isosurfaces so long as memory capacity is enough (see Appendix A). Drawing
boundaries are dened by ranges along x, y, and z axes as well as lattice planes.
As Fig. 1.1 illustrates, it runs on Microsoft Windows (hereafter referred to as Windows),
Mac OS X, and Linux. With VESTA, we can deal with multiple data in the same window, using
a tab user interface. VESTA further supports multiple windows, each of which may contain
multiple tabs corresponding to les.
1.3.2 Structural models
Structural models are represented as ball-and-stick (Fig. 1.2), space-lling (Fig. 1.3), poly-
hedral (Fig. 1.4), wireframe (Fig. 1.5), stick (Fig. 1.6), and thermal-ellipsoid (Fig. 1.7)
models. Ball-and-stick, wireframe, and stick models can be overlapped with dotted surfaces to
accentuate outer surfaces of atoms. Polyhedra may be made translucent so as to make inside
atoms and bonds visible.
We can insert movable lattice planes with variable opacity into a structural model. Vectors
(arrows) can be attached to atoms to represent magnetic moments or directions of static and
dynamic displacements.
2
http://www.opengl.org/
2
A B
C
Figure 1.1: Screenshots of VESTA running on (A) Windows, (B) Linux, and (C) Mac OS X.
Bond-search algorithm in VESTA is highly sophisticated; a variant of the cell index method
by Quentrec and Brot [18, 19] was adopted. This approach is widely used in programs for
molecular dynamics simulation that needs to deal with a large number of atoms.
1.3.3 A variety of structural information derived on selection of objects
Selection of objects (atoms, bonds, and coordination polyhedra) by clicking with a mouse pro-
vides us with a variety of structural information:
fractional coordinates,
symmetry operations and translation vectors,
site multiplicities plus Wycko letters [20],
3
site symmetries,
interatomic distances, bond angles, and dihedral angles,
information about coordination polyhedra including volumes [21], Baurs distortion indices
[22], quadratic elongations [23], bond angle variances [23], eective coordination number
[24, 25, 26], charge distribution [25, 26, 27], bond valence sums of central metals [28, 29, 30],
and bond lengths expected from bond valence parameters.
3
Figure 1.2: Crystal structure of a variant of mica group minerals, masutomilite K(Li,Al,Mn
2+
)
3
-
[(Si,Al)
4
O
10
](F,OH)
2
[31].
Figure 1.3: Crystal structure of a polymorph of vitamin B1 [32].
3
http://homepage.mac.com/fujioizumi/rietan/book/book.html#BVP
4
Figure 1.4: Crystal structure of beryl, Be
3
Al
2
Si
6
O
18
[33].
Figure 1.5: Crystal structure of sodalite, Na
4
Al
3
(SiO
4
)
3
Cl [34].
5
Figure 1.6: Crystal structure of the tetragonal form of melanophlogite,
46SiO
2
6M
14
2M
12
(M
14
= N
2
, CO
2
; M
12
= CH
4
, N
2
) [35]. Bright-blue
and pink spheres in cages of the SiO
4
framework represent M
14
and M
12
sites for guest molecules, respectively.
Figure 1.7: A thermal-ellipsoid model of 17-(2H-indazol-2-yl)androsta-5,16-dien-3-ol having
an indazole substituent at the C17 position [1]. C: brown, N: green, O: red, H: sky-blue. The
probability for atoms to be included in the ellipsoids was set at 50 % except for the small and
spherical H atoms.
1.3.4 Visualization of volumetric data in two and three dimensions
Electron and nuclear densities, wave functions, and electrostatic potentials, Patterson functions,
etc. are visualized as isosurfaces, birds-eye views, and two-dimensional (2D) maps.
The calculation of isosurface geometry has been appreciably accelerated in VESTA by virtue
6
Mg
B
0
a
b
Figure 1.8: Electron-density distribution in MgB
2
. Four hexagonal unit cells are
shown with an isosurface level of 0.11a
3
0
(a
0
: Bohr radius).
0.0a
0
-3
0.05
0.1
0.15
0.25
0.35
0.45
0.2
0.3
0.4
0.5
Figure 1.9: A (001) slice illustrating electron-density distribution on the
z = 1/2 plane in MgB
2
. Contours are plotted up to 0.5a
3
0
with an interval
of 0.05a
3
0
(a
0
: Bohr radius).
of new algorithm adopted in VESTA [16]. VESTA has a feature of surface coloring to illus-
trate another kind of a physical quantity at each point on isosurfaces; this feature has been
thoroughly redesigned to improve the quality of images [16]. Translucent isosurfaces can be
overlapped with a structural model. In VESTA, the visibility of both outlines of isosurfaces
and an internal structural model has been surprisingly improved by introducing two opacity
parameters. Further, we can add 2D slices of volumetric data in their 3D image. The quality
of rendering isosurfaces, boundary sections, and slices by VESTA is very high even when the
resolution of data is relatively low [16].
Figure 1.8 illustrates isosurfaces of electron densities calculated with WIEN2k [36] for a
superconductor MgB
2
[37]. A network of highly covalent BB bonds on the z = 1/2 plane and
the ionic nature of bonds between Mg
2+
ions (z = 0) and B atoms are clearly visualized in this
gure. A (001) slice of electron densities at the z = 1/2 level is depicted in Fig. 1.9.
7
O
H
O
0
a
b
c
Figure 1.10: Nuclear-density distribution in the paraelectric
phase of KH
2
PO
4
.
Figure 1.10 shows isosurfaces of scattering-length densities determined from neutron pow-
der diraction data of KH
2
PO
4
(paraelectric phase) at room temperature by MEM [38]. In
this way, two dierent colors are asigned to positive and negative isosurfaces. Blue isosurfaces
(density: 2.5 fm/

A
3
) for H atoms are elengated toword yellow ones (density: 2.5 fm/

A
3
) for O
atoms because of double minimum potential.
1.3.5 Cooperation with external programs
VESTA collaborates closely with external programs such as ORFFE [39], STRUCTURE TIDY
[40], RIETAN-FP [17], and MADEL (see 12.6). On selection of a bond (2 atoms) or a bond
angle (3 atoms) in a dialog box relevant to geometrical parameters output by ORFFE, the
corresponding object in a ball-and-stick model is highlighted in a graphic window, and vice
versa. STRUCTURE TIDY allows us to standardize crystal-structure data and transform the
current unit cell to a Niggli-reduced cell. RIETAN-FP makes it possible to simulate powder
diraction patterns from lattice and structure parameters. With MADEL, electrostatic site
potentials and a Madelung energy can be calculated from occupancies, fractional coordinates,
and oxidation states of all the sites.
1.3.6 Input and output les
VESTA can read in les with 38 kinds of formats such as CIF, ICSD, and PDB and output les
with 12 kinds of formats such as CIF and PDB (see chapter 15). Users of RIETAN-FP [17] must
be pleased to learn that standard input les, *.ins, can be both input and output by VESTA.
In addition, program ELEN [41] was built into VESTA for conversion of 3D electron densities
into electronic-energy densities and Laplacians [42].
8
The entire crystal data and various settings can be saved in a small text le, *.vesta, without
incorporating huge volumetric data. File *.vesta with the VESTA format contains relative paths
to volumetric-data les and optionally a crystal-data le that are automatically read in when
*.vesta is reopened. VESTA also makes it possible to export graphic-data les with 14 image
formats including 4 vector-graphic ones (see 15.5).
1.4 Supported Platforms
1.4.1 Minimum requirements of hardware
CPU: MMX Pentium 233 MHz or faster.
RAM: 64 MB or more.
Video RAM: 16 MB or more is desirable.
Video card: A video card capable of hardware acceleration of the OpenGL instruction set is
recommended.
Display: A minimum resolution of 1024 768 pixels with ca. 65,000 or ca. 16.7 million colors.
Because VESTA utilizes OpenGL technology, the use of an OpenGL-capable graphics processing
unit (GPU) is highly recommended but not mandatory.
1.4.2 Windows
Both 32- and 64-bit versions of VESTA are available for Windows. They were built with Mi-
crosoft Visual C++ 2008.
4
The 32-bit version was tested on Windows 2000, XP, and Vista (32-
and 64-bit editions) while the 64-bit version only on Windows Vista 64-bit edition. VESTA
has never been tested on Windows ME or earlier; it may or may not work on these operating
systems.
1.4.3 Mac OS X
The Mac OS X version of VESTA is a universal-binary application, which runs on two types
of central processing units (CPUs): PowerPC and Intel x86 CPUs. Mac OS X 10.4 or later is
required. This was build with GCC
5
4.0 included in Xcode 2.5.
1.4.4 Linux
Both 32- and 64-bit versions are available for Linux platforms. The 32-bit version was built with
GCC 3.3 for PCs equipped with Intel x86 CPUs. The 64-bit version was built with GCC 4.0 for
PCs equipped with x86-64 CPUs.
32-bit version
1. Dependency
(a) Gkt+ > 2.4
(b) libstdc++.so.5
(c) libjpeg
(d) libpng
(e) libti
4
http://msdn2.microsoft.com/en-us/visualc/default.aspx
5
http://gcc.gnu.org/
9
(f) Mesa OpenGL library
2. Distributions where VESTA is known to work
(a) Fedora 3, 4, 5, 6, 7, 8, 9, 10
(b) Mandriva 2007.1
(c) openSUSE 10.0, 10.2, 11.0, 11.1
(d) Red Hat Enterprise Linux 4
(e) Ubuntu 6.06, 7.10
6
(f) Vine Linux 3.2, 4.1
64-bit version
1. Dependency
(a) Gkt+ > 2.4
(b) libstdc++.so.6
(c) Mesa OpenGL library
2. Distributions where VESTA is known to work
(a) Fedora 3, 4, 5, 6, 7, 8, 9, 10
(b) Mandriva 2006, 2007.1
(c) openSUSE 10.0, 10.2, 11.0, 11.1
(d) Red Hat Enterprise Linux 4
(e) Ubuntu 7.04 7.10
1.5 Programming Concept
The source code of VESTA comprises GUI and core parts. In this section, only the fundamen-
tal concept of programming is briey described. For further details in algorithmic techniques
adopted in VESTA, read Ref. [16].
1.5.1 GUI parts
To overcome the faults described in 1.2 for VICS and VEND, we at rst upgraded VICS to
VICS-II employing a modern C++ GUI framework wxWidgets
7
[43] to build a new state-of-art
GUI. Later, we further integrated VICS-II and VEND into the next-generation 3D visualization
system VESTA. wxWidgets is a cross-platform application framework (toolkit) written in the
C++ language; it is one of the best toolkits for cross-platform GUI programming. It provides
us with a consistent look-and-feel inherent in each operating system. The license agreement of
wxWidgets,
8
an LGPL-like license with some exceptions allowing binary distribution without
source code and copyright, is exible enough to permit us to develop any types of applications
incorporating wxWidgets.
The GUI parts are carefully separated from the core ones to make it easier to reuse the latter
code. The core parts are basically controlled from the GUI ones. However, functions provided by
wxWidgets are exceptionally called in some core parts. In such a case, the function is wrapped
by another function in a separate le to make the core parts quite independent of GUI toolkits
and to make clear which functions depend on external libraries.
6
Need to create /usr/lib/libti.so.3 as a symbolic link to /usr/lib/libti.so.4
7
http://www.wxwidgets.org/
8
http://www.wxwidgets.org/about/newlicen.htm
10
1.5.2 Core parts
In contrast to the GUI framework, we tried to reuse other parts of the source code as much
as possible. All the global variables related to crystal data and graphic settings are capsulated
into a Scene class to allow object-oriented programming. An instance of the Scene class is
dynamically generated whenever new data are created or read in from les so that multiple
windows and tabs may work simultaneously. The previous programs, VICS and VEND, can
deal with only a limited number of objects and consume a large amount of memory even when
handling a relatively small number of data because both of them use native ANCI C arrays. To
eliminate this limitation with minimum eort and without any appreciable overhead, C arrays
were replaced with a small wrapper class of std::vector.
We also prepared the same kind of a wrapper class, aryVector, for pointers to arrays. This
class permits type-safe programming without unnecessary code duplication of the std::vector
template class. Further, it can be accessed by a [ ] operator in exactly the same way as with C
arrays on a source-code level despite great dierences in real operations on a binary level. We
should only note that all the objects must be deletable; that is, they should be generated by a
new operator. Because they are automatically deleted by a destructor when wasting an array, we
need not concern about memory leaks. We have another genuine reason why the std::vector
class is indirectly used in nearly all cases. Instances of objects can be shared by two or more
arrays of the same form. Elements in an array can be quickly sorted only by manipulating
pointers without copying a large amount of data for real objects.
11
Chapter 2
GETTING STARTED
2.1 Execution of VESTA
2.1.1 Windows
The Windows versions of VESTA are archived in the zip format. To get the 32-bit version of
VESTA for Windows, download
http://www.geocities.jp/kmo mma/crystal/download/VESTA.zip
To get the 64-bit version of VESTA for Windows, download
http://www.geocities.jp/kmo mma/crystal/download/VESTA-win64.zip
Note that the 32-bit version of VESTA can also be used on 64-bit Windows. Extract the whole
contents of the archive le in the same directory. VESTA can be launched in the following four
fashions:
1. Double click the icon of VESTA.exe.
2. Type a command in a command line:
VESTA file1 file2 ...
This command will start VESTA and open specied les, le1, le2, etc. Input les may
be omitted. When the current directory is dierent from the directory of VESTA, you
must input the absolute path of the executable binary le of VESTA. To execute VESTA
by simply typing VESTA regardless of the current directory, set environment variable PATH
to include absolute path of the directory where VESTA is placed.
3. Drag & drop le icons of supported le types on the icon of VESTA.exe.
4. Double click les with extensions associated with VESTA.
To associate an extension with VESTA, right click a le with the extension and select Proper-
ties. Then change the Program section in the Properties dialog box.
Beware that VESTA cannot be executed unless the zip le is decompressed. By default,
double-clicking a zip le on Windows XP or later will open it with Explore, i.e., the standard
le manager of Windows. However, the contents of the archive le seen in the Explore window
are actually not extracted but just previewed. In such a case, copy the whole contents of
the previewed archive into an appropriate directory, typically in C:\Program Files. Then,
VESTA.exe in the VESTA folder can be executed as described above.
12
2.1.2 Mac OS X
The Mac OS X version of VESTA is contained in VESTA.dmg with a disc image format. Down-
load
http://www.geocities.jp/kmo mma/crystal/download/VESTA.dmg
and mount its disc image by double-clicking the icon of VESTA.dmg. Copy the whole contents
of the disc image into an appropriate directory, e.g., /Applications/VESTA. VESTA can be
launched in the following four fashions:
1. Double click the icon of VESTA (or VESTA.app).
2. Type a command in a command line:
open -a VESTA file1 file2 ...
3. Drag & drop le icons of supported le types on the icon of VESTA.
4. Double click les with extensions associated with VESTA.
Note that VESTA in the mounted disc image cannot be executed.
2.1.3 Linux
Linux versions of VESTA are archived in the tar.bz2 format. For 32-bit OSs running on PCs
equipped with Intel x86 CPUs, download
http://www.geocities.jp/kmo mma/crystal/download/VESTA-i686.tar.bz2
For 64-bit OSs running on PCs equipped with AMD 64 CPUs, download
http://www.geocities.jp/kmo mma/crystal/download/VESTA-x86 64.tar.bz2
Extract the whole contents of the archive le to a directory. Then, execute VESTA in that
directory by double-clicking the VESTA le on a le manager or by typing the following
command in a command line:
VESTA file1 file2 ...
This command will start VESTA to open the specied les, le1, le2, etc. Of course, the input
les may be omitted. When the current directory is dierent from that of VESTA, you must
input the absolute path of the executable binary le of VESTA. To launch VESTA by simply
typing VESTA regardless of the current directory, create a symbolic link of VESTA in a directory
where environment variable path is set. You can optionally add VESTA to a panel, dock, or
the application menu.
VESTA does not run under some locale settings of OSs where commas are used instead of dec-
imal points. Users of such a locale setting should set environment variable LANG to en_US.UTF-8
before running VESTA. A quick solution for this problem is to execute VESTA by entering the
following command:
LANG=en_US.UTF-8 VESTA
VESTA crashes on Ubuntu 6.06 with some locale settings. In this case, try the following com-
mand:
GTK_IM_MODULE=xim VESTA
13
2.2 Trouble Shooting
If you fail in execution of VESTA on a Linux platform, try to execute it in a command line.
Then you will probably get some information about the cause of the problem. Most likely, the
failure is caused by missing libraries. In such a case, the names of missing libraries are output
to the terminal.
VESTA may not function properly with part of video cards; most of such troubles seem to
arise from bugs of their drivers. Updating drivers to latest ones may solve the problems. If some
troubles are encountered on the use of VESTA, please try to run it on another machine to check
whether or not the trouble is caused by a bug in it.
2.3 Notes on GUI Widgets
Throughout this document, the following symbols are used to show kinds of input data:
[Button] : A button (dotted lines appear after clicking it).
Option : A radio button and a check box (words following it may be clicked to select
it).
{Text Box} : An input item (e.g., a value or a name) including spinners and list boxes.
Tab : A page in a multiple-page user interface (an underscore is drawn below the
tab name).
<Key> : A key in the keyboard.
The text box supports four types of basic arithmetic operations: +, , , and /, which
means that you can input, for example, 1/3 instead of 0.333333. Pressing the <Tab> or
<Enter> key will focus a next control, i.e., text box, button, ratio button, or check box, in a
dialog box. Press <Shift> + <Tab> to focus the preceding control.
14
Chapter 3
MAIN WINDOW
3.1 Components of the Main Window
Figure 3.1 shows the main window of VESTA running on Windows XP. The Main Window
consists of the following six parts:
Figure 3.1: Main window of VESTA running on Windows XP.
1. Menu bar: File, Edit, View, Objects, Utilities, and Help menus are placed on
the menu bar. The menu bar is placed at the top of the Main Window in Windows and
Linux while menus are displayed at the top of the screen in Mac OS X.
2. Toolbar: Tools used frequently.
15
3. Tab: Two or more les loaded in a single Main Window are switched with tabs.
4. Side Panel: Tools and options used frequently.
5. Graphics Area: An area where structural models and volumetric data are displayed in three
dimensions.
6. Text Area: Two types of texts are displayed here. Information on symmetry operations
plus translations, geometrical parameters, physical quantities related to coordination poly-
hedra, etc. is output in the Output page while comments on data displayed currently can
be input under the Comment page.
The Text Area corresponds to a standard output stream equivalent to the Command Prompt
window on Windows and a terminal window on Mac OS X or Linux. Just after launching
VESTA, the Text Area displays information about the PC system you are using. The item
OpenGL version denotes the maximum version of OpenGL implementation supported by the
system. Video configuration provides information about the GPU. For example, in a Win-
dows PC equipped with Quadro FX 4600, a message
Video configuration: Quadro FX 4600/PCI/SSE2
is displayed in the Text Area. SSE2 means that Streaming SIMD Extensions 2 is supported in
this video card. In the case of a Power Mac G5 (Dual 2.5 GHz) equipped with ATI Radeon
9600 XT, a message
Video configuration: ATI Radeon 9600 XT OpenGL Engine
is issued in the Text Area. If the GPU of your PC does not support any hardware acceleration of
OpenGL, Video configuration would be GDI Generic on Windows, and Software Rasterizer
or Mesa GLX Indirect on Linux.
3.2 Menus
3.2.1 File menu
New Structure...: Open a Structure dialog box to input crystal data of a new structure.
New Window...: Open a new Main Window.
Open...: Open les with a le selection dialog box.
Import Data...: Import data les into the currently displayed data. An Import Data
dialog box will be opened.
Save...: Save current data to a le, *.vesta, with the VESTA format. If the current data
has been input from *.vesta or once saved as *.vesta, *.vesta is overwritten. Otherwise, a
le selection dialog box will appear to prompt you to input a new le name.
Save As...: Save current data to a new le, *.vesta, with the VESTA format. A le
selection dialog box will appear to prompt you to input a new le name.
Export Data...: Export current data to a le with formats other than the VESTA format.
Export Raster Image...: Export a graphic image to a le with a raster (pixel-based)
format.
16
Export Vector Image...: Export a graphic image to a le with a vector format.
Save Output Text...: Save text data in the Text Area as a text le.
Close: Close the current page. When no data are displayed, this menu is practically the
same as the Exit menu.
Exit: Exit VESTA.
3.2.2 Edit menu
Structure...: Open a Structure dialog box to edit structural data.
Bonds...: Open a Bonds dialog box to create or edit data related to bonds.
Vectors...: Open a Vectors dialog box.
Lattice Planes...: Open a Lattice Planes dialog box.
Preferences...
1
: Open a Preferences dialog box.
3.2.3 View menu
Antialiasing: Enable or disable antialiasing in the Graphics Area.
Parallel: Switch to parallel representation.
Perspective: Switch to perspective representation.
Zoom In: Zoom in objects.
Zoom Out: Zoom out objects.
Fit to Screen: Scale and center justify objects to t to the Graphics Area.
Overall Appearance...: Open an Overall Appearance dialog box.
Clear Text Are: Clear all the text in the Text Area.
3.2.4 Objects menu
Structural Model: items of this menu can be also selected in the Side Panel.
Show Model: Show or hide a structural model.
Show Dot Surface: Show or hide outer surfaces of atoms as dotted surfaces.
Ball-and-Stick: Show a ball-and-stick model.
Space-lling: Show a space-lling model.
Polyhedral: Show a a polyhedral model.
Wireframe: Show a wireframe model.
Stick: Show a stick model.
Volumetric Data: items of this menu can be also selected in the Side Panel.
Show Isosurfaces: Show or hide isosurfaces.
1
In the Mac OS X version, the Preferences item is placed under the VESTA menu.
17
Show Sections: Show or hide sections of isosurfaces.
Surface Coloring: Enable or disable the surface coloring of isosurfaces.
Smooth Shading: Show isosurfaces with a smooth-shading model.
Wireframe: Show isosurfaces with a wireframe model.
Dot Surface: Show isosurfaces as dotted surface.
Properties
General...: Open a Properties dialog box to select the General page.
Atoms...: Open a Properties dialog box to select the Atoms page.
Bonds...: Open a Properties dialog box to select the Bonds page.
Polyhedra...: Open a Properties dialog box to select the Polyhedra page.
Isosurfaces...: Open a Properties dialog box to select the Isosurfaces page.
Sections...: Open a Properties dialog box to select the Sections page.
Boundary...: Open a Boundary dialog box.
Orientation...: Open an Orientation dialog box.
3.2.5 Utilities menu
Equivalent Positions...: Open an Equivalent Positions dialog box to list general equivalent
positions.
Geometrical Parameters...: Open a Geometrical Parameters dialog box.
Standardization of Crystal Data...: Standardize the space-group setting and fractional
coordinates.
Niggli-Reduced Cell: Transform the current unit cell to a Niggli-reduced cell.
Powder Diraction Pattern...: Simulate a powder diraction pattern with RIETAN-
FP [17] and display the result with an external graph-plotting program.
Site Potentials and Madelung Energy...: Calculate site potentials and the Madelung
energy of the currently displayed crystal by an external program, MADEL.
2D Data Display...: Open a 2D Data Display window.
Line Prole...: Calculate a line prole of volumetric data between two positions and
output them in a text le.
Peak Search...: Search peaks in volumetric data.
Conversion of Electron Densities...: Convert 3D electron densities into electronic-
energy densities and Laplacians.
18
3.2.6 Help menu
Manual...: Open the users manual of VESTA in a PDF viewer.
2
Check for Updates...: Open the web page of VESTA
3
with a browser to check whether
or not a new version of VESTA has been released.
About VESTA...
4
: Show information about VESTA.
3.3 Tools in Toolbar
3.3.1 Alignment
View along the a axis
View along the b axis
View along the c axis
View along the a* axis
View along the b* axis
View along the c* axis
The above six buttons are used to align objects parallel to the a, b, or c axis, or parallel to the
a

, b

, or c

axis, respectively.
3.3.2 Rotation
Rotate around the x axis
Rotate around the x axis
Rotate around the y axis
Rotate around the y axis
Rotate around the z axis
Rotate around the z axis
These six buttons are used to rotate objects around the x, y, or z axis. The step width of
rotation (in degrees) is specied in the text box next to the sixth button:
2
The present PDF le, VESTA Manual.pdf, need to share the same folder with a binary executable le of
VESTA.
3
http://www.geocities.jp/kmo mma/crystal/en/vesta.html
4
In the Mac OS X version, this item is placed under the VESTA menu.
19
3.3.3 Translation
Translate upward
Translate downward
Translate leftward
Translate rightward
These four buttons are used to translate objects upward, downward, leftward, and rightward,
respectively. The step width of translation (in pixels) is specied in the text box next to the
fourth button:
3.3.4 Scaling
Zoom in
Zoom out
Fit to the screen
These three buttons are used to change object sizes. The step width of zooming (in %) is
specied in a text box next to the third button:
20
Chapter 4
REPRESENTATION OF STRUC-
TURE AND VOLUMETRIC DATA
4.1 Structural Models
The Structural model frame box contains frequently used tools to control representation of struc-
tural models. The same options are also placed in the Model item under the Objects menu.
This frame box is disabled for data with no structural model.
4.1.1 Objects to be displayed
Show model
This option controls the visibility of a structural model. When this option is checked (default),
a structural model is visible; otherwise, no structural model is shown. Uncheck this option when
you want to see only isosurfaces and sections for data containing both structural and volumetric
ones.
Show dot surface
In ball-and-stick, wireframe, and stick models, dot-surface spheres are added with radii corre-
sponding to 100 % of user specied ones in the same manner as with a space-lling model if
Show dot surface is checked (Fig. 4.1). This mode is designed to accentuate outer surfaces of
Figure 4.1: Crystal structure of quartz [44] represented
by a stick model with dot surfaces. Si: blue, O: red.
21
atoms. Each sphere is represented as though it were a hollow shell with numerous dots placed
on the surface. The combination of dot surface with a stick model is useful for understanding
how atoms are combined with each other in a molecule. The density of dots is controlled by
two parameters, {Stacks} and {Slices} in the Atoms page of the Properties dialog box as with
the same manner as solid spheres.
4.1.2 Styles
As described in 1.3.2, VESTA represents crystal structures by the ve dierent styles: ball-and-
stick, space-lling, polyhedral, wireframe, and stick models. When atoms are drawn as spheres,
they are rendered with radii corresponding to 40 % of actual atomic radii in all but the space-
lling model, where atoms are rendered with the actual atomic radii. Default radii of atoms are
selected from three types: atomic, ionic, and van der Waals radii. The radius of each element
and a type of radii are specied at the Atoms page in the Properties dialog box. Features in
each structural model are described below with screenshots of the structure for quartz [44] on
the right side.
Ball-and-stick
In the Ball-and-stick model, all the atoms are expressed as solid
spheres or thermal ellipsoids. Bonds are expressed as either cylinders
or lines.
Space-lling
In the Space-lling model, atoms are drawn as interpenetrating
solid spheres, with radii specied at the Atoms page in the Properties
dialog box. This model is useful for understanding how atoms are
packed together in the structure.
Polyhedral
In the polyhedral model, crystal structures are represented by coor-
dination polyhedra where central atoms, bonds, and apex atoms may
also be included. Bonds between central and apex atoms have to be
searched with the Bonds dialog box to display coordination polyhedra
comprising them. Atoms are expressed as solid spheres or thermal
ellipsoids. Bonds are expressed as either cylinders or lines. Needless
to say, the transparency of the coordination polyhedra must be high
enough to make it possible to see the central atoms and bonds. One
of six dierent styles for representing polyhedra is specied at the
Polyhedra page in the Properties dialog box.
22
Wireframe
In the Wireframe model, atoms having no bonds are drawn as wire-
frame spheres whereas those bonded to other atoms are never drawn.
All the bonds are presented as lines with gradient colors. This model
is useful for seeing and manipulating complex and/or large structures
because this is usually the fastest model for rendering structures on
the screen.
Stick
In the Stick model, atoms with no bonds are drawn as solid spheres
while atoms bonded to other atoms are never drawn. All the bonds
are expressed as cylinders, whose properties can be changed at the
Bonds page in the Properties dialog box. This model serves to see
frameworks or molecular geometry.
Thermal ellipsoids
In the ball-and-stick and polyhedral models, atoms can be rendered
as thermal ellipsoids. The probability for atoms to be included
in the ellipsoids is also specied in the Properties dialog box (see 10.2).
4.2 Volumetric Data
The Volumetric data frame box contains frequently used tools to control representation of volu-
metric data. The same options are also placed in the Volumetric Data item under the Objects
menu. This frame box is disabled for data containing no volumetric ones.
Show sections
This option controls whether or not sections of isosurfaces are visible. When this option
is checked (default), sections are visible, otherwise they are not shown.
Show isosurfaces
This option controls whether or not isosurfaces of volumetric data are visible. When this
option is checked (default), isosurfaces are visible, otherwise they are not shown. When
you want to see only a structural model for data containing both structural and volumetric
ones, uncheck this option in addition to the Show sections option.
Surface coloring
Surface coloring means that colors of isosurfaces drawn from one data set are determined
by a secondary data set. A typical example is coloring of electron-density isosurfaces on
the basis of electrostatic potentials (see Fig. 10.7). This option is enabled only when
the secondary volumetric data for surface coloring has been loaded using the Import Data
dialog box (see 5.2).
23
Styles of isosurfaces are chosen from the following three representations:
Smooth shading
Wireframe
Dot surface
Isosurfaces are drawn as solid surfaces with variable opacity in the Smooth shading mode whereas
isosurface are represented by lines and points, respectively, in the Wireframe and Dot surface
modes.
24
Chapter 5
OPEN AND IMPORT FILES
5.1 Open New Files
5.1.1 Four modes of opening les
Files can be loaded in four dierent fashions. Files to be opened can be specied on execution
of VESTA in a command line. Files can be specied in a le selection dialog box, which is
opened from the Open item in the File menu. In the le selection dialog box, only les with
supported formats can be selected. You can also drag and drop a le icon onto the main window
of VESTA from a le manager (Finder in Mac OS X or Explorer in Windows). If VESTA is
associated with some extensions, les having these extensions can be opened with a le manager
by double-clicking their icons.
5.1.2 A list output after getting new structural data
After structural-data les described in 15.4.1 and 15.4.3 have been input, crystal-structure data
contained in them are displayed in the Text Area, including lattice parameters, a unit-cell vol-
ume, and structure parameters (fractional coordinates, occupancies, and isotropic atomic dis-
placement parameters). For instance, a CIF of PbSO
4
aords the following output:
Title O4 Pb1 S1
Lattice type P
Space group name P n m a
Space group number 62
Setting number 1
Lattice parameters
a b c alpha beta gamma
8.51600 5.39900 6.98900 90.0000 90.0000 90.0000
Unit-cell volume = 321.339447
Structure parameters
x y z g B Site Sym.
1 Pb Pb1 0.18820 0.25000 0.16700 1.000 1.200 4c .m.
2 S S1 0.43700 0.75000 0.18600 1.000 0.700 4c .m.
3 O O1 0.59500 0.75000 0.10000 1.000 1.600 4c .m.
4 O O2 0.31900 0.75000 0.04300 1.000 1.600 4c .m.
5 O O3 0.41500 0.97400 0.30600 1.000 1.600 8d 1
It should be noted that VESTA is capable of deriving multiplicities plus Wycko letters (Site)
25
and site symmetries (Sym.) automatically for atoms in the asymmetric unit even if they are not
included in structural-data les. Of course, these data are also given for nonstandard lattice
settings (see 12.3).
Such a type of a list is output when
inputting structural-data les,
changing a lattice-setting number in the Structure dialog box (see p. 30),
transforming the coordinate system using a transformation matrix (see 6.2),
standardizing crysta data (see 12.3).
5.2 Import Data Dialog Box
With the Import Data dialog box (Fig. 5.1), structural data and two or more volumetric data
for isosurfaces and surface coloring can be input from les specied by the user.
Figure 5.1: Import Data dialog box.
5.2.1 Structural data
Click the [Browse...] button at the top right of the Import Data dialog box to import structural
data into the current page. Then, select a le with a supported format of structural data in the
le selection dialog box.
Option Link is used to link the structural data le when saving the current data as a
le, *.vesta, with the VESTA format. Structural data are usually recorded in *.vesta whereas
for volumetric data, relative paths to the data les are recorded in *.vesta instead of directly
recording the volumetric data. When option Link is checked, structural data are also recorded
as a relative path to the data le. This means that even if the structural data le is modied after
*.vesta is saved, the modications are reected in VESTA when *.vesta is reopened. This option
is useful to use *.vesta format as a template le for settings of objects and overall appearances.
Option Link is ineective unless you save the current data in *.vesta.
5.2.2 Volumetric data to draw isosurfaces
VESTA enables us to deal with more than two volumetric data sets. Clock [Browse...] to select
a le in the le selection dialog box. Only les with extensions of supported formats are visible
in the le selection dialog box. After selecting a volumetric data le, a Choose operations dialog
box appears (Fig. 5.2).
26
Figure 5.2: A dialog box to choose operations for volumetric data.
Operation
In the Operation radio box, you can select one of the following ve data operations in addition
to conversion of data units.
Add to current data
Subtract from current data
Replace current data
Multiply to current data
Divide current data by new data
Multiply to current data is convenient when squaring wave functions to obtain electron densities
(existing probabilities for electrons). The last two options are not displayed when no volumetric
data are contained in the current data. In such a case, the rst and third options have also the
same eect.
Convert the unit
In the Convert the unit radio box, the unit of data can be converted from

A
3
to bohr
3
, and
vice versa. Select option Other factor to multiply the data by an arbitrary factor. Then, click
[OK] to import data. Repeating the above procedures allows you to import multiple data, as
exemplied in Fig. 5.3.
5.2.3 Volumetric data for surface coloring
Volumetric data for surface coloring can be imported in the same manner as with data for
isosurfaces. A typical application of surface coloring is to color isosurfaces of electron densities
according to electrostatic potentials.
27
A B
C
Figure 5.3: Distributions of electron densities and eective spin densities in an O
2
molecule.
(A) up-spin electron densities,

, (B) down-spin electron densities,

, and (C) eective spin

densities, =

, calculated with VESTA. Both

and

were calculated with DVSCAT

[45]. Isosurface levels were set at 0.01a
3
0
in (A) and (B), and at 0.001a
3
0
in (C). where a
0
is
the Bohr radius.
28
Chapter 6
CREATE AND EDIT CRYSTAL
STRUCTURES
6.1 Structure Dialog Box
The Structure dialog box (Fig. 6.1) is used to input and edit crystal data of a compound. It
also serves to convert an axis setting into another one.
Figure 6.1: Structure dialog box.
29
6.1.1 Space-group symmetry
Selection of space groups
There are two ways of specifying the space group: using spinner {Number} and list box {Std.
Symbol}. In the spinner, use up and down buttons, or input a space group number directly in the
text box and press <Enter>. Then, the crystal system and space group symbol is automatically
updated. To select a space group symbol in the list box, select the crystal system from list box
{System} at rst, and then click list box {Std. Symbol}. List box {Std. symbol} is automatically
ltered to show only space groups belonging to the selected crystal system.
The space group on the use of Cartesian coordinates
For convenience, space group P1 (triclinic, No. 1) having only one equivalent position, (x, y,
z), in the unit cell is always assigned to structural data recorded on the basis of the Cartesian
coordinate system.
Lattice settings
The number of available settings for the selected space group is listed in list box {Setting}. If a
non-standard setting of the selected space group is preferred to the standard one, your cell choice
has to be specied in list box {Setting}. A space-group symbol of the current setting is displayed
below the setting number. This may be dierent from the standard symbol selected in list box
{Std. Symbol} when the setting number is not 1. Lattice settings available in list box {Setting}
are basically those compiled in International Tables for Crystallography, volume A [20]. In the
monoclinic system, for example, settings with unique axes b and c are available whereas settings
with a unique axis a need to be specied using a transformation matrix (see 6.2.1). When
changing a setting number, consider whether Update structure parameters should be checked
or unchecked, particularly in orthorhombic, tetragonal, and cubic crystals (see Behavior when
changing a lattice setting in p. 35).
In some space groups, additional settings are contained in list box {Setting}. In the or-
thorhombic system, any of six settings, i.e., abc, bac, cab, cba, bca, and acb listed in Table
4.3.2.1 in Ref. [20] may be selected with setting numbers of 1, 2, 3, 4, 5, and 6, respectively.
In orthorhombic space groups having two origin choices, odd and even setting numbers adopts
origin choices 1 and 2, respectively. Then, settings 1 and 2 are of axis choice abc, settings 3
and 4 are of bac, and so on (Table 6.1).
In the triclinic system, complex lattices, A, B, C, I, F, and R, may be selected with setting
numbers of 2, 3, 4, 5, 6, and 7, respectively (Table 6.2).
On changes in crystal axes, e.g., rhombohedral or hexagonal axes in a trigonal crystal, lattice
parameters are automatically varied.
6.1.2 Lattice parameters
In this frame box, lattice parameters are input in the unit of

A (a, b, and c) or degrees (,
, and ). Lattice parameters input by the user are automatically constrained on the basis
of a crystal system. For structural data based on the Cartesian coordinates, virtual lattice
parameters are appropriately determined by VESTA. Estimated standard deviations (e.s.d.s)
of lattice parameters, which are used to calculate e.s.d.s of geometrical parameters, are also
input here.
30
Table 6.1: Axis choices in the orthorhombic space groups.
Space groups
Axis choice No. 48, 50, 59, 68, and 70 Others
Origin choice 1 Origin choice 2
abc 1 2 1
bac 3 4 2
cab 5 6 3
cba 7 8 4
bca 9 10 5
acb 11 12 6
Table 6.2: Non-standard settings in two triclinic space groups.
Setting number Space group (P1) Space group (P

1)
1 P1 P

1
2 A1 A

1
3 B1 B

1
4 C1 C

1
5 I1 I

1
6 F1 F

1
7 R1 R

1
6.1.3 Fundamental equations in structure analysis
Descriptions of structure parameters (see 6.1.5) have to be preceded by those of fundamental
equations in structure analysis (here) and the eect of thermal motion on the diraction intensity
(see 6.1.4).
Let I(h) be the integrated intensity for the reection h with diraction indices of hkl. I(h)
is determined from the observed X-ray or neutron diraction intensity by correcting for it with
Lorentzpolarization, absorption, and extinction factors. I(h) is proportional to the squared
absolute value of the crystal-structure factor, F(h):
I(h) = s|F(h)|
2
, (6.1)
where s is the scele factor that consists of various constants dependent on the diractometer,
measurement conditions, and sample.
In X-ray diraction, F(h) is computed by
F(h) =
n

j=1
g
j
_
f
j
(h) +f

+ if

T
j
(h) exp
_
2i
_
hx
j
+ky
j
+lz
j
_
, (6.2)
where j is the atom number, n is the total number of atoms in the unit cell, g
j
is the occupancy,
f
j
(h) is the atomic scattering factor, the complex number, f

+if

, is the dispersion correction,

T
j
(h) is the DebyeWaller factor, and x
j
, y
j
, and z
j
are fractional coordinates. The summation
in Eq. (6.2) is formally carried out over all the atoms in the unit cell with the volume of V .
31
In the case of neutron diraction, f
j
(h) +f

+if

has to be replaced by the bound coherent

scattering length, b
c
:
F(h) =
n

j=1
g
j
b
c
T
j
(h) exp
_
2i
_
hx
j
+ky
j
+lz
j
_
. (6.3)
The value of b
c
is constant regardless of h because neutrons are scattered not by electrons but
by atomic nuclei (in the absence of magnetic atoms).
The electron density, (x, y, z), at the fractional coordinate of (x, y, z) in the unit cell with
the volume of V is calculated from the F(h)s of all the reections by Fourier synthesis:
(x, y, z) =
1
V

l
F(h) exp[2i(hx +ky +lz)]. (6.4)
The determination of 3D disctribution of (x, y, z)s in the unit cell corresponds to the rough
estimation of atomic coordinates.
6.1.4 Parameters related to thermal motion
Anisotropic and isotropic models
The DebyeWaller factor, T
j
(h), which is often referred to as the temperature factor, is included
in Eqs. (6.2) and (6.3) to represent the eect of thermal displacement for atom j on F(h). The
thermal displacement of the atom is formulated in two dierent ways: anisotropic and isotropic
atomic displacement.
The anisotropic DebyeWaller factor, T(h), is dened as
T(h) = exp
_
2
2
_
h
2
a
2
U
11
+k
2
b
2
U
22
+l
2
c
2
U
33
+ 2hka

U
12
+ 2hla

U
13
+ 2klb

U
23
_
= exp
_

_
h
2

11
+k
2

22
+l
2

33
+ 2hk
12
+ 2hl
13
+ 2kl
23
_
,
(6.5)
where a

, b

, c

, and

are reciprocal lattice parameters. On renement of

ij
s by the
Rietveld method with RIETAN-FP [17],
ij
s for each site have to satisfy linear constraints
imposed on them [46, 47].
The isotropic DebyeWaller factor, T(h), is given by
T(h) = exp
_
B
_
sin

_
2
_
= exp
_

B
4d
2
_
= exp
_

2
2
U
d
2
_
,
(6.6)
where B and U are the isotropic atomic displacement parameters (U = B/8
2
), is the Bragg
angle, is the X-ray or neutron wavelength, and d is the lattice-plane spacing. B and U are
related to the mean square displacement,

u
2

, along the direction perpendicular to the reection

plane with
B = 8
2
U = 8
2

u
2

. (6.7)
For example, a B value of 3

A
2
corresponds to a displacement of about 0.2

A.
Isotropic and anisotropic atomic displacement parameters dened in dierent ways must be
converted into U
ij
,
ij
, B, and U dened above.
32
If only B or U is input for each site by the user, VESTA does not recalculate corresponding
anisotropic atomic displacement parameters. Then, we can easily x B or U for hydrogen atoms
at a small value to show them as balls with a small radius while representing other atoms as
thermal ellipsoids, as illustrated in Fig. 1.7.
Two options
Check Use anisotropic atomic displacement parameters provided that anisotropic atomic dis-
placement parameters are assigned to at least one site and that thermal ellipsoids are to be
drawn.
The type of anisotropic atomic displacement parameters must be specied in list box {Type}:
U to input U
ij
(U
11
, U
22
, U
33
, U
12
, U
13
, and U
23
), and beta to input
ij
(
11
,
22
,
33
,
12
,

13
, and
23
). Using this list box, we can convert U
ij
into
ij
and vice versa.
6.1.5 Structure parameters
Input data
The following structural information is input in the Structure parameters frame box for each site
in the asymmetric unit:
[Symbol...]: Symbol of an element (up to two characters).
{Label}: Site name (up to six characters).
{Charge}: Formal charge (oxidation number; optional)
{x}, {y}, and {z}: Fractional coordinates, x, y, and z (dimensionless).
{esd(x)}, {esd(y)}, and {esd(z)}: Estimated standard deviations of the fractional coordi-
nates (default values: zero).
{g}: Occupancy, g (dimensionless; optional).
{B} or {U}: Isotropic atomic displacement parameter, B or U (in

A
2
; optional).
{U11}, {U22}, {U33}, {U12}, {U13}, and {U23}: Anisotropic atomic displacement param-
eters, U
ij
(in

A
2
, optional).
{beta11}, {beta22}, {beta33}, {beta12}, {beta13}, and {beta23}: Anisotropic atomic dis-
placement parameters,
ij
(dimensionless; optional).
Formal charge
Strictly speaking, formal charges are not structure parameters. They are used to estimate a
bond length from a bond valence parameter [28, 29, 30] (see 9.6.3), determine charge distribution
[25, 26, 27] (see 9.6.3), and electrostatic site potentials and Madelung energy per asymmetric
unit (see 12.6).
Fractional coordinates
For special positions, input fractional numbers, e.g., 1/4, 1/2, and 1/3, or a sucient number
of digits should be given, e.g., 0.333333 and 0.666667.
When thermal ellipsoids are drawn from anisotropic atomic displacement parameters rened
with RIETAN-FP [17], fractional coordinates corresponding to the rst equivalent position de-
scribed in International Tables for Crystallography, volume A [20] for each Wycko position
33
have to be entered. For example, input not (1/2, 0, z) but (0, 1/2, z) for a 4i site in space group
P4/mmm (No. 123).
Figure 6.2: Crystal structure of Cs
6
C
60
[48], with C
60
rep-
resented by translucent polyhedra. A virtual site X with an
occupancy of g = 0 was added at the center of C
60
. Polyhedra
of C
60
were created by specifying XC bonds in the Bonds dia-
log box. CC bonds are also searched by a Search molecules
mode to represent polyhedral edges by solid cylinders. See
chapter 7 for details in creating bonds and polyhedra.
Occupancy
The occupancy is unity for full occupation and zero for virtual sites. If the occupancy of a site is
less than unity, atoms occupying there are displayed as circle graphs for occupancies. If unicolor
balls are preferred to bicolor ones, change the occupancy to unity for convenience. On the use of
some video cards, dirty dots may appear on spheres for partially occupied sites. In such cases,
change their occupancies at unity for convenience.
If the occupancy of a site is zero, the site is treated as a virtual site that is not occupied
by any atoms in the real structure. Neither the virtual sites nor bonds connecting them are
displayed on the screen. Virtual sites are used to visualize, for example, cage structures of
porous crystals as solid polyhedra without creating a large number of unnecessary bonds (Fig.
6.2).
Atomic displacement parameters
The four kinds of the atomic displacement parameters have been dened in 6.1.4. The type
of the isotropic atomic displacement parameters, U or B, is specied in check box Use B as
isotropic atomic displacement parameters. If this check box is unchecked, U is used. After the
type has been changed, values of isotropic atomic displacement parameters are automatically
converted in response to your preference.
34
In crystallographic sites for which anisotropic atomic displacement parameters, U
ij
or
ij
,
have been input, equivalent isotropic atomic displacement parameters, B
eq
or U
eq
, are calculated
from U
ij
, a

, b

, c

, and the metric tensor G [49, 50]:

B
eq
= 8
2
U
eq
(6.8)
with
U
eq
=
1
3

j
U
ij
a

i
a

j
a
i
a
j
=
1
3
_
U
11
(aa

)
2
+U
22
(bb

)
2
+U
33
(cc

)
2
+ 2U
12
a

ab cos + 2U
13
a

ac cos + 2U
23
b

bc cos

.
(6.9)
B
eq
and U
eq
are regarded as B and U in the Structure dialog box after reopening it.
Atom numbers
In addition to the above input data, the total number of atoms (sites) and an atom number for
each atom (site) are displayed in the Structure parameters frame box.
6.1.6 A list of sites
Below the Structure parameters frame box, all the sites (in the asymmetric unit) are listed with
the following four buttons placed at the right side (Fig. 6.1):
[Add]: Add a new site.
[Modify]: Modify the selected site.
[Delete]: Delete the selected site.
[Clear All]: Delete all the sites in the list.
The [Modify] and [Delete] buttons cannot be clicked unless a site in the list is selected.
6.1.7 Options and buttons
Reducing symmetry
Clicking the [Remove symmetry] button generates all the atoms in the
unit cell as independent sites by reducing the symmetry of the crystal
to P1. When the space group is changed to that with higher symmetry,
or when the unit cell is transformed to a smaller one (see 6.2.2), the
same atomic position may result from two or more sites. In such a case, the redundant sites
can be removed by clicking a [Remove duplicate atoms] button. Then, a dialog box (right gure)
opens, prompting you to input a threshold of distances between two atoms to be regarded as a
single site. The threshold value is input in the unit of

A. Increasing the threshold value enables
us, for example, to extract an average structure from a large cell calculated by computational
simulations.
Behavior when changing a lattice setting
The Update structure parameters option allows us to keep the structural geometry of a crystal
when the setting number of a space group is changed. If this option is unchecked, lattice
parameters and fractional coordinates are kept instead of the geometry, which is indispensable
for changing the crystal lattice from the rst setting, which has been arbitrarily assumed by
VESTA, to another one in orthorhombic, tetragonal, or cubic crystals after reading in crystal
35
data from structural-data les such as CIF. VESTA regards the crystal data as based on the rst
setting because it has not any means of learning the actual lattice setting. Even if the second
(third, fourth, .....) setting is adopted, the setting number of 1 is always assumed by VESTA.
On the other hand, correct fractional coordinates for the actual lattice setting are displayed at
the lower part of the Structure dialog box.
Suppose, for instance, a setting number of 2 with an inversion center at the origin is adopted
in crystal data of a compound belonging to space group P4
2
/nmc (No. 137, tetragonal). If
VESTA inputs a CIF of this compound, the rst setting having no inversion center at the origin
is assumed by VESTA. In such a case, the correct setting number should be selected after
unchecking the Update structure parameters option. Then, correct fractional coordinates on
the basis of the second setting remain unchanged. If a lattice setting other than the rst one is
preferred after inputting crystal data on the basis of the rst setting, check the Update structure
parameters option to convert current fractional coordinates into new ones.
Note that even if this option is unchecked, some lattice parameters are still modied as a
result of constraints imposed on the basis of the current space-group setting.
6.2 Additional Lattice Settings Dialog
Clicking the [Option...] button opens the Additional Lattice Settings dialog box (Fig. 6.3) for
transformation of crystal axes.
Figure 6.3: Additional Lattice Settings dialog box.
6.2.1 Non-conventional lattice settings
The Additional Lattice Settings dialog box is used to transform the coordinate system of a crystal
with a transformation matrix [51].
The 44 transformation matrix P consists of the 33 rotation matrix P, the 31 translation
vector p, and the row vector o = (0, 0, 0):
P =
_
P p
o 1
_
=
_
_
_
_
_
P
11
P
12
P
13
p
1
P
21
P
22
P
23
p
2
P
31
P
32
P
33
p
3
0 0 0 1
_
_
_
_
_
.
(6.10)
36
Primitive translation vectors a, b and c are transformed by P as
(a

, b

, c

) = (a, b, c)P
= (a, b, c)
_
_
_
P
11
P
12
P
13
P
21
P
22
P
23
P
31
P
32
P
33
_
_
_
= (P
11
a +P
21
b +P
31
c, P
12
a +P
22
b +P
32
c, P
13
a +P
23
b +P
33
c).
(6.11)
The shift of the origin, i.e., (0, 0, 0), is dened with a shift vector t as
t = (a, b, c)p
= (a, b, c)
_
_
_
p
1
p
2
p
3
_
_
_
= p
1
a +p
2
b +p
3
c.
(6.12)
Let A be the representation matrix of primitive translation vectors, a, b, and c, on the basis
of a set of orthonormal vectors, x, y, and z:
A = (a, b, c) =
_
_
_
a
x
b
x
c
x
a
y
b
y
c
y
a
z
b
z
c
z
_
_
_
. (6.13)
Then the transformation by Eq. (6.11) can be rewritten as
A

= AP. (6.14)
Alternatively, the use of the metric tensor, G, is often preferable to the direct use of A
owing to the unique denition of G. On the other hand, the number of representations for A
is innite because the orthonormal bases, x, y, and z, may be arbitrarily chosen. The metric
tensor (matrix) of the direct lattice is dened as
G = A
t
A =
_
_
_
a a a b a c
b a b b b c
c a c b c c
_
_
_
=
_
_
_
a
2
ab cos ac cos
ba cos b
2
bc cos
ca cos cb cos c
2
_
_
_
, (6.15)
where t denotes transposition. Then, G is transformed into G

by
G

= (A

)
t
A

= (AP)
t
(AP) = P
t
(A
t
A)P = P
t
GP. (6.16)
Similarly, the representation matrix of the reciprocal basis vectors a

, b

, and c

is dened as
A

= (a

, b

, c

) =
_
_
_
a

x
b

x
c

x
a

y
b

y
c

y
a

z
b

z
c

z
_
_
_
. (6.17)
The direct and reciprocal basis vectors are related by
A

= (A
1
)
t
= (A
t
)
1
(6.18)
and
(A

)
t
A = A(A

)
t
= A

A
t
= A
t
A

= E, (6.19)
37
where E is the identity matrix:
E =
_
_
_
1 0 0
0 1 0
0 0 1
_
_
_
. (6.20)
The metric tensor of the reciprocal lattice is dened as
G

= (A

)
t
A

=
_
_
_
a

_
_
_
=
_
_
_
a
2
a

cos

cos

cos

b
2
b

cos

cos

cos

c
2
_
_
_
.
(6.21)
The transformation of G

into G

is represented by
G

= (A

)
t
A

= (AP)
1
((AP)
t
)
1
= P
1
((A

)
t
A

)(P
1
)
t
= QG

Q
t
, (6.22)
where Q = P
1
.
The coordinates x, y, and z in the direct space are transformed by the 4 4 transformation
matrix Q as
_
_
_
_
_
x

1
_
_
_
_
_
= Q
_
_
_
_
_
x
y
z
1
_
_
_
_
_
, (6.23)
where
Q =
_
P
1
P
1
p
o 1
_
=
_
Q q
o 1
_
=
_
_
_
_
_
Q
11
Q
12
Q
13
q
1
Q
21
Q
22
Q
23
q
2
Q
31
Q
32
Q
33
q
3
0 0 0 1
_
_
_
_
_
= P
1
.
(6.24)
In short, Eq. (6.23) can be expressed as:
X

= QX. (6.25)
The 4 4 symmetry operation matrix W transforms the coordinates x, y, and z, of point X to
a symmetrically equivalent point

X with the coordinates, x, y, and z, by
_
_
_
_
_
x
y
z
1
_
_
_
_
_
=
_
_
_
_
_
W
11
W
12
W
13
w
1
W
21
W
22
W
23
w
2
W
31
W
32
W
33
w
3
0 0 0 1
_
_
_
_
_
_
_
_
_
_
x
y
z
1
_
_
_
_
_
=
_
_
_
_
_
W
11
x +W
12
y +W
13
z +w
1
W
21
x +W
22
y +W
23
z +w
2
W
31
x +W
32
y +W
33
z +w
3
1
_
_
_
_
_
,
(6.26)
38
or in short notation,

X = WX. (6.27)
The matrix Q transforms both X and

X to X

and

X

, respectively:

= W

, (6.28)
QWX = W

QX. (6.29)
Then, the symmetry operation W is transformed with the relation
W

= QWP. (6.30)
Let Q
t
be the transposed matrix of Q dened in Eq. (6.24). Then, anisotropic atomic
displacement parameters, , are transformed by

=
_
_
_

11

12

13

12

22

23

13

23

33
_
_
_
=
_
_
_
Q
11
Q
12
Q
13
Q
21
Q
22
Q
23
Q
31
Q
32
Q
33
_
_
_
_
_
_

11

12

13

12

22

23

13

23

33
_
_
_
_
_
_
Q
11
Q
21
Q
31
Q
12
Q
22
Q
32
Q
13
Q
23
Q
33
_
_
_
= QQ
t
.
(6.31)
When the determinant of P, det(P), is negative, the crystal-coordinate system is transformed
from right-handed to left-handed, and vice versa. The volume, V

, of the new unit cell is obtained

from the unit-cell volume, V , by
V

= det(P)V. (6.32)
Therefore, the unit-cell volume changes on transformation of the lattice unless det(P) is 1.
When Update structure parameters in the Structure dialog box is checked, both lattice
parameters and fractional coordinates are automatically transformed into new ones to keep the
current geometry of the crystal structure. If this option is unchecked, only symmetry operations
are transformed to those of the new lattice setting whereas lattice parameters and fractional
coordinates in the Structure dialog box remain unchanged.
6.2.2 Create super- and sub-lattices
The transformation matrix is also used for (primitive lattice)(complex lattice) conversions
(Fig. 6.4) and for creating superstructures.
When a transformation matrix with det(P) > 1 is specied, a dialog box appears to ask you
whether or not you want to search for additional sites lying in the resultant unit cell (Fig. 6.5).
Clicking [Yes] in the dialog box leads to creation of a superstructure by examining the following
equation to nd additional sites lying in between (0, 0, 0) and (1, 1, 1):
_
_
_
_
_
x

1
_
_
_
_
_
= Q
_
_
_
_
_
x +n
x
y +n
y
z +n
z
1
_
_
_
_
_
. (n
x
, n
y
, n
z
= 0, 1, 2, . . . ) (6.33)
Note that VESTA adds not new equivalent positions but new sites when creating super-
structures. That is, the space group of the new structure should be one of isomorphic subgroups
for the original space group. Otherwise, reduce symmetry to P1 by clicking [Remove symmetry]
before creating a superstructure, and input a new space group after the superstructure has been
created. If det(P) < 1, the same position may result from two or more sites. In such a case,
the redundant sites can be removed by clicking [Remove duplicate atoms].
39
Figure 6.4: An example of lattice transformation between primi-
tive and face-centered-cubic lattices.
Figure 6.5: A dialog box asking whether additional sites
are searched or not.
6.2.3 Options and buttons
The [View general positions] button is used to open the Equivalent Positions dialog box to check
the general equivalent positions and symmetry operations of the transformed lattice (see 12.1).
Clicking [Initialize current matrix] resets the transformation matrix to the identity matrix. Cur-
rently, this operation cannot be canceled. When option Normalize the fractional coordinates is
checked, the transformed coordinates are normalized within the range of 0 to 1.
40
Chapter 7
CREATE BONDS AND
POLYHEDRA
The Bonds dialog box (Fig. 7.1) is used to search for not only bonds but also atoms connected
by them, and to create coordination polyhedra.
Figure 7.1: Bonds dialog.
7.1 Three Search Modes
One of the following three search modes is selected in the Search mode frame box:
Search A2 bonded to A1
A2 atoms bonded to A1 atoms inside the drawing boundary, which will hereafter be referred
to as the box, are searched on the basis of the user-specied minimum and maximum
interatomic distances: {Min. length} and {Max. length}. If Search beyond the boundary is
checked, A2 atoms outside the box are searched in addition to those within the box. When
Search beyond the boundary is unchecked, A2 atoms only inside the box are searched.
Search atoms bonded to A1
All the atoms bonded to A1 atoms lying within the box are searched on the basis of the
user-specied {Min. length} and {Max. length} regardless of species of the A2 atoms. If
41
Search beyond the boundary is checked, A2 atoms outside the box are searched in addition
to those within the box.
Search molecules
All the atoms in molecules that contain at least one atom within the box are recursively
searched on the basis of the user-specied {Min. length} and {Max. length}. Neither A1
nor A2 is specied in this option.
7.2 Operating Instructions
7.2.1 Search Bonds and Atoms
Searching for atoms bonded to specied ones
To build up coordination polyhedra, either Search A2 bonded to A1 or Search atoms bonded
to A1 modes should be selected. The central atoms of the coordination polyhedra must be
specied as A1. In the Search molecules mode, no coordination polyhedra are created because
VESTA does not have any information about central atoms of polyhedra. In the rst two bond
search modes, option Search beyond the boundary enables us to search for all the A2 atoms
bonded to A1 atoms so that no coordination polyhedra are truncated even if A2 atoms lie outside
of the box. A1 atoms lying outside of the box are never searched.
Atoms A1 and A2 may be specied by entering either symbols for elements or site names.
If option Search by label is unchecked (default), the two list boxes {A1} and {A2} list atoms
with symbols of elements. On the other hand, {A1} and {A2} list atoms with labels of sites if
option Search by label is checked.
The minimum and maximum bond lengths, {Min. length} and {Max. length}, are input in
the unit of

A. Though {Min. length} is usually set at zero, it may be positive when dealing with
a disordered structure, whose split-atom model gives seemingly very short bonds.
Searching for molecules
The Search molecules mode is similar to 406 instruction in ORTEP-III
1
[52], i.e., reiterative
convoluting sphere of enclosure add. This mode is generally used in combination with option
Search beyond the boundary to avoid some atoms of molecules being truncated (omitted)
even if they lie outside the box (Fig. 7.2). In this mode, neither A1 nor A2 is specied; only
{Min. length} and {Max. length} are input.
A maximum distance of 1.6

A is appropriate in typical organic compounds containing H,
C, N, O, and F atoms having covalent radii of 0.32, 0.77, 0.74, 0.66, and 0.72

A, respectively.
Inputting a larger value might lead to generation of unreal bonds. If larger atoms such as P,
S, Cl, and Br having covalent radii of 1.10, 1.04, 0.99, and 1.14

A, respectively, are involved,
A1A2 bonds must be added by selecting Search A2 bonded to A1.
Applications of the Search molecules mode to inorganic crystals
In general, the Search molecules mode should not be applied to inorganic compounds or metals
because a network of bonds continues innitely in their structures. Theoretically, an innite
number of atoms should be searched in such a case. Nevertheless, VESTA actually searches and
shows a huge number of atoms and bonds in a nite range because searching within a certain
area ensures that all the atoms contained in molecules in the box are searched in the cases of
periodic structures.
1
http://www.ornl.gov/sci/ortep/doc/i400.html
42
A B
Figure 7.2: Crystal structure of anthraquinone [53]. Bonds are searched in the Search mole-
cules mode (A) with option Search beyond the boundary unchecked and (B) with that option
checked. In both cases, the box have a range of coordinates from (0, 0, 0) to (1, 1, 1).
There is one exception that the Search molecules mode is well-suited to search for bonds
in inorganic crystals, that is, searching for OH bonds. Usually, there are no bonds closer than
OH distances; thus, using this mode with a small value of {Max. length} is safe. The maximum
bond length of ca. 1.1

A leads to a search for all the H atoms bonded to O atoms in the Graphics
Area. If OH bonds are searched in the other modes, H atoms bonded to some O atoms lying
outside of the box are omitted because no O atoms occupy centers of coordination polyhedra
(Fig. 7.3).
7.2.2 A list of bond specications
All the bond specications are listed below the Search bonds and atoms frame box (Fig. 7.1).
The states of the text boxes, list boxes, and options are updated in response to selection of an
item in the list of bond specications. The following four buttons are placed at the right side of
the list:
[Add]: Add a new bond specication.
[Modify]: Modify the selected bond specication.
[Delete]: Delete the selected bond specication.
[Clear All]: Delete all the bond specications in the list.
The [Modify] and [Delete] buttons cannot be clicked unless a bond specication in the list is
selected.
43
A B
Figure 7.3: Crystal structure of -AlOOH, a high-pressure modication of aluminum oxide
hydroxide, with thermal-displacement ellipsoids at a 99 % probability level [54]. The structures
were drawn in a coordinate range from (0, 0, 0) to (1, 1.5, 1). OH bonds were searched (A) in
the Search A2 bonded to A1 mode and (B) in the Search molecules mode. Option Search
beyond the boundary was enabled in either case. In (A), some H atoms were omitted when O
atoms lay outside of the box specied in the Boundary dialog box. On the hand, all the H atoms
bonded to O atoms displayed in the screen were searched in (B).
7.3 Hydrogen Bonds
To display XH Y hydrogen bonds, select the Search A2 bonded to A1 mode in the Search
mode frame box, set A1 at H and A2 at Y, input the maximum H Y distance in {Max. length},
click [Add], check Show hydrogen bonds, and click [OK]. Then, XH bonds are plotted as
cylinders or lines in response to the selected bond style, and H Y bonds as solid, dotted, or
dashed lines according to the selected hydrogen bond style. Unless Search hydrogen bonds is
checked, the H Y bonds are drawn in the same way as with the XH bonds.
44
Chapter 8
ADDITIONAL OBJECTS
8.1 Vectors on Atoms
The Vectors dialog box (Fig. 8.1) is used to attach vectors (arrows) to part of atoms. These
arrows serve to represent magnetic moments or directions of static and dynamic displacements
of atoms.
Figure 8.1: Vectors dialog box showing a list of atoms in SrFeO
2
[55], with the atom
list ltered by elements. Atoms selected in the dialog box are highlighted in the
underlying Graphics Area displaying the magnetic structure of SrFeO
2
.
45
8.1.1 A list of atoms
A list of atoms in the Graphics Area is displayed in the left pane of the dialog box. Selecting
atoms in this dialog box highlights corresponding objects in the Graphics Area, and vice versa.
Atoms in the list can be ltered by elements, sites, or states of atoms. Therefore, this dialog
box is also used to locate certain types of sites in the Graphics Area.
8.1.2 A list of vectors
At the right pane of the Vectors dialog box, all the vectors are listed with the following three
buttons placed above the list (Fig. 8.1):
[New]: Add a new vector.
[Edit]: Edit the selected vector.
[Delete]: Delete the selected vector.
The [Edit] and [Delete] buttons cannot be clicked unless a vector in the list is selected.
8.1.3 How to attach a vector to atoms
Click the [New] button to create a new vector or select a vector in the list of vectors and click
the [Edit] button to edit it. Then, a dialog box appears to edit properties of the selected vector
(Fig. 8.2). To attach a vector to atoms, select atoms in the left pane of the Vectors dialog
box or in the Graphics Area. Then, select a vector from the list of vectors in the right pane of
the Vectors dialog box and click the [<< Set] button. In a similar manner, select atoms in the
left pane and click the [>> Remove] button to detach a vector from the atoms. When option
Preview is checked (default), changes in the dialog box are reected on the Graphics Area in
real time.
Figure 8.2: A dialog box to create or edit specications
of a vector.
8.2 Lattice Planes
The Lattice Planes dialog box (Fig. 8.3) is used to add lattice planes in a structural model, and
to add 2D slices of volumetric data in a 3D image. When option Preview is checked (default),
changes in the dialog box are reected on the Graphics Area in real time.
8.2.1 Appearance of lattice planes
When volumetric data are included in the current data, lattice planes are colored according
to volumetric data on the lattice planes. Saturation levels of colors are specied at Sections
page in the Properties dialog box (Fig. 8.4; see 10.6). When only structural information is
included in the current data, lattice planes are drawn with colors specied in this dialog box.
46
Figure 8.3: Lattice Plane dialog box.
Figure 8.4: A section of a dierence Fourier map inserted in a ball-and-stick
model of AlOOH [54]. The maximum in the section corresponds to a position
of an H atom, which is not included in the structural model.
If you desire to draw lattice planes with specied colors for data having both structural and
volumetric information, volumetric data should be deleted at the Import Data dialog box (see
5.2).
Material settings of lattice planes, i.e. specular color, shininess, and opacity are input in the
Material frame box. Drawing of edges for lattice planes can be controlled in the Edge frame box.
These settings are common to all the lattice planes.
47
8.2.2 Add lattice planes frame box
Input Miller indices hkl and the distance from the origin, (0, 0, 0), to the lattice plane. The
distance may be specied in the unit of either its lattice-plane spacing, d, or

A. The color of the
lattice plane is specied as three integers ranging from 0 to 255, or from a color selection dialog
box, which is opened after clicking the [Select...] button.
When a lattice plane is drawn with a color specied in this dialog box, the lattice plane
rendered in the Graphics Area is selectable and movable by mouse dragging (see 9.5).
8.2.3 A list of lattice planes
Miller indices and distances of lattice planes are listed below the Add lattice planes frame box
(Fig. 8.3). Selecting a lattice plane in the list will update values in the Add lattice planes frame
box. The following four buttons are placed at the right side of the list:
[Add]: Add a new lattice plane.
[Modify]: Modify the selected lattice plane.
[Delete]: Delete the selected lattice plane.
[Clear All]: Delete all the lattice planes in the list.
The [Modify] and [Delete] buttons cannot be clicked unless a lattice plane in the list is selected.
48
Chapter 9
INTERACTIVE MANIPULATIONS
9.1 Drawing Boundary
The drawing boundary (box) is specied in the Boundary dialog box (Fig. 9.1).
Figure 9.1: Boundary dialog.
Changing the boundary regenerates all the atoms in the Graphics Area, and will reset all
the properties of objects to the default values. Selected or hidden states of atoms, bonds, and
polyhedra are reset to the default states, and vectors on atoms are removed.
9.1.1 Ranges of fractional coordinates
Drawing boundaries are fundamentally specied by inputting ranges of fractional coordinates
along x, y, and z axes.
9.1.2 Cuto planes
In addition to ranges of fractional coordinates, we can further specify cuto planes, as exem-
plied in Fig. 9.2. After atoms, bonds, and isosurfaces within the x, y, and z ranges have been
generated, those lying outside of the cuto planes are excluded. Even though atoms, bonds, and
polyhedra can be hidden by another procedure described in 9.6, this is the only way to remove
part of isosurfaces and sections. This feature is, therefore, particularly useful for visualizing 2D
distribution of volumetric data on lattice planes in addition to isosurfaces. Each cuto plane
49
Figure 9.2: Electron density distribution determined for D-sorbitol
[56] by the maximum-entropy method from synchrotron X-ray pow-
der diraction data. Drawn in ranges from (0, 0, 0) to (1, 1, 1) with a
cuto plane (110).
is specied as Miller indices hkl and a distance from the origin. The distance of a cuto plane
from the origin may be specied in the unit of either its lattice-plane spacing, d, or

A.
9.1.3 A list of cuto planes
Below the Cuto planes frame box, all the cuto planes are listed with the following four buttons
placed at the right side (Fig. 9.1):
[Add]: Add a new cuto plane
[Modify]: Modify the selected cuto plane
[Delete]: Delete the selected cuto plane
[Clear All]: Delete all the cuto planes in the list
The [Modify] and [Delete] buttons cannot be clicked unless a cuto plane in the list is selected.
9.2 Dene a View Direction
The Orientation dialog box (Fig. 9.3) let you specify a direction of viewing objects.
9.2.1 Manner of specifying directions
Either a lattice vector, [uvw], or a reciprocal-lattice vector, [hkl]

, perpendicular to a lattice
plane (hkl) is specied as the direction of projection.
50
Figure 9.3: Orientation dialog box.
Project along [uvw]
The direction of projection is a lattice vector ua +vb +wc.
Project along the normal to (hkl)
The direction of projection is a reciprocal-lattice vector ha

+kb

+lc

.
9.2.2 Current orientation
A 3 3 rotation matrix of the current orientation is displayed.
9.2.3 View direction
Two directions, i.e., the direction of projection (direction from the viewpoint to the screen) and
the upward direction on the screen are specied by a set of a lattice vector and a reciprocal-lattice
vector. The two vectors are perpendicular to each other. When the lattice vector, ua +vb +wc,
lies on the (hkl) plane, u, v, w, h, k, and l must satisfy the condition:
hu +kv +lw = 0. (9.1)
This condition must, therefore, be satised in order to specify the upward direction on the
screen, otherwise the upward direction on the screen is automatically determined by VESTA.
Projection vector: Direction of projection.
In the Project along [uvw] mode, this vector is u, v, and w in ua +vb +wc.
In the Project along the normal to (hkl) plane mode, this vector is h, k, and l in ha

+
kb

+lc

.
Upward vector: Upward direction on the screen.
In the Project along [uvw] mode, this vector is h, k, and l in ha

+kb

+lc

.
In the Project along the normal to (hkl) plane mode, this vector is u, v, and w in ua +
vb +wc.
9.3 Rotate
9.3.1 Drag mode
In the Drag mode, drag the mouse while pressing the left mouse button to rotate objects
in the Graphics Area. The objects are rotated while dragging with the mouse. They are never
51
rotated after releasing the mouse button. In the [Free rotation] mode, the rotation axis becomes
normal to the direction along which the mouse is moved. To restrict the rotation axis, select
[Around X axis], [Around Y axis], or [Around Z axis] in the pull-down menu of the Rotation
modes frame box.
9.3.2 Animation mode
Three types of animation modes can be used in VESTA: Click, Push, and Random modes.
The step width of rotation (in degrees/frame) and intervals between frames (in ms) are specied
in the Preferences dialog box (see chapter 14).
Click mode
In the Click mode, click the left mouse button to rotate the objects. In the Click plus
[Free] rotation modes, the rotation axis is perpendicular to a line connecting the clicked position
and the central point. To restrict the rotation axis, select [Around X axis], [Around Y axis],
or [Around Z axis] in the pull down menu of the Rotation modes frame box. To stop rotating
the objects, select either the Drag or Push mode. The animation speed (intervals between
frames in ms) is specied in the Preferences dialog box.
Push mode
In the Push mode, press the left mouse button at point 1 and drag the mouse to point 2. In
the [Free rotation] mode, objects are rotated around an axis perpendicular to a line connecting
points 1 and 2. The rotation speed is proportional to the speed of moving the mouse. The
objects stop rotating immediately after releasing the left mouse button. To restrict the rotation
axis, select [Around X axis], [Around Y axis], or [Around Z axis] in the pull down menu of the
Rotation modes frame box.
Random mode
In the Random mode, the rotation axis is automatically set and it changes dynamically. To
stop the rotation of the objects, select either the Drag or Push mode.
9.4 Magnify
Objects are magnied in proportion to the distance of dragging the mouse upward. On the other
hand, they are shrunk in inverse proportion to the distance of dragging the mouse downward.
9.5 Translate
Drag the mouse in the Graphics Area to translate objects. When a lattice plane with a color
specied in the Lattice Planes dialog box is selected and dragged in this mode, it is interactively
moved. Otherwise the entire objects are translated with the mouse along the same direction.
9.6 Select
In this mode, left click on an object to select it. To select all objects in a certain area, press
the left mouse button and drag area to be selected. When new objects are selected, previously
selected objects are reset to the normal state. To select additional objects while keeping pre-
viously selected objects, press <Shift> while clicking or dragging the Graphics Area. Objects
52
other than atoms, bonds, and polyhedra cannot be selected. Accordingly, you can select, for
example, atoms behind isosurfaces. Objects may also be selected by clicking an object or select-
ing a certain area by dragging a mouse while pressing the right button regardless of the current
manipulation mode.
When a single object is selected, a variety of information about it is output in the Text
Area. The estimated standard deviation is also displayed for interatomic distance, bond angle,
and dihedral angle if those of lattice parameters and fractional coordinates have been supplied.
However, note that VESTA gives only rough estimates of standard deviations by neglecting
o-diagonal elements of the variance-covariance matrix because no o-diagonal elements are
included in crystal-data les.
Press the <Delete> key to hide selected objects. The hidden objects are not actually deleted
but just made invisible. To restore all the hidden objects, press the <Esc> key. By hiding part
of coordination polyhedra, you can easily mix polyhedra with a ball-and-stick model.
9.6.1 Atom
On selection of an atom, site number, site name, symbol of the element, fractional coordinates
(x, y, z), occupancy, site multiplicity plus Wycko symbol, site symmetry, translation vector,
and symmetry operations (coordinate triplet) are displayed for the atom. For example, selection
of Al and O atoms connected with each other in a ball-and-stick model of -Al
2
O
3
gives the
following lines:
Atom: 1 Al Al 1.00000 0.00000 0.15000 g=1.000 12c 3. ( 1, 0, 0)+ y, x, -z+1/2
Atom: 2 O O 1.00000 0.31000 0.25000 g=1.000 18e .2 ( 1, 0, 0)+ -y, x-y, z
9.6.2 Bond
When a bond is selected, its length is displayed with its estimated standard deviation, if any,
in the Text Area as well as the Status Bar. Site number, site name, symbol of the element,
fractional coordinates (x, y, z), translation vector, and symmetry operations are also displayed
for two atoms connected with the bond. For example, on selection of an AlO bond in the
ball-and-stick model of -Al
2
O
3
, the following three lines are output:
Bond: l(Al-O) = 1.96249(0) Angstrom
1 Al Al 1.00000 0.00000 0.15000 ( 1, 0, 0)+ y, x, -z+1/2
2 O O 1.00000 0.31000 0.25000 ( 1, 0, 0)+ -y, x-y, z
9.6.3 Coordination polyhedron
When a coordination polyhedron with a coordination number of n is selected, information on
the central atom and coordinating atoms (ligands), distances between the central atom and
coordinating atoms are listed in the Text Area. Further, geometric information described below is
displayed in the Text Area; the polyhedral volume, D, <>, and
2
are useful when investigating
structural changes under high/low temperature and high pressure [57].
Polyhedral volume
The volume of the selected coordination polyhedron is calculated as a fundamentally important
physical quantity [21]. For example, distortion in perovskite-type compounds, ABO
3
can be
quantied with polyhedral volume ratios, V
A
/V
B
, rather than with tilting angles [58].
53
Distortion index
A distortion index, D, based on bond lengths was dened by Baur [22] as
D =
1
n
n

i=1
|l
i
l
av
|
l
av
, (9.2)
where l
i
is the distance from the central atom to the ith coordinating atom, and l
av
is the average
bond length.
Quadratic elongation
The quadratic elongation, <> [23], is dened only for tetrahedra, octahedra, cubes, dodecahe-
dra, and icosahedra:
<> =
1
n
n

i=1
_
l
i
l
0
_
2
, (9.3)
where l
0
is the center-to-vertex distance of a regular polyhedron of the same volume. <> is
dimensionless, giving a quantitative measure of polyhedral disortion which is independent of the
eective size of the polyhedron.
Bond angle variance
The bond angle variance,
2
[23], is calculated only for tetrahedra, octahedra, cubes, dodecahe-
dra, and icosahedra:

2
=
1
m 1
m

i=1
(
i

0
)
2
, (9.4)
where m is (number of faces in the polyhedron)3/2 (i.e., number of bond angles),
i
is the
ith bond angle, and
0
is the ideal bond angle for a regular polyhedron (for example, 90

for an
octahedron or 109

28

for a tetrahedron).
Eective coordination number
The eective coordination number (ECoN) [24, 25, 26] is dened as
ECoN =

i
w
i
, (9.5)
where the quantity
w
i
= exp
_
1
_
l
i
l
av
_
6
_
(9.6)
is called the bond weight of the ith bond. In Eq. (9.6), l
av
represents a weighted average bond
length dened as
l
av
=

i
l
i
exp
_
1 (l
i
/l
min
)
6
_

i
exp
_
1 (l
i
/l
min
)
6
_ , (9.7)
where l
min
is the smallest bond length in the coordination polyhedron.
54
Charge distribution
Let q
X
be the formal charge (oxidation number) of the central atom, X. Then, the fraction of
the charge received by an ion at a corner of a coordination polyhedron is calculated as
q
i
=
w
i
q
X
ECoN
X
. (9.8)
The total charge, Q
A
, received by an ion A is obtained by summing the relevant charge fractions,
q
i
s, over its n bonds. Similarly, the charge, Q
X
, received by the ion at the center of a
coordination polyhedron is calculated by
Q
X
=
_

i
w
i
(q
A
/Q
A
)
i
ECoN
X
_
q
X
. (9.9)
Calculation of the charge distribution [25, 26, 27] in the crystal structure depends on the current
bond specications. In other words, the fraction of the charge is not given to or received by
ions that are not bonded to each other. Formal charges of ions are read in from data les when
ICSD les or CIFs (see 15.4.1) including oxidation states (oxidation numbers) are opened. For
other le formats, formal charges of ions have to be input in the Structure dialog box to calculate
the charge distribution. Note that charge distribution in nonstoichiometric compounds can be
calculated because occupancies are used in the calculation.
Bond valence sum
In addition to the above physical quantities, the bond valence sum, V , dened as
V =
n

i=1
exp
_
l
i
l
0
b
_
(9.10)
is also obtainable from the bond valence parameter, l
0
, of the central atom [28, 29, 30]. In
VESTA, the empirical constant, b, in Eq. (9.10) is xed at a typical value of 0.37

A.
The bond valence model, which is a development of Paulings rules, has been theoretically
described in terms of classical electrostatic theory without resorting to quantum mechanics.
Nevertheless, V serves, in practice, to estimate the oxidation state of the central atom only from
bond lengths determined by X-ray or neutron diraction.
A CIF, bvparm2009.cif, storing bond valence parameters of most chemical species is available
for download from
http://www.ccp14.ac.uk/solution/bond valence/
This CIF is also included in the folder where the executable binary le of VESTA is located so
as to refer to it when l
0
values are required. In bvparm2009.cif, an l
0
value for a pair of a cation
and an anion is obtained, e.g., l
0
= 1.620

A for Al
3+
and O
2
and l
0
= 2.172

A for La
3+
and
O
2
(b = 0.37

A). Never select an l
0
value for which b = 0.37

A.
If a coordination polyhedron is clicked in the Graphic Area while pressing the <Ctrl> key in
the manipulation mode of Select, you are asked to input a bond valence parameter, l
0
, for the
central metal of the selected coordination polyhedron. After entering l
0
, the bond valence sum
[28, 29, 30] for the central atom is calculated and displayed in the Text Area from all the bond
lengths, l
i
, for the current polyhedron and l
0
.
Expected bond length
Unless oxidation states of sites have been input in the Structure dialog box or from some kinds
of structural les, VESTA prompts you to enter an oxidation number corresponding to V in Eq.
55
(9.10); pressing the <Enter> key skips the subsequent calculation. After entering the value, an
bond length expected with Eq. (9.10) is calculated and given in the Text Area.
An example of getting information on a coordination polyhedron
Suppose that a TiO
6
octahedron is selected in a structural model of perovskite (CaTiO
3
) [59]
while pressing the <Ctrl> key. You are asked to enter the bond valence parameter of Ti
4+
(l
0
= 1.815

A) and the oxidation state of Ti (= +4) unless it has already been input in the
Structure dialog box. Then, the following data including the physical quantities described above
are output to the Text Area:
POLYHEDRON:
1 Ti1 Ti 1.00000 0.50000 0.00000 ( 1, 0, 0)+ x, y, z
----------------------------------------------------------------------------
4 O2 O 0.71030 0.71120 -0.03730 ( 1, 1, 0)+ -x, -y, -z
4 O2 O 0.78970 0.21120 -0.03730 ( 0, 0, 0)+ x+1/2, -y+1/2, -z
3 O1 O 0.92890 0.48390 0.25000 ( 1, 0, 0)+ -x+1/2, y+1/2, z
3 O1 O 1.07110 0.51610 -0.25000 ( 0, 0, 0)+ x+1/2, -y+1/2, -z
4 O2 O 1.28970 0.28880 0.03730 ( 1, 0, 0)+ x, y, z
4 O2 O 1.21030 0.78880 0.03730 ( 1, 0, 0)+ -x+1/2, y+1/2, z
----------------------------------------------------------------------------
l(Ti1-O2) = 1.9571(11) Angstrom
l(Ti1-O2) = 1.9569(11) Angstrom
l(Ti1-O1) = 1.9497(4) Angstrom
l(Ti1-O1) = 1.9497(4) Angstrom
l(Ti1-O2) = 1.9571(11) Angstrom
l(Ti1-O2) = 1.9569(11) Angstrom
---------------------------------------
Average bond length = 1.9546 Angstrom
Polyhedral volume = 9.9546 Angstrom^3
Distortion index (bond length) = 0.00168
Quadratic elongation = 1.0001
Bond angle variance = 0.3773 deg.^2
Effective coordination number = 5.9994
Charge distribution
----------------------------------------------------------------------------
delta_q: Fraction of the charge received by the ion
Q: Total charge received by the ion
q: Formal charge (oxidation number)
----------------------------------------------------------------------------
x y z delta_q Q q
4 O2 O 0.71030 0.71120 -0.03730 0.661 -1.994 -2.000
4 O2 O 0.78970 0.21120 -0.03730 0.662 -1.994 -2.000
3 O1 O 0.92890 0.48390 0.25000 0.677 -2.013 -2.000
3 O1 O 1.07110 0.51610 -0.25000 0.677 -2.013 -2.000
4 O2 O 1.28970 0.28880 0.03730 0.661 -1.994 -2.000
4 O2 O 1.21030 0.78880 0.03730 0.662 -1.994 -2.000
----------------------------------------------------------------------------
56
1 Ti1 Ti 1.00000 0.50000 0.00000 4.000 4.000
Input a bond valence parameter: 1.815000
Bond valence sum = 4.115
Oxidation state of the cation: +4
Expected bond length = 1.965 Angstrom
All the fractional coordinates, translations, equivalent positions, bond lengths with their esti-
mated standard deviations for the TiO
6
octahedron are output before the polyhedral volume.
9.7 Distance
In the Distance mode, only atoms can be selected. Select a pair of atoms A and B to obtain
the interatomic distance between A and B. Then, the two atoms are highlighted and a dashed
line connecting them is plotted on the screen. The AB distance is displayed on the Status
Bar with its estimated standard deviation [60], if any, enclosed in a pair of parentheses. The
estimated standard deviation is calculated only when those of lattice parameters and fractional
coordinates have been input.
To obtain a bond length in a ball-and-stick model, selecting the relevant bond (see 9.6.2) is
faster than clicking two atoms. More information on the interatomic distance is displayed in the
Text Area, where site number, site name, symbol of the element, fractional coordinates (x, y, z),
symmetry operations, and translation vector are displayed for each of atoms A and B.
For example, an interatomic distance for A = C7 and B = O2 in 2

-Hydroxyl-4

-dimethylami-
nochalcone [61] is output as follows:
l(C7-O2) = 1.254(4) Angstrom
15 C7 C 0.75640 0.11620 0.34510 ( 1,-1, 0)+ -x, y+1/2, -z+1/2
4 O2 O 0.80530 0.04440 0.43470 ( 1,-1, 0)+ -x, y+1/2, -z+1/2
9.8 Angle
In this mode, only atoms can be selected. For example, you can select an atom inside a coordi-
nation polyhedron.
9.8.1 Bond angle
Select three atoms A, B, and C to calculate the bond angle (in degrees) with atom B at the
apex. The angle is displayed on the Status Bar with its estimated standard deviation, if any, in
a pair of parentheses. The estimated standard deviation [60] is calculated only when those of
lattice parameters and fractional coordinates have been input.
More information on the bond angle is displayed in the Text Area, where site number, site
name, symbol of the element, fractional coordinates (x, y, z), translation vector, and symmetry
operations are displayed for each of atoms A, B, and C.
For example, a bond angle for A = O1, B = Al1, and C = O1 in Al
2
O
3
[62] is output as
follows:
phi(O1-Al1-O1) = 86.37(3) deg.
1 O1 O 1.33333 -0.02698 0.41667 ( 1, 0, 0)+ y+1/3, -x+y+2/3, -z+2/3
2 Al1 Al 1.00000 0.00000 0.35217 ( 1, 0, 0)+ x, y, z
1 O1 O 1.00000 -0.30635 0.25000 ( 1,-1, 0)+ -y, x-y, z
57
9.8.2 Dihedral angle
For a sequence of four atoms A, B, C, and D, the dihedral angle, , is dened as the positive
angle between ABC and BCD planes. Let = (BCD), = (BAD

), and =
(ABC), where D

denotes the D atom when CD is translated in such a way that the C

atom overlaps with the A atom. Then, cos is formulated as [63]
cos =
cos cos cos
sin sin
. (9.11)
Select four atoms, A, B, C, and D while pressing the <Shift> key. The dihedral angle (in
degrees) for atom D and a plane on which atoms A, B, and C lie is displayed on the Status Bar
with its estimated standard deviation, if any, in a pair of parentheses. The estimated standard
deviation [60] is calculated only when those of lattice parameters and fractional coordinates have
been input.
More information about the dihedral angle is output in the Text Area, where site number,
site name, symbol of the element, fractional coordinates (x, y, z), symmetry operations, and
translation vector are displayed for each of atoms A, B, C, and D.
omega(C1-C2-C3-C4) = 2.36(8) deg.
5 C1 C 0.58190 0.79990 0.21900 ( 0, 0, 0)+ x, y, z
7 C2 C 0.52280 0.72240 0.11990 ( 0, 0, 0)+ x, y, z
9 C3 C 0.41240 0.66560 0.09830 ( 0, 0, 0)+ x, y, z
10 C4 C 0.36010 0.68130 0.17850 ( 0, 0, 0)+ x, y, z
For examples, in the case of C1 (= A), C2 (= B), C3 (= C), and C4 (= D) atoms contained in
an aromatic ring of 3-[4-(dimethylamino)phenyl]-1-(2-hydroxyphenyl)prop-2-en-1-one [61] (Fig.
9.4), the following ve lines are output in the text area: Data in lines No. 25 (selected lines
C1
C2
C3
C4
Figure 9.4: Calculation of a dihedral angle for four carbon atoms in an aromatic
ring in 3-[4-(dimethylamino)phenyl]-1-(2-hydroxyphenyl)-prop-2-en-1-one
58
in Fig. 9.4) can be used when imposing nonlinear restraints on the dihedral angle in Rietveld
r0nement with RIETAN-FP [17].
59
Chapter 10
PROPERTIES OF OBJECTS
Properties of various objects are edited in the Properties dialog box. When Preview is checked,
changes in the Properties dialog box are reected on the Graphics Area in real time. Click [OK]
to apply all the changes or [cancel] to discard the changes. Clicking the [Save as default] button
saves current properties as default values in VESTA.
10.1 General
The rst page in the Properties dialog box is General (Fig. 10.1).
Figure 10.1: General page in the Properties dialog box.
10.1.1 Unit cell
The display of Unit cell edges in the Graphics Area is controled with the following three radio
buttons:
Do not show: Do not show unit cell edges.
Single unit cell: Show edges of a single unit cell.
All unit cells: Show all the edges of unit cells within the drawing boundary.
60
Line styles of the unit cell edges are selected from three styles: Solid lines, Dotted lines,
and Dashed lines. The width of lines is input in text box {Line width}. The color of lines is
specied either (A) by entering R, G, and B values ranging from 0 to 255 or (B) by picking a
color from a color selection dialog box opened after clicking the [Select] button.
10.1.2 Axes
Show Compass is used to turn on or o a display of three arrows indicating a, b, and c axes
(or x, y, and z axes in the case of Cartesian coordinates). Show Axis Labels is used to turn
on or o a display of axis labels, a, b, and c (or x, y, and z in the case of Cartesian
coordinates).
10.2 Atoms
The second page in the Properties dialog box is Atoms (Fig. 10.2).
Figure 10.2: Atoms page in the Properties dialog box.
10.2.1 Material
{Specular} is a color of reected light that bounces sharply in a particular direction in the
manner of a mirror. A highly specular light tends to cause a bright spot on the surface it
shines upon, which is called the specular highlight.
{Shininess} is a property, which species how small and focused the specular highlight. A
value of 0 species an unfocused specular highlight.
10.2.2 Resolution
{Stacks} and {Slices} are parameters common to all the atoms, allowing you to change the reso-
lution (quality) of atoms displayed on the screen. {Stacks} denotes the numbers of subdivisions
along the Z axis (similar to lines of latitude) while {Slices} denotes the number of subdivisions
around the Z axis (similar to lines of longitude). {Stacks} and {Slices} should be equal to each
61
other in ball-and-stick and space-lling models where atoms are represented by perfect spheres.
In general, decreasing {Stacks} and {Slices} accelerates the rendering of objects in the Graphics
Area.
10.2.3 Atom style
Select either of the following two radio buttons for the mode of displaying atoms on the Graphics
Area: (1) Show as balls or (2) Show as thermal ellipsoids.
Show as balls: Atoms are rendered as spheres. List box {Radii type} is used to select a
type of default atomic radii from the following three:
Atomic: Metallic or covalent radii, whose values were mostly taken from Refs. [64,
65, 66].
Ionic: Eective ionic radii compiled by Shannon [67] for representative oxidation
states and coordination numbers.
van der Waals: van der Waals radii [68].
The user may modify default values of atomic, ionic, and van der Waals radii by editing a
text le, elements.ini, in the program folder of VESTA.
Show as thermal ellipsoids: Atoms are rendered as ellipsoids to represent anisotropic
thermal motion of atoms whose shapes are calculated from anisotropic atomic displacement
parameters,
ij
or U
ij
(see 6.1.4). The probability (in percentage) for atomic nuclei to be
included in the ellipsoids is input in text box {Probability}. It is common to all the
atoms when drawing thermal ellipsoids. If option Show principal ellipses is checked, three
principal ellipses corresponding to three principal planes are plotted on the surface of each
ellipsoid (Fig. 10.3). The line width of the principal ellipses is input in text box {Line
width}.
A B
Figure 10.3: Thermal ellipsoids of an atom (A) with principal
ellipses and (B) without them.
Hide non-bonding atoms: This option hides atoms that are connected by no bonds but
the crystallographic site for those atoms have a coordination number larger than 0. In
other words, atoms are made invisible if all the coordinated atoms are lying outside of the
drawing boundary.
10.2.4 Radius and color
Select a symbol of an element from list box {Symbol} and then specify its {Radius}. The color
of an atom (element) is specied either (a) by entering R, G, and B values ranging from 0 to
62
255 or (b) by picking a color from a color selection dialog box opened when clicking the [Select]
button.
10.3 Bonds
The third page in the Properties dialog box is Bonds (Fig. 10.4).
Figure 10.4: Bonds page in the Properties dialog box.
10.3.1 Material
It is the same as that in Atoms (see 10.2).
10.3.2 Resolution
It is the same as that in Atoms (see 10.2).
10.3.3 Bond style
One of the following four types of bonds are specied in this frame box:
Unicolor cylinder: Each bond is drawn as a cylinder, whose color can be changed in the
Color panel.
Bicolor cylinder: Each bond is drawn as a cylinder with colors of two atoms connected
with each other. This style is the default setting for les *.cc1, *.cc2, *.mol, *.pdb, and
*.mld (see 15.4.1). The two colors of the bond are just the same as those of atoms connected
by bonds.
Color line: Each bond is drawn as a straight line, whose color is changed in the Color
panel.
Gradient line: Each bond is drawn as a line connecting two atoms with gradient distri-
bution of colors. Colors of each bond at the ends of the line are the same as those of the
atoms.
63
In addition, each H X bond in the XH X hydrogen bond can be plotted with one of the
following three line types in the Hydrogen bond style panel:
Solid line
Dotted line
Dashed line
10.3.4 Radius and color
The radius of each cylindrical bond in the stick model is changed in text box {Radius (cylinder)}.
The actual radius of cylinders is 40 % of the input value. The line width of line-style bonds is
changed in text box, {Width (line)}. The color of bonds in the single color style is changed in
the Color tool below these two text boxes.
10.4 Polyhedra
The fourth page in the Properties dialog box is Polydedra (Fig. 10.5).
Figure 10.5: Polyhedra page in the Properties dialog box.
10.4.1 Material
{Specular} and {Shininess} are just the same as those in Atoms page (see 10.2). The {Opacity}
of coordination polyhedra is input with 100 % corresponding to opaque planes (the inside of
each coordination polyhedron is invisible) and 0 % corresponding to fully transparent planes
(coordination polyhedra are invisible).
10.4.2 Polyhedral style
Styles of coordination polyhedra are selected from the following six types of expressions that are
schematically illustrated in the Polyhedral style frame box:
64
Show atoms and bonds without any coordination polyhedra.
Show atoms, bonds, and coordination polyhedra (default).
Show atoms and coordination polyhedra.
Show central atoms and coordination polyhedra.
Show only coordination polyhedra.
Show only bonds.
10.4.3 Planes
The surface color of each coordination polyhedron is specied here. Select an element from list
box {Central atom} to specify its color. Default colors of coordination polyhedra are the same
as those of the central atoms.
10.4.4 Edges
The visibility, line width, and color of edges of coordination polyhedra are input in the Edges
frame box.
10.5 Isosurfaces
The fth page in the Properties dialog box is Isosurface (Fig. 10.6).
Figure 10.6: Isosurfaces page in the Properties dialog box.
10.5.1 Isosurfaces
At the top of the Isosurfaces frame box, the minimum and maximum data values are displayed in
the unit of the raw data. Data values such as electron and nuclear densities, and wave functions
on isosurfaces are equal to {Isosurface level}, d(iso). All the points with densities larger than
d(iso) lie inside the isosurfaces whereas those with densities smaller than d(iso) are situated
outside the isosurfaces.
65
A B
Figure 10.7: Comparison between the two modes of rendering isosurfaces for a molybdenum-
cadmium cluster [Cd{S
4
Mo
3
(Hnta)
3
}
2
]
4
. (A) Render from behind (default) and (B) Render
from front to back. Composite images of electron-density isosurfaces colored according to
electrostatic potentials and a ball-and-stick model are shown with an isosurface level of 0.03a
3
0
and electrostatic potentials on the isosurface ranging from 0.814 Ry (blue) to 0.174 Ry (red).
Order of rendering polygons
The sequence of drawing polygons can be changed with option Render from front to back. By
default, polygons are rendered from behind so that isosurfaces behind the nearest surface are
visible through translucent isosurfaces. In some cases with complex isosurfaces, it may become
dicult to understand the shape of isosurfaces because they heavily overlap each other. When
option Render from front to back is checked, only the nearest surfaces are rendered. The use
of this option may sometimes improve the visibility of complex isosurfaces because neither back
surfaces nor internal ones are drawn.
Figure 10.7 illustrates a dierence between the two rendering modes in 3D visualization of
results of an electronic-state calculation for the [Cd{S
4
Mo
3
(Hnta)
3
}
2
]
4
ion (H
3
nta: nitrilotri-
acetic acid) [2] by a discrete variational X method with DVSCAT [45].
Kinds of isosurfaces
For volumetric data having both positive and negative values, you can select which surfaces are
visible:
{Positive and negative},
{Positive},
{Negative}.
Figure 10.8 exemplies isosurfaces of wave functions calculated with DVSCAT [45] for a com-
plex ion with a ball-and-stick model superimposed on the isosurfaces. Yellow and blue surfaces
show positive and negative values, respectively.
Resolution
The spacial resolution of volumetric data can be increased using the algorithm of cubic spline
interpolation. The interpolation level is specied in the {Interpolation} text box.
66
Figure 10.8: The 64a
1g
orbital for the {Cd[S
4
Mo
3
(Hnta)
3
]
2
}
4
ion
[2] with a ball-and-stick model. The isosurface levels of the wave
function were set at 0.01a
3/2
0
(yellow) and 0.01a
3/2
0
(blue),
where a
0
is the Bohr radius.
Opacity of isosurfaces
The opacity of isosurfaces is specied by two parameters, {Opacity 1} (O
1
) and {Opacity 2}
(O
2
), as exemplied in Fig. 10.9. O
1
is the opacity for polygons parallel to the screen, and O
2
is that for polygons perpendicular to the screen (Fig. 10.10). The opacity, O(p), for polygon p
with a normal vector of (x, y, z) is calculated with a linear combination of O
1
and O
2
:
A B
Figure 10.9: Composite images of electron-density isosurfaces and a ball-and-stick model
drawn for albatrossene (C
114
H
76
) [69] with two pairs of opacity parameters: (A) O
1
= 60 %
and O
2
= 100 %; (B) O
1
= 10 % and O
2
= 70 %. The electron densities were calculated with
DVSCAT [45] and visualized with an equi-density level of 0.01a
3
0
.
67
O
1
O
2
Screen
Z
Isosurface
Figure 10.10: Schematic representation of relations between orientation
of an isosurface and the two opacity parameters, O
1
and O
2
.
O
p
= O
1
z +O
2
(1 z). (10.1)
10.5.2 Surface coloring
After volumetric data for surface coloring have been loaded, isosurfaces can be colored according
to those data. The saturation level of colors is specied as (a) a value normalized between 0
and 100, and (b) a value corresponding to raw data. This is mostly equal to color settings for
lattice planes and sections of isosurfaces (see 10.6), except that the normalized values of 0 and
100 corresponds to the minimum and maximum data values on the current isosurfaces, not the
data set itself.
The color index, T, to determine the color of a data point with a value of d is calculated from
the unit of the raw data: the minumum saturation level, S
min
, and the maximum one, S
max
, in
the unit of the raw data:
T =
d S
min
S
max
S
min
. (10.2)
Data points with values larger than S
max
and smaller than S
min
are given the same colors as
those assigned to S
min
and S
max
, respectively.
10.6 Sections
The seventh (nal) page in the Properties dialog box is Sections (Fig. 10.11).
For volumetric data, both lattice planes and sections of isosurfaces are colored according
to numerical values on them (Fig. 10.12). The saturation level of colors is specied as (a) a
value normalized between 0 and 100 corresponding the minimum and maximum data values,
respectively, and (b) a value corresponding to raw data. The color index, T, for a data point
68
Figure 10.11: Sections page in the Properties dialog box.
O
K
D
D
D
Figure 10.12: Distribution of nuclear densities obtained for KOD
at 580 K by MEM from single-crystal neutron diraction data [3].
Coordinate ranges from (0, 0, 0) to (1, 1, 1) were drawn with a cuto
plane of (111). D atoms are highly disordered around O atoms.
with a value of d is calculated from the minumum saturation level, S
min
, and the maximum one,
S
max
, in the unit of the raw data:
T =
d S
min
S
max
S
min
. (10.3)
69
Data points with numerical values larger than S
max
and smaller than S
min
are given the same
colors as those assigned to S
min
and S
max
, respectively. The color of each point is determined
from T, depending on color modes. One of six color modes, that is, B-G-R, R-G-B, C-M-Y,
Y-M-C, gray scale (from black to white), and inverse gray scale, is selected from a list box. For
example, blue is assigned to 0, green to 0.5, and red to 1 in the B-G-R mode, as shown in Fig.
10.13.
Saturation level
0 %
Max.
Data range
Min.
100 %
Figure 10.13: Relations among saturation levels, data values, and colors of sections.
When option Absolute values is checked, colors are assigned on the basis of absolute values
of data.
Option Assign colors recursively assigns rainbow colors recursively to data points with values
smaller or larger than the saturation levels, which aords an eect similar to contour lines with
gradient colors in between them. The opacity of sections are specied in the {Opacity of drawn
sections} box, and that of lattice planes in the Lattice Planes dialog box (see 8.2). When {Cuto
level of lattice plane} is larger than 0, some parts of lattice planes with data values smaller than
the cuto level are omitted.
70
Chapter 11
OVERALL APPEARANCE
The Overall Appearance dialog box (Fig. 11.1) appears on selection of the Overall appearance...
item under the View menu.
Figure 11.1: Overall Appearance dialog box.
11.1 Background
The background color is specied with three values in between 0 to 255, or selected from a color
selection dialog box after clicking the [Select...] button.
11.2 Lighting
Ambient: A light that comes from all directions equally and is scattered in all directions
equally by objects. Specied by a value in between 0 to 100.
Diuse: A light that comes from a particular direction and hits objects with an intensity
that depends on the orientation of their surfaces. However, once the light hits a surface, it
71
reects evenly o a surface and radiates in all directions. Specied by a value in between
0 to 100.
Lighting direction: This can be changed by dragging the track ball placed below the
Lighting frame box.
Even though the ambient and diuse lights are evenly reected, some part of lights are reected
more in the manner of a mirror where most of the light bounces o in a particular direction.
A light having such a reection component is called specular light. In VESTA, the intensity
of the specular light is xed at 100 %, and a color of a reected specular light, i.e., specular
color is controlled by two properties of objects, {Specular} and {Shininess}. The nal color of
a surface is the sum of all three components of lights. A object surface appears to be brighter
than the color of object if the specular light is reected in the direction of viewpoint.
11.3 Projection Mode
Parallel: Objects are rendered by parallel projection.
Perspective:: Objects are rendered by perspective projection. Accentuation of the per-
spective view is controlled by the slider placed below the radio button.
Figure 11.2 shows the structure of post-perovskite, MgSiO
3
[70], visualized in the parallel
and perspective modes.
A B
Figure 11.2: Crystal structure of post-perovskite, a high-pressure form of MgSiO
3
, rendered by
(A) parallel and (B) perspective projections.
11.4 Depth-Cueing
Depth-cueing blends a fog color with the original color of each object using the blending factor
f. The factor f at depth z is computed by
f =
end z
end start
. (11.1)
72
Both of the starting depth, start, and the ending depth, end, are input by the user. VESTA
automatically assigns the background color of the Graphics Area to the fog color C
f
. Then, the
color of a rendering object, Cr, is replaced by
C

r
= f C
r
+ (1 f) C
f
. (11.2)
When VESTA renders objects in the Graphics Area, internal coordinates of the OpenGL scene
are normalized in such a way that a radius of the bounding sphere for the scene becomes 0.9 and
that the center of the scene is placed at 0. Objects at z < start are clearly rendered without
any fog whereas objects at z > end are completely invisible.
Depth-cueing can be enabled or disabled in check box Enable depth-cueing. The eect of
depth-cueing is schematically displayed below the two text boxes Starting depth and Ending
depth.
Figure 11.3 illustrates the eect of depth-cueing on images of mordenite [71] viewed along
the c axis in the perspective-projection mode.
A B
Figure 11.3: Crystal structure of mordenite rendered with and without depth-cueing. (A)
Depth-cueing enabled by setting start and end parameters at 1.2 and 1.0, respectively. (B)
Depth-cueing disabled.
73
Chapter 12
UTILITIES
12.1 Equivalent Positions
The Equivalent Positions dialog box (Fig. 12.1) appears on selection of the Equivalent Posi-
tions... item under the Utilities menu.
Figure 12.1: Equivalent Positions dialog box.
A list of general equivalent positions is displayed in this dialog box. When one of the equivalent
positions in the list is selected, the corresponding symmetry operation is displayed in a matrix
form at the upper left of the dialog box. In the right side of this dialog box, symmetry operation
W and transformation of fractional coordinates (x, y, z) with it are explicitly described (see
6.2.1).
12.2 Geometrical Parameters
This dialog box lists interatomic distances and bond angles recorded in a le *.e output by
ORFFE [39]. ORFFE calculates geometrical parameters from crystal data in le *.xyz created
by RIETAN-FP [17], outputting them in le *.e. When reading in input and/or output les
of RIETAN-FP (*.ins, *.lst), VESTA also inputs *.e automatically provided that *.e shares
the same folder with *.lst and/or *.ins. Otherwise, *.e can be input by clicking the [Read *.e]
button in the Geometrical Parameters dialog box.
74
Figure 12.2: The Geometrical Parameters dialog box showing a list of bonds recorded for uora-
patite [72] in FapatiteE.e.
VESTA allows us to locate the bonds and bond angles displayed in the Geometrical Parameters
dialog box in the Graphic Area. On selection of a bond (2 atoms) or a bond angle (3 atoms) in this
dialog box, the corresponding objects in a ball-and-stick model is selected (highlighted), and vice
versa. Thus, atom pairs and triplets associated with geometrical parameters on which restraints
are imposed in Rietveld analysis with RIETAN-FP are easily recognized in the ball-and-stick
model.
Figure 12.2 exemplies visualization of a bond in a ball-and-stick model of uorapatite; a
PO3 bond (grey line) selected in the dialog box is highlighted in the structural model in the
graphic window. The upper part of the dialog box displays detailed information on the PO3
bond.
Because ORFFE calculates estimated standard deviations (e.s.d.s) of geometrical param-
eters from both diagonal and o-diagonal terms in the variance-covariance matrix output by
RIETAN-FP, the resulting values of e.s.d.s are more accurate than those evaluated by VESTA
from only the diagonal terms. Accordingly, the e.s.d.s output by ORFFE should be described
in papers rather than those calculated by VESTA.
75
12.3 Standardization of Crystal Data
On the use of RIETAN-FP [17], it is highly desirable for an axis setting and fractional coordinates
to be standardized in compliance with denite rules [73]. In RIETAN-FP, the following lattice
settings are inhibited:
1. monoclinic system: c-axis unique setting ( = 90),
2. trigonal system: rhombohedral lattice (a = b = c and = = = 90),
3. cubic and tetragonal systems: an inversion center not at the origin.
Unless standard settings are adopted, standardization of crystal axes and fractional coordinates
is indispensable for Rietveld analysis using RIETAN-FP; otherwise LAZY PULVERIX [74] built
in RIETAN-FP fails in generating diraction indices hkl and their multiplicities.
Standardized crystal data are generally suitable for structure renement because atoms in
the asymmetric unit are conned in a relatively narrow region. VESTA is capable of launching
STRUCTURE TIDY [40] for this purpose.
This feature of standardization of crystal-structure data is particularly useful when searching
for compounds with similar structures using the Inorganic Crystal Structure Database (ICSD)
[75].
In both STRUCTURE TIDY and LAZY PULVERIX, the following three standard settings
are preferred to other settings:
1. monoclinic system: b-axis unique setting ( = 90),
2. trigonal system: hexagonal lattice (a = b = c and = 120),
3. centrosymmetric space groups: an inversion center at the origin.
Let n be the number of atoms in the asymmetric unit, and (x
j
, y
j
, z
j
; j = 1, 2, , n) their
fractional coordinates. Then, the standardization parameter, , is dened as
=
n

j=1
_
x
2
j
+y
2
j
+z
2
j
_1
2
. (12.1)
Note that this equation does not contain occupncies, g
j
. STRUCTURE TIDY selects a set
of x
j
, y
j
and z
j
(j = 1, 2, , n) minimizing the value. For better distinction between
dierent structure-type branches, STRUCTURE TIDY further outputs another standardization
parameter, CG, depending also on lattice parameters:
CG =
1
nV
_
_
_
a
n

j=1
x
j
_
2
+
_
b
n

j=1
y
j
_
2
+
_
c
n

j=1
z
j
_
2
+2ab cos
_
n

j=1
x
j
y
j
_
+ 2ac cos
_
n

j=1
x
j
z
j
_
+ 2bc cos
_
n

j=1
y
j
z
j
_
_
_
1
2
,
(12.2)
where V denotes the unit-cell volume.
VESTA automatically normalize the fractional coordinates between 0 to 1 before standard-
ization of crystal-structure data because the absolute value of each fractional coordinate to
be converted by STRUCTURE TIDY should be less than unity; otherwise, the corresponding
part of the output text becomes disordered. If lattice parameters and fractional coordinates
of atoms in the asymmetric unit are changed on the transformation of the crystal lattice, cur-
rent crystal data are replaced with the standardized ones. The resulting data are output in
76
the Text Area while standard input and output les of STRUCTURE TIDY are, respectively,
saved as data.stin and data.sto in directory tmp under a directory for user settings (see 15.2);
data.sto provides us with more detailed information on the standardization of the crystal data.
For example, suppose that a structure data le for Si (space group: Fd

3m) is created on
the basis of the rst setting where Si in the asymmetric unit occupies the 8a site at (0, 0, 0).
Subsequent standardization using STRUCTURE TIDY moves Si from (0, 0, 0) to (1/8, 1/8,
1/8) in such a way that a center of symmetry is present at the origin (second setting). When
lattice parameters (a and ) and fractional coordinates based on a rhombohedral lattice are
input in a trigonal compound, STRUCTURE TIDY converts them into lattice parameters (a
and c) and fractional coordinates based on a hexagonal lattice.
An example of standardization of crystal data is given below. The structure of a high-T
c
superconductor YBa
2
Cu
4
O
8
is usually represented with the c axis perpendicular to the CuO
2
conduction sheet and space group Ammm (No. 65) [76]. However, the standard setting described
in International Tables for Crystallography, volume A [20] is Cmmm; Rietveld analysis with
RIETAN-FP has to be carried out on the basis of Cmmm.
Running STRUCTURE TIDY by selecting the Standardization of Crystal Data item under
the Utilities menus in VESTA, we obtain optimum crystal data based on space group Ammm;
they are listed at the tail of data.sto:
Axes changed to : b,c,a
Setting x,y,z origin 0.00000 0.00000 0.00000 gamma = 2.9785
( 65) C m m m - j2 i5 c oS30
-------------------------------------------------------------------------
DATA YBa2Cu4O8 2.9785 0.7284
CELL 3.8708 27.2309 3.8402 *
ATOM O1 4(j) 0 0.05253 1/2 O2
ATOM Ba 4(j) 0 0.36498 1/2 Ba
ATOM Cu1 4(i) 0 0.06138 0 Cu2
ATOM O2 4(i) 0 0.14562 0 O1
ATOM Cu2 4(i) 0 0.21296 0 Cu1
ATOM O3 4(i) 0 0.28178 0 O4
ATOM O4 4(i) 0 0.44786 0 O3
ATOM Y 2(c) 1/2 0 1/2 Y
TRANS b,c,a
REMARK Transformed from setting A m m m.
STRUCTURE TIDY converted a, b, and c axes in Ammm into b, c, and a axes in Cmmm,
respectively. The value of 2.9785 is the standardization parameter gamma, the last data,
0.7284, in the DATA line is CG, and the last data of each site is the site name input by the
user.
12.4 Niggli-Reduced Cell
Low-symmetry unit cells can be selected in a variety of ways regardless of the method of deter-
mining the unit cell suitable for indexing the whole diraction pattern. To compare and analyze
dierent indexing solutions, the lattice has to be reduced to a certain unique, peferably standard
form, which is usually achieved in the following way [77]:
In the orthorhombic system, the lattice parameters should be a b c.
77
In the monoclinic system, the lattice parameters should be a c with the b-axis unique
setting.
In the triclinic system, the reduction becomes more complicated owing to possible multiple
choices of basis vectors in the lattice.
Nigglis approach to unit-cell reduction [77, 78] denes the reduced cell in terms of the
shortest vectors:

1
+
2
+
3
= minimum. (12.3)
STRUCTURE TIDY [40] transforms the given unit cell to a Niggli-reduced cell to describe the
structure in space group P1 or P

1 when * precedes a space-group symbol in an input le. If

the Niggli-reduction is successful, the solution is output in the Text Area.
From the reduced form, we can obtain the metric symmetry of the lattice, which is the
highest symmetry possible for the lattice on the basis of only geometric considerations. A
second application of the Niggli-redcued cell is that dierent deformation variants of a basis
type (arstotype), probably conventionally described with dierent space groups and probably
non-comparable unit cells, can be compared if they have been standardized in their corresponding
Niggli-reduced cells.
The output in the Text Area on transformation to a Niggli-reduced cell for a spinel-type oxide
CoAl
2
O
4
is given below:
Lattice type F
Space group name F d -3 m
Space group number 227
Setting number 1
Lattice parameters
a b c alpha beta gamma
8.09500 8.09500 8.09500 90.0000 90.0000 90.0000
Unit-cell volume = 530.457520
Structure parameters
x y z g B Site Sym.
1 Co Co1 0.00000 0.00000 0.00000 1.000 1.000 8a -43m
====================================================================================
18 atoms, 16 bonds, 0 polyhedra; CPU time = 0 ms
Transformation to a Niggli-reduced cell
====================================================================================
Setting x,y,z origin 0.00000 0.00000 0.00000 gamma = 0.2165
ratio volume Niggli-reduced cell/original cell : 0.25
( 2) P -1 - i aP2
-------------------------------------------------------------------------
DATA 0.2165 0.3437
CELL 5.7240 5.7240 5.7240 60.000 60.000 60.000 *
ATOM Co 2(i) 0.12500 0.12500 0.12500 Co1
TRANS 0.5a+0.5b,0.5b+0.5c,0.5a+0.5c
REMARK Niggli-reduced cell
78
Figure 12.3: Simulation of an X-ray powder diraction pattern of a high-T
c
superconductor,
YBa
2
Cu
4
O
8
[76], which was plotted with Igor Pro.
12.5 Powder Diraction Pattern
VESTA utilizes RIETAN-FP [17] to simulate powder diraction patterns. On selection of the
Powder Diraction Pattern... item under the Utilities menu, a series of procedures, i.e.,
(a) generation of an input le, *.ins, for RIETAN-FP, (b) execution of RIETAN-FP in the
simulation mode, and (c) graphic representation of the resulting data in le, *.itx, with a
graphing program such as Igor Pro,
1
gnuplot,
2
and WinPLOTR
3
are executed by VESTA as if
they were implemented in VESTA (Fig. 12.3). The resulting three les, *.ins, *.lst, and *.itx,
are saved in folder tmp under the settings directory (see 15.2).
As described in 14.1, binary executable les of RIETAN, a graph generator, and a template
input le (*.ins) are specied in Preferences dialog box.
For those who have not any suitable program for visualization of *.itx format, a small Java
program named PowderPlot,
4
which was developed by one of the authors (K.M.), is included in
the archive le of VESTA. Unfortunately, it cannot be run on Intel Macs.
It is preferred to set NMODE at 1 (simulation of a powder diraction pattern) in the template
le of *.ins. In addition, change the following data in the template le if necessary:
Radiation (neutron, characteristic X-ray, or synchrotron X-ray diraction).
1
http://www.wavemetrics.com/
2
http://gnuplot.sourceforge.net/
3
http://www-llb.cea.fr/winplotr/winplotr.htm
4
http://www.geocities.jp/kmo mma/crystal/en/powderplot.html
79
Wavelength (synchrotron X-ray and neutron diraction) or target (diraction using char-
acteristic X rays).
Geometry (BraggBrentano, transmission, or DebyeScherrer geometry).
Prole functions and prole parameters.
Range of diraction angles, 2.
Prole cuto.
VESTA automatically standardizes lattice settings when outputting *.ins les to satisfy
conditions required by LAZY PULVERIX [74], which is embedded in RIETAN-FP for generat-
ing diraction indices hkl and their multiplicities. Standard lattice settings adopted in LAZY
PULVERIX are just the same as that in STRUCTURE TIDY [73, 40]: second settings in cen-
trosymmetric space groups with more than one origin choice and rst settings in the other space
groups. Then, beware that diraction indices displayed in PowderPlot are incompatible with the
coordinate system adopted in VESTA if the lattice setting in VESTA is other than a standard
one.
A standard-input le, *.ins, and two output les of RIETAN-FP, *.itx (an Igor text le to
store data to plot a simulated powder-diraction pattern) and *.lst, are created in directory
tmp under the directory for user settings (see 15.2). The standard-output le, *.lst, serves as a
template le for subsequent Rietveld analysis. A list of reections output in *.lst if NPRINT > 0
provides us with a variety of information about reections included in the simulated pattern. For
instance, simulation of an X-ray diraction pattern for uorapatite gives the following output
(parts of lines are shown):
h k l Code 2-theta d Ical |F(nucl)| |F(magn)| POF FWHM m Dd/d
1 0 0 1 10.896 8.11365 12288 16.7811 - 1.001 0.0746 6 0.01365
1 0 0 2 10.923 8.11365 6113 16.7811 - 1.001 0.0746 6 0.01361
1 0 1 1 16.877 5.24908 4459 11.1485 - 0.999 0.0745 12 0.00876
1 0 1 2 16.919 5.24908 2218 11.1485 - 0.999 0.0745 12 0.00874
1 1 0 1 18.929 4.68442 2078 12.0897 - 1.001 0.0745 6 0.00780
1 1 0 2 18.977 4.68442 1034 12.0897 - 1.001 0.0745 6 0.00778
2 0 0 1 21.891 4.05683 9675 30.2905 - 1.001 0.0746 6 0.00673
2 0 0 2 21.946 4.05683 4813 30.2905 - 1.001 0.0746 6 0.00671
1 1 1 1 22.945 3.87276 8573 21.1788 - 1.000 0.0746 12 0.00642
1 1 1 2 23.003 3.87276 4264 21.1788 - 1.000 0.0746 12 0.00640
.....
where Code 1 and Code 2 are, respectively, K
1
and K
2
reections, Ical is the calculated
integrated intensity, F(nucl) and F(magn) are, respectively, the crystal-structure and magnetic-
structure factors, POF is the preferred-orientation function, FWHM is the full-width at the half-
maximum intensity, m is the multiplicity, and Dd/d is the resolution, d/d.
12.6 Site Potentials and Madelung Energy
VESTA utilizes an external program, MADEL, to calculate electrostatic site potentials,
i
, and
the Madelung energy, E
M
, of a crystal. Three methods are used to calculate Madelung energies:
Ewald, Evjen, and Fourier methods; the Fourier method is adopted in MADEL. MADEL was
originally written by Katsuo Kato (old National Institute for Research in Inorganic Materials)
and slightly modied later by one of the authors (F.I.). An advantage of using MADEL in
VESTA is that troublesome inputting of formatted data, particularly symmetry operations, is
avoidable.
80
The electrostatic potential,
i
, for site i is computed by

i
=

j
Z
j
4
0
l
ij
, (12.4)
where Z
j
is the valence (oxidation state) of the jth ion in the unit of the elementary charge, e
(= 1.60217710
19
C),
0
is the vacuum permittivity (= 8.85418810
12
Fm
1
), and l
ij
is the
distance between ions i and j; the summation is carried out over all the ions j (i = j) in the
crystal. In case site j is partially occupied, Z
j
should be multiplied by its occupancy, g
j
. E
M
per asymmetric unit is calculated by using the formula
E
M
=
1
2

i
Z
i
W
i
, (12.5)
with
W
i
=
(occupncy) (number of equivalent positions)
(number of general equivalent positions)
. (12.6)
The summation in Eq. (12.5) is carried out over all the sites in the asymmetric unit. To obtain
the Madelung energy for the unit cell, E
M
must be multiplied by the number of general equivalent
positions .
Prior to the execution of MADEL, the oxidation numbers of atoms in the asymmetric unit
must be input in the Structure dialog box. Just after launching MADEL, you are prompted to
input two parameters, RADIUS and REGION:
RADIUS: Radius of an ionic sphere, s, in

A. The charge-density distribution, r, is given
by
(r) =
0
_
1 6(r/s)
2
+ 8(r/s)
3
3(r/s)
4

, (12.7)
where r is the distance from the center of the ionic sphere (r < s and (r) = 0
for r s). When lines for interstitial sites are not given in the input le, set
RADIUS at a value that is large enough but less than the smallest interatomic
distance (not half of it!).
REGION: Reciprocal-space range (in

A
1
) within which Fourier coecients are summed
up. MADEL sums up the Fourier coecients with respect to all hkls within
a sphere having a radius equal to RADIUS. Choose an appropriate value
ranging from 2

A
1
to 4

A
1
according to the desired precision of calculation.
Also, check whether or not a curve for Madelung energy versus REGION is
nearly at around the selected value of REGION.
The standard output of MADEL is displayed in the Text Area. The unit of
i
is e/

A (1
e/

A= 14.39965 V). The accuracy of

i
and E
M
obtained using MADEL is limited to 3 or 4
digits.
The following lines give part of an output le when this feature is applied to investigating
distribution of hole carriers in a high-T
c
superconductor, YBa
2
Cu
4
O
8
[76, 79]:
Potentials of sites in the asymmetric unit
Charge W x y z phi
Ba 2.000000 0.250000 0.500000 0.134830 0.500000 -1.328184E+00
Y 3.000000 0.125000 0.500000 0.000000 0.500000 -1.641144E+00
Cu(1) 2.000000 0.250000 0.000000 0.212980 0.000000 -2.155481E+00
Cu(2) 2.500000 0.250000 0.000000 0.061530 0.000000 -1.799758E+00
O(1) -2.000000 0.250000 0.000000 0.145450 0.000000 1.310316E+00
81
O(2) -2.000000 0.250000 0.000000 0.052280 0.500000 1.926823E+00
O(3) -2.000000 0.250000 0.500000 0.052360 0.000000 1.919274E+00
O(4) -2.000000 0.250000 0.500000 0.218010 0.000000 1.135534E+00
Electrostatic energy per asymmetric unit
-3.314042 e**2/A = -47.72105 eV
Electrostatic energy per mole
-4.604381 MJ/mol
The electrostatic energy per mole is calculated by multiplying a factor of the Avogadro constant
(6.0221410
23
mol
1
) and E
M
together. In the above calculation, all the holes are assumed
to be doped into the Cu2 atoms on the CuO
2
conduction sheet with the Cu1 atoms having an
oxidation state of +2.
Standard input les, *.pme, (see p. 107) for MADEL can be output by selecting the Export
Data... item under the File menu. This le is helpful in appending fractional coordinates
of interstitial sites to obtain their site potentials using the RIETAN-FPVENUS integrated
assistance environment.
12.7 2D Data Display
We sometimes wish to visualize 2D distribution of a physical quantity on a plane. In such a
case, open the 2D Data Display window to display 2D images of volumetric data. Details in the
feature of 2D Data Display are described in chapter 13.
12.8 Line Prole
Numerical values on a line segment with an equal step are
calculated from volumetric data by linear interpolation of vol-
umetric data and output in a text le to plot a line prole by a
graphing program such as Igor Pro, gnuplot, and WinPLOTR.
After this item has been selected, a dialog box (right gure) is
opened to input fractional coordinates, (x, y, z), of both ends
and the total number of data points to be calculated. After
clicking the [OK] button, a le selection dialog box appears
and prompts you to input the name of an output le.
The top part of an output le when dealing with electron densities in rutile-type TiO
2
is
listed below:
TiO2 (rutile)
Point 1: 0.00000 0.00000 0.00000
Point 2: 0.30530 0.30530 0.00000
50
0.000000E+000 2.899010E+002
4.047903E-002 2.582953E+002
8.095806E-002 2.272991E+002
1.214371E-001 1.836505E+002
1.619161E-001 1.328268E+002
.....
82
The rst line gives a title, followed by two lines for the fractional coordinates of both ends,
those of Ti and O atoms in this case. The fourth line is the total number of data points (= 50),
followed by 50 lines. In each of the 50 lines, the rst and second columns give a distance from
Point 1 in the unit of

A and an interpolated value, respectively.
12.9 Peak Search
Peak values and positions in volumetric data are searched and
listed in the Text Area. When this item is selected, a dialog box
(right gure) is opened to restrict the maximum number of peaks
listed in the Text Area. Enter 0 to list all the peaks detected in the
volumetric data.
Peaks in electron densities in the unit cell of rutile-type TiO
2
are exemplied below:
Number of peaks in the 3D pixel data = 10
Value x y z
1 2.89901E+002 0.00000 0.00000 0.00000
2 2.89901E+002 0.50000 0.50000 0.50000
3 5.66020E+001 0.29688 0.29688 0.00000
4 5.66020E+001 0.70313 0.70313 0.00000
5 5.66020E+001 0.79688 0.20313 0.50000
6 5.66020E+001 0.20313 0.79688 0.50000
7 9.07678E-002 0.70313 0.29688 0.00000
8 9.07678E-002 0.29688 0.70313 0.00000
9 9.07678E-002 0.79688 0.79688 0.50000
10 9.07678E-002 0.20313 0.20313 0.50000
Each line after the second line aords a serial number, an electron density, and fractional coor-
dinates (x, y, and z) of a peak.
12.10 Conversion of Electron Densities
We can extract some important physical quantities from electron densities obtained experimen-
tally from X-ray diraction experiments. Let (r) be the electron density at the position vector
r, and (r) the gradient vector eld of (r). With VESTA, we can calculate Laplacians of
electron densities
2
(r), electronic kinetic-energy densities g(r), electronic potential-energy
densities (r), and electronic energy densities h
e
(r) from electron densities displayed in the cur-
rent page according to Tsirelsons procedures [42]. Consideration of these functions, particularly

2
(r) and h
e
(r), is not restricted by the bond critical point, r
b
, where (r) = 0, providing
us with a comprehensive characterization of chemical bonding in molecules and crystals.
The electronic kinetic-energy density is expressed as
g(r) =

2
m
_
3
10
_
3
2
_
2/3
_
(r)

5/3
+

_
(r)

2
(r)
+
k
2
(r)
2
_
, (12.8)
where
2
is Diracs constant, and m is the electron mass; the gradient -expansion of the Green
function around the classical Thomas-Fermi approximation [80] yields = 1/72 and k = 1/6.
The electronic-energy density equals the sum of g(r) and (r):
h
e
(r) = g(r) +(r) (12.9)
83
with
(r) =

2
4m

2
(r) 2g(r). (12.10)
When the unit of (r) is bohr
3
, that of
2
(r) is bohr
5
while that of g(r), (r), and h
e
(r) is
E
h
/bohr
3
, where 1 bohr = a
0
= 5.29177211 10
11
m (a
0
: Bohr radius), and E
h
= 4.35974417
10
18
J (E
h
: hartree).
The electronic-energy density at r
b
gives a straightforward criterion for the recognition of
the atomic interaction type. That is, h
e
(r) < 0 is observed in shared-type atomic interactions
whereas h
e
(r) > 0 is observed in closed-shell interactions. In a similar manner, if the electrons
are locally concentrated around r
b
[
2
(r
b
) < 0], the electrons are shared by both nuclei, which
is typical for shared or covalent atomic interactions. Otherwise the electrons are concentrated in
each of the atomic basins [
2
(r
b
) > 0], and the atomic interaction belongs to the closed-shell
type.
On selection of the Conversion of Electron Densities item
in the Utilities menu, a dialog box appears (right gure) to
specify les to be output. The user-selected les are output
in the general volumetric-data format (see p. 102) to either a
directory where the electron densities are recorded or a direc-
tory specied by the user if VESTA fails in writing the les
under that directory. By default, les storing volumetric data
of
2
(r) and h
e
(r) are, respectively, output to les *.led and
*.ted with their contents visualized in two pages.
84
Chapter 13
TWO-DIMENSIONAL DATA
DISPLAY
13.1 Components of the 2D Data Display Window
Figure 13.1 shows the 2D Data Display window of VESTA running on Windows XP.
Figure 13.1: 2D Data Display window.
This window comprises the following four components:
Menu bar: On Windows and Linux, the menu bar is placed at the top of the Main Window.
On Mac OS X, menus are displayed at the top of the screen. File and View menus are
available in this window.
Toolbar: Tools used frequently.
Side Panel: Colors, contours, and other properties of a 2D image are controlled in the Side
Panel. The minimum and maximum values in the 2D slice/projection data are displayed
at the top of the panel.
Graphics Area: Display an 2D image of volumetric data.
85
If a 2D image has been previously displayed, it is restored as soon as this window is opened.
The default preferences of colors and width of contour lines, and the background color can be
changed when no 2D image is displayed in this window. To change these default settings, edit
values in the Side Panel and then close the window to save the settings.
13.2 Menus
File menu
Export 2D Data...: Export 2D data as a text le.
Export Raster Image...: Export a graphic image as a le with a raster (pixel-
based) format.
Close: Close the 2D Data Display window.
View menu
Zoom In: Zoom in an image.
Zoom Out: Zoom out an image.
13.3 Tools in the Toolbar
13.3.1 Rotation
Rotate around the x axis
Rotate around the x axis
Rotate around the y axis
Rotate around the y axis
Rotate around the z axis
Rotate around the z axis
These six buttons are used to rotate objects around one of the x, y, and z axes. The step width
of rotation (in degrees) is specied in a text box next to the sixth button:
13.3.2 Translation
Translate upward
Translate downward
Translate leftward
Translate rightward
These four buttons are used to translate objects toward up, down, left, and right, respectively.
The step width of translation (in pixels) is specied in a text box next to the fourth button:
86
13.3.3 Scaling
Zoom in
Zoom out
Fit to the screen
These three buttons is used to change object sizes. The step width of zooming (in %) is specied
in a text box next to the third button:
13.4 Create and Edit a 2D Image
Click the [Slice...] button in the Side Panel to create a new image. Then, a dialog box named
Create New Slice is opened (Fig. 13.2).
Figure 13.2: A dialog box for specications of a 2D image.
Three dierent methods of creating a slice or a projection image are supported:
1. (hkl) plane in the bounding box
2. (hkl) plane dened by two vectors
3. project along [hkl] axis.
Figure 13.3 schematically illustrates how a 2D image is created in each mode.
On selection of option (hkl) plane dened by two vectors or project along [hkl] axis, two-
dimensional data on the specied plane can be output as a text le by selecting the 2D Data
Display item from the File menu (see 15.5.1).
13.4.1 (hkl) plane in the bounding box
In this mode, a slice that cuts a bounding box is created. The size of the bounding box is
specied by ranges along x, y, and z axes. The position of the slice is given as a distance from
the origin of 3D data in the unit of either its lattice-plane spacing, d, or

A.
87
[u
1
v
1
w
1
]
[u
2
v
2
w
2
]
B
A
C
Figure 13.3: Electron-density distribution drawn with the three dierent modes for the (111)
plane in rutile-type TiO
2
. (A) (hkl) plane in the bounding box mode, (B) (hkl) plane dened
by two vectors mode, and (C) Project along [hkl] axis mode. The electron densities were
determined by MEM from X-ray powder diraction data.
88
13.4.2 (hkl) plane dened by two vectors
In this mode, the size of the slice is dened by two lattice vectors: [u
1
v
1
w
1
] and [u
2
v
2
w
2
]. The
two lattice vectors should be parallel to the (hkl) plane but should not be parallel to each other.
The position of the slice is given as a distance from the origin of 3D data in the unit of either
its lattice-plane spacing, d, or

A. This mode should not be used in non-crystalline and aperiodic
data. In this mode, data points on the slice are recalculated by linear interpolation. When the
numbers of data points along x, y, and z directions in the volumetric data are NPIX[0], NPIX[1],
and NPIX[2], respectively, the numbers of data points along a [u
1
v
1
w
1
] direction is set at the
least common multiple of non-zero values of u
1
NPIX[0], v
1
NPIX[1], and w
1
NPIX[2]. The
same procedures are applied to the numbers of data points along [u
2
v
2
w
2
].
Such a manner is used to avoid averaging of peak values in the volumetric data by interpo-
lation. However, no fail-safe mechanism has been implemented so far to avoid generation of a
too large number of data points.
13.4.3 Project along [hkl] axis
This mode shows the cumulative data of a series of (hkl) slices summed up along the [hkl]
direction in a user-specied range. The number of data points is calculated in the same way as
with the (hkl) plane dened by two vectors mode.
In the Slice Properties dialog box of this mode, three kinds of data have to be input:
h, k, and l: Three integers in an equation to represent the lattice vector, R = ha + kb + lc,
where a, b, and c denote fundamental lattice vectors. The [hkl] direction is dened with
R.
Range of projection: A pair of values is input in the unit of the magnitude of the lattice
vector, i.e., |R|. The origin and the point corresponding to R have values of 0 and 1,
respectively.
Vectors parallel to the (hkl) plane: A pair of vectors, V
1
and V
2
, is specied in the expres-
sion of the lattice vector, ua + vb + wc, on the (hkl) lattice plane with the reciprocal
lattice vector of ha +kb +lc.
13.5 Controlling Properties of a 2D Image
In the General page in the Side Panel, colors of the background and plane are mainly controlled
(see Fig. 13.1).
Press a button at the right side of Background color: to change the background color. Check
box Birds eye view enables or disables Birds eye view (Fig. 13.4). If Draw grid edges is
checked, edges of grids are drawn with solid lines of a color and a width specied below the
check box (Fig. 13.5).
When Fill polygons is checked, the surface of the plane is lled with colors corresponding to
data values. Colors of the plane are controlled in the same manner as with sections of isosurfaces
(see 10.6).
In the Contours page (Fig. 13.6), various properties of contour lines are specied. Check
Draw contour lines to draw contour lines, which are plotted in two dierent modes: linear or
logarithmic.
Contours are plotted in linear and logarithmic modes as solid lines (style L1) and dashed
ones (style L2). The width and colors of these lines are specied in the text boxes and buttons
in the Style frame box.
89
Figure 13.4: Birds eye view of a (001) slice of rutile-type TiO
2
.
A B
Figure 13.5: Electron-density distribution on the (001) plane in rutile-type TiO
2
. (A) without
grid edges and (B) with grid edges drawn with solid lines.
A B
Figure 13.6: Contours page in the Side Panel. (A) Linear mode and (B) logarithmic mode.
90
In the linear mode, lines are drawn at every Interval in data ranging from Min. to Max. The
numerical value at the Nth line, F(N), is given by
F(N) = Min. +N Interval. (13.1)
In this mode, line style L1 (solid line) is applied to lines of positive values, and line style L2
(dashed line) to lines of negative values.
In the logarithmic mode, F(N) is computed by
F(N) = AB
N/Step
, (13.2)
where A and B are constants specied by user. Line style L1 is applied to lines with integer
values of N/Step, and line style L2 to lines with non-integer values of N/Step.
91
Chapter 14
PREFERENCES
Default settings for the behavior of VESTA are changed in the Preferences dialog box (Fig.
14.1).
Figure 14.1: Preferences dialog box.
14.1 Settings for RIETAN
Settings for simulation of X-ray and neutron powder diraction patterns are specied in frame
box Settings for RIETAN.
Specify the absolute path of the executable binary le of RIETAN-FP [17], which is used for
simulating powder diraction patterns, after clicking the [Browse...] button at the right of text
box {RIETAN}.
Click the [Browse...] button at the right of text box {Viewer} to specify an application to
plot the powder diraction pattern output to le *.itx by RIETAN-FP. This text box may be
left vacant on the use of Igor Pro because the extension, itx, is associated with Igor Pro by
default if it has been installed.
VESTA uses a template le to export a standard input le of RIETAN-FP. The template
le is specied in text box {Template (*.ins)} by clicking the [Browse...] button at the right of
the text box. In the exported le, data other than a space group, and lattice and structure
parameters are copied from the template le. Beware that the format of the template le
must be compatible with the version of RIETAN-FP used by VESTA. The use of *.ins for an
92
older version of RIETAN-FP may cause an error on execution of its latest version. The default
template le distributed with VESTA is that for the latest version of RIETAN-FP.
14.2 Font for Text Area
In frame box Font for Text Area, a text font for the Text Area is specied. Specify the text font
after clicking the [Select...] button at the right of text box {Font for Text Area}.
14.3 Open a New File in
When VESTA reads in new data, a new page may be assigned to them, or an existing page may
be reused. To input new data under a new page, select New tab in radio box Open a new le
in. When Current tab is selected, data assigned to the current page are discarded, and new
data are input in the current page. Unless changes in old data have been saved, VESTA asks
you whether or not the changed data should be saved in a le.
14.4 Animation
Two settings for animation speed, i.e., the rotation angle per frame and the interval between
frames are specied in frame box Animation. The rotation angle per frame is specied in text
box {Step (deg./frame)}. The interval between frames is specied in milliseconds in text box
{Interval (ms)}. The number of frames per second is
1/ (t
draw
+t
interval
) , (14.1)
where t
draw
is the time required to draw a single frame of an image on the Graphics Area, and
t
interval
is the interval between frames.
14.5 Start-up Search for Bonds
All the bonds for atoms whose Cartesian coordinates have been input from *.cc1,*.cube,*.mld,
*.mol,*.pdb,*.sca and *.xyz are automatically searched on the basis of {Max. length} specied
in frame box Bond search, provided that Start-up search for bonds is checked and that these
les have no information on bonds.
14.6 A Setting for Raster Image Export
Resolutions of atoms and cylindrical bonds are scaled when exporting pixel-based images. The
numbers of stacks and slices for the graphics image, R
g
, is scaled as
R
g
= R
d
s f, (14.2)
where R
d
is the number of stacks or slices for display on the screen, s is the scale of the image,
and f is the factor to increase R
g
specied in text box {Increasing factor for stacks/slices}.
14.7 Default isosurface level
The default isosurface level, d(iso), after opening a le storing volumetric data, , is calculated
as
d(iso) = || +n (||), (14.3)
93
where || is the average of ||s, (||) is the standard deviation of ||, and n is a parameter
to adjust d(iso).
94
Chapter 15
INPUT AND OUTPUT FILES
15.1 File Formats of Volumetric Data
Volumetric data are composed of regular grids in 3D space. A volume element, voxel, at each
grid point represents a value on the grid point. Voxels are analogous to pixels, which represent
2D image data.
In general, two types of formats are used to record volumetric data in les: general and
periodic grids. Figure 15.1 schematically illustrates the general concepts of the two kinds of
the formats.
The general grid is a uniform one spanned inside a bounding box for molecules and a unit
cell for crystals. For crystal structures, part of data in the general grid are redundant owing to
the periodicity of the data. For example, a numerical value at (1, 1, 1) are equal to that at the
origin, i.e, (0, 0, 0). Grids where these redundant points have been omitted are called periodic
ones.
VESTA distinguishes the grid types automatically from le extensions. For volumetric data
with the periodic grid format, VESTA internally generates the general grid by adding redundant
data points. On preparation of volumetric data using a self-made script or a program, the user
must pay attention to the grid type of a le by himself.
: irreducible data points : redundant points (periodic replicas)
A B
Figure 15.1: Two types of the grids for volumetric data on a plane. (A) general grid
and (B) periodic grid.
95
15.2 Directories for User Settings
VESTA loads and saves two les, VESTA.ini and style/default.ini in a directory for user settings
to store user settings for the program and for graphics, respectively. A directory, tmp, to store
temporary les is also created in the same directory. The location of the settings directory is
determined in the following order of priority:
If environmental variable VESTA_PREF has been dened, it is used. In the case of Mac
OS X, denition of environmental variables for GUI applications is rather troublesome
because they must be described in a text le named environment.plist under a hidden
folder, /.MacOSX. Therefore, a utility called set_VESTA_PREF.app is included in the the
RIETAN-VENUS package
1
for Mac OS X to make it easier to dene VESTA_PREF; refer to
Readme mac.pdf for details in set_VESTA_PREF.app.
If a directory named VESTA (Windows) or .VESTA (Mac OS X and Linux) is included in
the home directory of the user, it is used.
If the user has a permission to write in the program directory of VESTA, that directory is
used so as not to put the home directory in disorder. In the case of the Mac OS X version,
the above directory is included inside the application bundle, i.e., VESTA.app/Contents/
Resources/, which can be reached with Finder by (1) clicking VESTA while pressing the
control key, (2) selecting Show Package Contents, and (3) opening folder Contents.
A directory named VESTA (Windows) or .VESTA (Mac OS X and Linux) are automatically
created in the home directory of the user on the execution of VESTA for the rst time by
the user.
1
http://homepage.mac.com/fujioizumi/download/download.html
96
15.3 Files Used by VESTA
Table 15.1: Files used by VESTA.
element.ini A le storing default colors and radii of atoms.
spgra.dat A le storing information on 230 space groups, e.g., coordi-
nates of equivalent positions compiled in International Ta-
bles for Crystallography, volume A [20].
spgro.dat A le storing non-conventional symbols of orthorhombic
space groups.
style.ini Default le for styles of the graphic scene.
template.ins A template le for the standard input of RIETAN-FP [17].
It is used for exporting *.ins and simulating of powder
diraction patterns.
VESTA def.ini Default le for various settings of VESTA.
$USER DIR/VESTA.ini

A le for user settings of VESTA. This le is copied from

VESTA def.ini when VESTA is executed for the rst time
by the user.
$USER DIR/style/default.ini

User settings le for styles of the graphic scene. This le is

copied from style.ini when VESTA is executed for the rst
time by the user.

$USER DIR is a directory for user settings (see 15.2).

15.4 Input Files
15.4.1 Structural data
1. VESTA format (*.vesta)
Text les containing the entire structure data and graphic settings. File *.vesta saved by
VESTA may contain relative paths to volumetric data les and a crystal-data le that are
automatically read in when *.vesta is reopened. If any one of keywords, IMPORT_STRUCTURE,
IMPORT_ORFFE, IMPORT_DENSITY, and IMPORT_TEXTURE, is included in *.vesta and followed
by lines specifying relative paths to data les, these data les are also input by VESTA
when *.vesta are opened.
After Rietveld analysis with RIETAN-FP [17], lattice and structure parameters in *.vesta
are automatically updated provided that *.vesta and *.ins share the same folder.
2. VICS format (*.vcs)
http://homepage.mac.com/fujioizumi/visualization/VENUS.html
VICS, which is the predecessor of the structure-drawing part of VESTA, is now obsolete.
Nevertheless, this format is still supported in VESTA to maintain compatibility.
3. American Mineralogist Crystal Structure Database [81] (*.amc)
http://rru.geo.arizona.edu/AMS/amcsd.php
97
In part of AMCSD text les, extra characters are attached to space group names, and non-
standard space-group symbols are used. Some of such non-standard space-group settings
may not be read in correctly. In such a case, modify the space-group symbol appropriately
and then change the setting number in the Structure dialog box if necessary. Sometimes,
you have to convert fractional coordinates by yourself if a non-standard setting that is not
described in International Tables for Crystallography, volume A [20] is adopted.
4. asse (*.asse)
http://www.nims.go.jp/cmsc/sta/arai/asse/
5. Chem3D (*.cc1)
http://www.cambridgesoft.com/
6. Crystallographic Information File (CIF; *.cif)
http://www.iucr.org/resources/cif/
CIF has a variety of formats for crystal data. You can get detailed information about CIF
from the above Web site. For example, CIF les may contain Cartesian coordinates, but
VESTA cannot input them. Note that VESTA does not support all the formats allowed
in the CIF format. For example, Cartesian coordinates included in CIFs cannot be input
with VESTA; only fractional coordinates should be given in CIFs. For readable formats,
refer to example les of *.cif in VENUS/examples/VICS/CIF. In the case of *.cif containing
multiphase data, only the crystal data of the rst phase are input.
VESTA does not use any insucient space-group information from CIFs created by SHELX-
97 [82] through WinGX [83]. In this case, set the space group in the Structure dialog box
after opening such a le, or modify a line relevant to the space group in the CIF in the
following manner:
_symmetry_space_group_number 12
or
_symmetry_space_group_name_H-M P 21/n
7. CrystalMaker text le (*.cmt, *.cmtx)
http://www.crystalmaker.co.uk/
CrystalMaker is a commercial program for building, displaying, and manipulating all kinds
of crystal and molecular structures.
8. Crystal Structure Search and Retrieval (CSSR; *.cssr)
In a le with the CSSR format, a setting numbers may be given after OPT =. Unfortu-
nately, we have no information about setting numbers in this format. Then, please change
it in the Structure dialog box if necessary.
9. Cambridge Structural Database [84] (CSD/FDAT; *.csd, *.fdt))
http://www.ccdc.cam.ac.uk/products/csd/
10. DL POLY [85] format (CONFIG, REVCON, *.cong)
http://www.cse.scitech.ac.uk/ccg/software/DL POLY/
11. FEFF input le (fe.inp)
http://leonardo.phys.washington.edu/fe/
FEFF [86, 87] is an automated program for ab initio multiple-scattering calculations of
X-ray Absorption Fine Structure (XAFS) and X-ray Absorption Near-Edge Structure
98
(XANES) spectra for clusters of atoms. The names of input les for FEFF must be
either feff.inp or FEFF.INP.
12. Output le of the Elk FP-LAPW Code (GEOMETRY.OUT)
http://elk.sourceforge.net/
Elk is an all-electron full-potential linearised augmented-planewave (FP-LAPW) code for
determining the properties of crystalline solids.
13. Inorganic Crystal Structure Database [75] (ICSD; *.ics)
http://www.z-karlsruhe.de/icsd content.html
Two retrieval programs, RETRIEVE for MS-DOS and FindIt for Windows, of ICSD output
text les with quite dierent formats. VESTA is capable of reading in both types of the
crystal data les.
In these *.ics les, extra characters are sometimes attached to space group names, e.g.,
P 42/n m c S, which should be P 42/n m c (P4
2
/nmc). In addition, full Hermann-
Mauguin space-group symbols are sometimes given in ICSD text les. In such a case,
an error message appears in both the Text Area and a message box. Read it carefully
to proceed to a next operation. When you encounter this type of an error, it is strongly
recommended to output a CIF instead of *.ics.
14. ICSD-CRYSTIN: (*.cry)
15. MDL Molle (*.mol)
http://www.symyx.com/downloads/public/ctle/ctle.jsp
16. MINCRYST (Crystallographic Database for Minerals; *.min)
http://database.iem.ac.ru/mincryst/
In part of MINCRYST text les, extra characters are attached to space group names, and
non-standard space-group symbols are used. In such a case, an error message appears in
the Text Area. Such space group names need to be appropriately changed. Change the
setting number in the Structure dialog box if necessary.
17. MOLDA [88] (*.mld)
http://www3.u-toyama.ac.jp/kihara/cc/mld/index.html
The Web site of MOLDA has been closed because the author, Hiroshi Yoshida passed away
in 2005. Then, the MODRAST/MOLDA format for *.mld is briey explaned here. This
format consists of the following lines:
(a) Line No. 1: A comment about the compound, e.g., its name
(b) Line No. 2: Number of atoms, n
a
, in the compound
(c) Lines No. 3 (3 +n
a
): Cartesian coordinates (x, y, and z) and an atomic number
(d) Line No. (4 +n
a
): Number of bonds, n
b
, in the compound
(e) Lines No. (5 +n
a
) (5 +n
a
+n
b
): A pair of atom numbers
For example, in the case of ethylene, where n
a
= 6 and n
b
= 5, the following lines are
required:
"Ethylene (CH2=CH2)"
6
.66958, 0, 0, 6
99
1.23873, -.94397, 0, 1
1.23873, .94397, 0, 1
-.66958, 0, 0, 6
-1.23873, -.94397, 0, 1
-1.23873, .94397, 0, 1
5
1, 4
1, 2
1, 3
4, 5
4, 6
18. Protein Data Bank (PDB; *.pdb)
http://www.wwpdb.org/
PDB has a variety of formats for crystal data. You can get detailed information on PDB
in a Web page http://www.wwpdb.org/docs.html. Note that VESTA does not support
all the formats allowed in these two formats. For readable formats, refer to *.pdb in
VENUS/examples/VICS/PDB.
19. Input le of RIETAN-FP/2000 [17, 89] (*.ins)
http://homepage.mac.com/fujioizumi/download/download.html
VESTA cannot input *.ins for versions of RIETAN (e.g., RIETAN-94) earlier than RIETAN-
2000. In the case of *.ins containing multiphase data, only the crystal data of the rst
phase are input.
In *.ins, the volume name of International Tables should be not I but A in accordance
with a specication in RIETAN-FP. For example, A-230-2 is input for the second setting
of space group Fd

3m. The input of I-230-2 causes an error. Lattice parameters must be

given within one line in the following way:
CELLQ 9.36884 9.36884 6.88371 90.0 90.0 120.0 0.0 1010000
20. Output le of RIETAN-FP [17] (*.lst)
http://homepage.mac.com/fujioizumi/download/download.html
Beware lest *.lst output by RIETAN-2000 [89] is input.
21. WIEN2k [36] (*.struct)
http://www.wien2k.at/
22. XMol XYZ (*.xyz)
http://hackberry.chem.trinity.edu/IJC/Text/xmolxyz.html
XMol developed at the Minnesota Supercomputer Center is a utility for creating and
viewing graphic images of molecules.
23. F01 for SCAT [45, 90] and C04D for contrd
http://www.dvxa.org/
VESTA need to read in c04d, which is an input le of contrd, in addition to f01 if a
structural model is to be overlap with volumetric data. For this purpose, the dimensions
of a boundary box (an area where volumetric data are output to text les) in c04d are
required. For details in contrd, refer to Readme_contrd.txt in the VENUS package [10,
100
11]. Of course, c04d and f01 should be placed in the same folder. If c04d has not been input
by VESTA, atomic coordinates are treated in the Cartesian coordinate system, as is the
case for *.xyz les. In this case, volumetric data, *.scat and *.sca, cannot be overlapped
with the structural model.
All the atoms recorded in f01 must be included within the above boundary box. Otherwise,
atomic coordinates are normalized within the boundary box by assuming periodicity, which
leads to the appearance of incorrect structures in the Graphics Area.
The assistance environment for the DV-X method
2
using Hidemaru Editor is very con-
venient when carrying out a series of electronic-state calculations.
The Web site of Genta Sakane of Okayama University of Science:
http://www.chem.ous.ac.jp/%7Egsakane/
is very useful for those who like to visualize physical quantities calculated with contrd.
Introduction to the assistance environment for the DV-X method
3
distributed there
is a detailed Japanese document suitable for beginners in the DV-X method as well as
VESTA.
24. MXDORTO/MXDTRICL [91, 92] (FILE06.DAT, FILE07.DAT)
http://www.geo.titech.ac.jp/lab/kawamura/download/Download.html
25. XTL format (*.xtl)
Text les used in Cerius2 (Accelrys, Inc.
4
). GULP [93] and GSAS [13] are capable of
outputting crystal data with this format.
15.4.2 Volumetric data
26. PRIMA (*.pri, *.prim)
http://homepage.mac.com/fujioizumi/visualization/VENUS.html
Binary les of 3D electron and nuclear densities and Patterson functions output by PRIMA
[8] and ALBA [12], respectively. The unit of electron densities (strictly speaking, number
densities of electrons) recorded in these les is

A
3
, and that of nuclear densities is fm

A
3
.
27. MEED/PRIMA text data (*.den)
Text les of 3D electron and nuclear densities output by PRIMA, MEED [94], MEEDCAB
and MEND [95], and ENIGMA [96]. The unit of electron densities recorded in these les
is

A
3
while that of nuclear densities is fm

A
3
.
28. Energy Band (*.eb)
Text les having practically the same format with *.rho. Files *.eb are used to visualize
Fermi surfaces from results obtained with programs for band-structure calculations, e.g.,
WIEN2k [36]. As a matter of convenience, a constant is added to all the energy eigenvalues
in *.eb so as to make them greater or equal to zero. Therefore, the isosurface level has to
be set with this padding in mind.
Masao Arai of NIMS provides us with detailed information about *.eb in his Web site:
http://www.nims.go.jp/cmsc/sta/arai/
2
http://homepage.mac.com/fujioizumi/visualization/VENUS.html#assistance environment
3
http://www.chem.ous.ac.jp/%7Egsakane/dvxa assistance environment 1.html
4
http://accelrys.com/
101
29. General volumetric-data format (*.?ed)
A le with the general volumetric-data format stores one of the following physical quantities
converted from electron densities according to procedures proposed by Tsirelson [42] (see
12.10):

2
(r): Laplacian of the electron density (*.led).
g(r): Electronic kinetic-energy density (*.ked).
(r): Electronic potential-energy density (*.ped).
h
e
(r): Electronic energy density (*.ted).
with the format (common to all the les with the extension of ?ed):
Title: Title up to 80 characters.
a, b, c, , , : Lattice parameters with at least one space between two pa-
rameters (free format).
N1+1, N2+1, N3+1: Numbers of voxels along the a, b, and c axes, respectively,
with at least one space between two integers (free format).
followed by elements of a three-dimensional array, D
(((D(I1, I2, I3), I3 = 1, N3+1), I2 = 1, N2+1), I1 = 1, N1+1)
with any number of data in each line and at least one space between two real data (free
format). Note that voxels at N1+1, N2+1, and N3+1 lie on x = 1, y = 1, and z = 1,
respectively. A top part of an example of *.ted is given below:
TiO2 (rutile)
4.59393 4.59393 2.95886 90.00000 90.00000 90.00000
65 65 65
-3.700291E+004 -3.355551E+004 -2.515189E+004 -1.581991E+004 -8.551052E+003 -4.096268E+003
-1.802126E+003 -7.564262E+002 -3.147750E+002 -1.347187E+002 -6.129182E+001 -3.045389E+001
-1.683229E+001 -1.043066E+001 -7.226008E+000 -5.529954E+000 -4.588889E+000 -4.036150E+000
-3.669965E+000 -3.365605E+000 -3.043458E+000 -2.664371E+000 -2.229370E+000 -1.771544E+000
-1.337361E+000 -9.658093E-001 -6.762105E-001 -4.680259E-001 -3.280885E-001 -2.393913E-001
.....
VESTA allows you to save a le, *.?ed, with the general volumetric-data format. Accord-
ingly, VESTA serves as a converter from a binary le to a text one.
30. Periodic volumetric-data format (*.grd)
http://www.ncnr.nist.gov/xtal/software/gsas.html
Text les for 3D Fourier maps output by GSAS [13]. To create *.grd les that can be input
by VESTA, select option C Select section (X, Y, or Z) selection in Fourier calculation
setup and then enter X for prompt Enter section desired (X,Y,Z - choose Z for DSN6
maps).
The le format is essentially the same as the general volumetric-data format, except that
the three-dimensional data array, D, are output in the following range:
(((D(I1, I2, I3), I3 = 1, N3), I2 = 1, N2), I1 = 1, N1)
VESTA also allows you to export volumetric data in this format.
102
31. SCAT volumetric-data les (*.sca, *.scat)
http://www.dvxa.org/
Electron densities, electrostatic potentials, and wave functions calculated with contrd from
les F09 and F39 output by SCAT [45, 90]. Text les (CHG3D.SCA, POT3D.SCA,
WXXX-3D.SCA, WXXXU-3D.SCA, and WXXXU-3D.SCA) created with a batch le
named contrd.bat can directly be input by VESTA, where XXX denotes an integer assigned
to a wave function. To learn details in *.SCA, refer to Readme_contrd.txt distributed
as a part of package [10, 11]. Three-dimensional numerical data recorded in *.SCA or
*.SCAT are drawn without any conversion.
Units for electron densities, electrostatic potentials, and wave functions are, respectively,
bohr
3
, Ry (rydberg), and bohr
3/2
, where bohr is the atomic unit for length, i.e., 1 bohr
= a
0
= 5.29177211 10
11
m = 0.529177211

A (a
0
: Bohr radius), and 1 Ry = E
h
/2 =
2.179 871 9 10
18
J (E
h
: hartree).
Refer to No. 23 in 15.4.1 for the assistance environment for the DV-X method and its
detailed document written in Japanese.
32. WIEN2k (*.rho)
http://www.wien2k.at/ (WIEN2k)
http://www.nims.go.jp/cmsc/sta/arai/wien/venus.html (wien2venus.py)
A script, wien2venus.py, coded in Python
5
by Masao Arai of NIMS makes it possible to
export electron densities calculated with WIEN2k [36] to a text le, *.rho, which is in turn
visualized by VESTA. The unit of electron densities stored in this le is bohr
3
.
33. WinGX 3D Fourier map (*.fou)
http://www.chem.gla.ac.uk/louis/software/wingx/
Text les of 3D Fourier Maps output by WinGX [83]. To create *.fou that can be input with
VESTA, open FOURIER MAP Control Panel of WinGX from Maps menu FOURIER
MAP Slant plane. Select 3D Fouier (Beevers-Lipson) and Write MarchingCubes
File options and set Projection at Z axis. The minimum and maximum of Summation
limits should be set at 0 and 1 for all of X, Y, and Z axes. A resolution for each axis
should be carefully set in view of the following matter.
In this format, data points along each axis are not uniformly distributed but just placed
at intervals of given resolutions. When length of an axis is L and resolution is set at d, the
number of data points, NPIX, is set at the integer part of L/d+1. The output les have a
general grid format with nearly correct periodicity only if L/d is close to an integer. It is
recommended that special positions lie exactly on data grid. For example, if mirror planes
lie at x = 1/4 and x = 3/4, L/d should be multiples of 4.
34. X-PLOR/CNX (*.xplor)
http://accelrys.com/products/datasheets/cnx.pdf (CNX)
http://superip.fzu.cz/ (Superip)
Superip [97] is a computer program for ab initio structure solution of crystal structures
by the charge-ipping method [98, 99, 100]. Superip outputs electron densities in the
unit cell in le *.x-plor with the X-PLOR format [101], which can be directly visualized
by VESTA.
5
http://www.python.org/
103
15.4.3 Structural and volumetric data
35. CASTEP [102, 103] (*.cell, *.charg frm)
http://www.castep.org/
File *.cell contains crystal-structure data while le *.charg frm stores electron densities in
the unit of

A
3
. Only a structural model is visualized when *.cell is opened by VESTA.
On the other hand, both a structural model and electron-density distribution are displayed
when opening *.charg frm. Because no unit-cell dimensions are recorded in *.charg frm,
it must be accompanied by *.cell.
36. GAMESS [104] input and volumetric data les output by MacMolPlt [105]
http://www.msg.ameslab.gov/GAMESS/GAMESS.html (GAMESS)
http://www.scl.ameslab.gov/brett/MacMolPlt/ (MacMolPlt)
A GAMESS input le, *.inp, can readily be obtained from a GAMESS log le, *.log,
with MacMolPlt. At rst, examine the unit of nal Cartesian coordinates of atoms after
convergence in this le using a text editor. Then, run MacMolPlt to open *.log. In the
Windows menu, select Coordinates through the name of *.log and check whether the unit
of Cartesian coordinates is

A or bohr (au). The unit should be

A, as is usual with VESTA.
If the unit is Bohr, select Convert to Angstroms under the Molecule menu. Then, select
Input Builder through the name of *.log in the Windows menu and click [Write File] to
create *.inp storing atomic symbols and Cartesian coordinates.
Next, a volumetric data le, *.mmp, in which the origin of the 3D grid is recorded, has to be
output. In the Windows menu, select Surfaces through the name of *.log. Specify an item
from 3D Orbital, 3D Total Electron Density, and 3D Molecular Electrostatic Potential.
In the dialog box that appears subsequently, change the number of grid points and grid
size appropriately, select an orbital (in the case of 3D Orbital), and click [Update]. An
isosurface plus a ball-and-stick model appear in the *.log window. The number of grid
points, origin, and grid increment are displayed by clicking [Parameters...]. Then, click
[Export...]. Specify the name and location of *.mmp with the same basename as *.inp.
Note that *.inp and *.mmp must share the same folder.
37. Gaussian Cube format (*.cube, *.cub)
http://www.gaussian.com/
Text les storing electron densities, spin densities, electrostatic potentials, wave functions,
and so forth calculated with Gaussian [106, 107] with a keyword of Cube.
6
.
Cube les can also be created by PC GAMESS.
7
38. VASP (*.vasp, CHG, CHGCAR, PARCHG, LOCPOT, ELFCAR, POSCAR, CONTCAR)
http://cms.mpi.univie.ac.at/vasp/
http://www.materialsdesign.com/medea (commercial software MedeA including VASP as
its component)
All the above les are text les storing crystal-structure and volumetric data output by
VASP [108, 109].
CHG stores lattice vectors, atomic coordinates, and total charge densities multiplied by
the unit-cell volume, V . PAW one-center occupancies are added to them in CHGCAR.
Though both CHG and CHGCAR provide us with the same information on the valence
6
Units used in Gaussian 03 are described in http://www.gaussian.com/g tech/g ur/k constants.htm
7
http://classic.chem.msu.su/gran/gamess/
104
charges, the le size of CHG is smaller than that of CHGCAR owing to the lower accuracy
of the numerical data. PARCHG has the same format as CHG has, storing partial charge
densities of a particular k-point and/or band. When these les are read in to visualize
isosurfaces and sections, data values are divided by V in the unit of bohr
3
. The unit of
charge densities input by VESTA is, therefore, bohr
3
.
LOCPOT contains lattice vectors, atomic coordinates, and Coulomb potentials (unit:
eV), i.e., total potentials without exchange-correlation contributions (unless the line of
LEXCHG=1 is commented out in main.F). ELFCAR, which has the same format as
CHG has, stores dimensionless electron localization functions (ELF). POSCAR and CON-
TCAR include lattice vectors, atomic coordinates, and optionally starting velocities and
predictor-corrector coordinates for molecular-dynamics calculation. POSCAR and CON-
TCAR, respectively, correspond to the initial structure and the nal one output by VASP
at the end of a job; CONTCAR can be used to continue the job.
Owing to the absence of symbols of elements or atomic numbers in these les, they must
accompany another le, OUTCAR, to display structural models. OUTCAR may be re-
named *.out having the same basename as *.vasp has. Only lines following POTCAR:
at the top part of OUTCAR are required to obtain symbols of elements.
When les are read in for surface coloring, no data values are scaled unless the le name
is CHGCAR or PARCHG.
39. XCrySDen XSF format (*.xsf)
http://www.xcrysden.org/ (XCrySDen)
http://www.abinit.org/ (ABINIT)
http://www.abinit.org/documentation/helples/for-v6.4/users/cut3d help.html (Cut3D)
http://www.pwscf.org/ (Quantum ESPRESSO)
The ABINIT package [110] for the ab initio computation of material properties has a
feature to output binary les storing electron densities, electrostatic potentials, and wave
functions. They can be converted into text les, *.xsf, having the XSF format of XCrySDen
[111] using a converter named Cut3D. The unit of electron densities is bohr
3
. Cut3D sup-
ports data Type 13 (XCrySDen/VENUS wavefunction real data), whereby wave functions
can be directly output in *.xsf.
Quantum ESPRESSO [112] for quantum simulations of materials has also a feature to
output les with the XSF format.
Detailed information about the XSF format can be obtained from
http://www.xcrysden.org/doc/XSF.html
In general, *.xsf consists of some sections that begins with keywords. VESTA reads in
volumetric data from a BEGIN_BLOCK_DATAGRID section. To make it possible to overlap
isosurfaces with a structural model, *.xsf should also contain (1) PRIMEVEC and PRIMCOORD
sections or (2) CONVVEC and CONVCOORD sections. Furthermore, lattice vectors in the
PRIMVEC or CONVVEC section must agree with spanning vectors in the BEGIN_DATAGRID
section. The unit of primitive lattice vectors (PRIMVEC) and Cartesian coordinates is

A in
the XSF format.
105
15.5 Output Files
15.5.1 Data les
Structure data
1. Original format of VESTA (*.vesta)
The entire information about data displayed currently are saved in a text le, *.vesta, with
the VESTA format. Files with the VESTA format contains all the structural data and
graphic settings whereas volumetric data are not directly recorded in *.vesta but imported
from external les. Directories and names of les storing volumetric data are recorded in
*.vesta as relative paths to the volumetric-data les from the directory of *.vesta, which
makes it possible to save the entire information about the current data in a small le
without duplicating huge volumetric data.
Structural data can be optionally imported from an external le in the same manner. To
prevent the structural data from being recorded directly in *.vesta, import structural data
using the Import Data dialog box with option Link (see 5.2.1). A name of le *.e output
by ORFFE [39] is also recorded in *.vesta so that geometrical parameters recorded in *.e
are automatically listed in the Geometrical Parameters dialog box after reopening *.vesta
(see 12.2). After Rietveld analysis with RIETAN-FP [17], lattice and structure parameters
in *.vesta are automatically updated provided that *.vesta and a standard input le, *.ins,
share the same folder.
2. Chem3D (*.cc1)
http://www.cambridgesoft.com/
3. Crystallographic Information File (CIF; *.cif)
http://www.iucr.org/resources/cif/
4. DL POLY input le (*.cong)
http://www.cse.scitech.ac.uk/ccg/software/DL POLY/
5. Protein Data Bank (PDB; *.pdb)
http://www.wwpdb.org/
6. User input le of RIETAN-FP/2000 (*.ins)
http://homepage.mac.com/fujioizumi/download/download.html
The feature of outputting *.ins enables us to read in crystal data les of various formats
with VESTA, simulate powder diraction patterns, and carry out subsequent Rietveld
renements with RIETAN-FP [17]. Note that standard input les of RIETAN-2000 [89]
and RIETAN-FP are incompatible with each other. VESTA uses a template le to ex-
port input les of RIETAN-FP/2000, and the default template le, template.ins, has the
RIETAN-FP format. If a standard input le with the RIETAN-2000 format is desired,
change the template le to a le with the RIETAN-2000 format in the Preferences dialog
box (see 14).
In Rietveld renement or simulations of powder diraction patterns with RIETAN-FP, the
following standard lattice settings of STRUCTURE TIDY [73, 40] embedded in RIETAN-
FP should be selected:
Monoclinic system: b-axis unique setting ( = 90),
Trigonal system: hexagonal lattice (a = b = c and = 120),
106
Centrosymmetric space groups: an inversion center at the origin (origin choice 2 in
Ref. [20]).
Unless these three standard settings are adopted, LAZY PULVERIX [74], which is also
incorporated in RIETAN-FP, fails in generating correct diraction indices hkl and their
multiplicities. Accordingly, VESTA automatically transforms the lattice setting of a crystal
structure to the standard one when exporting *.ins with the RIETAN-FP/2000 [89, 17]
format.
It should be pointed out that fractional coordinates in part of crystal data recorded in
databases have too small signicant digits, e.g., 0.3333 instead of 0.333333. In such a case,
please increase the signicant digit to 6 or 7 before outputting *.ins.
7. XMol XYZ (*.xyz)
http://hackberry.chem.trinity.edu/IJC/Text/xmolxyz.html
8. VRML (*.wrl)
http://www.web3d.org/x3d/specications/vrml/
9. P1 structure (*.p1)
Simple text les storing lattice parameters and atomic positions of all the atoms in the
unit cell. Atomic positions are written in fractional coordinates, and the space group of
the crystal is converted into P1 (triclinic, No. 1). This format is useful when preparing
initial atomic congurations in computational simulations.
10. Fractional coordinates (*.xtl)
Simple text les with the xtl le format storing atomic positions of all the atoms currently
displayed in the Graphics Area. Atomic positions are written in fractional coordinates. The
space group of the crystal is treated as P1 (triclinic, No. 1) for convenience.
11. MADEL input les (*.pme)
Although MADEL can be executed from VESTA (see 12.6), input les of MADEL, *.pme,
should be manually edited to calculate electrostatic site potentials of interstitial (vacant)
sites, whose fractional coordinates have to be input in FORMAT(3F9.6) at the tail of *.pme.
The use of the RIETAN-FPVENUS integrated assistance environment included in the
distribution le of the RIETAN-FP system is highly recommended to edit *.pme and run
MADEL to input it on Hidemaru Editor.
8
For details in the above assistance environment,
read Readme macros.txt contained in the distribution le.
When exporting les with this format, you are prompted to input two parameters, RADIUS
and REGION (see 12.6). Refer to the users manual of MADEL for details in these
parameters to be input in *.pme.
Volumetric data
11. General volumetric-data format (*.3ed)
12. Periodic volumetric-data format (*.grd)
8
http://hide.maruo.co.jp/software/hidemaru.html
107
Two-dimensional data
Suppose that the 2D Data Display window has been opened to choose option (hkl) plane dened
by two vectors or project along [hkl] axis in the Create New Slice dialog box (see 13.4). Then,
two-dimensional data shown in 2D Data Display window can be output by selecting the 2D Data
Display item from the File menu. Note that this function is disabled on selection of the rst
option: (hkl) plane in the bounding box.
For example, electron densities on the (100) plane with a distance of 0.5d from the origin in
rutile-type TiO
2
are output as follows:
TiO2
(hkl) : 1 0 0
Vector 1: 0 1 0
Vector 2: 0 0 1
Distance from the origin: 2.29696 2.29696
65 65
0.000000 0.000000 1.771768E-02
0.000000 0.046232 1.777942E-02
0.000000 0.092464 1.796342E-02
0.000000 0.138697 1.826639E-02
0.000000 0.184929 1.868208E-02
.....
The two integers (65 and 65) in the sixth line are numbers of divisions along the x and y axes.
Three data in each data line are X, Y , and density, where X and Y are, respectively, x and y
coordinates (unit:

A) from the origin in the two-dimensional plot.
15.5.2 Raster images
1. BMP
2. EPS
3. JPEG
9
4. JPEG 2000
10
5. PNG
11
6. PPM
7. RAW
8. RGB (SGI)
9. TGA
10. TIFF
Images of structural and volumetric data displayed in the Graphics Area and the 2D Data Display
window can be recorded in les with a variety of graphics formats. When saving les storing
bit-mapped images, select Export Raster Image... in the File menu. The image size of the
9
http://www.jpeg.org/jpeg/index.html
10
http://www.jpeg.org/jpeg2000/index.html
11
http://www.libpng.org/pub/png/
108
exported graphics les is scaled up as a multiple of the size of the Graphics Area. The scale factor
is input after the le name has been entered in a le selection dialog box. VESTA can export
huge images that exceed the maximum width and height of the viewport, which are displayed
in the Text Area just after launching VESTA, e.g.,
Maximum supported width and height of the viewport: 4096 x 4096
The resolution of objects, i.e., the number of stacks and slices for atoms and bonds, are also
scaled by two factors {Scale} and {Increasing factor for stacks/slices} specied in the Preferences
dialog box (see 10.2, 10.3, and 14).
15.5.3 Vector images
1. Encapsulated PostScript (EPS)
2. Portable Document Format (PDF)
3. PostScript (PS)
4. Scalable Vector Graphics (SVG)
Images of structural and volumetric data displayed in the Graphics Area of the Main Window
can be saved as vector images. When saving les storing vector images, select Export Vector
Image... under the File menu. There are some limitations on exporting vector images.
Colors of isosurfaces and sections colored according to volumetric data cannot be represented
with vector image formats. Translucent polygons of polyhedra, lattice planes, and isosurfaces
become opaque in all the formats except for PDF.
15.5.4 Output text
The content of the Text Area can be saved as a text le by selecting Save Output Text... in
the File menu.
109
CLOSING REMARKS
Feedback and Suggestions
The development of VESTA has been led by user feedback. Its phenomenal success results from
the response from around the country. We assume that things are working ne provided that
no feedback is received. If its users were so selsh as to behave like black holes, VESTA
would not be improved very much. Please feel free to inform us of any bugs and your thoughts.
Sending us les relevant to your troubles could facilitate the debug.
Encouraging responses would present us with driving forces for further development and
renement of VESTA.
Acknowledgments
We wish to thank Ruben A. Dilanian of Melbourne University for his dedication to a great deal
of programming for VICS and VEND, the two predecessors of VESTA. Masao Arai of NIMS and
Masataka Mizuno of Osaka University kindly coded le converters, wien2venus.py for WIEN2k
and contrd for SCAT, respectively. Hiroshi Ohki of Shinshu University helped us improve VICS,
VEND and VICS-II signicantly. Genta Sakane of Okayama University of Science, Takuji Ikeda
of AIST, and Kazuki Komatsu of the University of Tokyo gave us many suggestions and helped
us debug the program.
Finally, we thank for all who have helped us improve the program by reporting bugs and
oering valuable ideas.
110
Appendix A
Scalability of VESTA
Table A.1: Maximum numbers of objects allowed in VESTA.
Object Maximum number of objects
Atoms (in the asymmetric unit) Unlimited

Atoms (total) Unlimited

Chemical species 99
Bonds Unlimited

Bonds per atom Unlimited

Bonds specications Unlimited

Polyhedra Unlimited

Polyhedral faces per polyhedron Unlimited

Vertices per polyhedral face Unlimited

Lattice planes (slices) Unlimited

Selected objects Unlimited

Hidden objects Unlimited
Polygons on isosurfaces Unlimited

Polygons on sections Unlimited

The theoretical limit is about 1 billion (2

30
) in 32-bit versions and about 2 quintillion (2
61
)
in 64-bit ones. These limits are, respectively, evaluated at the maximum sizes of pointer arrays
that can be expressed in 32- and 64-bit address spaces. Nevertheless, the actual limits are less
than them; the maximum numbers of objects are practically determined only by the amount of
free memory space in PCs.
111
Appendix B
Keyboard Shortcuts
Table B.1: Keyboard shortcuts.
Global (Menu)
Ctrl + N File menu - New Structure...
Ctrl + O File menu - Open...
Ctrl + S File menu - Save...
Ctrl + Shift + S File menu - Save as
Ctrl + W File menu - Close
Ctrl + Q File menu - Exit
Ctrl + E Edit menu - Structure...
Ctrl + B Edit menu - Bonds...
Ctrl + + View menu - Zoom In
Ctrl + View menu - Zoom Out
Ctrl + Shift + A View menu - Overall Appearance...
Ctrl + Shift + P Objects menu - Properties...
Ctrl + Shift + B Objects menu - Boundary...
Ctrl + Shift + O Objects menu - Orientation...
When the Graphics Area has focus
r Change Manipulation to the Rotate mode
m Change Manipulation to the Magnify mode
t Change Manipulation to the Translate mode
s Change Manipulation to the Select mode
d Change Manipulation to the Distance mode
a Change Manipulation to the Angle mode
Left Rotate along y
Right Rotate along +y
Up Rotate along x
Down Rotate along +x
Ctrl + Up Rotate along z
Ctrl + Down Rotate along +z
Shift + Left Rotate along y (1/10 step)
Shift + Right Rotate along +y (1/10 step)
Shift + Up Rotate along x (1/10 step)
112
Table B.1: Continued from the previous page.
Shift + Down Rotate along +x (1/10 step)
Ctrl + Shift + Up Rotate along z (1/10 step)
Ctrl + Shift + Down Rotate along +z (1/10 step)
Alt + Left Translate along x
Alt + Right Translate along +x
Alt + Up Translate along +y
Alt + Down Translate along y
113
Bibliography
[1] M. R. Silva, V. M. Moreira, C. Cardoso, A. M. Beja, and J. A. R. Salvador, Acta Crystal-
logr., Sect. C: Cryst. Struct. Commun., 64, o217 (2008).
[2] G. Sakane, H. Kawasaki, T. Oomori, M. Yamasaki, H. Adachi, and T. Shibahara, J.
Cluster Sci., 13, 75 (2002).
[3] K. D. Schotte, U. Schotte, H. J. Bleif, and R. Papoular, Acta Crystallogr., Sect. A: Found.
Crystallogr., 51, 739 (1995).
[4] W. I. F. David, J. Appl. Crystallogr., 20, 316 (1987).
[5] M. Sakata and M. Sato, Acta Crystallogr., Sect. A: Found. Crystallogr., 46, 263 (1990).
[6] G. Bricogne and C. J. Gilmore, Acta Crystallogr., Sect. A: Found. Crystallogr., 46, 284
(1991).
[7] R. J. Papoular, Acta Crystallogr., Sect. A: Found. Crystallogr., 47, 293 (1991).
[8] F. Izumi and R. A. Dilanian, Recent Research Developments in Physics, Vol. 3, Part II,
Transworld Research Network, Trivandrum (2002), pp. 699726.
[9] F. Izumi, Solid State Ionics, 172, 1 (2004).
[10] F. Izumi, Rigaku J., 36, 18 (2005).
[11] F. Izumi and R. A. Dilanian, Commission on Powder Diraction, IUCr Newsletter, No.
32, 59 (2005).
[12] F. Izumi, Spectroscopy and Diraction III, Vol. 11: the 5th Series of Chemistry, Maruzen,
Tokyo (2006), p. 220.
[13] A. C. Larson and R. B. V. Dreele, General Structure Analysis System (GSAS), Report
LAUR 86-748, Los Alamos National Laboratory, Los Alamos (2004).
[14] J. Rodriguez-Carvajal, Phys. B (Amsterdam, Neth.), 192, 55 (1993).
[15] F. Izumi and Y. Kawamura, Bunseki Kagaku, 55, 391 (2006).
[16] K. Momma and F. Izumi, Commission on Crystallographic Computing, IUCr Newsletter,
No. 7, 106 (2006).
[17] F. Izumi and K. Momma, Solid State Phenom., 130, 15 (2007).
[18] B. Quentrec and C. Brot, J. Comput. Phys., 13, 430 (1973).
[19] M. P. Allen and D. J. Tildesley, Computer Simulation of Liquids, Clarendon Press,
Oxford (1987), p. 149.
114
[20] International Tables for Crystallography, Vol. A, 5th ed., ed. by Th. Hahn, Kluwer,
Dordrecht (2005).
[21] D. K. Swanson and R. C. Peterson, Can. Mineral., 18, 153 (1980).
[22] W. H. Baur, Acta Crystallogr., Sect. B: Struct. Sci., 30, 1195 (1974).
[23] K. Robinson, G. V. Gibbs, and P. H. Ribbe, Science (Washington, DC, U. S.), 172, 567
(1971).
[24] R. Hoppe, Z. Kristallogr., 150, 23 (1979).
[25] R. Hoppe, S. Voigt, H. Glaum, J. Kissel, H. P. M uller, and K. Bernet, J. Less-Common
Met., 156, 105 (1989).
[26] C. Ferraris, Fundamentals of Crystallography, 2nd ed., ed. by C. Giacovazzo, Oxford
University Press, Oxford (2002), p. 526.
[27] M. Nespolo, G. Ferraris, and H. Ohashi, Acta Crystallogr., Sect. B: Struct. Sci., 55, 902
(1999).
[28] I. D. Brown and D. Altermatt, Acta Crystallogr., Sect. B: Struct. Sci., 41, 244 (1985).
[29] N. E. Brese and M. OKeee, Acta Crystallogr., Sect. B: Struct. Sci., 47, 192 (1991).
[30] I. D. Brown, The Chemical Bond in Inorganic Chemistry, Oxford University Press,
Oxford (2001).
[31] M. F. Brigatti, A. Mottana, D. Malferrari, and G. Cibin, Am. Mineral., 92, 1395 (2007).
[32] T. Balasubramanian, S. R. Jebas, S. Thamotharan, G. Rheinwald, and A. G. Lang, Acta
Crystallogr., Sect. E: Struct. Rep. Online, 62, o290 (2006).
[33] R. M. Hazen, A. Y. Au, and L. W. Finger, Am. Mineral., 71, 977 (1986).
[34] I. Hassan, S. M. Antao, and J. B. Parise, Am. Mineral., 89, 359 (2004).
[35] T. Nakagawa, K. Kihara, and K. Harada, Am. Mineral., 86, 1506 (2001).
[36] P. Blaha, K. Schwarz, G. Madsen, D. Kvasnicka, and J. Luitz, WIEN2k, an Augmented
Plane Wave + Local Orbitals Program for Calculating Crystal Properties, Technische
Universitat Wien, Wien (2001).
[37] J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zenitani, and J. Akimitsu, Nature (London,
U. K.), 410, 63 (2001).
[38] S. Kumazawa, S. Yamamura, E. Nishibori, M. Takata, M. Sakata, F. Izumi, and Y. Ishii,
J. Phys. Chem. Solids, 60, 1407 (1999).
[39] W. R. Busing, K. O. Martin, and H. A. Levy, A FORTRAN Crystallographic Function
and Error Program, Report ORNL-TM-306, Oak Ridge National Laboratory, Oak Ridge
(1964).
[40] L. M. Gelato and E. Parthe, J. Appl. Crystallogr., 20, 139 (1987).
[41] F. Izumi, J. Crystallogr. Soc. Jpn., 44, 380 (2002).
[42] V. G. Tsirelson, Acta Crystallogr., Sect. B: Struct. Sci., 58, 632 (2002).
115
[43] J. Smart, K. Hock, and S. Csomor, Cross-Platform GUI Programming with wxWidgets,
Prentice Hall (2005).
[44] K. Kihara, Eur. J. Mineral., 2, 63 (1990).
[45] H. Adachi, M. Tsukada, and C. Satoko, J. Phys. Soc. Jpn., 45, 875 (1978).
[46] W. J. A. M. Peterse and J. H. Palm, Acta Crystallogr., 20, 147 (1966).
[47] E. Prince, L. W. Finger, and J. H. Konnert, International Tables for Crystallography,
Vol. C, 3rd ed., ed. by E. Prince, Kluwer, Dordrecht (2004), p. 695.
[48] O. Zhou and D. E. Cox, J. Phys. Chem. Solids, 53, 1373 (1992).
[49] Y. Ohashi, X-Ray Crystal Structure Analysis, Syokabo, Tokyo (2005), p. 103.
[50] F. Izumi, Spectroscopy and Diraction III, Vol. 11: the 5th Series of Chemistry, Maruzen,
Tokyo (2006), p. 161.
[51] H. Arnold, International Tables for Crystallography, Vol. A, 5th ed., ed. by Th. Hahn,
Kluwer, Dordrecht (2006), pp. 7885.
[52] M. N. Burnett and C. K. Johnson, ORTEP-III: Oak Ridge Thermal Ellipsoid Plot Pro-
gram for Crystal Structure Illustrations, Report ORNL-6895, Oak Ridge National Lab-
oratory, Oak Ridge (1996).
[53] A. Prakash, Acta Crystallogr., 22, 439 (1967).
[54] K. Komatsu, T. Kuribayashi, A. Sano, E. Ohtani, and Y. Kudoh, Acta Crystallogr., Sect.
E: Struct. Rep. Online, 62, i216 (2006).
[55] Y. Tsujimoto, C. Tassel, N. Hayashi, T. Watanabe, H. Kageyama, K. Yoshimura, M.
Takano, M. Ceretti, C. Ritter, and W. Paulus, Nature (London, U. K.), 450, 1062 (2007).
[56] M. Rakiah, J. Refebvre, O. Hernandez, W. van Beek, and M. Serpelloni, J. Appl. Crystal-
logr., 37, 766 (2004).
[57] R. M. Hazen, R. T. Downs, and C. T. Prewitt, High-Temperature and High-Pressure
Crystal Chemistry, Vol. 41: Reviews in Mineralogy, Mineral. Soc. Am., Washington, D.C.
(2000), Chap. 1.
[58] M. Avdeev, E. N. Caspi, and S. Yakovlev, Acta Crystallogr., Sect. B: Struct. Sci., 63, 363
(2007).
[59] A. Beran, E. Libowitzky, and T. Armbruster, Can. Mineral., 34, 803 (1996).
[60] C. Giacovazzo, Fundamentals of Crystallography, 2nd ed., ed. by C. Giacovazzo, Oxford
University Press, Oxford (2002), p. 135.
[61] Z.-Q. Liu, Q. Fang, W.-T. Yu, G. Xue, D.-X. Cao, and M.-H. Jiang, Acta Crystallogr.,
Sect. C: Cryst. Struct. Commun., 58, o445 (2002).
[62] V. G. Tsirelson, M. Y. Antipin, R. G. Gerr, R. P. Ozerov, and Y. T. Struchkov, Phys.
Status Solidi A, 87, 425 (1985).
[63] C. Giacovazzo, Fundamentals of Crystallography, 2nd ed., ed. by C. Giacovazzo, Oxford
University Press, Oxford (2002), p. 81.
116
[64] D. F. Shriver and P. W. Atkins, Inorganic Chemistry, 3rd ed., Oxford University Press,
Oxford (1999), p. 24 and p. 71.
[65] B. K. Vainshtein, V. M. Fridkin, and V. L. Indenbom, Structure of Crystals, 3rd ed.,
Springer, Berlin (2000), p. 69.
[66] A. F. Wells, Structural Inorganic Chemistry, 5th ed., Clarendon Press, Oxford (1984),
p. 1288.
[67] R. D. Shannon, Acta Crystallogr., Sect. A: Found. Crystallogr., 32, 751 (1976).
[68] A. Bondi, J. Phys. Chem., 68, 441 (1964).
[69] L. Tong, D. M. Ho, N. J. Vogelaar, C. E. Schutt, and J. R. A. Pascal, J. Am. Chem. Soc.,
119, 7291 (1997).
[70] M. Murakami, K. Hirose, K. Kawamura, N. Sata, and Y. Ohishi, Science (Washington,
DC, U. S.), 304, 855 (2004).
[71] E. Passaglia, G. Artioli, A. Gualtieri, and R. Carnevali, Eur. J. Mineral, 7, 429 (1995).
[72] M. C. J. M. Hughes and K. D. Crowley, Am. Mineral., 74, 870 (1989).
[73] E. Parthe and L. M. Gelato, Acta Crystallogr., Sect. A: Found. Crystallogr., 40, 169
(1984).
[74] K. Yvon, W. Jeitschko, and E. Parthe, J. Appl. Crystallogr., 10, 73 (1977).
[75] A. Belsky, M. Hellenbrandt, V. L. Karen, and P. Luksch, Acta Crystallogr., Sect. B: Struct.
Sci., B58, 364 (2002).
[76] P. Lightfoot, S. Pei, J. D. Jorgensen, Y. Yamada, T. Matsumoto, F. Izumi, and Y. Kodama,
Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 47, 1143 (1991).
[77] V. K. Pecharsky and P. Y. Zavalij, Fundamentals of Powder Diraction and Structural
Characterization of Materials, 2nd ed., Springer, New York (2009), pp. 447-450.
[78] C. Giacovazzo, Fundamentals of Crystallography, 2nd ed., ed. by C. Giacovazzo, Oxford
University Press, Oxford (2002), pp. 8587.
[79] F. Izumi, Phys. C (Amsterdam, Neth.), 190, 35 (1991).
[80] D. Kirzhnits, Sov. Phys. JETP, 5, 64 (1957).
[81] R. T. Downs and M. Hall-Wallace, Am. Mineral., 88, 247 (2003).
[82] G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 64, 112 (2008).
[83] L. J. Farrugia, J. Appl. Crystallogr., 32, 837 (1999).
[84] F. H. Allen, Acta Crystallogr., Sect. B: Struct. Sci., 58, 380 (2002).
[85] W. Smith and T. Forester, J. Mol. Graph., 14, 136 (1996).
[86] J. Rehr and R. Albers, Phys. Rev. B: Condens. Matter Mater. Phys., 41, 8139 (1990).
[87] J. Rehr and R. Albers, Rev. Mod. Phys., 72, 621 (2000).
117
[88] H. Yoshida, Molecular Modeling on Computers A Guidebook of Molda , Science
House, Tokyo (2000).
[89] F. Izumi and T. Ikeda, Mater. Sci. Forum, 321, 198 (2000).
[90] Y. Kowada, I. Tanaka, H. Nakamatsu, and M. Mizuno, Hajimete no Denshi Jotai Keisan,
Sankyo Shuppan, Tokyo (1998).
[91] K. Hirao and K. Kawamura, Material Design Using Personal Computer, Syokabo, Tokyo
(1994).
[92] K. Kawamura, Jpn. Chem. Prog. Exchange Newsletter, 6, No. 4, 91 (1995).
[93] J. D. Gale, J. Chem. Soc., Faraday Trans., 93, 692 (1997).
[94] S. Kumazawa, Y. Kubota, M. Takata, M. Sakata, and Y. Ishibashi, J. Appl. Crystallogr.,
26, 453 (1993).
[95] K. Burger and W. Prandl, Acta Crystallogr., Sect. A: Found. Crystallogr., 55, 719 (1999).
[96] H. Tanaka, M. Takata, E. Nishibori, K. Kato, T. Iishi, and M. Sakata, J. Appl. Crystallogr.,
35, 282 (2002).
[97] L. Palatinus and G. Chapuis, J. Appl. Crystallogr., 40, 786 (2007).
[98] G. Oszlanyi and A. S uto, Acta Crystallogr., Sect. A: Found. Crystallogr., 60, 134 (2004).
[99] G. Oszlanyi and A. S uto, Acta Crystallogr., Sect. A: Found. Crystallogr., 61, 147 (2005).
[100] G. Oszlanyi and A. S uto, Acta Crystallogr., Sect. A: Found. Crystallogr., 63, 156 (2006).
[101] A. T. Br unger, X-Plor Version 3.1: A System for X-Ray Crystallography and NMR,
Yale University Press, New Haven (1993).
[102] M. D. Segall, P. J. D. Lindan, M. J. Probert, C. J. Pickard, P. J. Hasnip, S. J. Clark, and
M. C. Payne, J. Phys.: Condens. Matter, 14, 2717 (2002).
[103] S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. J. Probert, K. Refson, and M. C.
Payne, Z. Krystallogr., 220, 567 (2002).
[104] M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S. Gordon, J. J. Jensen,
S. Koseki, N. Matsunaga, K. A. Nguyen, S. Su, T. L. Windus, M. Dupuis, and J. A.
Montgomery, J. Comput. Chem., 14, 1347 (1993).
[105] B. M. Bode and M. S. Gordon, J. Mol. Graph. Mod., 16, 133 (1998).
[106] . Frisch, M. J. Frisch, and G. W. Trucks, Gaussian 03 Users Reference, Gaussian,
Inc., Pittsburgh (2003).
[107] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.
Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark,
J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K.
Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega,
J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
118
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J.
Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J.
Cioslowski, and D. J. Fox, Gaussian 09 Revision A.1, Gaussian Inc., Wallingford CT
(2009).
[108] G. Kresse and J. Furthm uller, Phys. Rev. B: Condens. Matter Mater. Phys., 54, 11169
(1996).
[109] G. Kresse and J. Furthm uller, Comput. Mater. Sci., 6, 15 (1996).
[110] X. Gonze, J.-M. Beuken, R. Caracas, F. Detraux, M. Fuchs, G.-M. Rignanese, L. Sindic,
M. Verstraete, G. Zerah, F. Jollet, M. Torrent, A. Roy, M. Mikami, P. Ghosez, J.-Y. Raty,
and D. C. Allan, Comput. Mater. Sci., 25, 478 (2002).
[111] A. Kokalj, J. Mol. Graph. Mod., 17, 176 (1999).
[112] P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G. L.
Chiarotti, M. Cococcioni, I. Dabo, A. D. Corso, S. de Gironcoli, S. Fabris, G. Fratesi,
R. Gebauer, U. Gerstmann, C. Gougoussis, A. Kokalj, M. Lazzeri, L. Martin-Samos, N.
Marzari, F. Mauri, R. Mazzarello, S. Paolini, A. Pasquarello, L. Paulatto, C. Sbraccia, S.
Scandolo, G. Sclauzero, A. P. Seitsonen, A. Smogunov, P. Umari, and R. M. Wentzcovitch,
J. Phys.: Condens. Matter, 21, 395502 (2009).
119
Index
2D Data Display, 82
ABINIT, 105
ALBA, 2
Alchemy, 2
AMCSD, 98
American Mineralogist Crystal Structure Database,
97
anisotropic atomic displacement parameter,
32
asse, 98
atomic scattering factor, 31
axis setting, 76
ball-and-stick model, 2, 17, 2123, 53, 57, 62,
75
BMP, 108
Bohr radius, 7, 28, 67, 84, 103
bond angle, 4, 53, 57
bond angle variance, 4, 54
bond valence parameter, 4
bond valence sum, 4, 55
bound coherent scattering length, 32
Cartesian coordinate, 101
CASTEP, 104
centrosymmetric space group, 76
CG, 76
charge distribution, 4, 55
charge-ipping method, 103
Chem3D, 98, 106
CHG, 104
CHGCAR, 104, 105
CIF, 8, 55, 98, 106
CNX, 103
complex lattice, 30, 39
CONTCAR, 104
contour line, 70, 86, 89
coordination polyhedron, 53
crystal-structure factor, 31
Crystallographic Information File, 98
CrystalMaker, 98
CSD, 98
CSSR, 98
Cut3D, 105
cuto plane, 49
DebyeWaller factor, 31
depth-cueing, 72
diraction index, 31
dihedral angle, 4, 53, 58
dispersion correction, 31
distortion index, 4, 54
DL POLY, 98, 106
dot surface, 17, 18, 21, 24
DV-X, 101
DVSCAT, 28
dynamic displacement, 45
eective coordination number, 4, 54
electron density, 32
electronic kinetic-energy density, 83, 102
electronic potential-energy density, 83, 102
electronic-energy density, 83, 102
electrostatic site potential, 33
electrostatic site potentials, 80
ELFCAR, 104
Elk FP-LAPW Code, 99
encapsulated PostScript, 109
Energy Band, 101
ENIGMA, 101
EPS, 108, 109
equivalent isotropic atomic displacement
parameter, 35
FDAT, 98
FEFF, 98
FindIt, 99
formal charge, 33
Fourier map, 102, 103
Fourier synthesis, 32
fractional coordinate, 3, 31, 76
GAMESS, 104
gamma, 77
Gaussian, 104
120
general equivalent position, 18, 40, 74
general grid, 95
general volumetric-data format, 107
GLUI, 2
GLUT, 2
gnuplot, 79, 82
Graphic Area, 75
GSAS, 101, 102
GULP, 101
hexagonal lattice, 76
ICSD, 8, 55, 76, 99
ICSD-CRYSTIN, 99
Igor Pro, 79, 82
integrated intensity, 31
interatomic distance, 4
inversion center, 76
isotropic atomic displacement parameter, 32
JPEG, 108
JPEG 2000, 108
Laplacian, 83
Laplacian of electron density, 83, 102
lattice parameter, 30, 76
lattice plane, 46
lattice-plane spacing, 32
LAZY PULVERIX, 76, 80
Line Prole, 82
Linux, 2, 9, 1315, 85, 96
LOCPOT, 104
Mac OS X, 2, 9, 13, 17, 96
MacMolPlt, 104
MADEL, 18, 80, 107
Madelung energy, 33, 80
magnetic moment, 2, 45
maximum-entropy method, 1
maximum-entropy Patterson method, 2
MDL Molle, 99
mean square displacement, 32
MEED, 101
MEM-based pattern tting, 1
metric tensor, 35, 37, 38
MINCRYST, 99
MOLDA, 99
MS-DOS, 99
MXDORTO, 101
MXDTRICL, 101
neutron diraction, 32
Niggli-reduced cell, 77
occupancy, 31, 34, 53
opacity, 2, 7, 24, 47, 64, 67, 70
OpenGL, 2, 9, 10, 73
ORFFE, 74, 75, 106
ORTEP-III, 42
OUTCAR, 105
oxidation state, 8, 55
PARCHG, 104, 105
Patterson function, 6, 101
PDB, 8, 100, 106
PDF, 109
peak search, 83
periodic grid, 95
periodic volumetric-data format, 107
PNG, 108
polyhedra, 53
polyhedral model, 2, 17, 22, 23
polyhedral volume ratio, 53
portable document format, 109
POSCAR, 104
POTCAR, 105
powder diraction pattern, 18, 79, 92, 97
PowderPlot, 79
PPN, 108
PRIMA, 2, 101
primitive lattice, 39
primitive translation vector, 37
protein data bank, 100
PS, 109
quadratic elongation, 4, 54
RAW, 108
reciprocal basis vector, 37
reciprocal-lattice, 32, 50, 51
RETRIEVE, 99
RGB (SGI), 108
rhombohedral lattice, 76
RIETAN-2000, 100, 106
RIETAN-FP, 8, 18, 59, 74, 75, 79, 92, 97, 100,
106, 107
rotation matrix, 36
SCAT, 100, 103, 110
SHELX-97, 98
shift of the origin, 37
shininess, 47, 61, 64, 72
space-lling model, 2, 4, 17, 21, 22, 62
121
specular, 47, 61
standard setting, 76, 77
standardization of crystal data, 76
standardization parameter, 76
static displacement, 45
stick model, 2, 17, 2123, 64
STRUCTURE TIDY, 76, 78, 80
sublattice, 39
super-lattice, 39
Superip, 103
surface coloring, 7, 18, 23, 26, 27, 68, 105
SVG, 109
symmetry operation, 3, 3840, 53, 57, 58, 74
temperature factor, 32
TGA, 108
thermal ellipsoid, 2, 22, 23, 33, 62
TIFF, 108
transformation matrix, 30, 36, 38
translation vector, 53
unit-cell volume, 39
VASP, 104
VEND, 1, 2, 10, 110
VENUS, x, 1, 100, 103, 105
VICS, 1, 2, 10, 97, 110
VICS-II, 10, 110
voxel, 2, 95
VRML, 107
WIEN2k, 7, 100, 101, 103
Windows, 2, 9, 12, 15, 85, 96, 99
WinGX, 98, 103
WinPLOTR, 79, 82
wireframe model, 2, 17, 18, 2123
wxWidgets, 10
X-PLOR, 103
X-ray diraction, 31
XAFS, 98
XANES, 99
XCrySDen, 105
XMol XYZ, 100, 107
XSF format, 105
XTL format, 101
122

An integrated three-dimensional visualization system VESTA using
wxWidgets

Koichi Momma
1
and Fujio Izumi
2

1
Institute of Mineralogy, Petrology, and Economic Geology, Tohoku University, Aoba, Sendai, Miyagi
980-8578, Japan;
2
Quantum Beam Center, National Institute for Materials Science, 1-1 Namiki, Tsu-
kuba, Ibaraki 305-0044, Japan.
E-mail: monmakou@ganko.tohoku.ac.jp - WWW: http://www.geocities.jp/kmo_mma/

1. Introduction

Progress in modern structure-refinement techniques of the maximum-entropy method (MEM) and MEM-
based pattern fitting (MPF) [1], has made it easier and more popular to determine three-dimensional (3D)
distribution of electron densities from X-ray diffraction data and densities of coherent-scattering lengths
(nuclear densities), b
c
, from neutron diffraction data. On the other hand, rapid developments of computer
hardware and software have accelerated and facilitated electronic-structure calculations affording
physical quantities including electron densities, wave functions, and electrostatic potentials.

Such technological advances in recent years bring demands for integrated 3D visualization systems to
deal with both structural models and 3D pixel data such as electron and nuclear densities. The crystal
structures and spatial distribution of various physical quantities obtained experimentally and by computer
simulation should be understood three-dimensionally. Despite the availability of many structure-drawing
programs, cross-platform free software capable of visualizing both crystal and electronic structures in
three dimensions is very few; if any, they are not very suitable for displaying those of inorganic and
metallic compounds.

To improve such a situation, we have recently developed a new integrated system VESTA (Visualization
for Electronic and STructural Analysis) for 3D visualization of crystal structures and pixel data on
personal computers. This article at first gives a brief overview of VESTA, which is followed by more
detailed descriptions of features and algorithm that have been newly implemented in it.

2. Circumstances behind the development of VESTA

VESTA is a successor to two 3D visualization programs, VICS and VEND, in the VENUS (Visualization
of Electron/NUclear densities and Structures) software package [1-4]. VENUS, which was developed by
Dilanian, Izumi, and Kawamura with help from Ohki and Momma during 20012006, comprises the
following five programs:

1. VICS (VIsualization of Crystal Structures) for displaying and manipulating crystal structures,
2. VEND (Visualization of Electron/Nuclear Densities) for displaying and manipulating 3D pixel
data,
3. PRIMA (PRactice Iterative MEM Analyses) for MEM analysis from X-ray and neutron
diffraction data,
4. ALBA (After Le Bail Analysis) for the maximum-entropy Patterson method,
5. Alchemy: a file converter to make it possible to analyze observed structure factors, which result
from Rietveld analysis using GSAS and FullProf, by MEM with PRIMA.

VICS and VEND are GLUT- and GLUI-based applications written in the C language with full use of
OpenGL technology. They saw the light of day at the end of 2002 and, since then, continued their growth
to be used widely in a variety of studies.

Page 106

However, we never get full satisfaction from their usability and performance for the following four
reasons. First, the combined use of them to visualize both crystal and electronic structures via text files is
rather troublesome; on-the-fly visualization of these two kinds of images is highly desired. Second, their
graphical user interface (GUI) is not very user-friendly because they are based on the old-fashioned
toolkits, GLUT and GLUI, which have been no longer upgraded. GLUT and GLUI offer only a limited
number of widgets. For example, they do not support pull-down menus and tabs for windows, which are,
at present, very popular in applications with GUI. One of the most serious troubles in VICS and VEND is
that they terminate suddenly without any warning messages whenever the close button at the upper right
corner of the title bar in each window/bar has been clicked. This problem results from a bug in the
toolkits. Third, VICS and VEND can deal with only a limited number of objects such as atoms, bonds,
polyhedra, and polygons on isosurfaces. Fourth, they require large system resources owing to unrefined
programming.

To overcome the above faults in VICS and VEND, we at first upgraded VICS to VICS-II employing a
modern C++ GUI framework wxWidgets [5,6] to build a new state-of-art GUI and further integrated
VICS-II and VEND into the next-generation 3D visualization system VESTA, adding new capabilities.

3. Overview

VESTA is a 3D graphic application written in the C++ language on the basis of OpenGL technology. It
runs fast on personal computers equipped with video cards accelerating OpenGL. Windows, Mac OS X,
and Linux versions are available. For each platform, both 32- and 64-bit applications will be distributed
on the Web.

Thanks to wxWidgets, we can open multiple files using tabbed graphic windows; pull-down menus and
tabbed dialog boxes are also supported. Needless to say, the annoying bug related to the close button
described above has now been fixed. VESTA allows us to deal with a practically unlimited number of
objects as far as memory size goes. It requires much less system resources than VICS and VEND.

VESTA represents crystal structures as ball-and-stick, space-filling, polyhedral, wireframe, stick, and
thermal-ellipsoid models. Ball-and-stick, wireframe, and stick models can be overlapped with dotted
surfaces corresponding to van der Waals radii. Polyhedra may be made translucent so as to make inside
atoms and bonds visible. We can insert a movable lattice plane with variable opacities into a structural
model. Drawing boundaries can be defined by ranges along x, y, and z axes as well as lattice planes.

Electron/nuclear densities, wave functions, and electrostatic potentials are visualized as isosurfaces,
birds-eye views, and two-dimensional (2D) maps. VESTA has a feature of surface colorization to show
another kind of a physical quantity at each point on isosurfaces. Translucent isosurfaces can be
overlapped with a structural model.

VESTA can read in files with 36 kinds of formats such as CIF, ICSD, and PDB and output files with 11
kinds of formats such as CIF and PDB. Users of RIETAN-FP [7,8] must be pleased to learn that standard
input files,
*
.ins, can be both input and output by VESTA. The entire crystal data and graphic settings can
be saved in a small text file, *.vesta, without duplicating huge 3D pixel data. File *.vesta with the VESTA
format contains relative paths to 3D data files and optionally a crystal-data file that are read in
automatically when *.vesta is reopened. VESTA also makes it possible to export graphic files with 14
image formats including 4 vector-graphic ones.

4. Programming models

4.1 GUI parts
wxWidgets [5,6], which was formerly referred to as wxWindows, is one of the best toolkits for cross-
platform GUI programming. It provides us with a look-and-feel inherent in each operating system. The
Page 107

license agreement of wxWidgets, an LGPL-like license with some exceptions allowing binary distribution
without source code and copyright, is flexible enough to permit us to develop any types of applications
incorporating wxWidgets

During the course of GUI reconstruction, we carefully separated the source code of GUI parts from that
of other core parts to make it easier to reuse the latter in combination with other GUI toolkits. The core
parts are basically controlled from the GUI ones. However, there are exceptions where functions provided
by wxWidgets are called in some core parts. In such a case, the function is wrapped by another function
to make the core parts quite independent of GUI toolkits and to show which functions depend on external
libraries.

VESTA supports multiple windows, each of which may contain multiple tabs corresponding to files.
Various kinds of information is output to the bottom area, where comments can also be input after
clicking the comment tab.

Figure 1 shows the main window of VESTA running on Windows XP. A brief hierarchy of widget
classes to build it is illustrated in Fig. 2. An OpenGL canvas is placed on a wxNotebook widget giving
the tab interface. Notebook widgets are usually designed to handle multiple-window components in the
same window area. However, we used some tricks to minimize system resources. VESTA was designed
in such a way that all the notebook pages have the same kind of a child, i.e., a wxGLCanvas widget.
Because the notebook widget ensures that no multiple pages are simultaneously displayed, we actually
need no multiple OpenGL canvases. Hence, VESTA assigns a single wxGLCanvas recursively to every
notebook page.



Fig 1: A screenshot of the VESTA main window.

wxPanel
wxGLCanvas
wxTextCtrl
wxFrame
wxNotebook
wxNotebook


Fig 2: Schematic image of widgets-class hierarchy on the main window.

Page 108

To maximize graphic performance, VESTA uses a cache mechanism of OpenGL display command lists
that increase in size with increasing number of drawing objects in the canvas; a command list for only one
canvas is always created. If each tab had its own canvas, the total size of memory consumed for the
command lists of all the OpenGL canvases would become huge. The layout of widgets in VESTA,
therefore, reduces memory usage dramatically. For the Linux and Mac OS X ports, some additional
tweaks are required to reuse the same OpenGL canvas in different tabs.

4.2 Core parts
In contrast to the GUI framework, we tried to reuse other parts of the source code as much as possible.
All the global variables related to crystal data and graphic settings are capsulated into a Scene class to
allow object-oriented programming. An instance of the Scene class is dynamically generated whenever
new data are created or read in from files so that multiple windows and tabs may work simultaneously.
The previous programs, VICS and VEND, can deal with only a limited number of objects and consume a
large amount of memory even when handling a relatively small number of data because both of them use
native ANCI C arrays. To eliminate this limitation with minimum effort and without any appreciable
overhead, C arrays were replaced with a small wrapper class of std::vector. The main part of the
wrapper class is coded as follows:

template<typename T> class objVector
{
public:
~objVector(){
clear();
}
inline T& operator[](size_t i) {
return (T&)*(T*)v[i];
}
inline void add(T* item){
v.push_back( item );
}
void clear(){
for(size_t i=0; i<v.size(); i++) delete (T*)v[i];
v.clear();
}
void remove(size_t i){
delete (T*)v[i];
v.erase(v.begin()+i);
}
std::vector<void*> v;
};

We also prepared the same kind of a wrapper class, aryVector, for pointers to arrays. This class
permits type-safe programming without unnesessary code duplication of the std::vector template
class. Further, it can be accessed by a [] operator in exactly the same way as with C arrays on a source-
code level despite great differences in real operations on a binary level. We should only note that all the
objects must be deletable; that is, they should be generated by a new operator. Because they are
automatically deleted by a destructor when wasting an array, we need not concern about memory leaks.
We have another genuine reason why the std::vector class is indirectly used in nearly all cases.
Instances of objects can be shared by two or more arrays of the same form. Elements in an array can be
quickly sorted only by manipulating pointers without copying a large amount of data for real objects.

5. Dealing with structural models

5.1 Searching for bonds
Bond-search algorithm was much improved in VESTA. The most primitive way to search for bonds is to
examine the entire pairs of atoms within drawing boundaries. We adopted a similar approach with some
improvements in VICS. Although this algorithm is very simple, the calculation time increases drastically
Page 109

with increasing number of atoms, N, in the drawing region. The calculation cost is proportional to the
square of N, which is expressed as O(N
2
) with big O notation.
12


On the other hand, the calculation time is linearly proportional to N, O(N), in VESTA as a result of
checking only atom pairs where distances between them are smaller than a certain value. To minimize the
number of atoms to be examined, a variant of the cell index method by Quentrec and Brot [9,10] was
adopted in VESTA. This approach is widely used in programs for molecular dynamics simulation that
needs to deal with a large number of atoms [11]. The basic idea of the cell index method is to divide a
large box into subboxes (subcells in the present case) and make up a list of atoms in each subcell. When
drawing a crystal structure, the large box corresponds to drawing boundaries, and sizes of a subcell along
x, y, and z directions are chosen so as to be slightly larger than the largest bond length. This manner
ensures that all the bonds for an atom can be found within the same cell plus neighboring 26 subcells,
each of which typically contains only a few atoms.

VESTA can search for bonds in three modes. In the first mode, both A1 and A2 atoms bonded to each
other are specified. The second mode is used to search for all the atoms located between minimum and
maximum distances from an A1 atom. In the third Search molecule mode, a set of atoms within a
specified bond length is reiteratively found. Atoms A1 and A2 may be specified by entering either
symbols for elements or site names.

5.2 Changes in drawing boundaries
Drawing boundaries for structural models can be changed in sophisticated ways similar to convoluting
and reiterative-convoluting spheres in ORTEP-III [12]. For example, we usually prefer that coordination
polyhedra are not omitted but drawn when their central atoms are situated within the drawing boundaries.
For this purpose, all the atoms bonded to these central atoms are added with option Beyond the
boundary (Fig. 3). To draw Fig. 3, bonds were searched in two different ways. One was to use the
Search A2 bonded to A1 mode with A1 = Al and A2 = O and check Show polyhedra and Search
beyond the boundary options. The minimum and maximum limitations of bond lengths were 0 and 2.1
, respectively. The other was to use the Search molecules mode and apply the Search beyond the
boundary option to OH bonds. Bond lengths were set between 0 and 1.0 , and drawing boundaries
between (0, 0, 0) and (1, 1.5, 1).



Fig 3: Crystal structure of AlOOH, a high pressure modification of aluminum oxide hydroxide, with
thermal-displacement ellipsoids at a 99 % probability level [13]. H atoms are represented by spheres of
an arbitrary radius.


12
The big O notation is a mathematical one to describe the behavior of a function for a very large (or a very small) number of
input data, in this case, N atoms. More precisely, f(n) is O(g(n)) means that there exists a certain constant m and a posi-
tive constant c such that f(n) c*g(n) for n > m.
Page 110

In organic compounds including metal complexes, the combination of the Search molecule mode and
option Beyond the boundary is used to search for the entire atoms of molecules with at lease one atom
lying within drawing boundaries.

5.3 Information about objects
Selection of objects (atoms, bonds, and coordination polyhedra) by clicking with a mouse provides us
with a variety of crystallographic information in the output window at the bottom:
1) fractional coordinates,
2) symmetry operations and translation vectors,
3) site multiplicities plus Wyckoff letters derived by STRUCTURE TIDY [14] embedded in
RIETAN-FP (for files of some formats)
4) interatomic distances, bond angles, and torsion angles,
5) information about coordination polyhedra including volumes, Baurs distortion indices [15],
quadratic elongations [16], bond angle variances [16], bond valence sums [17] of central
metals, and bond lengths expected from bond valence parameters [17].

5.4 Lattice transformation
VESTA has a feature to convert general equivalent positions in a conventional setting into those in a non-
conventional one with a transformation matrix, which is also used for (primitive lattice)(complex lattice)
conversions and for creating superstructures.
The ( ) 4 4 transformation matrix

=
1 0
p P
consists of the ( ) 3 3 rotation matrix P and the ( ) 1 3
translation vector p. Primitive translation vectors a, b, c are transformed by P as

( ) ( )
( )
( ). , ,
33 23 13 32 22 12 31 21 11
33 32 31
23 22 21
13 12 11
c b a c b a c b a
c b a
P c b a c b a
P P P P P P P P P
P P P
P P P
P P P
, ,
, , ' ', ',
+ + + + + + =

=
=


The shift of the origin is defined with the shift vector

( ) ( ) c b a c b a p c b a
3 2 1
3
2
1
, , , , p p p
p
p
p
+ + =

= = p .

The symmetry operation is transformed by
= ,

where Q

=
1 0 1 0
1 1
p P P q Q
- -
-
=
1
.

The lattice vector

=
1
z
y
x
is transformed by

= .
Page 111


If the determinant of P, det(P), is negative, the crystal-coordinate system is transformed from right-
handed to left-handed, and vice versa. If det(P) is not 1, the unit-cell volume, V, changes on the lattice
transformation. When det(P) is larger than 1, VESTA allows us to create a superstructure by examining
the following equation to find additional sites lying in between (0,0,0) and (1,1,1).

1
'
'
'
x
y
x
=

+
+
+
1
z
y
x
n z
n y
n x
...) , 2 , 1 , 0 , , ( =
z y x
n n n .

Note that VESTA adds no new equivalent positions but new sites when creating superstructures. That is,
the space group of the new structure should be a subgroup of the original space group. When det(P) is
smaller than 1, the same position may result from two or more sites. In such a case, the redundant sites
can be removed by clicking a Remove duplicate atoms button.

6. Dealing with 3D pixel data

6.1 Three kinds of representation
Isosurfaces are represented as smooth-shaded polygons, wireframes, and dot surfaces. Physical quantities,
e.g., wave functions and nuclear densities, having both positive and negative values are expressed by
isosurfaces with two different colors.

6.2 Surface colorization
Another attractive feature is surface colorization, whose typical application is to colorize isosurfaces of
electron densities according to electrostatic potentials. Fi gure 4 exemplifies an application of this
feature to results of an electronic-state calculation for the [Cd{S
4
Mo
3
(Hnta)
3
}
2
]
4
ion (H
3
nta:
nitrilotriacetic acid) [18] by a discrete variational X method with DVSCAT [19]. The isosurface level
was set at 0.03a
0
3
, and electrostatic potentials on the isosurface range from 0.814 Ry (blue) to 0.174 Ry
(red). Opacity parameters, C
1
and C
2
, for the isosurface (see 6.4) were 80 % and 100 %, respectively.



Fig 4: A composite image of an electron-density isosurface colorized on the basis of electrostatic
potentials and a ball-and-stick model for the [Cd{S
4
Mo
3
(Hnta)
3
}
2
]
4
ion [18].
Page 112

6.3 Two-dimensional distribution on lattice planes and boundary sections
Lattice planes (slices) can be inserted to display 2D distribution of a physical quantity in addition to its
isosurfaces. Both boundary sections and lattice planes are colorized on the basis of numerical values on
them. The saturation level of color is specified as (a) a value normalized between 0 and 100
corresponding minimum and maximum values, respectively, and (b) a value corresponding to raw data.
The color index, T, for a data point with a value of d is calculated from two saturation levels, S
min
and
S
max
, in the unit of the raw data:
min max
min
S S
S d
T

= .
Data points with values larger than S
max
and smaller than S
min
are given the same colors as those assigned
to S
max
and S
min
, respectively. The color of each point is determined from T, depending on color modes:
B-G-R, R-G-B, C-M-Y, Y-M-C, gray scale (from black to white), and inversed gray scale. For example,
blue is assigned to 0, green to 0.5, and red to 1 in the B-G-R mode.

6.4 How to calculate opacities of polygons
With VESTA, the visibility of both outlines of isosurfaces and an internal structural model was
surprisingly improved in comparison with that with VEND. The opacity, C(p), for polygon p with a
normal vector of (x, y, z) is calculated with a linear combination of two opacity parameters, C
1
and C
2
, for
polygons parallel and perpendicular to the screen, respectively:
). 1 ( ) (
2 1
z C z C p C + =

6.5 Order of drawing polygons
When drawing translucent objects with the OpenGL API, polygons should be usually drawn from behind.
VESTA sorts a set of polygons before drawing them, making it possible to render images properly
regardless of orientation of the objects. The sequence of drawing polygons can optionally be reversed.
The use of this option may sometimes improve the visibility of complex isosurfaces because neither back
surfaces nor internal ones are drawn; otherwise, they heavily overlap each other to give an improper
image.

6.6 Pixel operations
VESTA supports pixel operations between more than two 3D data sets. For example, this feature enables
us (a) to subtract calculated electron densities from observed ones obtained by MEM analysis to detect
atoms missing in a structural model and (b) to subtract down-spin electron densities from up-spin ones to
visualize effective spin densities (Fig. 5).

A
B
C


Fig 5: Distributions of electron densities and effective spin densities calculated with DVSCAT for the O
2

molecule. A: up-spin electron density; B: down-spin electron density; C: effective spin density (A B)
calculated with VESTA. Isosurface levels were set at 0.01a
0
3
(A and B) and 0.001a
0
3
(C), respectively.

Page 113

6.7 Improvements in rendering quality
The quality of rendering objects such as isosurfaces, boundary sections, and lattice planes in VESTA is
much higher than that in VEND, even when dealing with a relatively small number of pixels.

6.7.1 Smooth-shading model
A flat-shading model to draw isosurfaces was used in VEND. It assigns a normal vector to each polygon,
which causes the discontinuity of reflection colors near vertices and edges shared with neighboring
polygons. To settle this problem, we adopted a smooth-shading model, which assigns a normal vector to
each vertex of a polygon, in VESTA. The normal vector at each point on the polygon is calculated with a
shading model, GL_SMOOTH, provided by the OpenGL API using a linear combination of the normal
vectors at the vertices.

6.7.2 Colors for boundary sections and lattice planes
VEND assigns RGB values to vertices of polygons on boundary sections and lattice planes. However,
these colored polygons are problematic when data values mainly change in a narrow region. Suppose a
situation where blue is assigned to 0, green to 0.5, and red to 1. If the three vertices of a triangular
polygon have values of 0, 0.5, and 1, the color for the center of the polygon becomes a mixture of the
three colors:
3 / 1 3 / ) 1 0 0 ( = + + = R
3 / 1 3 / ) 0 1 0 ( = + + = G
3 / 1 3 / ) 0 0 1 ( = + + = B .
The resultant color is gray that does not represent any physical quantities. To avoid the appearance of
such a meaningless color, VESTA prepares a rainbow-colored one-dimensional texture, glTexImage1D,
and assigns coordinates in the texture to vertices. For RGB texture, 0 is assigned to blue texture positions,
0.5 to green ones, and 1 to red ones. Then, for any sets of three physical quantities at vertices and for any
positions of polygons, texture coordinates represent values at those positions. In the above case, the
texture coordinate (= color index), T, at the center of the polygon is:
5 . 0 3 / ) 1 5 . 0 0 ( = + + = T .

The internal unit of saturation levels for colors is changed from a normalized one to a raw-data one so that
saturation levels close to 0 or 1 (in a normalized unit) can work correctly. Assume that we specify
saturation levels normalized between 0 and 1 corresponding to minimum and maximum values in a data
set, respectively. Then, a problem arises from floating-point arithmetic errors if 3D data have extremely
large (or small negative) values as in the case of electrostatic potentials. For instance, suppose a case
where minimum and maximum data values are respectively 10
20
and 0 with saturation levels from 10 to
0 in the raw-data unit. The normalized value for 10 is (10 + 10
20
)/(0 + 10
20
) 1 while that for 0 is (0 +
10
20
)/(0 + 10
20
) = 1. Thus, with the normalized representation, values between 10 and 0 cannot be
distinguished within the precision of floating-point arithmetic. This problem does not occur in VESTA
because the internal representation of saturation levels, 10 and 0 in the above case, has the same
precision as with raw data. Thus, VESTA generates no factitious lines on colored isosurfaces.

6.7.3 Surface colorization
For colorization of isosurfaces, unrefined code was written in VEND. That is, data values at vertices on
the isosurfaces are approximated at those of the nearest pixels while different values are assigned to those
vertices of neighboring polygons which have just the same coordinates. We thoroughly redesigned the
surface-colorization feature in VESTA. With VESTA, the data value, U(v), for vertex v between two
pixels, p
1
and p
2
, is evaluated using a linear combination of U(p
1
) and U(p
2
):
2 1
1 2 2 1
) ( ) (
) (
l l
l p U l p U
v U
+
+
=
where l
1
and l
2
are the distance between v and p
1
, and that between v and p
2
, respectively; l
1
+ l
2
is equal
to the distance between p
1
and p
2
. Thus, VESTA shows much smoother isosurfaces without interpolation
of pixel data than VEND.

Page 114

6.8 Algorithm for geometry calculations
The calculation of isosurface geometry has been appreciably accelerated in VESTA compared with that in
VEND.

Figure 6 shows a 2D projection of 3D pixel data and grids defined by pixel rows along x, y, and z axes.
Vertices and polygons on isosurfaces are calculated for all the subgrids with eight pixels per grid. For
each subgrid, its twelve edges are checked whether they intersect the isosurfaces or not. For example, if a
pixel data, p(i,j,k), is larger than the isosurface level whereas p(i-1,j,k) is smaller than it, a vertex between
p(i,j,k) and p(i-1,j,k) for a polygon is generated. Every polygon on the isosurface generated in this way
has its own three vertices. However, all the pixels except those on drawing boundaries are shared with
neighboring eight subgrids, which means that values at most pixels are evaluated as many as eight times.
Furthermore, this method generates many duplicate vertices with the same coordinates because most
vertices on the isosurface are common among neighboring six polygons. The sluggish speed of
calculating isosurface geometry with VEND is explained in terms of these facts. The simple but
inefficient approach to calculate isosurface geometry is given below:

p(0,0,k) p(1,0,k) p(i,0,k)
. . . . . .
g(1,1,k) g(2,1,k) g(i,1,k)
g(1,2,k)
g(1,j,k)
g(2,2,k)
p(0,1,k)
p(0,j,k)
.

.

.

.

.

.
p(i-1,0,k)
p(0,j-1,k)


Fig 6: The relation between 3D pixel data p and grid g defined by nearest eight pixels. Polygons for the
isosurface are calculated for each grid.

calculate_isosurface_geometry_old(
float *rho, // 3D pixel data
float isolevel // isosurface level
int bbox[6], // Drawing boundaries
int NPIX[3], // Number of pixels along x, y, z in unit cell
){
int i, j, k, m;
int r[8][3];
float grid[8][4];
float *vertlist[12];
objVector<Polygon> poly; // list of polygons on isosurface
unsigned char cubeindex; // bit flags for a grid recording which pixels are outside/inside of
// isosurface
...

r[0][0] = 0; r[0][1] = 0; r[0][2] = 0;
r[1][0] = 1; r[1][1] = 0; r[1][2] = 0;
r[2][0] = 1; r[2][1] = 1; r[2][2] = 0;
r[3][0] = 0; r[3][1] = 1; r[3][2] = 0;
r[4][0] = 0; r[4][1] = 0; r[4][2] = 1;
r[5][0] = 1; r[5][1] = 0; r[5][2] = 1;
r[6][0] = 1; r[6][1] = 1; r[6][2] = 1;
r[7][0] = 0; r[7][1] = 1; r[7][2] = 1;

for (k=bbox[4]; k<bbox[5]; k++){
for (j=bbox[2]; j<bbox[3]; j++){
Page 115

for (i=bbox[0]; i<bbox[1]; i++){
for (m=0; m<8; m++){
int id = pos(i, j, k); // pos(i, j, k) reterns integer index of pixel(i, j, k)
grid[m][0] = i + r[m][0];
grid[m][1] = j + r[m][1];
grid[m][2] = k + r[m][2];
grid[m][3] = rho[id];
if (rho[id] < isolevel) cubeindex |= 1 << m;
}
// All the eight pixels are evaluated. Now calculate vertices for tweleve edges.
// edgeTable is a list of IDs used to judge wether the edge intersects the
isosurface or not.
// calc_vtx() returns coordinates of a vertex.
if(edgeTable[cubeindex] & 1) vertlist[0] = calc_vtx(isolevel, grid[0], grid[1]);
...
if(edgeTable[cubeindex] & 1024) vertlist[10] = calc_vtx(isolevel, grid[2], grid[6]);
if(edgeTable[cubeindex] & 2048) vertlist[11] = calc_vtx(isolevel, grid[3], grid[7]);

// Build polygons using vertices stored in vertlist.
...
} // end loop(i)
} // end loop(j)
} // end loop(k)
}

What is worse, if we try to render an isosurface with the smooth-shading model, redundant vertices of the
isosurface should be reduced to a single vertex to calculate a normal vector. Such a situation should be
avoided because of the following three reasons: pixel values are evaluated as many as eight times per
pixel; vertices on the isosurface are generated six times greater than needed; a large number of vertices
generated in this way are compared with each other to check duplication.

Our new algorithm introduced into VESTA eliminates all of these redundant calculations by a refinement
of the order of loops, as can be appreciated from Fig. 6.

calculate_isosurface_geometry_new(
float *rho, // 3D pixel data
float isolevel // isosurface level
int bbox[6], // Drawing boundaries
int NPIX[3], // Number of pixels along x, y, z in unit cell
){
int i, j, k;
aryVector<float*> vtx; // list of vertices on isosurface
unsigned char *cubeindex; // bit flags for each grid recording which pixels are outside/inside of
// isosurface
long *edgeList; // edge tables recording integer identifier of vertices on the edges
...
cubeindex = new unsigned char[2*(bbox[1]-bbox[0] +1)* (bbox[3]-bbox[2]+1)];
edgeList = new long[6*(bbox[1]-bbox[0] )* (bbox[3]-bbox[2])]

for (k=bbox[4]; k<=bbox[5]; k++){
... // copy cubeindex(i,j,k) to cubeindex(i,j,k-1), reset cubeindex(i,j,k).
// copy edgeList(i,j,k) to edgeList(i,j,k-1).
for (j=bbox[2]; j<=bbox[3]; j++){
for (i=bbox[0]; i<=bbox[1]; i++){
int M = pos(i, j, k); // pos(i, j, k) reterns integer index of pixel(i, j, k)
if (rho[M] > isolevel){
...
// register bit flags to 8 grids around this pixel that this pixel is inside
// of isosurface
}
if(i > bbox[0]){
...
// Calculate vertices between p(i-1, j, k) and p(i, j, k),
Page 116

// add to vtx and register it to edgeList.
}
if(j > bbox[2]){
...
// Calculate vertices between p(i, j-1, k) and p(i, j, k),
// add to vtx and register it to edgeList.
}
if(k > bbox[4]){
...
// Calculate vertices between p(i, j, k-1) and p(i, j, k),
// add to vtx and register it to edgeList.
}

if(i > bbox[0] && j > bbox[2] && k > bbox[4]){
...
// Generate polygons for grid(i, j, k) using the bit flag cubeindex(i, j, k)
// and edgeList.
}
} // end loop(i)
} // end loop(j)
} // end loop(k)
}

Note that cubeindex is now changed to an array of indices for two sheets with k=k-1 and k=k in grids
arranged three-dimensionally. Instead of carrying out a loop for grids to check pixel values, our new
algorithm directly carries out loops for pixels and updates the indices of eight grids around the pixel in
every cycle. In the same cycle, the coordinates of vertices on the three edges of grid g(i,j,k) are calculated
and stored in an array vtx. For example, in cycle (i=2, j=2, k=k), the coordinates of vertices between
p(1,2,k) and p(2,2,k), p(2,1,k) and p(2,2,k) (bold lines in Fig. 6), and p(2,2,k-1) and p(2,2,k) are calculated.
Other vertices of grid g(2,2,k) have already been calculated in other cycles, (i=1,j=2,k=k), (i=2,j=1,k=k),
etc. Then, in the same cycle, polygons for grid g(2,2,k) are generated using a list of vertices stored in
vtx.

This algorithm has another advantage that the memory size for data at vertices is significantly reduced
owing to the reuse of data at vertices among neighboring polygons.

7. Other new features

7.1 Cutoff planes
Drawing boundaries are fundamentally specified by inputting ranges along x, y, and z axes in the
Boundary dialog box, where we can further specify Cutoff planes. Each cutoff plane is specified with
Miller indices h, k, and l, and a central distance from the origin (0,0,0). After atoms, bonds, and
isosurfaces within the x, y, and z ranges have been generated, those lying outside of the cutoff planes are
excluded. Therefore, this feature is particularly useful to visualize distribution of 3D pixel data on lattice
planes in addition to isosurfaces. The central distance of a cutoff plane may be specified in the unit of
either its lattice-plane spacing, d, or .

7.2 Hiding objects efficiently
Two or more atoms, bonds, and polyhedra can now be selected with mouse dragging and hidden by
pressing the Delete key. Pressing the Esc key resumes the hidden objects.

7.3 Improvement in depth-cueing
Depth-cueing was revised to allow more flexible settings. An OpenGL API, glFog, with a
GL_LINEAR mode is used to enable depth-cueing. It blends a fog color with the original color of
each object using the blending factor f. Both of the starting depth, start, and the ending depth, end, are
defined by users. The factor f at depth z is computed by
Page 117

start end
z end
f

= .
VESTA automatically assigns the background of the canvas to the fog color Cf. Then, the color, Cr, of a
rendering object is replaced by
Cf f Cr f Cr * ) 1 ( * ' + = .
When VESTA displays objects on the OpenGL canvas, the depth of each object, z, is normalized in such
a way that the radius of a bounding sphere for the scene becomes 0.9. Then, start = 0.0 and end = 1.0, for
example, clearly render the half of the scene closer than z = 0.0 without any depth-cueing whereas the
farthest objects (z = 0.9) are completely invisible.

8. Collaboration with external programs

8.1 ORFFE
ORFFE [20] calculates geometrical parameters from data in file *.xyz created by RIETAN-FP, outputting
them in file *.ffe. When reading in *.lst and/or *.ins, VESTA also inputs *.ffe automatically provided that
*.ffe shares the same folder with *.lst and/or *.ins. Otherwise, *.ffe can be input by pressing the Read
*.ffe button in the Geometrical Parameters dialog box. This dialog box lists interatomic distances and
bond angles recorded in
*
.ffe.

VESTA allows us to locate the bonds and bond angles displayed in the Geometrical Parameters dialog
box on a graphic window. On selection of a bond (2 atoms) or a bond angle (3 atoms) in this dialog box,
the corresponding object in a ball-and-stick model is highlighted, and vice versa (Fig. 7). Thus, atom pairs
and triplets associated with geometrical parameters on which restraints are imposed in Rietveld analysis
with RIETAN-FP are easily recognized in the ball and stick model.



Fig 7: "Geometrical Parameters" dialog box showing a list of bond angles recorded for fluorapatite in
*.ffe. An O3PO2 angle is highlighted in the graphic window.

Page 118

Because ORFFE calculates estimated standard deviations (e.s.d.s) of geometrical parameters with both
diagonal and off-diagonal terms in the variance-covariance matrix, the resulting values of e.s.d.s are
improved compared with those given by VESTA using only diagonal terms.

8.2 RIETAN-FP
wxWidgets provides us with function wxExecute and class wxProcess to call external programs
easily from existing processes. Therefore, we can virtually extend VESTA by running other applications
as a child process, yet keeping its core code simple and small. VESTA utilizes this framework to simulate
powder diffraction patterns with RIETAN-FP. On selection of the Powder Diffraction Pattern item
under the Utilities menu, a series of procedures, i.e., generation of an input file, *.ins, for RIETAN-FP,
execution of RIETAN-FP, and graphic representation of the resulting data in file, *.itx, with a graphing
program such as Igor Pro and gnuplot, is executed by VESTA as if they were implemented in VESTA.

9. Summary

With all the advanced features and high performance described above, VESTA is expected to contribute
many investigations of crystal and electronic structures. It will act as a mediator between structure
analyses and electronic-structure calculations. We will polish it up further to make it more useful and
more powerful.

References

[1] F. Izumi and R. A. Dilanian, in: Recent Research Developments in Physics, vol. 3, part II,
Transworld Research Network, Trivandrum (2002) pp. 699726,.
[2] F. Izumi: Rigaku J., 36, No. 1 (2005) 18.
[3] F. Izumi and R. A. Dilanian: Commission on Powder Diffr., IUCr Newslett., No. 32 (2005) 59.
[4] F. Izumi and Y. Kawamura: Bunseki Kagaku, 55 (2006) 391.
[5] Information obtainable from http://www.wxwidgets.org/
[6] J. Smart, K. Hock, and S. Csomor: Cross-Platform GUI Programming with wxWidgets, Prentice
Hall (2005).
[7] F. Izumi and T. Ikeda: Mater. Sci. Forum, 321324 (2000) 198.
[8] F. Izumi and K. Momma: Proc. XX Conf. Appl. Crystallogr., Solid State Phenom., in press.
[9] B. Quentrec and C. Brot: J. Comp. Phys., 13 (1975) 430.
[10] M. P. Allen and D. J. Tildesley: Computer simulation of liquids, Clarendon Press, Oxford (1987) p.
149.
[11] K. Refson: Computer Phys. Commun., 126 (2000) 310.
[12] M. N. Burnett and C. K. Johnson: Report ORNL-6895, Oak Ridge National Laboratory (1996).
[13] K. Komatsu, T. Kuribayashi, A. Sano, E. Ohtani, and Y. Kudoh: Acta Crystallogr., Sect. E, 62
(2006) 216
[14] L. M. Gelato and E. Parth: J. Appl. Crystallogr., 20 (1987) 139.
[15] W. H. Baur: Acta Crystallogr., Sect. B, 30 (1974) 1195.
[16] K. Robinson, G. V. Gibbs, and P. H. Ribbe: Science,172 (1971) 567.
[17] I. D. Brown and D. Altermatt: Acta Crystallogr., Sect. B, 41 (1985) 244.
[18] G. Sakane, H. Kawasaki, T. Oomori, M. Yamasaki, H. Adachi, and T. Shibahara: J. Cluster Sci., 13
(2002) 75.
[19] H. Adachi, M. Tsukada, and C. Satoko: J. Phys. Soc. Jpn., 45 (1978) 875.
[20] W. R. Busing, K. O. Martin, and H. A. Levy: Report ORNL-TM-306, Oak Ridge National
Laboratory (1964).

Page 119
USERS MANUAL FOR MADEL
1. WHAT IS MADEL?
MADEL is a FORTRAN 77 program to calculate electrostatic site potentials,
i
, and Madelung
energies, E
M
, of ionic crystals by the Fourier method. It was programmed by Dr. Katsuo Kato,
National Institute for Research in Inorganic Materials (NIRIM), 1-1 Namiki, Tsukuba, Ibaraki
305, Japan.
This manual was translated from Japanese into English with some modifications and additions
by Dr. Fujio Izumi of NIRIM during his short stay at the Argonne National Laboratory in
March, 1991.
2. HOW TO COMPILE MADEL
This program should be compiled using an automatic double-precision option of a FORTRAN
77 compiler. If the compiler of your computer, unfortunately, does not support such a
convenient feature, declare IMPLICIT DOUBLE PRECISION (AH, OZ) in the main program
and all the subprograms of MADEL.

3. CREATING AN INPUT FILE FOR MADEL
Line 1. Title (A72)
Line 2. Radius of an ionic sphere and Fourier-summation range (2F6.2)
RADIUS: radius of an ionic sphere, s, in . The charge-density distribution, r, is given by
(r) =
0
[1 6(r/s)
2
+ 8(r/s)
3
3(r/s)
4
],
where r is the distance from the center of the ionic sphere (r < s and (r) = 0 for r
s). When Lines 7 are not given, set RADIUS at a value that is large enough but
less than the smallest interatomic distance (not half of it!).
REGION: reciprocal-space range (in
1
) within which Fourier coefficients are summed up.
MADEL sums up the Fourier coefficients with respect to all hkl s within a sphere
having a radius equal to RADIUS. Choose an appropriate value within the range
2.04.0
1
according to the desired precision of calculation. Also, check whether
or not a curve for Madelung energy versus REGION is nearly flat around the
1
selected value of REGION.
Line 3. Three integers related with the crystal structure (3I3)
NS: number of symmetry operations. Inverted positions (x, y, z) in a centrosymmetric
cell are not required, whereas translated positions in a complex lattice must be input.
NA: number of atoms contained in the asymmetric unit.
IZ: = 0, non-centrosymmetric cell.
= 1, centrosymmetric cell.
Line 4. Lattice constants (3F9.4, 3F9.3)
A(1): a () or a* (
1
).
A(2): b () or b* (
1
).
A(3): c () or c* (
1
).
A(4): (degree) or cos.
A(5): (degree) or cos.
A(6): (degree) or cos.
Lines 5. Symmetry operations (F11.6, 2I2, F11.6, 2I2, F11.6, 2I2)
Input NS lines (j = 1, NS).
Column
1 - 11: translational part of x
j
or blank.
12 - 13: 1, 2, 3, 1, 2, 3, or blank for x, y, z, x, y, z, or blank, respectively, as used in
the expression for the transformed x
j
.
14 - 15: 1, 2, 3, 1, 2, 3, or blank for x, y, z, x, y, z, or blank, respectively, as used in
the expression for the transformed x
j
. Columns 12 and 13 are exactly equivalent to
Columns 14 and 15. Also, note that an expression such as x
j
= 2x must be treated as
x
j
= x + x.
16 - 26: translational part of y
j
or blank.
27 - 30: integers representing plus or minus x, y, or z in the expression for transformed y
j
as
described above.
31 - 41: translational part of z
j
or blank.
42 - 45: integers representing plus or minus x, y, or z in the expression for transformed z
j
as
described above.
These lines have the same format as those in ORFLS (least-squares structure-refinement
program) and ORFFE (program to calculate interatomic distances, bond angles, etc.) developed
at the Oak Ridge National Laboratory. Therefore, they can be copied easily from input files
for these two programs.
2
Lines 6. Sites in the asymmetric unit (A8, 1X, 5F9.4)
Input NA lines (I = 1, NA).
ATOM(I): name of an atom occupying the ith site.
Z(I): valence (oxidation state) of the atom. Input a positive value for a cation and a
negative one for an anion. In case this site is partially occupied, Z(I) must be
multiplied by its occupation factor (occupation probability).
W(I): (occupation factor for the ith site) (number of equivalent positions for the ith
site) / (number of general equivalent positions).
X(1,I): fractional coordinate, x
i
.
X(2,I): fractional coordinate, y
i
.
X(3,I): fractional coordinate, z
i
.
Lines 7. Fractional coordinates of vacant sites whose potentials are to be calculated (3F9.6)
Input lines as many as you like (I = NA+1, NA+2, .....). These lines are optional.
X(1,I): fractional coordinate, x
i
.
X(2,I): fractional coordinate, y
i
.
X(3,I): fractional coordinate, z
i
.
In addition to the occupied sites input in Lines 6, electrostatic potentials containing no contribution
from each site input in Lines 7 can be calculated automatically. In such a case, RADIUS
should be set in such a way as not to overlap with each other: less than half the smallest
interatomic distance in the whole structure.
4. OUTPUT
The output of MADEL is described in German in the present version. The unit of the
electrostatic site potential,
i
, for the ith site is e/ (1 e/ = 14.399652 V), where e is the
elementary charge (= 1.602177310
19
C). The precision of results obtained using MADEL is
limited to 3 or 4 digits.
When the site input in Lines 7 is located within an ionic sphere, MADEL prints out a potential
(POBBA) excluding contribution of the sphere and, in addition, a potential (PMBBA) calculated
by substituting an original point charge for the sphere. If the site is not contained in any ionic
sphere, a potential in which contribution of the nearest neighbor (either one if two or more
nearest neighbors are present) is subtracted is output as POBBA.
3
The Madelung energy for the asymmetric unit, E
M
, is calculated by using the formula
E
M
=
1
2
i
i
Z
i
W
i
where Z
i
is the valence (oxidation state) of the ith site in the unit of e, and W
i
is W(I). E
M
must
be multiplied by the number of general equivalent positions to obtain the Madelung energy for
the unit cell.
4

1
How to uso oontrd and makooO4d

oontrd IDVSOAT 77|]/'.1fO9 f89 n|E,;0,m`:1,
}7+.1I77]0lortran .1^O'``1.'^:`)
77.oontr 08 ,;|7|j||].

makooO4d Ioontrd 0/7.1oO4d 0JJ#}|j7]0'77.

71Opon Watoom lortran Oomp1!orhttp`//www.oponwatoom.or/)7i1J|]],W1ndows
V1sta ]QJ77.


1. #

oontrd ISOAT n`7+117PO 7||J1IJ|A.7J|(,')
I:7')|7 dvd1r n]17,nQ/5.1}O]
.q15DVXAYLXLO I`J|/7nj17n7.

oontrd IE0.1 fO9,f89,oO4d[)}777.H|J}|j7[II,
MAKLlO6 )/7.1fO1 }VLSTA n/7|7.

@}|||77I,p>:07}|]I fO1
oO4d }[.q15IQ,p} VLSTA 77J1IJ|A.fO1 II]/]07
[)nn17J]77.oO4d n[.q15I|{7I[,VLSTA I fO1 }
7][IoO4d [%I777.

7+.1fO1,/'.1fO9[(^^+10)f8950)
I7I,,0@}@|7]

n}E,, E7 |, 7|], A]@,I|70[#`)[|@],E+
|1998).

oontrd 0|#5.1I,[07I|7 ODROM IO17 DVSOAT n7|] fO9
f89 }`J.1`.+})77}J]|77.

W1ndows )70OS],IMao OS X,UN!X)7DVSOAT }||7|]fO9 f89 I,/'
50.q+nW1ndows )DVSOAT })]/J],W1ndows )oontrd 7I7
|A.0I,oontrd 0`7.1^O'}= OS 7i1J7n,W1ndows ) DVSOAT
2
7 fO9 f89 }7|]|7].7+.17 fO1 I,|z] ORLl I7
1I777.


2. oontrd )/7.1oO4d)

oO4d 7I,05}/7|7

1 || |8!6) , , /0i^1
2 || |8l1O.6) , , /0[t` bohr,
O
.
O
= O.629177 7)
8 || |8l1O.6) 7(|t`
O
)
4 || |16!6) .1I7700
6 || |X6,16!6) 07i`01` up, 2` down)

1 ||n|8 ||I7,8D 5}77}]/]7]|7.fO1 I|n[17
7^70[n0]/]JI/7J1IJ|A.VLSTA I0]/]n]nt
7n0_I7J|7.]/]07jItQ7[n|{7,VLSTA IQj|I
7]/]0JjI?J[][})|7|7.

]/]70i^1II,[n7.[17 676 OOO i^1/]26O i^
1I|=)}Q7,Numbor of p1xo!s out of rano_7O>:n71,oontrd
n{L|7.

7(|0tQn]/]07(,7J|(=|7 |1/2, 1/2, 1/2) IQn17.],I,fO6
70[0|}7(|Ij7,0[n]/]07(ItQ7IJ7.

m`:1II717.m`:1 POT I1`^m`:
1APOT }|7[#|7`

POT = OPOT XPOT APOT

]]|OPOT I|@7^1`0},XPOT I5Qm`:1}|77.

}7|]I4 ||7]|7.71O j7777.

6 ||I7i`0}]707jD|7.

/7.1]},I|7
8
91 91 91
2O.OOOOO 2O.OOOOO 2O.OOOOO
O.OOOOO O.OOOOO O.OOOOO
2 87 886

0,fO6 I[}7(|,7 2O
O
2O
O
2O
O
0]/]70 919191 j0
E,;5}7|7.

770I2 7,87 886 0n717.

0I7II|70[#]01O9 ^:,J0
(f0|}7 }@|7].

7i`[#0I,00]Iup down })7}jD|7

41 41 41
2O.OOOOO 2O.OOOOO 2O.OOOOO
O.OOOOO O.OOOOO O.OOOOO
2 87 886
1 2

0,87 Iup 7i`,886 Idown 7i`0n717.

J VLSTA 7|@}7[,fO1 ]7J oO4d 70 | 7(|71
7.0,fO1 oO4d n[.q15I17J,7On)|707,.
].

J,oontrd n[%Iopon 7.10I71O6,0I71OO 77.


8. makooO4d

|IoO4d 0JJ#}|j7`JmakooO4d ])77.DVXAYLXLO .q15IQn17
/5.1makooO4d.bat }7makooO4d }||7.

fO9,f89,oO4d }O].q15}J1`+1^+'|7n|

makooO4d
4
`JO`7/7|7.@I?77.

`Numbor of atoms to bo ana!1zod |O`a!! atoms)

_[II,O`}/7|7.

I[|7,oO4d 70, , /0i^1}=JI|7].


4. 77+.1

oontrd n777+.1I)077

OHG8D.soat
OHG8DU.soat Up 7i`07i`|007)
OHG8DD.soat Down 7i`07i`|007)
POT8D.soat m`:1
WXXX8D.soat XXX 0
WXXXU8D.soat 7i`[#00up 7i`0
WXXXD8D.soat 7i`[#00down 7i`0

jsoat Isoa [77.

VLSTA 7I,l1!o >.00 Opon...}J,.1.q+ SOAT }]|7 +.soat
}77,}8D |7.VLSTA II,|I7j}|||] +.vosta }
77,||||n7.|I,0}m
`:10I[|7]777.[|I VLSTAManua!.pdf 0 1O.6 !sosurfaoos
}7].


6. oontrd 0|/

DVXAYLXLO .q15IQn17/5.1oontrd.bat }7oontrd }||7.fO9,
f89,oO4d }O].q15}J1`+1^+'|7n|

oontrd

`JO`7/7|7.@I?77.
6
6. 7.1#5

1 || |8!6,8l1O.6) >:.0[
2 ||)| |L14.6) | 502 ,;5 )

/0>:.n 72 )0,|I 1O 505n |1OL14.6) 0.q+77
17.

7150tI7I,VLSTAManua!.pdf 7016.4.2 Vo!umotr1o data }@|7
].

2OO2.O9.8O A]@
2OO8.O8.14 D]