Manual update as of 5/30/2014 5:10PM
(Goto TOC, KEYW, KEYA)

DOCUMENTATION FOR StoBe2014 (Version 3.3)

K. Hermann (Fritz-Haber-Institut der MPG, Berlin, Germany) L. G. M. Pettersson (University of Stockholm, Stockholm, Sweden) This manual is prepared by Klaus Hermann and Lars G.M. Pettersson using an original version prepared by Mark E. Casida, Annick Goursot, Emil Proynov, and Alain St-Amant. For questions and comments contact Klaus Hermann (hermann@fhi-berlin.mpg.de) Lars G.M. Pettersson (lgm@physto.se) The following header is included at the beginning of each StoBe output : ******************************************************************************* StoBe-deMon SOFTWARE Stockholm-Berlin version 3.3 of deMon (StoBe2014 dated May 2014) Copyright 1996-2014 by the authors. All rights reserved. ******************************************************************************* The deMon software is based on programs originally written under the direction of Professor Dennis R. Salahub of the University of Montreal. The codes are being developed and extended through an international collaborative effort. CITE THIS PROGRAM AS: StoBe-deMon version 3.3 (2014), K. Hermann and L.G.M. Pettersson, M.E. Casida, C. Daul, A. Goursot, A. Koester, E. Proynov, A. St-Amant, and D.R. Salahub. Contributing authors: V. Carravetta, H. Duarte, C. Friedrich, N. Godbout, M. Gruber, J. Guan, C. Jamorski, M. Leboeuf, M. Leetmaa, M. Nyberg, S. Patchkovskii, L. Pedocchi, F. Sim, L. Triguero, and A. Vela. ******************************************************************************* ******************************************************************************* (Goto TOP, KEYW, KEYA)

TABLE OF CONTENTS

1. Introduction and general description 1.1. Some history 2. Users guide 2.1. Introduction 2.1.1. Tables of keywords 2.1.1.1. Logical grouping 2.1.1.2. Alphabetic order 2.2. Detailed description of keywords 2.2.1. Keywords for geometry basics 2.2.2. Keywords for electronic states 2.2.3. Keywords for scf iteration part 2.2.4. Keywords for geometry optimization 2.2.5. Keywords for properties 2.2.6. Miscellaneous keywords 2.2.7. Untested / obsolete keywords 2.3. Basis set input 2.3.1. Global basis set definitions 2.4. Symmetry basis set input 2.4.1. Format of symmetry group file 2.5. Format of restart file 2.5.1. Previous restart file formats 2.6. Input/output file units 2.7. Utilities 2.7.1. Analyze restart files (anlyz) 2.7.2. Convert/copy/compare restart files (transf) 2.7.3. Plotting utility (khpltps) 2.7.4. Doing RIXS analysis 2.7.5. Total/partial densities-of-states (doscalc) 2.7.6. Total/angle-resolved Xray spectra (xrayspec) 2.7.7. Total/angle-resolved Infrared spectra (irspec) 2.7.8. Symmetry operations (symfind, symgen) 2.7.9. Combine restart files (combine) 2.7.10. Expand molecular orbitals (expand) 2.7.11. Remove symmetry in restart file (desymm) 2.7.12. Generate point charge sets (latsph, pcfilt) 2.7.13. Linear paths between images (linpath) 2.8. Program structure and dimensions 2.9. Compilation and error handling 3. Bibliography 1. INTRODUCTION AND GENERAL DESCRIPTION (Goto TOC, KEYW, KEYA) 1.1. SOME HISTORY (Goto TOC, KEYW, KEYA) The first version of deMon, deMon0, was described in Alain St-Amant's thesis [A92]. deMon0 was subsequently substantially modified for commercialization by BIOSYM Technologies Inc. The beta-release of this modified version became the basis of the deMon1 series. Meanwhile deMon0 was also undergoing modification by Annick Goursot (deMon3) in Montpellier and, later, by Lars G.M. Pettersson (deMon2) in Stockholm. deMon-KS3 represents the fusion of deMon2 and deMon3. deMon-KS3p0 permitted the use of symmetry in the parts of the program involving the grid, an aspect which can lead to a considerable saving of time. In addition, deMon-KS3p1 permits the use of symmetry to block-diagonalize the KS-Fock matrices, which can aid both convergence and the interpretation of the MOs. Particular emphasis was put on both computational efficiency and SCF convergence to allow very large computational models to be treated. 2. USERS GUIDE (Goto TOC, KEYW, KEYA) The present manual describes the most recent StoBe version of the deMon package which is developed and maintained by Lars G.M. Pettersson (Stockholm University, Sweden) and Klaus E. Hermann (Fritz-Haber-Institut, Berlin, Germany) to deal with very large molecules and surface clusters with specific implementations for inner-shell spectroscopies, see http://www.fhi-berlin.mpg.de/KHsoftware/StoBe/index.html. 2.1. INTRODUCTION (Goto TOC, KEYW, KEYA) The program StoBe is a realization of the Linear Combination of Gaussian Type Orbitals - MO solution of the Kohn-Sham DFT equations. The theory and numerical details of this realization can be found in refs.[SCP93,SCF+95]. The StoBe input file consists of a series of keywords. By calling a keyword, the user can activate or deactivate the option(s) associated with that keyword. For certain options, the user must also assign numbers, such as the convergence criteria for an SCF procedure. To make the input as simple as possible, the input is almost everywhere in free fortran format and each keyword or option may be truncated to four letters (both upper and lower case characters are allowed). In addition, a default value is assigned to each option, so at a minimum, it suffices to specify the molecular geometry and the basis sets to be used (please never do!). The input file may contain lines between keyword lines starting with '#' which are interpreted as comment lines and ignored in the input processing. In the following all input options to StoBe are described in detail including a brief introduction to the theory behind these options. 2.1.1. TABLES OF KEYWORDS (Goto TOC, KEYW, KEYA) There are currently about 60 keywords to control a StoBe calculation. The keywords together with their numerical parameters are explained in Sec. 2.2. The following two tables list all keywords (a) by logical grouping and (b) in alphabetical order. All keyword input is handled in subroutine DECODE called from the main CRAYCP driver. The list of keywords is contained in the character array OPT in the DECODE routine. ----------------------------------------------------------------------------- 2.1.1.1. Keywords by logical grouping (Goto TOC, KEYW, KEYA) ----------------------------------------------------------------------------- *** a. Geometry basics *** a.1) CARTesian - geometry input in cartesian coordinates a.2) ZMATrix - geometry input in internal coordinates a.3) NOSYmmetry - do not use symmetry a.4) SYMMetry - use of spatial symmetry and unrestricted calculations *** b. Electronic state *** b.1) MULTiplicity - spin multiplicity of N-electron state b.2) CHARge - total charge of the system b.3) POTEntial - type of Exc functional to be used b.4) NOFSsymmetry - do not symmetry block the KS matrix b.5) FSYMmetry - use symmetry blocking of KS matrix b.6) SMEAr - level occupation smearing b.7) CONF - define shell configuration in atom calculations b.8) EXCI - define excited hole states b.9) FIELd - apply external electric field b.10) SPINcontamination - calculate spin contamination b.11) REORder - reorder orbitals in restart b.12) LOCAlize - localize orbitals at specific atom centers *** c. SCF iteration *** c.1) RUNType - type of calculation c.2) SCFType - type of SCF algorithms (disk, memory, direct) c.3) MAXCycles - maximum number of SCF cycles c.4) ECONvergence - SCF energy convergence parameter c.5) DCONvergence - SCF electron density convergence parameter c.6) DMIXing - mixing scheme in extrapolation between SCF steps c.7) DIIS - direct iterative inversion (SCF extrapolation) c.8) LEVElshift - level shifting for improving SCF convergence c.9) MAXOverlap - maximum overlap orbital selection c.10) INTEgration - method for angular grid integration c.11) GRID - choice of grid for integration and fitting c.12) WEDGe - restrict angular grid of Gauss-Legendre integration c.13) SLATer - modify Slater radii used in numerical integration c.14) PTCHarges - include external point charges (no basis functions) c.15) BSSError - select centers without nucleus for BSSE evaluation c.16) SUPSym - restrict orbital mixings in SCF c.17) TRRH - trust-region Roothan-Hall extrapolation *** d. Geometry optimization *** d.1) GCONvergence - energy gradient convergence (geometry opt.) d.2) GSTEpsize - maximum step size (geometry optimizations) d.3) MAXGeometries - maximum number of geometry steps (optimization) d.4) COORdinate - constrain atoms in geometry optimization d.5) ATOM - constrain atoms in geometry opt. (fix, free, dir. move) d.6) GROUp - constrain groups of atoms in geometry opt. d.7) HESSian - method of Hessian matrix update (geometry optimization) d.8) GRADients - use Versluis correction in energy gradients d.9) STRBalsac - BALSAC structure plot at each geometry step *** e. Properties *** e.1) MULLiken - include Mulliken population analysis e.2) LOEWdin - include Loewdin population analysis e.3) MEPFit - calculate eff. atomic charges from MEPS fitting (n.a.) e.4) BALPopulation - orbital, Fukui, population output (BALSAC format) e.5) DRAW - output data for plotting MOs, MEPS, or other properties e.6) PSCRatch - evaluate polarizabilities by SCF scratch runs e.7) XRAYspectra - compute X-ray spectra (absorption, emission, RIXS) e.8) TOPOlogical - Topological charge/orbital analysis (Bader, etc.) e.9) NMROut - output for subsequent NMR calculations e.10) EPROut - output for subsequent EPR calculations e.11) VIBRations - vibrational analysis details e.12) NEBparameters - compute NEB paths e.13) RESInput - redirect restart input file units e.14) RESOutput - redirect restart output file units *** f. Miscellaneous keywords *** f.1) TITLe - title line of StoBe run f.2) PRINtout - extended print output f.3) SAVE - save restart file periodically f.4) ORBItalchoice - choice of orbital representation (d functions) f.5) ECPRead - use alternative one-electron hamiltonian f.6) PNTGenerator - gridpoint generator for angular integration f.7) ALCHem - geometry, basis, orbital output to ALCHEM format files f.8) FILEoutput - analysis file output, used for Molden, Molekel f.9) VIRTuals - include virtual orbitals in print output f.10) CTRLoption - interactive control parameter in/output (unit 7) f.11) DOSOutput - file output for DOS calculations (units 96/97) f.12) SHoRTrestartfile - short StoBe restart file output (unit 2) f.13) FORCes - evaluate atom forces after single point SCF run f.14) DYNAmics - short-time dynamics simulation f.15) GEOFormat - set format for geometry output files ----------------------------------------------------------------------------- ----------------------------------------------------------------------------- 2.1.1.2. Keywords in alphabetical order (Goto TOC, KEYW, KEYA) ----------------------------------------------------------------------------- ALCHem - geometry, basis, orbital output to ALCHEM format files ..... f.7) ATOM - constrain atoms in geometry opt. (fix, free, dir. move) ...... d.5) BALPopulation - orbital, Fukui, population output (BALSAC format) ... e.4) BSSError - select centers without nucleus for BSSE evaluation ....... c.15) CARTesian - geometry input in cartesian coordinates ................. a.1) CHARge - total charge of the system ................................. b.2) CONF - define shell configuration in atom calculations .............. b.7) COORdinate - constrain atoms in geometry optimization ............... d.4) CTRLoption - interactive control parameter in/output (unit 7) ....... f.10) DCONvergence - SCF electron density convergence parameter ........... c.5) DIIS - direct iterative inversion (SCF extrapolation) ............... c.7) DMIXing - mixing scheme in extrapolation between SCF steps .......... c.6) DOSOutput - file output for DOS calculations (units 96/97) .......... f.11) DRAW - output data for plotting MOs, MEPS, or other properties ...... e.5) DYNAmics - short-time dynamics simulation ........................... f.14) ECONvergence - SCF energy convergence parameter ..................... c.4) ECPRead - use alternative one-electron hamiltonian .................. f.5) EPROut - output for subsequent EPR calculations ..................... e.10) EXCI - define excited hole states ................................... b.8) FIELd - apply external electric field ............................... b.9) FILEoutput - analysis file output, used for Molden, Molekel ......... f.8) FORCes - evaluate atom forces after single point SCF run ............ f.13) FSYMmetry - use symmetry blocking of KS matrix ...................... b.5) GCONvergence - energy gradient convergence (geometry opt.) .......... d.1) GEOFormat - set format for geometry output files .................... f.15) GRADients - use Versluis correction in energy gradients ............. d.8) GRID - choice of grid for integration and fitting ................... c.11) GROUp - constrain groups of atoms in geometry opt. .................. d.6) GSTEpsize - maximum step size (geometry optimizations) .............. d.2) HESSian - method of Hessian matrix update (geometry optimization) ... d.7) INTEgration - method for angular grid integration ................... c.10) LEVElshift - level shifting for improving SCF convergence ........... c.8) LOCAlize - localize orbitals at specific atom centers ............... b.12) LOEWdin - include Loewdin population analysis ....................... e.2) MAXCycles - maximum number of SCF cycles ............................ c.3) MAXGeometries - maximum number of geometry steps (optimization) ..... d.3) MAXOverlap - maximum overlap orbital selection ...................... c.9) MEPFit - calculate eff. atomic charges from MEPS fitting (n.a.) ..... e.3) MULLiken - include Mulliken population analysis ..................... e.1) MULTiplicity - spin multiplicity of N-electron state ................ b.1) NMROut - output for subsequent NMR calculations ..................... e.9) NEBparameters - compute NEB paths ................................... e.12) NOFSsymmetry - do not symmetry block the KS matrix .................. b.4) NOSYmmetry - do not use symmetry ................................... a.3) ORBItalchoice - choice of orbital representation (d functions)....... f.4) PNTGenerator - gridpoint generator for angular integration .......... f.6) POTEntial - type of Exc functional to be used ....................... b.3) PRINtout - extended print output .................................... f.2) PSCRatch - evaluate polarizabilities by SCF scratch runs ............ e.6) PTCHarges - include external point charges (no basis functions) ..... c.14) REORder - reorder orbitals in restart ............................... b.11) RESInput - redirect restart input file units ........................ e.13) RESOutput - redirect restart output file units ...................... e.14) RUNType - type of calculation ....................................... c.1) SAVE - save restart file periodically ............................... f.3) SCFType - type of SCF algorithms (disk, memory, cpu intensive) ...... c.2) SHoRTrestartfile - short StoBe restart file output (unit 2) ......... f.12) SLATer - modify Slater radii used in integration scheme ............. c.13) SMEAr - level occupation smearing ................................... b.6) SPINcontamination - calculate spin contamination .................... b.10) STRBalsac - BALSAC structure plot at each geometry step ............. d.9) SUPSym - restrict orbital mixings ................................... c.16) SYMMetry - use of spatial symmetry, broken symmetry calculations .... a.4) TITLe - title line of StoBe run ..................................... f.1) TOPOlogical - Topological charge/orbital analysis (Bader, etc.) ..... e.8) TRRH - trust-region Roothan-Hall extrapolation ...................... c.17) VIBRations - vibrational analysis details ........................... e.11) VIRTuals - include virtual orbitals in print output ................. f.9) WEDGe - restrict angular grid of Gauss-Legendre integration ......... c.12) XRAYspectra - compute X-ray spectra (absorption, emission, RIXS) ... e.7) ZMATrix - geometry input in internal coordinates .................... a.2) ----------------------------------------------------------------------------- 2.2. DETAILED DESCRIPTION OF KEYWORDS (Goto TOC, KEYW, KEYA) There are currently about 60 keywords to define input parameters. Keywords are, in most cases, followed by numbers on the same line or on following lines. The keywords together with their numerical parameters are explained in detail by logical groups - Geometry basics - Electronic states - SCF iteration - Geometry optimization - Properties - Miscellaneous keywords - Untested keywords - Unclear and obsolete keywords in the following sections. In the following, the description of each main keyword starts with its calling sequence where the shortest acceptable word form (short notation) is shown by upper case characters. In the actual input, all keywords start with their short notation (in any mixture of upper and lower case characters) but may be longer. Main/option keywords, text or numbers are separated on the same line by (at least) one blank character. 2.2.1. KEYWORDS FOR GEOMETRY BASICS (Goto TOC, KEYW, KEYA) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) a.1) CARTesian [ANGstrom, BOHr] [nimg] ----------------------------------------------------------------------------- This keyword defines geometry input by cartesian coordinates. The keyword ANGstrom must be used if the geometry is defined in Angstroms. Otherwise, atomic units (Bohr) are used. Integer number nimg refers to multi-image input (which is essential for nudged-elastic-band (NEB) calculations, see c.1), e.12)) where different geometric structures of a cluster/molecule (so-called images) are treated at a common basis. Here nimg denoted the number of images to be given in the following (default is nimg = 1). After the keyword line CARTesian ..., each line defines one atom (I) of the system. The input format is Nm X Y Z ZZ Ngrid [Mass] where Nm = element symbol for atom I, followed by up to two numbers. NOTE that the element name (one or two characters) must be a valid name of the element table H, He, Li, Be, B, C, N, O, F, Ne, Na, Mg ( 1-12) Al, Si, P, S, Cl, Ar, K, Ca, Sc, Ti, V, Cr (13-24) Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Kr (25-36) Rb, Sr, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Cd (37-48) In, Sn, Sb, Te, I, Xe, Cs, Ba, La, Ce, Pr, Nd (49-60) Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Hf (61-72) Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl, Pb, Bi, Po (73-84) At, Rn, Fr, Ra, Ac, Th, Pa, U, Np, Pu, Am, Cm (75-96) Bk, Cf, Es, Fm, Md, No, Lr, Ku, Ha (97-105) The additional numbers permit you to distinguish different atoms of the same element, e.g. H1 H2 H3 ...H21. In addition, the atomic symbol Du refers to dummy atoms, i. e. atoms with basis sets but no nuclear charge. Note that dummy atoms are meaningful only for single point SCF and polarizability calculations (integration parameters are those of hydrogen, still testing!). X, Y, Z = x , y , z coordinates of atom I given in Angstrom or Bohr units, see line CARTesian ... above. ZZ = effective nuclear charge of atom I. This may be different from the atomic number if model core potentials are being used. Basis set superposition error (BSSE) corrections are calculated using ZZ=0.0 (set internally) for the respective atom centers while their basis sets are retained. Ngrid = number of radial grid points used in the evaluation of the exchange-correlation potentials and energy density of atom I. Mass = (optional) atomic mass of atom I, used to correct forces for translational/rotational invariance in geometry optimizations without symmetry. If no mass value appears, a default from an internal table is used. The atom definition list of each image is terminated by one line containing the keyword 'END'. The above list is repeated for multi-image input. Examples: >CARTESIAN >O 0.0000 0.00000 0.0000 8. 32 >H1 0.0000000 1.4499633 1.1214967 1. 32 >H2 0.0000000 -1.4499633 1.1214967 1. 32 >END >CARTESIAN 2 >O 0.0000 0.00000 0.0000 8. 32 >H1 0.0000000 1.4499633 1.1214967 1. 32 >H2 0.0000000 -1.4499633 1.1214967 1. 32 >END >O 0.0000 0.00000 0.2000 8. 32 >H1 0.0000000 1.4499633 1.1214967 1. 32 >H2 0.0000000 -1.4499633 1.1214967 1. 32 >END >CARTESIAN ANGSTROM >O 0.0000 0.00000 0.0000 8. 32 >H1 0.0000000 0.7672872 0.5934703 1. 32 >H2 0.0000000 -0.7672872 0.5934703 1. 32 >END >CARTESIAN ANGSTROM 2 >O 0.0000 0.00000 0.0000 8. 32 >H1 0.0000000 0.7672872 0.5934703 1. 32 >H2 0.0000000 -0.7672872 0.5934703 1. 32 >END >O 0.0000 0.00000 0.2000 8. 32 >H1 0.0000000 0.7672872 0.5934703 1. 32 >H2 0.0000000 -0.7672872 0.5934703 1. 32 >END ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) a.2) ZMATrix [ANGstrom] ----------------------------------------------------------------------------- This keyword defines geometry input by internal coordinates based on the Z-matrix notation. The ANGSTROM option is chosen if the geometry is defined in Angstroms. Otherwise, atomic units (Bohr) are used. After the keyword, each line defines one atom (I) of the system. The line input format is (atom 1) Nm ZZ Ngrid [Mass] (atom 2) Nm J R(I,J) ZZ Ngrid [Mass] (atom 3) Nm J R(I,J) K ALPHA(I,J,K) ZZ Ngrid [Mass] (atom >3) Nm J R(I,J) K ALPHA(I,J,K) L PHI(I,J,K,L) ZZ Ngrid [Mass] where Nm = element symbol for atom I, followed by up to two numbers. The definition is identical to that of keyword CARTesian, a.1) J = index of previous atom J R(I,J) = bond distance between atoms I, J K = index of previous atom K (.ne. J) ALPHA(I,J,K) = bond angle between atoms I, J, K (angle I - J - K ) L = index of previous atom L (.ne. J, .ne. K) PHI(I,J,K,L) = dihedral angle formed by atoms I, J, K, L (angle between planes through J,K,I and through J,K,L: PHI > 0 for clockwise rotation of plane J,K,I towards J,K,L about J - K, else PHI < 0). ZZ = effective nuclear charge of atom I. This may be different from the atomic number if model core potentials are being used. Basis set superposition error (BSSE) corrections are calculated using ZZ=0.0 (set internally) for the respective atom centers while their basis sets are retained. Ngrid = number of radial grid points used for the least-squares fit of the exchange-correlation potentials and energy density of atom I. Mass = (optional) atomic mass of atom I, used to correct forces for translational/rotational invariance in geometry optimizations without symmetry. If no mass value appears, a default from an internal table is used. The atom definition list is terminated by one line containing the keyword END. The Z-matrix notation is converted internally to cartesian coordinates where - atom # 1 defines the origin, i. e. R1 = (0, 0, 0) - atom # 2 points along the (positive) z axis, i. e. R2 = (0, 0, R(1,2)) - atom # 3 lies in the xz plane with x > 0, i. e. R3 = (x, 0, y) - atoms # p>3 are defined accordingly Examples: (a) H2O molecule >ZMATRIX ANGSTROM >O 8. 32 >H 1 0.97 1. 32 >H 1 0.97 2 105.5 1. 32 >END (b) NH3 molecule >ZMATRIX BOHR >N 7. 32 >H 1 1.9419 1. 32 >H 1 1.9419 2 107.06 1. 32 >H 1 1.9419 2 107.06 3 114.54 1. 32 >END corresponding to >CARTESIAN BOHR >N 0.00000000 0.00000000 0.00000000 7. 32 >H 0.00000000 0.00000000 1.94190000 1. 32 >H 1.85645263 0.00000000 -0.56970100 1. 32 >H -0.77103753 1.68876212 -0.56970100 1. 32 >END Note, that option ZMATrix can be used only if no symmetry is applied to coordinates, see keyword NOSYm in a.3). Further, the Z-matrix format does not support multi-image input available with cartesian coordinates, see a.1). ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) a.3) NOSYmmetry [UNREstricted] ----------------------------------------------------------------------------- This keyword excludes the use of spatial symmetry in the calculations. The keyword must be specified BEFORE the coordinate input and coordinates of ALL atoms have to be given. The (optional) keyword UNREstricted forces a spin-unrestricted calculation even in cases where the number of alpha and beta spin orbitals is identical. Examples: >NOSYMMETRY >NOSYMMETRY UNRESTRICTED For more details on the use of symmetry, see also keywords SYMMetry, FSYMmmetry, NOFSymmetry. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) a.4) SYMMetry symlabel [UNREstricted, NOSYmmetry] ----------------------------------------------------------------------------- This keyword allows the use of spatial symmetry in the calculations to speed up the calculations [DGS93]. With symmetry, the numerical integration is performed only for symmetry non-equivalent atoms and only non-equivalent atoms can appear in the coordinate input, see keyword CARTesian. Therefore, the keyword line 'SYMMetry ...' must be specified BEFORE the coordinate input and only coordinates (and orbital/auxiliary basis sets) of symmetry non-equivalent atoms have to be included. At present, the following 48 point symmetry groups are supported (see also file symbasis): 1. C1 2. Ci 3. Cs 4. Csxz 5. Csy 6. Csyz 7. Csx 8. Csxy 9. Csz 10. C2 11. C3 12. C4 13. C5 14. C6 15. S4 16. S6 17. D2 18. D3 19. D4 20. D5 21. D6 22. C2v 23. C2vb 24. C3v 25. C4v 26. C5v 27. C6v 28. C2h 29. C3h 30. C4h 31. C5h 32. C6h 33. D2h 34. D3h 35. D4h 36. D5h 37. D6h 38. D2d 39. D3d 40. D4d 41. D5d 42. D6d 43. T 44. Th 45. Td 46. O 47. Oh 48. ATOM Symmetry label ATOM which denotes the full rotational point group (up to d functions) is restricted to free atoms. For further definitions see also notes on symmetries, section 2.4. Note that the above given symmetry labels can be given by mixed case (e.g. C4v) or only upper case characters (e.g. C4V), see examples below. Note that inside StoBe spatial coincidence of atoms as a result of symmetrization is detected and corrected accordingly. Here "coincidence" is defined by an interatomic distance < 0.001 atomic units (0.000529177 Angstrom). All atom centers at distances > 0.001 au will be treated as separate centers. The (optional) keywords UNREstricted forces a spin-unrestricted calculation even in cases where the number of alpha and beta spin orbitals in each symmetry representation is identical. NOSYmmetry applies the symmetry group only while generating atom centers of a cluster/molecule whereas all SCF calculations (single point, geometry optimizations, vibrations) are carried out without symmetry (C1), requiring input parameters which refer to C1. Note that with this option all basis sets (orbital, auxiliary, model potentials) have to be provided only for symmetry non-eqivalent centers, see example m) below. Examples: a) H2O, no symmetry >SYMM C1 or >NOSYMM >CARTESIAN ANGSTROM >O 0.0 0.0 0.0 8.0 35 >H 0.0 0.763 0.5861 1.0 32 >H 0.0 -0.763 0.5861 1.0 32 b) H2O, C2v symmetry >SYMM C2v >CARTESIAN ANGSTROM >O 0.0 0.0 0.0 8.0 35 >H 0.0 0.763 0.5861 1.0 32 c) O2 (other homonuclear diatomics), D2h symmetry >SYMM D2h >CARTESIAN ANGS > O 0.0 0.0 0.61 8.0 35 d) HF (other heteronuclear diatomics), C4v symmetry >SYMM C4V >CARTESIAN ANGS >H 0.0 0.0 0.000 1.0 32 >F 0.0 0.0 0.916 9.0 35 e) NH3, C3v symmetry >SYMM C3V >CARTESIAN ANGSTROM > N 0.000 0.0000 0.0677 7.0 35 > H 0.9406 0.0000 -0.2993 1.0 32 f) Pd3, D3h symmetry >SYMM D3h >CART ANGSTRO > PD 2.38 0.0 0.0 46.0 42 g) C2H6, D3h symmetry >SYMM D3H >CARTESIAN BOHR > C 0.0 0.00000 1.465 6.0 35 > H 1.915 0.0000 2.215 1.0 32 h) C6H6, D6H symmetry >SYMM D6H >CARTESIAN BOHR > C 0.00000 2.6518633 0.000 6.0 35 > H 0.00000 4.6936720 0.000 1.0 32 i) CH4, Td symmetry >SYMM TD >CARTE ANGSTR >C 0.0 0.0 0.0 6.0 35 >H 0.633 0.633 0.633 1.0 32 j) CH4, Td symmetry, spin-unrestricted calculation >SYMM TD UNRESTRICTED >CARTE ANGSTR >C 0.0 0.0 0.0 6.0 35 >H 0.633 0.633 0.633 1.0 32 k) Pd8, Oh symmetry >SYMM OH >CART ANGSTRO > PD 2.38 2.38 2.38 46.0 42 l) Ni atom, full rotational atom symmetry >SYMM ATOM >CART ANGSTRO > Ni 0.000 0.000 0.000 28.0 32 m) NH3 molecule, C1 symmetry but coordinate and basis set input with C3v symmetry >SYMM C3V NOSYMM >CARTESIAN BOHR > N 0.000000 0.000000 0.000000 7. 32 > H 1.787874 0.000000 0.734852 1. 32 >END > ... >END >A-NITROGEN (4,4;4,4) >A-HYDROGEN (4,0;4,0) diffuse >O-NITROGEN (631/31/1) >O-HYDROGEN (41/1*/1+) (TZVP-FIP2) >END 2.2.2. KEYWORDS FOR ELECTRONIC STATES (Goto TOC, KEYW, KEYA) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.1) MULTiplicity [number] ----------------------------------------------------------------------------- Spin multiplicity of the N-electron state. Numbers 1, 2, 3, etc. refer to singlet, doublet, triplet, etc. states. By default, a singlet state is assumed (leading to an error message for systems where the total number of electrons is odd). The spin multiplicity together with the net charge, see CHARge, determines the total number of spin alpha and beta orbitals to be occupied and the multiplicity / charge values are checked for consistency. Note: Since only the occupations are given, the multiplicity is only formally correct for high-spin states, i.e. when in each symmetry the number of alpha(beta) electrons exceeds that of the other spin. Otherwise the resulting spin state is a mixture of the possible spin states which have a projection equal to (MULT-1)/2. In practice the formalism thus corresponds to a single-determinant Hartree-Fock with regards to spin-coupling. Example (triplet state): >MULTIPLICITY 3 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.2) CHARge [number] ----------------------------------------------------------------------------- Net charge of the system, positive (negative) for cations (anions). By default, the system is assumed neutral (number = 0). The net charge together with the spin multiplicity, see MULTiplicity, determines the total number of spin alpha and beta orbitals to be occupied and the multiplicity / charge values are checked for consistency. Example (singly charged cation): >CHARGE 1 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.3) POTEntial [FTYP] [EXCH] [CORR] ----------------------------------------------------------------------------- The exchange-correlation functional used in the calculations is defined by three keywords characterizing the - overall type of functional FTYP (LOCAl, NONLocal, MIXEd), - the exchange part EXCH (NORMal, PD86, PD91, PVS1, BE88, BECKe, DEPK, BE86, BA86, PBE, PBE2, RPBE), and - the correlation part CORR (XALPha, VWN, PD91, PVS1, PD86, PD91, LAP, PBE). All meaningful combinations are given in the following table, together with the corresponding pointers (NLCL, NLEX and NLEC) in the code. ___________________________________________________________________ | local | nonlocal | mixed | |---------------------|---------------------|---------------------| | NLCL=1 | NLCL=2 | NLCL=3 | |---------------------|---------------------|---------------------| |exch |nlex|corr |nlec|exch |nlex|corr |nlec|exch |nlex|corr |nlec| |----------|----------|----------|----------|----------|----------| | |XALP | 0 | |XALP | 0 | |XALP | 0 | |NORM | 1 |VWN | 1 |PD86 | 1 |PD86 | 1 |PD86 | 1 |PD86 | 1 | |PVS1 | 2 |PD91 | 2 |BE88 | 2 |PD91 | 2 |BE88 | 2 |PD91 | 2 | | |PVS1 | 3 |PD91 | 3 |PVS1 | 3 |PD91 | 3 |PVS1 | 3 | | |VBH | 4 |BA86 | 4 |LAP | 4 |BA86 | 4 |LAP | 4 | | |DEPK | 5 |PBE | 5 |DEPK | 5 |PBE | 5 | | |BE86 | 6 | |BE86 | 6 | | | |PBE | 7 | |PBE | 7 | | | |PBE2 | 8 | |PBE2 | 8 | | | |RPBE | 9 | |RPBE | 9 | | ------------------------------------------------------------------- NOTE: For backwards compatibility keyword "BE88" can be replaced by "BECK". The various choices for exchange and correlation parts have the following meaning: (A) Exchange part EXCH (LOCAl) NORMal - standard local spin density exchange [D30] (LOCAl) PVS1 - modified gradientless SD exchange [PVS94]. (NONLocal) BE88 - GGA exchange functional of Becke [BE88]. (NONLocal) BECK - identical to BE88 (for backwards compatibility). (NONLocal) PD86 - GGA exchange of Perdew and Wang [PW86]. (NONLocal) PD91 - improved GGA exchange functional of Perdew / Wang [P91]. (NONLocal) PBE - GGA exchange of Perdew / Burke/ Ernzerhof [PBE96]. (NONLocal) PBE2 - revised GGA exchange of Perdew / Burke/ Ernzerhof [ZY98], using different fit parameter choices. (NONLocal) RPBE - alternative revised PBE exchange according to Hammer et al. [HHN99]. (NONLocal) BE86 - GGA exchange version of Becke as of 1986 [B86]. (NONLocal) BA86 - GGA exchange version of Becke as of 1986 (alt.) [B86a]. (NONLocal) DEPK - GGA exchange by Daul [GDZ86]. (B) Correlation part CORR (LOCAL) XALP - No explicit correlation, included for historic reasons only. (LOCAL) VWN - LSD correlation functional by Vosko, Wilk and Nusair [VWN80]. (LOCAL) VBH - LSD correlation functional by v. Barth, Hedin and others [BH72, HL71, GL76]. (LOCAL) PD91 - Accurate analytical representation of the homogeneous gas correlation energy by Perdew and Wang [PW91]. (LOCAL) PVS1 - Gradientless correlation functional based on (inhomogeneous) pair-correlation function of Colle-Salvetti type [PVS94]. (NONLocal) PD86 - GGA correlation functional of Perdew [P86]. (NONLocal) PD91 - New GGA correlation functional of Perdew and Wang [P91]. (NONLocal) LAP - Nonlocal generalization of the correlation functional PVS1 involving the kinetic energy density and the Laplacian of the electron density [PRV95]. (NONLocal) PBE - GGA correlation of Perdew / Burke/ Ernzerhof [PBE96]. The exchange/correlation combinations tested in detail are POTENTIAL LOCAL NORM VWN (default functional) POTENTIAL LOCAL PVS1 PVS1 POTENTIAL NONLOCAL PD86 LAP POTENTIAL NONLOCAL BE88 LAP POTENTIAL NONLOCAL PBE PBE POTENTIAL NONLOCAL PBE2 PBE POTENTIAL NONLOCAL RPBE PBE POTENTIAL MIXED PD86 LAP POTENTIAL MIXED BE88 LAP. Some choices involving the options PVS1 and LAP are yet not debugged and should not be used at present. In particular, for the LAP correlation option, at present only the following parameter choices are meaningful 1. BLAP1 (does not involve any parallel spin correlation beyond Ex) >POTENTIAL NONLOCAL BE88 LAP 0.197 1.255 1.48 0.00 2. PLAP1 (similar to BLAP1 in essence ) >POTENTIAL NONLOCAL PD86 LAP 0.197 1.255 1.48 0.00 3. BLAP3 (includes some parallel spin correlation beyond Ex) >POTENTIAL NONLOCAL BE88 LAP 0.197 1.276 1.477 0.04 4. PLAP3 (similar to BLAP3) >POTENTIAL NONLOCAL PD86 LAP 0.197 1.26 1.48 0.01 5. PLAP4 (similar to PLAP3 but with a refined share of the parallel spin correlation, so as to achieve optimal results for weak interactions) >POTENTIAL NONLOCAL PD86 LAP 0.1962 1.2557 1.48 0.003 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.4) NOFSymmetry [AUFBau, SCFOccupation] ----------------------------------------------------------------------------- The StoBe code uses symmetry in two ways: to generate the equivalent centers and only treat the non-equivalent centers in the numerical integration of the exchange-correlation potentials and to generate a symmetry-adapted basis for the expansion of the Kohn-Sham orbitals. Using symmetry on the grid can substantially reduce the time for each iteration, while symmetrizing the orbitals results in a block-diagonal Kohn-Sham matrix which improves convergence and leads to more easily identified symmetry orbitals. With the NOFSymmetry keyword, block diagonalization of the Kohn-Sham matrix according to molecular symmetry, see keyword FSYMmetry, is not used while the symmetry may still be used in the numerical integration, see keyword SYMMetry. NOTE that in atom calculations with full rotational symmetry and explicit shell occupation by option CONFig, see b.7), this option is ignored, see also b.5). The keyword line NOFSsymetry AUFBau [ALphabeta] inhibits symmetry-based block-diagonalization of the Kohn-Sham matrix and determines the occupation of orbitals by the Aufbau principle (orbitals are filled with increasing KS 1-electron energy up to the HOMO). The Aufbau principle can be used together with the level smear option, see keyword SMEAr. The optional keyword ALphabeta does an overall aufbau using both alpha and beta spin orbitals together. This procedure does not necessarily preserve the total spin between iterations. NOFSymmetry SCFOccupation [EXCIted] inhibits symmetry based block diagonalization of the Kohn-Sham matrix and defines a fixed occupation scheme. The scheme on new lines as ALFA nocca leva na1 occa1 na2 occa2 ... nleva occleva BETA noccb levb nb1 occb1 nb2 occb2 ... nlebv occlevb where on the line starting with 'ALFA', nocca is the total number of alpha orbitals to be occupied (including those whose occupations are explicitly given!) and leva denotes the number of alpha orbitals (< 10) whose occupations are to be given explicitly, followed by leva pairs of numbers (nai, occai) defining the orbital index nai and the orbital occupation occai. The following line starting with 'BETA' contains the analogous input for beta orbitals. Both 'ALFA' and 'BETA' lines are required in the input. The occupation numbers are kept fixed in the SCF calculation irrespective of the Aufbau principle (which may result in an excited state or in bad convergence). Note that the definition of the numbers following the 'ALFA' and 'BETA' keywords are somewhat different when used for atoms and with symmetry, see below. Fixed occupations are incompatible with the level smear option, see keyword SMEAr. An excited configuration (within the single KS determinant representation) can also be specified using the optional keyword EXCIted on the keyword line. This requires additional input lines where for the Kohn-Sham matrix all orbitals for which the occupations will be specified are given as: SYM 1 ALFA 0 na num1 occ1 num2 occ2 ... numna occna BETA 0 nb num1 occ1 num2 occ2 ... numnb occnb END If na is greater than 9 then the input is simply continued as a numerical input on following lines (10 orbital specifications i.e. sets of (num,occ) to a line). Note that all orbitals occurring in the global occupation scheme (see above) but not specified explicitly by occupations retain an occupation of 1. The excited configuration scheme can also be used to evaluate Fukui functions and indices [PY84,PY89] within the finite difference approximation. This requires a restart from an initial configuration (with or without excitations), see c.1), where an artificial excitation is introduced in addition by changing the occupation of a specific orbital by a small amount (e.g. 0.01 electrons, Fukui charge). A comparison of the self-consistent solution with Fukui charge with the initial restart input is used to evaluate Fukui functions and indices. The Fukui charge excitation is defined by a keyword line FUKUi [ALFA, BETA] qfukui norb iorb(1) worb(1) iorb(2) worb(2) ... (after the NOFSymmetry ... keyword line) which includes parameters ALFA, BETA determines whether the charge addition/subtraction is to occur in a spin alpha or beta orbital. Default is spin alpha. qfukui actual charge to be included in the calculation. Positive values add charge to an unoccupied orbital (LUMO in most cases) referring to f+(r), nucleophilic attack. Negative values remove charge from an occupied orbital (HOMO in most cases) referring to f-(r), electrophilic attack. A meaningful value for f+/- is +/-0.01. norb number of orbitals involved in the calculation. All norb orbitals to be defined in the following are varied in their charge where the total (summed) charge amounts to qfukui. Up to 20 orbitals (up to 5 per line, where continuation lines simply continue with 'iorb(n) worb(n) up to iorb(norb) occ(norb)...') can be included. iorb(i) index of each orbital occupation to be changed in the Fukui analysis. (Index iorb denotes the position in the restart file.) For qfukui < 0 each orbital must be fully or partially occupied in the initial input configuration (including usually the HOMO). For qfukui > 0 each orbital must be empty or partially occupied in the initial input configuration (including usually the LUMO). worb(i) relative weight of each orbital whose occupation is to be changed in the Fukui analysis. The sum of all weights, W, is determined and the charge q(i) of each orbital is calculated as qfukui*worb(i)/W. The Fukui input can be mixed with the above excitation definitions but must appear always after all excitation definitions. The complete configuration definition of the Fukui scheme must be terminated by a keyword line 'END'. NOTE: If combined with specific excitations using keyword EXCIted, then the END statement should appear only after both excitation and Fukui input are complete. The Fukui analysis is included in the output after the population analysis (if selected) where atom-projected Fukui indices are listed for both relaxed and frozen orbitals (in the latter case orbitals of the restart input file are used). In addition, both Fukui matrices are saved on unit 33 using format Record 1 (6I4) NFUKUI Spin selection ( = 1 for alpha, = 2 for beta) QFUKUI Charge used to evaluate f+/- matrix ISYM,IORB Symmetry and sequence index of orbital used for analysis, here ISYM = 1. NCNTRT Dimension of Fukui matrix NCNTSQ = (NCNTRT+1)*NCNTRT/2 Record 2 (NCNTSQ*R8) Fukui matrix elements (relaxed values) in tridiagonal form Record 3 (NCNTSQ*R8) Fukui matrix elements (frozen orbital values) in tridiagonal form Further, the DRAW option allows to plot contour lines of the Fukui function, see e.5). The Fukui option can be used also in a properties-only run (parameter CPROp, see c.1) if output unit 33 from a previous Fukui run is provided as input. Examples of molecules: a) H2O, C2v symmetry, no block-diagonal (i.e. no symmetrization) KS matrix and Aufbau principle occupation >TITLE >H2O with symmetry but no block diag. KS matrix, Aufbau principle >SYMM C2V >CART BOHR > O 0.0 0.0 0.0 8.0 35 > H 1.4596 0.0 1.1354 1.0 32 >END >RUNTYPE STARTUP NO-OPT >MULTIPLICITY 1 >CHARGE 0.0 >NOFSYM AUFBAU >DMIX MDENS 0.45 >DIIS ON >END >A-OXYGEN (5,2;5,2) >A-HYDROGEN (5,1;5,1) >O-OXYGEN (5211/411/1) >O-HYDROGEN (41/1) >END b) H2O, C2v symmetry, no block-diagonal KS matrix and fixed occupation (taken from the output of the previous run). >TITLE >H2O with symmetry and no block-diag. KS matrix, fixed occupations >SYMM C2V >CART BOHR > O 0.0 0.0 0.0 8.0 35 > H 1.4596 0.0 1.1354 1.0 32 >END >RUNTYPE STARTUP NO-OPT >MULTIPLICITY 1 >CHARGE 0.0 >NOFSYM SCFOCCUP >ALFA 5 0 >BETA 5 0 >DMIX MDENS 0.45 >END >A-OXYGEN (5,2;5,2) >A-HYDROGEN (5,1;5,1) >O-OXYGEN (5211/411/1) >O-HYDROGEN (41/1) >END c) H2O, C2v symmetry, no block-diagonal KS matrix and fixed occupation including a HOMO-LUMO excitation (singlet state) with fractional occupancy of 0.5 (Slater transition state method). >TITLE >H2O with symmetry and no block-diag. KS matrix, fixed occupations >WATER HOMO-LUMO excitation with Slater TSM >SYMM C2V >CART BOHR > O 0.0 0.0 0.0 8.0 35 > H 1.4596 0.0 1.1354 1.0 32 >END >RUNTYPE STARTUP NO-OPT >MULTIPLICITY 1 >CHARGE 0.0 >MAXCYC 35 >NOFSYM SCFOCCUP >ALFA 6 2 5 0.5 6 0.5 >BETA 5 0 >DMIX MDENS 0.45 >DIIS ON >END >A-OXYGEN (5,2;5,2) >A-HYDROGEN (5,1;5,1) >O-OXYGEN (5211/411/1) >O-HYDROGEN (41/1) >END d) An alternative to the above using EXCITED keyword with SCFOCC: H2O, C2v symmetry, no block-diagonal KS matrix and fixed occupation including a HOMO-LUMO excitation (singlet state) with fractional occupancy of 0.5 (Slater transition state method). >TITLE >H2O with symmetry and no block-diag. KS matrix, fixed occupations >WATER HOMO-LUMO excitation with Slater TSM >SYMM C2V >CART BOHR > O 0.0 0.0 0.0 8.0 35 > H 1.4596 0.0 1.1354 1.0 32 >END >RUNTYPE STARTUP NO-OPT >MULTIPLICITY 1 >CHARGE 0.0 >MAXCYC 35 >NOFSYM SCFOCCUP EXCITED >ALFA 6 >BETA 5 >SYM 1 >ALFA 0 2 5 0.5 6 0.5 >BETA 0 0 >END >DMIX MDENS 0.45 >DIIS ON >END >A-OXYGEN (5,2;5,2) >A-HYDROGEN (5,1;5,1) >O-OXYGEN (5211/411/1) >O-HYDROGEN (41/1) >END e) An evaluation of atom-projected Fukui indices (f-, electrophilic indices) of CO. Restart from neutral CO with occupation (7,7) >TITLE >CO molecule ground state, Fukui charge -0.01 >NOSYMM >CARTESIAN BOHR > C 0.0000 0.00000 0.00000 6. 32 > O 0.0000 0.00000 2.15000 8. 32 >END >RUNTYPE RESTART NEWOCC >SCFTYPE DIRECT >POTENTIAL NONLOCAL PD91 PD91 >GRID FINE NONRANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 50 >ECONVERGENCE .0000001 >DCONVERGENCE .0000001 >DMIXING MDENS 0.40 >DIIS NEW 7 >NOFSYM SCFOCCUP EXCITED >ALFA 7 >BETA 7 >FUKUI ALFA -.01 1 7 1.0 >END >VIRT ALL >MULL ON FULL >ORBI 5D >END >A-CARBON (4,4;4,4) >A-OXYGEN (4,4;4,4) >O-CARBON (631/31/1) >O-OXYGEN (631/31/1) >END f) An evaluation of atom-projected Fukui indices (f+, nucleophilic indices) of CO. Restart from neutral CO with occupation (7,7) >TITLE >CO molecule ground state, Fukui charge +0.01 >NOSYMM >CARTESIAN BOHR > C 0.0000 0.00000 0.00000 6. 32 > O 0.0000 0.00000 2.15000 8. 32 >END >RUNTYPE RESTART NEWOCC >SCFTYPE DIRECT >POTENTIAL NONLOCAL PD91 PD91 >GRID FINE NONRANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 50 >ECONVERGENCE .0000001 >DCONVERGENCE .0000001 >DMIXING MDENS 0.40 >DIIS NEW 7 >NOFSYM SCFOCCUP EXCITED >ALFA 8 >BETA 7 >SYM 1 >ALFA 0 1 8 0.0 >BETA 0 0 >FUKUI ALFA .01 1 8 1.0 >END >VIRT ALL >MULL ON FULL >ORBI 5D >END >A-CARBON (4,4;4,4) >A-OXYGEN (4,4;4,4) >O-CARBON (631/31/1) >O-OXYGEN (631/31/1) >END Examples of atoms: a) C atom, no block-diagonal KS matrix and Aufbau principle occupation, "broken symmetry" (s and d-sigma can mix) s,p,d orbitals with non-spherical electron density. >TITLE >C nonspherical atom with NOFSYM >SYMM D2H >CARTESIAN BOHR > C 0.00000 0.00000 0.00000 6.0 35 >END >RUNTYPE STARTUP NO-OPTI >GRID FINE RANDOM >MULTIPLICITY 3 >CHARGE 0. >DMIXING MDENS 0.3 >DIIS ON >NOFSYM AUFBAU >END >A-CARBON (5,2;5,2) >O-CARBON (621/41/1*) >END ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.5) FSYMmetry [AUFBau, SCFOccupation] ----------------------------------------------------------------------------- The StoBe code uses symmetry in two ways: to generate the equivalent centers and only treat the non-equivalent centers in the numerical integration of the exchange-correlation potentials and to generate a symmetry-adapted basis for the expansion of the Kohn-Sham orbitals. Using symmetry on the grid can substantially reduce the time for each iteration, while using the keyword FSYM to symmetrize the orbitals results in a block-diagonal Kohn-Sham matrix which improves convergence and leads to more easily identified symmetry orbitals and facilitates the analysis of the resulting molecular or atomic state. Block-diagonalization of the Kohn-Sham matrix may lead to small increases in the total energy compared to the case without symmetry-blocking (keyword NOFSym) depending on whether or not orbital symmetry-breaking occurs in the case without symmetry-blocking. The keyword line FSYMmetry AUFBau [ALphabeta] (default setting) uses symmetry-based block-diagonalization of the Kohn-Sham matrix and determines the occupation of orbitals by the Aufbau principle (orbitals are filled with increasing KS 1-electron energy up to the HOMO). Occupations are allowed to change between different symmetry representations in the course of an SCF iteration which is indicated in the print output. The Aufbau principle can be used together with the level smear option, see keyword SMEAr. NOTE that in atom calculations with full rotational symmetry and explicit shell occupation by option CONFig, see b.7), keywords AUFBau and ALphabeta are ignored and corresponding fixed occupations are used, see below. The optional keyword ALphabeta does an overall aufbau using both alpha and beta spin orbitals together. This procedure does not necessarily preserve the total spin between iterations. FSYMmetry SCFOccupation [EXCIted] uses symmetry-based block-diagonalization of the Kohn-Sham matrix and defines a fixed occupation scheme by giving the number of occupied alpha and beta orbitals for each symmetry representation. The occupation information starts on new lines as ALFA na1 na2 na3 ... nansym BETA nb1 nb2 nb3 ... nbnsym where nai, nbi are the occupation numbers of alpha and beta orbitals of each symmetry representation i, i = 1 ... nsym (including orbitals whose occupations are given explicitly, see below!). These occupation numbers are kept fixed in the SCF calculation irrespective of the Aufbau principle. Note that the definition of the numbers following the 'ALFA' and 'BETA' keywords are somewhat different when used for atoms, see below. Fixed occupations are incompatible with the level smear option, see keyword SMEAr. Starting from the above occupation scheme, excited configurations (within the single KS determinant representation) can be defined. Here the occupation of orbitals in each symmetry representation can be defined individually. Examples are excited states via the Slater transition state method (fractional occupations are allowed) or broken symmetry cases (see option keyword UNREstricted). This option requires the optional keyword EXCIted on the keyword line and additional lines of input as follows SYMmetry isym ALFA idum leva na1 occa1 na2 occa2 ... nleva occleva BETA idum levb nb1 occb1 nb2 occb2 ... nlebv occlevb where the line starting with 'SYMmetry' defines the index isym of a symmetry representation. On the following line starting with 'ALFA', leva denotes the number of alpha orbitals (< 10) in representation isym whose occupations are to be given explicitly, followed by leva pairs of numbers (na1, occai) defining the orbital index nai and the orbital occupation occai. The following line starting with 'BETA' contains the analogous input for beta orbitals. (Parameters idum are dummies left for compatibility reasons.) Both 'ALFA' and 'BETA' lines are required in the input. The occupation information lines can be repeated if more than 9 alpha or beta orbital occupations are to be defined. The complete configuration definition of excited states must be terminated by a keyword line 'END'. Note that all symmetry orbitals occurring in the global occupation scheme (see above) but not specified explicitly by occupations retain an occupation of 1. The excited configuration scheme can also be used to evaluate Fukui functions and indices [PY84,PY89] within the finite difference approximation. This requires a restart from an initial configuration (with or without excitations), see c.1), where an artificial excitation is introduced in addition by changing the occupation of a specific orbital by a small amount (e.g. 0.01 electrons, Fukui charge). A comparison of the self-consistent solution with Fukui charge with the initial restart input is used to evaluate Fukui functions and indices. The Fukui charge excitation is defined by a keyword line FUKUi [ALFA, BETA] qfukui norb isym(1) iorb(1) worb(1) isym(2) ... (after the FSYM ... keyword line) which includes parameters ALFA, BETA determines whether the charge addition/subtraction is to occur in a spin alpha or beta orbital. Default is spin alpha. qfukui actual charge to be included in the calculation. Positive values add charge to an unoccupied orbital (LUMO in most cases) referring to f+(r), nucleophilic attack. Negative values remove charge from an occupied orbital (HOMO in most cases) referring to f-(r), electrophilic attack. A meaningful value for f+/- is +/-0.01. norb number of orbitals involved in the calculation. All norb orbitals to be defined in the following are varied in their charge where the total (summed) charge amounts to qfukui. Up to 20 orbitals (up to 5 per line, where continuation lines simply continue with 'iorb(n) worb(n) up to iorb(norb) occ(norb)...') can be included. isym(i), iorb(i) symmetry and orbital index of each orbital whose occupation is to be changed in the Fukui analysis. (Indices isym/iorb refer to the position in the restart file.) For qfukui < 0 each orbital must be fully or partially occupied in the initial input configuration (including usually the HOMO). For qfukui > 0 each orbital must be empty or partially occupied in the initial input configuration (including usually the LUMO). worb(i) relative weight of each orbital whose occupation is to be changed in the Fukui analysis. The sum of all weights, W, is determined and the charge q(i) of each orbital is calculated as qfukui*worb(i)/W. The Fukui input can be mixed with the above excitation definitions but must appear always after all excitation definitions. The complete configuration definition of the Fukui scheme must be terminated by a keyword line 'END'. NOTE: If combined with specific excitations using keyword EXCIted, then the END statement should appear only after both excitation and Fukui input are complete. The Fukui analysis is included in the output after the population analysis (if selected) where atom-projected Fukui indices are listed both relaxed and frozen orbitals (in the latter case orbitals of the restart input file are used). In addition, both Fukui matrices are saved on unit 33 using format Record 1 (6I4) NFUKUI Spin selection ( = 1 for alpha, = 2 for beta) QFUKUI Charge used to evaluate f+/- matrix ISYM,IORB Symmetry and sequence index of orbital used for analysis NCNTRT Dimension of Fukui matrix NCNTSQ = (NCNTRT+1)*NCNTRT/2 Record 2 (NCNTSQ*R8) Fukui matrix elements (relaxed values) in tridiagonal form Record 3 (NCNTSQ*R8) Fukui matrix elements (frozen orbital values) in tridiagonal form Further, the DRAW option allows to plot contour lines of the Fukui function, see e.5). The Fukui option can be used also in a properties-only run (parameter CPROp, see c.1) if output unit 33 from a previous Fukui run is provided as input. Examples of molecules: a) H2O, C2v symmetry with block-diagonalization (i.e. symmetrization) of the KS matrix and Aufbau principle occupation >TITLE >H2O with symmetry and block-diag. KS matrix, Aufbau principle >SYMM C2V >CART BOHR > O 0.0 0.0 0.0 8.0 35 > H 1.4596 0.0 1.1354 1.0 32 >END >RUNTYPE STARTUP NO-OPT >MULTIPLICITY 1 >CHARGE 0.0 >FSYM AUFBAU >DMIX MDENS 0.45 >DIIS ON >END >A-OXYGEN (5,2;5,2) >A-HYDROGEN (5,1;5,1) >O-OXYGEN (5211/411/1) >O-HYDROGEN (41/1) >END b) H2O, C2v symmetry with block-diagonalization of the KS matrix and fixed occupation (taken from the output of the previous run). >TITLE >H2O with symmetry and block-diag. KS matrix, fixed occupations >SYMM C2V >CART BOHR > O 0.0 0.0 0.0 8.0 35 > H 1.4596 0.0 1.1354 1.0 32 >END >RUNTYPE STARTUP NO-OPT >MULTIPLICITY 1 >CHARGE 0.0 >FSYM SCFOCCUP >ALFA 3 1 1 0 >BETA 3 1 1 0 >DMIX MDENS 0.45 >END >A-OXYGEN (5,2;5,2) >A-HYDROGEN (5,1;5,1) >O-OXYGEN (5211/411/1) >O-HYDROGEN (41/1) >END c) H2O, C2v symmetry with block-diagonalization of the KS matrix and fixed occupation including a HOMO-LUMO excitation (singlet state) with fractional occupancy of 0.5 (Slater transition state method). Note that ALL orbitals involved must be specified in the occupation input following the SCFOccup line, i.e. the alfa spin a1 occupation here becomes four. >TITLE >H2O with symmetry and block-diag. KS matrix, fixed occupations >WATER HOMO-LUMO excitation with Slater TSM >SYMM C2V >CART BOHR > O 0.0 0.0 0.0 8.0 35 > H 1.4596 0.0 1.1354 1.0 32 >END >RUNTYPE STARTUP NO-OPT >MULTIPLICITY 1 >CHARGE 0.0 >MAXCYC 35 >FSYM SCFOCCUP EXCITED >ALFA 4 0 2 0 >BETA 3 0 2 0 >SYM 1 >ALFA 0 2 3 0.5 4 0.5 >BETA 0 0 >END >DMIX MDENS 0.45 >DIIS ON >END >A-OXYGEN (5,2;5,2) >A-HYDROGEN (5,1;5,1) >O-OXYGEN (5211/411/1) >O-HYDROGEN (41/1) >END d) An evaluation of atom-projected Fukui indices (f-, electrophilic indices) of CO. Restart from neutral CO with occupation (50002, 50002) >TITLE >CO molecule ground state, Fukui charge -0.01 >SYMM C4V >CARTESIAN BOHR > C 0.0000 0.00000 0.00000 6. 32 > O 0.0000 0.00000 2.15000 8. 32 >END >RUNTYPE RESTART NEWOCC >SCFTYPE DIRECT >POTENTIAL NONLOCAL PD91 PD91 >GRID FINE NONRANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 50 >ECONVERGENCE .0000001 >DCONVERGENCE .0000001 >DMIXING MDENS 0.40 >DIIS NEW 7 >FSYM SCFOCC EXCITED >ALFA 5 0 0 0 2 >BETA 5 0 0 0 2 >FUKUI ALFA -.01 1 1 5 1. >END >VIRT ALL >MULL ON FULL >ORBI 5D >END >A-CARBON (4,4;4,4) >A-OXYGEN (4,4;4,4) >O-CARBON (631/31/1) >O-OXYGEN (631/31/1) >END e) An evaluation of atom-projected Fukui indices (f+, nucleophilic indices) of CO. Restart from neutral CO with occupation (50002, 50002) >TITLE >CO molecule ground state, Fukui charge +0.01 >SYMM C4V >CARTESIAN BOHR > C 0.0000 0.00000 0.00000 6. 32 > O 0.0000 0.00000 2.15000 8. 32 >END >RUNTYPE RESTART NEWOCC >SCFTYPE DIRECT >POTENTIAL NONLOCAL PD91 PD91 >GRID FINE NONRANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 50 >ECONVERGENCE .0000001 >DCONVERGENCE .0000001 >DMIXING MDENS 0.40 >DIIS NEW 7 >FSYM SCFOCC EXCITED >ALFA 5 0 0 0 3 >BETA 5 0 0 0 2 >SYM 5 >ALFA 0 1 3 0.0 >BETA 0 0 >FUKUI ALFA .01 1 5 3 1. >END >VIRT ALL >MULL ON FULL >ORBI 5D >END >A-CARBON (4,4;4,4) >A-OXYGEN (4,4;4,4) >O-CARBON (631/31/1) >O-OXYGEN (631/31/1) >END Calculations for atoms are more difficult than those for molecules. This is due to the rotational symmetry and the more general "multiplet problem" in DFT discussed, for example, in Refs. [D94] and [S95]. In all cases it is suggested to use the generic atom symmetry (SYMMetry ATOM), see a.4). Open shell atoms can be treated either as "spherical" atoms or non-spherical atoms: 1) Atoms with closed atomic shells or only s-type open shells are spherical by their nature. Others can be forced to exhibit spherical total electron density by introducing fractional occupation numbers. This is particularly simple with the atom symmetry (SYMM ATOM) together with the CONFig option, see b.7). Note that the resulting solution in general does NOT correspond to a pure atomic state. 2) Non-spherical atoms (with non-spherical total electron density) can be treated by integer occupation of pure (real) s,p,d orbitals. Since the state is determined by the occupations in a SINGLE-DETERMINANT picture and the program uses real orbitals, one must be careful to give occupations that produce a single-determinant spin and space representation of the desired atomic state. This is always possible for the highest spin-state, but low-spin open-shell states must be treated by considering the state-mixing induced by giving only the spin-orbital occupations. For fixed atom occupations using the generic atom symmetry, the occupation numbers given on the two lines following the keyword line with keyword 'SCFOccupation' are different in their definitions from those of the molecular case. The lines ALFA ns npx npy npz ndz2 ndx2-y2 ndxy ndxz ndyz BETA ns npx npy npz ndz2 ndx2-y2 ndxy ndxz ndyz define the total occupation of all s, px, py, pz, dz2, dx2-y2, dxy, dxz, dyz orbitals of alpha and beta spin, respectively. Examples of atoms: a) C atom with block-diagonalization of the KS matrix and Aufbau principle occupation, pure s,p,d orbitals with non-spherical electron density. >TITLE >C atom, non-spherical >SYMM ATOM >CARTESIAN BOHR > C 0.00000 0.00000 0.00000 6.0 35 >END >RUNTYPE STARTUP NO-OPTI >GRID FINE RANDOM >MULTIPLICITY 3 >CHARGE 0. >DMIXING MDENS 0.3 >DIIS ON >FSYM AUFBAU >END >A-CARBON (5,2;5,2) >O-CARBON (621/41/1*) >END b) C atom with block-diagonalization of the KS matrix and fixed occupations, pure s,p,d orbitals with non-spherical electron density. >TITLE >C atom, non-spherical, fixed occupations >SYMM ATOM >CARTESIAN BOHR > C 0.00000 0.00000 0.00000 6.0 35 >END >RUNTYPE STARTUP NO-OPTI >GRID FINE RANDOM >MULTIPLICITY 3 >CHARGE 0. >DMIXING MDENS 0.3 >DIIS ON >FSYM SCFOCC >ALFA 2 1 0 1 0 0 0 0 0 >BETA 2 0 0 0 0 0 0 0 0 >END >A-CARBON (5,2;5,2) >O-CARBON (621/41/1*) >END c) C atom with block-diagonalization of the KS matrix and fixed occupations, fractional p occupations with spherical electron density. This produces equivalent orbitals, but corresponds to a mixture of atomic states. >TITLE >C atom, non-spherical, fixed occupations >SYMM ATOM >CARTESIAN BOHR > C 0.00000 0.00000 0.00000 6.0 35 >END >RUNTYPE STARTUP NO-OPTI >GRID FINE RANDOM >MULTIPLICITY 1 >CHARGE 0. >DMIXING MDENS 0.3 >DIIS ON >FSYM AUFBAU >CONFIG 2 1 0 2 1 0 >END >A-CARBON (5,2;5,2) >O-CARBON (621/41/1*) >END d) C atom with block-diagonalization of the KS matrix and fixed occupations, pure s,p,d orbitals (s and d-sigma can mix) with non-spherical electron density, 2s -> 2p excitation, total spin projection Ms=1 (corresponding to a triplet state if it had been high-spin). >TITLE >C atom, non-spherical, fixed occupations >SYMM D2H >CARTESIAN BOHR > C 0.00000 0.00000 0.00000 6.0 35 >END >RUNTYPE STARTUP NO-OPTI >GRID FINE RANDOM >MULTIPLICITY 3 >CHARGE 0. >DMIXING MDENS 0.3 >DIIS ON >FSYM SCFOCC >ALFA 1 1 1 1 0 0 0 0 0 >BETA 2 0 0 0 0 0 0 0 0 >END >A-CARBON (5,2;5,2) >O-CARBON (621/41/1*) >END e) C atom with block-diagonalization of the KS matrix and fixed occupations, pure s,p,d orbitals (s and d-sigma can mix) with non-spherical electron density, alpha 2s -> 3s excitation, triplet state. Note that ALL orbitals involved must be specified in the occupation input following the SCFOccup line, i.e. the alfa spin a1 occupation here becomes 3. >TITLE >C atom, non-spherical, fixed occupations >SYMM D2H >CARTESIAN BOHR > C 0.00000 0.00000 0.00000 6.0 35 >END >RUNTYPE STARTUP NO-OPTI >GRID FINE RANDOM >MULTIPLICITY 3 >CHARGE 0. >DMIXING MDENS 0.3 >DIIS ON >FSYM SCFOCC EXCITED >ALFA 3 1 0 1 0 0 0 0 0 >BETA 2 0 0 0 0 0 0 0 0 >SYM 1 >ALFA 0 2 2 0. 3 1. >BETA 0 0 >END >END >A-CARBON (5,2;5,2) >O-CARBON (621/41/1*) >END f) C atom with block-diagonalization of the KS matrix and fixed occupations, pure s,p,d orbitals (s and d-sigma can mix) with non-spherical electron density, HOMO hole state, doublet. >TITLE >C atom: Ionization from HOMO >SYMM D2H >CARTESIAN BOHR > C 0.00000 0.00000 0.00000 6.0 35 >END >RUNTYPE STARTUP NO-OPTI >MULTIPLICITY 2 >CHARGE 1. >DMIXING MDENS 0.3 >DIIS ON >FSYM AUFBAU >END >A-CARBON (5,2;5,2) >O-CARBON (621/41/1*) >END g) C atom with block-diagonalization of the KS matrix and fixed occupations, pure s,p,d orbitals (s and d-sigma can mix) with non-spherical electron density, C1s hole state, quartet. Note that ALL orbitals involved must be specified in the occupation input following the SCFOccup line, i.e. the beta spin a1 occupation here is still 2, but the occupation of the 1s is set to zero. (The MULTIPLICITY is the formal one based on the initial occupations only, before modification of the occupation numbers. In the program only the spin projection is used.) >TITLE >C atom, Ionization from 1s, quartet state >SYMM D2H >CARTESIAN BOHR > C 0.00000 0.00000 0.00000 6.0 35 >END >RUNTYPE STARTUP NO-OPTI >MULTIPLICITY 3 >CHARGE 1. >DMIXING MDENS 0.3 >DIIS ON >FSYM SCFOCCUP EXCITED >ALFA 2 1 0 1 0 0 0 0 0 >BETA 2 0 0 0 0 0 0 0 0 >SYM 1 >ALFA 0 0 >BETA 0 1 1 0. >END >END >A-CARBON (5,2;5,2) >O-CARBON (621/41/1*) >END ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.6) SMEAr [number] ----------------------------------------------------------------------------- For large systems with many orbital levels near the Fermi-level Ef, convergence can be improved by assigning fractional occupation numbers to orbitals of levels close to Ef. Using the keyword SMEAr with a window parameter w (in eV) places an energy window [Ef-w/2, Ef+w/2] around the Fermi-level in which orbitals are fractionally occupied [Fxx]. For w > 0, the occupations of all levels within the energy window are replaced by values of a monotonic occupation function n(E) where n(Ef-w)=1 and n(Ef+w)=0. (Thus, an orbital with a lower energy will be assigned a larger occupation number.) The function n(E) is determined iteratively to guarantee conservation of the total number of electrons. If the separation between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) is larger than w/2, SMEAR will have no effect. A reasonable smear value is w = 0.3 eV where w = 0 (no fractional occupation) is the default. Note that the smear option is incompatible with the DIIS extrapolation scheme, see keyword DIIS. Example: >SMEAR 0.3 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.7) CONFiguration [SPIN] nas nap nad nbs nbp nbd ----------------------------------------------------------------------------- For StoBe calculations on free atoms using the full rotational symmetry (symmetry label ATOM, see a.4) this keyword allows to define a global configuration by the occupation numbers nas of s electrons (from all shells) with spin alpha nap of p electrons (from all shells) with spin alpha nad of d electrons (from all shells) with spin alpha nbs of s electrons (from all shells) with spin beta nbp of p electrons (from all shells) with spin beta nbd of d electrons (from all shells) with spin beta The occupation numbers are averaged over the different orbitals of the p (px,py,pz) and d (dz2,dx2-y2,dxy,dxz,dyz) shells using fractional occupations if required. Averaging over shells irrespective of spin results in a singlet state. With the optional keyword "SPIN" the average is performed over spin-shells (alpha, beta) separately and allows to treat higher multiplet states in an approximate way. NOTE that the keyword CONFiguration MUST be combined with the full rotational symmetry (symmetry label ATOM, see a.4). Further, other explicit orbital occupations given by options 'NOFSym SCFOccupation' or 'FSYM SCFOccupation', see b.4) and b.5), must omitted. Examples: a) Averaged singlet state of nickel >TITLE >Ni spherical atom 3d9 4s1 (shell average) >SYMM ATOM >CARTESIAN BOHR > Ni 0.00000 0.00000 0.00000 28.0 32 >END >RUNTYPE START >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 100 >GRID FINE NONRANDOM >ECONVERGENCE 0.0000001 >DCONVERGENCE 0.0000001 >DIIS NEW 7 >ORBITAL 5D >FSYM AUFBAU >CONFIG 4 6 5 3 6 4 >VIRT ALL >END >A-NICKEL (5,5;5,5) >O-NICKEL (63321/5211*/311+) >END b) Averaged triplet state of nickel >TITLE >Ni spherical atom 3d9 4s1 (spin shell average) >SYMM ATOM >CARTESIAN BOHR > Ni 0.00000 0.00000 0.00000 28.0 32 >END >RUNTYPE START >MULTIPLICITY 3 >CHARGE 0 >MAXCYCLES 100 >GRID FINE NONRANDOM >ECONVERGENCE 0.0000001 >DCONVERGENCE 0.0000001 >DIIS NEW 7 >ORBITAL 5D >FSYM AUFBAU >CONFIG SPIN 4 6 5 3 6 4 >VIRT ALL >END >A-NICKEL (5,5;5,5) >O-NICKEL (63321/5211*/311+) >END ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.8) EXCIted [ALFA, BETA] ----------------------------------------------------------------------------- This keyword allows to calculate excited states where holes are created in specific orbitals achieved by level shifting. Keyword line EXCIted ALFA na1 na2 ... creates empty alpha orbitals where indices na1, na2, ... refer to the energetic order of KS one-electron energies. Up to 20 orbitals can be included on the keyword line. EXCIted BETA nb1 nb2 ... creates empty beta orbitals where indices nb1, nb2, ... refer to the energetic order of KS one-electron energies. Up to 20 orbitals can be included on the keyword line. For the holes to be created, level shifts, see keyword LEVElshift, are required to be larger than the energy difference between the HOMO level and the levels to be emptied. A more efficient method to calculate these states is based on using fixed occupations with the EXCIted option, see keywords FSYMmetry or NOFSymmetry. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.9) FIELd Fx fy fz ----------------------------------------------------------------------------- This keyword includes an external electric field (Fx, Fy, Fz) (in cartesian coordinates and atomic units, H/Bohr) in the electronic structure evaluation. If a symmetry other than C1 is assumed the field direction will be checked for compatibility (the field must not lower the symmetry). By default, no field is applied. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.10) SPINcontamination [FULL] ----------------------------------------------------------------------------- This keyword evaluates the spin contamination by calculating the expectation value <S2> and an effective multiplicity based on the Kohn-Sham MOs (assuming occupancy = 1 for each occupied orbital). The optional parameter 'FULL' includes a listing of the overlap matrix <alpha|beta> of all occupied orbitals used to calculate the spin contamination Example: >SPINCONT FULL ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.11) REORder ----------------------------------------------------------------------------- Allows the reordering of orbitals from a previous SCF calculation to be used to build the starting density in a restart run. The starting configuration is defined by na triplets for spin up (alpha) orbitals and nb triplets for spin down (beta) orbitals which are given on the lines after the keyword line. Each triplet (isym, i, k) selects alpha/beta orbitals i, i+1, ... k of symmetry representation isym in the restart file. Up to 6 triplets per line are used (up to a total of 100 triplets). The orbital indices reflect the order within each symmetry as given in the output listing of the SCF calculation whose restart file is to be used as input. The orbital order may also be checked with utility anlyz (option PORBitals), see Sec. 2.7.1. The general format reads as follows >REORDER >ALFA na >iasym(1) ia(1) ka(1) iasym(2) ia(2) ka(2) ... iasym(6) ia(6) ka(6) > ... > ... iasym(na) ia(na) ka(na) >BETA nb >ibsym(1) ib(1) kb(1) ibsym(2) ib(2) kb(2) ... ibsym(6) ib(6) kb(6) > ... > ... ibsym(na) ib(na) kb(na) Example: > REORDER > ALFA 7 > 1 1 2 1 4 4 1 3 3 1 5 7 1 9 9 2 1 7 > 3 1 3 > BETA 4 > 1 2 9 2 1 7 3 1 2 3 4 4 This defines a (hypothetical) configuration with orbital sets alpha: { 1 2 4 3 5 6 7 9 / 1 2 3 4 5 6 7 / 1 2 3} beta: { 2 3 4 5 6 7 8 9 / 1 2 3 4 5 6 7 / 1 2 4} in three symmetry representations. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) b.12) LOCAlize nloc ----------------------------------------------------------------------------- Localize orbitals at nloc specific atom centers. This requires a previous run to define the orbitals to localize. The localization is performed by diagonalizing the R**2 matrix (in MO basis) with respect to each of the requested nloc centers in sequence. The localization at the first center uses all available occupied MO’s and selects the requested number of orbitals as localized based on the eigenvalues of the diagonal R**2 operator (“iorban(isym)” lowest eigenvalues in each symmetry “isym”). For the next center the R**2 matrix is set up with respect to that center, transformed to MO basis (excluding the previously localized orbitals) and diagonalized. This procedure preserves the previously localized orbitals and results in a set of MO’s where the first “iorban(isym)” in irreducible representation “isym” are localized on the first specified center, the next set on the next requested center etc. Note that the energetic order is not preserved, i.e. the orbitals are now reordered according to the R**2 eigenvalues with respect to each center for the selected orbitals and with respect to the last center for the remaining ones. This can thus be used to put the core orbitals in a predefined order in core excitation or core hole calculations, see examples below. A parameter RMAX can be used to specify the maximum R**2 eigenvalue, e.g., in the example below 9 localized orbitals are requested for a water molecule embedded in a larger cluster of waters. The molecule itself carries 5 orbitals and the additional 4 would be accepting lone-pairs or donated H-bonds associated with neighboring waters, but only accepted if within a squared distance of 26 (Bohr)**2. Orbitals that should be excluded from the localization procedure can be specified by the FROZen keyword which can be applied to either ALPHA or BETA spin or the same parameters applied to both. Note that, by applying the procedure to only the occupied orbitals the total density is preserved. In case of degeneracies the program will try to resolve them by diagonalizing in the sequence (1) R**2 (2) X**2-Y**2 (3) 2xZ**2-X**2-Y**2 (4) X**2 To preserve the localized character of the orbitals, e.g., in an XAS or XPS spectrum calculation, all localized core orbitals on centers other than the active one (where the excitation/ionization originates) should be frozen by using the SUPSYM keyword (note that the order of the orbitals is determined by the order of localization). The keyword line is followed by nloc sets of lines 1 to 5 Line 1 (A4) Atomlabel(i) This defines the atom center i considered for orbital localization. The label MUST be identical (except for leading blanks) to the label of the center used in the input. Line 2 (A4,20I) 'ALFA', (iorb, iorb = 1,nsym) This defines the number of alpha (spin up) orbitals to be localized in each symmetry on center Atomlabel(i). Line 3 (A4,20I) 'BETA', (iorb, iorb = 1,nsym) This defines the number of beta (spin down) orbitals to be localized in each symmetry on center Atomlabel(i). Additional options: Line 4 (2A4) FROZen [ALFA, BETA] This specifies orbitals that should be excluded from the localization procedure. ALFA and/or BETA spin can be given separately or, if left out, the same parameters apply to both spin manifolds. The parameters are given in the same format as used with the SUPSym option for freezing orbitals in an SCF calculation. The FROZen keyword is followed by a specification of the number num1 of orbitals to be frozen and corresponding sequence numbers iorb(i), i = 1, num1 Line 4a (20I) num1 Number orbitals that should be frozen. Line(s) 4b (20I) iorb(1), iorb(2), ...,iorb(num1) Sequence numbers of orbitals that should be frozen. If more than 20 sequence numbers are to be provided they can be given on additional lines. Line 5 (A4,20I) 'RMAX' R2max This defines the maximum R**2 eigenvalue accepted for a localized orbital. This can be used to guarantee that all relevant orbitals are found in cases of varying distances. The LOCALIZE section needs to be terminated by a line reading 'END'. Example 1: > LOCALIZE > O1 > ALFA 7 > BETA 0 > RMAX 26. > END defines a localization of 7 spin up orbitals at the center labeled 'O1' where a radius square of 26. (radius slightly above 5) is considered. Example 2: XPS energy of carbon #6 in cytidine. Localize in sequence all the heavy atom 1s orbitals. Freeze all of them except C6 using supersymmetry (supsym) with one orbital per representation. To compute C7 the same input is used, but with the hole in orbital 10 (EXCITED) and orbital 9 frozen and 10 relaxed in the SUPSYM input. title cytidine – compute xps energy for all heavier atoms (one at a time) nosymm cartesian angstrom N1 0.38646326 -1.54649985 0.66531157 7. 32 N2 -0.08738194 -3.78380815 -0.08598576 7. 32 N3 0.65678252 -4.51297581 -2.12464257 7. 32 C1 0.93086113 -0.81560931 3.03141808 6. 32 C2 0.27418739 -0.48659819 1.66662656 6. 32 C3 -1.25360574 0.13338725 3.33272145 6. 32 C4 0.15064289 0.13118493 3.95253745 6. 32 C5 1.08672425 -1.30871792 -0.48041873 6. 32 C6 1.24580924 -2.26630584 -1.43488579 6. 32 C7 0.59055776 -3.52087309 -1.19374197 6. 32 C8 -0.17465968 -2.86962981 0.92715897 6. 32 C9 -2.22004073 -0.86362975 3.98181078 6. 32 O1 -1.09122955 -0.17376422 1.92312879 8. 32 O2 -0.68010463 -3.09724861 2.02059654 8. 32 O3 -2.50948314 -0.48787428 5.33869194 8. 32 O4 0.80848843 1.40890755 3.81360550 8. 32 O5 2.33322485 -0.62732755 3.03905717 8. 32 H 1.05002970 -4.35307374 -3.04719524 1. 32 H 0.11021281 -5.35529433 -1.95511616 1. 32 H 1.82446618 -2.06603579 -2.34003132 1. 32 H 1.52962973 -0.31372474 -0.58072057 1. 32 H 0.42704726 2.02012457 4.46876354 1. 32 H -1.68265352 1.15141678 3.40568261 1. 32 H -1.77243659 -1.86689718 4.01524978 1. 32 H -3.13215705 -0.93155240 3.35777074 1. 32 H -3.10596573 0.28600658 5.31318650 1. 32 H 0.14045634 -0.19711390 5.00652857 1. 32 H 0.72465240 -1.86086239 3.29378764 1. 32 H 2.46204101 0.34144288 3.14070687 1. 32 H 0.79849013 0.39233230 1.23364306 1. 32 end nofsym scfocc excited alfa 64 beta 64 sym 1 alfa 0 1 9 0. beta 0 0 end supsym alfa range 16 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 end localize 17 N1 alfa 1 beta 0 N2 alfa 1 beta 0 N3 alfa 1 beta 0 C1 alfa 1 beta 0 C2 alfa 1 beta 0 C3 alfa 1 beta 0 C4 alfa 1 beta 0 C5 alfa 1 beta 0 C6 alfa 1 beta 0 C7 alfa 1 beta 0 C8 alfa 1 beta 0 C9 alfa 1 beta 0 O1 alfa 1 beta 0 O2 alfa 1 beta 0 O3 alfa 1 beta 0 O4 alfa 1 beta 0 O5 alfa 1 beta 0 end runtype newocc no-opti scftype direct potential nonlocal RPBE PBE grid fine random multiplicity 1 charge 0 maxcycles 75 diis new 7 dmix mdens 0.1 econv 0.0000001 dconv 0.005 maxgeo 75 hessian 7 gconvergence 0.0001 mulliken full ctrl printout off file gmolden end Example 3: Evaluation of carbon #2 (C2) 1s excitations in pyridine using the transition potential approach to calculate the corresponding XAS spectrum. Localize in sequence the 1s orbitals of all other carbons (C1, C3, C4, C5) and freeze them using supersymmetry (supsym) with one orbital per representation. This calculation requires a restart output file from a previous groundstate calculation. In the following, we list three input files, for (a) the ground state, (b) the transition potental, and (c) the relaxed hole state calculation. The restart file of (a) needs to be given as input to (b) and (c). (a) GROUND STATE CALCULATION title Pyridine C5NH5 ground state for NEXAFS symm C1 cartesian angstrom N 1.39612966 0.00000000 0.00000000 7.0000 35 C1 0.68882311 1.14824537 0.00000000 6.0000 35 C2 0.68882311 -1.14824537 0.00000000 6.0000 35 C3 -0.71283021 1.20331341 0.00000000 6.0000 35 C4 -0.71283021 -1.20331341 0.00000000 6.0000 35 C5 -1.42987148 0.00000000 0.00000000 6.0000 35 H 1.27938895 2.07088516 0.00000000 1.0000 32 H 1.27938895 -2.07088516 0.00000000 1.0000 32 H -1.22662201 2.16743392 0.00000000 1.0000 32 H -1.22662201 -2.16743392 0.00000000 1.0000 32 H -2.52352446 0.00000000 0.00000000 1.0000 32 end runtype start scftype direct potential nonlocal RPBE PBE grid fine nonrandom multiplicity 1 charge 0 maxcycles 100 econvergence 0.0000001 dconvergence 0.0000001 dmixing mdens 0.1 diis new 7 virt all print default nofsym scfo alfa 21 beta 21 end GENA4 O-NITROGEN iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) O-NITROGEN iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) end (b) TRANSITION POTENTIAL CALCULATION title Pyridine C5NH5 TP state for C2 1s NEXAFS symm C1 cartesian angstrom N 1.40270041 0.00000000 0.00000000 7.0000 35 C1 0.69106191 1.15092381 0.00000000 6.0000 35 C2 0.69106191 -1.15092381 0.00000000 6.0000 35 C3 -0.71373934 1.20353633 0.00000000 6.0000 35 C4 -0.71373934 -1.20353633 0.00000000 6.0000 35 C5 -1.43503389 0.00000000 0.00000000 6.0000 35 H 1.28137718 2.07468149 0.00000000 1.0000 32 H 1.28137718 -2.07468149 0.00000000 1.0000 32 H -1.22770180 2.16899074 0.00000000 1.0000 32 H -1.22770180 -2.16899074 0.00000000 1.0000 32 H -2.52940901 0.00000000 0.00000000 1.0000 32 end runtype restart newocc scftype direct potential nonlocal RPBE PBE grid fine nonrandom multiplicity 1 charge 0 maxcycles 100 econvergence 0.0000001 dconvergence 0.0000001 dmixing mdens 0.1 diis new 7 virt all print default localize 4 C1 alfa 1 beta 0 C3 alfa 1 beta 0 C4 alfa 1 beta 0 C5 alfa 1 beta 0 end supsym alfa 1 1 1 2 1 3 1 4 end nofsym scfo alfa 21 1 6 0.5 beta 21 0 xray xas end end GENA4 O-NITROGEN iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) O-NITROGEN iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) end (c) RELAXED HOLE STATE CALCULATION title Pyridine C5NH5 C2 1s hole state ionization potential symm C1 cartesian angstrom N 1.40270041 0.00000000 0.00000000 7.0000 35 C1 0.69106191 1.15092381 0.00000000 6.0000 35 C2 0.69106191 -1.15092381 0.00000000 6.0000 35 C3 -0.71373934 1.20353633 0.00000000 6.0000 35 C4 -0.71373934 -1.20353633 0.00000000 6.0000 35 C5 -1.43503389 0.00000000 0.00000000 6.0000 35 H 1.28137718 2.07468149 0.00000000 1.0000 32 H 1.28137718 -2.07468149 0.00000000 1.0000 32 H -1.22770180 2.16899074 0.00000000 1.0000 32 H -1.22770180 -2.16899074 0.00000000 1.0000 32 H -2.52940901 0.00000000 0.00000000 1.0000 32 end runtype restart newocc scftype direct potential nonlocal RPBE PBE grid fine nonrandom multiplicity 1 charge 0 maxcycles 100 econvergence 0.0000001 dconvergence 0.0000001 dmixing mdens 0.1 diis new 7 virt all print default localize 4 C1 alfa 1 beta 0 C3 alfa 1 beta 0 C4 alfa 1 beta 0 C5 alfa 1 beta 0 end supsym alfa 1 1 1 2 1 3 1 4 end #nofsym scfo #alfa 21 #beta 21 nofsym scfo alfa 21 1 6 0.0 beta 21 0 end GENA4 O-NITROGEN iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo GENA4 O-CARBON iii_iglo A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) A-HYDROGEN (3,1;3,1) O-NITROGEN iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-CARBON iii_iglo O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) O-HYDROGEN (311/1) end 2.2.3. KEYWORDS FOR SCF ITERATION (Goto TOC, KEYW, KEYA) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.1) RUNType [STARtup, RESTart, CONTinue, CPROperties, NO-Optimize, (NOOPtimize,) OPTimize, NEWGeometry, NEWOccupation, VIBRations, POLArizability, NEBcalculation] ----------------------------------------------------------------------------- This keyword defines the type of calculation to be performed. Depending on the run type up to two option keywords (which must be consistent) are needed where all meaningful combinations are discussed in the following. If starting from scratch, the initial density is generated from orbitals obtained from a diagonalization of the one-electron matrix (i.e. neglecting electron repulsion and the exchange-correlation potential). ---- NOTE ----------------------------------------------------------------- In single-image runs the final results are written onto a save file (fort.2). In a restart this information is read from fort.1 so that the save file must be copied onto fort.1 before a restart run. ---- NOTE ----------------------------------------------------------------- (A) Single point SCF calculations for a fixed geometry RUNType STARtup NO-Optimize this starts a single point (fixed geometry) SCF calculation from scratch ignoring any input from previous StoBe runs. This keyword line may be shortened to 'RUNType STARtup'. RUNType STARtup NOOPtimize here keyword NOOPtimize means the same as NO-Optimize above and is included only for compatibility with previous deMon format input. RUNType RESTart this restarts a single point (fixed geometry) SCF calculation if a previous run was not converged. Orbital occupation numbers are taken from the input file if defined (see keywords FSYM, NOFS), else from the restart file (in the latter case total charge and spin multiplicity of the input and restart files must agree). Atom coordinates are taken from the restart file. This run type requires additional file input (unit 1) from a previous StoBe run (single point SCF or geometry optimization). Using occupation or geometry input different from that of the restart file should/must be handled using keywords NEWOcc or NEWGeometry (see below). RUNType RESTart NEWGeometry this restarts a single point (fixed geometry) SCF calculation allowing a system geometry different from that of the previous SCF run. Atom coordinates are taken from the input file. Orbital occupation numbers are also taken from the input file if defined (see options FSYM, NOFS), else from the restart file (in the latter case total charge and spin multiplicity of the input and restart files must agree). This run type requires additional file input (unit 1) from a previous StoBe run (single point SCF or geometry optimization). This keyword line may be shortened to 'RUNType NEWGeometry'. RUNType RESTart NEWOccupation this restarts a single point (fixed geometry) SCF calculation allowing an orbital occupation different from that of the previous SCF run. Orbital occupation numbers must be provided explicitly in the input file (see options FSYM, NOFS), while atom coordinates are taken from the restart file. This run type requires additional file input (unit 1) from a previous StoBe run (single point SCF or geometry optimization). This keyword line may be shortened to 'RUNType NEWOccupation'. (B) Geometry optimizations involving SCF and force field iterations RUNType STARtup OPTimize this starts a geometry optimization from scratch and does not use any input from previous StoBe runs. This keyword line may be shortened to 'RUNType OPTimize'. RUNType CONTinue this continues a geometry optimization where the starting geometry in the input file is ignored and the actual geometry of the restart file (unit 1) is taken. Orbital occupation numbers, if defined (see keywords FSYM, NOFS), and spin multiplicity of the input and restart files must agree. Force information of a previous geometry optimization restart file is used. This run type requires additional file input (unit 1) from a previous StoBe run (geometry optimization). NOTE that keyword CONTinue must NOT be combined with other keywords (like NEWGeometry or NEWOccupation) except VIBRations, see below. RUNType OPTimize NEWGeometry this starts a geometry optimization using a single point SCF restart file (unit 1) or a previous geometry optimization restart file. Initial orbital occupation numbers are taken from the input file if defined, see keywords FSYM, NOFS, else from the restart file. Initial atom coordinates are also taken from the input file. This run type requires additional file input (unit 1) from a previous StoBe run (single point SCF or geometry optimization). RUNType OPTimize NEWOccupation this starts a geometry optimization using a single point SCF restart file (unit 1) or a previous geometry optimization restart file. Initial orbital occupation numbers must be provided expicitly in the input file, see keywords FSYM, NOFS. Initial atom coordinates are taken from the restart file. This run type requires additional file input (unit 1) from a previous StoBe run (single point SCF or geometry optimization). This keyword line may also read 'RUNType NEWOccupation OPTimize'. (C) Reaction path optimization using the nudged-elatic-band (NEB) method [MJ94, MJG95, JMJ98, HUJ00] involving SCF and force field iterations along a path RUNType STARtup NEBcalculation this starts a NEB path optimization from scratch and does not use any input from previous StoBe runs. Additional control parameters determining the NEB run are provided by the keyword NEBParameters, see e.12). This keyword line may be shortened to 'RUNType NEBcalculation'. RUNType RESTart NEBcalculation this continues a NEB path optimization where the starting geometries of the nimg different images are ignored in the input file and replaced by those of the restart files (units nrbaseinp to nrbaseinp+nimg-1, see e.13) and a.1)) from a previous NEB run. Additional control parameters determining the NEB run are provided by the keyword NEBParameters, see e.12). This run type requires additional file input (units nrbase to nrbase+nimg-1) from a previous StoBe run (NEB path optimization) (D) Vibrational analysis RUNType STARtup VIBRations this starts a single point SCF calculation from scratch. This is followed by a vibrational analysis (vibrational eigenmodes and frequencies) based on finite differences of forces, where the initial input geometry is assumed to reflect the equilibrium geometry. Further parameters defining the vibrational analysis may be provided with keyword VIBRations, see e.11). See also notes below. RUNType RESTart VIBRations this starts a single point SCF calculation where orbital occupation numbers and atom coordinates are taken from the restart file (total charge and spin multiplicity of the input and restart files must agree). This is followed by a vibrational analysis, see above. This run type requires additional file input (unit 1) from a previous StoBe run (single point SCF, geometry optimization, no vibrational analysis!) and (optionally) a displacement vector file (unit 60), see e.11). For further comments on input parameters see NOTE above. See also notes below. RUNType CONTinue VIBRations this continues a vibrational analysis where the starting geometry in the input file is ignored and geometries of the restart file (unit 1) is taken. Orbital occupation numbers, if defined (see keywords FSYM, NOFS), and spin multiplicity of the input and restart files must agree. This run type requires additional file input (unit 1) from a previous StoBe run (vibrational analysis, displacement vector file, unit 60, is created from the restart file). For further comments on input parameters see NOTE above. See also notes below. The vibrational analysis creates, in addition to list output of vibrational frequencies (in harmonic approximation), dynamical dipoles, and displacement vectors, corresponding output in Molekel format on file unit 59 which can be used to visualize vibrational modes with the interactive graphics package Molekel [MOL03]. Note that StoBe creates Molekel format files for vibrational properties only. NOTE that the vibrational analysis is subject to constraints in the input parameters - The SCF runs for vibrational distortions must not use symmetry in the construction/diagonalization of the Kohn-Sham matrix even if the molecule/cluster has symmetry (defined by keyword SYMMetry, see a.4)). Thus, keyword NOFSymmetry (together with options AUFBau or SCFOccupation) must be applied explicitly, see b.4). - For large systems the orbital selection according to maximum overlap, keyword MAXOverlap, see c.9), is suggested to avoid state transitions. - Atom or group constraints, see d.5) and d.6), can be included in the vibrational analysis but have to be treated with care in order to avoid geometric incompatibilities. For example, constraining atom centers to motions which are symmetry forbidden should be avoided. NOTE that vibrational modes of a system with symmetry higher than C1, if calculated in one analysis run, are restricted to the highest symmetry representation A (or A1, A1g). A full vibrational analysis, including all symmetry representations, requires two subsequent analysis runs, - a first run with "RUNType ... VIBRations" where the full symmetry of the system is specified with keyword SYMMetry, option a.4). Further, the VIBRations keyword line, see e.11), has to read "VIBRations BASis". The vectors of symmetry-adapted vibrational distortions, file unit 60, have to be saved for the subsequent run. - a second run with "RUNType STARtup VIBRations" where the previous system is treated in C1 symmetry with all atom centers defined in the same order as in the first run. Further, the VIBRations keyword line, see e.11), has to read "VIBRations [modes] INPut [maxd]" where option [modes] can be left out or has to read "ALL" or "SELected". NOTE that vibrational analyses and polarizability calculations, see (D) below, have to be performed in separate StoBe runs. (E) Polarizability calculations RUNType [STARtup, RESTart] POLArizability this starts (or restarts) a single point SCF calculation, followed by three other SCF runs for the same geometry where homogeneous fields (E, 0, 0), (0, E, 0), (0, 0, E) are applied. The resulting dipole moments are evaluated and used to determine the static polarizability tensor of the system. The strength of the electric test field E (defaulted to 0.0005 au.) can be set separately by defining a homogeneous external electric field (Ex, Ey, Ez), see FIELd option, b.9), where E = sqrt(Ex*Ex+Ey*Ey+Ez*Ez) will be used in the polarizability calculations. By default, the SCF calculations following the first will be carried out as subsequent restart cases (except for free atoms/ions) which saves time but can introduce slight errors depending of the SCF convergence threshold. These SCF runs can be forced to start from scratch by option PSCRatch, see e.6). In polarization calculations on free atoms/ions SCF runs no. 2 - 4 will always start from scratch. NOTE that polarizability calculations and vibrational analyses, see (C) above, have to be performed in separate StoBe runs. Example: >RUNType STARtup POLArization >FIELd 0 0 .0002 >PSCRatch (F) Properties calculations RUNType CPROperties this uses the restart file (unit 1) to do properties only: plots (keyword DRAW), population analyses and X-ray spectra, bypassing the SCF iteration and geometry optimization. No restart file will be written. NOTE that property options SYMMetry, NEWSymmetry, VIBRation, FREQuencies, RAMAn, which are available with keyword RUNType are not available so far. The following table lists all keyword combinations that are accepted (and have been tested) at present >RUNType STARtup ( = RUNType STARtup NO-Optimize) >RUNType RESTart >RUNType RESTart NEWGeometry ( = RUNType NEWGeometry) >RUNType RESTart NEWOccupation ( = RUNType NEWOccupation) >RUNType STARtup OPTimize ( = RUNType OPTimize) >RUNType CONTinue >RUNType NEWGeometry OPTimize ( = RUNType OPTimize NEWGeometry) >RUNType NEWOccupation OPTimize ( = RUNType OPTimize NEWOccupation) >RUNType CPROperties >RUNType STARtup VIBRations >RUNType RESTart VIBRations >RUNType CONTinue VIBRations (to be implemented!) >RUNType STARtup POLArization >RUNType RESTart POLArization >RUNType NEBcalculation >RUNType STARtup NEBcalculation >RUNType RESTart NEBcalculation ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.2) SCFType [CONventional,DIRect,MEMory] [NUMerical] ----------------------------------------------------------------------------- This keyword defines the method of handling 2-electron integrals to construct the Kohn-Sham matrices and allows to force numerical evaluation of the exchange-correlation potentials and total energy in the SCF-cycles rather than using the fitted potentials based on the auxiliary basis. The option keywords are CONventional the 2-electron integrals are calculated at the beginning of the SCF procedure and stored on disk. The integrals are retrieved from disk at each iteration. This requires large disk storage for large systems. This option is used as default. DIRect the 2-electron integrals are recalculated at each iteration. No integral disk storage is performed. This option is suggested for smaller systems. MEMory the 2-electron integrals are calculated at the beginning of the SCF procedure and as many integrals as possible are fit into memory, the rest is stored on disk. The second option keyword NUM forces the numerical evaluation of the exchange-correlation contributions to the Kohn-Sham matrix over the grid at every SCF cycle instead of using an auxiliary basis to fit the potentials and energy. The NUM keyword is equivalent to providing an auxiliary basis with functions only for the fit of the Coulomb interaction, which also forces a strict numerical integration of the exchange and correlation contributions. However, neither can be combined with the use of symmetry. The use of the fitting basis results in cpu savings at the expense of accuracy in the energy in the iterations. The final energy is always computed numerically on an enhanced grid. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.3) MAXCycles [number] ----------------------------------------------------------------------------- This keyword defines the maximum number of iterations, NSCF, calculated in any SCF procedure (single point calculation or geometry optimization). The default value is NSCF=50. Note that restarting a run with NSCF=0 calculates only 1-electron properties without iterating and no restart file will be written. Example: >MAXCYCLES 50 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.4) ECONvergence [number] ----------------------------------------------------------------------------- This keyword defines the energy convergence for an SCF procedure where the number gives the convergence threshold Econv (in Hartree). An SCF procedure is terminated if three consecutive values of the total energies differ by less than Econv. The default value is Econv 1.0E-6 . Example: >ECONVERGE 1.D-6 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.5) DCONvergence [number] ----------------------------------------------------------------------------- This keyword defines the electron density convergence for an SCF procedure where the number gives the convergence threshold Dconv (in a.u.). An SCF procedure is terminated if three consecutive values of the densities differ by less than Dconv. The default value is Dconv = 0.001. Example: >DCONVERGE 1.D-3 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.6) DMIXing [NORMal nd nx , MDENsity nm] ----------------------------------------------------------------------------- This keyword defines the extrapolation between successive SCF iteration steps by linear mixing schemes. The recommended option keyword MDENsity, together with parameter nm, allows mixing of density matrices by Pin(m+1) = (1 - nm) * Pin(m) + nm * Pout(m) 0 < nm < 1 where Pin(m), Pin(m+1) are input density matrices of SCF steps m, m+1 and Pout(m) is the output density matrix of SCF step m. Reasonable mixing factors range between 0.05 (difficult convergence) and 0.5 (easy convergence) with nm = 0.2 being the default. Density matrix mixing with nm = 0.2 is set by default. NOTE that density matrix mixing leads to erroneous results if changing fractional occupations (see level occupation, smearing keyword SMEAr) are used. If an SCF calculation is started from scratch (i.e. without input density) the fixed mixing of densities is replaced by "dynamical" mixing for the beginning N iterations using mixing factors nd(i) = nd + (N-i+1)/N * dnd for iterations i < N+1 = nd for iterations i > N where at present N = 4, dnd = 0.3. The option keyword NORMal, together with two parameters nd, nx, allows mixing of electron densities and of exchange/correlation potentials by rhoin(m+1) = (1 - nd) * rhoin(m) + nd * rhoout(m) 0 < nd < 1 fxcin(m+1) = (1 - nx) * fxcin(m) + nx * fxcout(m) 0 < nx < 1 where rhoin(m), rhoin(m+1) are input electron densities of SCF steps m, m+1 and rhoout(m) is the output electron density of SCF step m. Further, fxcin(m), fxcin(m+1), fxcout(m) are input and output exchange/correlation potentials of steps m, m+1. Reasonable mixing factors range between 0.05 (difficult convergence) and 0.5 (easy convergence) with nd = nx = 0.2 being defaults. If the direct iterative inversion scheme [P82] is used to speed up convergence, see keyword DIIS, then the mixing factor nd is automatically reset to unity, when the DIIS scheme is turned on. Examples: >DMIXING MDENSITY 0.3 >DMIXING NORMAL 0.5 0.4 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.7) DIIS [OFF, ON [esh], NEW nstep [esh, [thrd]]] ----------------------------------------------------------------------------- This keyword defines the use of the DIIS algorithm (Direct Inversion of Iterative Subspace [P82]) to accelerate the SCF convergence, see also keyword DMIXing. Note that the DIIS scheme is incompatible with the smear option, keyword SMEAr. The keyword combination DIIS OFF does not use the DIIS scheme in the SCF iteration cycle. This is required if the smear option, keyword SMEAr, is to be used to determine orbital occupations. DIIS ON [eshd] uses the DIIS scheme in the SCF iteration cycle (set as default) where all previous iteration steps are used for the extrapolation between cycles. (This may result in very large intermediate scratch files for large/long runs.) This scheme includes downwards/upwards shifts of all occupied/empty orbital levels by eshd, analogous to keyword LEVElshift, see c.8), where eshd can be defined on the keyword line (eshd in eV units), see below. As a default, eshd = 0.5 eV is used and has to be modified if the HOMO/LUMO gap falls below 0.5 eV. DIIS NEW nstep [eshd, [thrd]] uses the DIIS scheme in the SCF iteration cycle where only up to nstep previous iteration steps are used for the extrapolation between cycles. (Reasonable nstep values are nstep = 5 - 10 with nstep = 7 being default.) This scheme includes downwards/upwards shifts of all occupied/empty orbital levels by eshd, analogous to keyword LEVElshift, see c.8), where eshd can be defined on the keyword line (eshd in eV units), see below. As a default, eshd = 0.5 eV is used and has to be modified if the HOMO/LUMO gap falls below 0.5 eV. Further, parameter thrd allows to set the threshold for the DIIS error, see below, where thrd is defaulted to 0.05. The DIIS algorithm is turned on only after the largest element of the DIIS error MATRIX (computed from the second iteration on) is below a threshold thrd (default = 0.05, may be reset with option DIIS NEW). This happens typically after 8 - 15 iterations. Before DIIS is turned on a linear mixing scheme is used for the extrapolation between successive SCF iteration steps, see c.6). Note that the eshd value of the DIIS option will be overridden by any level shift > 0 provided explicitly with keyword LEVElshift, see c.8). The eshd value will be used only if the keyword LEVElshift is omitted in the input or included by a keyword line "LEVElshift 0.0". ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.8) LEVElshift esh ----------------------------------------------------------------------------- This keyword includes the levelshift algorithm which may improve SCF convergence. Here the Kohn-Sham hamiltonian is modified such that all occupied/empty orbital levels are shifted downwards/upwards by esh. This increases the HOMO/LUMO gap artificially by 2*esh while the occupied orbitals and all calculated properties remain unchanged. (Level shifting may not be suitable in cases where the virtual orbital space is used, as in NMR or EPR calculations.) The procedure can avoid oscillatory convergence or divergence due to swapping between virtual and occupied orbitals. The default level shift is esh = 0 (applies no shift). For level shift values esh > 0 (in eV units, reasonable values are 5 - 10 eV) level shifting is applied where information about occupied and empty orbitals is assumed to be reliable. Therefore, level shifting should be used only after the energy ordering of the orbitals is stable, i. e. it should NOT be used in calculations starting from scratch. If the DIIS algorithm is used to accelerate the SCF convergence, see keyword DIIS, level shifting is always applied where the present esh value provided with keyword LEVElshift overrides any eshd value given (explicitly or defined implicitly) with keyword DIIS, see c.7). The eshd value is used only for esh = 0 (default shift). Note that level shifting is incompatible with the smear option, keyword SMEAr. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.9) MAXOverlap [LIST] ----------------------------------------------------------------------------- With this keyword, the maximum orbital overlap criterion is used to select orbitals of an SCF iteration: if in iteration n the set of occupied output orbitals is {f1, f2, ... fn}, the occupied orbitals of iteration n+1, {g1, g2, ... gn}, are selected such that the overlap integral (fi|gj) is largest (only orbitals fi, gj of the same symmetry representation are compared). This guarantees that the orbital character is approximately conserved and allows smoother convergence in some cases and also allows the calculation of specific excited and hole states. Example: >MAXOVERLAP The keyword sequence MAXOverlap LIST prints for each iteration the orbital selection and maximum overlap values. In the listing "S(i,k)= ovlp" means that occupied orbital i refers to the kth root from the diagonalization with the overlap value oclp. NOTE that this option may generate a lot of output and is meant only for testing purposes in case of convergence problems. The maximum overlap option may be used together with the REORder option which allows a reordering of orbitals from a previous run to be used to build the starting density in a restart run. For details see option b.11). Note that using the maximum orbital overlap option may result in orbital sequences in the output listing which do not reflect the natural sequence (increasing orbital energies within each symmetry representation). The relevant sequence is always given by the orbital listing of the output and refers also to the sequence used in the restart file output. This has taken into account in subsequent runs using the restart file as input. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.10) INTEgration [LEBEdev, GAUSs n1 n2 n3 n4 n5] ----------------------------------------------------------------------------- The angular integration can be performed using either Lebedev (default) or Gauss-Legendre quadrature. The keyword combination INTEgration LEBEdev (default) uses the Lebedev integration scheme [L75] which is highly accurate and more efficient than the Gauss-Legendre method but is restricted to a maximum of 194 grid points. INTEgration GAUSs n1 n2 n3 n4 n5 uses the Gauss-Legendre integration scheme which allows angular grids with more than 194 points but is less efficient than the Lebedev method (requiring about 20% more points for the same accuracy). The numbers n1 ... n5 define the order of the integration polynominals. For n2=n3=n4=n5=0 (or omitted in the input) one order n1 polynominal is used globally while finite values for n2 ... n5 allow to use different orders in 5 different radial regions. The number of points generated for each radial region is ` 2*(2ni+1)*(2ni+1). In practice, ni=5 leads to a good quality grid, ni=8 or 9 allow calculations of weak interaction energies (together with 128 radial grid points). Example: >INTEGRATE GAUSS 5 5 5 5 5 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.11) GRID [COARse,FINE,MEDIum,EXTRafine] [RANDom,NONRandom] [SAME] ----------------------------------------------------------------------------- The user specifies the quality and type of the grids used for the least squares fit of the exchange-correlation potentials and final energy density. Keyword GRID comes with up to three additional option keywords from the following three groups (a) the size of the angular grid is defined by keywords COARse using 6 points per radial shell MEDIum using 12 points per radial shell FINE using 26 points per radial shell (default) EXTRafine using 194 points per radial shell The radial grid consists of 64 points by default, i.e., each atom is surrounded by 64 shells of either 194, 26, 12, or 6 points. The number of radial points can be given for each atom with the Ngrid parameter in the coordinate list, see CARTesian or ZMATrix. (b) the mesh type is defined by keywords NONRandom (default) forces the angular points of the radial shells to line up. RANDom all angular points of a radial shell are subject to a common (random) rotation where rotation angles differ between shells. As a result, total energy variations due to coordinate rotations are somewhat reduced and possible bias introduced by the lineup of angular points is (hopefully) minimized. However, randomization may destroy the molecular symmetry slightly (position differences within 1.D-3 Angstrom). Perfect symmetry can be maintained by choosing the NONRandom option with a properly oriented molecule. (c) the additional keyword SAME specifies that the same angular grid is used for fitting and for the numerical integration to evaluate total energies. By default, the numerical integration is performed with a larger grid. Example: >GRID FINE NONRANDOM ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.12) WEDGe [TETA, PHI] (UNTESTED!) ----------------------------------------------------------------------------- This keyword allows to restrict the angular integration grid of selected atoms to a smaller range in case of the Gauss-Legendre integration scheme and symmetry, see keyword INTEgration. This results in a larger effective grid. The keyword line WEDGE TETA i1 nt1 i2 nt2 ... restricts the theta range of integration for atom ii to [0, pi/nti] where up to 10 atom centers ii can be included on the keyword line. The keyword line can be repeated for n > 10 atom centers. WEDGE PHI i1 np1 i2 np2 ... restricts the phi range of integration for atom ii to [0, pi/npi] where up to 10 atom centers ii can be included on the keyword line. The keyword line can be repeated for n > 10 atom centers. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.13) SLATer_radii ----------------------------------------------------------------------------- In the numerical integration scheme the points are distributed between atoms based on their size. The radii used as default are the Slater atomic radii. For strongly ionic systems (e.g. oxides such as MgO) the numerical accuracy can be improved by making the division based on the ionic radii instead. In the limit of a very large set of points per atom this should only have minor effects on the accuracy. By adapting the radii to the specific situation the standard numerical grids in StoBe can still be used. An indicator when this option can be used is a calculation where the total number of electrons obtained from the density integrated over the grid differs "substantially" (>0.01 electron) from the correct number of electrons. SLATer i1 rad1 i2 rad2 .. .... END where i1, i2,... are the sequence numbers (same order as in the coordinate input) of the symmetry independent centers for which the radius will be changed to the values rad1, rad2, ... The input is terminated by the END line. Example: >TITLE >CO/MgO(001) >SYMM C2V >CARTESIAN BOHR >O 0. 0. 7.13 8. 32 >C 0. 0. 5.00 6. 32 >Mg 0.00000 0.00000 0.0000 12. 32 >O 0.00000 0.00000 -3.97906 8. 32 >O 3.97906 0.00000 0.00000 8. 32 >O 0.00000 3.97906 0.00000 8. 32 >O 3.97906 3.97906 -3.97906 8. 32 >Mg 0.00000 3.97906 -3.97906 12. 32 >Mg 3.97906 0.00000 -3.97906 12. 32 >Mg 3.97906 3.97906 0.00000 12. 32 >END >RUNTYPE START NOOPTIMIZE >SCFTYPE DIRECT >SLATER RADII >3 0.86 >4 1.26 >5 1.26 >6 1.26 >7 1.26 >8 0.86 >9 0.86 >10 0.86 >END >POTENTIAL NONLOCAL PD86 PD86 >GRID FINE NONRANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 65 >ECONVERGENCE 0.000001 >DCONVERGENCE 0.00001 >MAXGEOMETRIES 15 >GCONVERGENCE 0.0001 >GSTEPSIZE 0.25 >HESSIAN 7 >DMIXING MDENS 0.15 >MULLIKEN ON >DIIS ON >FSYM SCFOCC >ALFA 37 23 23 14 >BETA 37 23 23 14 >END >A-OXYGEN (4,4;4,4) >A-CARBON (4,4;4,4) >A-MAGNESIUM (5,4;5,4) >A-OXYGEN (4,4;4,4) >A-OXYGEN (4,4;4,4) >A-OXYGEN (4,4;4,4) >A-OXYGEN (4,4;4,4) >A-MAGNESIUM (5,4;5,4) >A-MAGNESIUM (5,4;5,4) >A-MAGNESIUM (5,4;5,4) >O-OXYGEN (631/31/1) >O-CARBON (631/31/1) >O-MAGNESIUM (6321/411*) >O-OXYGEN (631/31/1) >O-OXYGEN (631/31/1) >O-OXYGEN (631/31/1) >O-OXYGEN (631/31/1) >O-MAGNESIUM (6321/411*) >O-MAGNESIUM (6321/411*) >O-MAGNESIUM (6321/411*) >END In the following, numerical values of all internally defined Slater radii are listed, radii are given in Angstrom. -------------------------------------------------------------------------- element Rslater element Rslater element Rslater -------------------------------------------------------------------------- 1 H 0.5000 36 Kr 2.7500 71 Lu 1.7500 2 He 2.0000 37 Rb 2.3500 72 Hf 1.5500 3 Li 1.4500 38 Sr 2.0000 73 Ta 1.4500 4 Be 1.0500 39 Y 1.8000 74 W 1.3500 5 B 0.8500 40 Zr 1.5500 75 Re 1.3500 6 C 0.7000 41 Nb 1.4500 76 Os 1.3000 7 N 0.6500 42 Mo 1.4500 77 Ir 1.3500 8 O 0.6000 43 Tc 1.3500 78 Pt 1.3500 9 F 0.5000 44 Ru 1.3000 79 Au 1.3500 10 Ne 2.2500 45 Rh 1.3500 80 Hg 1.5000 11 Na 1.8000 46 Pd 1.4000 81 Tl 1.9000 12 Mg 1.5000 47 Ag 1.6000 82 Pb 1.8000 13 Al 1.2500 48 Cd 1.5500 83 Bi 1.6000 14 Si 1.1000 49 In 1.5500 84 Po 1.9000 15 P 1.0000 50 Sn 1.4500 85 At 1.6500 16 S 1.0000 51 Sb 1.4500 86 Rn 3.2500 17 Cl 1.0000 52 Te 1.4000 87 Fr 2.8000 18 Ar 3.0000 53 I 1.4000 88 Ra 2.1500 19 K 2.2000 54 Xe 3.0000 89 Ac 1.9500 20 Ca 1.8000 55 Cs 2.6000 90 Th 1.8000 21 Sc 1.6000 56 Ba 2.1500 91 Pa 1.8000 22 Ti 1.4000 57 La 1.9500 92 U 1.7500 23 V 1.3500 58 Ce 1.8500 93 Np 1.7500 24 Cr 1.4000 59 Pr 1.8500 94 Pu 1.7500 25 Mn 1.4000 60 Nd 1.8500 95 Am 1.7500 26 Fe 1.4000 61 Pm 1.8500 96 Cm 1.7500 27 Co 1.3500 62 Sm 1.8500 97 Bk 1.7500 28 Ni 1.3500 63 Eu 1.8500 98 Cf 1.7500 29 Cu 1.3500 64 Gd 1.8000 99 Es 1.7500 30 Zn 1.3500 65 Tb 1.7500 100 Fm 1.7500 31 Ga 1.3000 66 Dy 1.7500 101 Md 1.7500 32 Ge 1.2500 67 Ho 1.7500 102 No 1.7500 33 As 1.1500 68 Er 1.7500 103 Lr 1.7500 34 Se 1.1500 69 Tm 1.7500 104 1.5500 35 Br 1.1500 70 Yb 1.7500 105 1.5500 -------------------------------------------------------------------------- ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.14) PTCHarges [FILE [ir], FIL& [ir]] ----------------------------------------------------------------------------- This keyword allows to include external point charges (without basis functions) in the calculation. At present, up to 100000 (noutch) external charges can be included where coordinates and charges are read from a separate charge file (unit ir) and/or from the input file. The keyword line(s) PTCHarges npc xpc ypc zpc qpc ... where the first line gives the number of point charges, npc, to be included. The following npc lines give for each point charge its coordinates (xpc, ypc, zpc) and its charge qpc. PTCHarges FILE [ir] defines input only from a separate file (unit ir, default = 9) where the charge information is given in the same ASCII format as discussed above. PTCHarges FIL& [ir] defines input from a separate file (unit ir, default = 9). Additional point charges can be read in from the input file. The input formats are those described above. Coordinates of atom charges have to be defined in Angstrom or Bohr units depending on the definitions used to describe atom positions, see keywords CARTesian and ZMATrix. In calculations with molecular symmetry, all point charge centers are checked for the existence of symmetry equivalent centers and centers are added if needed to conserve symmetry. Therefore, with molecular symmetry, the input of external charges can be resticted to non-equivalent centers. Symmetry equivalent centers, if provided with the input, will be detected and processed accordingly. Example: >PTCHARGES > 5 > 0 0 10 .1 > 5 0 0 -.1 > -5 0 0 -.1 > 0 5 0 .1 > 0 -5 0 .1 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.15) BSSError na i1 i2 ... ina ----------------------------------------------------------------------------- This keyword defines an SCF calculation where na selected atom centers (indices i1 i2 ... ina) are treated as "ghosts", i. e. their basis sets and numerical grids are included but their nuclear charges are ignored in the calculation. This type of SCF calculation allows an estimate of the basis set superposition error (BSSE). If symmetry is used, the atom numbering ii refers to non-equivalent centers only. The keyword line can contain up to 19 atom indices. If na > 19 atoms are to be included, the index list can be continued on the following line(s) with up to 20 indices per line. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.16) SUPSym [ALFA, BETA] ----------------------------------------------------------------------------- This restricts orbital rotations to avoid undesired mixings in, for instance, calculations where the energetic effects of different interactions should be investigated in CSOV (Constrained Space Orbital Variation) calculations [BHB84a,BHB84b]. By adding an additional (formal) symmetry label to specific orbitals an additional blocking of the Kohn-Sham matrix is done and only orbitals with the same (formal) symmetry label are allowed to mix with each other. By specifying the optional keyword ALFA or BETA this will be applied only to the specified spin symmetry. If this additional keyword is left out it will be applied to both the alpha and beta spin-orbitals. NOTE: this option can also be used to determine excited states variationally in a stepwise procedure by eliminating excitations into lower-lying orbitals through: 1) Assigning a different SUPSym to orbitals that should be eliminated 2) Defining an occupation zero to these orbitals in the EXCIted specification when restarting from the converged result for the preceding excited state. This option requires additional numerical input following the keyword line and reading SUPSym num1 i1 i2 i3 ... num2 j1 j2 j3 .... num... ... END where the first numerical line gives the number num1 of orbitals to be grouped together as "supersymmetry" 1. The following lines (20 entries to a line) specify which orbitals, i1, i2, ... are assigned to supersymmetry 1, where the numbering is from the first orbital in molecular symmetry 1 to the last orbital in the last molecular symmetry (irrespective of occupation, order given in the listing of orbitals in the output - same numbering for alpha and beta). Additional lines can be added if num1 > 20. The input for supersymmetry 2 is analogous to that of supersymmetry 1. Up to nsup different supersymmetries are allowed, where nsup + nsym (nsym = sum of basic symmetry representations) must not exceed 20. Example: >TITLE >h2o restart with frozen 1s core hole in alpha spin KS-matrix >SYMM C2V >CARTESIAN BOHR > O 0.0000 0.00000 0.0000 8. 32 > H 0.0000000 2.0504030 1.1309461 1. 32 >END >RUNTYPE NEWOCC NOOPTIMIZE >SCFTYPE DIRECT >POTENTIAL NONLOCAL PD86 PD86 >GRID FINE NONRANDOM >MULTIPLICITY 2 >CHARGE 1 >MAXCYCLES 75 >ECONVERGENCE 0.000001 >DCONVERGENCE 0.00001 >DMIXING MDENS 0.40 >MULLIKEN ON >DIIS ON >FSYM SCFOCC EXCITED >ALFA 3 1 1 0 >BETA 3 1 1 0 >SYM 1 >ALFA 0 1 1 0. >BETA 0 0 >END >SUPSYM ALFA >1 >1 >END >END >A-OXYGEN (4,4;4,4) >A-HYDROGEN (4,2;4,2) >O-OXYGEN (631/31/1) >O-HYDROGEN (311/1) >END There is an alternative way to assign orbitals to supersymmetries which is particularly useful when longer sequences of consecutive orbital indices are to be given. This option starts, for each supersymmetry, with the keyword RANGe as follows RANGe npairs where npairs denotes the number of index pairs to be given in the following line(s) (20 entries to a line, additional lines can be added if npairs > 10). These numerical lines reading ia1 ie1 ia2 ie2 ... define ranges of orbital indices selecting orbitals ia1, ia1+1, ie1, and ia2, ia2+1, ie2, ... for a given supersymmetry. As an example, the input > SUPSYM > 9 > 1 2 3 4 5 6 7 10 11 > 3 > 12 13 14 > END which sets two supersymmetries of 9 and 3 orbitals each can be written alternatively as > SUPSYM > RANGE 2 > 1 7 10 11 > RANGE 1 > 12 14 > END ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) c.17) TRRH [DIIS, nswitch] ----------------------------------------------------------------------------- Trust-region Roothan-Hall extrapolation from Thogersen et al., J. Chem. Phys. 123, 074103 (2005). This approach can be applied to cases that are very difficult to converge, oscillate or even diverge. A model Hessian for the orbital rotations is set up and an attempted step in the direction of the orbital gradient is attempted if the error function increases a shorter step is attempted. These microiterations are continued until a suitable step is found. The time for each iteration cycle can thus vary strongly depending on how many microiterations are required at each step. However, this is a very powerful procedure that “almost always” leads to convergence. The input specifies at which SCF iteration step the extrapolation should be turned on. For standard cases where it goes into the DIIS, but the DIIS has difficulties the TRRH extrapolation can be turned on automatically after the DIIS has started by TRRH DIIS This switches to the trust-region Roothan-Hall extrapolation two iteration steps after the direct iterative inversion scheme (DIIS) [P82], see also keyword DIIS, has been activated. This is considered the default option. For cases where the divergence or convergence difficulties appear at a given step it is recommended to turn on the TRRH a few iterations before reaching that point. For this to be well-defined the calculation has to be started again from the same point (so that the behavior at each iteration is identical) by specifying at which iteration the Trust-region Roothan-Hall procedure should be started. The corresponding input is TRRH nswitch This switches to the trust-region Roothan-Hall extrapolation at iteration step nswitch. 2.2.4. KEYWORDS FOR GEOMETRY OPTIMIZATION (Goto TOC, KEYW, KEYA) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.1) GCONvergence [number] ----------------------------------------------------------------------------- This keyword defines the gradient convergence for a geometry optimization where the number gives the convergence threshold Gconv (in a.u.). A geometry optimization is terminated when the energy gradient (absolute value) is less than Gconv. The default value is Gconv 1.0D-4. Example: >GCONVERGE 1.D-4 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.2) GSTEpsize [maxstep] [fscale] ----------------------------------------------------------------------------- This keyword defines an upper bound maxstep to the allowed average displacement of atoms between successive steps in a geometry optimization (maxstep in Bohr or Angstrom depending on the initial coordinate input, see a.1). Larger displacements will be reduced to the maximum while their direction remains unchanged. A reasonable value is 0.25 Bohr with the default = 0.5 Bohr. The second (optional) value allows to scale all displacements evaluated according to forces globally by a constant factor fscale (with default = 1.0). For values fscale < 1.0 this may result in an increased number of convergence steps but smoother convergence. Example: >GSTEPSIZE 0.25 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.3) MAXGeometries [number] ----------------------------------------------------------------------------- This keyword defines the maximum number of geometry steps, MAXG, in a geometry optimization procedure. The default value is MAXG=10. Example: >MAXGEOMETRIES 30 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.4) COORdinate FIX ----------------------------------------------------------------------------- This keyword allows to include geometry constraints for selected atom coordinates in a geometry optimization. This can be used as an alternative to keyword 'ATOM' (which is more general, see d.5). As an example, consider an adsorbate/cluster model where all cluster atoms are kept fixed while the adsorbate atom is allowed to equilibrate along the z axis or parallel to the surface (xy plane). The keyword line COORdinate FIX X=Y na constrains atom center na to movements along a plane through its starting position and spanned by vectors (1,1,0), (0,0,1). This has the same effect as keyword line 'ATOM DIR na 1 1 0 0 0 1'. COORdinate FIX X=Z na constrains atom center na to movements along a plane through its starting position and spanned by vectors (1,0,1), (0,1,0). This has the same effect as keyword line 'ATOM DIR na 1 0 1 0 1 0'. COORdinate FIX Y=Z na constrains atom center na to movements along a plane through its starting position and spanned by vectors (0,1,1), (1,0,0). This has the same effect as keyword line 'ATOM DIR na 0 1 1 1 0 0'. COORdinate FIX X na constrains atom center na to movements along a plane through its starting position and spanned by vectors (0,1,0), (0,0,1). This has the same effect as keyword line 'ATOM DIR na 0 1 0 0 0 1'. COORdinate FIX Y na constrains atom center na to movements along a plane through its starting position and spanned by vectors (1,0,0), (0,0,1). This has the same effect as keyword line 'ATOM DIR na 1 0 0 0 0 1'. COORdinate FIX Z na constrains atom center na to movements along a plane through its starting position and spanned by vectors (1,0,0), (0,1,0). This has the same effect as keyword line 'ATOM DIR na 1 0 0 0 1 0'. In all cases, the atom number na refers to a non-equivalent center with its symmetry equivalents being subject to corresponding constraints where symmetry relations are accounted for. Note that coordinate constraints cannot be combined. The last keyword line for a given atom center determines its constraints. Further, the present constraints are ignored for atom centers which are fixed by a keyword line 'ATOM FIX', see keyword ATOM, d.5). Example: >COOR FIX X=Y 4 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.5) ATOM [FIX, FREE, DIRection] ----------------------------------------------------------------------------- This keyword allows to include geometry constraints for selected atoms in a geometry optimization or vibrational analysis. It can be used and is more general than keyword 'COORdinate', see d.4). As an example, consider an adsorbate/cluster model where all cluster atoms are kept fixed at the experimental bulk geometry while the adsorbate structure is optimized. The keyword line ATOM FIX na i1 i2 ... ina fixes na atom centers i1 ... ina at their position given in the input file (or in the restart file if option CONTinue is used, see keyword RUNType, c.1)). All other atom centers are allowed to move freely in the optimization/vibration, see also below. If symmetry is used, the atom numbers ii refer to non-equivalent centers with their symmetry equivalents being fixed as well. The keyword line can contain up to 19 atom indices. If na > 19 atoms are to be included, the index list can be continued on the following line(s) with up to 20 atom indices. Fixing an atom center overrides other geometry constraints set by keyword COORdinate. ATOM FIX ALL fixes all atom centers of the cluster/molecule at their position given in the input file (or in the restart file if option CONTinue is used, see keyword RUNType, c.1)). This keyword line is meaningful only for geometry optimizations where group constraints (allowing groups of atoms to move) are given in addition, see d.6). ATOM FREE na i1 i2 ... ina keeps all atom centers fixed at their position given in the input file (or in the restart file if option CONTinue is used, see keyword RUNType, c.1)) while na atom centers i1 ... ina are allowed to move freely in the optimization/vibration, see also below. If symmetry is used, the atom numbering ii refers to non-equivalent centers with their symmetry equivalents being allowed to move such that the global symmetry is conserved. The keyword line can contain up to 19 atom indices. If na > 19 atoms are to be included, the index list can be continued on the following line(s) with up to 20 atom indices. Freeing an atom center allows it to be imposed by other geometry constraints set by keyword COORdinate. ATOM DIRection na x1 y1 z1 [x2 y2 z2] constrains the movement of atom center na (and its symmetry partners) during the optimization/vibration. Here v1 = (x1, y1, z1) and v2 = (x2, y2, z2) are direction vectors in cartesian coordinates. If only v1 is given on the keyword line, center na is constrained to move along the line defined by its starting position and direction vector v1. If both v1 and v2 appear on the keyword line, center na is constrained to move along the plane defined by its starting position and the two direction vectors v1, v2. If symmetry is used, the atom number na refers to a non-equivalent center with its symmetry equivalents being allowed to move such that the global symmetry is conserved. The present constraint is overridden by subsequent constraints set by keyword COORdinate. A keyword line ATOM FIX ... overrides all previous lines ATOM FIX, ATOM FREE, and ATOM DIR appearing in the input file. A keyword line ATOM FREE ... overrides all previous lines ATOM FREE, ATOM FIX and ATOM DIR appearing in the input file. A keyword line ATOM DIR ... constrains only those atom centers which have not been fixed in a previous ATOM FIX or ATOM FREE keyword line appearing in the input file. Examples: >ATOM FIX 4 1 2 3 4 >ATOM DIR 5 1. 0. 0. Fixes atom centers 1, 2, 3, 4 and constrains atom center 5 to move along (1,0,0) from its initial position. >ATOM FREE 3 5 6 7 >ATOM DIR 5 1. 0. 0. 0. 0. 1. Fixes all atom centers except centers 5, 6, 7 and constrains atom center 5 to move along a plane through its initial position and parallel to the xz plane. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.6) GROUp [DEFine, FIX, TRAnslate, ROTate, MOVe] ----------------------------------------------------------------------------- IN PROGRESS! This option is still experimental. This keyword allows to include geometry constraints for selected groups of atoms in a geometry optimization or vibrational analysis. A group of atoms is considered a rigid subunit which is allowed to move only by translation of its center of gravity (COG) or by rotation about its COG. As an example, consider a surface cluster model where (rigid) surface layers are allowed to relax in an optimization. Atom (see d.5)) and group constraints can be mixed as long as they are consistent. In doubtful cases (e.g. atoms fixed by atom constraints but allowed to move by a group constraint, atoms move in different directions by an atom and a group constraint) group constraints take precedence over atom constraints. NOTE that group definitions and operations are allowed (and meaningful) only for systems with > 2 atoms. Further, group rotation is ignored for systems with symmetry > C1. The keyword line GROUp DEFine na i1 i2 ... ing defines a group of na atoms i1 ... ing. If symmetry is used, atom numbers i1 ... ing refer to non-equivalent centers with their symmetry equivalents being also included such that the global symmetry is conserved. The keyword line can contain up to 19 atom indices. If more than 19 atoms are to be included in the group, the index list can be continued on the following line(s) (up to 20 indices per line). The group is assigned a group number ng according to the sequence of group definitions. This number is refered to below. If no other constraint of group ng (after its definition) is given, the group is allowed to move rigidly by translation of its COG as well as by free rotation about its COG. Note that group definitions with only one atom center (counting also symmetry equivalents) ignore rotation and consider translation only. In this case only combinations GROUP DEF / GROUP FIX and GROUP TRA are meaningful and can be dealt with by atom constraints, see keyword ATOM in d.5). Group rotation is also ignored for systems with symmetry > C1, i. e. GROUP ROT => GROUP FIX, GROUP MOV => to GROUP TRA. GROUp FIX ng fixes all atom centers of group ng at their positions given in the input file (or in the restart file if option CONTinue is used, see keyword RUNType, c.1)). This does not affect atom centers outside the group. Fixing atom centers overrides other geometry constraints. GROUp TRAnslate ng [x1 y1 z1 [x2 y2 z2]] allows group ng to move rigidly by translation of its COG only without rotation about its COG. The translation can be restricted further by direction contraints where v1 = (x1, y1, z1) and v2 = (x2, y2, z2) are direction vectors in cartesian coordinates. If only v1 is given on the keyword line, the COG is constrained to move along the line defined by its starting position and direction vector v1. If both v1 and v2 appear on the keyword line, the COG is constrained to move along the plane defined by its starting position and the two direction vectors v1, v2. GROUp ROTate ng [x1 y1 z1] allows group ng to rotate rigidly about its COG only without translation of its COG. The rotation can be restricted further by an axis contraint where v1 = (x1, y1, z1) is the direction vector (in cartesian coordinates) of a fixed rotation axis through the COG. Group rotation is ignored for systems with symmetry > C1, i. e. GROUP ROT => GROUP FIX. GROUp MOVe ng [DIR x1 y1 z1 [x2 y2 z2]] [AXIs x3 y3 z3] allows rigid group ng to both translate its COG and to rotate about its COG. The movement can be restricted further by direction and axis contraints analogous to those for translation or rotation only. Here v1 = (x1, y1, z1) and v2 = (x2, y2, z2), following keyword DIR, are direction vectors in cartesian coordinates. If only v1 is given, the COG is constrained to move along the line defined by its starting position and direction vector v1. If both v1 and v2 appear, the COG is constrained to move along the plane defined by its starting position and the two direction vectors v1, v2. Vector v1 = (x1, y1, z1), following keyword AXIs, is a direction vector (in cartesian coordinates) of a fixed rotation axis through the COG. Group rotation is ignored for systems with symmetry > C1, i. e. GROUP MOV ... => GROUP TRA x1 y1 z1 [x2 y2 z2]. Note that atoms can appear in only one group definition (different group definitions must not overlap). Keyword lines with additional constraints (FIX, TRAnslate, ROTate, MOVe) must appear always after the group definition line(s) in the input file. If the input file includes different constraints (FIX, TRAnslate, ROTate, MOVe) for the same atom group the constraint appearing last will be used. Examples: >GROUP DEFINE 5 1 3 4 5 7 Defines a rigid group of 5 atom centers # 1, 3, 4, 5, 7 and allows it to move rigidly by translation of its COG as well as by rotation about its COG. >ATOM FIX ALL >GROUP DEFINE 2 2 6 >GROUP TRANSLATE 1 0. 0. 1. >GROUP DEFINE 5 1 3 4 5 7 >GROUP TRANSLATE 2 >GROUP DEFINE 2 8 9 Fixes all atom centers except those of the three groups defined subsequently. Defines a rigid group (# 1) of 2 atom centers # 2, 6 which is allowed to move by translation of its COG along the (0,0,1) direction without rotation. Further, a rigid group (# 2) of 5 atom centers # 1, 3, 4, 5, 7 is defined allowing for translation in any direction. Finally, a rigid group (# 3) of 2 atom centers # 8, 9 is defined allowing for both translation and rotation. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.7) HESSian [number [scale]] ----------------------------------------------------------------------------- This keyword defines the method used to update the Hessian matrix between successive steps in a geometry optimization. The following methods are available: number = 0 Steepest descent (extrapolation mixing alpha = 1.0) 1 Steepest descent (extrapolation mixing alpha = 0.5) 2 Conjugate gradient (extrapolation mixing alpha = 1.0) 3 Conjugate gradient (extrapolation mixing alpha = 0.5) 4 Conjugate gradient, Polak-Ribiere variant 5 Quasi-Newton, Davidson-Fletcher-Powell variant 6 Quasi-Newton, Murtaugh - Sargent variant 7 Quasi-Newton, Broyden-Fletcher-Goldfarb-Shanno variant (default) 8 Quasi-Newton, BFGS variant with line search (experimental!) With update methods number 0, 1, 2, 3 the extrapolation mixing factor alpha may be set explicitly using the (optional) parameter scale. For values scale < 0.5 this may result in an increased number of convergence steps but smoother convergence. Examples: >HESSIAN 7 >HESSIAN 3 0.7 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.8) GRADients [VERSluis, STANdard] ----------------------------------------------------------------------------- In the energy gradient evaluation, all contributions are calculated analytically except for the exchange-correlation (XC) contribution. The latter is determined numerically and introduces an error which can be compensated partly by the Versluis correction [VZ88]. The keyword line GRADients VERSluis eliminates the spurious 1-center XC contributions to the energy gradients. This Versluis correction leads, in many cases, to slightly more exact energy gradients at the expense of increased computational effort (cpu time increased by about 30%) GRADients STANdard does not include the Versluis correction. This is the default setting. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) d.9) STRBalsac ----------------------------------------------------------------------------- This keyword refers to BALSAC [H00] format file output of the cluster/molecule geometry during a geometry optimization. By default, BALSAC structure plot files are generated at the beginning (file struc000.plt) and at the end (file strucfin.plt) of a geometry optimization run. If keyword STRBalsac is included, a BALSAC plot file (file strucXXX.plt) will be generated at each geometry step where XXX denotes the step number (filled by zeros to form a 3 digit number, an example file name is struc007.plt). The output files can be used as input to BALSAC where the atoms of a cluster/molecule are displayed as shaded balls. 2.2.5. KEYWORDS FOR PROPERTIES (Goto TOC, KEYW, KEYA) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.1) MULLiken [OFF / ON [FULL, ORBItal] ] ----------------------------------------------------------------------------- This keyword includes a listing of the gross Mulliken population analysis and Mayer bond orders at the end of SCF calculations or geometry optimizations. By default, the total Mulliken population analysis and total Mayer bond orders, i. e. values summed over all occupied orbitals, are included. In addition, orbital resolved data can be requested where the analysis is performed for all occupied and virtual orbitals included in the listing, see also f.9). For NEXAFS calculations all excitation orbitals with intensity greater than 0.001 are included in the listing. The analysis is performed in spherical or cartesian gaussians depending on the internal representation of d basis functions used in the electronic state calculation, see option ORBItalchoice (f.4). Keyword line MULLiken ON (or MULLiken) lists total Mulliken populations and total Mayer bond orders. This is the default if no keyword is given. MULLiken OFF lists neither populations nor bond orders. MULLiken ON ORBItal lists total and orbital resolved Mulliken populations as well as total Mayer bond orders. MULLiken ON FULL lists total and orbital resolved Mulliken populations as well as total and orbital resolved Mayer bond orders. Examples: >MULLIKEN ON FULL full Mulliken population and Mayer bond order analysis. >MULLIKEN OFF no Mulliken population nor bond order analysis. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.2) LOEWdin [ ON [FULL, ORBItal] ] ----------------------------------------------------------------------------- This keyword includes a listing of the Loewdin population analysis and Mayer bond orders at the end of SCF calculations or geometry optimizations. By default, no Loewdin population analysis is included. In addition, orbital resolved data can be requested where the analysis is performed for all occupied and virtual orbitals included in the listing, see also f.9). For NEXAFS calculations all excitation orbitals with intensity greater than 0.001 are included in the listing. The analysis is performed in spherical or cartesian gaussians depending on the internal representation of d basis functions used in the electronic state calculation, see option ORBItalchoice (f.4). Keyword line LOEWdin OFF lists neither populations nor bond orders. This is the default if no keyword is given. LOEWdin ON (or LOEWdin) lists total Loewdin populations and total Mayer bond orders. LOEWdin ON ORBItal lists total and orbital resolved Loewdin populations as well as total Mayer bond orders. LOEWdin ON FULL lists total and orbital resolved Loewdin populations as well as total and orbital resolved Mayer bond orders. Examples: >LOEWDIN ON FULL full Loewdin population and Mayer bond order analysis. >LOEWDIN OFF no Loewdin population nor bond order analysis. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.3) MEPFit [VANDerwaals] (not available!) ----------------------------------------------------------------------------- This keyword includes the calculation of effective atom charges based on electrostatic potentials evaluated from fitted and numerically evaluated charge densities (Method of Singh and Kollman [KLS96]). The methods uses default Van der Waals radii which are stored internally. Modified Van der Waals radii can be chosen together with option keyword VANDerwaals as follows MEPFit VANDerwaals nat1 c nat2 radVdW2 ... END redefines the Van der Waals radii of atomic number nat1, nat2, ... by radVdW1, radVdW2, ... where the definition list is terminated by one line containing the keyword END. Example: >MEPF evaluates effective atom charges based on default Van der Waals radii. >MEPF VANDER evaluates effective atom charges based on > 1 1.0 default Van der Waals radii except for H > 5 2.0 (Z=1) and B (Z=5) atoms whose radii are >END redefined as 1.0 and 2.0, respectively. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.4) BALPopulation [TOTAl, FUKUi, ALFA, BETA] ----------------------------------------------------------------------------- This keyword initiates BALSAC [H00] format file output to visualize and analyze atom charges derived from Mulliken/Loewdin populations as well as atom-projected Fukui indices. The output files can be used as input to BALSAC where the atoms of a cluster/molecule are displayed as shaded balls. The ball radii are defined by the actual atom charges and positive/negative charges are shown by different colors (red/blue by default) in the plot. The selection of Mulliken or Loewdin populations is determined by the print output choice, see options MULLiken (e.1), LOEWdin (e.2). Note that the present option requires appropriate population (either Mulliken or Loewdin or both, see examples below) print output. The keyword line BALPOP TOTAL [ALFA, BETA, DIFF] generates BALSAC file output, poptotM.plt and/or poptotL.plt, with atom radii given by total atom charges (absolute values) calculated from Mulliken and/or Loewdin populations. Negative / positive atom charges are shown by different colors (green / blue by default, coresponding to ball "charges" Z = 1 / = 2) in the plot. Optional keyword 'ALFA' uses spin alpha projected atom charges, defined by qnuc/2 - pop(alpha) for the plot. Optional keyword 'BETA' uses spin beta projected atom charges, defined by qnuc/2 - pop(beta) for the plot. Optional keyword 'DIFF' uses spin differences per atom, defined by pop(alpha) - pop(beta) for the plot. BALPOP FUKUI generates BALSAC file output, fukrlax.plt, fukfroz.plt, with atom radii given by atom-projected Fukui indices (relaxed and frozen orbital values, respectively). Negative / positive atom charges are shown by different colors (green / blue by default, coresponding to ball "charges" Z = 1 / = 2) in the plot. NOTE that this option requires the FUKUI option to be activated, see b.4), b.5). BALPOP ALFA na i1 i2 ... ina generates a set of BALSAC output files, populMAiii.plt and/or populLAiii.plt, iii= i1, i2 ... ina, for na alpha orbitals. Here, for each alpha orbital ii, atom radii are determined by respective Mulliken and/or Loewdin populations (<0 and <1) describing the orbital character and location inside the system. The orbital indices ii reflect the energetic order of the levels (over all symmetries) and are identical to those used in the Mulliken/Loewdin analysis of the print output. The keyword line can contain up to 19 orbital indices. If na > 19 alpha orbitals are to be included, the index list can be continued on the following line(s). BALPOP BETA nb i1 i2 ... inb generates a set of BALSAC output files, populMBiii.plt and/or populLBiii.plt, iii= i1, i2 ... inb, for nb beta orbitals. Information and input format are identical to that of alpha orbitals described above. Examples: >MULLIKEN ON performs short Mulliken analysis printout >BALPOP TOTAL and generates a BALSAC file poptotM.plt >LOEWDIN ON performs short Loewdin and full Mulliken >MULLIKEN ON FULL analysis printout, then generates BALSAC >BALPOP BETA 2 101 102 files populMB101.plt, populMB102.plt (Mulliken output) and populLB101.plt, populLB102.plt (Loewdin output) for orbitals 101, 102. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.5) DRAW [MEPS, DENSity, WAVEfunction, SFUNction, XCENergy, XCPOtential, FUKUifunction, FFUKuifunction, VOROnoi, BADEr] ----------------------------------------------------------------------------- This keyword allows to calculate electrostatic potentials, orbital and total densities, spin densities, orbitals, XC densities, and more on a regular grid. The corresponding function values are stored in files (binary unit 88, ASCII unit 98, format details are given below) and can be used later on for graphical representations or analysis purposes, see e. g. utility khpltps. All function values are given in atomic units whereas coordinates and lengths saved in file units 88, 98 refer to Angstrom or Bohr units depending on the definitions used to describe atom positions, see keywords CARTesian in a.1) and ZMATrix in a.2). Several DRAW options can be combined in a single StoBe run except for electrostatic potential plots inside 3-dimensional cubes/polyhedra, see (A) below. NOTE that restart files of converged SCF runs or geometry optimizations can be used for generating plot output files without further iterations by setting >RUNType CPROperties , see keyword RUNType or alternatively >RUNType RESTart , see keyword RUNType >MAXCycles 0 , see keyword MAXCycles (A) Electrostatic potentials The keyword line and following numerical line(s) DRAW MEPS POINts np x1 y1 z1 ... determines the electrostatic potential at np specific points (xi,yi,zi), i = 1, np, whose coordinates are on the following lines (one line per point). The coordinates are defined in Angstrom or Bohr units depending on the definitions used to describe atom positions, see keywords CARTesian and ZMATrix. Note that this option does not generate binary file output. For details of the output files see below. DRAW MEPS PLANe xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines the electrostatic potential on an equidistant mesh inside a rectangular section on a plane. The plane is defined by the three positions R0=(x0,y0,z0), R1=(x1,y1,z1), R2=(x2,y2,z2) where R0 forms the drawing origin, the x axis points along (R1-R0), and the y axis along (R2-R0)' = [(R1-R0)x(R2-R0)]x(R1-R0) such that (R1-R0)*(R2-R0)'=0. The path from R0 to R0+xlen*(R1-R0)/|(R1-R0)| is subdivided into nx points, that from R0 to R0+ylen*(R2-R0)'/|(R2-R0)'| into ny points, forming altogether an nx*ny point mesh over an area of size xlen by ylen. The position vectors and lengths are defined in Angstrom or Bohr units depending on the definitions used to describe atom positions, see keywords CARTesian and ZMATrix. Any of the three triplets (xi, yi, zi) can be replaced by an integer number ni pointing to an atom center ni whose coordinates will be used for (xi, yi, zi). The area section with respect to the origin R0=(x0,y0,z0) can be defined more generally by an option line DRAW MEPS PLANe xa xe ya ye nx ny where the section refers to an area [xa, xe] x [ya, ye] wrt. the origin. (The above short form uses an area [0, xlen] x [0, ylen].) Note that this option generates binary file output. For details of the output files see below. DRAW MEPS CUBE x0 y0 z0 xlen ylen zlen nx ny nz determines the electrostatic potential on an equidistant mesh inside a cube. The cube is defined by its origin R0=(x0,y0,z0) and its lengths along x (xlen), y (ylen), and z (zlen), i. e. a [x0,x0+xlen] x [y0,y0+ylen] x [z0,z0+zlen] block. The block sides are subdivided into nx, ny, nz points, respectively, forming altogether an nx*ny*nz point mesh. The position vectors are defined in Angstrom or Bohr units depending on the definitions used to describe atom positions, see keywords CARTesian and ZMATrix. The printout includes the minimum value of the MEP within the cube. Note that this option does not generate binary file output. For details of the output files see below. NOTE that this DRAW option cannot be combined with other DRAW task in the same StoBe run. DRAW MEPS GCUBE x0 y0 z0 x1 y1 z1 n1 x2 y2 z2 n2 x3 y3 z3 n3 determines the electrostatic potential on an equidistant mesh inside a polyhedron spanned by vectors R1 = (x1, y1, z1), R2 = (x2, y2, z2), and R3 = (x3, y3, z3) starting from the origin at R0 = (x0, y0, z0). (If the vectors R1, R2, R3 are mutually perpendicular the polyhedron is rectangular.) The mesh points are given by r = R0 + i*R1/(n1-1) + j*R2/(n2-1) + k*R3/(n3-1) where i, j, k run between 0 and n1-1, n2-1, and n3-1, respectively which yields an n1*n2*n3 point mesh. The position vectors are defined in Angstrom or Bohr units depending on the definitions used to describe atom positions, see keywords CARTesian and ZMATrix. The ASCII output (unit 98) is given in the Gaussian Cube file format (see http://astronomy.swin.edu.au/~pbourke/geomformats/cube/) and can be used for example as input to the visualisation tool Molekel [MOL03]. This option does not generate binary file output. NOTE that this DRAW option cannot be combined with other DRAW task in the same StoBe run. DRAW MEPS SURF [isur] determines the electrostatic potential on any surface defined by grid points which are provided separately in a file (units isur, default unit = 9) with fixed format (I5) > ntot (3F12.6) > xi yi zi , i = 1, ntot The position vectors (xi,yi,zi) are defined in Angstrom or Bohr units depending on the definitions used to describe atom positions, see keywords CARTesian and ZMATrix. Note that this option does not generate binary file output. For details of the output files see below. All electrostatic potential values are give in atomic units (Hartree, corresponding to 27.211608 eV). (B) Densities The keyword line and following numerical line(s) DRAW DENSity ORBItal xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) norb n1 n2 ... determines densities of norb (up to 200) orbitals on an equidistant mesh inside a rectangular section on a plane where orbitals are defined by their indices n1, n2, ... The input line can contain up to 20 indices pointing to orbitals according to their energetic order given in the listing. As an example, n1 = 4 selects the fourth orbital in energetic sequence, irrespective of its symmetry label. In the output listing of a drawing task the orbital indexing is shown as Alpha dens. ni ( nfia): min= ... Beta dens. ni ( nfib): min= ... where ni refers to the energetic order (given in the input) while numbers nfia, nfib in parentheses denote the positions of the alpha, beta spin orbitals in the restart file. If norb > 20 orbitals are to be included, the index list can be continued on the following line(s). The plane is defined as described for electrostatic potentials. The last input line "norb n1 n2 ..." may be replaced by a line HOLU nocc nempty which uses orbital densities of the nocc highest occupied levels and nempty lowest unoccupied levels for input (nocc = 0, nempty = 0 is accepted). For details of the output files see below. DRAW DENSity MIXOrbitals xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) norb n1 ns1 q1 n2 ns2 q2 ... determines a weighted superposition of norb (up to 200) orbital densities on an equidistant mesh inside a rectangular section on a plane. The corresponding orbital contributions are each defined by index triples (ni nsi qi) where ni gives the orbital index analogous to the definition used for orbital densities, see above. Parameter nsi defines the spin assignment where nsi = 1 (2) selects the alpha (beta) spin orbital and nsi = 3 uses both alpha and beta spin orbitals. Each orbital contribution is weighted by a factor qi. As an example, the triple (17 2 0.5) defines a density contribution of beta spin orbital no. 17 (energetic order, see above) with a weight of 0.5. The input line can contain up to 6 index triples. If norb > 6 orbital contributions are to be included, the index list can be continued on the following line(s). The plane is defined as described for electrostatic potentials. DRAW DENSity TOTAl xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines the total electron density (numerical and fitted) on an equidistant mesh inside a rectangular section on a plane. The plane and section are defined as described for electrostatic potentials. For details of the output files see below. DRAW DENSity EXchangecorrelation xlen ylen nx ny [ xa xe ya ...] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines the Ex, Ec, Exc energy densities on an equidistant mesh inside a rectangular section on a plane. The plane is defined as described for electrostatic potentials. For details of the output files see below. All density values are give in atomic units (e/Bohr**3, corresponding to 6.748343 e/Angstrom**3, for orbital and total densities, H/Bohr**3, corresponding to 183.633252 eV/Angstrom**3, for energy densities). (C) Orbitals The keyword line and following numerical line(s) DRAW ORBItal xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) norb n1 n2 ... determines wavefunction values of norb (up to 200) orbitals on an equidistant mesh inside a rectangular section on a plane where orbitals are defined by their indices n1, n2, ... The input line can contain up to 20 indices pointing to orbitals according to their energetic order given in the listing. As an example, n1 = 4 selects the fourth orbital in energetic sequence, irrespective of its symmetry label. In the output listing of a drawing task the orbital indexing is shown as Alpha orb. ni ( nfia): min= ... Beta orb. ni ( nfib): min= ... where ni refers to the energetic order (given in the input) while numbers nfia, nfib in parentheses denote the positions of the alpha, beta spin orbitals in the restart file. If norb >20 orbitals are to be included, the index list can be continued on the following line(s). The plane and section are defined as described for electrostatic potentials. The last input line "norb n1 n2 ..." may be replaced by a line HOLU nocc nempty which uses orbitals of the nocc highest occupied levels and nempty lowest unoccupied levels for input (nocc = 0, nempty = 0 is accepted). For details of the output files see below. All orbital values are give in atomic units (Bohr**(-3/2), corresponding to 2.597757 Angstrom**(-3/2)). (D) Spin difference densities DRAW SFUNction xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines the total spin density difference (rho(alpha)-rho(beta) from numerical spin densities) on an equidistant mesh inside a rectangular section on a plane. The plane and section are defined as described for electrostatic potentials. For details of the output files see below. All spin density values are give in atomic units (spin/Bohr**3, corresponding to 6.748343 spin/Angstrom**3). (E) Exchange and correlation functions DRAW XCENergy xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines Ex, Ec and Exc energies on an equidistant mesh inside a rectangular section on a plane. The plane and section are defined as described for electrostatic potentials. For details of the output files see below. DRAW XCPOtential xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines Vx, Vc and Vxc potentials on an equidistant mesh inside a rectangular section on a plane. The plane and section are defined as described for electrostatic potentials. All energy values are give in atomic units (Hartree, corresponding to 27.211608 eV). (F) Fukui functions DRAW FUKUifunction xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines the (relaxed) Fukui function f+(r) or f-(r) (depending on the sign of parameter qfukui, see b.5)) on an equidistant mesh inside a rectangular section on a plane. Here f+(r) / f-(r) include orbital relaxation. The plane and section are defined as described for electrostatic potentials. For details of the output files see below. NOTE that plots of Fukui functions require the FUKUI option to be activated, see b.4), b.5). DRAW FFUKuifunction xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines the (frozen) Fukui function f+(r) or f-(r) (depending on the sign of parameter qfukui, see b.5)) on an equidistant mesh inside a rectangular section on a plane. Here f+(r) / f-(r) include only occupation changes without orbital relaxation (orbitals of the restart input file are used). The plane and section are defined as described for electrostatic potentials. For details of the output files see below. NOTE that plots of Fukui functions require the FUKUI option to be activated, see b.4), b.5). All Fukui function values are give in atomic units (e/Bohr**3, corresponding to 6.748343 e/Angstrom**3). (G) Miscellaneous DRAW VOROnoi xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines the index of the closest atom for each point on an equidistant mesh inside a rectangular section on a plane. This can be used to draw the boundaries of the molecular Voronoi cells along the plane. The plane and section are defined as described for electrostatic potentials. For details of the output files see below. DRAW BADEr xlen ylen nx ny [ xa xe ya ye nx ny] x0 y0 z0 (or nc0) x1 y1 z1 (or nc1) x2 y2 z2 (or nc2) determines the index of the Bader attractor atom for each point on an equidistant mesh inside a rectangular section on a plane. This can be used to draw the boundaries of the molecular Bader cells along the plane. The plane and section are defined as described for electrostatic potentials. For details of the output files see below. FORMAT DETAILS OF DRAW OUTPUT FILES The combination of DRAW tasks of a StoBe run generates output in binary and ASCII files which can be inspected and used as input to subsequent applications, see e. g. utility khpltps. The file formats are described in the following (a) Binary output files (unit 88) The basic file format consists of ntask sets of records 1 - nx described below where ntask is the total number of DRAW tasks defined in the StoBe input (parameter types are I4=INTEGER*4, R8=REAL*8, Cn=CHARACTER*n): Record 1 (C80) TASKTITL task title. Possible titles start with the run title followed by 'elstat. potential' for plots of the electrostatic potential along a plane. 'rho alpha orb.# nn' and 'rho beta orb.# nn' for orbital density plots of alpha / beta orbital no. nn where nn denotes the orbital index in energetic sequence (shown in the listing). 'total density (num.)' and 'total density (fit.)' for total density plots (all occupied orbitals) evaluated numerically or by fitting. 'Exchange energy density', 'Correlation energy density' and 'Combined XC energy density' for plots of Ex, Ec, Exc energy densities. 'wvf alpha orb.# nn' or 'wvf beta orb.# nn' for orbital plots of alpha / beta orbital no. nn where nn denotes the orbital index in energetic sequence (shown in the listing) 'spin difference density' for plots of the total spin density difference. 'Exchange energy function', 'Correlation energy function' and 'Combined XC energy function' for plots of Ex, Ec and Exc energy distributions. 'Exchange potential', 'Correlation potential' and 'Combined XC potential' for plots of Vx, Vc and Vxc potential distributions. Record 2 (2*R8) XA, XE starting and ending point along the x direction (in Bohr or Angstrom defined by the coordinate input, see a.1) or a.2)) of the rectangular plot section. The length of the section along x is given by (XE - XA). (2*R8) YA, YE starting and ending point along the y direction (in Bohr or Angstrom defined by the coordinate input, see a.1) or a.2)) of the rectangular plot section. The length of the section along y is given by (YE - YA). (I4) NX number of points in the equidistant mesh used along the x direction (I4) NY number of points in the equidistant mesh used along the y direction (I4) NPTS block size (e.g. 128) used in the blocking of the function array of record 3. Records 3 (NX*NY*R8) ARRAY(I) function array (in atomic units) in subrecords for all grid points where a blocking of NPTS numbers per subrecord is used. The last subrecord may contain less than NPTS numbers. Record 4 (I4) NCENTER number of atom centers of the cluster / molecule. Record 5 (C4) LABCEN label of each atom center (R8) XPROJ x coordinate of each atom center wrt. plot plane (R8) YPROJ y coordinate of each atom center wrt. plot plane (R8) ZPROJ z coordinate of each atom center wrt. plot plane where all centers of the cluster / molecule are collected in one record (LABCEN(I),XPROJ(I),YPROJ(I),ZPROJ(I), I=1,NCENTER). The coordinates are given in Bohr or Angstrom defined by the coordinate input, see a.1) or a.2). (b) ASCII output files (unit 98) This output format is self-explanatory and needs no further documentation. All coordinates and lengths are given in Bohr or Angstrom defined by the coordinate input, see a.1) or a.2). ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.6) PSCRatch ----------------------------------------------------------------------------- Four SCF calculations are used to determine the polarization of a system, see option RUNtype, c.1), where, by default, the three SCF runs no. 2 - 4 are carried out as subsequent restart cases (except for free atoms/ions). This saves time but can introduce slight errors depending of the SCF convergence threshold. SCF runs no. 2 - 4 can be forced to start from scratch by including the keyword line >PSCRatch In polarization calculations on free atoms/ions SCF runs no. 2 - 4 will always start from scratch. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.7) XRAYspectra [XAS, XES, RIXS] ----------------------------------------------------------------------------- This keyword allows the calculation of core - valence and core - unoccupied transition energies and corresponding dipole transition probabilities which can be used to evaluate theoretical X-ray spectra (XAS = X-ray absorption, XES = X-ray emission, RIXS = resonant ionization Xray scattering). The three types of calculations are described in the following. (A) Xray absorption spectra (XAS, NEXAFS, XANES) The keyword line XRAYspectra XAS allows the calculation of X-ray absorption spectra (XAS, NEXAFS) based on the transition state approach using the Kohn-Sham orbitals of the present StoBe run. The option requires additional keyword input, described below, as well as additional basis set input for the augmentation basis to be used to describe the excitation orbitals of the final states. The augmentation basis is specified at the end of the basis set input (section 2.3). Further, the electronic configuration of the StoBe run must correspond to that of a transition state where the corresponding core hole level is occupied with q = 0.5 electrons. For further background on the techniques to compute X-ray spectra consult references [TPA98], [TLP99], [FBH00], [PAK00], [KPP01]. Additional keyword lines are REMThreshold remthr defining the threshold (remthr) for linear dependencies in the augmentation basis set used in the NEXAFS calculation. The default value of remthr (used if this line is omitted) is 0.000001. SHIFtenergy eshft defining an energy eshft (in eV) which is used to shift ionization potentials and all excitation energies by the same amount. This rigid shift can be used to correct for the difference between orbital energy and real ionization potential. This is done in most cases in the subsequent graphics program. The default value of eshft (used if this line is omitted) is 0.0. END terminating the XAS input. This line must be present even if default values for the parameters remthr, eshft are chosen and the corresponding lines are omitted. NOTE that the 'END' line must always be present in the input. The XAS data are, apart from print output, also saved in a spectrum file on unit 11 (fort.11, ASCII format) containing excitation energies, oscillator strengths, and transition moments. The format is Line 1 (A5,I5) Type,Ntrans Type = 'XAS ' characterizes the subsequent data as referring to Xray absorption. Ntrans number of core excitations listed below. Line(s) 2 (F22.10,2F22.16,9D20.12) (Eex,Osctot,OscR2,Dx,Dy,Dz, Qxx,Qxy,Qxz,Qyy,Qyz,Qzz), I=1,Ntrans describing Ntrans core excitations with Eex excitation energy in Hartree units. The energy may be shifted wrt the level energy difference by eshft, see above. Osctot total dipole oscillator strength of the excitation in a.u. given by Osctot = 2/3 * (Eex-eshft) * ( <final|x|core>**2 + <final|y|core>**2 + <final|z|core>**2 ) OscR2 transition moment of the hypothetical R**2 operator, given by OscR2 = 2/3 * (Eex-eshft) * <final|x**2+y**2+z**2|core>**2 This quantity can be helpful in placing the s-components while it does not correspond to a physical operator. Dx, Dy, Dz cartesian components of the dipole transition moment, given by Dx = <final|x|core> , Dy = <final|y|core>, Dz = <final|z|core> These components allow the evaluation of total and angular resolved spectra using e. g. utility xrayspec, see Sec. 2.7.6. Qxx, Qxy, Qxz, Qyy, Qyz, Qzz cartesian components of the quadrupole transition moment tensor, given by Qxx = <final|xx|core> , Qxy = <final|xy|core>, Qxz = <final|xz|core> , Qyy = <final|yy|core>, Qyz = <final|yz|core> , Qzz = <final|zz|core>, These components allow the evaluation of total and angular resolved spectra using e. g. utility xrayspec, see Sec. 2.7.6. Example #1: >TITLE >h2o O1s excitation based on transition potential for 1s orbital >SYMM C2V >CARTESIAN ANGSTROM >O 0.000000 0.000000 0.212830 8. 32 >H 0.699763 0.000000 -0.451321 1. 32 >END >RUNTYPE START NOOPTIMIZE >SCFTYPE DIRECT >POTENTIAL NONLOCAL PD86 PD86 >GRID FINE NONRANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 25 >ECONVERGENCE 0.000001 >DCONVERGENCE 0.00001 >DMIXING MDENS 0.40 >MULLIKEN ON >DIIS ON >FSYM SCFOCC EXCITED >ALFA 3 0 1 1 >BETA 3 0 1 1 >SYM 1 >ALFA 3 1 1 0.5 >BETA 3 0 >END >XRAY XAS >END >END >A-OXYGEN (4,4;4,4) >A-HYDROGEN (4,2;4,2) >O-OXYGEN (521/41) >O-HYDROGEN (311/1) >X-FIRST >X-DUMMY >END For this example the unit 11 output file contains a first line 'XAS 153' followed by 153 lines listing quantities Eex(H), Osctot, OscR2, Dx, Dy, Dz, Qxx, Qxy, Qxz, Qyy, Qyz, Qzz for each transition (all numbers of the first 6 and last 3 lines are cut off after 4 digits, quadrupole moments are listed on a separate line): 19.76416 0.0067 0.0051 -0.5028D-09 -0.5612D-16 0.2271D-01 -0.8124D-02 -0.1282D-15 0.1727D-10 -0.4181D-02 0.3573D-16 -0.7436D-02 19.84634 0.0160 0.0000 -0.3487D-01 0.7084D-16 -0.3753D-09 0.1766D-09 0.3402D-13 0.8079D-03 0.5579D-10 -0.9250D-16 0.1185D-09 19.89767 0.0012 0.0000 0.2237D-15 0.9888D-02 0.8847D-16 -0.1832D-15 -0.1031D-09 -0.1956D-15 -0.3439D-16 0.4446D-02 -0.2149D-15 19.90631 0.0011 0.0012 -0.3289D-09 0.8310D-16 0.9328D-02 -0.4724D-02 -0.5224D-15 -0.4374D-10 -0.6298D-02 0.4344D-16 0.1506D-02 19.91228 0.0008 0.0297 0.4479D-09 0.1210D-15 -0.8097D-02 0.1254D-01 0.1532D-15 -0.2152D-09 0.1552D-01 -0.6309D-17 0.1930D-01 19.92696 0.0014 0.0000 -0.1050D-01 0.5735D-15 -0.4928D-09 -0.5297D-10 0.4479D-13 0.2830D-02 0.9416D-09 -0.9011D-15 0.5393D-09 ........................................................................... 34.12209 0.0936 0.0000 0.1447D-10 0.1645D-11 -0.6417D-01 0.1502D-03 0.2221D-12 -0.7141D-12 -0.1666D-03 -0.3533D-13 0.3377D-03 34.17390 0.0931 0.0000 -0.4366D-11 -0.6393D-01 -0.1644D-11 0.7875D-13 -0.6558D-07 0.8241D-13 -0.6524D-13 0.9139D-04 0.3560D-14 75.92744 0.0000 0.0104 -0.9598D-13 0.4462D-14 0.1602D-03 0.4808D-02 -0.1312D-12 0.1086D-13 0.4775D-02 0.2016D-13 0.4806D-02 where the quantities have been described above. From this file total or angle resolved NEXAFS spectra can be generated by convolution using utility xrayspec or freefwhm (for the latter see subdirectory XRAY/Convolute of the present release). NOTE: In the output listing both the transition moment OscR2 and the expectation value of R**2 for the final excited orbital are given. The latter can be used to estimate the spatial extent of the excited state which can provide some information on the valence versus Rydberg character of the excitation. Example #2: >TITLE >Four water molecules in bilayer of ice. >Use ECP for the 1s on the oxygens in the waters that are >not considered for the excitation. >SYMM C1 >CARTESIAN ANGSTROM >O -2.686528 4.321350 .000000 6. 32 >O 1.343264 4.321350 2.326602 6. 32 >O 1.343264 4.321350 -2.326602 6. 32 >O1 .031838 3.470107 .035066 8. 32 >H -3.075511 3.967239 .803056 1. 32 >H -1.785863 4.025866 -.022379 1. 32 >H 2.248521 4.001804 2.327158 1. 32 >H .891118 4.001804 3.110856 1. 32 >H 2.233222 3.967239 -2.261942 1. 32 >H .876184 4.034834 -1.552115 1. 32 >H1 .059902 2.527293 .094088 1. 32 >H1 .464274 3.812810 .810402 1. 32 >END >FSYM SCFOCC EXCITED >ALFA 17 >BETA 17 >SYM 1 >ALFA 17 1 1 0.5 >BETA 17 0 >END >RUNTYPE START NOOPT >SCFTYPE DIRECT >POTENTIAL NONLOCAL BE88 PD86 >GRID FINE RANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 100 >XRAY XAS >REMTHR 1.E-07 >END >ECONV 0.00001 >DCONV 0.005 >DMIX MDENS 0.05 >MAXOVERLAP >MULLIKEN ON >DIIS ON >PRINTOUT OFF >END >A-OXYGEN (4,4;4,4) >A-OXYGEN (4,4;4,4) >A-OXYGEN (4,4;4,4) >A-OXYGEN (4,4;4,4) >A-HYDROGEN (3,1;3,1) >A-HYDROGEN (3,1;3,1) >A-HYDROGEN (3,1;3,1) >A-HYDROGEN (3,1;3,1) >A-HYDROGEN (3,1;3,1) >A-HYDROGEN (3,1;3,1) >A-HYDROGEN (3,1;3,1) >A-HYDROGEN (3,1;3,1) >O-OXYGEN(+6) (311/211) >O-OXYGEN(+6) (311/211) >O-OXYGEN(+6) (311/211) >O-OXYGEN iii_iglo >O-HYDROGEN (311/1) >O-HYDROGEN (311/1) >O-HYDROGEN (311/1) >O-HYDROGEN (311/1) >O-HYDROGEN (311/1) >O-HYDROGEN (311/1) >O-HYDROGEN (311/1) >O-HYDROGEN (311/1) >P-OXYGEN(+6) (3,1:10,0) >P-OXYGEN(+6) (3,1:10,0) >P-OXYGEN(+6) (3,1:10,0) >X-DUMMY >X-DUMMY >X-DUMMY >X-FIRST >X-DUMMY >X-DUMMY >X-DUMMY >X-DUMMY >X-DUMMY >X-DUMMY >X-DUMMY >X-DUMMY >END For this example the unit 11 output file contains a first line 'XAS 243' followed by 153 lines listing quantities Eex(H), Osctot, OscR2, Dx, Dy, Dz, Qxx, Qxy, Qxz, Qyy, Qyz, Qzz for each transition (all numbers of the first 6 and last 3 lines are cut off after 4 digits, quadrupole moments are listed on a separate line): 19.71494 0.0034 0.0013 0.4055D-02 -0.1376D-01 0.7492D-02 0.3415D-02 0.1375D-04 0.5500D-04 0.3237D-02 0.2835D-04 0.3477D-02 19.75885 0.0038 0.0003 0.1858D-02 0.1663D-01 0.3095D-02 -0.1888D-02 0.5575D-04 0.7507D-04 -0.1526D-02 0.8809D-04 -0.1811D-02 19.79333 0.0013 0.0000 -0.8845D-02 0.3899D-04 0.5179D-02 0.2081D-03 -0.2086D-03 0.1182D-03 0.1486D-05 0.1167D-03 -0.2076D-03 19.80733 0.0023 0.0000 -0.1743D-02 0.1288D-01 -0.2865D-02 0.4559D-03 -0.3060D-04 -0.8058D-04 0.6798D-03 -0.5317D-04 0.3615D-03 19.82692 0.0002 0.0000 -0.3958D-02 -0.8391D-04 0.2365D-02 -0.7199D-04 -0.6073D-04 -0.6435D-04 0.3672D-04 0.3434D-04 0.1613D-03 19.83206 0.0005 0.0006 -0.3175D-02 -0.2437D-03 -0.5834D-02 -0.2343D-02 0.1125D-03 0.9331D-04 -0.2507D-02 0.2090D-03 -0.2230D-02 .......................................................................... 77.03001 0.0481 0.0000 -0.2679D-01 0.5788D-03 0.1481D-01 0.4522D-05 -0.3402D-05 0.2821D-05 -0.2157D-06 0.1723D-05 -0.3485D-05 81.83401 0.0000 0.0025 -0.4240D-04 0.5289D-04 -0.7886D-04 0.2276D-02 -0.5125D-06 -0.2065D-05 0.2276D-02 -0.1020D-05 0.2275D-02 281.08093 0.0000 0.0001 -0.7156D-06 0.8284D-06 -0.1247D-05 -0.3370D-03 -0.4186D-07 -0.4766D-06 -0.3354D-03 -0.1162D-06 -0.3369D-03 (B) Xray emission spectra (XES) The keyword line XRAYspectra XES [ncore icore(1) icore(2) ... icore(ncore)] allows the calculation of X-ray emission spectra (XES) corresponding to valence -> core level transitions based on ground state, transition state, or relaxed core hole state Kohn-Sham orbitals where the electronic configuration of the StoBe run (ground, transition, or relaxed core hole state) determines the type of approach. The calculation can include several initial core hole orbitals where ncore denotes the number of orbitals and icore(1), icore(2), ..., icore(ncore) gives their orbital indices. If no core orbitals are included in the keyword line (or ncore = 0) then only the lowest core orbital with occupation < 1 is used in the calculation. In this case the electronic configuration of the StoBe run must correspond to a transition state or a relaxed core hole state. NOTE that in XES practice the choice of ground state orbitals is preferable over that of transition state or core hole state orbitals. For further background on the techniques to compute X-ray spectra consult references [TPA98], [TLP99], [FBH00], [PAK00], KPP01]. NOTE that available orbital indices icore(i) are given in the energy level listing of the print output, column "(pos.)". Spin beta orbitals have to be incremented by an offset NLA which gives the number of occupied spin alpha orbitals. As an example, in H2O (5 occupied spin alpha orbitals) the O1s orbital has icore = 1 for spin alpha and = 6 for spin beta. The option requires additional keyword input with keyword lines SHIFtenergy eshft defining an energy eshft (in eV, default = 0.0) which is used to shift ionization potentials and all de-excitation energies by the same amount. This rigid shift can be used to correct for the difference between orbital energy and real ionization potential. This is done in most cases in the subsequent graphics program. END terminating the XES input. This line must be present even if default values for the parameters fwhm, eshft, emax are chosen and the corresponding lines are omitted. NOTE that the 'END' line must always be present in the input. The XES data are, apart from print output, also saved in a spectrum file on unit 11 (fort.11, ASCII format) containing excitation energies, intensities, and transition moments. The format is Line 1 (A5,I5) Type,Ntrans Type = 'XES ' characterizes the subsequent data as referring to Xray emission. Ntrans number of valence->core transitions listed below. Line(s) 2 (F22.10,2F22.16,9D20.12) (Eex,Osctot,OscR2,Dx,Dy,Dz, Qxx,Qxy,Qxz,Qyy,Qyz,Qzz), I=1,Ntrans describing Ntrans valence->core transitions with Eex transition energy in Hartree units. The energy may be shifted wrt the level energy difference by eshft, see above. Osctot total dipole emission intensity of the transition in a.u. given by Osctot = 1/3 * (Etr-eshft)**3 * ( <final|x|core>**2 + <final|y|core>**2 + <final|z|core>**2 ) OscR2 transition moment of the hypothetical R**2 operator, given by OscR2 = 1/3 * (Etr-eshft)**3 * <final|x**2+y**2+z**2|core>**2 This quantity can be helpful in placing the s-components while it does not correspond to a physical operator. Dx, Dy, Dz cartesian components of the dipole transition moment, given by Dx = <final|x|core> , Dy = <final|y|core>, Dz = <final|z|core> These components allow the evaluation of total and angular resolved spectra using e. g. utility xrayspec, see Sec. 2.7.6. Qxx, Qxy, Qxz, Qyy, Qyz, Qzz cartesian components of the quadrupole transition moment tensor, given by Qxx = <final|xx|core> , Qxy = <final|xy|core>, Qxz = <final|xz|core> , Qyy = <final|yy|core>, Qyz = <final|yz|core> , Qzz = <final|zz|core>, These components allow the evaluation of total and angular resolved spectra using e. g. utility xrayspec, see Sec. 2.7.6. If more than one core level is included in the XES calculation the above lines 1, 2 are saved on unit 11 for each core level sequentially. Example: >TITLE >NiCO Frozen XES calculation for carbon and oxygen 1s >SYMM C4V >CARTESIAN BOHR >Ni 0.00000 0.00000 0.00000 28.0 64 >C 0.00000 0.00000 3.17 6.0 64 >O 0.00000 0.00000 5.37 8.0 64 >END >RUNTYPE START NOOPTIMIZATION >SCFTYPE CONVENTIONAL >POTENTIAL NONLOCAL PD86 PD86 >GRID FINE RANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 70 >ECONVERGENCE 0.000001 >DCONVERGENCE 0.00001 >DMIXING MDENS 0.10 >MULLIKEN ON FULL >DIIS ON >FSYM SCFOCC >ALFA 11 0 1 1 8 >BETA 11 0 1 1 8 >XRAY XES 2 4 5 >END >END >A-NICKEL (5,8;5,8) >A-CARBON (5,2;5,2) >A-OXYGEN (5,2;5,2) >O-NICKEL (63321/5211*/311+) >O-CARBON iii_iglo >O-OXYGEN iii_iglo >END For this example the unit 11 output file contains a first line 'XES 15' describing the first set of transitions (O1s) followed by 15 lines listing quantities Etr(H), Osctot, OscR2, Dx, Dy, Dz, Qxx, Qxy, Qxz, Qyy, Qyz, Qzz for each transition (all numbers are cut off after 4 digits, quadrupole moments are listed on a separate line): 8.84003 0.0000 0.0010 0.0000D+00 0.0000D+00 -0.1393D-03 0.1329D-04 0.0000D+00 0.0000D+00 0.1329D-04 0.0000D+00 0.1382D-03 15.03307 0.0000 0.0000 0.0000D+00 0.0000D+00 0.2648D-06 0.1071D-04 0.0000D+00 0.0000D+00 0.1071D-04 0.0000D+00 0.2426D-04 16.37857 0.0000 0.0000 0.0000D+00 0.0000D+00 0.2234D-04 0.7839D-06 0.0000D+00 0.0000D+00 0.7839D-06 0.0000D+00 -0.1560D-03 16.42006 0.0000 0.0000 0.1464D-04 -0.3219D-17 0.0000D+00 0.0000D+00 0.0000D+00 -0.7093D-04 0.0000D+00 0.6555D-18 0.0000D+00 16.42006 0.0000 0.0000 -0.4121D-17 0.1464D-04 0.0000D+00 0.0000D+00 0.0000D+00 0.7471D-18 0.0000D+00 -0.7093D-04 0.0000D+00 17.82309 0.6794 3.1796 0.0000D+00 0.0000D+00 -0.1897D-01 0.1224D-01 0.0000D+00 0.0000D+00 0.1224D-01 0.0000D+00 0.1328D-01 18.34703 2.3970 0.6128 0.0000D+00 0.0000D+00 -0.3412D-01 -0.6830D-02 0.0000D+00 0.0000D+00 -0.6830D-02 0.0000D+00 -0.5992D-02 18.44251 3.0763 0.2583 0.0000D+00 0.0000D+00 0.3835D-01 0.2521D-02 0.0000D+00 0.0000D+00 0.2521D-02 0.0000D+00 0.1713D-02 18.45278 5.1045 0.0000 0.4936D-01 -0.4186D-15 0.0000D+00 0.0000D+00 0.0000D+00 -0.5723D-03 0.0000D+00 0.8086D-17 0.0000D+00 18.45278 5.1045 0.0000 0.9127D-15 0.4936D-01 0.0000D+00 0.0000D+00 0.0000D+00 -0.7844D-17 0.0000D+00 -0.5723D-03 0.0000D+00 18.63657 0.8159 0.0000 -0.1944D-01 0.2582D-15 0.0000D+00 0.0000D+00 0.0000D+00 -0.1222D-04 0.0000D+00 -0.4148D-17 0.0000D+00 18.63657 0.8159 0.0000 -0.3009D-15 -0.1944D-01 0.0000D+00 0.0000D+00 0.0000D+00 -0.1079D-16 0.0000D+00 -0.1222D-04 0.0000D+00 18.65392 0.0811 0.0001 0.0000D+00 0.0000D+00 -0.6123D-02 0.4758D-03 0.0000D+00 0.0000D+00 0.4758D-03 0.0000D+00 0.5143D-03 18.67967 0.0000 0.0000 0.0000D+00 0.0000D+00 0.5421D-19 0.3964D-04 0.0000D+00 0.0000D+00 -0.3964D-04 0.0000D+00 -0.5407D-17 18.67967 0.0000 0.0000 0.0000D+00 0.0000D+00 0.0000D+00 0.0000D+00 0.3964D-04 0.0000D+00 0.0000D+00 0.0000D+00 0.0000D+00 after which the line 'XES 14' describing the second set of transitions (C1s) followed by 14 lines listing quantities Etr(H), Osctot, OscR2, Dx, Dy, Dz, Qxx, Qxy, Qxz, Qyy, Qyz, Qzz for each transition (all numbers are cut off after 4 digits, quadrupole moments are listed on a separate line): 6.19303 0.0000 0.0000 0.0000D+00 0.0000D+00 -0.6250D-03 0.1042D-03 0.0000D+00 0.0000D+00 0.1042D-03 0.0000D+00 0.6068D-03 7.53851 0.0006 0.0008 0.0000D+00 0.0000D+00 -0.2146D-02 0.4583D-03 0.0000D+00 0.0000D+00 0.4583D-03 0.0000D+00 0.1588D-02 7.58003 0.0000 0.0000 0.2846D-03 0.5266D-17 0.0000D+00 0.0000D+00 0.0000D+00 -0.1246D-03 0.0000D+00 0.2290D-17 0.0000D+00 7.58003 0.0000 0.0000 0.3824D-17 0.2846D-03 0.0000D+00 0.0000D+00 0.0000D+00 -0.6859D-19 0.0000D+00 -0.1246D-03 0.0000D+00 8.98305 0.2990 0.4685 0.0000D+00 0.0000D+00 0.3518D-01 0.1222D-01 0.0000D+00 0.0000D+00 0.1222D-01 0.0000D+00 0.1927D-01 9.50700 0.0096 0.9993 0.0000D+00 0.0000D+00 -0.5795D-02 0.1916D-01 0.0000D+00 0.0000D+00 0.1916D-01 0.0000D+00 0.2086D-01 9.60247 1.0327 0.1540 0.0000D+00 0.0000D+00 -0.5915D-01 0.8436D-02 0.0000D+00 0.0000D+00 0.8436D-02 0.0000D+00 0.5812D-02 9.61274 0.6664 0.0000 0.4744D-01 -0.8171D-15 0.0000D+00 0.0000D+00 0.0000D+00 0.2612D-02 0.0000D+00 -0.2260D-16 0.0000D+00 9.61274 0.6664 0.0000 0.4536D-15 0.4744D-01 0.0000D+00 0.0000D+00 0.0000D+00 0.5113D-16 0.0000D+00 0.2612D-02 0.0000D+00 9.79653 0.1026 0.0000 0.1809D-01 0.2605D-15 0.0000D+00 0.0000D+00 0.0000D+00 -0.2397D-02 0.0000D+00 0.3710D-16 0.0000D+00 9.79653 0.1026 0.0000 0.7764D-15 0.1809D-01 0.0000D+00 0.0000D+00 0.0000D+00 -0.4050D-16 0.0000D+00 -0.2397D-02 0.0000D+00 9.81389 0.0061 0.0063 0.0000D+00 0.0000D+00 0.4412D-02 -0.2571D-02 0.0000D+00 0.0000D+00 -0.2571D-02 0.0000D+00 0.6986D-03 9.83964 0.0000 0.0000 0.0000D+00 0.0000D+00 0.7589D-18 0.3805D-03 0.0000D+00 0.0000D+00 -0.3805D-03 0.0000D+00 0.1084D-17 9.83964 0.0000 0.0000 0.0000D+00 0.0000D+00 0.0000D+00 0.0000D+00 0.3805D-03 0.0000D+00 0.0000D+00 0.0000D+00 0.0000D+00 SPECIAL NOTES for normalization of XES data: In some cases it is desirable to use a normalization which makes the XES intensity correspond to the number of p-electrons on that specific atom. This can be done by summing up the squares of the dipole transition moments for an atom and putting up an equation e.g. for carbon A*sum(m)=2 where A is a normalization constant which is specific for each atom, m are the dipole transition moments and 2 comes from that carbon has 2 p-electrons. The constants for some atoms are: C 148.2 N 198.0 O 249.5 The normalization constants were tested for different excited states of the atoms and were constant to within about 5%. A simple procedure is to convolute the (square of the) transition moments and scale with the factor A. This does not include the energy dependence of the XES intensities, which will make a difference of a few percent. (C) Resonant Inelastic X-ray Scattering (RIXS) The keyword line XRAYspectra RIXS [ncore icore(1) icore(2) ... icore(ncore)] allows the calculation of resonant inelastic X-ray scattering (RIXS) corresponding to combinations of core -> unocc. and valence -> core transitions (resulting in a final state with an electron in a formerly unoccupied orbital and a hole in a valence orbital). The calculation is based on Kohn-Sham orbitals determined by the electronic configuration of the StoBe run (ground, transition, or relaxed core hole state). The RIXS analysis can include several initial core hole orbitals where ncore denotes the number of orbitals and icore(1), icore(2), ..., icore(ncore) gives their orbital indices. If no core orbitals are included in the keyword line (or ncore = 0) then only the lowest core orbital with occupation < 1 is used in the calculation. In this case the electronic configuration of the StoBe run must correspond to a transition state or a relaxed core hole state. NOTE that in RIXS practice the choice of ground state orbitals is preferable over that of transition state or core hole state orbitals. For further background on the techniques to compute X-ray spectra consult references [TPA98], [TLP99], [FBH00], [PAK00], KPP01]. NOTE that available orbital indices icore(i) are given in the energy level listing of the print output, column "(pos.)". Spin beta orbitals have to be incremented by an offset NLA which gives the number of occupied spin alpha orbitals. As an example, in H2O (5 occupied spin alpha orbitals) the O1s orbital has icore = 1 for spin alpha and = 6 for spin beta. The option requires additional keyword input with keyword lines SHIFtenergy eshft defining an energy eshft (in eV, default = 0.0) which is used to shift ionization potentials and all transition energies by the same amount. This rigid shift can be used to correct for the difference between orbital energy and real ionization potential. This is done in most cases in the subsequent graphics program. END terminating the RIXS input. This line must be present even if default values for the parameters fwhm, eshft, emax are chosen and the corresponding lines are omitted. NOTE that the 'END' line must always be present in the input. The RIXS data are, apart from print output, also saved in a spectrum file on unit 11 (fort.11, ASCII format) containing excitation energies, intensities, and transition moments. The format is Line 1 (A5,I5) Type,Ntrans Type = 'RIXA ' characterizes the subsequent data as referring to the first part of the RIXS analysis describing valence->core transitions. NtransA number of valence->core transitions listed below. Line(s) 2 (F22.10,2F22.16,3D20.12) (Eex,Osctot,OscR2,Dx,Dy,Dz, Qxx,Qxy,Qxz,Qyy,Qyz,Qzz), I=1,NtransA describing NtransA valence->core transitions with Eex transition energy in Hartree units. The energy may be shifted wrt the level energy difference by eshft, see above. Osctot total dipole emission intensity of the transition in a.u. given by Osctot = 1/3 * (Etr-eshft)**3 * ( <final|x|core>**2 + <final|y|core>**2 + <final|z|core>**2 ) OscR2 transition moment of the hypothetical R**2 operator, given by OscR2 = 1/3 * (Etr-eshft)**3 * <final|x**2+y**2+z**2|core>**2 This quantity can be helpful in placing the s-components while it does not correspond to a physical operator. Dx, Dy, Dz cartesian components of the dipole transition moment, given by Dx = <final|x|core> , Dy = <final|y|core>, Dz = <final|z|core> These components allow the evaluation of total and angular resolved spectra using e. g. utility xrayspec, see Sec. 2.7.6. Qxx, Qxy, Qxz, Qyy, Qyz, Qzz cartesian components of the quadrupole transition moment tensor, given by Qxx = <final|xx|core> , Qxy = <final|xy|core>, Qxz = <final|xz|core> , Qyy = <final|yy|core>, Qyz = <final|yz|core> , Qzz = <final|zz|core>, These components allow the evaluation of total and angular resolved spectra using e. g. utility xrayspec, see Sec. 2.7.6. Line 3 (A5,I5) Type,Ntrans Type = 'RIXB ' characterizes the subsequent data as referring to the second part of the RIXS analysis describing core->unocc. transitions. NtransB number of core->unocc. transitions listed below. Line(s) 2 (F22.10,2F22.16,9D20.12) (Eex,Osctot,OscR2,Dx,Dy,Dz, Qxx,Qxy,Qxz,Qyy,Qyz,Qzz), I=1,NtransB describing NtransB core->unocc. transitions with Eex excitation energy in Hartree units. The energy may be shifted wrt the level energy difference by eshft, see above. Osctot total oscillator strength of the excitation in a.u. given by Osctot = 2/3 * (Etr-eshft)**3 * ( <final|x|core>**2 + <final|y|core>**2 + <final|z|core>**2 ) OscR2 transition moment of the hypothetical R**2 operator, given by OscR2 = 2/3 * (Etr-eshft)**3 * <final|x**2+y**2+z**2|core>**2 This quantity can be helpful in placing the s-components while it does not correspond to a physical operator. Dx, Dy, Dz cartesian components of the dipole transition moment, given by Dx = <final|x|core> , Dy = <final|y|core>, Dz = <final|z|core> These components allow the evaluation of total and angular resolved spectra using e. g. utility xrayspec, see Sec. 2.7.6. Qxx, Qxy, Qxz, Qyy, Qyz, Qzz cartesian components of the quadrupole transition moment tensor, given by Qxx = <final|xx|core> , Qxy = <final|xy|core>, Qxz = <final|xz|core> , Qyy = <final|yy|core>, Qyz = <final|yz|core> , Qzz = <final|zz|core>, These components allow the evaluation of total and angular resolved spectra using e. g. utility xrayspec, see Sec. 2.7.6. If more than one core level is included in the XES calculation the above lines 1, 2 are saved on unit 11 for each core level sequentially. Example: >TITLE >NiCO Frozen RIXS calculation for carbon and oxygen 1s >SYMM C4V >CARTESIAN BOHR >Ni 0.00000 0.00000 0.00000 28.0 64 >C 0.00000 0.00000 3.17 6.0 64 >O 0.00000 0.00000 5.37 8.0 64 >END >RUNTYPE START NOOPTIMIZATION >SCFTYPE CONVENTIONAL >POTENTIAL NONLOCAL PD86 PD86 >GRID FINE RANDOM >MULTIPLICITY 1 >CHARGE 0 >MAXCYCLES 70 >ECONVERGENCE 0.000001 >DCONVERGENCE 0.00001 >DMIXING MDENS 0.10 >MULLIKEN ON FULL >DIIS ON >FSYM SCFOCC >ALFA 11 0 1 1 8 >BETA 11 0 1 1 8 >XRAY RIXS 2 4 5 >END >END >A-NICKEL (5,8;5,8) >A-CARBON (5,2;5,2) >A-OXYGEN (5,2;5,2) >O-NICKEL (63321/5211*/311+) >O-CARBON iii_iglo >O-OXYGEN iii_iglo >END For this example the unit 11 output file contains a first line 'RIXA 15' describing part A (valence->core transitions) of the first set of transitions (O1s) followed by 15 lines listing quantities Etr(H), Osctot, OscR2, Dx, Dy, Dz, Qxx, Qxy, Qxz, Qyy, Qyz, Qzz for each transition. The numerical output is identical to that of the XES example in Sec. (B) above. Then the line 'RIXB 81' describes part B (core->unocc. transitions) of the first set of transitions (O1s) followed by 81 lines listing quantities Eex(H), Osctot, OscR2, Oscx, Oscy, Oscz for each transition (all numbers of the first 6 and last 3 lines are cut off after 4 digits, quadrupole moments are listed on a separate line): 18.72435 0.0004 0.0000 0.0000D+00 0.0000D+00 0.8122D-02 -0.8592D-03 0.0000D+00 0.0000D+00 -0.8592D-03 0.0000D+00 -0.7161D-03 18.76761 0.0055 0.0000 -0.2968D-01 0.1416D-15 0.0000D+00 0.0000D+00 0.0000D+00 -0.3365D-03 0.0000D+00 0.2276D-17 0.0000D+00 18.76761 0.0055 0.0000 -0.9443D-16 -0.2968D-01 0.0000D+00 0.0000D+00 0.0000D+00 -0.9960D-18 0.0000D+00 -0.3365D-03 0.0000D+00 18.90101 0.0015 0.0000 0.1555D-01 -0.1022D-16 0.0000D+00 0.0000D+00 0.0000D+00 0.4172D-03 0.0000D+00 -0.5461D-18 0.0000D+00 18.90101 0.0015 0.0000 -0.3656D-17 0.1555D-01 0.0000D+00 0.0000D+00 0.0000D+00 0.1945D-17 0.0000D+00 0.4172D-03 0.0000D+00 18.90602 0.0000 0.0002 0.0000D+00 0.0000D+00 -0.1885D-02 -0.3042D-02 0.0000D+00 0.0000D+00 -0.3042D-02 0.0000D+00 -0.2674D-02 .......................................................................... 74.07447 0.0000 0.0018 0.0000D+00 0.0000D+00 -0.3368D-03 -0.2851D-02 0.0000D+00 0.0000D+00 -0.2851D-02 0.0000D+00 -0.2909D-02 151.43965 0.0000 0.0000 0.0000D+00 0.0000D+00 -0.3646D-04 -0.9415D-05 0.0000D+00 0.0000D+00 -0.9415D-05 0.0000D+00 0.1656D-03 273.00328 0.0000 0.0000 0.0000D+00 0.0000D+00 0.1377D-04 0.3482D-03 0.0000D+00 0.0000D+00 0.3482D-03 0.0000D+00 0.3323D-03 after which the line 'RIXA 14' describes part A (valence->core transitions) of the second set of transitions (C1s) followed by 14 lines listing quantities Etr(H), Osctot, OscR2, Dx, Dy, Dz, Qxx, Qxy, Qxz, Qyy, Qyz, Qzz for each transition. The numerical output is identical to that of the XES example in Sec. (B) above. Then the line 'RIXB 81' describes part B (core->unocc. transitions) of the second set of transitions (C1s) followed by 81 lines listing quantities Eex(H), Osctot, OscR2, Oscx, Oscy, Oscz for each transition (all numbers of the first 6 and last 3 lines are cut off after 4 digits, quadrupole moments are listed on a separate line): 9.88432 0.0008 0.0011 0.0000D+00 0.0000D+00 -0.1595D-01 0.7321D-02 0.0000D+00 0.0000D+00 0.7321D-02 0.0000D+00 0.3828D-02 9.92757 0.0075 0.0000 0.4782D-01 -0.3529D-15 0.0000D+00 0.0000D+00 0.0000D+00 -0.1676D-02 0.0000D+00 0.1287D-16 0.0000D+00 9.92757 0.0075 0.0000 -0.4396D-17 0.4782D-01 0.0000D+00 0.0000D+00 0.0000D+00 -0.3201D-17 0.0000D+00 -0.1676D-02 0.0000D+00 10.06097 0.0093 0.0000 -0.5278D-01 -0.4493D-16 0.0000D+00 0.0000D+00 0.0000D+00 0.3556D-03 0.0000D+00 0.3323D-18 0.0000D+00 10.06097 0.0093 0.0000 -0.1071D-15 -0.5278D-01 0.0000D+00 0.0000D+00 0.0000D+00 0.1382D-17 0.0000D+00 0.3556D-03 0.0000D+00 10.06599 0.0002 0.0000 0.0000D+00 0.0000D+00 0.8920D-02 -0.6149D-03 0.0000D+00 0.0000D+00 -0.6149D-03 0.0000D+00 -0.3945D-02 .......................................................................... 65.23444 0.0000 0.0000 0.0000D+00 0.0000D+00 0.3562D-03 -0.2217D-03 0.0000D+00 0.0000D+00 -0.2217D-03 0.0000D+00 0.5204D-03 142.59961 0.0000 0.0001 0.0000D+00 0.0000D+00 -0.5483D-05 0.8882D-03 0.0000D+00 0.0000D+00 0.8882D-03 0.0000D+00 0.2517D-03 264.16325 0.0000 0.0000 0.0000D+00 0.0000D+00 -0.4110D-05 -0.9673D-04 0.0000D+00 0.0000D+00 -0.9673D-04 0.0000D+00 0.2216D-03 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.8) TOPOlogical [ORBital] [BALsac] [qcore rhomin qstep] ----------------------------------------------------------------------------- This keyword starts a topological charge (and orbital) analysis and prints atom charges according to Becke (Becke atom integrals), Voronoi (charges inside Voronoi cells of the cluster), and Bader (charges inside Bader atom cells [B95]). ORBital (optional) includes an orbital projection of atom charges (all occupied orbitals only). Note that for systems with symmetry and representations of dimension > 1 (e. g. C4v symmetry and 2-dim. E representation) corresponding symmetry orbitals may give erroneous atom charges. Orbital projections are not included by default. BALsac (optional) initiates BALSAC [H00] format file output to visualize and analyze total (i.e. not orbital-specific) atom charges derived from Becke, Voronoi, and Bader analyses. The different charge plots are saved in files - BeckeCh.plt for total Becke atom charges - VoronCh.plt for total Voronoi atom charges - BaderCh.plt for total Bader atom charges The output files can be used as input to BALSAC where the atoms of a cluster/molecule are displayed as shaded balls. The ball radii are defined by the actual atom charges and positive/negative charges are shown by different colors (green/blue by default, corresponding to ball "charges" Z = 1 / = 2) in the plot. qcore (optional, default = 0.8) determines the radius of a core sphere about each atom inside which charge is uniquely assigned to the atom. The definition of the core radius, as used in the Bader analysis, is given by Rcore = qcore * Rslater where Rslater is the Slater radius of the corresponding atom. Slater radii are defaulted internally but may be redefined with option SLATer_radii, see c.13). rhomin (optional, default = 1.D-6 e/Bohr^3) defines the smallest value of the (weighted) charge density of an integration point to be included in the charge evaluation. secmin (optional, default = 0.1 Bohr) defines the smallest section length of all polygons used to approximate Bader attractor paths. For a given point in space the length of a local polygon section depends non-linearly on the local charge gradient whith the section length ranging inside [secmin, 6*secmin]. NOTE that, depending on the system and parameter choices, this analysis can be very time consuming. Examples: >TOPO >TOPO 0.9 >TOPO 0.8 1.D-6 0.05 >TOPT ORB 2 7 8 >TOPO BAL 0.7 >TOPO BAL ORB .7 1.D-4 .2 equivalent with >TOPO ORB BAL .7 1.D-4 .2 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.9) NMROut ----------------------------------------------------------------------------- This keyword prepares output for subsequent NMR spectral data calculations on units 70, 71, 72. Requires additional program (V. Malkin and O. Malkina). NOTE that this option forces a diis level shift = 0 and uses the largest possible angular integration grid. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.10) EPROut ----------------------------------------------------------------------------- This keyword prepares output for subsequent EPR spectral data calculations on units 73, 74, 75. Requires additional program (V. Malkin and O. Malkina). NOTE that this option forces a diis level shift = 0 and uses the largest possible angular integration grid. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.11) VIBRations [modes] [basis] [maxd] ----------------------------------------------------------------------------- This keyword defines details of the vibrational analysis, see also keyword RUNType, option c.1). NOTE that in the input line any of the three options may be left out but the option sequence, [modes] before [basis] before [maxd], must be kept. Options modes = ALL evaluates vibrational modes of all symmetry representations. This is the default if no modes keyword is given. NOTE that evaluating all vibrational modes may be rather time consuming!) = SELected evaluates vibrational modes of specific representations of the point symmetry group of the cluster (the group is defined with the cluster coordinates, see a.4)). The representations have to be provided by corresponding symmetry labels of the point symmetry group on the subsequent input line (format is 12A, see examples). NOTE that including the highest symmetry representation (labels A, A1, or A1g) in the selectron may lead to improved anharmonicity treatment in the analysis. = BASis evaluates vibrational displacement vectors of all representations of the point symmetry group of the cluster but does not calculate vibrational excitations. The displacement vectors are saved on unit 60 and can be used as input in a subsequent calculation of vibrational excitations, see option basis = INPut below. NOTE that this option, if combined with the restart option (RUNType RESTart VIBRations, see c.1), does not start an SCF run but takes geometry, symmetry, and electronic parameter information from the restart file. basis = GUEss constructs a basis of the (symmetry adapted) displacement vectors using an intelligent guess based on atom charges and bond orders. This is the default if no basis keyword is given. = INPut reads a basis of the (symmetry adapted) displacement vectors from file (unit 60). maxd maximum atom displacement used to evaluate force derivatives by finite differences (maxd in Bohr or Angstrom depending on the initial coordinate input, see a.1). The default value, if no value given, is 0.01 Bohr). The vibrational analysis creates, in addition to list output of vibrational frequencies (in harmonic approximation), dynamical dipoles, and displacement vectors, corresponding file output - in a generic StoBe format on file unit 65 which collects frequencies, dynamical dipoles, and symmetry labels of all vibrational excitation modes. These files can be used as input to utility irspec, see 2.7.7, to evaluate total and angle-resolved vibrational (infrared, IR) spectra for graphical output or for further processing. The file format is Line 1 (I5) NLEVL NLEVL number of vibrational levels Line(s) 2 (F10.3,4D15.6,A8) EV,D2,Dx,Dy,Dz,SYLBL , i= 1, ... NLEVL EV energy (in cm-1) of vibrational level i D2 = Dx**2 + Dy**2 + Dz**2 Dx, Dy, Dz Cartesian components of the dynamical dipole moment (in a.u.) of vibrational level i SYLBL symmetry label of vibrational level i - in Molekel format on file unit 59 which can be used to visualize vibrational modes with the interactive graphics package Molekel [MOL03]. Note that StoBe creates animation files for vibrational properties only in Molekel format Examples: >VIBRATE BASIS (determine distorsion vector basis only, no mode calculation) >VIBRATE ALL GUESS 0.01 (default setting) >VIBRATE SEL 0.005 (selected modes of representations > A2 B2 A2, B2 only (e.g. in a cluster of C4V symmetry; max. atom displacement = 0.005 Angstrom) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.12) NEBparameters ----------------------------------------------------------------------------- This keyword defines control parameters of a reaction path optimization using the nudged-elastic-band (NEB) method, see c.1), where both the standard NEB [MJ94, MJG95, JMJ98] and the climbing-image-NEB (CI-NEB) approach [HUJ00] are implemented. Additional keyword lines are NIMAge nneb defining the number of NEB images (integer nneb >= nimg) of the NEB path evaluated in each optimization step. The default value of nneb (used if this line is omitted) is nneb = nimg where nimg >= 2 is obtained with the cartesian coordinate input, see a.1). For NEB runs from scratch, nneb images are calculated by linear coordinate interpolation between the nimg initial images given in the input file, see a.1). For NEB restart runs nneb = nimg is assumed and the nneb images are taken from the restart input files (ignoring the explicit coordinates of the input file). MAXPath maxpath defining the (integer valued) maximum number of path iterations of a NEB optimization to obtain a converged path. The actual number of iterations may be smaller than maxpath if convergence is reached earlier, see below. The default value is maxpath = 1. NEBOpt [number [scale]] defining the local optimization and path extrapolation method used in the NEB path optimization. The following methods are available: number = 0 Steepest descent (extrapolation mixing alpha= 1.0) 1 Steepest descent (extrapolation mixing alpha= 0.5) 2 Conjugate gradient (extrapolation mixing alpha= 1.0) 3 Conjugate gradient (extrapolation mixing alpha= 0.5) 4 Conjugate gradient, Polak-Ribiere variant 5 Quasi-Newton, Davidson-Fletcher-Powell variant 6 Quasi-Newton, Murtaugh - Sargent variant 7 Quasi-Newton, Broyden-Fletcher-Goldfarb-Shanno variant (default), no line search available so far With extrapolation methods number 0, 1, 2, 3 the extrapolation mixing factor alpha may be set explicitly using the (optional) parameter scale. For values scale < 0.5 this may result in an increased number of convergence steps but smoother convergence. SPRIng sprng sprngv defining the spring constants (R*8 values) used in the elastic coupling between adjacent images along the NEB reaction path. Here constant sprng is the basic spring constant while sprngv gives the variation of the spring constant for the lowest images (i.e. value sprng is replaced by sprng - sprngv). For sprngv = 0.0 a fixed spring constant will be applied. The default values of sprng, sprngv (used if this line is omitted) are sprng = 0.1, sprngv = 0.0. NEBConvergence cvgneb cvgnebci defining the convergence thresholds (R*8 values) used in the comparison of paths between subsequent path iterations of a NEB optimization. Here cvgneb is the NEB gradient convergence threshold which refers to Cartesian path gradients of dimension ntot*3 (x, y, z components of atoms 1 ... ntot in a.u.). Parameter cvgnebci denotes the climbing image convergence threshold (ntot*3 dimesional gradient in a.u.) which becomes active after NEB gradient convergence is reached. For cvgnebci = 0.0 no climing image calculation will be performed. The default values of cvgneb, cvgnebci (used if this line is omitted) are cvgneb = 0.001, cvgnebci = 0.0. MAXStep [stepmax] [fscale] defining an upper bound to the allowed maximum displacement of atoms between successive steps in a NEB path optimization (stepmax in Bohr or Angstrom depending on the initial coordinate input, see a.1). Larger displacements will be reduced to the maximum while their direction remains unchanged. A reasonable value is 0.25 Bohr with the default = 0.5 Bohr. The second (optional) value allows to scale all displacements (evaluated initially according to forces) by a constant factor fscale (with default = 1.0). For values fscale < 1.0 this may result in an increased number of convergence steps but smoother convergence. NEBInput [FILE, READ] defining geometric (and electronic) structure of all images used to determine the initial path. By default, nimg different initial images, provided in the coordinate section of the input file (see a.1)), are used to calculate a first path of nneb > nimg images starting an NEB run from scratch. These initial images can also be provided by separate restart files from earlier StoBe runs, such as geometry optimizations. The keyword line NEBInput FILE assumes all nimg initial images to be defined by StoBe restart files available in file units nrbaseinp to nrbaseinp+nimg-1 (nrbasinp = 100 by default), see e.13). However, (dummy) coordinates of all images have to be present in the input file and will be overwritten by those of the restart files. Further, the electroinc parameters of the restart files will be used as inital input of the corresponding initial images (restart case). NEBInput READ assumes the atom coordinates of all nimg initial images to be defined in the input file. The electronic structure of the initial images will be calculated from scratch. This is the default if keyword NEBInput is not given in the input. NEBPrint [COORdinates, FORCes, ALL] defining print output during the NEB path calculation. The keyword line NEBPrint COORdinates prints atom coordinates of all images after completion of a path step of nneb images. NEBPrint COORdinates prints atom forces (both initial Pulay and NEB corrected forces) of all images after completion of a path step of nneb images. NEBPrint ALL prints both atom coordinates and atom forces of all images after completion of a path step of nneb images. END terminating the NEB parameter input. This line must be present even if default values for all NEB control parameters are chosen and the corresponding lines are omitted. NOTE that the 'END' line must always be present in the input. Example: >NEBP >NIMAGE 9 >MAXPATH 50 >NEBOPT 2 1. >SPRINGS .5 .1 >NEBCONVERGENCE 1.D-3 1.D-5 >MAXSTEPS .1 >NEBINPUT READ >END ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.13) RESInput nrbasinp ----------------------------------------------------------------------------- This allows to redirect the fortran unit number(s) of one or more restart files. For single image runs (nimg = 1, see a.1)), nrbasinp (defaulted to nrbasinp = 1) defines the (redirected) unit number nrbasinp assuming the restart file to be linked to fort.{nrbasinp}. For multiple image runs, such as NEB path runs (nimg > 1, see e.12)), nrbasinp (defaulted to nrbasinp = 100) defines the offset of subsequent unit numbers, i.e. units {nrbasinp}, {nrbasinp+1}, ..., {nrbasinp+nimg-1} will be used. This assumes the restart files to be linked to fort.{nrbasinp}, fort.{nrbasinp+1}, ..., fort.{nrbasinp+nimg-1}. NOTE that redirecting unit numbers may conflict with existing units used for other parameter files, see Sec. 2.6. Example: >RESINPUT 100 ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) e.14) RESOutput nrbasout ----------------------------------------------------------------------------- This allows to redirect the fortran unit number(s) of one or more restart output files used for subsequent StoBe runs. For single image runs (nimg = 1, see a.1)), nrbasout (defaulted to nrbasout = 2) defines the (redirected) unit number nrbasout for restart file output assuming fort.{nrbasout} as output file. For multiple image runs, such as NEB path runs (nimg > 1, see e.12)), nrbasout (defaulted to nrbasout = 200) defines the offset of subsequent unit numbers, i.e. units {nrbasout}, {nrbasout+1}, ..., {nrbasout+nn-1} will be used for output where nn = nimg applies to regular multiple image runs and nn = nneb to NEB runs, see e.12). NOTE that redirecting unit numbers may conflict with existing units used for other parameter files, see Sec. 2.6. Example: >RESOUTPUT 200 2.2.6. MISCELLANEOUS KEYWORDS (Goto TOC, KEYW, KEYA) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.1) TITLe ----------------------------------------------------------------------------- Brief descriptive title of the present StoBe run on the lines following the keyword. This title appears in the output file(s). Example: >TITLE >h2o local opt >test run Note that, while the title input in the input file allows any number of lines, restart files save only the first line of the title. In the above example only the title line 'h2o local opt' would be saved. Continued title lines must not start with a word identical to a keyword of StoBe input. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.2) PRINtout [DISTances, BASIs, ORBItals, FORBitals, IORBitals, SORBitals, DIPOles, QUADrupoles, MULTipoles, POINtcharges, HESSian, ZMATrix, TIMIng, SPACe, DEFAult, SHORt, FULL] ----------------------------------------------------------------------------- This keyword allows more detailed print output for SCF runs and geometry optimizations where any combination of option keywords from the above list is allowed. Option keyword DISTances includes a listing of all interatomic distances in the output. BASIs includes a listing of all orbital, auxiliary, model potential, and augmented basis sets in the output. ORBItals includes a listing of all orbitals at the end of SCF calculations or geometry optimizations. If an orbital component is < .000001 in absolute value for all orbitals within a symmetry respresentation it will not be listed. FORBitals same as option ORBItals but all orbital components are listed. IORBitals includes a listing of all initial input orbitals at SCF startup (restart or start from scratch). In geometry optimizations this listing is given for each extrapolation geometry. If both keywords IORBitals and SORBitals appear in the input only keyword SORBitals will be used since the orbital listing of the first iteration step gives the initial input orbitals. SORBitals includes a listing of all input orbitals at each iteration step of SCF calculations or geometry optimizations. In particular, the listing at the first iteration step refers to the input orbitals at startup (restart or start from scratch). DIPOles includes a listing of all dipole transition moments (cartesian components, no quadrupole components) at the end of SCF calculations where XAS or XES transitions are evaluated. QUADrupoles includes a listing of all quadrupole transition moments (cartesian components, no dipole components) at the end of SCF calculations where XAS or XES transitions are evaluated. MULTipoles includes a listing of all dipole and quadrupole transition moments (cartesian components) at the end of SCF calculations where XAS or XES transitions are evaluated. POINtcharges includes a listing of all external point charges in the input of SCF calculations or geometry optimizations (place keyword BEFORE point charge listing, see keyword PTCH). HESSian includes a listing of eigenvectors of the Hessian matrix at each iteration step of geometry optimizations. ZMATrix includes a listing of all atom centers in Z-matrix notation (internal coordinates) irrespective of the input format, see a.1), a.2). For internal coordinate input this listing is always included but does not appear for cartesian coordinate input. Note that for cartesian coordinate input dihedral angles may not be meaningful for all atoms depending on the atom sequence (e. g. if three subsequent atoms are collinear the fourth atom canot be assigned a dihedral angle). In this case the dihedral angle is given a value = 0.0 deg. TIMIng includes additional timing information which may be useful to optimize cpu usage. The timing printout mixes with the actual StoBe results and is useful only for testing. SPACe includes information about the space the restart output file (in bytes, unit 2) which may be useful to organize file usage. DEFAult combines option keywords ORBItals, MULTipoles, POINtcharges and excludes HESSian, SORBitals, TIMIng. This is the default if no PRINtout keyword line appears in the input. The keyword DEFAult overrides all previous settings. Thus, the following keyword lines are equivalent: >PRINTOUT TIMING SORBITALS DEFAULT HESSIAN >PRINTOUT DEFAULT HESSIAN >PRINTOUT ORBITALS MULTIP HESSIAN SHORt includes none of the above listing options. The keyword SHORt overrides all previous settings. Thus, the following keyword lines are equivalent: >PRINTOUT TIMING SORBITALS SHORT HESSIAN >PRINTOUT SHORT HESSIAN FULL includes all of the above listing options. This option generates an extremely large output listing file and is meaningful only for testing purposes. The keyword FULL overrides all previous settings. Thus, the following keyword lines are equivalent: >PRINTOUT TIMING SORBITALS FULL HESSIAN >PRINTOUT FULL Example: >PRINTOUT SORBITALS TIMING SPACE ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.3) SAVE ns ----------------------------------------------------------------------------- This option allows periodic updates of the restart file (unit 2) in addition to the restart output at the end of an SCF calculation or geometry optimization. With this keyword line the restart file is updated (overwritten by the most recent result) every ns iteration steps in the SCF cycle and for geometry optimizations a restart file update is performed at the end of each geometry step. By default, restart file output happens only at the end of a calculation. Intermediate restart file update is strongly suggested for large systems and can save time in case of system crashes. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.4) ORBItalchoice [5D, 6D, 5D1S] ----------------------------------------------------------------------------- This keyword determines the internal representation of d basis functions. Option keyword 6D uses the 6 cartesian gaussian d functions (x2, y2, z2, xy, xz, yz) to represent d orbital basis functions. 5D1S uses the 5 atomic gaussian d functions (dz2,dx2-y2, dxy, dxz, dyz) to represent d orbital basis functions and adds the 6th function (x2+y2+z2) to the s orbital basis function set. This representation is equivalent to option 6D but the distinction between atomic d and s function is more explicit. 5D (default) uses the 5 atomic gaussian d functions (dz2,dx2-y2, dxy, dxz, dyz) to represent d orbital basis functions. This eliminates linear dependencies associated with the s-component of the d-functions, facilitates analysis of the wave function and is strongly recommended! Total energies from calculations using options 6D and 5D1S yield the same results which are somewhat different from those using option 5D since the latter uses in effect a smaller, but much safer, basis set. For atoms, the choice of keywords 'FSYM' and option 6D switches to option 5D1S due to rotational symmetry. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.5) ECPRead ----------------------------------------------------------------------------- This keyword allows the use of an alternative set of one-electron integrals (and nuclear repulsion contributions) in a StoBe calculation. The data are read in from unit 52. Examples are calculations where the core potential is modified from a previous calculation. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.6) PNTGenerator ng ----------------------------------------------------------------------------- This keyword allows the selection of different gridpoint generators in the angular integration. For ng = 0 (default) a Gauss-Legendre grid for theta and an equidistant grid for phi is used. = 1 a Gauss-Legendre grid for both theta and phi is used. = 2 a Gauss-Legendre grid for theta and an equidistant grid for phi with twice the number of grid points (compared with ng=0) is used ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.7) ALCHem (NOT IMPLEMENTED) ----------------------------------------------------------------------------- This keyword initiates output of geometry, basis, and orbital coefficients on a separate file, named ALCHEM.inp. The data is stored in ALCHEM format for further processing by the visualization and analysis software ALCHEM. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.8) FILEoutput [MOLDen] [GMOLDen] [MOLEkel] ----------------------------------------------------------------------------- This keyword generates analysis output files to be used as input for the interactive visualization packages Molden [MOD03] and Molekel [MOL03]. Keyword line FILEoutput MOLDen generates file Molden.molf of generic Molden format which can be used as input for the visualization software Molden. FILEoutput GMOLDen generates file GMolden.mdf of Gaussian-94 format which can be used as input for the visualization software Molden. This option was used as default in previous StoBe versions and had to be disabled by keyword NOMOlden (no Molden file output). The obsolete keyword NOMOlden leads to a warning. FILEoutput MOLEkel generates file Molekel.mkl of generic Molekel format which can be used as input for the visualization software Molekel. The keyword line allows also multi-file output of any combination of the three format types. For example, keyword line > FILE MOLDEN MOLEKEL produces files Molden.molf and Molekel.mkl. By default all analysis file output is disabled. The interactive graphics package Molden [MOD03] visualizes geometric and electronic properties as well as convergence parameters. The software has been developed by Gijs Schaftenaar, CAOS/CAMM Center. The interactive graphics package Molekel [MOL03] visualizes geometric, electronic and vibrational properties. The software has been developed by Stefan Portmann and Peter F. Fluekiger at CSCS/ETHZ/UNIGeneva. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.9) VIRTuals [ALL, NONE, nvirt] ----------------------------------------------------------------------------- This keyword allows to define the number of virtual (unoccupied) orbitals to be included in the output listing of orbital energies, orbital coefficients, and in a full Mulliken population analysis. By default, output listings contain all occupied and empty orbitals with KS energies below the HOMO energy. Setting VIRTuals ALL includes all virtual orbitals of all symmetry representations in the output listing. VIRTuals NONE includes no virtual orbitals with KS energies above the HOMO energy in the output listing. Only occupied orbitals (and empty orbitals below the HOMO) are listed. VIRTuals nvirt (where nvirt > 0 is an integer) includes, for each symmetry representation, up to nvirt orbitals in the output listing. If in a given representation there are nempt > 0 empty orbitals below the HOMO these orbitals will be listed and included in the virtual orbital count such that max(nvirt,nempt) virtual orbitals appear in the listing. If the total number of orbitals in a given symmetry representation is too small the actual number of included virtuals may be smaller than nvirt. In spin polarized cases, orbitals which are occupied in only one spin subspace (alpha or beta) are included also with their (unoccupied) spin complement in the listing of orbital energies and Mulliken populations, but not in the listing of orbital coefficients. By default up to 5 virtual orbitals are included in the output listing (corresponding to 'VIRTuals 5'). ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.10) CTRLoption [ctrlfn] ----------------------------------------------------------------------------- With this keyword, StoBe saves a collection of important control parameters from the input in a separate ASCII file (fortran unit 7). The (optional) parameter ctrlfn denotes a file name (default is 'fort.7') to be used for subsequent processing after the StoBe run. During the StoBe run these control parameters are read in from the file before each SCF iteration step (NOTE: lines 6, 7 are not yet included). Editing the ASCII file allows changes of the control parameters while the job is running. The ASCII file has the following structure Line 1 (A80) RUNTIT Title of the StoBe run, see f.1). Only the first line of the title is saved. Line 2 (I5,2D20.10) NSCF,ECONV,DCONV NSCF = maximum number of SCF iterations, see c.3). ECONV = energy convergence of the SCF procedure, see c.4). DCONV = electron density convergence of the SCF procedure, see c.5). Line 3 (I5,3D20.10) MAXG,GCONV,GSTEP,GSCALE MAXG = maximum number of geometry steps in a geometry optimization procedure, see d.3). Note that in a restart run this number includes previous geometry steps. GCONV = convergence threshold of gradients (in a.u.) in a geometry optimization procedure, see d.1). GSTEP = upper bound to the allowed average displacement (in Bohr) of atoms between successive steps in a geometry optimization , see d.2). GSCALE = scaling factor for atom displacements between successive steps in a geometry optimization , see d.2). Line 4 (6I5) KPRORB,KPRDQP,KPRPCH,KPRHES,KPRSOR,KPRTIM KPRORB = 1 prints all orbitals at the end of a StoBe run, see f.2). = 0 does not print orbitals. KPRDQP = 1 prints all dipole and quadrupole moments (cartesian components) at the end of a StoBe run, see f.2). = 0 does not print moments. KPRPCH = 1 prints all external point charges in the input of a StoBe run, see f.2). = 0 does not print point charges. KPRHES = 1 prints eigenvectors of the Hessian matrix at each iteration step of a geometry optimization, see f.2). = 0 does not print eigenvectors of the Hessian matrix. KPRSOR = 1 prints all orbitals at each iteration step of a StoBe run, see f.2). = 0 does not print orbitals at each step. KPRTIM = 1 prints additional timing information, see f.2). = 0 does not print timing. Line 5 (4I5) NMULL,NLOEW,IVIRT,NCYCLE NMULL = 0 prints neither Mulliken populations nor bond orders. = -1 prints total Mulliken populations and total Mayer bond orders at the end of a StoBe run, see e.1). = 1 prints total and orbital resolved Mulliken populations as well as total Mayer bond orders at the end of a StoBe run. = 2 prints total and orbital resolved Mulliken populations as well as total and orbital resolved Mayer bond orders at the end of a StoBe run. NLOEW = -1, 0, 1, 2 Loewdin population analysis and Mayer bond orders with parameter values analogous to NMULL, see e.2). IVIRT = number of virtual orbitals to be printed in the final output, see e.1). A value IVIRT = -1 includes all computed virtuals, see f.9). NCYCLE = frequence of periodic updates of the restart file (unit 2), see f.3). Line 6 (3I5,2F10.5) MAXPATH,NEBOPT,NEBPRT,CVGNEB,CVGNEBCI (see e.12)) MAXPATH = maximum number of path iterations of a NEB optimization. NEBOPT = type of NEB path extrapolation method where = 0 Steepest descent (extrap. mixing alpha= 1.0), = 1 Steepest descent (extrap. mixing alpha= 0.5), = 2 Conjugate gradient (extrap. mixing alpha= 1.0), = 3 Conjugate gradient (extrap. mixing alpha= 0.5), = 4 Conjugate gradient, Polak-Ribiere variant, = 5 Quasi-Newton, Davidson-Fletcher-Powell variant, = 6 Quasi-Newton, Murtaugh - Sargent variant, = 7 Quasi-Newton, Broyden-Fletcher-Goldfarb-Shanno. NEBPRT = printing during the NEB path calculation where = 0 Neither atom coordinates nor forces printed, = 1 Atom coordinates printed, = 2 Atom forces (Pulay,NEB) printed, = 3 Atom coordinates and atom forces printed. CVGNEB = convergence threshold comparing subsequent NEB path iterations. CVGNEBCI = convergence threshold comparing maximum images of subsequent NEB paths in a climing image NEB calculation. Line 7 (4F10.5) STPNEB,SCALNEB,SPRING,SPRINGV (see e.12)) STPNEB = largest allowed displacement of atoms between successive NEB steps (parameter stepmax in NEB input, see e.12). SCALNEB = additiponal scaling factor applied to all atom displacements according to forces (parameter fscale in NEB input, see e.12). SPRING = spring constant used in the elastic coupling between adjacent NEB images. SPRINGV = variation of the spring constant for the lowest images. Line 8 (4F10.5) DMIX,DMIXC,ESHIFT,EFRAC (not yet implemented!) DMIX = density mixing factor (nd, keyword NORMAL) or density matrix mixing factor (nd, keyword MDEN), see c.6). DMIXC = exchange/correlation potential mixing factor (nx, keyword NORMAL), see c.6). ESHIFT = level shift value (esh in Hartree) separating occupied and empty orbitals, see c.8). EFRAC = smear value (efrac in Hartree) determining energy region of fractional orbital occupation numbers, see b.6). Line 9 (2I5,F10.5) IDIIS,NISTEP,ESHIFD (not yet implemented!) IDIIS = 0 DIIS scheme turned off = 1 DIIS scheme turned on, see c.7). NISTEP = number of previous iteration steps used for the DIIS scheme, see c.7). ESHIFD = level shift value (eshd in Hartree) separating occupied and empty orbitals while the DIIS scheme is applied, see c.7). Example: >CTRL (control file on unit 7) >CTRL Control.txt (control file on unit 7 named 'Control.txt') ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.11) DOSOutput [MULLiken, LOEWdin] ----------------------------------------------------------------------------- With this keyword, StoBe saves Kohn-Sham levels and Mulliken/Loewdin populations on external files (unit 96 for Mulliken populations, keyword MULLiken; unit 97 for Loewdin populations, keyword LOEWdin) to be used for subsequent densities-of-states calculations with utility doscalc, see Sec. 2.7.5. Example: >DOSOUTPUT MULLIKEN (stores Mulliken data on unit 96) >DOSOUTPUT LOEWDIN MULLIKEN (stores Mulliken data on unit 96 and Loewdin data on unit 97) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.12) SHoRTrestartfile ----------------------------------------------------------------------------- With this keyword (spelled as 'SHRT'), StoBe saves restart data on unit 2 in a shortened format (short StoBe file format) where records 8, 9, 19 are replaced by one-value dummies. This can reduce the restart file size considerably. However, using restart files with short file format for input will recalculate the missing data at startup which takes (little) more time compared to using the full StoBe format. Example: >SHRT (uses short StoBe format for restart file output, unit 2) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.13) FORCes ----------------------------------------------------------------------------- With this keyword, StoBe evaluates forces on all atom centers after a single point SCF run. The forces are listed in the output file. This is useful to check whether an SCF run refers to an equilibrium geometry of the cluster/molecule. In geometry optimizations or property runs this keyword is ignored, see c.1). Example: >FORCES (evaluates atom forces after a single point SCF run) ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.14) DYNAmics ----------------------------------------------------------------------------- Short-time dynamics simulation. This is NOT a full molecular dynamics (MD) procedure, i.e. there is no equilibration or thermalization. The procedure only computes forces on the atoms (as in ab initio MD) and then integrate forward in time (as in MD). Initial velocities can be given, typically as a sampling of the zero-point distribution. Applications have so far been to XES on hydrogen-bonded systems. See, e.g., M.P. Ljungberg, A. Nilsson, and L.G.M. Pettersson, Phys. Rev. B 82, 245115 (2010); M.P. Ljungberg, L.G.M. Pettersson, and A. Nilsson, J. Chem. Phys. 134, 044513 (2011)]. Additional keyword lines are STEP deltat defining the time between successive steps of the simulation in femtoseconds. TEND tstop defining the end time after which the simulation is finshied in femtoseconds. VELO ncent defining the number icent of centers of the cluster which are allowed to move during the simulation. This is followed by ncent lines (I,3F) (i, vx,vy,vz), i=1,ncent defining the starting velocities of the selected centers of the where i denotes the center index and (vx,vy,vz) is the velocity vectorcluster in atomic units (a.u.). END terminates the simulation input. This line must always be present in the input. ----------------------------------------------------------------------------- (Goto TOC, KEYW, KEYA) f.15) GEOFormat [BALsac, XYZ, PDB, ALL] ----------------------------------------------------------------------------- Atom coordinates of molecules/clusters evaluated in StoBe runs can be saved in geometry output files where different formats are available and are defined by the parameters of the present option. Parameter BALsac refers to the Balsac/Plot3D [H00] file format to be used for output of the cluster/molecule geometry. Corresponding files will contain the postfix ".plt", e.g. "strucfin.plt". This is the default setting if the option is not included in the input. XYZ refers to the XYZ file format to be used for output of the cluster/molecule geometry. Corresponding files will contain the postfix ".xyz", e.g. "strucfin.xyz". PDB refers to the Protein Data Bank format to be used for output of the cluster/molecule geometry. Corresponding files will contain the postfix ".pdb", e.g. "strucfin.pdb". ALL combines output of with all three formats, i.e. each structure output creates three files, fname.plt, fname.xyz, and fname.pdb, to be used for postprocessing. NONE ignores all geometry file output. The different formats can be used simultaneously to create files of the same structure. Option 'NONE' ignores all output irrespective of other choices in the keyword line. Option 'ALL' (and no 'NONE') saves with all output formats irrespective of other choices in the keyword line. Examples: >GEOFORMAT XYZ only xyz format geometry file output >GEOFORMAT BALSAC PDB both Balsac/Plot3D and pdb format geometry file output >GEOFORMAT ALL Balsac/Plot3D, xyz, and pdb format geometry file output 2.2.7. UNTESTED / OBSOLETE KEYWORDS (Goto TOC, KEYW, KEYA) The following keywords have not been carefully tested or not yet implemented in StoBe : INTEgration - method for angular grid integration LENNard - include Lennard-Jones potentials with external point charges MEPFit - calculate eff. atomic charges from MEPS fitting (not implemented) PNTGenerator - choice of gridpoint generator for angular integration WEDGe - restrict angular grid of Gauss-Legendre integration 2.3. BASIS SET INPUT (Goto TOC, KEYW, KEYA) The list of all StoBe keywords and corresponding numerical parameters (explained in Sec. 2.2.) is terminated by the END line. The remaining part of the input file defines the basis sets (orbital bases, auxiliary bases, model core potential (MCP or ECP) definitions, and augmentation bases for core electron spectroscopy, NEXAFS, if applicable) to be used in the calculations. The general sequence of basis sets is - auxiliary basis sets (A-...) for all (symmetry non-equivalent) atom centers using the order defined by the coordinate list, see a.1) and including all dummy atoms, see a.1), or using a global definition, see below. - orbital basis sets (O-...) for all (symmetry non-equivalent real and dummy) atom centers using the same order as for auxiliary basis sets, or using a global definition, see below. - orbital basis sets combined with (parametrized) Dolg effective core potentials (D-...) for all (symmetry non-equivalent real and dummy) atom centers using the same order as for auxiliary basis sets. No global definition available do far. - (optional) model core potential definitions (P-...) for selected atom centers using the same order as for auxiliary basis sets (omitting centers without model core potentials), see below. A model core potential is assumed for an atom center whose effective nuclear charge ZZ defined in the coordinate list, see a.1), is different from its atomic number and is non-zero. - (required for XAS) augmentation basis sets including all (symmetry non-equivalent) atom centers using the order defined by the coordinate list, see below and keyword XRAY, e.7). The basis set input allows different formats depending on the basis set type as described in the following. (A) Orbital basis sets (A.1) Explicit definition Here the basis set of each atom center is defined by a one-line title which refers to an entry in the basis set file provided with the input (fort.3). Possible titles are of the form [A92,GSA92] >O-ELEMENT (n1s,n2s,...,nKs/n1p,n2p,...,nLp/n1d,n2d,...,nMd) where ELEMENT is a name which helps the user to identify the basis (typically the element name). Further, K denotes the number of s-type Gaussians with n1s,n2s,...,nKs primitives respectively, L contractions of p-type Gaussians with n1p, n2p,...,nLp primitives, and M contractions of d-type Gaussians with n1d,n2d,...nMd primitives respectively. Examples of orbital basis sets (the star at the end denotes that a diffuse function is added) are >O-OXYGEN (621/41/1) >O-HYDROGEN (41/1*) >O-NITROGEN (7111/411/1) (A.2) Global definition Here basis set definitions are given by global keywords which apply to all atom centers and refer to a general basis set quality (exceptions for selected centers can be defined, see below). The definition consists of one keyword line >GLOBal [orbbasglobal auxbasglobal] where orbbasglobal and auxbasglobal define the quality of orbital and auxiliary basis sets with valid names (for corresponding explicit definitions see 2.3.1.) ---------------------------------------------------------------- orbbasglobal quality ---------------------------------------------------------------- DZVP double zeta + valence polarization (default) DEF same as DZVP TZVP triple zeta + valence polarization ---------------------------------------------------------------- and ---------------------------------------------------------------- auxbasglobal quality ---------------------------------------------------------------- A1 A1 type (default) A5 A5 type DEF same as A1 GENAx Ax type, x=2, 3, 4, computed from the corresponding orbital basis set GENAx+ Ax+ type, x=2, 3, 4, computed from the corresponding orbital basis set ---------------------------------------------------------------- NOTE that keywords orbbasglobal and auxbasglobal can only appear as pairs or have to be left out both. These keywords have to be given with upper case characters only. Examples of global definitions are >GLOBAL TZVP A5 >GLOBAL DZVP GENA4 >GLOBAL DZVP GENA2+ >GLOBAL (equivalent with GLOBAL DZVP A1) Selected atom centers (defined by their atom labels) can be exempt from the global basis set definition. Their orbital and auxiliary basis sets have to be defined separately. This requires a keyword line >SUBStitute atomlabel where atomlabel defines the label of the corresponding atom which must appear (with the same notation) in the previous list of coordinates, see keyword a.1) of 2.2.1. This has to be followed by two lines giving the explicit titles of auxiliary and orbital basis sets (the former may also be a computed auxiliary basis set, see B.3). An example is >GLOBAL DZVP A1 >SUBSTITUTE O >A-OXYGEN (4,4;4,4) >O-OXYGEN (721/51/1*) The global basis set input has to be finished by an END line. After this all basis set input of model core potentials and augmentation bases follows, using the explicit basis set definitions, see C, D below. (A.3) Explicit definition of valence basis with Dolg effective core potential There is a full collection of valence basis sets together with (parametrized) relativistic effective core potentials according to Dolg et al. [DD82-95] for all elements from Li (Z = 3) to Fermium (Z = 100) following the parametrization/representation of the Toulouse group (J.-L. Heully [?]). The basis set of each atom center is defined by a one-line title which refers to an entry in the basis set file provided with the input (fort.3). Possible titles are of the form >D-TITLE where TITLE is a name which helps the user to identify the basis. Examples of orbital basis sets are >D-B.ECP.Dolg.4s4p.2s2p.3e-MWB. >D-Fe.ECP.Dolg.8s7p6d1f.6s5p3d1f.16-MDF. >D-Pu.ECP.Dolg.12s11p10d8f.8s7p6d4f.34e-MWB. Note that combined Dolg basis sets and effective core potentials do not require additional definitions of effective core parameters as opposed to valence basis sets combined with model core potential, see (C.1) below. Further, Dolg orbital basis sets do not come with explicit auxiliary basis sets which have to be generated, see (B.3) below. (B) Auxiliary basis sets (B.1) Explicit definition Here the basis set of each (non-equivalent) atom center is defined by a one-line title which refers to an entry in the basis set file provided with the input. Possible titles are of the form [A92,GSA92] >A-ELEMENT (ns(CD),nspd(CD); ns(XC),nspd(XC)) where ELEMENT is a name which helps the user to identify the basis (typically the element's name), ns(CD) is the number of s-type Gaussians used to fit the charge density (CD), nspd(CD) is the number of s-, p- and d- type Gaussians (sharing the same exponent) used to fit the valence charge density; ns(XC) and nspd(XC) have definitions similar to ns(CD) and nspd(CD), while referring to the Gaussians used to fit the exchange and correlation potentials. Examples of auxiliary basis sets are >A-NITROGEN (4,4;4,4) >A-OXYGEN (5,2;5,2) >A-HYDROGEN (5,1;5,1) (B.2) Global definition Here basis set definitions are given by global keywords which apply to all atom centers and refer to a general basis set quality (exceptions for selected centers can be defined, see below). The definition consists of one keyword line >GLOBal [orbbasglobal] [auxbasglobal] at the beginning of the basis set section as described in A.2 above. (B.3) Internal construction from orbital basis sets Auxiliary basis sets can be generated from corresponding orbital basis sets following procedures described in [KH01] and below. The input line for a computed auxiliary basis set reads >GENAxy orbitalbasistitle where orbitalbasistitle is the title of an orbital basis set which refers to an entry in the basis set input file and xy = 2, 3, 4 or 2+, 3+, 4+ denotes the quality of the auxiliary basis set. NOTE that keyword GENAx and the orbital basis title must be separated by at least one blank character. Examples of computed auxiliary basis sets are >GENA2 O-OXYGEN (621/41/1) >GENA4+ O-OXYGEN (721/51/1*) The internal generator of auxiliary basis sets from orbital basis produces even tempered basis sets (linear progression) according to empirical recipes (a) GENAx, x = 2,3,4 This procedure determines the range of all exponents [zmin, zmax] of the orbital basis set irrespective of quantum number. Then the range [2*zmin, 2*zmax] is spanned by an even tempered sequence of N exponents from the top. Then the first N/2 exponents are assigned to an s basis set and the second to a mixed s, p, d basis set describing the density basis. The corresponding exchange/correlation basis is determinated from the density basis by dividing all exponents by 3. Number x = 2,3,4 determines the progression factor in the exponent sequence (alp(i)= alp(i-1)/R, R = 6 - x). (b) GENAx+, x = 2,3,4 This procedure according to [H03] determines in addition to [zmin, zmax], see (a), also the range of all p, d exponents [zmin', zmax'] of the orbital basis set. Then the range [2*zmin', 2*zmax'] is spanned by an even tempered sequence of N exponents from the top. These exponents are assigned to a mixed s, p, d basis set describing the density basis. The sequence is continued at the top and bottom to cover the full range [2*zmin, 2*zmax] where additional exponents, not included in the spd basis are assigned to a separate s basis. The corresponding exchange/correlation basis is determinated as in (a). Number x = 2,3,4 determines the progression factor in the exponent sequence (alp(i)= alp(i-1)/R, R = 6 - x). A line >GENAxy ALL at the beginning of the basis set input section (with explicit orbital, MCP, and/or augmentation basis set definitions following) computes, for all atoms, auxiliary basis sets of Ax or Ax+ quality, x=2,3,4, from corresponding orbital bases. NOTE that this definition does not allow further explicit auxiliary basis set definitions. (C) Model core potential definitions (C.1) Explicit definition The model core potential (MCP) of each (non-equivalent) atom center is defined by a one-line title which refers to an entry in the basis set file provided with the input. The definition of the MCPs follows Huzinaga's procedure as described e. g. in [PWG87]. At present, this is the only way to define MCP's. Possible MCP titles are of the form >P-LABEL(Zeff) (nv(mcp): ns(mcp),np(mcp),nd(mcp)) where nv(mcp) is the number of s-type Gaussians used to fit the potential seen by the valence electrons, ns(mcp),np(mcp) and nd(mcp) are respectively the number of s-, p-, and d-type Gaussians used to describe the s,p and d core orbitals and Zeff is the effective nuclear charge after removal of the core electrons. Examples of MCP basis sets are >P-PALLADIUM(+16) (6:9,6,4) >P-ALUMINUM(+3) (4:6,4) >P-PHOSPHORUS(+5) (4:6,4) >P-GERMANIUM(+14) (3:7,5) >P-SULFUR(+6) (4:6,4) >P-CHLORINE(+7) (4:6,4) >P-NICKEL(+16) (5:7,4) >P-SILICON(+4) (4:6,4) >P-RHODIUM(+15) (6:9,6,4) >P-NIOBIUM(+11) (5:8,6,4) >P-PLATINUM(+16) (7:12,9,7,5) >P-MOLYBDENUM(+14) (5:8,6,4) (D) Augmentation basis sets (D.1) Explicit definition The model augmentation basis set of each (non-equivalent) atom center is defined by a one-line title which refers to an entry in the basis set file provided with the input. At present, this is the only way to define augmentation basis sets. Possible titles are >X-FIRST or >X-SECOND or >X-DUMMY where X-FIRST is sufficient for first-row systems and X-SECOND is somewhat more extended. X-DUMMY is used for all atom centers except those where the core excitation originates from. NOTE: The order of the augmentation basis input must follow that of the geometry input, i.e. as for the auxiliary and orbital bases. After the basis set for the core-excited center has been specified (i.e. X-FIRST or X-SECOND) additional "trailing" X-DUMMY need not be given. 2.3.1. GLOBAL BASIS SET DEFINITIONS (Goto TOC, KEYW, KEYA) The following tables list explicit basis set titles for each element (if available) which are used in the global definitions. (a) Orbital basis sets (DZVP, TZVP) ------------------------------------------------------------------------------ element DZVP TZVP ------------------------------------------------------------------------------ 1 H O-HYDROGEN (41/1*) --- 2 He O-HELIUM (51) --- 3 Li O-LITHIUM (621/1/1) --- 4 Be O-BERYLLIUM (621/1/1) --- 5 B O-BORON (621/41/1) --- 6 C O-CARBON (621/41/1) O-CARBON (7111/411/1) 7 N O-NITROGEN (621/41/1) O-NITROGEN (7111/411/1) 8 O O-OXYGEN (621/41/1) O-OXYGEN (7111/411/1) 9 F O-FLUORINE (621/41/1 O-FLUORINE (7111|1+/411/1*|1+) (TZVP-FIP1) 10 Ne O-NEON (621/41/1) --- 11 Na O-SODIUM (6321/411/1) O-SODIUM (6321/411/1) 12 Mg O-MAGNESIUM (6321/411/1) O-MAGNESIUM (6321/411/1) 13 Al O-ALUMINUM (6321/521/1) O-ALUMINUM (73111/6111/1) 14 Si O-SILICON (6321/521/1) O-SILICON (73111/6111/1) 15 P O-PHOSPHORUS (6321/521/1) O-PHOSPHORUS (73111/6111/1) 16 S O-SULFUR (6321/521/1) O-SULFUR (73111/6111/1) 17 Cl O-CHLORINE (6321/521/1) O-CHLORINE (73111/6111/1) 18 Ar O-ARGON (6321/521/1) O-ARGON (73111/6111/1) 19 K O-POTASSIUM (63321/5211/1) O-POTASSIUM (63321/5211/1) 20 Ca O-CALCIUM (63321/5211/41) O-CALCIUM (63321/5211/41) 21 Sc O-SCANDIUM (63321/531/311) O-SCANDIUM (63321/531/311) 22 Ti O-TITANIUM (63321/531/311) O-TITANIUM (63321/531/311) 23 V O-VANADIUM (63321/531/311) O-VANADIUM (63321/531/311) 24 Cr O-CHROMIUM (63321/531/311) O-CHROMIUM (63321/531/311) 25 Mn O-MANGANESE (63321/531/311) O-MANGANESE (63321/531/311) 26 Fe O-IRON (63321/531/311) O-IRON (63321/531/311) 27 Co O-COBALT (63321/531/311) O-COBALT (63321/531/311) 28 Ni O-NICKEL (63321/531/311) O-NICKEL (63321/531/311) 29 Cu O-COPPER (63321/531/311) O-COPPER (63321/531/311) 30 Zn O-ZINC (63321/531/311) O-ZINC (63321/531/311) 31 Ga O-GALLIUM (63321/5321/41) --- 32 Ge O-GERMANIUM (63321/5321/41) --- 33 As O-ARSENIC (63321/5321/41) --- 34 Se O-SELENIUM (63321/5321/41) --- 35 Br O-BROMINE (63321/5321/41) --- 36 Kr O-KRYPTON (63321/5321/41) --- 37 Rb O-RUBIDIUM (633321/53211/51) --- 38 Sr O-STRONTIUM (633321/53211/531) --- 39 Y O-YTTRIUM (633321/53211/531) --- 40 Zr O-ZIRCONIUM (633321/53211/531) --- 41 Nb O-NIOBIUM (633321/53211/531) --- 42 Mo O-MOLYBDENUM (633321/53211/531) --- 43 Tc O-TECHNETIUM (633321/53211/531) --- 44 Ru O-RUTHENIUM (633321/53211/531) --- 45 Rh O-RHODIUM (633321/53211/531) --- 46 Pd O-PALLADIUM (633321/53211/531) --- 47 Ag O-SILVER (633321/53211/531) --- 48 Cd O-CADMIUM (633321/53211/531) --- 49 In O-INDIUM (633321/53321/531) --- 50 Sn O-TIN (633321/53321/531) --- 51 Sb O-ANTIMONY (633321/53321/531) --- 52 Te O-TELLURIUM (633321/53321/531) --- 53 I O-IODINE (633321/53321/531) --- 54 Xe O-XENON (633321/53321/531) --- ------------------------------------------------------------------------------ (b) Auxiliary basis sets (A1, A5) ------------------------------------------------------------------------------ element A1 A5 ------------------------------------------------------------------------------ 1 H A-HYDROGEN (4,2;4,2) A-HYDROGEN (5,1;5,1) 2 He A-HELIUM (3,1;3,1) A-HELIUM (3,1;3,1) 3 Li A-LITHIUM (4,3;4,3) A-LITHIUM (4,3;4,3) 4 Be A-BERYLLIUM (4,3;4,3) A-BERYLLIUM (4,3;4,3) 5 B A-BORON (4,3;4,3) A-BORON (5,5;5,5) 6 C A-CARBON (4,3;4,3) A-CARBON (5,2;5,2) 7 N A-NITROGEN (4,3;4,3) A-NITROGEN (5,2;5,2) 8 O A-OXYGEN (4,3;4,3) A-OXYGEN (5,2;5,2) 9 F A-FLUORINE (4,3;4,3) A-FLUORINE (5,2;5,2) 10 Ne A-NEON (4,3;4,3) A-NEON (4,4;4,4) 11 Na A-SODIUM (5,4;5,4) A-SODIUM (5,4;5,4) 12 Mg A-MAGNESIUM (5,4;5,4) A-MAGNESIUM (5,4;5,4) 13 Al A-ALUMINUM (5,4;5,4) A-ALUMINUM (5,4;5,4) 14 Si A-SILICON (5,4;5,4) A-SILICON (5,4;5,4) 15 P A-PHOSPHOROUS (5,4;5,4) A-PHOSPHOROUS (5,4;5,4) 16 S A-SULFUR (5,4;5,4) A-SULFUR (5,4;5,4) 17 Cl A-CHLORINE (5,4;5,4) A-CHLORINE (5,4;5,4) 18 Ar A-ARGON (5,4;5,4) A-ARGON (5,4;5,4) 19 K A-POTASSIUM (5,5;5,5) A-POTASSIUM (5,5;5,5) 20 Ca A-CALCIUM (5,5;5,5) A-CALCIUM (5,5;5,5) 21 Sc A-SCANDIUM (5,5;5,5) A-SCANDIUM (5,5;5,5) 22 Ti A-TITANIUM (5,5;5,5) A-TITANIUM (5,5;5,5) 23 V A-VANADIUM (5,5;5,5) A-VANADIUM (5,5;5,5) 24 Cr A-CHROMIUM (5,5;5,5) A-CHROMIUM (5,5;5,5) 25 Mn A-MANGANESE (5,5;5,5) A-MANGANESE (5,5;5,5) 26 Fe A-IRON (5,5;5,5) A-IRON (5,5;5,5) 27 Co A-COBALT (5,5;5,5) A-COBALT (5,5;5,5) 28 Ni A-NICKEL (5,5;5,5) A-NICKEL (5,5;5,5) 29 Cu A-COPPER (5,5;5,5) A-COPPER (5,5;5,5) 30 Zn A-ZINC (5,5;5,5) A-ZINC (5,5;5,5) 31 Ga A-GALLIUM (5,5;5,5) A-GALLIUM (5,5;5,5) 32 Ge A-GERMANIUM (5,5;5,5) A-GERMANIUM (5,5;5,5) 33 As A-ARSENIC (5,5;5,5) A-ARSENIC (5,5;5,5) 34 Se A-SELENIUM (5,5;5,5) A-SELENIUM (5,5;5,5) 35 Br A-BROMIUM (5,5;5,5) A-BROMIUM (5,5;5,5) 36 Kr A-KRYPTON (5,5;5,5) A-KRYPTON (5,5;5,5) 37 Rb A-RUBIDIUM (5,5;5,5) A-RUBIDIUM (5,5;5,5) 38 Sr A-STRONTIUM (5,5;5,5) A-STRONTIUM (5,5;5,5) 39 Y A-YTTRIUM (5,5;5,5) A-YTTRIUM (5,5;5,5) 40 Zr A-ZIRCONIUM (5,5;5,5) A-ZIRCONIUM (5,5;5,5) 41 Nb A-NIOBIUM (5,5;5,5) A-NIOBIUM (5,5;5,5) 42 Mo A-MOLYBDENUM (5,5;5,5) A-MOLYBDENUM (5,5;5,5) 43 Tc A-TECHNETIUM (5,5;5,5) A-TECHNETIUM (5,5;5,5) 44 Ru A-RUTHENIUM (5,5;5,5) A-RUTHENIUM (5,5;5,5) 45 Rh A-RHODIUM (5,5;5,5) A-RHODIUM (5,5;5,5) 46 Pd A-PALLADIUM (5,5;5,5) A-PALLADIUM (5,5;5,5) 47 Ag A-SILVER (5,5;5,5) A-SILVER (5,5;5,5) 48 Cd A-CADMIUM (5,5;5,5) A-CADMIUM (5,5;5,5) 49 In A-INDIUM (5,5;5,5) A-INDIUM (5,5;5,5) 50 Sn A-TIN (5,5;5,5) A-TIN (5,5;5,5) 51 Sb A-ANTIMONY (5,5;5,5) A-ANTIMONY (5,5;5,5) 52 Te A-TELLURIUM (5,5;5,5) A-TELLURIUM (5,5;5,5) 53 I A-IODINE (5,5;5,5) A-IODINE (5,5;5,5) 54 Xe A-XENON (5,5;5,5) A-XENON (5,5;5,5) ------------------------------------------------------------------------------ 2.4. SYMMETRY BASIS SET INPUT (Goto TOC, KEYW, KEYA) Symmetry group file 'symbasis' (associated with fort.4 in the program) contains symmetry transformations and irreducible representation tables of a large number of point symmetry groups. At present, the following 48 point groups are included 1. C1 2. Ci 3. Cs 4. Csxz 5. Csy 6. Csyz 7. Csx 8. Csxy 9. Csz 10. C2 11. C3 (!) 12. C4 (!) 13. C5 (!) 14. C6 (!) 15. S4 (!) 16. S6 (!) 17. D2 18. D3 19. D4 20. D5 21. D6 22. C2v 23. C2vb 24. C3v 25. C4v 26. C5v 27. C6v 28. C2h 29. C3h (!) 30. C4h (!) 31. C5h (!) 32. C6h (!) 33. D2h 34. D3h 35. D4h 36. D5h 37. D6h 38. D2d 39. D3d 40. D4d 41. D5d 42. D6d 43. T (!) 44. Th (!) 45. Td 46. O 47. Oh 48. ATOM Note that file symbasis uses symmetry labels in upper case notation (e.g. C4V) only while in StoBe input these labels can be given by mixed case (e.g. C4v) or only upper case characters (e.g. C4V). Note further that - All point symmetry groups (except C1) refer to a symmetry origin at (0, 0, 0) in absolute cartesian coordinates. - Point groups with labels (!) are not fully reducible for real valued basis functions and are incompletely reduced. - Symmetry label ATOM which denotes the full rotational point group (up to d functions) is restricted to free atoms. - Symmetry label C2vb denotes the C2v symmetry group where the two mirror planes contain the z axis and the xy diagonal axes. - Groups with only mirror symmetry (Cs..) refer to mirror plane normals (0, 1, 0) for Cs, Csxz, Csy (corresponding to the xz plane), (1, 0, 0) for Csyz, Csx (corresponding to the yz plane), (0, 0, 1) for Csxy, Csz (corresponding to the xy plane), in absolute cartesian coordinates. - Groups with one rotational axis only (groups # 10-16, 22-32) assume that this axis points along (0, 0, 1), i. e. the z axis in absolute cartesian coordinates. - Groups with more than one rotational axis (i. e. groups # 17-21, 33-47) assume that (one of) the rotational axis (axes) of highest order points along (0, 0, 1), i. e. the z axis in absolute cartesian coordinates. - For groups with a rotational axis (of highest order, pointing along the z direction (0, 0, 1)) and with mirror planes only parallel or perpenducular to the rotational axis (groups # 22,24-42, for group # 23 see exception above), one of the mirror planes containing the axis is assumed to form the xz plane, i. e. the mirror plane is spanned by unit vectors (1, 0, 0) and (0, 0, 1). - Tetrahedral groups (groups # 43-45) assume 2-fold rotation axes along the cartesian axes (1, 0, 0), (0, 1, 0), (1, 0, 0), and 3-fold rotation axes along diagonals (+/-1, +/-1, +/-1). Possible mirror plane normals point along (1, 0, 0), (0, 1, 0), (1, 0, 0). - Octahedral group # 46 (O) assumes 4-fold rotation axes along the cartesian axes (1, 0, 0), (0, 1, 0), (1, 0, 0), and 3-fold rotation axes along diagonals (+/-1, +/-1, +/-1). - Octahedral group # 47 (Oh) assumes 4-fold rotation axes along the cartesian axes (1, 0, 0), (0, 1, 0), (1, 0, 0), and 3-fold rotation axes along diagonals (+/-1, +/-1, +/-1). Mirror plane normals point along (1, 0, 0), (0, 1, 0), (1, 0, 0) and along +/-(1, 1, 0), +/-(1, 0, 1), +/-(0, 1, 1). Format details of the ASCII type symmetry group file are discussed in Sec. 2.4.1. 2.4.1. FORMAT OF SYMMETRY GROUP FILE (Goto TOC, KEYW, KEYA) The ASCII type symmetry group file 'symbasis', see previous Sec. 2.4., contains symmetry transformations and irreducible representation tables of a large number of point symmetry groups. The file format for a given symmetry group is as follows (notation according to standard Fortran 77) Line 1. (A4) SYMLAB SYMLAB := label of the symmetry group in upper case characters or numbers. At present 48 different symmetry groups are included, see previous Sec. 2.4. Line 2. (2I5) NSYM,INVER NSYM := total number of symmetry operations. INVER := number of symmetry operations which conserve the handedness of the Cartesian coordinate system (identity, proper rotations). The listing of all symmetry operations in the following line(s) 3 is given in the sequence of INVER operations conserving handedness (identity, proper rotations), followed by (NSYM-INVER) operations changing handedness (improper rotations: inversion, mirroring). NSYM lines 3 follow. Line(s) 3. (3F,A) E1,E2,E3,SYMOP Free Fortran input format with fractions allowed as 'K/L' (K, L integer; example '4/3' = 1.333...). E1,E2,E3 := Euler rotation angles of each symmetry operation in multiples of Pi. The three Euler angles A (alpha), B (beta), C (gamma) are given as A = E1 x Pi, B = E2 x Pi, C = E3 x Pi, Pi = 3.1415926535898 Then the general (3x3) transformation matrix T of a symmetry operation transforming Cartesian coordinates is defined by a product of three rotation matrices and a prefactor q, i. e. by ( t11 t12 t13 ) ( cos A -sin A 0 ) T = ( t21 t22 t23 ) = q x ( sin A cos A 0 ) x ( t31 t32 t33 ) ( 0 0 1 ) ( cos B 0 sin B ) ( cos C -sin C 0 ) x ( 0 1 0 ) x ( sin C cos C 0 ) (-sin B 0 cos B ) ( 0 0 1 ) The prefactor q is set to q = 1 for proper rotations, operations I = 1, INVER and to q = -1 for operations I = INVER+1, NSYM (improper rotations: inversion, mirroring). SYMOP := ASCII label of each symmetry operation. The identity is labeled E, inversion is labeled I, rotation labels start with Cn (n denoting the foldedness), mirror operations start with S. Line 4. (I5) NREP NREP := Number of irreducible representations of the symmetry group. NREP lines 5 follow. Line(s) 5. (A4,I3) REPOP,IRPDIM REPOP := ASCII label of irreducible representation. IRPDIM := Dimension of irreducible representation. NREP sequences of (IRPDIM x IRPDIM) lines 6 follow. Each sequence addresses one representation, labeled REPOP, with NSYM representation matrices RMAT of dimension (IRPDIM x IRPDIM). The format is Line(s) 6. (NSYM*A) (((RMAT(I,J,ISY),ISY=1,NSYM), J=1,IRPDIM), I=1, IRPDIM) RMAT := component (I,J) of representation matrix reflecting symmetry operation ISY. Components are given with only 7 digits in the symmetry group file but are corrected to yield machine accuracy (function FRATIO) to reflect the most accurate symmetry description. As examples, fractions, square roots, and some trigonometric function values (e.g. cos(72)) are corrected. Line 7. (A4) 'STOP' This finishes a symmetry group description in the symmetry group file As an example the different lines of symmetry group C4v as given in the file are shown with line comments to the right. C4V line 1 8 4 line 2 0 0 0 E 8 lines 3 0 0 1 C2Z ... 0 0 -1/2 C4Z- ... 0 0 1/2 C4Z+ ... 0 1 0 SIGY ... 0 1 1 SIGX ... 0 1 1/2 SGDA ... 0 1 -1/2 SGDB ... 5 line 4 A1 1 5 lines 5 A2 1 ... B1 1 ... B2 1 ... E 2 ... 1 1 1 1 1 1 1 1 line 6, 1-dim rep. 1: 1,1 1 1 1 1 -1 -1 -1 -1 line 6, 1-dim rep. 2: 1,1 1 1 -1 -1 1 1 -1 -1 line 6, 1-dim rep. 3: 1,1 1 1 -1 -1 -1 -1 1 1 line 6, 1-dim rep. 4: 1,1 1 -1 0 0 1 -1 0 0 line 6, 2-dim rep. 5: 1,1 0 0 1 -1 0 0 -1 1 line 6, 2-dim rep. 5: 1,2 0 0 -1 1 0 0 -1 1 line 6, 2-dim rep. 5: 2,1 1 -1 0 0 -1 1 0 0 line 6, 2-dim rep. 5: 2,2 STOP 2.5. FORMAT OF RESTART FILE (Goto TOC, KEYW, KEYA) This section describes the format of restart files generated by StoBe runs and saved on unit 2. Restart files can be used with input unit 1 to continue / restart unfinished runs or to start new runs with different input parameters. The basic format of a restart file is binary, see also utility transf of Sec. 2.7.2, consisting of different records as follows (parameter types are I4=INTEGER*4, R8=REAL*8, Cn=CHARACTER*n): Record 1 (I4) MAXG max. number of geometries (+) (C80) RUNTIT run title (R8) ENERGY total energy of present configuration Record 2 (R8) CONSTD measure of density convergence (I4) NRESFM restart file format = 0 for old StoBe file format, see Sec. 2.5.1. = 1 for new StoBe file format (default) = 2 for new short StoBe file format (recs. 8, 9, 19 replaced by dummy values, see f.12) (I4) NVBPOS position index of the SCF to be continued in the vibrational analysis (only for vibrational analysis, NVBPOS-1 SCF runs are assumed to be completed, see records 28) (I4) NVBALL total number of SCF runs needed for the vibrational analysis to be continued in the vibrational analysis (vibrational analysis continued only) Record 3 (I4) NLA total number of occupied alpha orbitals, for excited states this may include zero occupation orbitals, see record 4. (I4) NLB total number of occupied beta orbitals for excited states this may include zero occupation orbitals, see record 4. (I4) NCNTRT total number of basis contractions (I4) NCENTR total number of centers (including symmetry equivalents) (I4) NBCH total number of aux. basis functions for charge density fit (I4) NBXC total number of aux. basis functions for XC potential fit (NCENTR*R8) CHARGE atom charges (valence charges if ECPs used) (NCENTR*C4) LABCEN atom labels (element names) (I4) NBCHS number of aux. basis functions (separate s functions only) for charge density fit (I4) NBXCS number of aux. basis functions (separate s functions only) for XC potential fit Record 4 (NCNTRT*R8) DOCA occupations of all alpha orbitals. The internal order follows the scheme described for the MO coefficients, see record 11. Record 5 (NCNTRT*R8) ELGALP Kohn-Sham energies of all alpha orbitals. The internal order follows the scheme described for the MO coefficients, see record 11. Record 6 (NCNTRT*R8) DOCB occupations of all beta orbitals. The internal order follows the scheme described for the MO coefficients, see record 12. Record 7 (NCNTRT*R8) ELGBET Kohn-Sham energies of all beta orbitals. The internal order follows the scheme described for the MO coefficients, see record 12. Record 8 (NT*R8) AMAT (extrapolated) density matrix of alpha orbitals in triangular form, NT = NCNTRT*(NCNTRT+1)/2 This matrix is replaced by a dummy value (=0.D0) for short StoBe file format, see f.12). Record 9 (NT*R8) BMAT (extrapolated) density matrix of beta orbitals in triangular form, NT = NCNTRT*(NCNTRT+1)/2 This matrix is replaced by a dummy value (=0.D0) for short StoBe file format, see f.12). Record 10 (3*NCENTR*R8) CENT (x,y,z) of each center (including symmetry equivalents), latest geometry step (NCENTR*R8) AMASS atom mass of each center (including symmetry equivalents) Records 11 (NCNTRT*R8) CALPHA(i,j) coefficients of alpha orbitals, latest geometry step, j=1,...NCNTRT. The orbitals appear in the order NOCCA(1) occ. orbitals, symmetry 1 NOCCA(2) occ. orbitals, symmetry 2 ... NOCCA(NSYM) occ. orbitals,symmetry NSYM NBAS(1)-NOCCA(1) virt. orbitals, symmetry 1 NBAS(2)-NOCCA(2) virt. orbitals, symmetry 2 ... NBAS(NSYM)-NOCCA(NSYM) virt. orbitals, symmetry NSYM NOTE that for excited states empty orbitals of a symmetry representation below the representation HOMO are considered as occupied orbitals. The application of supersymmetries (see c.16) does not affect the orbital order. Records 12 (NCNTRT*R8) CBETA(i,j) coefficients of beta orbitals, latest geometry step, j=1,...NCNTRT. The orbitals appear in the order NOCCB(1) occ. orbitals, symmetry 1 NOCCB(2) occ. orbitals, symmetry 2 ... NOCCB(NSYM) occ. orbitals,symmetry NSYM NBAS(1)-NOCCB(1) virt. orbitals, symmetry 1 NBAS(2)-NOCCB(2) virt. orbitals, symmetry 2 ... NBAS(NSYM)-NOCCB(NSYM) virt. orbitals, symmetry NSYM NOTE that for excited states empty orbitals of a symmetry representation below the representation HOMO are considered as occupied orbitals. The application of supersymmetries (see c.16) does not affect the orbital order. Record 13 (NBCH*R8) CDCOEF coefficients of charge density fit, latest geometry step Record 14 (NBXC*R8) XCOEFA coefficients of XC potential fit (spin alpha), latest geometry step Record 15 (NBXC*R8) XCOEFB coefficients of XC potential fit (spin beta), latest geometry step Record 16 (NBXC*R8) XCOEFE coefficients of XC energy fit, latest geometry step Record 17 (I4) NSYM number of irreducible representations (24*I4) NBAS number of orbital components for each representation (I4) NTOT total number of orbital components, sum(NBAS) = NCNTRT - NDEL, where NDEL denotes symmetry removals (e.g. x2+y2+z2) (24*I4) NOCCA number of occupied alpha orbitals for each representation (for excited states empty orbitals of a symmetry representation below the representation HOMO are considered as occupied, see records 11) (24*I4) NOCCB number of occupied beta orbitals for each representation (for excited states empty orbitals of a symmetry representation below the representation HOMO are considered as occupied, see records 12) Record 18 (C4) SYMLAB point group symmetry label (NSYM*C4) IRRSYM label of each irreducible representation Record 19 (NT*R8) SXC charge fitting matrix, triangular matrix <<gi(r) gj(r') /|r-r'| >>, NT = NBCH*(NBCH+1)/2 This matrix is replaced by a dummy value (=0.D0) for short StoBe file format, see f.12). Record 20 (3*NCENTR*R8) CENT0 (x,y,z) of each center (including symmetry equivalents), 2nd latest geometry step in optmization or initial equilibrium geometry in vibrational analysis Record 21 (3*NCENTR*R8) HF forces (fx,fy,fz) on each center (including symmetry equivalents), latest geometry step Record 22 (3*NCENTR*R8) HF0 forces (fx,fy,fz) on each center (including symmetry equivalents), 2nd latest geometry step Record 23 (R8) GRADKM1 average gradient between latest and 2nd latest geometry step (R8) GRADMAX1 maximum gradient between latest and 2nd latest geometry step Records 24 (3*NCENTR*R8) HESS(i,j) Hessian matrix, j=1,...3*NCENTR Record 25 (I4) NNNS number of s components in orbital basis (I4) NNNP number of p components in orbital basis (I4) NNND number of d components in orbital basis (NCNTRT*I4) NCNTBK center pointer for each component in orbital basis, NCNTRT = NNNS + 3*NNNP + 6*NNND Record 26 (NT*R8) OVLP overlap matrix in triangular form, NT = NCNTRT*(NCNTRT+1)/2 The following records 27 to 28 appear only if the restart file results from a vibrational analysis, see options e.11) and c.1). Record 27 (I4) NSYMV number of symmetry representations used in the vibrational analysis. Record 27 is followed by I = NSYMV groups of records 27a, b, c (I = 1, NSYMV in the following). Record(s) 27a (C4) IRRSYM(I) symmetry label of representation I, I = 1, NSYMV (I4) NBAS(I) number of vibrational distorsion vectors (true vibrations only) of representation I. Each record 27a is followed by NBAS(I) groups of records 27b, c (J = 1, NBAS(I) in the following). Record(s) 27b (I4) KSYMOP vibrational symmetry index, see below. (3*NCENTR*R8) VBAS(J,I) vibrational distorsion vectors with 3*NCENTR components each Records(s) 27c (appears only if KSYMOP > 0 in record 27b) (9*R8) AROT 3x3 symmetry transformation matrix of the vibrational mode (NCENTR*I4) MAPP atom center pointers of symmetry transformation AROT Record(s) 28 ( N = 1, NVBPOS, see record 2 ) (R8) ENERGY total energy of N-th SCF run (3*NCENTR*R8) HF forces (fx,fy,fz) on each center (including symmetry equivalents) of N-th SCF run (3*R8) DIPOLE(I) dipole moment (dx, dy, dz) of N-th SCF run Restart cases for single point SCF calculations use only records 1 - 19 for input, continued geometry optimizations use the extended set, records 1 - 26, and contunued vibrational analyses use the full set, records 1 - 28. The properties calculation, see RUNType CPROperties, of StoBe and the external analysis of the restart file, see utility anlyz in Sec. 2.7.1, uses the extended set, records 1 - 26. Records 27 - 27c are exact copies of the displacement vector file (unit 60) while records 28 copy the internal vibrational analysis file (unit 61). 2.5.1. PREVIOUS RESTART FILE FORMATS (Goto TOC, KEYW, KEYA) In the following we describe the format used in previous StoBe runs. This format is obsolete and of interest only if very old restart files of large systems are to be revived for runs with the new StoBe version. Note that a previous local Berlin version of StoBe used the new format described in 2.5.1, however, with records 20a,b of the old format (obsolete). The old (obsolete) file format is identical with the latest format in records 1, 3, 8, 9, 11-18, 20, 21-26. Different records and/or formats: Record 2 (R8) CONSTD measure of density convergence (I4) NRESFM restart file format ( = 0 for old format) Record 4 (NLA*R8) DOCA occupations of occupied alpha orbitals Record 5 (NLA*R8) ELGALP Kohn-Sham energies of occupied alpha orbitals Record 6 (NLB*R8) DOCB occupations of occupied beta orbitals Record 7 (NLB*R8) ELGBET Kohn-Sham energies of occupied beta orbitals Record 10 (3*NCENTR*R8) CENT (x,y,z) of each center (including symmetry equivalents), latest geometry step Record 19 (NT*R8) SXC exchange-correlation fitting matrix, full square matrix NT = NBCH*NBCH Record 19a (NT*R8) SXCINV inverse of SXC matrix, full square matrix NT = NBCH*NBCH Records 20a (NLA*R8) GSHMTA(i,j) Difference vector on alpha orbitals used in the vibrational analysis, j=1,...NLA Records 20b (NLB*R8) GSHMTB(i,j) Difference vector on beta orbitals used in the vibrational analysis, j=1,...NLB 2.6. INPUT/OUTPUT FILE UNITS (Goto TOC, KEYW, KEYA) The following list contains all predefined input/output units used by StoBe. In addition, separate units for file input/output may be defined with the input. Unit types are input to StoBe (I), output of StoBe (O), or both input and output (I/O). ----------------------------------------------------------------------------- Units Type Definition ----------------------------------------------------------------------------- 1 I Restart input file for SCF or optimization runs 2 O Restart output file from SCF or optimization runs 3 I Basis set file 4 I Symmetry information file (symbasis) 7 I/O Interactive control parameter file, see f.10) 8 I/O Three-center Coulomb integrals in conventional SCF 9 I/O Exchange-correlation fitting integrals in conv. SCF 9 I/O Default unit for surface definition in drawing MEPs 11 O XAS/XES/RIXS spectrum file (required as input to xrayspec) 12 O Structure file output (pdb, balsac format) I/O Intermediate input/output Dolg pseudopotential titles 17 I/O Auxiliary Coulomb matrix for total energy evaluation 18 I/O Scratch unit for disk-version of matrix inversion (MATINV1) used to build charge-density fit matrix and its inverse; Inverse of charge density fit matrix 19 I/O Scratch unit for symmetrization I (ASYM) 20 I/O Inverse of exchange-correlation fit matrix (SSXC) 21 I/O Scratch unit for geometry optimization I (BROYD3) 22 I/O Scratch unit for geometry optimization II (BROYD3) 23 I/O Scratch unit for geometry optimization III (BROYD3) 24 I/O Scratch unit for geometry optimization IV (BROYD3) 25 I/O Scratch unit for geometry optimization V (BROYD3) 26 I/O Scratch unit for geometry optimization VI (BROYD3) 30 I/O Intermediate scratch unit for geometry optimization 31 I/O Scratch unit for point grids (DENSTY) 32 O Scratch unit for BALSAC population output 33 O Fukui matrix file 41 I/O Square root of overlap matrix (for Loewdin analysis) 42 O Scratch unit for Gaussian-94 format (to be used in Molden) and generic Molden format file output as well as Molekel format file output, intermediate file store for vibrational analysis. 43 I/O Overlap matrix 44 I/O Orbital symmetrization matrix 52 I Input unit for external 1-electron integrals (SCFDOO, STEX) 59 O Output of vibrational modes in Molekel [MOL03] format to be used for (animated) visualization. 60 I/O Input/output unit of displacement basis (vibrational analysis) 61 I/O Scratch unit for vibrational analysis (energies, forces) 62 I/O Scratch unit for vibrational analysis (sel. displacements) 63 I/O Scratch unit for vibrational analysis (sel. energies, forces) 64 I/O Scratch unit for vibrational analysis (copy of unit 60) 65 O Vibrational (IR) spectrum file (required as input to irspec) 66 I/O Scratch unit for plotting output (SURFAC) 70 O File output # 1 (binary) used for NMR analysis 71 O File output # 2 (binary) " 72 O File output # 3 (binary) " 73 O File output # 1 (binary) used for EPR analysis 74 O File output # 2 (binary) " 75 O File output # 3 (binary) " 81 I/O Scratch unit for spin alpha fock matrix (DIIS) 82 I/O Scratch unit for spin alpha DIIS error matrix (DIIS) 88 O Binary file output of plotting (DRAW, MEP) 89 O Scratch unit for plotting output (DENSDES) 91 I/O Scratch unit for spin beta fock matrix (DIIS) 92 I/O Scratch unit for spin beta DIIS error matrix (DIIS) 95 O File output (ASCII) of MO energies and coefficients 96 O File output (ASCII) of MO energies and Mulliken populations 97 O File output (ASCII) of MO energies and Loewdin populations 98 O File output (ASCII) of plotting (DRAW, MEP) 99 O Scratch unit (binary) for plotting (DRAW, MEP) 100... I Restart input files 100 ... 100+nimg for continuing a multi-image SCF run or a NEB path optimization, see a.1), e.12). The file units can be redirected by resetting the unit offset (default = 100), see e.13). 200... I/O Restart intermediate and output files 200 ... 200+nn for a subsequent multi-image SCF run (nn = nimg, see a.1)) or for continuing a NEB path optimization (nn = nneb, see e.12)). The file units can be redirected by resetting the unit offset (default = 200), see e.14). Named files: GMolden.mdf Output file (Gaussian-94 format) to be used in Molden [MOD03] Molden.molf Output file (Molden format) to be used in Molden [MOD03] Molekel.mkl Output file to be used in Molekel [MOL03] struc000.plt Structure plot for BALSAC strucfin.plt Structure plot for BALSAC (final geometry) strucfin.pdb Structure file (Protein Data Bank format) XrayTnnn.out XAS/XES/RIXS (total curves) spectrum file from xrayspec, nnn = task number in xrayspec input (e.g. XrayT002.out) XrayAnnn.out XAS/XES/RIXS (angle-resolved curves) spectrum file from xrayspec, nnn = task number in xrayspec input (e.g. XrayA003.out) XrayPnnn.out XAS/XES/RIXS (polarization-resolved curves) spectrum file from xrayspec, nnn = task number in xrayspec input (e.g. XrayP004.out) dosTnnn.out DOS output (total DOS's) file from doscalc, nnn = task number in doscalc input (e.g. dosT004.out) dosPnnn.out PDOS output (partial DOS's) file from doscalc, nnn = task number in doscalc input (e.g. dosP003.out) ----------------------------------------------------------------------------- 2.7. UTILITIES (Goto TOC, KEYW, KEYA) 2.7.1. ANALYZE RESTART FILES (ANLYZ) (Goto TOC, KEYW, KEYA) Utility anlyz allows the analysis of a calulation based on its (binary) restart file. The calling line reads anlyz.x -[options] sourcefile or anlyz.x sourcefile < inputfile where 'sourcefile' is a StoBe restart file and 'options' are combinations of characters c, e, o, m, M, l, L, d, D, s, i or a with c : listing of all centers, Balsac/Plot3D (file Structure.plt), XYZ (file Structure.xyz), PDB (files Structure.pdb) format output files for postprocessing. e : listing of all level energies. o : listing of all orbital coefficients. m5 (m6) : Short Mulliken, no Loewdin population analysis (5 true component (6 cartesian) d basis functions) M5 (M6) : Full Mulliken, no Loewdin population analysis, no bond orders for orbitals (5 true component (6 cartesian) d basis functions) l5 (l6) : Short Loewdin, no Mulliken population analysis (5 spherical (6 cartesian) d basis functions) L5 (L6) : Full Loewdin, no Mulliken population analysis, no bond orders for orbitals (5 true component (6 cartesian) d basis functions) ml5 (ml6) : Short Mulliken, short Loewdin population analysis (5 true component (6 cartesian) d basis functions) Ml5 (Ml6) : Full Mulliken, short Loewdin population analysis, no bond orders for orbitals (5 true component (6 cartesian) d basis functions) mL5 (mL6) : Short Mulliken, full Loewdin population analysis, no bond orders for orbitals (5 true component (6 cartesian) d basis functions) ML5 (ML6) : Full Mulliken, full Loewdin population analysis, no bond orders for orbitals (5 true component (6 cartesian) d basis functions) dx : saves energies and populations on units 96/97 for subsequent (P)DOS calculations (using 5 true component d basis functions). x = 1 Mulliken analysis, output on unit 96. = 2 Loewdin analysis, output on unit 97. = 3 both Mulliken / Loewdin analysis, output on units 96, 97. Dx : same as option dx but using 6 cartesian d basis functions). a : combines options c, e, o, M5, L5, d3. s : short analysis of restart file (syntax check only). i : include records 20a,b in restart file read. Option parameters can also be provided by an ASCII input file (unit 5) consisting of a series of keywords analogous to the original StoBe input and described in the following. The keyword input has to be finished by a line reading 'END'. , ----------------------------------------------------------------------------- 1) CENTers ----------------------------------------------------------------------------- This keyword includes a listing of all center coordinates. In addition, Balsac/Plot3D (file Structure.plt), XYZ (file Structure.xyz), PDB (files Structure.pdb) format output files are saved for postprocessing. This keyword is equivalent to option 'c' in the command line input. ----------------------------------------------------------------------------- 2) LEVEls ----------------------------------------------------------------------------- This keyword includes a listing of all energy levels, both in their order included in the restart file and in straight energetic order. This is equivalent to option 'e' in the command line input. ----------------------------------------------------------------------------- 3) PORBitals ----------------------------------------------------------------------------- This keyword includes a listing of all orbital coefficients and is equivalent to option 'o' in the command line input. ----------------------------------------------------------------------------- 4) MULLiken [OFF / ON [FULL, ORBItal] ] ----------------------------------------------------------------------------- This keyword includes a listing of the gross Mulliken population analysis and Mayer bond orders. Both total and orbital resolved data can be requested where, in the case of orbital values, the analysis is performed for all occupied and virtual orbitals included in the listing, see also StoBe option f.9). By default, no populations nor bond orders are included. The analysis is performed in spherical or cartesian gaussians depending on the internal representation of d basis functions used in the electronic state calculation, see option ORBItalchoice 6). Keyword line MULLiken OFF lists neither populations nor bond orders. This is the default if no keyword is given. MULLiken ON (or MULLiken) lists total Mulliken populations and total Mayer bond orders. MULLiken ON ORBItal lists total and orbital resolved Mulliken populations as well as total Mayer bond orders. MULLiken ON FULL lists total and orbital resolved Mulliken populations as well as total and orbital resolved Mayer bond orders. ----------------------------------------------------------------------------- 5) LOEWdin [ ON [FULL, ORBItal] ] ----------------------------------------------------------------------------- This keyword includes a listing of the Loewdin population analysis and Mayer bond orders. Both total and orbital resolved data can be requested where, in the case of orbital values, the analysis is performed for all occupied and virtual orbitals included in the listing, see also StoBe option f.9). By default, no populations nor bond orders are included. The analysis is performed in spherical or cartesian gaussians depending on the internal representation of d basis functions used in the electronic state calculation, see option ORBItalchoice 6). Keyword line LOEWdin OFF lists neither populations nor bond orders. This is the default if no keyword is given. LOEWdin ON (or LOEWdin) lists total Loewdin populations and total Mayer bond orders. LOEWdin ON ORBItal lists total and orbital resolved Loewdin populations as well as total Mayer bond orders. LOEWdin ON FULL lists total and orbital resolved Loewdin populations as well as total and orbital resolved Mayer bond orders. ----------------------------------------------------------------------------- 6) ORBItalchoice ----------------------------------------------------------------------------- This keyword determines the internal representation of d basis functions used in the Mulliken or Loewdin population analysis, see 4), 5). Option keyword 5D (default) uses the 5 true atomic gaussian d functions (dz2,dx2-y2, dxy, dxz, dyz) to represent d orbital basis functions. 6D uses the 6 cartesian gaussian d functions (x2, y2, z2, xy, xz, yz) to represent d orbital basis functions. ----------------------------------------------------------------------------- 7) DOSOutput [MULLiken, LOEWdin] [5D, 6D] ----------------------------------------------------------------------------- This keyword saves Kohn-Sham levels and Mulliken/Loewdin populations on external files (unit 96 for Mulliken populations, keyword MULLiken; unit 97 for Loewdin populations, keyword LOEWdin, see also StoBe option f.11) to be used for subsequent densities-of-states calculations with utility doscalc, see Sec. 2.7.5. Option keywords 5D, 6D determine the internal representation of d basis functions used in the Mulliken or Loewdin population analysis, see 6). ----------------------------------------------------------------------------- 8) REC20include ----------------------------------------------------------------------------- This keyword allows to process old restart files which still contain records 20a, b, see 2.5.1. ----------------------------------------------------------------------------- 9) SYNTax ----------------------------------------------------------------------------- This keyword performs a syntax check of the restart file ignoring all other option keywords. ----------------------------------------------------------------------------- 10) ALL ----------------------------------------------------------------------------- This keyword performs a global analysis of the restart file combining options 1) - 7) and corresponding to an input sequence > CENTERS > LEVELS > PORBITALS > MULLIKEN ON FULL > LOEWDIN ON FULL > ORBITALCHOICE 5D > DOSOUTPUT MULLIKEN LOEWDIN 5D Examples: > anlyz.x -d3ceM5l5o NH3.res > NH3anlyz.out or equivalently > anlyz.x NH3.res < infile > NH3anlyz.out with an input file 'infile' reading ----infile---------------- DOSO MULLIKEN LOEWDIN CENTERS LEVELS MULLIKEN ON ORBITAL LOEWDIN ON ORBITAL 5D PORBITAL END ----infile---------------- 2.7.2. CONVERT/COPY/COMPARE RESTART FILES (TRANSF) (Goto TOC, KEYW, KEYA) Utility transf allows to convert restart files from binary to ASCII format, to copy respective binary / ASCII format restart files, or to compare restart files. This option is useful if restart files have to be transfered between systems with incompatible binary number representation. The calling line reads transf.x -[options] sourcefile targetfile where 'sourcefile' is a StoBe restart file, 'targetfile' is the output or another StoBe restart file, and 'options' define the input/output file format as aa : source ASCII, target ASCII ab : source ASCII, target binary ba : source binary, target ASCII bb : source binary, target binary If a character "l" is attached to the option ("ab" -> "abl") then the output listing will include transfer details. "c" is attached to the option ("bb" -> "bbc") then the utility will compare all records of the input file with those of the (existing output file) and list numerical differences. "s" is attached to the option ("ab" -> "abs") then the output file is saved in short StoBe restart file format. NOTE that this utility works only with the new (short) StoBe restart file format detailed in Sec. 2.5. 2.7.3. PLOTTING UTILITY (KHPLTPS) (Goto TOC, KEYW, KEYA) Utility khpltps (executable khpltps.x) allows to use binary output files from StoBe runs to generate PostScript output of (shaded) contour plots showing variations of - electrostatic potentials (IPROP= 2, DRAW MEPS PLANe) - orbital wavefunctions (IDESS= 1, DRAW DENSity ORBItal) - orbital densities (IDESS= 2, DRAW DENSity ORBItal) - total electron densities (IDESS= 3, DRAW DENSity TOTAl) - Ex, Ec, Exc energy densities (IDESS= 4, DRAW DENSity EXchange) - XC energies (Ex*rho,Ec*rho,Exc*rho) (IDESS= 5, DRAW XCENergy) - spin difference density (IDESS= 6, DRAW SFUNction) - X, C and XC potentials (IDESS= 7, DRAW XCPOtential) - Fukui function (relaxed orbitals) (IDESS= 9, DRAW FUKUI) - Fukui function (frozen orbitals) (IDESS= 10, DRAW FFUKUI) - Voronoi indices (IDESS= 11, DRAW VOROnoi) - Bader indices (IDESS= 12, DRAW BADEr) in planar sections of a molecule/cluster, see keyword DRAW, e.5). The utility requires binary input (unit 4) from StoBe runs (StoBe output unit 88 created by DRAW options) as well as ASCII input (unit 5) consisting of a series of keywords analogous to the original StoBe input and described in the following. The ASCII input has to be finished by a line 'END'. ----------------------------------------------------------------------------- 1) ANALyze ----------------------------------------------------------------------------- This keyword (used by itself) performs an analysis of the binary input file (unit 4) from a StoBe run to be used for plotting. ----------------------------------------------------------------------------- 2) [PRINtplot, CONTours, SHADes, SHCOntours] nset ----------------------------------------------------------------------------- Any combination of these keywords uses set no. nset of the binary input file to generate PostScript or print plot output with PRINtplot giving a graphical representation of the function of set no. nset. CONTours producing PostScript file output (unit 10) with a contour plot of the function of set no. nset. SHADes producing PostScript file output (unit 10) with a shading plot of the function of set no. nset. SHCOntours producing PostScript file output (unit 10) with a shaded contour plot of the function of set no. nset. PostScript file output (shading, contours) requires one additional line following the keyword line with KEY KZERO IZENT KNUCC KPRT VALMIN DELTA VALMAX RELY where KEY = 'ABS' the absolute values of parameters VALMIN, DELTA, VALMAX are used for contour values, |DELTA| is used as an additive increment. = 'REL' VALMIN, DELTA, VALMAX are taken as fractions of the largest absolute value of the function calculated for the plot, |DELTA| is used as an additive increment. = 'LOG' the absolute values of parameters VALMIN, DELTA, VALMAX are used for contour values, |DELTA| > 1 is used as a multiplicative increment. KZERO = 1 suppress zero value contour line. = 0 draw all contour lines as defined. IZENT = 1 if VALMIN and VALMAX differ in sign, the values are shifted internally such that at least one contour line corresponds to the exact value zero. = 0 option not used. KNUCC = 1 mark centers on a plot. = 0 no marked centers on a plot. KPRT = 0 no coordinate printing. = 1 print value and number of points for each contour line = 2 print excplicit contour line coordinates (for testing purposes only, large output!). VALMIN smallest value of contour lines to be plotted. For keyword LOG +/-|VALMIN| will be used for positive/negative array values, only VALMIN > 0 is meaningful. DELTA increment value between adjacent contour lines to be plotted. For keywords ABS/REL the value must be > 0.0, incrementing happens additively evaluating contours for VALMIN, VALMIN+DELTA, VALMIN+2*DELTA, ... VALMAX. For keyword LOG the increment value must be > 1.0, incrementing happens multiplicatively evaluating contours for +/-|VALMIN|, +/-|VALMIN|*DELTA, +/-|VALMIN|*DELTA**2, ... +/-|VALMAX|. VALMAX largest value of contour lines to be plotted. For keyword LOG +/-|VALMAX| will be used for positive/negative array values, only VALMAX > VALMIN > 0 is meaningful. RELY absolute value of plot size along y direction given in cm (default = 25.0 cm). ----------------------------------------------------------------------------- 3) PSFIle ----------------------------------------------------------------------------- This keyword followed by one ASCII line (A72) denotes the file name for PostScript print output of the present run. If no name is given the standard name of fortran unit 10, e.g. fort.10, is used instead. ----------------------------------------------------------------------------- 4) ORIEntation [SEPArate] ----------------------------------------------------------------------------- This keyword allows to print an orientation plot showing the positions of all atom centers with projections inside the plot area of the previous plot. The option keyword SEPArate prints the orientation plot on a separate page. Note that this option requires a PostScript plot to be finished before it can be used. ----------------------------------------------------------------------------- 5) NEWPage ----------------------------------------------------------------------------- This keyword starts a new page in the PostScript output allowing to separate groups of plots. The keyword NEWPage must appear directly after a PostScript plot request, see options CONTours, SHADes, SHCOntours in 2). The binary file used as input (unit 4) is created by a StoBe run (unit 88) when the DRAW keyword is used, see e.5). It consists of different records as follows (parameter types are I4=INTEGER*4, R8=REAL*8, Cn=CHARACTER*n): Record n.1 (C80) RUNTIT run title of set n Record n.2 (2*R8) XA, XE start and end point of plot section along x (real space molecular coordinates). The section length along x is given by (XE - XA). (2*R8) YA, YE start and end point of plot section along y (real space molecular coordinates). The section length along y is given by (YE - YA). (I4) NX number of calculated points along x (I4) NY number of calculated points along y (I4) NPTS block size of storing function array f(i,j) Records n.3 (NX*NY*R8) FUNCT function array f(i,j), i=1,NX, j=1,NY in blocks of NPTS values (except last block) Record n.4 (I4) NCENT total number of atom centers of the system Record n.5 (NCENT*(C4,3R8)) LAB,X,Y,Z atom labels and coordinates of all atom centers wrt. plot plane where a sequence of records n.1, n.2, n.3 occurs in the file depending on the number of plot arrays generated. Example input: >ANALYZE >PRINTPLOT 3 >CONTOUR SHCONTOUR 4 >ABS 0 0 0 .01 .01 .1 10. >PRINT SHCONT 1 >ABS 0 0 0 .01 .01 .1 15. >ORIENT SEPARATE >SHADE 2 >REL 1 0 0 .01 .01 .1 12. >PSFILE >newfile.eps >END produces - a print plot of set 3, - a contour plot and a shaded contour plot of set 4 (absolute contours between 0.01 and 0.1 with increments 0.01), real plot size along y = 10 cm, - a print plot and shaded contour plot of set 1 (absolute contours between 0.01 and 0.1 with increments 0.01), real plot size along y = 15 cm, - an orientation plot of all atom centers on a separate page, - a shading plot of set 2 (relative shding values between 0.01 and 0.1 with increments 0.01), real plot size along y = 12 cm, - and all PostScript print output is saved on file newfile.eps . 2.7.4. DOING RESONANT INELASTIC X-RAY SCATTERING (RIXS) ANALYSIS (Goto TOC, KEYW, KEYA) The keyword RIXS prints both the occupied and unoccupied dipole transition moments. If a true RIXS treatment of the data is desired, this is done in separate programs e.g. RIXS_detune.x as supplied on Source/XRAY/RIXS. Here, the spectrum is spherically averaged according to the group theory formulation by Y. Luo et al (J. Phys. B27, 4169 (1994); Phys. Rev. B52, 14478 (1995)). Input example: First line: Gamma, Gamm_L, where Gamma is the life time broadening (eV) and Gamm_L is the fwhm width of the incoming photon distribution (eV). Second line: 0.5 eV detuning, 2 core orbs, 2 occ orbs, 5 unocc orbs, start, end, atom for which the excitation is considered, resonant level number, shift (should be 0.0 except for test cases) The example below will give 0.5 eV detuning from the unoccupied level at -3.956 eV. -------------------------------------------------------------------------- 0.08 0.2 %Gamma, Gamm_L 0.5 2 2 5 242. 265. 3 1 0.0 % See explanation above -267.37 %Core level energy level 1 -267.36 %Core level energy level 2 -4.876 -0.0022 0.0000 0.0000 %Unocc orbs on atom 1 (5 in this case) -4.188 0.0000 0.0040 0.0096 -3.956 0.0000 0.0021 -0.0011 -3.685 0.0000 0.0080 0.0013 -3.495 0.0000 -0.0073 -0.0114 -10.991 0.0000 0.0052 -0.0113 %Occ orbs on atom 1 (2 in this case) -5.648 0.0000 -0.0028 -0.0047 -4.876 -0.0011 0.0000 0.0000 %Unocc orbs on atom 2 (5 in this case) -4.188 0.0000 0.0030 0.0044 -3.956 0.0000 0.0011 -0.0021 -3.685 0.0000 0.0080 0.0023 -3.495 0.0000 0.0073 0.0104 -10.994 0.0010 0.0000 0.0000 %Occ orbs on atom (2 in this case) -5.648 0.0000 0.0028 0.0037 -------------------------------------------------------------------------- 2.7.5. TOTAL/PARTIAL DENSITIES-OF-STATES (DOSCALC) (Goto TOC, KEYW, KEYA) Utility doscalc (executable doscalc.x) allows you to use output files from StoBe runs to generate arrays of (projected) densities-of-states, (P)DOS, for graphical output or for further processing. Total DOSs are generated by gaussian level broadening where the broadening width (FWHM) can be chosen freely. PDOSs are also generated by gaussian level broadening where different projections are available based on the Mulliken analysis. The projections allow you to select atoms, orbitals, orbital angular components, spin, and symmetry in any combination. File output of (P)DOSs consists of array listings combining each energy of a given range with one or three (spin polarized case) (P)DOS values. Each array listing is saved in a separate file dosT(k).out (for total DOS's) or dosP(k).out (for partial DOS's) where k numbers the (P)DOS curves according to their sequence given in the input and '(k)' is a 3-digit version of k padded with leading zeros. The (P)DOS listings can also be printed, see option PRINt below. The utility requires input (unit 1) from StoBe runs (StoBe output unit 96, created by option DOSO) as well as ASCII input (unit 5) consisting of a series of keywords analogous to the original StoBe input and described in the following. The keyword input has to be finished by a line reading 'END'. ----------------------------------------------------------------------------- 1) TITLe ----------------------------------------------------------------------------- Brief descriptive title of the present doscalc run on the line following the keyword. This title appears in the output file(s). If no title is given doscalc takes the title of the StoBe input file (unit 1) Example: >TITLE >h2o local opt ----------------------------------------------------------------------------- 2) RANGe [opt] ebot etop ----------------------------------------------------------------------------- Range of energies [ebot, etop] (in eV) for which (P)DOSs are to be calculated. The optional keyword opt is any combination of the three characters h, lg g. For option opt = h Ehomo is used as only reference energy: range = [Ehomo+ebot, Ehomo+etop] = hg same as h with exact zero (P)DOS inside HOMO-LUMO gap = l Elumo is used as only reference energy: range = [Elumo+ebot, Elumo+etop] = lg same as l with exact zero (P)DOS inside HOMO-LUMO gap = lh Elumo and Elumo are used as reference energies: range = [Ehomo+ebot, Elumo+etop] = lhg same as lh with exact zero (P)DOS inside HOMO-LUMO gap = g absolute range [ebot, etop], exact zero (P)DOS inside HOMO-LUMO gap Omitting keyword RANGe correponds to the default range = [Ehomo - 20, Ehomo + 10] with exact zero (P)DOS inside HOMO-LUMO gap Examples: >RANGE -100.0 .5 : range between -100.0 eV and +0.5 eV >RANGE hlg -20 10 : range between Ehomo - 20 eV and Elumo + 10 eV density inside HOMO-LUMO gap suppressed ----------------------------------------------------------------------------- 3) POINts npts ----------------------------------------------------------------------------- Number of energies within the range [ebot, etop] for which (P)DOSs are to be calculated. The energy sequence is equidistant, given by e = ebot + (etop-ebot)/(npts-1)*(i-1), i = 1, npts If this keyword is omitted the default value npts = 1000 is used. Example: >POINTS 500 ----------------------------------------------------------------------------- 4) WIDTh broad1 [ broad2 ewid1 ewid2 ] ----------------------------------------------------------------------------- Width (full-width-at-half-maximum, FWHM) (in eV) used in the gaussian level broadening. The short version uses only a global broadening of width broad1. The full version uses an energy dependent broadening where a width broad1 is used for level energies < ewid1, width broad2 is used for level energies > ewid2 ( > ewid1), and for level energies between ewid1 and ewid2 a linear interpolation broad(E) = broad1 + (broad2-broad1)*(E-ewid1)/ewid2-ewid1) is used. If this keyword is omitted the default value broad = 1 eV is used. Examples: >WIDTH .5 >WIDTH .5 .8 -10.0 -5.0 ----------------------------------------------------------------------------- 5) PRINt ----------------------------------------------------------------------------- This keyword includes lists of the (P)DOS arrays in the standard printout in addition to file output. ----------------------------------------------------------------------------- 6) TOTAldos [SPIN] ----------------------------------------------------------------------------- This keyword initializes a total DOS calculation. By default ALL atoms, orbitals, orbital angular components, and symmetries are selected. Up to 10 total and partial DOSs can be calculated in one doscalc run. The file (and print) output lists one DOS array (for combined spin up/down). If the optional keyword SPIN is included, the total DOS is evaluated for spin up and spin down orbitals separately. The file (and print) output lists three DOS arrays (for combined spin up/down, for spin up, for spin down orbitals). ----------------------------------------------------------------------------- 7) PARTialdos [SPIN] ----------------------------------------------------------------------------- This keyword initializes a partial (projected) DOS calculation. By default NO atoms, orbitals, orbital angular components, and symmetries are selected and have to be selected separately (with keywords ALL, ATOMs, ORBItal, COMPonent, SYMMetry). Up to 10 total and partial DOSs can be calculated in one doscalc run. The file (and print) output lists one DOS array (for combined spin up/down). If the optional keyword SPIN is included, the PDOS is evaluated for spin up and spin down orbitals separately. The file (and print) output lists three PDOS arrays (for combined spin up/down, for spin up, for spin down orbitals). ----------------------------------------------------------------------------- 8) ALL ----------------------------------------------------------------------------- This keyword (not needed in total DOS calculations, see keyword TOTAldos) selects all atoms, orbitals, orbital angular components, and symmetries in a PDOS calculation. ----------------------------------------------------------------------------- 9) NONE ----------------------------------------------------------------------------- This keyword (not needed in total DOS calculations, see keyword TOTAldos) deselects all atoms, orbitals, orbital angular components, and symmetries in a PDOS calculation. ----------------------------------------------------------------------------- 10) ATOMselect {ALL, NONE, INCLude, EXCLude} ----------------------------------------------------------------------------- This keyword (not needed in total DOS calculations, see keyword TOTAldos) selects or deselects atoms in a PDOS calculation. Several keyword lines can be added and will be evaluated sequentially to achieve the final atom selection. This keyword line comes with different versions ATOMselect ALL [weight] selects all atoms available from the StoBe input file. The optional parameter weight uses a scaling factor to scale atom contributions. If no weight value is given, a default weight = 1.0 is used. ATOMselect NONE deselects all atoms available from the StoBe input file. ATOMselect INCLude elmnt [weight] selects all atoms of a given element name elmnt (e.g. H, O, V, Na) available from the StoBe input file. The optional parameter weight uses a scaling factor to scale the atom contributions. If no weight value is given, a default weight = 1.0 is used. ATOMselect INCLude n1 n2 n3 ... weight selects atoms no. n1 n2 n3 ... (up to 10 atom numbers are allowed) available from the StoBe input file. The parameter weight uses a scaling factor to scale the atom contributions. Note that with this keyword choice the parameter weight MUST be given at the end of the list. This option line can be repeated if more than 10 atom numbers are requested. ATOMselect EXCLude elmnt deselects all atoms of a given element name elmnt (e.g. H, O, V, Na) from the present atom list. ATOMselect EXCLude n1 n2 n3 ... deselects atoms no. n1 n2 n3 ... (up to 10 atom numbers are allowed) from the present atom list. This option line can be repeated if more than 10 atom numbers are requested. Examples: >ATOM ALL >ATOM NONE >ATOM INCL V .5 >ATOM INCL 1 2 3 .8 >ATOM EXCL O >ATOM EXCL 12 1 5 ----------------------------------------------------------------------------- 11) ORBItalselect {ALL, NONE, INCLude, EXCLude} ----------------------------------------------------------------------------- This keyword (not needed in total DOS calculations, see keyword TOTAldos) selects or deselects orbitals in a PDOS calculation. Several keyword lines can be added and will be evaluated sequentially to achieve the final orbital selection. This keyword line comes with different versions ORBItalselect ALL [ALFA, BETA] selects all orbitals available from the StoBe input file. The optional keywords ALFA, BETA allow you to select only alpha or beta spin orbitals. ORBItalselect NONE deselects all orbitals available from the StoBe input file. ORBItalselect INCLude n1 n2 n3 ... selects orbitals no. n1 n2 n3 ... (up to 10 orbital numbers are allowed) available from the StoBe input file. This option line can be repeated if more than 10 orbital numbers are requested. ORBItalselect EXCLude n1 n2 n3 ... deselects orbitals no. n1 n2 n3 ... (up to 10 orbital numbers are allowed) from the present orbital list. This option line can be repeated if more than 10 orbital numbers are requested. Examples: >ORBITAL ALL >ORBITAL NONE >ORBITAL INCL 1 13 124 >ORBITAL EXCL 1 2 3 ----------------------------------------------------------------------------- 12) COMPonentselect {ALL, NONE, INCLude, EXCLude} ----------------------------------------------------------------------------- This keyword (not needed in total DOS calculations, see keyword TOTAldos) selects or deselects orbital angular components in a PDOS calculation. Several keyword lines can be added and will be evaluated sequentially to achieve the final component selection. This keyword line comes with different versions COMPonentselect ALL selects all orbital angular components available from the StoBe input file. For atoms with s, p, d functions this includes altogether 10 components: s, x, y, z, xx, xy, xz, yy, yz, zz. COMPonentselect NONE deselects all orbital angular components available from the StoBe input file. COMPonentselect INCLude c1 c2 c3 ... selects orbital angular components available from the StoBe input file using their generic names ci = S, P (combines x,y,z), D (combines xx, xy, xz, yy, yz, zz), X, Y, Z, XX, XY, XZ, YY, YZ, ZZ. Note that entries ci have to be given by upper case characters. The definition of d components depends also on the choice of the d representation in the preceding StoBe run (5d vs. 6d). For a 5d representation 'XX' is meaningless, 'YY' refers to dz2, and 'ZZ' refers to dx2-y2. COMPonentselect EXCLude c1 c2 c3 ... deselects orbital angular components from the present component list using their generic names ci = S, P (combines x,y,z), D (combines xx, xy, xz, yy, yz, zz), X, Y, Z, XX, XY, XZ, YY, YZ, ZZ. Note that entries ci have to be given by upper case characters. The definition of d components depends also on the choice of the d representation in the preceding StoBe run(5d vs. 6d). For a 5d representation 'XX' is meaningless, 'YY' refers to dz2, and 'ZZ' refers to dx2-y2. Examples: >COMPONENT ALL >COMPONENT NONE >COMPONENT INCL S P XX YY ZZ >COMPONENT EXCL D Z ----------------------------------------------------------------------------- 13) SYMMetryselect {ALL, NONE, INCLude, EXCLude} ----------------------------------------------------------------------------- This keyword (not needed in total DOS calculations, see keyword TOTAldos) selects or deselects symmetry components in a PDOS calculation. Several keyword lines can be added and will be evaluated sequentially to achieve the final symmetry component selection. This keyword line comes with different versions SYMMetryselect ALL selects all symmetry components available from the StoBe input file (e.g. A1, A2, E for NH3 in C3v symmetry). SYMMetryselect NONE deselects all symmetry components available from the StoBe input file (e.g. A1, A2, E for NH3 in C3v symmetry). SYMMetryselect INCLude s1 s2 s3 ... selects symmetry components available from the StoBe input file using their generic names si (e.g. A1, A2, E for NH3 in C3v symmetry). Note that entries ci have to conform with the notation used in StoBe. SYMMetryselect EXCLude s1 s2 s3 ... deselects symmetry components from the present component list using their generic names si (e.g. A1, A2, E for NH3 in C3v symmetry). Note that entries ci have to conform with the notation used in StoBe. Examples: >SYMMETRY ALL >SYMMETRY NONE >SYMMETRY INCL A1 E >SYMMETRY EXCL A1 ----------------------------------------------------------------------------- 14) NOOCcupations ----------------------------------------------------------------------------- This keyword (used in both total and partial DOS calculations) ignores level occupations given with the StoBe file (input unit 1) and sets all orbitals occupations to 1.0. This allows to calculate (P)DOS's for any energy range including occupied as well as empty (virtual) orbitals. By default, level occupations are taken into account in the (P)DOS calculations. ----------------------------------------------------------------------------- 15) VIRTualorbitalsonly ----------------------------------------------------------------------------- This keyword (used in both total and partial DOS calculations) includes only unoccupied levels (referring to virtual orbitals) of the StoBe file (input unit 1) in the (P)DOS calculation, setting occupations of all virtual orbital to 1.0 and of all occupied orbitals to 0.0. By default, the level occupations of the input file are taken into account in the (P)DOS calculations. ----------------------------------------------------------------------------- 16) BALPopulation ----------------------------------------------------------------------------- This keyword initiates BALSAC [H00] format file output to visualize and analyze atom populations determined in total and partial DOS calculations, see also e.4). The output files can be used as input to BALSAC where the atoms of a cluster/ molecule are displayed as shaded balls. The ball radii are defined by the actual atom populations where positive/negative populations are shown by different colors (red/blue by default) in the plot. The selection of Mulliken or Loewdin populations is determined by the DOS constraints, see above. If spin selection in the DOS output is defined, see keyword SPIN in options 6), 7), the present option generates three files, popAxxx.plt (spin alpha result), popBxxx.plt (spin beta result), popTxxx.plt (total (spin alpha + beta) result). Here xxx is the sequence number of the total / partial DOS calculation in the present DOSCALC run. If no spin selection is defined, only file output on popTxxx.plt (total (spin alpha + beta) result) is made. By default, no BALSAC file output is generated. The following full example starts from a NH3 calculation and calculates a total DOS with spin resolution together with BALSAC file output on file popA001.plt, popB001.plt, popT001.plt, followed by a PDOS calculation including only hydrogen s contributions in all symmetry representations. >TITLE >NH3 local opt density of states >RANGE -386 0 >POINTS 2000 >WIDTH .7 >PRINT >TOTAL SPIN >BALPOP >PARTIAL >ATOM INCL H >ORBITAL ALL >COMP INCL S >SYMM ALL >END 2.7.6. TOTAL/ANGLE-RESOLVED XRAY SPECTRA (XRAYSPEC) (Goto TOC, KEYW, KEYA) Utility xrayspec (executable xrayspec.x) allows you to use output files from StoBe runs to generate total and angle-resolved Xray (XAS, XES, RIXS) spectra for graphical output or for further processing. The spectra include gaussian broadening where the broadening width (FWHM) can be chosen freely. Angle-resolved spectra can be computed for any polarization direction or incident/exit beam given by angles theta, phi. File output of spectra consists of array listings combining each transition energy with the corresponding absorption (given in a.u.). Each spectrum is saved in a separate file XrayTnnn.out (for total spectra), XrayAnnn.out (angle-resolved spectra), or XrayPnnn.out (polarization- resolved spectra) where 'nnn' is a 3-digit number (padded with leading zeros id needed) counting the sequence of spectra defined by the input. For example, the second total spectrum file is named XrayT002.out. The spectra listings can also be printed, see option PRINt below. The utility requires input (unit 1) from StoBe runs (StoBe output unit 11, created by option XRAY) as well as ASCII input (unit 5) consisting of a series of keywords analogous to the original StoBe input and described in the following. The keyword input has to be finished by a line reading 'END'. ----------------------------------------------------------------------------- 1) TITLe ----------------------------------------------------------------------------- Brief descriptive title of the present xrayspec run on the line following the keyword. This title appears in the output file(s). Example: >TITLE >H2O O1s -> virt Xray spectrum ----------------------------------------------------------------------------- 2) XRAY [type] ----------------------------------------------------------------------------- Type of spectrum to be calculated where type = 'XAS' refers to an Xray absorption spectrum. This is the default if no spectrum type is defined. = 'XES' refers to an Xray emission spectrum. = 'RIXS' refers to a RIXS (resonant ionization Xray spectrum) spectrum. NOTE that so far only the two RIXS steps, core->unocc. excitation, valence->core deexcitation, can be accounted for in separate spectra. Spectra of the complete RIXS process are not available. The RIXS analysis allows an additional keyword (OCC, VIRT, ALL) where XRAY RIXS OCC includes only valence->core deexcitations in the spectrum. XRAY RIXS VIRT includes only core->unocc. excitations in the spectrum. XRAY RIXS ALL superimposes core->unocc. as well as valence->core transitions in the spectrum. This is the default if the additional keyword is omitted. The program checks the data input (unit 1) for consistency with the spectrum type, see also option XRAY, e.7). NOTE that this keyword line must appear before any spectrum definition (TOTAlspectrum, ANGLeresolvedspectrum, POLArizedspectrum), see below. ----------------------------------------------------------------------------- 3) TOTAlspectrum [ipos] ----------------------------------------------------------------------------- This spectrum keyword initializes the evaluation of a total (angle-averaged and unpolarized) Xray spectrum calculation using data from previous StoBe runs. The spectral weight of each transition is given by (a) dipole transitions W(i) = 2/3*f * E(i) * (Dx**2 + Dy**2 + Dz**2) where D(i) = (Dx, Dy, Dz) are dipole transition matrix elements of transition i, E(i) is the transition energy, and f (= 1000) is an empirical scaling factor. (b) quadrupole transitions W(i) = 1/15*f * E(i) * [ 3*(Qxx + Qyy + Qzz)**2 - 4*(Qxx*Qyy + Qxx*Qzz + Qyy*Qzz) + 4*(Qxy**2 + Qxz**2 + Qyz**2) ] where Q(i) = (Qxx, Qxy, Qxz, Qyy, Qyz, Qzz) are quadrupole transition matrix elements of transition i, E(i) is the transition energy, and f (= 1000) is an empirical scaling factor. Set no. ipos of the StoBe spectrum file (output unit 65) is used for input (as file unit 1). If ipos is omitted, ipos = 1 is assumed. Corresponding spectrum parameters (keywords RANGe, POINts, WIDTh, PRINt, SCALe, OFFSet), if different from default values, have to be given immediately after the keyword line. Output files are named XrayTnnn.out where nnn = 001, 002, ... Example: >TOTAL 1 ----------------------------------------------------------------------------- 4) ANGLeresolvedspectrum theta phi [ipos] ANGLeresolvedspectrum xx yy zz ipos (alternative format) ----------------------------------------------------------------------------- This keyword initializes the evaluation of an angle-resolved Xray spectrum using data from a previous StoBe run. The spectrum refers to a given direction of the incident/exciting light beam where the absorption is an average over all possible polarization vectors perpendicular to the beam direction e given by e = ( sin(theta)*cos(phi), sin(theta)*sin(phi), cos(theta) ) . Here theta, phi (in degrees) are correponding polar and azimuthal angles. The alternative input uses a vector R = (xx, yy, zz) which is converted to angles theta, phi according to R = |R| ( sin(theta)*cos(phi), sin(theta)*sin(phi), cos(theta) ) . The spectral weight of each transition is given by (a) dipole transitions W(i) = f * E(i) * 1/2 [ (Dx**2 + Dy**2 + Dz**2) - (Dx*ex + Dy*ey + Dz*ez)**2 ] where D(i) = (Dx, Dy, Dz) are dipole transition matrix elements of transition i, E(i) is the transition energy, and f (= 1000) is an empirical scaling factor. (b) quadrupole transitions W(i) = f * E(i) * 1/8 { 3 [ Qxx*(1-ex**2) + Qyy*(1-ey**2) + Qzz*(1-ez**2) - 2*Qxy*ex*ey - 2*Qxz*ex*ez - 2*Qyz*ey*ez ]**2 + 4 [ (Qyz**2-Qyy*Qzz)*ex**2 + (Qxz**2-Qxx*Qzz)*ey**2 + (Qxy**2-Qxx*Qyy)*ez**2 - 2*(Qxz*Qyz-Qxy*Qzz)*ex*ey - 2*(Qxy*Qyz-Qxz*Qyy)*ex*ez - 2*(Qxy*Qxz-Qxx*Qyz)*ey*ez ] } where Q(i) = (Qxx, Qxy, Qxz, Qyy, Qyz, Qzz) are quadrupole transition matrix elements of transition i, E(i) is the transition energy, and f (= 1000) is an empirical scaling factor. Set no. ipos of the StoBe spectrum file is used for input (unit 1). If ipos is omitted, ipos = 1 is assumed. Note that for direction input in Cartesian coordinates xx, yy, zz the position index ipos MUST be given. Corresponding spectrum parameters (keywords RANGe, POINts, WIDTh, PRINt, SCALe), if different from default values, have to be given immediately after the keyword line. Output files are named XrayAnnn.out where nnn = 001, 002, ... Examples: >ANGLE 50 20 ----------------------------------------------------------------------------- 5) POLArizedspectrum theta phi [ipos] POLArizedspectrum xx yy zz ipos (alternative format) ----------------------------------------------------------------------------- This keyword initializes the evaluation of an polarization-resolved Xray spectrum using data from a previous StoBe run. The spectrum refers to a given direction of the polarization vector of the incident/exciting light. The polarization direction is defined by two angles, theta, phi of a direction vector e with (angles in degrees) e = ( sin(theta)*cos(phi), sin(theta)*sin(phi), cos(theta) ) where theta, phi (in degrees) are correponding polar and azimuthal angles. The spectrum can be averaged over all angles theta for given phi or phi for given theta by using keyword 'AVG' instead of an explict numerical value for the angles, see example below. The alternative input uses a vector R = (xx, yy, zz) which is converted to angles theta, phi according to R = |R| ( sin(theta)*cos(phi), sin(theta)*sin(phi), cos(theta) ) . Note that with this input format angle averaging is not available. The spectral weight of each transition is given by (a) dipole transitions W(i) = f * E(i) * (Dx*ex + Dy*ey + Dz*ez)**2 where D(i) = (Dx, Dy, Dz) are dipole transition matrix elements of transition i, E(i) is the transition energy, and f (= 1000) is an empirical scaling factor. (b) quadrupole transitions W(i) = f * E(i) * [ Qxx*ex**2 + Qyy*ey**2 + Qzz*ez**2 + 2*Qxy*ex*ey + 2*Qxz*ex*ez + 2*Qyz*ey*ez) ]**2 where Q(i) = (Qxx, Qxy, Qxz, Qyy, Qyz, Qzz) are quadrupole transition matrix elements of transition i, E(i) is the transition energy, and f (= 1000) is an empirical scaling factor. Set no. ipos of the StoBe spectrum file is used for input (unit 1). If ipos is omitted, ipos = 1 is assumed. Note that for direction input in Cartesian coordinates xx, yy, zz the position index ipos MUST be given. Corresponding spectrum parameters (keywords RANGe, POINts, WIDTh, PRINt, SCALe), if different from default values, have to be given immediately after the keyword line. Output files are named XrayPnnn.out where nnn = 001, 002, ... Examples: >POLAR 70 10 >POLAR AVG 45 >POLAR 39 AVG ----------------------------------------------------------------------------- 6) MULTipole [type] [ratio] ----------------------------------------------------------------------------- Definition of multipole transitions to be included in the spectrum. If this keyword line is omitted a default transition type is used. If the keyword line appears before any spectrum task (see keywords TOTAlspectrum, ANGLeresolvedspectrum, POLArizedspectrum) the initial default transition type will be reset and applied to all spectrum tasks without an explicit MULTipole line. Otherwise, the multipole keyword applies to the spectrum (TOTAl, ANGLe, POLArized) defined before. Options are multipole = 'DIPL' refers to an Xray spectrum where only dipole transition are evaluated. This is the default if no multipole type is defined. = 'QUAD' refers to an Xray spectrum where only quadrupole transition are evaluated. = 'DIPQ' [fmix] refers to an Xray spectrum where both dipole and quadrupole transition are evaluated and corresponding intensities are superimposed. Parameter fmix defines the mixing of dipole and quadrupole intensities in the superposition I(tot) = I(dip) + fmix * I(quad) where fmix = 1.0 is the default. Example: >MULTIPOLE DIPQ 100. ----------------------------------------------------------------------------- 7) RANGe ebot etop ----------------------------------------------------------------------------- Range of energies [ebot, etop] (in eV) for which the spectrum of the present spectrum task is to be evaluated. If this keyword line is omitted a default range is used. If the keyword line appears before any spectrum task (see keywords TOTAlspectrum, ANGLeresolvedspectrum, POLArizedspectrum) the initial default values [500.0, 1000.0] will be reset to [ebot, etop] and applied to all spectrum tasks without an explicit RANGe line. Example: >RANGE 530 600 ----------------------------------------------------------------------------- 8) POINts npts ----------------------------------------------------------------------------- Number of energies within the range [ebot, etop] for which the spectrum of the present spectrum task is to be evaluated. The energy sequence is equidistant, given by e = ebot + (etop-ebot)/(npts-1)*(i-1), i = 1, npts If this keyword line is omitted a default value is used. If the keyword line appears before any spectrum task (see keywords TOTAlspectrum, ANGLeresolvedspectrum, POLArizedspectrum) the initial default value (= 1000) will be reset to npts and applied to all spectrum tasks without an explicit POINts line. Example: >POINTS 2000 ----------------------------------------------------------------------------- 9) WIDTh broad1 [ broad2 ewid1 ewid2 ] ----------------------------------------------------------------------------- Width (full-width-at-half-maximum, FWHM) (in eV) used for the gaussian broadening of the spectrum of the present spectrum task. The short version uses only a global broadening of width broad1. The full version uses an energy dependent broadening where a width broad1 is used for level energies < ewid1, width broad2 is used for level energies > ewid2 ( > ewid1), and for level energies between ewid1 and ewid2 a linear interpolation broad(E) = broad1 + (broad2-broad1)*(E-ewid1)/ewid2-ewid1) is applied. If this keyword line is omitted default values are used. If the keyword line appears before any spectrum task (see keywords TOTAlspectrum, ANGLeresolvedspectrum, POLArizedspectrum) the initial defaults (broad1 = broad2 = 1 eV, ewid1 = ewid2= 1.D30) will be reset and applied to all spectrum tasks without an explicit WIDTh line. Examples: >WIDTH .5 >WIDTH 0.2 2.5 544.09 554.09 ----------------------------------------------------------------------------- 10) SCALe scalfac ----------------------------------------------------------------------------- Definition of a factor used to scale the spectrum of the present spectrum task. All intensities will be enhanced/reduced by a factor scalfac. If this keyword line is omitted a default factor scalfac is used. If the keyword line appears before any spectrum task (see keywords TOTAlspectrum, ANGLeresolvedspectrum, POLArizedspectrum) the initial default (scalfac = 1.0) will be reset and applied to all spectrum tasks without an explicit SCALe line. Example: >SCALE 3.0 ----------------------------------------------------------------------------- 11) OFFSetenergy eoffst ----------------------------------------------------------------------------- Definition of an energy offset (in eV) used to shift the spectrum of the present spectrum task. All spectral energies will be shifted by adding the offset eoffst. If the keyword line appears before any spectrum task (see keywords TOTAlspectrum, ANGLeresolvedspectrum, POLArizedspectrum) the initial default (eoffst = 0.0) will be reset and applied to all spectrum tasks without an explicit OFFSetenergy line. Example: >OFFSET 1.54 ----------------------------------------------------------------------------- 12) PRINt [ON, OFF] ----------------------------------------------------------------------------- This keyword allows to include lists of the spectrum curves in the standard printout in addition to file output (option ON). If this keyword line is omitted a default setting is used (no spectrum listing in printout). If the keyword line appears before any spectrum task (see keywords TOTAlspectrum, ANGLeresolvedspectrum, POLArizedspectrum) the initial default (no print output, file output only) will be reset and applied to all spectrum tasks without an explicit PRINt line. Examples: >PRINT ON includes spectrum listing (same as line >PRINT) >PRINT OFF omits spectrum listing (but saves spectra to files) The following example starts from an H2O transition state calculation for an O1s hole and evaluates four spectra with 2000 points in the energy range (eV) [530, 600] for a variable (energy dependent) broadening: 1) a total Xray absorption spectrum, where the absorption is reduced to 50%, yielding spectrum file XrayT001.out. The spectrum is also printed. 2) a polarized Xray absorption spectrum with polarization angle theta = 35, phi = 10, yielding spectrum file XrayP002.out. The spectrum is not printed. 3) a polarized Xray absorption spectrum with polarization angle theta = 20, phi = 75, where the absorption is enhanced to 200%, yielding spectrum file XrayP003.out. The spectrum is also printed. 4) an angle-resolved Xray absorption spectrum with incident angle theta = 10, phi = 55, yielding spectrum file XrayA004.out. The spectrum is also printed. >TITLE >H2O O1s -> virt Xray absorption spectrum >RANGE 530 600 >POINTS 2000 >WIDTH 0.2 2.5 544.09 554.09 >PRINT >XRAY XAS >TOTAL >SCALE .5 >POLAR 35 10 >NOPRINT >POLAR 20 75 >SCALE 2 >ANGLE 10 55 >END The spectrum output files (XrayTnnn.out, XrayAnnn.out, XrayPnnn.out, nnn = 001, 002, ...) are ASCII type and of the general format Lines 1 ... N (F20.8,D20.8) E(I),SPEC(I), I=1, ... N E(I) energy inside the spectrum range (in eV) SPEC(I) intensity/absorption for energy E(I) where N denotes the number of energies requested for the spectrum. 2.7.7. TOTAL/ANGLE-RESOLVED INFRARED SPECTRA (IRSPEC) (Goto TOC, KEYW, KEYA) Utility irspec (executable irspec.x) allows you to use output files from StoBe runs (see e.11) to generate total and angle-resolved vibrational (infrared, IR) spectra for graphical output or for further processing. The spectra include gaussian broadening where the broadening width (FWHM) can be chosen freely. Angle-resolved spectra can be computed for any polarization direction or incident/exit beam given by polar angles theta, phi. Further, filters may be applied to extract spectra of specific types or mixtures of vibrational symmetry modes. File output of spectra consists of array listings combining each vibrational excitation energy with the corresponding intensity (given in a.u.). Each spectrum is saved in a separate file IrTnnn.out (for total spectra), IrAnnn.out (angle-resolved spectra), or IrPnnn.out (polarization- resolved spectra) where 'nnn' is a 3-digit number (padded with leading zeros id needed) counting the sequence of spectra defined by the input. For example, the second total spectrum file is named IrT002.out. The spectra listings can also be printed, see option PRINt below. The utility requires input (unit 1) from StoBe runs (StoBe output unit 65, created by option RUNType VIBTations) as well as ASCII input (unit 5) consisting of a series of keywords analogous to the original StoBe input and described in the following. The keyword input has to be finished by a line reading 'END'. ----------------------------------------------------------------------------- 1) TITLe ----------------------------------------------------------------------------- Brief descriptive title of the present irspec run on the line following the keyword. This title appears in the output file(s). Example: >TITLE >H2O vibrational (IR) spectrum ----------------------------------------------------------------------------- 2) TOTAlspectrum ----------------------------------------------------------------------------- This spectrum keyword initializes the evaluation of a total (angle-averaged and unpolarized) IR spectrum using data from a previous StoBe run. The spectral weight of each vibrational excitation is given by W(i) = 2/3*f * E(i) * (Dx**2 + Dy**2 + Dz**2) where D(i) = (Dx, Dy, Dz) are dynamical dipole moments of excitations i, E(i) is the excitation energy, and f (= 1.0) is an empirical scaling factor. The StoBe vibrational spectrum file (output unit 65) is used for input (as file unit 1). Corresponding spectrum parameters (keywords SYMFilter, RANGe, POINts, WIDTh, PRINt, SCALe, OFFSet), if different from default values, have to be given immediately after the keyword line. Output files are named IrTnnn.out where nnn = 001, 002, ... Example: >TOTAL ----------------------------------------------------------------------------- 3) ANGLeresolvedspectrum theta phi ANGLeresolvedspectrum xx yy zz (alternative format) ----------------------------------------------------------------------------- This keyword initializes the evaluation of an angle-resolved IR spectrum using data from a previous StoBe run. The spectrum refers to a given direction of the incident/exciting light beam where the absorption is an average over all possible polarization vectors perpendicular to the beam direction e given by e = ( sin(theta)*cos(phi), sin(theta)*sin(phi), cos(theta) ) . Here theta, phi (in degrees) are correponding polar and azimuthal angles. The alternative input uses a vector R = (xx, yy, zz) which is converted to angles theta, phi according to R = |R| ( sin(theta)*cos(phi), sin(theta)*sin(phi), cos(theta) ) . The spectral weight of each vibrational excitation is given by W(i) = f * E(i) * 1/2 [ (Dx**2 + Dy**2 + Dz**2) - (Dx*ex + Dy*ey + Dz*ez)**2 ] where D(i) = (Dx, Dy, Dz) are dynamical dipole moments of excitations i, E(i) is the excitation energy, and f (= 1.0) is an empirical scaling factor. The StoBe vibrational spectrum file (output unit 65) is used for input (as file unit 1). Corresponding spectrum parameters (keywords SYMFilter, RANGe, POINts, WIDTh, PRINt, SCALe, OFFSet), if different from default values, have to be given immediately after the keyword line. Output files are named IrAnnn.out where nnn = 001, 002, ... Examples: >ANGLE 50 20 >ANGLE 1. 1. 0. ----------------------------------------------------------------------------- 4) POLArizedspectrum theta phi POLArizedspectrum xx yy zz (alternative format) ----------------------------------------------------------------------------- This keyword initializes the evaluation of an polarization-resolved IR spectrum using data from a previous StoBe run. The spectrum refers to a given direction of the polarization vector of the incident/exciting light. The polarization direction is defined by two angles, theta, phi of a direction vector e with (angles in degrees) e = ( sin(theta)*cos(phi), sin(theta)*sin(phi), cos(theta) ) where theta, phi (in degrees) are correponding polar and azimuthal angles. The spectrum can be averaged over all angles theta for given phi or phi for given theta by using keyword 'AVG' instead of an explict numerical value for the angles, see example below. The alternative input uses a vector R = (xx, yy, zz) which is converted to angles theta, phi according to R = |R| ( sin(theta)*cos(phi), sin(theta)*sin(phi), cos(theta) ) . Note that with this input format angle averaging is not available. The spectral weight of each vibrational excitation is given by W(i) = f * E(i) * (Dx*ex + Dy*ey + Dz*ez)**2 where D(i) = (Dx, Dy, Dz) are dynamical dipole moments of excitations i, E(i) is the excitation energy, and f (= 1.0) is an empirical scaling factor. The StoBe vibrational spectrum file (output unit 65) is used for input (as file unit 1). Corresponding spectrum parameters (keywords SYMFilter, RANGe, POINts, WIDTh, PRINt, SCALe, OFFSet), if different from default values, have to be given immediately after the keyword line. Output files are named IrPnnn.out where nnn = 001, 002, ... Examples: >POLAR 70 10 >POLAR AVG 45 >POLAR 39 AVG >POLAR 3. 5. 4. ----------------------------------------------------------------------------- 5) SYMFilter ff1 ff2 ff3 ... ----------------------------------------------------------------------------- Filter of vibrational levels to be used for the present spectrum calculation. In the StoBe output file (see e.11) each vibrational level is characterized by a symmetry label SYLBL, e.g. nA1u (nA if the system has no symmetry), describing the its energetical sequence n and the irreducible representation of the corresponding symemtry group. These lables can be used to select specific sets of modes for which vibrationional spectra are to be evaluated. The sequence of ASCII labels ff1, ff2, ff3, ... allows to select vibrational levels (symmetry filter) for the present spectrum calculation as follows - All levels whose symmetry label SYLBL includes the character group given by ff1, or by ff2, etc. will be selected. Note that the character comparison is not case sensitive. - For ffi = 'NONE' no levels will be selected irrespective of the previous ffk k < i filters. However, subsequent filters ffk k > i will be evaluated. - For ffi = 'ALL' all levels will be selected irrespective of the previous ffk k < i filters. Subsequent filters ffk k > i will be ignored. If this keyword line is omitted the symmetry filter defined for the previous spectrum of the present run (or a default filter selecting all vibrational levels if the present spectrum is the first) will be applied. Examples: Assume CO2 with partial D4h symmetry: >SYMFILTER A (selects all A1g, A2g, A1u, A2u modes) >SYMFILTER eu (selects all Eu modes) >SYMFILTER a1g eu (selects all A1g and Eu modes) >SYMFILTER 1A1g (selects only the 1A1g mode) >SYMFILTER all (all modes: 1Eu, 2Eu, 1A1g, 1A2u) ----------------------------------------------------------------------------- 5) RANGe ebot etop ----------------------------------------------------------------------------- Range of energies [ebot, etop] (in cm-1) for which the present spectrum is to be evaluated. If this keyword line is omitted the range defined for the previous spectrum of the present run (or a default range [500.0, 1000.0] if the present spectrum is the first) will be used. Example: >RANGE 500 2500 ----------------------------------------------------------------------------- 6) POINts npts ----------------------------------------------------------------------------- Number of energies within the range [ebot, etop] for which the present spectrum is to be evaluated. The energy sequence is equidistant, given by e = ebot + (etop-ebot)/(npts-1)*(i-1), i = 1, npts If this keyword line is omitted the number of energies defined for the previous spectrum of the present run (or a default value, 1000, if the present spectrum is the first) will be used. Example: >POINTS 2000 ----------------------------------------------------------------------------- 7) WIDTh broad1 [ broad2 ewid1 ewid2 ] ----------------------------------------------------------------------------- Width (full-width-at-half-maximum, FWHM) (in eV) used for the gaussian broadening of the present spectrum to be evaluated. The short version uses only a global broadening of width broad1. The full version uses an energy dependent broadening where a width broad1 is used for level energies < ewid1, width broad2 is used for level energies > ewid2 ( > ewid1), and for level energies between ewid1 and ewid2 a linear interpolation broad(E) = broad1 + (broad2-broad1)*(E-ewid1)/ewid2-ewid1) is applied. If this keyword line is omitted the broadening parameters defined for the previous spectrum of the present run (or default values, broad1 = broad2 = 5 cm-1, ewid1 = ewid2= 1.D30, if the present spectrum is the first) will be used. Examples: >WIDTH 10.0 >WIDTH 5.0 10.0 1000.0 1200.0 ----------------------------------------------------------------------------- 8) SCALe scalfac ----------------------------------------------------------------------------- Definition of a factor used to scale the present spectrum to be evaluated. All intensities will be enhanced/reduced by a factor scalfac. If this keyword line is omitted the scaling factor defined for the previous spectrum of the present run (or a default value, 1.0 (no scaling), if the present spectrum is the first) will be used. Example: >SCALE 3.0 ----------------------------------------------------------------------------- 9) OFFSetenergy eoffst ----------------------------------------------------------------------------- Definition of an energy offset (in cm-1) used to shift the present spectrum to be evaluated. All spectral energies will be shifted by adding the offset eoffst. If this keyword line is omitted the offset defined for the previous spectrum of the present run (or a default value, 0.0 cm-1 (no offset), if the present spectrum is the first) will be used. Example: >OFFSET 100 ----------------------------------------------------------------------------- 10) PRINt [ON, OFF] ----------------------------------------------------------------------------- This keyword includes lists of the spectrum curves in the standard printout in addition to file output (option ON). If this keyword line is omitted the print setting defined for the previous spectrum of the present run (or the default setting (PRINT OFF), if the present spectrum is the first) will be used. Examples: >PRINT ON includes spectrum listing (same as line >PRINT) >PRINT OFF omits spectrum listing (but saves spectra to files) The following example starts from an StoBe calculation of vibrational modes of a linear CO2 molecule (with its symmetry axis along the z direction) using D4h symmetry. This yields four non-trivial symmetry modes, 1Eu, 2Eu, 1A1g, 1A2u in energetic order. In the input below vibrational (IR) spectra are requested for 1) a total IR absorption spectrum in the range between 500 and 2500 cm-1 with 2001 energies on an equidistant mesh, a broadening width (FWHM) of 10 cm-1, and no offset. All intensities are to be increased by 50% and the spectrum is to be printed. 2) a total IR absorption spectrum with parameters as in 1) but selecting only the two energetically highest modes 1A1g, 1A2u. 3) an angle-resolved IR absorption spectrum with parameters as in 1) with the beam direction along the molecular axis (z axis) and selecting only the two energetically lowest modes 1Eu, 2Eu. 4) a polarization-resolved IR absorption spectrum with parameters as in 1) with the photon polarization direction along the molecular axis (z axis) and selecting only the two energetically lowest modes 1Eu, 2Eu. >title >CO2, D4h symmetry, vibrational (IR) spectrum >total >range 500 2500 >points 2001 >width 10.0 >scale 1.5 >offset 0.0 >print on >total >symfilter a >angle 0 0 1 >symfilter e >pola 0 0 1 >end The spectrum output files (IrTnnn.out, IrAnnn.out, IrPnnn.out, nnn = 001, 002, ...) are ASCII type and of the general format Lines 1 ... N (F20.8,D20.8) E(I),SPEC(I), I=1, ... N E(I) energy inside the spectrum range (in cm-1) SPEC(I) intensity/absorption for energy E(I) where N denotes the number of energies requested for the spectrum. 2.7.8. SYMMETRY OPERATIONS (SYMFIND, SYMGEN) (Goto TOC, KEYW, KEYA) (A) Utility symfind (executable symfind.x) allows you to identify symmetry elements (inversion, mirror planes, rotation and roto-reflection axes) and point symmetry groups (axial and cubic groups only) of a given set of atoms forming a cluster or molecule. The calling sequence is symfind.x < infile > outfile where ASCII file infile lists the atom coordinates and element names using format Line 1 (I5) NATOM number of atoms to be listed in the following (F20.10) ACCU (optional) tolerance used to determine coinciding atom positions and symmetry elements. Default value is 1.D-5. NOTE that inside StoBe this tolerance is set to 1.D-3 (0.529177D-3) for coordinate input in atomic (Angstrom) units. Thus, ACCU has to be set accordingly to comply with StoBe coordinate input. Line(s) 2 (A4) ELMNT(I) atom label of center no. I. (3F20.10) X(I),Y(I),Z(I) cartesian coordinates of atom center no. I. where NATOM lines 2 appear in the file. The output lists - all atom coordinates indicating symmetry non-equivalent atoms by "#" - the symmetry center of the cluster/molecule - all symmetry elements and corresponding direction vectors - angles between mirror plane normals em(i) and rotation axis directions er(i) - the full point symmetry group name including C1, Ci, Cs, axial groups (C%v, D%h (% = infinity), Cn, Cnv, Cnh, Dn, Dnd, Dnh, Sn), cubic groups (O, Oh, T, Th, Td), and icosahedral groups (I, Ih) In addition, a StoBe format list of all non-equivalent atoms is saved on output unit 4. This file complies with the group definitions in the symmetry library of StoBe can, thus, be used as a basis to generate a StoBe input file. Example input file >4 1.D-6 >V 0.000000000 0.000000000 0.000000000 >V 1.000000000 1.000000000 0.000000000 >V 1.000000000 0.000000000 1.000000000 >V 0.000000000 1.000000000 1.000000000 resulting in a output listing > ATOM COORDINATES ( 4 atoms, 1 inequivalent (#) ) : > ------------------------------------------------------------- > Atom x y z ref > ------------------------------------------------------------- > 1) V 0.000000 0.000000 0.000000 1 # > 2) V 1.000000 1.000000 0.000000 1 > 3) V 1.000000 0.000000 1.000000 1 > 4) V 0.000000 1.000000 1.000000 1 > ------------------------------------------------------------- > > SYMMETRY CENTER : > Rsym= ( 0.500000, 0.500000, 0.500000 ) > > 3-DIM. MOLECULE, SYMMETRY OPERATIONS (accuracy = 1.0000D-06 ): > A) unit operation > B) no inversion > C) 6 mirror plane(s) : > 1) n= ( 0.707107, -0.707107, 0.000000 ), err.= 3.1402D-16 > 2) n= ( 0.707107, 0.707107, 0.000000 ), err.= 3.1402D-16 > 3) n= ( 0.707107, 0.000000, -0.707107 ), err.= 3.1402D-16 > 4) n= ( 0.707107, 0.000000, 0.707107 ), err.= 3.1402D-16 > 5) n= ( 0.000000, 0.707107, -0.707107 ), err.= 3.1402D-16 > 6) n= ( 0.000000, 0.707107, 0.707107 ), err.= 3.1402D-16 > D) 7 rotation axis(axes) : > 1) 3-fold, n= ( 0.577350, 0.577350, 0.577350 ), err.= 4.9651D-16 > 2) 3-fold, n= ( 0.577350, 0.577350, -0.577350 ), err.= 4.9651D-16 > 3) 3-fold, n= ( 0.577350, -0.577350, 0.577350 ), err.= 4.9651D-16 > 4) 3-fold, n= ( 0.577350, -0.577350, -0.577350 ), err.= 4.9651D-16 > 5) 2-fold, n= ( 1.000000, 0.000000, 0.000000 ), err.= 1.7554D-16 > 6) 2-fold, n= ( 0.000000, 1.000000, 0.000000 ), err.= 1.7554D-16 > 7) 2-fold, n= ( 0.000000, 0.000000, 1.000000 ), err.= 1.7554D-16 > E) 3 roto-refl. axis(axes) : > 1) 4-fold, n= ( 0.000000, 0.000000, 1.000000 ), err.= 2.2204D-16 > 2) 4-fold, n= ( 0.000000, 1.000000, 0.000000 ), err.= 2.2204D-16 > 3) 4-fold, n= ( 1.000000, 0.000000, 0.000000 ), err.= 2.2204D-16 > > ANGLES BETWEEN SYMMETRY AXES > Angles em(i)/em(k), i,k= 1 - 6 : > 1 0.0000 > 2 90.0000 0.0000 > 3 60.0000 60.0000 0.0000 > 4 60.0000 60.0000 90.0000 0.0000 > 5 60.0000 60.0000 60.0000 60.0000 0.0000 > 6 60.0000 60.0000 60.0000 60.0000 90.0000 0.0000 > Angles er(i)/er(k), i,k= 1 - 7 : > 1 0.0000 > 2 70.5288 0.0000 > 3 70.5288 70.5288 0.0000 > 4 70.5288 70.5288 70.5288 0.0000 > 5 54.7356 54.7356 54.7356 54.7356 0.0000 > 6 54.7356 54.7356 54.7356 54.7356 90.0000 0.0000 > 7 54.7356 54.7356 54.7356 54.7356 90.0000 90.0000 0.0000 > Angles em(i)/er(k), i=1 - 6, k= 1 - 7 : > 1 90.0000 90.0000 35.2644 35.2644 45.0000 45.0000 90.0000 > 2 35.2644 35.2644 90.0000 90.0000 45.0000 45.0000 90.0000 > 3 90.0000 35.2644 90.0000 35.2644 45.0000 90.0000 45.0000 > 4 35.2644 90.0000 35.2644 90.0000 45.0000 90.0000 45.0000 > 5 90.0000 35.2644 35.2644 90.0000 90.0000 45.0000 45.0000 > 6 35.2644 90.0000 90.0000 35.2644 90.0000 45.0000 45.0000 > > Total of 17 symmetry operations, order = 24 > > Symmetry group = Td > > SYMMETRY-ADAPTED COORDINATE SYSTEM > Origin at Rsym= ( 0.500000, 0.500000, 0.500000 ) > Unit vectors: > e1= ( 1.000000, 0.000000, 0.000000 ) > e2= ( 0.000000, 1.000000, 0.000000 ) > e3= ( 0.000000, 0.000000, 1.000000 ) > > Symmetry-adapted atom coordinates > 4 atoms, 1 non-equivalent (#) : > ------------------------------------------------------------- > Atom x y z ref > ------------------------------------------------------------- > 1) V -0.500000 -0.500000 -0.500000 1 # > 2) V 0.500000 0.500000 -0.500000 1 > 3) V 0.500000 -0.500000 0.500000 1 > 4) V -0.500000 0.500000 0.500000 1 > ------------------------------------------------------------- and an output file fort.4 reading > SYMM TD > V -0.5000000000 -0.5000000000 -0.5000000000 zz ngrid > END (B) Utility symgen (executable symgen.x) allows you to take a set of atoms and generate all symmetry equivalents according to a given point symmetry group completing a cluster or molecule. The symmetrization procedure is also used inside StoBe to complete a molecule with symmetry. At present, the following 47 point symmetry groups are supported (see also file symbasis): C1, CI, CS, CSXZ, CSY, CSYZ, CSX, CSXY, CSZ, S4, S6, C2VB, Cn, CnV, CnH, Dn, DnH, DnD, n=2-6, T, TH, TD, O, OH This utility requires the symmetry library file to be present as unit 4 in addition to ASCII file input. The calling sequence is symgen.x < infile > outfile where ASCII file infile lists the atom coordinates and element names using format Line 1 (A4) SYMLAB label of the symmetry group as given above. Lines 2 (A4) ELMNT(I) atom label of center no. I. (3F20.10) X(I),Y(I),Z(I) cartesian coordinates of atom center no. I. where the atom list is terminated with end-of-file. The output lists the full set of symmetrized atom coordinates including all atoms where symmetry non-equivalent atoms (given in the input) are labelled by "#". Example input file >C3V >N 0.000000000 0.000000000 0.000000000 >H 1.801764605 0.000000000 0.705656897 resulting in output > Spatial symmetry = C3V > > INPUT ATOM COORDINATES ( 2 non-equivalent atoms ) : > ----------------------------------------------------- > Atom x y z > ----------------------------------------------------- > 1) N 0.000000 0.000000 0.000000 > 2) H 1.801765 0.000000 0.705657 > ----------------------------------------------------- > SYMMETRIZED ATOM COORDINATES ( 4 atoms ) : > ------------------------------------------------------------- > Atom x y z ref > ------------------------------------------------------------- > 1) N 0.000000 0.000000 0.000000 1 # > 2) H 1.801765 0.000000 0.705657 2 # > 3) H -0.900882 1.560374 0.705657 2 > 4) H -0.900882 -1.560374 0.705657 2 > ------------------------------------------------------------- 2.7.9. COMBINE RESTART FILES (COMBINE) (Goto TOC, KEYW, KEYA) This utility allows you to combine restart files of two molecules to yield a restart file of the compound system. This combined restart file can provide a very good starting guess for the compound system, in particular, if the interaction between the components is weak and/or if the second component is only a small addition to the first. Examples are small molecules attached to large substrate clusters (e. g. Cu49 + CO --> Cu49CO). The restart file of the compound system is of short format, see option SHRT, f.12) and Sec. 2.5, and is truncated (including only recs. 1-19, see Sec. 2.5). It can be used as input for single point SCF runs of the compound system as well as for geometry optimizations (with keyword line RUNT OPT NEWG, see c.1)) The utility requires restart files (StoBe output unit 2) of the two component systems as input (file units 1/2 for the first/second components) while the restart file of the compound system is saved on file unit 3. Additional ASCII input (unit 5) consists of a series of keywords analogous to the original StoBe input and described in the following. The keyword input has to be finished by a line reading 'END'. ----------------------------------------------------------------------------- 1) TITLe ----------------------------------------------------------------------------- Brief descriptive title of the compound system on the line following the keyword. This title appears in the output file(s). Example: >TITLE >NH3CO: combination of NH3 with CO ----------------------------------------------------------------------------- 2) AUXBasissize nbchs1 nbxcs1 nbchs2 nbxcs2 ----------------------------------------------------------------------------- Number of aux. s basis functions (separate s functions only, of all atoms including symmetry equivalents) for charge density and XC potential fits referring to component 1 (nbchs1, nbxcs1) and component 2 (nbchs2, nbxcs2). These numbers are included with the latest StoBe restart file format (rec. 3, see Sec. 2.5.) and need not be given explicitly. They are, however, required if truncated or old format restart files are to be combined. Correct values of nbchs, nbxcs can be determined from the aux. basis set titles (given by (ns(CD), ns(XC)) in the StoBe output listing, see also Sec. 2.3 Example: >AUXBASIS 16 16 8 8 ----------------------------------------------------------------------------- 3) ORBBasissize nnns1 nnnp1 nnnd1 nnns2 nnnp2 nnnd2 ----------------------------------------------------------------------------- Number of orbital basis functions (s, p, d components) referring to component 1 (nnns1, nnnp1, nnnd1) and component 2 (nnns2, nnnp2, nnnd2). These numbers are included with the latest StoBe restart file format (rec. 25, see Sec. 2.5.) and need not be given explicitly. They are, however, required if truncated restart files are to be combined. The StoBe output listing of each molecular component shows correct values for nnns, nnnp, nnnd. Example: >ORBBASIS 9 2 1 6 4 2 ----------------------------------------------------------------------------- 4) SHIFt xs ys zs ----------------------------------------------------------------------------- Cartesian components of vector R = (xs, ys, zs) used to shift the origin of the second molecular component with respect to that of the first. If (identical) symmetry is applied to both molecular components the vector shift must not destroy the symmetry of the compound system. For example, if both components are of C3v symmetry then only shift vectors (0, 0, z) are allowed. Example: >SHIFT 0.0 0.0 -10.0 NOTE that this utility is still experimental and is not guaranteed to work in all cases. - At present, the restart file of the first molecular component must refer to a singlet state. - Combinations of components with model potentials have not been testet so far. - If the molecular components allow for symmetry their symmetry group must be identical and the vector shift must not destroy the symmetry of the compound system. 2.7.10. EXPAND MOLECULAR ORBITALS (EXPAND) (Goto TOC, KEYW, KEYA) This utility allows you to expand cluster orbitals of one configuration (contained in restart file of unit #2) by those of another (restart file of unit #1) of the same system. Examples are excited configurations where the orbitals are expanded by ground state orbitals to find out about excitation induced hybridization. Another example is a compound cluster AB where orbitals of the separate subunits A, B (after being combined with utility combine, see Sec. 2.7.8) are projected onto the orbitals of AB. The utility requires restart files (StoBe output unit #2) of the configuration whose orbitals are to be expanded (input file file2, unit #2) and of the restart file containing the expansion orbitals (input file file1, unit #1). The calling sequence is expand.x > expand.out where the ASCII file expand.out lists all orbitals of file1 and file2 by their LCAO basis components as well as those of file2 by their expansion by orbitals of file1. 2.7.11. REMOVE SYMMETRY IN RESTART FILE (DESYMM) (Goto TOC, KEYW, KEYA) This utility allows you to reduce the symmetry group used in the restart file (orbital classification, etc.) of a given system to C1 (no symmetry). The removal of symmetry can be necessary in calculations of vibrational excitations, see c.1), e.11), of the polarizability, see c.1), e.6), or in geometry optimizations where optimization constraints are relaxed. Only binary restart files are accepted for input and the output files will be binary. The calling sequence is desymm.x symmfile nosymmfile where "symmfile" denotes a complete binary restart file describing the molecule/cluster using symmetry group X, see 2.4.). This file will be used as input to generate a new restart file "nosymmfile" describing the system without using symmetry (the symmetry group will be set to C1) where all electronic parameters such as multiplicity, occupations etc. will be maintained. The restart format of file nosymmfile will be short, see f.12). Note that the transformation affects only records no. 2, 4 - 9, 11, 12, 17 - 19 of the restart file, see section 2.5. 2.7.12. GENERATE POINT CHARGE SETS (LATSPH, PCFILT) (Goto TOC, KEYW, KEYA) Utility LATSPH calculates coordinates of all atom centers in a 3-dim. lattice (defined by lattice vectors and lattice basis vectors with atom charges) which lie inside a sphere of given center and radius. In addition, a planar constraint may be imposed to include only those centers which lie below a plane of given normal direction and distance from the sphere center. This utility can be used to generate a point charge environment about a cluster based on a crystal lattice in studies where surface or bulk clusters with a point charge embedding are considered, see also option c.14). Lattice centers which coincide with cluster centers can be identified and removed from the list of lattice centers with utility PCFILT, see below. All calculated centers (ordered according to distance from sphere origin) are output on unit unit # 1 in a format ready to be used as an external file for point charge input to StoBe, see keywords 'FILE' and 'FIL&' of option c.14). Here cordinates are given in absolute units of the lattice. unit # 2 in a format ready to be used as an external file for point charge input to StoBe, see keywords 'FILE' and 'FIL&' of option c.14). Here cordinates are given with respect to the sphere center. unit # 3 in PLOT3D type format, see BALSAC manual at http://www.fhi-berlin.mpg.de/KHsoftware/Balsac/index.html. unit # 4 XYZ type format. The calling sequence is latsph.x < inputfile where "inputfile" denotes a valid file describing the lattice input and control parameters defined by 6 lines given as follows: Line 1. (A80) TITLE TITLE := Title of the LATSPH run. Lines 2. (3(3F15.9/)) ((R(I,J),I=1,3),J=1,3) R(I,J) := Lattice vectors (Ith component if Jth vector in cartesian coordinates). Line 3. (I5) NBG NBG := The absolute value |NBG| defines the number of lattice basis vectors while positive and negative NBG values distinguish between absolute and relative vector coordinates, see lines 4. For primitive lattices set NBG = 1. Line(s) 4. (A4,4F15.9) (ELNM(J),(RBG(I,J),I=1,3),CHG(J),J=1,|NBG|) ELNM(J) := Element name of center J. RBG(I,J) := Lattice basis vector Rj of center J. NBG > 0 : lattice basis vectors are given in absolute cartesian coordinates where RBG(I,J) denotes the Ith component of Jth lattice basis vector, J=1,NBG. For primitive lattices set RBG(I,1) = 0.0. NBG < 0 : lattice basis vectors are given by linear combinations of lattice vectors where RBG(I,J) denotes the weight of the Ith lattice vector contributing to the Jth lattice basis vector, J = 1,NBG. CHG(J) := Atom (element) charge of atoms assigned to center J. Line 5. (I5,5F15.9) ICENT,(CENT(J),J=1,3),RADMX,EPS,IPLAN ICENT := Index of atom center J = ICENT to act as center of the spherical crystal section. CENT(J) := Absolute cartesian coordinates C = (X1, X2, X3) of the center of the spherical crystal section. ICENT = 0 : coordinates CENT(J) are used for the section center. 0 < ICENT <= |NBG| : coordinates CENT(J) are ignored and replaced by those of center ICENT (given above) defining the section center. RADMX := Radius of the center of the spherical crystal section. EPS := Accuracy to define atom shells (determined internally) about the sphere center. IPLAN := Flag to set additional planar constraint, see line 6. = 0 no additional planar constraint. Without planar constraint all atom centers enclosed by the sphere about center CENT(J) and of radius RADMX are calculated and saved on an output file. > 0 additional planar constraint defined by line 6. Line 6 is required only for IPLAN > 0, see above. Line 6. (4F15.9) (PLAN(J),J=1,3),DPLAN PLAN(J) := Absolute cartesian coordinates P = (X(1), X(2), X(3)) of the plane normal vector used for lattice building with a planar constraint, see below. Vector P does not need to be normalized. DPLAN := Pseudo-distance D of plane from sphere center to confine lattice building, see below. With a planar constraint only those atom centers of the sphere are calculated which are below the plane of plane normal PLAN(J) at a distance |DPLAN| of the plane from the sphere center. The mathematical condition for an atom center to be included is (Rj - C)(P/|P|) <= D where Rj is the position vector of atom center j, C is the position vector of the sphere center, P is the plane normal vector, and D denotes the pseudo-distance. Note that D can assume both positive and negative values. As an example we list below an input file for the quasi-rhombohedral (0001) oriented V2O3 crystal ================= input starts ============================================== V2O3 quasi-rhombohedral (0001) oriented, Angstrom units 2.852449000 0.000000000 4.655656000 -1.426225000 2.470209400 4.655656000 -1.426225000 -2.470209400 4.655656000 -10 V .346300000 .346300000 .346300000 3.000000000 V .653700000 .653700000 .653700000 3.000000000 V .846300000 .846300000 .846300000 3.000000000 V .153700000 .153700000 .153700000 3.000000000 O .565000000 .935000000 .250000000 -2.000000000 O .935000000 .250000000 .565000000 -2.000000000 O .250000000 .565000000 .935000000 -2.000000000 O .435000000 .065000000 .750000000 -2.000000000 O .065000000 .750000000 .435000000 -2.000000000 O .750000000 .435000000 .065000000 -2.000000000 1 .000000000 .000000000 .000000000 11.000000000 .01 1 .000000000 .000000000 1.000000000 .800000000 ================= input ends ================================================ In many cases the list of lattice centers has to be reduced by removing atom centers which appear in a cluster (whose electronic structure is evaluated with StoBe) and coincide with lattice centers. This is possible using utility PCFILT. This utility reads a list of lattice centers generated by LATSPH as well as a list of cluster atom centers. By comparing the two lists coinciding centers (defined by inter-atomic distances < 1.D-8) are removed and a reduced list of lattice centers is produced. The following input/output files are required unit # 1 (input) list of lattice centers (cartesian coordinates, charges) generated by latsph (latsph output units # 1, 2) or using the corresponding ASCII file format given by Line 1. (I10,A) NPTC, TITLEL NPTC := Number of lattice centers. TITLEL := (optional title of the lattice center list. Lines 2. (4F15.9) X(I), Y(I), Z(I), Q(I), I = 1, NPTC X(I),Y(I), := Cartesian coordinates of lattice center I. Z(I) Q(I) := Atom charge of lattice center I. unit # 2 (input) list of cluster centers (cartesian coordinates, charges) in Plot3D format (see balsac manual at http://www.fhi-berlin.mpg.de/KHsoftware/Balsac/index.html) or using the corresponding ASCII file format given by Line 1. (A) TITLEC TITLEC := Title of the cluster (can be an empty line). Line 2. (I5) NKTOT NKTOT := Number of cluster centers. Lines 3. (4F15.9) (X(I), Y(I), Z(I), R(I), Q(I)), I = 1, NKTOT X(I),Y(I), := Cartesian coordinates of cluster center I. Z(I) RAD(I) := Atom radius of cluster center I (can be zero). Q(I) := Atom charge of cluster center I. unit # 3 (output) list of lattice centers (cartesian coordinates, charges) of the input list on unit # 1 (using the same ASCII format) reduced by those coinciding with cluster centers of unit # 2. The removed cluster centers are added at the end of the list on unit # 3. If not all cluster centers coincided with lattice centers a warning will be issued. The calling sequence is pcfilt.x where input files on units # 1, 2 have to be provided with the call. 2.7.13. LINEAR PATHS BETWEEN IMAGES (LINPATH) (Goto TOC, KEYW, KEYA) Utility LINPATH calculates coordinates of all atom centers in a cluster along a linear path connecting a set of Nini initial geometric structures (initial images, up to 100) and generates a final set of Nfin < 1000 images. These images refer to a linear interpolation between initial images on an equidistant distance along the path where a 3n-dimensional Cartesian length (n denotes the number of cluster atoms) is used as a measure. Atom coordinates of the initial images are provided by single- or multi-image StoBe files, see also a.1), and the final interpolation images can be saved both as single- or multi-image StoBe files for subsequent StoBe runs. The interpolation files can be used for multi-image SCF calculations or as a first step of a nudged-elastic-band (NEB) optimization. The algorithm is used also inside StoBe if a NEB optimization is started from scratch, see c.1) and e.12). Utility LINPATH can also be used to generate a set of Nfin image files by linear interpolation from a set of Nini initial image files where both input and output files are of Balsac or XYZ format. The utility requires either a set of Nini StoBe restart files (single-image type) or a multi-image restart file, or Nini Balsac or XYZ format files describing all Nini images. The output file format is always identical to that of the input files. Additional ASCII input (unit 5) consists of a series of keywords analogous to the original StoBe input and described in the following. The keyword input has to be finished by a line reading 'END'. ----------------------------------------------------------------------------- 1) NINPutimages Nini ----------------------------------------------------------------------------- Definition of the number Nini < 1000 of initial images (StoBe single-image, Balsac, or XYZ format) used for the linear interpolation. If a StoBe multi-image format file is provided as input, Nini must agree with the number of images of that file. Example: >NINPUTIMAGES 3 (3 images provided by input file(s)) ----------------------------------------------------------------------------- 2) NPATh Nfin ----------------------------------------------------------------------------- Definition of the number Nfin of final images (StoBe single- or multi-image, Balsac, or XYZ format) calculated by linear interpolation. The default value of Nfin is set equal to Nini. Thus final and initial images agree. Example: >NPATH 7 (7 images calculated and saved in output file(s)) ----------------------------------------------------------------------------- 3) OFFSet Noffi Noffo ----------------------------------------------------------------------------- Noffi Definition of the offset used for file unit numbers pointing to the initial StoBe restart, Balsac, or XYZ format file(s). For StoBe single-image, Balsac, or XYZ format input with Nini images, units Noffi, Noffi+1, ... Noffi+Nini-1 have to be assigned to existing StoBe restart, Balsac, or XYZ files. For StoBe multi-image input unit Noffi has to be assigned to one StoBe restart file containing all Nini images. Noffo Definition of the offset used for file unit numbers pointing to final StoBe restart, Balsac, or XYZ format file(s). For StoBe single-image, Balsac, or XYZ format output with Nfin images, units Noffo, Noffo+1, ... Noffo+Nfin-1 will be used for file output. For StoBe multi-image output, only unit Noffo will be used for file output where the StoBe restart file contains all Nfin images. For named file output, see keyword FNAMe below, only unit Noffo will be used. Values of Noffi, Noffo must be larger than 10 where default values are Noffi = 10, Noffo = 100. Unit numbers starting with Noffi and those starting with Noffo must not overlap. Example: >OFFSET 10 20 ----------------------------------------------------------------------------- 4) MULTi-image ----------------------------------------------------------------------------- This keyword defines StoBe multi-image output of the Nfin interpolated images in one Stobe restart file (unit Noffo). If the keyword is missing (default setting) the output refers to Nfin different Stobe restart files (units Noffo, Noffo+1, ... Noffo+Nfin-1). This keyword is ignored for Balsac and XYZ format file output. Example: >MULTI-IMAGE ----------------------------------------------------------------------------- 5) FTYPe [StoBe, Balsac, XYZ] ----------------------------------------------------------------------------- Utility LINPATH can calculate sequences of images by linear interpolation where input and output refer to (multi- and single-image) StoBe, (single-image) Balsac [H00], and XYZ format files. Balsac output files are used as input to BALSAC, a visualization and graphical analysis software [H00]. XYZ format output files are accepted as input to other graphics software. The present keyword FTYPe defines the corresponding format. The default format (omitting keyword FTYPe) refers to StoBe restart file handling. NOTE that input and output files refer always the the same file format. Examples: >FTYPE STOBE (default if keyword is omitted) >ftype Balsac >FTYP XYZ ----------------------------------------------------------------------------- 5) FNAMe Ftemplate ----------------------------------------------------------------------------- Named file output is an an alternative way to define output files. This is very useful for handling large numbers of interpolation files where file output on different file units becomes impractical. This output option uses only output unit Noffo together with explicit file names generated internally and based on a text template Ftemplate. The output of Nfin interpolation files will be saved on files Ftemplate_001.xxx, Ftemplate_002.xxx, ..., Ftemplate_{Nfin}.xxx where the numbers 1, 2, ..., Nfin are padded to 3 digits and xxx gives the file format, xxx = stb for (single-image) StoBe format files, = plt for Balsac format files, = xyz for XYZ format files. StoBe multi-image output with the named file option saves the interpolation data on file Ftemplate_000.stb . Example: >FTYPE StoBe >FNAME KHout (saves files KHout_001.stb, KHout_002.stb, KHout_003.stb, ... ) ----------------------------------------------------------------------------- 6) PRINt [NONE, INPUT, OUTPUT, ALL] ----------------------------------------------------------------------------- This keyword determines whether and which coordinates are included in the output listing. Option NONE does not list any coordinates but includes headers of each in- and output structure file. INPUT lists all coordinates of the input structure files but only headers of the output structure files. OUTPUT lists only headers of the input structure files but all coordinates of the output structure files. ALL (default setting) lists coordinates of all in- and output structure files. 2.8. PROGRAM STRUCTURE AND DIMENSIONS (Goto TOC, KEYW, KEYA) The main driver of StoBe is the CRAYCP routine from which all main calls are performed. The input is processed in DECODE where also some additional comments on input options etc may be found. The SCF procedure is controlled by SCFRUN which in turn calls SCFDOO for the actual construction and diagonalization of the K-S matrices. The numerical grid work is handled by DENSTY with allocations done in the control routine CPGRSCF. Geometry optimization is controlled by DYNNRM. All relevant dimensions are controlled by parameter statements and separated from the code in two include files: workvec.h and param.h workvec.h : gives the size, NDIM, of the requested work area A(NDIM). This is used for all dynamical allocations in the program. Two suggested sizes are included (small and big). Only included in CRAYCP. param.h : gives sizes for a number of different arrays and vectors stored in common blocks throughout the program. Two suggested sizes are included (small and big). The program will check that the given size is sufficient and will inform which parameter needs to be increased if necessary. If param.h is modified the entire program must be recompiled. 2.9. COMPILATION AND ERROR HANDLING (Goto TOC, KEYW, KEYA) The executables (StoBe.x and the different utility programs) are generated by running the Makefile located in the appropriate directories (Source, Utilities). Compilation commands for a number of systems are included, but not guaranteed to work (systems tested include DEC Alpha and IBM RS, SUN, SGI). Below are a few hints concerning the compilation on Compaq Alpha workstations running True64unix. The subroutine tag.f causes errors when the compiler optimization level O3 is used. Compile separately using >f77 -r8 -O1 -c tag.f before the overall make (as included in the Makefile). The executable that is generated tends to be quite big and requires substantial (virtual) memory. For interactive runs use the command >unlimit to increase the allowed heap. Otherwise an error similar to that below can occur: 29625:../Source/StoBe.x: /sbin/loader: Fatal Error: set_program_attributes failed to set heap start 3. BIBLIOGRAPHY (Goto TOC, KEYW, KEYA) The following bibliography is rather incomplete and is given in alphabetical order. [A92] A. St-Amant, Ph.D. Thesis Universit'e de Montreal, 1992. [B86] A. D. Becke, J. Chem. Phys. 84 (1986) 4524. [B86a] A. D. Becke, J. Chem. Phys. 85 (1986) 7184. [B95] R. F. W. Bader, "ATOMS IN MOLECULES, A Quantum Theory", Intern. Series of Monographs on Chemistry, Vol. 22, Clarendon Press, Oxford 1990. [BE88] A.D. Becke, Phys. Rev. A 38 (1988) 3098. [BH72] U. von Barth and L. Hedin, J. Phys. C 5 (1972) 1629. [BHB84a] P.S. Bagus, K. Hermann, and C.W. Bauschlicher, J. Chem. Phys. 80 (1984) 4378. [BHB84b] P.S. Bagus, K. Hermann, and C.W. Bauschlicher, J. Chem. Phys. 81 (1984) 1966. [D30] P. A. M. Dirac, Proc. Cambridge Phil. Soc. 26 (1930) 376. [D94] C. Daul, Int. J. Quant. Chem. 52 (1994) 867. [DD82] P. Fuentealba, H. Preuss, H. Stoll, L. v. Szentpaly, Chem. Phys. Lett. 89 (1982) 418. [DD87] M. Dolg, U. Wedig, H. Stoll, H. Preuss, J. Chem. Phys. 86 (1987) 866. [DD89] M. Dolg, H. Stoll, A. Savin, H. Preuss, Theor. Chim. Acta 75 (1989) 173. [DD90] D. Andrae, U. Haeussermann, M. Dolg, H. Stoll, H. Preuss, Theor. Chim. Acta 77 (1990) 123. [DD91a] M. Kaupp, P. v. R. Schleyer, H. Stoll, H. Preuss, J. Chem. Phys. 94 (1991) 1360. [DD91b] M. Dolg, M. Fischer, H. Preuss, A. Alex, T. Clark, J. Chem. Phys. 94 (1991) 3011. [DD91c] W. Kuechle, M. Dolg, H. Stoll, H. Preuss, Mol. Phys. 74 (1991) 1245; J. Chem. Phys. 94 (1991) 3011. [DD93] A. Bergner, M. Dolg, W. Kuechle, H. Stoll, H. Preuss, Mol. Phys. 80 (1993) 1431. [DD95] A. Nicklass, M. Dolg, H. Stoll, H. Preuss, J. Chem. Phys. (1995) ??? [DD95] X. Cao, M. Dolg, H. Stoll, J. Chem. Phys. 118 (2003) 487. [DGS93] C. Daul, A. Goursot, and D.R. Salahub in "NUMERICAL GRID METHODS AND THEIR APPLICATION TO SCHRODINGER'S EQUATION", C. Cerjan, Ed., NATO ASI series, vol 412, p. 153, 1993. "Numerical Calculation of Multicentre Integrals for Polyatomic Molecules". [FBH00] A. Foehlisch, P. Bennich, J. Hasselstroem, O. Karis, A. Nilsson, L. Triguero, M. Nyberg and L.G.M Pettersson, Phys. Rev. B 61 (2000) 16229. [Fxx] R. Fournier, private communication. [GDZ86] J. N. Gex, C. Daul, A. von Zelewsky, Chem. Phys. Lett. 132 (1986) 276. [GL76] O. Gunnarsson and x. Lundqvist, Phys. Rev. B 13 (1976) 4274. [GSA92] N. Godbout, D.R. Salahub, J. Andzelm and E. Wimmer, Can. J. Chem. v.70, p.560 (1992). [H00] BALSAC (Build and Analyze Lattices, Surfaces, and Clusters), software package by K. Hermann, distributed by Garching Innovation GmbH, Munich (Germany), see also http://www.fhi-berlin.mpg.de/KHsoftware/Balsac/index.html. [H03] T. Heine, 2003, private communication. [HHN99] B. Hammer, L. B. Hansen, and J. K. Norskov, Phys. Rev. B 59 (1999) 7413. [HL71] L. Hedin and B. Lundqvist, J. Phys. C 2064 (1971). ??? [HUJ00] G. Henkelman, B. P. Uberuaga, and H. Jonsson, J. Chem. Phys. 113 (2000) 9901. [JMJ98] H. Jonsson, G. Mills, and K. W. Jacobson, “Nudged elastic band method for finding minimum energy paths and transitions,” in Classical and Quantum Dynamics in Condensed Phase Simulations, edited by B. J. Berne, G. Ciccotti, and D. F. Coker (World Scientific Publishing Company, 1998). [KH01] A. Koester, T. Heine, 2001, unpublished. [KLS96] A. M. Koester, M. Leboeuf and D. R. Salahub in "MOLECULAR ELECTROSTATIC POTENTIALS: CONCEPTS AND APPLICATIONS", Ed. J. S. Murray and K. Sen, Theor. and Comp. Chem., Vol. 3, (Elsevier, Amsterdam, 1996), p. 105. [KPP01] C. Kolczewski, R. Puettner, O. Plashkevych, H. Agren, V. Staemmler, M. Martins, G. Snell, M. Sant'anna, G. Kaindl and L.G.M. Pettersson, J. Chem. Phys. 115 (2001) 6426. [L75] V. I. Lebedev, Zh. Vychisl. Mat. Fiz. 15 (1975) 48. [MJ94] G. Mills and H. Jonsson, Phys. Rev. Lett. 72 (1994) 1124. [MJG95] G. Mills, H. Jonsson, and G. K. Schenter, Surf. Sci. 324 (1995) 305. [MOD03] The interactive graphics package Molden can be downloaded from the web at http://www.cmbi.kun.nl/~schaft/molden/molden.html. [MOL03] The interactive graphics package Molekel can be downloaded from the web at http://www.cscs.ch/molekel. [P82] P. Pulay, J. Comp. Chem. 3 (1982) 556. [P86] J.P. Perdew, Phys. Rev. B 33 (1986) 8822. [P91] J. P. Perdew, in "Electronic Structure of Solids '91", edited by P. Ziesche and H. Eschrig (Akademie Verlag, 1991), p. 11; J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson, D. J. Singh, and C. Fiolhais, Phys. Rev. B 46 (1992) 6671. [PAK00] O. Plashkevych, H. Agren, L. Karlsson and L.G.M. Pettersson, J. Electron Spec. Rel. Phen. 106 (2000) 51. [PBE96] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett 77 (1996) 3865. [PRV95] E.I. Proynov, E.Ruiz, A. Vela, and D.R. Salahub, Int. J. Quant. Chem. Symp. 29 (1995) 61. [PVS94] E.I. Proynov, A. Vela, and D.R. Salahub, Phys. Rev. A 50 (1994) 3766. [PW86] J.P. Perdew and Y. Wang, Phys. Rev. B 33 (1986) 8800. [PW91] J. P. Perdew and Y. Wang, Phys. Rev. B 45, 13244 (1992). [PWG87] L.G.M. Pettersson, U. Wahlgren and O. Gropen, J. Chem. Phys. 86 (1987) 2176. [PY84] R. G. Parr and W. Yang, J. Am. Chem. Soc. 106 (1984) 4049. [PY89] R. G. Parr and W. Yang, "DENSITY FUNCTIONAL THEORY OF ATOMS AND MOLECULES", Oxford University Press, New York 1989. [S95] A. Savin, in RECENT ADVANCES IN DENSITY FUNCTIONAL METHODS, PART I, D.P. Chong, Ed. (World Scientific, Singapore, 1995), p. 129. "Beyond the Kohn-Sham Determinant". [SCF+95] D.R. Salahub, M.E. Castro, R. Fournier, P. Calaminici, N. Godbout, A. Goursot, C. Jamorski, H. Kobayashi, A. Martinez, I. Papai, E. Proynov, N. Russo, S. Sirois, J. Ushio, and A. Vela in "THEORETICAL AND COMPUTATIONAL APPROACHES TO INTERFACE PHENOMENA", H. Sellers and J.T. Golab, Eds. (Plenum Press, New York, 1995), p.187. [SCP93] D. R. Salahub, M. E. Castro and E. I. Proynov in "RELATIVISTIC AND ELECTRON CORRELATION EFFECTS IN MOLECULES AND SOLIDS", vol.318 of NATO ASI Series B: Physics, ed. G. L. Malli (Plenum Press, New York, 1994), p.411. "Principles and Applications of DFT". [TLP99] L. Triguero, Y. Luo, L.G.M. Pettersson, H. Agren, P. Vaeterlein, M. Weinelt, A. Foehlisch, J. Hasselstroem, O. Karis and A. Nilsson, Phys. Rev. B 59 (1999) 5189. [TPA98] L. Triguero, L.G.M. Pettersson and H. Agren, Phys. Rev. B 58 (1998) 8097; J. Phys. Chem. 102 (1998) 10599. [VWN80] S.H. Vosko, L. Wilk, and M. Nusair, Can. J. Phys. 58 (1980) 1200. [VZ88] L. Versluis and T. Ziegler, J. Chem. Phys. 88 (1988) 322. [ZY98] Y. Zhang and W. Yang, Phys. Rev. Lett. 80 (1998) 890; see also Perdew, Burke, Ernzerhof, PRL 80 (1998) 891.