# Input Description

This is the contents of file 'mndo2020.txt':

Input Description for Program MNDO. Version 8.0 of 15 August 2019. By Walter Thiel, Max-Planck-Institut fuer Kohlenforschung, Kaiser-Wilhelm-Platz 1, D-45470 Muelheim, Germany. Contents: A. Selected references B. General overview C. Outline of standard input D. Description of standard input E. Outline of keyword-oriented input F. MNDO keywords G. MOPAC keywords H. File dictionary I. Input of external parameters J. Input of external data for HDLC optimizer K. Input for molecular dynamics and surface hopping driver L. Input for Born-Oppenheimer ground-state molecular dynamics ****************************************************************** A. Selected references. MNDO M.J.S.Dewar and W.Thiel, J.Am.Chem.Soc. 99, 4899 (1977). MNDOC W.Thiel, J.Am.Chem.Soc. 103, 1413, 1420 (1981). MNDO/d W.Thiel and A.A.Voityuk, Theor.Chim.Acta 81, 391 (1992). W.Thiel and A.A.Voityuk, J.Phys.Chem. 100, 616 (1996). AM1 M.J.S.Dewar, E.G.Zoebisch, E.F.Healy and J.J.P.Stewart, J.Am.Chem.Soc. 107, 3902 (1985). PM3 J.J.P.Stewart, J.Comput.Chem. 10, 209 (1989). OM1 M.Kolb and W.Thiel, J.Comput.Chem. 14, 775 (1993). OM2 W.Weber, Ph.D. Thesis, University of Zurich, 1996. W.Thiel and W.Weber, Theor.Chem.Acc. 103, 495 (2000). OM3 M.Scholten, Ph.D. Thesis, University of Dusseldorf, 2003. OMx P.O.Dral, X.Wu, L.Spoerkel, A.Koslowski, W.Weber, R.Steiger, M.Scholten and W.Thiel, J.Chem.Theory Comput. 12, 1082 (2016). OMx P.O.Dral, X.Wu, L.Spoerkel, A.Koslowski and W.Thiel, J.Chem.Theory Comput. 12, 1097 (2016). ODMx P.O.Dral, X.Wu and W.Thiel, J.Chem.Theory Comput. 15, 1743 (2019). MINDO/3 R.C.Bingham, M.J.S.Dewar and D.H.Lo, J.Am.Chem.Soc. 97, 1285 (1975). CNDO/2 J.A.Pople and G.A.Segal, J.Chem.Phys. 44, 3289 (1966). ****************************************************************** B. General overview. There are three modes of input for this program. - Standard input based on standard MNDO-style numerical data. - Keyword input based on keywords and MNDO-style numerical data. - MOPAC input based on keywords and MOPAC-style numerical data. The standard input allows the user to access all options of the program. It is fully specified in sections C-D (see below). The keyword input is equivalent to the standard input. It consists of a keyword section (up to ten lines of MNDO keywords and two lines of text) and standard MNDO-style numerical data for the remainder. MNDO keywords have the format= where denotes the name of any input variable in the standard input (as described below), and is the actual input value (see section F). The MOPAC input is provided to allow the use of MOPAC input files. It consists of a keyword section (up to three lines of keywords and two lines of text) and MOPAC-style data for the remainder. Most common keywords are recognized by this program and converted to options of the standard input. Some keywords are not available. In this case the program will print an appropriate message and then either stop or ignore the keyword. The treatment of MOPAC keywords is specified in section G of this input description. The MOPAC input for geometry and symmetry (etc) is not described here since it is the same as in MOPAC(6.0), see MOPAC manual for details. MNDO keywords can be combined with MOPAC keywords. The program automatically determines the mode of input from the contents of the input file. It checks the first three lines for characters A-Z or a-z (columns 1-80 of lines 1-2 and columns 1-30 of line 3) and assigns the mode of input as follows: - Standard input: No characters A-Z or a-z are found. - Keyword input: Characters A-Z or a-z appear, and there are only MNDO keywords. - MOPAC input: Characters A-Z or a-z appear, and there is at least one genuine MOPAC keyword. It is recommended to use keyword-oriented input for the options. The more complicated formatted input is described first in the following sections C-D, and the trivial relation to keywords is addressed thereafter in sections E-G. ****************************************************************** C. Outline of standard input. In this section, the standard input is summarized briefly. A more detailed description will be given in the next section. The standard input consists of the following parts. 1. First line with general options. 2. Second line with general options. 2.1 Special options for optimization etc (inp21=1). 2.2 Special options for eigenvector following (inp22=1). 2.3 Special options for analytic derivatives (inp23=1). 2.4 Special options for linear scaling SCF (inp24=1). 2.5 Special options for conventional SCF (inp25=1). 3. Data for the first molecule. 3.1 Title line, including special options for the molecule. 3.2 Atomic numbers and coordinates, one line per atom. 3.3 Symmetry conditions (ksym=1). 3.4 Definition of a reaction path or grid (kgeom=1-4). 3.5 Definition of fragments for SCF (ktrial=30-39). 3.6 Definition of orbital occupations (abs(iuhf)=3). 3.7 Data for configuration interaction (abs(kci)=1). 3.8 Data for perturbation treatment (abs(kci)=2-4). 3.9 Data for GUGA configuration interaction (abs(kci)=5). 3.10 Data for CIS and RPA excited-state module (abs(kci)=6-8). 3.11 Data for reference properties (inrefd.ne.0). 3.12 Data for COSMO solvation treatment (abs(icosmo)>2). 3.13 Data for external points (mminp=1,2). 3.14 Data for NMR chemical shifts (nmrnuc=6). 3.15 Definition of atomic masses (jop=2-6 and kmass>0). 3.16 Data for HDLC optimizer (ief<0). 4. Data for the following molecules (optional). Computations for an arbitrary number of molecules can be carried out in a single job. The input (4) for each new molecule normally consists of a new set of data (3) (see also option nexmol). The job is terminated if the title line (3.1) for the next molecule has 99 in columns 1-2 or if an end-of-file is encountered when reading the title line. As indicated above, input for 2.1-2.5 and 3.3-3.16 is only needed when requested by the corresponding input option (see section D). Input for continuation jobs (options jop=0-2). In order to continue an uncompleted calculation, the input file for the corresponding molecule is resubmitted after setting option middle on line 2. The continuation job makes use of the information saved on file nb4 by the preceding job. ****************************************************************** D. Description of standard input. The standard input is generally formatted (Fortran conventions). Free format (Fortran) is available in several parts of the input. The relevant chapters (see 3.2-3.4, 3.6-3.13, and 3.15 below) start with a corresponding reminder. Input for options is normally described in section D as follows: First, a brief overview table of a given input line containing - Option: Name of the MNDO keyword for which input is provided. - No.: Internal number of keyword (in IN1/IN2 or XN1/XN2). - Columns: Location of input data on input line. - Format: Fortran input format. - Short description. Thereafter, a complete description of each option containing - Option: Name of the MNDO keyword for which input is provided. - Full description specifying all possible input values. Frequently used input options are often starred in the overview table. It is generally recommended to use the default values for options that are not starred. Default values are defined for all input options. The description of other input data normally contains: - Variable: Internal name under which the data are stored. - Columns: Location of input data on input line. - Format: Fortran input format. - Full description specifying all possible input values. 1. ***** First line with general options ************************* Option No. Columns Format Short description limit 1 1-5 i5 Time limit in seconds. * iop 2 6-10 i5 Choice of semiempirical SCF method. * jop 3 11-15 i5 Type of calculation. * igeom 4 16-20 i5 Type of geometrical coordinates. * iform 5 21-25 i5 Input format for molecular data. nexmol 6 26-30 i5 Choice of input for next molecule. mplib 7 31-32 i2 Option for using parallel code. * ief 8 33-34 i2 Choice of special geometry optimizers. * idiis 9 35-36 i2 DIIS extrapolation procedure for SCF. * inrefd 10 37-38 i2 Input and evaluation of reference data. iparok 11 39-40 i2 Definition of non-standard parameters. mminp 12 41-42 i2 Definition of external points. * nmr 13 43-44 i2 Computation of NMR chemical shifts. * nsav7 14 45-45 i1 Generation of a new input file nb7. nsav8 15 46-46 i1 Generation of an output file nb8. nsav9 16 47-47 i1 Generation of an output file nb9. nsav13 17 48-48 i1 Generation of an output file nb13. nsav15 18 49-49 i1 Generation of an output file nb15. nsav16 19 50-50 i1 Generation of an output file nb16. immok 20 51-52 i2 Molecular mechanics correction for peptides. ihbond 21 53-54 i2 Automatic recognition of hydrogen bonds. ifld1 22 55-56 i2 Calculation of electric properties. ifld2 23 57-58 i2 Definition of the applied electric field. ifld3 24 59-60 i2 Finite-field SCF convergence criterion. icuts 25 61-62 i2 OM2 cutoff for three-center terms (all). icutg 26 63-64 i2 OM2 cutoff for three-center terms (gradient). iexbas 27 65-66 i2 Choice of polarized basis set. * icosmo 28 67-68 i2 COSMO solvation model. * ipsana 29 69-70 i2 Global option for analytic derivatives. immdp 30 71-75 i5 Option for dispersion corrections. inac 31 76-80 i5 Enforce special modes of NACME calculation. Option Full description limit Time limit in seconds. = 0 Effectively no time limit imposed. Default 10,000,000,000 seconds which is more than 300 years. = -1 Useful for testing restarts. iop Choice of semiempirical SCF method. =-23 ODM3. =-22 ODM2. =-10 MNDO/d. =- 8 OM3. =- 7 PM3. =- 6 OM2. =- 5 OM1. =- 2 AM1. =- 1 MNDOC. = 0 MNDO. = 1 MINDO/3. = 2 CNDO/2. = 5 SCC-DFTB. = 6 SCC-DFTB with Jorgensens parameters for reproduction of heats of formation. Special options for MNDO and MNDO/d. =- 3 MNDO/H with special treatment of hydrogen bonds. =- 4 MNDO with non-standard parameters which may be selected by option iparok (see description of iparok for details). Equivalent to iop=0 for iparok=0. =-13 MNDO/dH with special treatment of hydrogen bonds. jop Type of calculation. =-3 Born-Oppenheimer ground state dynamics, starting geometry. Further input required in the form of a namelist "dynvar.in". See section L for a detailed description. =-2 Gradient calculation, input geometry. =-1 Standard calculation, input geometry. = 0 Optimization for energy minimum. = 1 Optimization for transition state. = 2 Force constant analysis for input geometry. = 3 Optimization for energy minimum and force constant analysis for optimized geometry. = 4 Optimization for transition state and force constant analysis for optimized geometry. = 5 Same as jop=2 initially. However, use jop=3 if the cartesian gradient norm turns out to be too large (see option minitg). = 6 Same as jop=2 initially. However, use jop=4 if the cartesian gradient norm turns out to be too large (see option minitg). *** Special convention for inrefd.ne.0: *** Input for jop may be overwritten when vibrational frequencies *** are used as reference data (id1=16,26,27). *** See chapter 3.11 for details. igeom Type of geometrical coordinates used in the input file. = 0 Internal coordinates. = 1 Cartesian coordinates. *** igeom=1 is imposed by default for PDB input (ingeom=1). iform Input format for molecular data (see sections 3.2-3.16). = 0 Formatted input as described below (Fortran). = 1 Input in free format (Fortran). nexmol Choice of input for next molecule. =-1 Skip input of general options (before section 3) and read only the new molecular data (section 3). = 1 Read complete set of data (sections 1-3). = 0 Use the default conventions of the program: nexmol=-1 for standard input, nexmol= 1 for keyword input and MOPAC input. *** When using standard input, nexmol=1 must always be included in order to read complete data sets for each molecule. mplib Option for using parallel code. Choice of message passing library and granularity. = 0 Sequential code is used. = 1 Use fine-grained parallel code with MPI library. = 2 Use fine-grained parallel code with PVM library. =-1 Use coarse-grained parallel code with MPI library. =-2 Use coarse-grained parallel code with PVM library. Fine-grained parallelization for: - Direct SCF treatment. Coarse-grained parallelization for: - Reaction paths and grids, - finite-difference gradient, - finite-difference force constants. *** Option valid only for parallel code. *** Input ignored for sequential code except for using *** simplified control routines in coarse-grained runs. ief Choice of special geometry optimizers. = 0 Default procedures for optimizations, do not use eigenvector following (ief > 0) or HDLC (ief < 0). = 1 Use eigenvector following in geometry optimizations for energy minima and transition states. Default values for all options. No additional input required. = 2 Use eigenvector following in geometry optimizations for energy minima and transition states, Newton-Raphson search. Default options chosen for Newton-Raphson search. No additional input required. = 3 Use eigenvector following in geometry optimizations for energy minima and transition states. Read input for all options on one separate line of input. Detailed control of optimization strategy for expert users. Input description see chapter 2.2. =-1 Use HDLC optimizer for energy minima and transition states. Default values for all options. No additional input required from section 2.2. =-3 Use HDLC optimizer for energy minima and transition states. Read input for all options on one separate line of input. Detailed control of optimization strategy for expert users. Input description see chapter 2.2. idiis DIIS extrapolation procedure for SCF. =-1 Do not use DIIS. = 0 Do not use DIIS by default, but apply DIIS if the SCF procedure does not converge without using DIIS. In this case, idiis=1 is set automatically. = 1 Use standard DIIS procedure. *** Positive input values of idiis define two control variables: n = mod(idiis,10): DIIS starts at SCF cycle n and is applied in each SCF cycle thereafter (n=1...9). m = idiis/10 determines whether the diagonal elements of the error matrix are scaled by 1.02 (for numerical stability) and whether the error matrix must be computed with minimum memory (DSPMM, not DGEMM, see code). 0 Scaling, enough memory (DGEMM). 1 Scaling, minimum memory (DSPMM). 2 No scaling, enough memory (DGEMM). 3 No scaling, minimum memory(DSPMM). The recommended option idiis=1 implies n=1 and m=0. DIIS restarts for every 12 iterations (determined by experience). inrefd Input and evaluation of reference data. = 0 None. = N Input and evaluation for nonzero N. =-1 Output of results only on file nb19. = 1 Standard output. = 2 Detailed output. *** See chapter 3.11 for more information. iparok Definition of non-standard parameters. = 0 No such definition. =-1 Use MNDO parameters, if available, for elements which do not have their final parameters listed in the blockdata section. Only for AM1 and PM3. = 1 Read parameters from file nb14 using conventions analogous to MOPAC(6.0). The contents of file nb14 and their formats are described in section I. The new parameters overwrite the standard parameters in working common blocks. = 2 Read parameters from file nb14 using numerical input and numbering scheme 1. For additional remarks see iparok=1. = 3 Read parameters from file nb14 using numerical input and numbering scheme 2. For additional remarks see iparok=1. = 4 Use original MNDO parameters for Si and S, see J.Am.Chem.Soc. 100, 3607 (1978). Available with iop=0 and iop=-4. = 5 Use MNDO parameters copied from MNDO87-MNDO89, very small numerical deviations from option iop=0 with parameters copied from MOPAC(6.0). Only available with iop=-4. = 6 Use special fullerene MNDO parameters from Theor.Chim.Acta 92, 269 (1995). Parameter set B for carbon atoms. Only available with iop=-4. = 7 Do not use the predefined values of MNDO-type dependent parameters from the BLOCKDATA section, compute these parameters internally from scratch. Option to check numerical precision. Available with iop=0,-1,-2,-3,-4,-7 and iop.le.-10. = 8 Use parameters from PDDG approach (Jorgensen, JCC 2002). Available with MNDO (iop=0,-4) and PM3 (iop=-7). mminp Definition of external points. = 0 No such definition. = 1 Properties such as the electrostatic potential or the electric field are computed at the points defined. = 2 Quantum-chemical (QM) calculation in the field of external point charges. Used in QM/MM hybrid treatments where the molecular-mechanics (MM) atoms are located at the external points. *** Additional input is required for options mminp=1-2. *** See section 3.13 for details. nmr Computation of NMR chemical shifts. = 0 No such computation. =-1 Compute NMR shifts using standard parameters (see iop). Default options, no extra input. = 1 Compute NMR shifts using a special MNDO parametrization (A). Default options, no extra input. = 2 Compute NMR shifts using a special MNDO parametrization (B). Default options, no extra input. When using keyword input, special options can be defined by appropriate keywords. When using standard input, this can be accomplished as follows. =-2 Compute NMR shifts using standard parameters (see iop). Define options by additional input. See chapter 2.2 for more information. =11 Compute NMR shifts using a special MNDO parametrization (A). Define options by additional input. See chapter 2.2 for more information. =12 Compute NMR shifts using a special MNDO parametrization (B). Define options by additional input. See chapter 2.2 for more information. nsav7 Generation of a new input file nb7. Data from the following input sections are included: section 1, 2, 2.1-2.4, 3.1-3.3, 3.5, 3.9, 3.11. Useful for converting input data to a different format and for saving an optimized geometry for subsequent jobs. = 0 Do not write such a file. = 1 Write a file with standard input in standard format. = 2 Write a file with standard input in free format. = 3 Write a file with keyword input in standard format. = 4 Write a file with keyword input in free format. = 5 Write a file with MOPAC input. = 6 Analogous to nsav7=2, file contains cartesian coordinates. = 7 Analogous to nsav7=4, file contains cartesian coordinates. = 8 Analogous to nsav7=5, file contains cartesian coordinates. *** For nsav7=1-5 the geometry is generated in the same type of *** coordinates (internal or cartesian) as in the current job. nsav8 Generation of an output file nb8. = 0 Do not write such an output file. = 1 Write a geometry-charge file (nb8) corresponding to a MOPAC input file, with the atomic charge given at the end of the input for each atom. The geometry is saved in internal coordinates without symmetry. nsav9 Generation of an output file nb9. = 0 Do not write such an output file. = 1 Write a pdb-compatible file (nb9) which can be used as input file for evaluation programs that accept the pdb format. Only a small subset of the pdb format is implemented. pdb = protein data base. nsav13 Generation of an output file nb13 for postprocessing. = 0 Do not write such an output file. = 1 Write a graphics file (nb13) according to MOPAC conventions. = 2 Write an auxiliary file (nb13) and convert it to an input file for the MOLDEN program (nb93=molden.dat). nsav15 Generation of an output file nb15 collecting the current results after each energy and gradient calculation. Useful as an interface to other separate programs and for debugging purposes. = 0 Do not write such an output file. = 1 Save current Cartesian coordinates. = 2 Also save energy and gradient norms. = 3 Also save Cartesian and internal gradient (if available). = 4 Also save electrostatic potential and electric field (if available). = 9 Save everything. < 9 Always rewind file nb15 to save only the current results. = 9 Never rewind file nb15 to save all results consecutively. *** The data saved include any results available for external *** points (MM), for use in QM/MM treatments. *** For single-point multi-state calculations (icross=1,2,7) *** multiple energies and gradients will be saved. nsav16 Generation of an output file nb16 for evaluations using SYBYL. = 0 Do not write such an output file. = 1 Write a SYBYL input file (nb16) after generating all data required, including MOPAC-style bond orders and valencies as well as Mulliken populations and charges. = 2 Generate all data as for nsav16=1, without writing an output file nb13. This option makes sense only when these data are printed which requires the option nprint.ge.2 (see below). immok Molecular mechanics correction for peptides (MOPAC conventions). = 0 No such correction. = 1 Apply same correction as in MOPAC. ihbond Criteria for an automatic recognition of hydrogen bonds X...H-Y. Useful in MNDO/H calculations, iop=-3. = 0 Default criteria. Minimum distance rhxmin=1.1 Angstrom. Maximum distance rhxmax=5.0 Angstrom. Minimum angle angmin= 90 degree. = 1 Alternative criteria. Minimum distance rhxmin=1.1 Angstrom. Maximum distance rhxmax=2.5 Angstrom. Minimum angle angmin=100 degree. = n Input of distance criteria (n.gt.1). Minimum distance rhxmin=1.1 Angstrom. Maximum distance rhxmax= n Angstrom. Minimum angle angmin= 90 degree. ifld1 Calculation of electric properties. = 0 No such calculation. = n Finite-field calculation (n.gt.0). Note that the finite-field calculation is possible for all values of jop, except for jop=-2 which is treated as jop=-1. ifld2 Definition of the applied electric field. Relevant only for ifld1.gt.0. = 0 Default field of 0.001 au. = n Field of ifld1*10**(-n) au. ifld3 Finite-field SCF convergence criterion for density matrix. Relevant only for ifld1.gt.0. = 0 Default criterion of 10**(-10). = n Specific criterion of 10**(-n). icuts Cutoff for three-center orthogonalization corrections in OMx and ODMx (energy and gradient). Only relevant in combination with iop=-6,-8,-22,-23 (OM2, OM3, ODM2, ODM3). =-1 No cutoffs used. = 0 Default, equivalent to icuts=12. = n Three-center corrections neglected for resonance integrals between atoms I-J, if the product of the ss overlap integrals for the pairs I-K and J-K is below 10**(-n). icutg Cutoff for three-center orthogonalization corrections in OMx and ODMx (only for gradient). Only relevant in combination with iop=-6,-8,-22,-23 (OM2, OM3, ODM2, ODM3). =-1 No cutoffs used. = 0 Default, equivalent to icutg=6. = n Three-center corrections neglected for a given atom pair I-J during gradient evaluation, if the relative magnitude of the contribution from this pair to the total correction at the reference geometry is below 10**(-n). iexbas Choice of polarized basis set for the calculation of electric properties. = 0 Standard minimal basis. = 1 Add 2p polarization functions at H and 3d polarization functions at the first-row atoms. *** Experimental option *** Parameters for polarization functions must be provided via blockdata or input. Current default: Parameters for 2p (H) from blockdata, all others from input. Polarization functions are not included for elements which do not have the required parameters defined. *** Option is presently only implemented for MNDO and MNDO/d. icosmo COSMO solvation model. = 0 No treatment of solvation applied. = 1 Electrostatic COSMO treatment in its standard version. Water as solvent, default options, no further input. = 2 Electrostatic COSMO treatment with additional cavitation and dispersion interactions, currently available only for AM1. Water as solvent, default options, no further input. = 3 Nonstandard electrostatic COSMO treatment. Explicit input required, see chapter 3.12 for details. = 4 Nonstandard electrostatic COSMO treatment with additional cavitation and dispersion interactions. Explicit input required, see chapter 3.12 for details. =-n Single-point COSMO calculation (using icosmo=n, with n=1-4) after a standard gas phase calculation (with or without geometry optimization, see jop; unavailable in reaction paths or grids). *** Special convention for geometry optimizations (icosmo=1-4): Optimizations always use an electrostatic COSMO treatment only. For icosmo=2,4 single-point solvation energies at the optimized geometry are evaluated that include cavitation and dispersion. *** Special convention for inrefd.ne.0: *** Input for icosmo may be overwritten when solvation energies *** are used as reference data (id1=28 or id1=29). *** See chapter 3.11 for details. ipsana Global option for analytic derivatives. = 0 Do not use analytic derivatives unless selected by default. In the current version, this is done for the half-electron ROHF gradient, for the minimal CI gradient, for the GUGACI gradient, and for available second derivatives. Default options, no extra input. = 1 Use analytic derivatives whenever available. Default options, no extra input. = 2 Use analytic derivatives whenever available. Define options by additional input. See chapter 2.2 for more information. =-1 Do not use analytic derivatives. =-2 Do not use analytic derivatives, but use an input file with explicit input (i.e. like ipsana=+2). This option is included merely for convenience. Note: Fully analytic gradients with analytic integral and Fock matrix derivatives are implemented for all MNDO-type methods. Generic analytic gradients with numerical integral and Fock matrix derivatives are available for all NDDO-based methods. The fully analytic mode is faster and selected by default for MNDO-type methods, but use of the generic analytic mode can be enforced via option igrad=-3. The generic analytic mode is the default for other NDDO-based methods such as OM2, OM3, ODM2, and ODM3. Note: ipsana may be reset within the program when selecting the computation of analytic derivatives by default or when avoiding their computation in cases where they are not available. immdp Option for dispersion corrections. This option is only for AM1, PM3, OM2, OM3, ODM2, and ODM3. =-1 Do not include dispersion function corrections. = 0 Equivalent to immdp=-1 for all methods except for ODMx. Equivalent to immdp=-3 for ODM2 and ODM3. For AM1 and PM3: = 1 include dispersion corrections (PCCP 9, 2362 (2007)). No dispersion corrections included for other values. For OM2 and OM3: = 1 Include the D2 dispersion correction from Grimme with Elstner's damping function, see JCP 114, 5149 (2001); called D1 in Xin Wu, Ph.D. thesis (2013). = 2 Include the D2 dispersion correction from Grimme with Yang's damping function (II), see JCP 116, 515 (2002); D2 references: JCC 25, 1463 (2004); JCC 27, 1787 (2006); JCTC 12, 1082 (2016). For OM2, OM3, ODM2, and ODM3: = 3 Include the D3 dispersion correction from Grimme with Becke-Johnson damping and without three-body terms. Analytical gradients available for OM2, OM3, ODM2, and ODM3. D3 references: JCP 132, 154104 (2010); JCC 32, 1456 (2011); JCTC 9, 1580 (2013); JCTC 12, 1082 (2016). =-3 Include the D3 dispersion correction from Grimme with Becke-Johnson damping and with three-body terms. Analytical gradients available for OM2, OM3, ODM2, and ODM3. D3 references see above. *** Note: This is the default value for ODM2 and ODM3. inac Special modes of computing non-adiabatic coupling matrix element. = 0 Use best available method. Currently this involves the solution of the Z vector equations using the LAPACK DGESV routine which may require a significant amount of memory for systems with 50 or more non-hydrogen atoms. = n Use fully numerical method involving one SCF calculation per point of the two-sided differences. 2. ***** Second line with general options ************************ Option No. Columns Format Short description maxend 32 1-2 i2 Maximum number of SCF calcs per optimization. maxlin 33 3-4 i2 Maximum number of SCF calcs per line search. maxrtl 34 5-6 i2 Maximum number of optimization cycles per job. * iscf 35 7-7 i1 SCF convergence criterion for the energy. * iplscf 36 8-8 i1 SCF convergence criterion for the density. * middle 37 9-10 i2 Option for job continuation. * iprint 38 11-12 i2 Printing flag for optimization. kprint 39 13-14 i2 Printing flag for force constants. lprint 40 15-16 i2 Printing flag for vibratrional analysis. mprint 41 17-18 i2 Printing flag for gradients. * jprint 42 19-20 i2 Flag for printing of input data. * iprec 43 21-24 i4 Convergence criteria for optimization: values. * iconv 44 25-26 i2 Convergence criteria for optimization: type. ihess 45 27-28 i2 Definition of initial hessian. idfp 46 29-30 i2 Update of inverse hessian matrix. nrepet 47 31-32 i2 Special convergence criterion for f and x. linitg 48 33-34 i2 Check for vanishing initial gradient. lconvg 49 35-38 i4 Check for acceptable gradient norm (force). lgdum 50 39-40 i2 Check for gradient: Not used. ihdlc1 51 41-42 i2 Coordinates for HDLC optimizer: general. ihdlc2 52 43-44 i2 Coordinates for HDLC optimizer: core. ihdlc3 53 45-46 i2 Extra input for HDLC optimizer. ingeom 54 47-50 i4 Flag for special geometry input. intdir 55 51-52 i2 Integral direct SCF procedure. lindms 56 53-54 i2 Linear scaling CG-DMS approach. lindia 57 55-56 i2 Choice of full diagonalization after CG-DMS. linfrg 58 57-58 i2 Initial density from fragment calculations. inpfrg 59 59-60 i2 Read extra input for fragments. inp21 60 61-62 i2 Read extra input from section 2.1. inp22 61 63-64 i2 Read extra input from section 2.2. inp23 62 65-66 i2 Read extra input from section 2.3. inp24 63 67-68 i2 Read extra input from section 2.4. inp25 64 69-70 i2 Read extra input from section 2.5. iaterg 119 71-72 i2 Convention for atomization energy. Option Full description maxend Maximum number of SCF calculations for each optimization. Default 9999. = 1 Special option for jop=0,1,3,4: No geometry optimization, only one SCF calculation is carried out at the input geometry (as with jop=-1). maxlin Maximum number of SCF calculations for each line search. Default 4 for lsub=0 (see below). Default 10 for lsub=1,2 (see below). *** Not counting any extra SCF calculations *** required after an MO mapping failure (see maxmap) maxrtl Maximum number of optimization cycles per job. Default 9999. iscf SCF convergence criterion for the energy = 10**(-iscf) eV. Default 6 normally. Default 9 for GUGA-CI (abs(kci)=5) with keyword input. Recommended minimum 5. iplscf SCF convergence criterion for the diagonal elements of the density matrix, test on 10**(-iplscf). Default 6 normally. Default 9 for GUGA-CI (abs(kci)=5) with keyword input. middle Option for job continuation option. Conventions for jop=0-2. =-1 No job continuation since no restart information is saved. This may be useful for minimizing disk I/O operations. = 0 Normal job. = 1 Continuation of a previous job starting with a new cycle and using information saved on file 4 via middle=0. Additional possibility for jop=0,1. = 2 Continuation of a previous job using more stringent convergence criteria. This option allows the continuation of jobs which have converged using less stringent criteria. *** middle=-1 is enforced with the following combination of options: impar=1, kgeom=1-3, jop=0-1. iprint Printing flag for the optimization. =-5 No output. =-1 Small output. = 0 Standard output. = 1 Detailed output. Required to print the final interatomic distances for molecules with 100 or more atoms. = 5 Debug print. kprint Printing flag for the force constants. =-5 No output. =-1 Small output. = 0 Standard output. = 1 Detailed output. = 5 Debug print. lprint Printing flag for the vibrational analysis. =-5 No output. =-2 Minimum output. =-1 Small output. = 0 Standard output. = 1 Detailed output. = 5 Debug print. mprint Printing flag for the gradients. =-1 No output. = 0 Standard output. = 1 Detailed output including gradient norms during optimization. iprint=1 sets mprint=1 if the input value is mprint.le.0. = 2 More detailed output including dispersion contributions. = 5 Debug print. jprint Flag for printing of input data. =-1 No output. = 0 Standard output. = 1 Detailed output. = 2 Detailed output including memory allocation and rotational constants. = 5 Debug print. Required to print the initial interatomic distances for molecules with 100 or more atoms. = 6 Extended debug print including a printout of the standard input file. = 7 More output including documentation of the available options. iprec Convergence criteria for geometry optimization. Option to increase the precision of the convergence criteria. Default 1. Suggested maximum 100. Convergence tests refer to tolend(i). The default values for tolend(i) are divided by iprec to define the actual tolerances used. *** Conventions for jop=0,3,5. i=1 Test on norm of variables x. i=2 Test on function value f. i=3 Test on gradient components g or on gradient norm (iconv=3). i=4 Test on predicted decrease in f. Default values for tolerances. tolend(1)=1.D-04 tolend(2)=2.D-03 kcal/mol tolend(3)=1.D+00 kcal/(mol*Angstrom) tolend(4)=1.D-03 kcal/mol *** Conventions for jop=1,4,6. i=1 Test on absolute change of x. i=2 Test on relative change of x. i=3 Test on gradient components g. Default values for the tolerances. tolend(1)=1.D-08 tolend(2)=1.D-08 tolend(3)=1.D+00 kcal/(mol*Angstrom) iconv Type of convergence criteria for optimizations (jop=0,3,5). = 0 Successful termination if either - test on g and x satisfied - test on g and f satisfied - test on f or x satisfied for nrepet consecutive cycles - test on alpha.p.g (predicted decrease in f) satisfied = 1 Successful termination if either - test on g and x satisfied - test on g and f satisfied - test on f or x satisfied for nrepet consecutive cycles = 2 Successful termination if either - test on g satisfied - test on f or x satisfied for nrepet consecutive cycles = 3 Successful termination if - test on gnorm satisfied = 4 Successful termination if - standard Gaussian98 convergence criteria are satisfied * currently only available with HDLC optimizer For backward compatibility: =-1 Equivalent to iconv=2. =-2 Equivalent to iconv=3. =-3 Equivalent to iconv=4. ihess Definition of initial hessian. = 0 Same as ihess=1 except for the second and following points on a reaction path where ihess=2 is assumed. = 1 Initial hessian is estimated from a finite-difference approximation using gradient calculations at the initial point and a neighboring one. = 2 Initial hessian read from file 4. = 3 Initial hessian taken as unit matrix. Useful for testing purposes only. idfp Update of inverse hessian matrix. = 0 BFGS update. = 1 DFP update. nrepet Special convergence criterion for f and x, see above (iconv). Default 3. linitg Check for a vanishing initial gradient in an optimization. =-1 No such check. = 0 Set linitg=n=10 and check. = n Apply a criterion of tolend(3)/n. An optimization is considered to be unnecessary if the gradient is below this criterion (see description for iprec and iconv above). lconvg Check for an acceptable gradient norm to ensure that force constants can be computed in a meaningful manner, either at the final geometry after geometry optimization (jop=3,4) or at the initial input geometry (jop=5,6). =-1 No such check, any gradient norm is accepted. = 0 Set lconvg=n=10 and check. = n Check whether the final optimized gradient norm is below tolend(3)*n for jop=3,4 or whether the Cartesian gradient norm is less than n kcal/(mol*Angstrom) for jop=5,6. Force constants are computed in these cases, otherwise their calcalution is skipped (jop=3,4) or the geometry is optimized (jop=5,6). lgdum Check for gradient norm. Not used presently. ihdlc1 Coordinates used by the HDLC optimizer (ief < 0) in general. = 0 HDLC internal coordinates. = 1 Cartesian coordinates. ihdlc2 Coordinates used by the HDLC optimizer (ief < 0) for the reaction core (microiterative transition state search) or for the first and only fragment defined. = 0 HDLC primitive internal coordinates (reaction core). = 1 HDLC total connection scheme (reaction core). = 2 Cartesian coordinates (reaction core). = 3 DLC primitive internal coordinates (only one fragment). = 4 DLC total connection scheme (only one fragment). ihdlc3 Additional input for the HDLC optimizer (ief < 0). = 0 Read additional input from separate input file nb42. See section J for details. Treat as ihdlc3=2 if this file does not exist. = 1 Read additional input from standard input file nb5. See section 3.16 for details. = 2 Use default HDLC options, no input. Try to define constraints based on the standard coordinate input, provided that there are no dummy atoms and no symmetry relations: experimental option (may be unsafe). ingeom Flag for special geometry input. = 0 Nothing special. = 1 Read geometry and associated data from PDB-type input which replaces sections 3.2-3.5 of the standard input. See last part of section 3.2 for more information. =-1 Skip input for geometry and associated data. Sections 3.2-3.5 of the standard input are omitted. This may be useful when this code is combined with another code that provides these input data. intdir Choice of integral-direct SCF procedure. = 0 Conventional integral handling. = 1 Direct approach, without thresholds. = 2 Direct approach, default thresholds. = 3 Direct approach, special thresholds, inp24=1. *** Main limitations of the current implementation: a) No methods with orthogonalization corrections (OM1, OM2). b) No electron correlation. c) No COSMO solvation treatment. d) No SCF with external point charges. lindms Choice of linear scaling CG-DMS approach: conjugate gradient density matrix search. = 0 Conventional SCF treatment. = 1 CG-DMS, square matrix, without thresholds. = 2 CG-DMS, square matrix, default thresholds. = 3 CG-DMS, square matrix, special thresholds, inp24=1. = 4 CG-DMS, sparse matrix, default thresholds. = 5 CG-DMS, sparse matrix, special thresholds, inp24=1. *** Main limitations of the current implementation: a) No open-shell systems. b) lindia=1 required for some applications (e.g. NMR). c) linfrg=1 often required. lindia One conventional diagonalization after CG-DMS convergence to obtain MO eigenvalues and eigenvectors. = 0 No such diagonalization. = 1 Allow such diagonalization. linfrg Block-diagonal initial density matrix from separate RHF-SCF calculations on user-defined fragments using the same convergence criteria as in the molecular case. = 0 Do not build such an initial density matrix. = n Number of SCF iterations allowed for each fragment. Recommended values are n=1 and n=2 to obtain a sufficiently accurate initial guess for CG-DMS. Normally (unless inpfrg=-1) the definition of the fragments requires some extra input (section 3.5). *** Some variables are fixed internally during a fragment SCF calculation: inout=0, iuhf=-1, ifast=2, nstart=-1. inpfrg Input to define the fragments for CG-DMS in section 3.5. =-1 Do not read such input, treat the whole molecule as a single fragment (only useful for testing purposes). = 0 Read such extra input using the default format. = n Read such extra input using other formats. Not yet implemented. lindum Reserved for linear scaling integral-direct methods. Not used presently. inp21 One line of extra input as described in section 2.1. Special options for optimizations and force constants. = 0 Do not read such extra input, use default options. = 1 Read such extra input. inp22 One line of extra input as described in section 2.2. Special options for eigenvector following. = 0 Do not read such extra input, use default options. = 1 Read such extra input. Set internally for ief=3 or ief=-3 (see above). inp23 Two lines of extra input as described in section 2.3. Special options for analytic derivatives. = 0 Do not read such extra input, use default options. = 1 Read such extra input. Set internally for ipsana>1, nmr>10, or nmr<-2 (see above). inp24 One or two lines of extra input as described in section 2.4. Special options for linear scaling and direct SCF methods. = 0 Do not read such extra input, use default options. = 1 Read one line of extra input. Set internally for intdir=3 or lindms=3,5 (see above). = 2 Read two lines of extra input. Not set internally, explicit input required. inp25 One line of extra input as described in section 2.5. Special options for conventional SCF methods. = 0 Do not read such extra input, use default options. = 1 Read such extra input. iaterg Convention for atomization energy. =-1 Treat SCF atomization energy with post-SCF corrections as ZPVE-exlusive atomization energies at 0 K. Default in ODM2 and ODM3. = 0 Equivalent to iaterg=-1 for ODM2 and ODM3 and to iaterg=1 for all other methods. = 1 Treat SCF atomization energy with post-SCF corrections as if they were enthalpies of atomization at 298 K. Default in MNDO, MNDO/d, MNDO/H, MNDO/dH, MNDOC, AM1, PM3, OM1, OM2, and OM3 as well as in MINDO/3, CNDO/2, and SCC-DFTB (special conventions for iop=6). 2.1 ***** Special options for optimization etc ***** inp21=1 ***** The input in this section is needed only if special options are chosen for optimizations or force constant calculations. In most applications, the default options are sufficient, and there is no such extra input. Option No. Columns Format Short description nrst 191 1-4 i4 Reset of hessian matrix to initial values. ldrop 192 5-8 i4 Criterion for restarting an optimization. ldell 193 9-10 i2 Change of variables in such a restart. lsub 194 11-12 i2 Choice of line search routine. lalpha 195 13-14 i2 Initial step size for line search. lconv 196 15-16 i2 Convergence criterion for line search: step. ltolf 197 17-18 i2 Convergence criterion for line search: energy. lmaxst 198 19-20 i2 Step size limit for line search. igrad 199 21-22 i2 Enforce special modes of gradient calculation. lpoint 200 23-24 i2 Points for numerical gradient calculation. lfac 201 25-28 i4 Step size for numerical gradient calculation. lldum 202 29-30 i2 Option presently not used. kpoint 203 31-32 i2 Points for numerical force constants. kfac 204 33-36 i4 Step size for numerical force constants. kmass 205 37-38 i2 Atomic masses for the vibrational analysis. kkdum 206 39-40 i2 Option presently not used. ntemp 207 41-42 i2 Number of temperatures for properties. ntemp1 208 43-46 i4 Lowest temperature (in Kelvin). ntemp2 209 47-50 i4 Temperature increment (in Kelvin). Option Full description nrst Number of cycles between resetting the hessian matrix to its initial values. Default 999. =-1 Do not reset the hessian matrix even if the gradient and the search direction are almost orthogonal. Define nrst=99999 for internal use. ldrop A restart in the geometry optimization is carried out if the heat of formation drops by more than ldrop kcal/mole in two consecutive cycles (jop=0,3,5). Default 10. ldell The geometrical variables in such a restart are changed by 0.001*ldell units (Angstrom or radian). Default 10. lsub Choice of line search routine. = 0 Quadratic search via fstmin. = 1 Quadratic search via locmin. = 2 Cubic search via linmin if gradients are always available, otherwise quadratic search via locmin. lalpha Initial step size for line search. = 0 Taken from preceding cycle (times pnormlast/pnorm for lsub=1 or 2). For variational wavefunctions and lconv.gt.25 (see below), the program uses lalpha=1 by default. = 1 Use step size alpha=1 initially. = 2 Taken from preceding cycle without any change. lconv Convergence criterion for line search: step size. = 0 Convergence if the predicted step size differs from the previous step size told by less than tcrit=0.01*told+0.01 = n Multiply default value by n. For variational wavefunctions and lsub=0, the program uses lconv=50 by default. For nonvariational wavefunctions, the program always uses lconv=1 regardless of the actual input. ltolf Convergence criterion for line search: energy. = 0 Convergence if the energy drops by less than 0.5*tolend(2) for two consecutive points. = n Multiply default value by n. Negative ltolf values turn off this convergence criterion. lmaxst Step size limit for line search. Maximum allowed change of the variables for two consecutive points in the line search. = 0 Use default values of 0.1 Angstrom and 0.1 radian. = n Multiply default values by n. igrad Special modes of gradient calculation. = 0 Internal choice of best method for gradient calculation depending on the type of wavefunction. = 1 Enforce numerical finite-difference calculation via full energy evaluations for all types of wavefunctions. This implies quadratic line searches. =-3 Enforce numerical finite-difference evaluation of integral and Fock matrix derivatives even when the corresponding analytic derivatives are available (for testing purposes). This implies use of the PSOMX driver (rather than PSDRV). lpoint Number of points for numerical gradient calculation. = 0 Central difference approach: compute gradient from two energy evaluations at both sides of the reference point. = 1 One-sided differentiation: compute gradient from the energies at the reference point and one distorted point. *** lpoint=1 is significantly less accurate than lpoint=0. *** lpoint=1 is not available in this version of the code. *** lpoint=0 is used regardless of the actual input. lfac Step size for numerical gradient calculation: lfac multiplies the predefined minimum step size of 0.00001 Angstrom for distances, 0.002 degree for bond angles, 0.005 degree for dihedral angles, and 0.00001 Angstrom for x,y,z Cartesian coordinates. Default 100. kpoint Number of points for numerical force constants during the calculation of the offdiagonal force constants by numerical differentiation of the gradient. = 0 Average from two calculations for f(i,j) and f(j,i). = 1 One single calculation for f(i,j). *** kpoint=1 is significantly less accurate than kpoint=0. *** kpoint=1 not available with impar=1. *** kpoint=1 is ignored for variational wavefunctions with fast gradients or for analytic gradients, in these cases kpoint=0 is always used regardless of the actual input. *** kpoint=1 is allowed in all other cases to speed up slow numerical force constant calculations when accuracy is not of major concern to the user. kfac Step size for numerical force constants. Displacement of cartesian coordinates in units of 0.00001 au to generate distorted geometries for computing the gradient. The numerical force constants are obtained from such gradients by finite difference. Default 1000. kmass Atomic masses used for the vibrational analysis. =-1 Average atomic weight. = 0 Mass of the most abundant isotope. = n In addition to the default case, n isotopomers are treated which are defined by input (see section 3.15). A vibrational analysis for isotopic substitution (defined via kmass=n) can be done by using a previously computed force constant matrix from the restart file (option middle=1). ntemp Number of temperatures for calculation of thermodynamic properties after force constant analysis. Default 10. Maximum 25. Default temperatures for ntemp=0 are 273.15, 298.15, 300, 400, ... 1000 K regardless of input for ntemp1 and ntemp2. ntemp1 Lowest temperature (in Kelvin). Default 100. Minimum 100. ntemp2 Temperature increment (in Kelvin). Default 100. The default values defined above are generally used if there is no input from this section (inp21=0). These default values are also used in the case of explicit input (inp21=1) unless a specific nonzero input value is provided for a given option. 2.2 ***** Eigenvector following ***** inp22=1 ******************** The input in this section is needed only if special options for eigenvector following are required. In most applications, the default options are sufficient, and there is no extra input for eigenvector following (standard cases ief=-1, ief=1, or ief=2). Many general options for optimizations apply also to eigenvector following (see preceding chapter 2). This includes: maxrtl, iscf, ipl, middle, iprint, jprint, and iprec. Option No. Columns Format Short description mode 80 1-2 i2 Hessian eigenvector that is followed. ireclc 81 3-4 i2 Recalculation of Hessian matrix. iupd 82 5-6 i2 Update of Hessian matrix. igthes 83 7-8 i2 Determination of Hessian matrix. llamda 84 9-10 i2 Determination of current lambda: general. lnonr 85 11-12 i2 Determination of current lambda: details. lrscal 86 13-14 i2 Check on predicted step size. lgnmin 87 15-16 i2 Check on applied step size. lnoupd 88 17-18 i2 Check on current trust radius. lefdum 89 19-20 i2 Less tight convergence criterion. dmax 1 21-30 f10.5 Initial trust radius (Angstrom or rad). ddmin 2 31-40 f10.5 Minimum trust radius (Angstrom or rad). ddmax 3 41-50 f10.5 Maximum trust radius (Angstrom or rad). rmin 4 51-60 f10.5 Minimum acceptable ratio for energy change. rmax 5 61-70 f10.5 Maximum acceptable ratio for energy change. omin 6 71-80 f10.5 Minimum acceptable overlap for TS mode. Option Full description mode Hessian eigenvector that is followed. Not used in search for energy minimum. Default 1 in search for transition state. ireclc Number of cycles between explicit calculation of the Hessian matrix. Default 9999. iupd Update of Hessian matrix. =-1 No such update. = 0 Default behaviour (Powell or BFGS). = 1 Powell update (default for TS). = 2 BFGS update (default for minimum). = 3 Murtagh & Sargent update. igthes Determination of Hessian matrix. = 0 Diagonal matrix, empirical estimate. Default in Yarkony CI search, overriden in all other cases. = 1 Calculated numerically by one-sided finite difference. Default in all other cases. = 2 Read from file nb4. = 3 Calculated numerically by two-sided central differences. = 4 Diagonal matrix calculated numerically. = 5 Unit matrix. llamda Determination of current lambda value: general approach. = 0 Follow MOPAC-type default procedures: Try Newton-Raphson, P-RFO, and QA algorithms in this order. = 1 Follow procedures from J.Nichols et al, JCP 92, 340 (1990). = 2 Insist on lambda=0 (Newton-Raphson). lnonr Determination of current lambda value: details for llamda=0. = 0 Use default procedures and allow for pure Newton-Raphson step with lambda=0, if appropriate. = 1 Use default procedures, but do not allow for pure Newton-Raphson step. lrscal Check on predicted step size. = 0 No request for scaling. = 1 Scale the predicted step down if it is larger than the trust radius. lgnmin Check on applied step size. = 0 No check on gradient norm. = 1 Do not allow the gradient norm to increase during TS search. Reject previous step and try smaller step. lnoupd Check on current trust radius. = 0 Allow changes in the trust radius based mainly on the ratio between found and expected energy changes. = 1 Do not allow such changes. lefdum Less tight convergence criterion. = 0 Use default criterion for gradient norm: tol2 = 1.0 = n Use less tight criterion for gnorm: tol2 = n Note that more stringent criteria can be chosen using iprec. However, input for lefdum overrides input for iprec. dmax Initial trust radius (Angstrom or rad). Default 0.2. ddmin Minimum trust radius (Angstrom or rad) that is acceptable in the update. Default 0.001 (ief>0). Default 0.0002 (ief<0). ddmax Maximum trust radius (Angstrom or rad) that is acceptable in the update. Default 0.5 for minimum (ief>0). Default 0.3 for TS (ief>0). Default 1.0 for minimum and TS (ief<0). rmin Minimum ratio between found and expected energy changes that is acceptable. Default 0.0 (ief<0, not used for ief>0). rmax Maximum ratio between found and expected energy changes that is acceptable. Default 4.0 (ief>0, not used). Default 10.0 (ief<0). omin Minimum overlap between current and previous TS mode that is acceptable. Default 0.8 (ief>0). Default 0.6 (ief<0). The default values defined above are generally used if there is no input from this section (inp22=0; ief=-1 or ief=1 or ief=2). These default values are also used in the case of explicit input (inp22=1; ief=-3 or ief=3) unless a specific nonzero input value is provided for a given option. In the case of a Newton-Raphson search (ief=2, no extra input), the program internally employs the following definitions: ireclc=1, igthes=1, llamda=2, lnoupd=1 (other defaults unchanged). Some options are irrelevant with ief=2 (mode,iupd,lnonr,omin). 2.3 ***** Analytic derivatives ***** inp23=1 ********************* In standard applications with analytic derivatives, this input section is skipped because the program automatically assigns reasonable values to all input variables using internal optimization procedures. Specific input can be requested by the user (inp23=1; ipsana>2 or nmr>10 or nmr<-2, see earlier description of ipsana and nmr). In this case, two lines of input are expected as described below: 29 INTEGER options (first line), and 8 REAL options (second line). ***** First line ***** Option No. Columns Format Short description ipsprt 90 1-5 i5 Printing option for analytic derivatives. ienrg 91 6-10 i5 Debug option (see code). Do not use. icore 92 11-15 i5 Amount of memory available as a buffer. idisk 93 16-20 i5 Amount of disk space available. imix 94 21-22 i2 Computation of Fock matrix derivatives. idens 95 23-24 i2 Computation of density matrix derivatives. indsym 96 25-26 i2 Computation of various intermediate terms. iqswap 97 27-28 i2 Control of rhs swapping for CPHF equations. iaveit 98 29-30 i2 Estimated average number of CPHF iterations. imaxit 99 31-34 i4 Estimated maximum number of CPHF iterations. inrhs 100 35-36 i2 Number of CPHF equations solved together. ikrvec 101 37-38 i2 Debug option (see code). Do not use. irows 102 39-42 i4 Number of rows of CPHF K matrix held in core. iprcon 103 43-44 i2 Preconditioner in iterative CPHF solution. incpus 104 45-48 i4 Not yet implemented. Do not use. idstrp 105 49-50 i2 Fine tuning (see code). Do not use. ihlst 106 51-52 i2 Treatment of intermediate HE gradient terms. ihlwrp 107 53-54 i2 Treatment of redundant CPHF variables (HE). ikmode 108 55-56 i2 Method for computing CPHF K matrix. isolve 109 57-58 i2 Selection of iterative linear CPHF solver. ikrsav 110 59-62 i4 Number of shared basis vectors for solver. iumix 111 63-64 i2 Sequential file to store Fock matrix derivs. iurhs 112 65-66 i2 Direct access file to store rhs (CPHF). iuk 113 67-68 i2 Sequential file to store the CPHF K matrix. iures 114 69-70 i2 Sequential file to store the results. * nmrlev 127 71-72 i2 NMR: Choice of integral approximation. * intctl 128 73-76 i4 NMR: Calculation of three-center terms. incoff 129 77-78 i2 NMR: Cutoff for three-center integrals. * nmrnuc 130 79-80 i2 NMR: Centers where shieldings are computed. Option Full description ipsprt Printing option for analytic derivatives. = -5 No output, except for fatal errors. = -1 No output, except for fatal errors and warnings concerning the choice of unreasonable input options. = 0 Standard output. = 1 Detailed output. = 2 Collect and report more cpu times. = 5 Debug print. Additional debug print can be generated by adding increments: DELTA Print extra information on: 16 half-electron derivatives 32 solution of CPHF equations 64 data passed between stages 128 dynamic memory allocation 256 initial integral calculations 512 explicit CPHF K matrix formation 1024 summation of response quantities 2048 computation of density matrix derivatives by finite difference 4096 integral transformation 8192 hi-tech solvers (isolve=2,3,4,5,6) 16384 NMR-related code 32768 CI-related code Note: Setting ipsprt=65535 will produce the largest possible debug output (which will be HUGE). See code for further details. For ipsprt=0 (default) ipsprt may be redefined internally. For first derivatives, ipsprt=-1 is usually set internally, with ipsprt=1 for mprint=5. For second derivatives, ipsprt=kprint. ienrg Debug option (see code). Do not use. = 0 Normal calculation. icore Amount of memory available as a buffer. = 0 Use as much memory as possible. See PARAMETER statement for LEN in the main program for the predefined upper limit. = n Memory in Mbyte. Internally converted to number of 8-byte words (n*131072) and automatically reduced to LEN when exceeding the allowed upper limit. idisk Amount of disk space available. = 0 Unlimited disk space assumed. = n Disk space in Mbyte. Internally converted to number of 8-byte words (n*131072). Note: Slightly more disk space might be needed than expected due to headers from the Fortran run-time library. imix Computation of Fock matrix derivatives, or equivalently, of second derivatives of the energy wrt coordinates and density. = 0 Automatic optimum choice. = 1 Compute once, keep in core. = 2 Compute once, swap out during CPHF. = 3 Store atom-pair contributions and recompute derivatives as needed. = 4 Use numerical differentiation of the Fock matrix. Only supported in the PSOMX driver. idens Computation of first derivatives of the density matrix wrt coordinates. = 0 Automatic optimum choice. This will never be numerical. = 1 Numerical differentiation using extrapolated integrals (only for testing purposes). = 2 Select best analytic method. = 3 CPHF in MO basis, direct solution of linear system using LAPACK (DSPSV). = 4 CPHF in MO basis, iterative solution in-core. = 5 CPHF in MO basis, iterative solution out-of-core. = 6 CPHF in MO basis, iterative solution with on-the-fly recomputation of the K matrix (integral-direct approach). = 7 CPHF in AO basis with a separate solution phase. This may be slightly faster than idens=8, but requires quartic memory or disk storage. This option allows parallelization. = 8 CPHF in AO basis with no separate solution phase. This may run in N**2 memory without too much redundant work, if imix=3 and iqswap=3. indsym Calculation of force constants: Numerical evaluation of first derivatives of the density matrix wrt coordinates. = 0 Automatic optimum choice. = 1 Central finite differences using symmetric steps in both directions. = 2 One-sided finite differences. Never selected by default. Should be used with caution since numerical instabilities may occur. NMR: Calculation of one-electron terms. = 0 Automatic optimum choice. = 1 Explicitly evaluate all one-electron quantities (not using hermiticity). This is two times slower than option indsym=2, but may be used to find failures in integral routines since equivalent integrals are computed twice independently. = 2 Use hermiticity of the one-electron quantities (HAB, H0B) to reduce the number of integrals needed. **** Input option mainly for debugging. **** Should be avoided in production runs. iqswap Control of swapping for right-hand sides (rhs) of CPHF equations. = 0 Automatic optimum choice. = -1 Never swap. = 1 Build out-of-core, bring in for CPHF. Only useful with direct solver from LAPACK, see idens=3. = 2 Build out-of-core. = 3 Recompute rhs as needed and consume solution vectors on the fly. Only valid with idens=8. iaveit Estimated average number of CPHF iterations. = 0 Use reasonable heuristic estimate. Check code for details. = -n Use heuristic estimate*ABS(n). = n Input value. Note: iaveit will be reset to the average number observed during the actual run. The change is only visible to routines called from PSDRV, see code. imaxit Estimated maximum number of CPHF iterations. = 0 Use default: 10 + 1.5*iaevit. = -1 Use default, same as imaxit=0. = n Input value. Note: imaxit will be reset to the largest number observed during the actual run. The change is only visible to routines called from PSDRV, see code. inrhs Number of CPHF equations that are solved simultaneously. = 0 Automatic optimum choice. = -1 Solve for all variables at once. = -n ncpvrs/2**(n-1) variables at once. ncpvrs = number of perturbations. = n Input value: preferably 3, 6, or 12. ikrvec Debug option (see code). Do not use. = 0 Normal calculation. irows Number of rows of CPHF K matrix that are held in core during out-of-core solution. = 0 Automatic optimum choice. = -n maxrows/2**n. = n Input value. iprcon Preconditioner in iterative CPHF solution = 0 Automatic optimum choice. For idens=6,7,8: iprcon=3. For idens=4,5 : iprcon=4. = 1 J.A.Pople, R.Krishnan, H.B.Schlegel, J.S.Binkley, Int.J.Quant.Chem.Symp. 13, 225 (1979). = 2 M.J.S.Dewar, D.A.Liotard, Theochem 206, 123 (1990). Use dshift as the shift value, with a default shift of 0.15 au. In most cases iprcon=3 fares better. = 3 Estimate shift value for each active coupling block as an average of K(Last1,First2,Last1,First2) and K(First1,First2,First1,First2). = 4 Use exact diagonal value as a shift (available only for idens=4,5). = 5 Estimate shift value for each coupling block by interpolation on K(First1,First2,First1,First2), K(Last1,First2,Last1,First2) and K(Last1,Last2,Last1,Last2). Usually much worse than iprcon=3, but sometimes slightly better. = 6 Assume shift value to be equal to the corresponding Coulomb integral. Almost as efficient as iprcon=4. Should be used only if iprcon=3 is not satisfactory, because it needs an N**3 step while gaining marginal improvement over iprcon=3. **** Additional conventions: For iprcon=3-6, the estimate is scaled by dshift (default 1.0 for iprcon=3,5,6 and default 0.9 for iprcon=4). For iprcon=2,3,5,6, the scaled estimate is replaced by 0.7*DELTA (difference in corresponding orbital energies) if it exceeds that value. For iprcon=4, there is no such check. incpus Not yet implemented. Do not use. = 0 Normal calculation. idstrp Fine tuning (see code). Do not use. = 0 Automatic optimum choice. ihlst Treatment of intermediate terms in the static part of the half-electron gradient. = 0 Automatic optimum choice. = 1 Compute whenever needed. = 2 Compute once and keep in core. ihlwrp Wrapping redundant into nonredundant CPHF variables for the half-electron gradient. = 0 Automatic optimum choice. = 1 Explicit formation of rhs (CPHF) for the Z vector in MO basis. = 2 Transformation into AO basis. ikmode Method for computing CPHF K matrix. = 0 Automatic optimum choice. = 1 Use explicit transformation of integrals into MO basis. = 2 Use AO->Fock->MO procedure. isolve Selection of iterative linear solver, usually with basis vector sharing (bvs). = 0 Automatic optimum choice. = 1 Pople's solver = 2 Solver with bvs, minimize norm of the residual. = 3 Solver with bvs, make residual orthogonal to the basis vectors. = 4 CG-like solver with bvs, minimize 2-norm of the residual. = 5 CG-like solver with bvs, make residual orthogonal to the search directions. = 6 Quasi-CG solver with bvs. Reduces to exact CG when dbascr is less than zero or when solution for a single variable is attempted with ikrsav=0. Note: Options isolve=4-6 will use amount of memory independent of the number of iterations, but they may require more iterations than isolve=1-3 (especially when using iprcon=1). ikrsav Defines the number of basis vectors shared during iterative linear solution (N). ikrsav is actually required to be N+1. = 0 Automatic optimum choice. = 1 Do not share basis vectors. = n Input value. Note: For isolve=4-6, N corresponds to the size of the stabilization trail. iumix Number of sequential file to store the Fock matrix derivatives during CPHF. = 0 Use default (95) = n Input value, n.gt.20 required. iurhs Number of direct access file to store the right-hand sides and density matrix derivatives during CPHF. = 0 Use default (96) = n Input value, n.gt.20 required. iuk Number of sequential file to store the CPHF K matrix. = 0 Use default (97) = n Input value, n.gt.20 required. iures Number of sequential file to store the computed gradients and force constants. = 0 No such file. = n Input value, n.gt.20 required. Note: For positive iures, an additional sequential file is generated (with number iures+1) which holds dump information from various common blocks and certain cpu times (useful for code development). Note: iures=98 is recommended if these two files are needed (iures,iures+1). This choice assigns files 95-99 to hold the internal data from the analytical derivative code (assuming the default file numbers for iumix,iurhs,iuk). nmrlev NMR: Choice of integral approximation. = 0 Use default value (nmrlev=3). = 1 Include only one-center terms. = 2 Include also two-center terms. = 3 Include also three-center terms. intctl NMR: Calculation of three-center terms. = 0 Use default value (intctl=14). = 1 Use STOs with an expansion in incomplete gamma functions. This is faster than intctl=2, but numerically less stable. = 2 Use STOs with an expansion in derivatives of modified Bessel functions (slow and accurate). = 10 Use GTOs with an STO-4G expansion. This is much faster than intctl=1. = 11 Use GTOs with an STO-1G expansion. = 12 Use GTOs with an STO-2G expansion. = 13 Use GTOs with an STO-3G expansion. = 14 Use GTOs with an STO-4G expansion. = 15 Use GTOs with an STO-5G expansion. = 16 Use GTOs with an STO-6G expansion. incoff NMR: Cutoff for three-center integrals. = -1 Do not apply any cutoff, compute all three-center integrals explicitly. = 0 Use default value (incoff=6). = n Use cutoff of 10**(-n) atomic units. n=6 translates into an accuracy of typically better than 0.01 ppm for NMR chemical shifts. nmrnuc NMR: Centers where the shielding tensors are computed. = 0 Use default value (nmrnuc=4). = 1 Compute shieldings for all nuclei and all dummy atoms (NICS values). = 2 Compute shieldings for all nuclei that have been parametrized and for all dummy atoms (NICS values). = 3 Compute shieldings for all nuclei that have been parametrized, but not for dummy atoms. = 4 Compute shieldings for all carbon, nitrogen, and oxygen atoms. = 5 Compute shieldings only for all carbon atoms. = 6 Compute shieldings for all centers defined in input section 3.14. ***** Second line ***** Option No. Columns Format Short description dstep 8 1-10 g10.5 Step size for numerical density derivatives deconv 9 11-20 g10.5 SCF convergence (E) for density derivatives dpconv 10 21-30 g10.5 SCF convergence (P) for density derivatives dcphf 11 31-40 g10.5 Residual error norm for CPHF solution. dprec 12 41-50 g10.5 Desired precision of the derivatives. dcdiff 13 51-60 g10.5 Warning if MO coefficients are imprecise. dshift 14 61-70 g10.5 Shift parameter for CPHF preconditioners. dbascr 15 71-80 g10.5 Rejection criteria for orthogonalization. Option Full description dstep Step size (Angstrom) for the numerical evaluation of density matrix derivatives. Option for debugging and testing. = 0.0 Use default value (0.0002 Angstrom) deconv SCF convergence criterion for the energy during the numerical evaluation of density matrix derivatives (in eV). Option for debugging and testing. = 0.0 Automatic optimum choice. dpconv SCF convergence criterion for the density matrix during the numerical evaluation of density matrix derivatives. Option for debugging and testing. = 0.0 Automatic optimum choice. dcphf Residual error norm for iterative CPHF solution. = 0.0 Automatic optimum choice. Note: The default values for deconv and dpconv are defined to be consistent with the actual value of dcphf. dprec Desired precision of the derivatives. = 0.0 Use default values of 0.00001 au for gradient and 0.00001 au for force constants. Note: The default value for dcphf is defined to be consistent with the actual value of dprec. dcdiff Issue a warning if the maximum change in the MO coefficients exceeds this value during the numerical evaluation of density matrix derivatives. = 0.0 Use default (dcdiff=0.1). = 2.0 Disable the warning. dshift Shift parameter for CPHF preconditioners. See comments for iprcon. = 0.0 Automatic optimum choice. dbascr Rejection criteria for singular values of orthogonalized basis sets (see code). = 0.0 Use default values of 0.2 for isolve=1-3 and 1.0D-10 for isolve=4-6. 2.4 ***** Linear scaling and direct methods ***** inp24>0 ******** In standard applications, this input section may be skipped. The program will then use reasonable default values for all options. Specific input can be requested by the user (inp24>0; intdir>1 or lindms>1, see earlier description of intdir and lindms). One line of input is expected for inp24=1 or intdir=3 or lindms=3; two lines of input are expected for inp24=2 or intdir=4 or lindms=4: 16 INTEGER options (first line), and 8 REAL options (second line). ***** First line ***** Option No. Columns Format Short description maxcg 171 1-5 i5 Maximum number of CG cycles during DMS. maxpur 172 6-10 i5 Maximum number of McWeeny purifications. mcmax 173 11-15 i5 Convergence criterion for purification (P). midemp 174 16-20 i5 Convergence criterion for purification (PP). mpurif 175 21-25 i5 CG cycle where purification starts. mlroot 176 26-30 i5 Choice of root for the CG density update. mcgpre 177 31-35 i5 Preconditioning of CG gradient matrix. mcgupd 178 36-40 i5 Choice of update for search direction. mpscal 179 41-45 i5 Scaling of intermediate density matrices. mcutau 180 46-50 i5 Choice of units for cutoffs. mcutm 181 51-55 i5 Cutoff for intermediate products. mcutf 182 56-60 i5 Cutoff for Fock matrix. mcutp 183 61-65 i5 Cutoff for density matrix. mcut1 184 71-72 i5 Cutoff for one-electron integrals (eV). mcut2 185 73-74 i5 Cutoff for two-electron integrals (eV). mcutr 186 75-80 i5 Cutoff for interatomic distances (A). Option Full description maxcg Maximum number of conjugate gradient (CG) cycles during density matrix search. Default 2. maxpur Maximum number of McWeeny purifications during one CG cycle. Default 2. mcmax Convergence criterion for purification. Maximum allowed change of diagonal density matrix elements: pmcmax=10**(-mcmax). Default for mcmax.le.0: Ignore criterion, use pmcmax=pcgmax/2 where pcgmax is the corresponding global criterion. midemp Convergence criterion for purification. Maximum allowed violation of idempotency for diagonal density matrix elements: pidemp=10**(-midemp). Default for midemp.le.0: pidemp=1, i.e. the criterion is effectively ignored. mpurif CG cycle where purification starts. = n Purification starting at CG cycle n. = 0 Use default value of mpurif=99. =-1 Purification turned off, activated automatically only when the CG search approaches convergence (as measured by the magnitude of the CG update). *** For mpurif.ge.0, a purified density matrix is always used, available either from a single transformation (before CG cycle n) or from repeated transformations (thereafter). *** For mpurif.gt.maxcg, purification will be turned on in the last CG cycle (maxcg) even if CG convergence is not reached. *** For mpurif=-1, the linear CG update for the density matrix is used as long as the purification has not been activated. Option mpurif=-1 is NOT recommended. mlroot Choice of root for the CG density update. The step size for the CG density update is given by the root of a quadratic equation. The root-finding algorithm is as follows: (a) Check whether linear term dominates such that the solution of the linear equation can be adopted. (b) Reject any physically unacceptable root of the quadratic equation with any diagonal density matrix element below 0 or above 2. (c) Compute the functional value for both roots of the quadratic equation and adopt the lower root. Option mlroot controls the first step. = n Check (a) is done, and the solution of the linear equation is adopted if the absolute value of x in the term sqrt(1+x) of the quadratic equation is smaller than the threshold: xsqmax = 10**(-mlroot). Errors will then be of the order x**2. = 0 Use default value of mlroot=5. =-1 Check (a) is not done. Step size determined from (b)-(c). mcgpre Preconditioning of CG gradient matrix. = 0 Not used. = 1 Diagonal preconditioning applied. mcgupd Choice of update for search direction. = 0 Polak-Ribiere formula for CG. = 1 Fletcher-Reeves formula for CG. = 2 Hestenes-Stiefel formula for CG. = 3 Davidon-Fletcher-Powell update. mpscal Scaling of intermediate density matrices to enforce normalization which may be lost due to purification. = 0 No such scaling. = 1 Restore correct trace of the density matrix after each CG cycle by adding a constant to each diagonal element. = 2 Restore correct trace of the density matrix after each CG cycle by scaling each diagonal element. = 3 Analogous to mpscale=2, but apply the scaling to the complete matrix. *** Implemented only for lin4=1,2,3. mcutau Choice of units for cutoffs. = 0 Use eV for energies (default). = 1 Use atomic units for energies (rather than eV). *** Not yet implemented. mcutm Cutoff for intermediate matrices in sparse matrix code. =-1 No such cutoff. = 0 Use default value of mcutm=100. = n Use cutoff for Fock matrix divided by n. mcutf Cutoff for Fock matrix (eV). =-1 No such cutoff. = 0 Use default value of mcutf=20. = n Cutoff 10**(-n). *** May be superseded by input for fcutf on second line. mcutp Cutoff for density matrix. =-1 No such cutoff. = 0 Use default value of mcutp=20. = n Cutoff 10**(-n). *** May be superseded by input for fcutp on second line. mcut1 Cutoff for one-electron integrals (eV). =-1 No such cutoff. = 0 Use default value of mcut1=20. = n Cutoff 10**(-n) eV. *** May be superseded by input for fcut1 on second line. *** Not yet implemented. mcut2 Cutoff for two-electron integrals (eV). =-1 No such cutoff. = 0 Use default value of mcut2=20. = n Cutoff 10**(-n) eV. *** May be superseded by input for fcut2 on second line. *** Not yet implemented. mcutr Cutoff for interatomic distances (Angstrom). =-1 No such cutoff. = 0 Use default value of mcutr=10000. = n Cutoff of 0.1*n Angstrom. Two-center integrals are not computed if the corresponding distance exceeds the cutoff. *** May be superseded by input for fcutr on second line. *** Not yet implemented. ***** Second line (inp24>1) ***** The integer-based cutoffs defined on the first line may be replaced individually by explicit input of the real-valued cutoffs. Option No. Columns Format Short description fcutf 20 1-10 g10.5 Cutoff for Fock matrix (eV). fcutp 21 11-20 g10.5 Cutoff for density matrix. fcut1 22 21-30 g10.5 Cutoff for one-electron integrals (eV). fcut2 23 31-40 g10.5 Cutoff for two-electron integrals (eV). fcutr 24 41-50 g10.5 Cutoff for interatomic distances (Angstrom). If a given input value is zero, the corresponding integer-based cutoff from the first line is used. If a given input value is negative, the corresponding absolute value is used in atomic units (rather than eV or Angstrom). 2.5 ***** Integral evaluation and SCF treatment ***** inp25>0 **** IN PREPARATION ... PRELIMINARY ... PARTLY IMPLEMENTED In standard applications, this input section may be skipped. The program will then use reasonable default values for all options. Specific input can be requested by the user (inp25>0). One line of input is expected for inp25=1. Two lines of input are expected for inp25=2. 16 INTEGER options (first line), and 8 REAL options (second line). ***** First line ***** Option No. Columns Format Short description imode 211 1-5 i5 Handling of two-electron MNDO-type integrals. inout 212 6-10 i5 Storage of data during SCF treatment. ivbse 213 11-15 i5 Interface to valence bond treatment. ivbovr 214 16-20 i5 Option for overlap in VB treatment. ifermi 215 21-25 i5 Electronic temperature for Fermi smearing. nfloat 216 21-25 i5 Number of orbitals with floating occupation. ndocc 217 26-30 i5 Number of doubly occupied (non-floating) MOs. iop218 218 36-40 i5 Not used presently. iop219 219 41-45 i5 Not used presently. iop220 220 46-50 i5 Not used presently. icutzs 221 51-55 i5 GTO integrals: range of small arguments. icutsm 222 56-60 i5 GTO integrals: range of medium arguments. icutml 223 61-65 i5 GTO integrals: range of large arguments. maxfmt 224 66-70 i5 GTO integrals via interpolation. limfmt 225 71-75 i5 GTO integrals via asymptotic formulas. ihcorr 226 76-80 i5 Option for H bonding corrections. Option Full description imode Handling of two-electron MNDO-type integrals (iop.le.0) and choice of subroutine for fock matrix. *** Option for testing and debugging. =-n Integrals are stored in memory as a matrix - no alternative storage is considered if memory is too small. Fock matrix from subroutine fock if vector length lm6.ge.n, otherwise fock matrix from subroutine fockx. =-2 Integrals are stored in memory as a matrix - no alternative storage is considered if memory is too small. Fock matrix from subroutine fock. =-1 Integrals are stored in memory as a matrix - no alternative storage is considered if memory is too small. Fock matrix from subroutine fockx. = 0 Integrals are stored in memory as a matrix by default. However, if memory is too small, a treatment using imode=1 is attempted first. If memory is still too small, the integrals are handled via imode=10. Fock matrix from the most appropriate subroutine as determined internally: - fock or fockx for imode=0, - fock1/fock2 for imode.gt.0. = 1 Integrals are stored in memory as a linear array of unique integrals which requires about half the buffer of the storage as a matrix. Stop if memory is too small. Fock matrix from subroutines fock1 and fock2. = n Disk input/output of two-electron integrals is enforced using a buffer of 512*n words (n.gt.1). Fock matrix from subroutines fock1 and fock2. *** Option imode=0 is recommended since it will choose the *** optimum treatment for a given memory size. inout Storage of data during SCF treatment. *** Option for testing and debugging. =-1 All relevant matrices are stored in memory. Data may be recomputed to avoid memory contention. Presently, several specific OM2 matrices are recomputed. = 0 All relevant matrices are stored in memory. If memory is too small, the program will attempt a treatment using first inout=1, then inout=2. = 1 The density matrix is stored on disk. Several specific OM2 matrices are also stored on disk. = 2 The core hamiltonian matrix, the fock matrix, and the difference density matrix are stored on disk. *** Option imode=0 is recommended since it will choose the *** optimum treatment for a given memory size. Note that the memory requirements can be also reduced by certain other input options: - nstart=-1 turns off the extrapolation in the SCF iterations which may be tolerable in many applications. - ifast=2 and idiag=1 enforce the use of slower diagonalization routines which is, however, not recommended. ivbse Interface to valence bond (VB) program. = 0 Do not use the VB interface. = 1 Provide integrals and other relevant input for a VB treatment carried out with a separate VB program. Write these data on file nb3 and continue the run. See code for contents of file nb3. =-1 Provide integrals and other relevant input for a VB treatment carried out with a separate VB program. Write these data on file nb3 and stop thereafter. See code for contents of file nb3. ivbovr Option for overlap in VB treatment. = 0 Basis functions are assumed to be orthogonal. Use unit matrix for overlap in VB treatment. Use original one-electron and two-electron integrals. = 1 Basis functions are assumed to be non-orthogonal. Use full overlap matrix in VB treatment. Use original one-electron and two-electron integrals. = 2 Basis functions are assumed to be non-orthogonal. Use full overlap matrix in VB treatment. Use transformed one-electron integrals (OAO -> AO). Use original two-electron integrals. = 3 Basis functions are assumed to be non-orthogonal. Use full overlap matrix in VB treatment. Use transformed one-electron integrals (OAO -> AO). Use transformed two-electron integrals (only NDDO). = 4 Basis functions are assumed to be non-orthogonal. Use full overlap matrix in VB treatment. Use transformed one-electron integrals (OAO -> AO). Use transformed two-electron integrals (all, N**4). ifermi Electronic temperature (K) for Fermi smearing (abs(iuhf)=5). For ifermi.le.0 the default value of 20000 will be used. nfloat Number of molecular orbitals with floating occupation for floating occupation number SCF by Granucci and Toniolo (iuhf=-6). For nfloat.le.0, the value will default to all orbitals. Combination with GUGA-CI (kci=5) is mandatory to obtain a physically meaningful electronic energy. The defaults will be ici1=nfloat and ici2=0. ndocc For floating occupation number SCF by Granucci and Toniolo (iuhf=-6), the number of doubly occupied molecular orbitals with fixed (non-floating) occupation. Default (nelec-nfloat)/2. icutzs Evaluation of error function F(m,t) for small arguments. = 0 Use F(m,0) only for zero argument t=0. = n Use F(m,0) for arguments t < 10**(-n). *** The default is recommended. itolfm Evaluation of error function F(m,t) for medium arguments. Desired accuracy computed from the asymptotic expansion which is applied for arguments t > cutsm. = n Target accuracy 10**(-n), cutsm selected internally by testing values between 20 and 10. = 0 Use default itolfm=9 and cutsm=10. *** Option itolfm determines the choice of cutsm *** which may only adopt values between 10 and 20. *** The default is recommended. icutml Evaluation of error function F(m,t) for large arguments. = n Use simple asymptotic formula for t > icutml. = 0 Use default icutml=42. *** The default is recommended. maxfmt Evaluation of error function F(m,t) via interpolation. = n Number of precomputed interpolation points. = 0 Use default maxfmt=400. *** Maximum value allowed in this version: maxfmt=400. *** Choose maxfmt < 0 to avoid interpolation: *** this will enforce explicit integral computation *** which is more accurate and more expensive. *** The recommended default generates 400 equidistant *** interpolation points between t=0 and t=19.95. limfmt Asymptotic expressions for Gaussian integrals. = n Use such expressions for arguments x > n. = 0 Use default limfmt=20. *** Choose limfmt=1000 to avoid such expressions: *** this will enforce explicit integral computation *** which is more accurate and more expensive. ihcorr Option for hydrogen bond correction = 0 Do not include hydrogen bond corrections. = 1 Include hydrogen bond correction H4 from Hobza (see JCTC 8, 141 (2012)). 3.1 ***** Title line for the molecule *************************** Option No. Columns Format Short description * kharge 65 1-2 i2 Molecular charge. * imult 66 3-4 i2 Definition of multiplicity. ktrial 67 5-6 i2 Initial density matrix for SCF. * kgeom 68 7-8 i2 Geometry input for reaction paths and grids. ipubo 69 9-10 i2 Option for saving SCF results on file. iuhf 70 11-12 i2 Type of SCF treatment (RHF/UHF). * kitscf 71 13-16 i4 Maximum number of SCF iterations. * nprint 72 17-18 i2 Printing flag for SCF. ifast 73 19-19 i1 Option for pseudo-diagonalizations in SCF. idiag 74 20-20 i1 Option for standard diagonalizations in SCF. * ksym 75 21-22 i2 Input of symmetry conditions. numsym 76 23-24 i2 Symmetry number for partition function. * kci 77 25-26 i2 Choice of correlation treatment. nstart 78 27-28 i2 First SCF cycle with extrapolation. nstep 79 29-30 i2 SCF extrapolation every nstep SCF cycles. * ktitle - 31-78 a48 Title for the molecule. Option Full description kharge Molecular charge. =99 To terminate the job. imult Definition of multiplicity. *** Options for RHF calculations. = 0 Closed-shell singlet. = 1 Open-shell singlet with two singly occupied orbitals. This usually corresponds to an excited singlet state. = 2 Doublet. = 3 Triplet. *** Options for UHF calculations. = 0 Singlet. = 1 Singlet (same as imult=0). = 2 Doublet. = 3 Triplet. = 4 Quartet (etc). *** Note on RHF and UHF calculations. By default, RHF for imult=0 and UHF for imult.gt.0 which may be changed using option iuhf (see below). imult.gt.3 is possible for UHF only. ktrial Initial density matrix for SCF. = 0 Standard diagonal matrix. The electrons are distributed such that each atom is initially neutral. For main-group elements, electrons are distributed evenly over s and p orbitals without populating d AOs. For transition-metal elements, only s and d AOs are populated. = 1 Simplified diagonal matrix. The electrons are distributed evenly over all orbitals. = 2 Modified diagonal matrix. The electrons are distributed such that each atom is initially neutral. They are then distributed evenly over the available orbitals at each atom. Equivalent to ktrial=0 for an sp basis. =11 Density matrix will be read in from file nb11. =12 Eigenvectors will be read in from file nb12 to compute density matrix. =13 RHF density matrix will be read in from file nb11 to form initial alpha and beta UHF density matrices. =20 From standard SCF calculation without external charges, external fields or COSMO solvation terms, using the chosen semiempirical method (iop). =21 Analogous to ktrial=20 using MNDOC. =22 Analogous to ktrial=20 using AM1. =23 Analogous to ktrial=20 using MNDO/H =24 Analogous to ktrial=20 using MNDO. =25 Analogous to ktrial=20 using OM1. =26 Analogous to ktrial=20 using OM2. =27 Analogous to ktrial=20 using PM3. =30 Block-diagonal density matrix from separate RHF-SCF calculations on user-defined fragments (section 3.5) using the same convergence criteria as in the molecular calculation. Some variables are fixed internally: inout=0, iuhf=-1, ifast=2, nstart=-1. =31 Same as ktrial=30, only 1 SCF cycle. =32 Same as ktrial=30, only 2 SCF cycles. =33 Same as ktrial=30, only 3 SCF cycles. =34 Same as ktrial=30, only 4 SCF cycles. =35 Same as ktrial=30, only 5 SCF cycles. =36 Same as ktrial=30, only 6 SCF cycles. =37 Same as ktrial=30, only 7 SCF cycles. =38 Same as ktrial=30, only 8 SCF cycles. =39 Same as ktrial=30, only 9 SCF cycles. *** Molecule-specific options ktrial=30-39 prevail over *** the general option linfrg=0 (see section 2). =41 Use initial guess charges from file CHR.dat to accelerate convergence of SCC-DFTB electron density optimizations. Use of any other value of ktrial>0 is disabled with SCC-DFTB (IOP=5,6) and terminates execution of MNDO99. kgeom Geometry input for reaction paths and grids. = 0 Standard. = 1 Additional input for reaction path. See chapter 3.4 for more information. = 2 Additional input for reaction grid. See chapter 3.4 for more information. = 3 Additional input for reaction grid. See chapter 3.4 for more information. = 4 Additional input for interpolating on pathways. See chapter 3.4 for more information. =-1 Program terminates after computing coordinates and distances (useful for checking the input data). ipubo Save SCF results on file. = 0 Do not save. = 1 Save density matrix on file nb11. = 2 Save eigenvectors on file nb12. = 3 Save density matrix on file nb11 and eigenvectors on file nb12. iuhf Type of SCF calculation. =-N Always RHF. = 0 Without correlation treatment (abs(kci).eq.0): - RHF for closed-shell systems, - UHF for open-shell systems. With correlation treatment (abs(kci).gt.0): - RHF for closed-shell systems, - half-electron ROHF for open-shell systems. = N Always UHF. The input value of N=abs(iuhf) may be used to define the occupations of the molecular orbitals. N=0 Standard occupations (implicitly). N=1 Standard occupations (implicitly). N=2 Standard occupations (explicitly), only useful for testing the code employing occupation numbers. N=3 Occupation numbers are read in for each irreducible representation. See chapter 3.6 for details. N=4 Average occupation numbers are defined internally for quasi-degenerate levels (see subroutine OCCDEF). The user may also consider to provide corresponding own code in subroutines OCCDEF and/or HECOR for an explicit definition of occupation numbers. *** Experimental option *** *** Not recommended for general use *** N=5 Determine fractional occupation numbers using the Fermi smearing technique. The default electronic temperature is 20000 K. It can be changed using option ifermi. N=6 Fractional occupation numbers are determined using the floating occupation number SCF procedure by Granucci and Toniolo, see CPL 325, 79 (2000). nfloat and ndocc may be used to select the number of orbitals with floating and fixed occupation. The default of 5.442 eV (0.2 au) for the half-width of the Gaussians may be changed using option domega which sets the half-width in eV. To obtain a physically meaningful electronic energy, a full CI calculation including all orbitals with floating occupation in the active space is required. By default, the program sets the options to perform a full CI calculation: ici1=nfloat and ici2=0. *** Not implemented for UHF (do not use iuhf=6). *** Available for RHF (please use iuhf=-6 and kci=5). For N > 1, the program uses ifast=2 internally (see below) which slows down the calculation. Hence, N > 1 should be avoided in standard applications. kitscf Maximum number of SCF iterations. Default 200. =-1 No SCF iterations: perform one energy evaluation using a given density matrix which needs to be provided in a suitable manner (for example via ktrial=11, see above). A subsequent gradient calculation is also possible. MO eigenvectors are not computed so that half-electron and CI calculations are not feasible for ktrial=-1. This option may be useful in QM/MM treatments. =-2 No SCF iterations: perform one energy evaluation using given MO eigenvectors which need to be provided in a suitable manner (for example via ktrial=12, see above). A subsequent gradient calculation is also possible. Half-electron and CI calculations can be carried out. This option may be useful in QM/MM treatments. =-3 No SCF iterations: evaluate only energy and gradient contributions arising from the electrostatic coupling of the QM system to external MM point charges, using a given density matrix which needs to be provided in a suitable manner (for example via ktrial=11, see above). This option is useful only in QM/MM treatments. nprint Printing flag for SCF. =-5 Prints no SCF information. =-1 Prints eigenvalues, SCF energies, net charges, and dipole moment. = 0 Prints eigenvalues, eigenvectors, SCF energies, symmetry labels, net charges, and dipole moment. = 1 Also prints density matrix, and UHF spin densities. = 2 Also prints SCF iterations, core hamiltonian, and final fock matrix. For nsav16.gt.0 (see above) also prints MOPAC-style bond orders and Mulliken populations. = 5 Debug printing. For nprint.gt.0, the eigenvalues and eigenvectors are printed in a format which is suitable for 80-column screens, at the expense of more lines of output. ifast Fast diagonalizations in SCF. = 0 Allowed (whenever possible). = 1 Allowed after initial full diagonalizations. = 2 Not allowed (for ifast.gt.1). idiag Standard diagonalizations in SCF. See documentation for details. = 0 In most program versions treated as idiag=9 (evvrsp). = 1 Using subroutine tdiag for linear fock matrix (eispack), calls to tred3,tql2,trbak3. = 2 Using subroutines tred2 and tql2 for square fock matrix (eispack). = 3 Using subroutine dspev for linear fock matrix (lapack). = 4 Using subroutine dspevx for linear fock matrix (lapack). = 5 Using subroutine dsyev for square fock matrix (lapack). = 6 Using subroutine dsyevx for square fock matrix (lapack). = 7 Using subroutine dspevd for linear fock matrix (lapack). = 8 Using subroutine dsyevd for packed fock matrix (lapack). = 9 Using subroutine evvrsp for square fock matrix (eispack-based), calls to tred1,eqlrat,einvit,trbak1. ksym Input of symmetry conditions. = 0 No symmetry conditions. = 1 Read symmetry conditions. See chapter 3.3 for more information. numsym Symmetry number for calculation of thermodynamic properties. = 0 Automatic determination of the symmetry number by the program. This works for most point groups, but may fail in some complicated cases (e.g. Dn, Dnd with even n). Explicit input is then required. = 1 C1,Ci,Cs,C0v. = 2 C2,C2v,C2h,D0h. = 3 C3,C3v,C3h,S6. = 4 C4,C4v,C4h,D2,D2d,D2h. = 5 C5,C5v,C5h. = 6 C6,C6v,C6h,D3,D3d,D3h. = 8 D4,D4d,D4h. =10 D5,D5d,D5h. =12 D6,D6d,D6h,T,Td. =24 Oh. kci Correlation treatment. = 0 None. = 1 Minimal configuration interaction involving two RHF MOs. See chapter 3.7 for more information. = 2 Brillouin-Wigner perturbation method with one main configuration. BWEN. See chapter 3.8 for more information. = 3 Brillouin-Wigner perturbation method with two main configurations. BWEN1. See chapter 3.8 for more information. = 4 Brillouin-Wigner perturbation method with two main configurations. BWEN2. See chapter 3.8 for more information. = 5 General configuration interaction based on the graphical unitary group approach (GUGA-CI). See chapter 3.9 for more information. = 6 Configuration interaction with single excitations and spin-flip CIS. Dedicated module for use with large active MO spaces. See chapter 3.10 for more information. = 7 Spin-adapted spin-flip CIS methods (including SF-XCIS, i.e. spin-flip extended CIS) using dedicated code. See chapter 3.10 for more information. = 8 Random phase approximation (RPA) method using dedicated code. See chapter 3.10 for more information. For negative kci, geometry optimizations are done at the SCF level, followed by a single correlated calculation at the final geometry, according to abs(kci). nstart First SCF cycle with extrapolation or damping of density matrix. =-1 No extrapolation or damping. The program also sets nstart=-1 if nstart.gt.kitscf. = 0 Use nstart=4 by default, unless DIIS is requested (idiis.gt.0) which leads to nstart=-1 by default. = n Start extrapolation or damping in the n-th SCF cycle, even if DIIS is requested. nstep Choice between extrapolation or damping and input of additional information. = n Perform SCF extrapolations every n SCF cycles (after cycle nstart). =-n Perform damping in every SCF cycle starting with SCF cycle nstart and using the damping factor n/10. = 0 Use nstep=4 by default. Extrapolation is selected by default. Damping requires input for nstep. title Title for the molecule. 3.2 ***** Molecular geometry ***** One line per atom ************* The input in this section is either formatted (option iform=0) or in free format (option iform.gt.0). The column numbers and the formats given below refer to the option iform=0. Alternatively, it is possible to read a formatted PDB file (option ingeom=1, see description at the end of this chapter). Depending on the value of variable igeom (see first line), the geometry may be defined in internal coordinates or in cartesian coordinates. The geometrical data are stored in a(j,i) where i is the number of the atom and j the type of coordinate as given in the following table. j internal coordinate cartesian coordinate 1 bond length x-coordinate 2 bond angle y-coordinate 3 dihedral angle z-coordinate In the case of internal coordinates, each atom i is defined with respect to three reference atoms na(i),nb(i),nc(i) as follows. 1 bond length i-na(i) 2 bond angle i-na(i)-nb(i) 3 dihedral angle i-na(i)-nb(i)-nc(i) For the first three atoms, there are special conventions. Atom 1 is put into the origin, no geometry input needed. Atom 2 is put on the positive x-axis, bond length 2-1 as input. Atom 3 is put into the xy-plane with positive y-coordinate, bond length 3-na(3) and bond angle 3-na(3)-nb(3) as input, with default values na(3)=2 and nb(3)=1. In the case of cartesian coordinates, reference atoms are not needed for the definition of the geometry. Cartesian coordinates and bond lengths are given in Angstrom, and angles in degree. Variable Columns Format Description nat(i) 1-2 i2 Atomic number of atom (i). =99 for a dummy atom which only assists in the definition of the geometry. =86 for a specially parametrized link atom called connection atom and used in QM/MM hybrid treatments; currently available for MNDO, MNDO/d, AM1 and PM3 (published) as well as for OM1 and OM2 (unpublished). = 0 to end input of geometry. a(1,i) 11-20 f10.5 First coordinate. la 23-24 i2 Optimization of first coordinate. = 0 a(1,i) is not optimized. = 1 a(1,i) is optimized. a(2,i) 31-40 f10.5 Second coordinate. lb 43-44 i2 Optimization of second coordinate. = 0 a(2,i) is not optimized. = 1 a(2,i) is optimized. a(3,i) 51-60 f10.5 Third coordinate. lc 63-64 i2 Optimization of third coordinate. = 0 a(3,i) is not optimized. = 1 a(3,i) is optimized. na(i) 71-72 i2 Number of first reference atom. nb(i) 73-74 i2 Number of second reference atom. nc(i) 75-76 i2 Number of third reference atom. End input of geometry by nat(i).le.0. In the case of formatted input a blank line may be used for this purpose. NOTE: la = 0 for the first atom in internal coordinates. lb = 0 for the first and second atom in internal coordinates. lc = 0 for the first, second and third atom in internal coordinates. ***** Alternative input from PDB file ***** ingeom=1 ************ It is possible to read a standard PDB file as part of a standard input file. Only the data on the ATOMS records are evaluated. The following information from these records is processed. record 1-6 a6 Identifier. 'ATOM ' Read this record. 'END ' Terminate PDB input. For any other identifier, skip this line. res 18-20 a3 Residue name. ires 23-26 i4 Residue sequence number. x 31-38 f8.3 Cartesian x coordinate (Angstrom). y 39-46 f8.3 Cartesian y coordinate (Angstrom). z 47-54 f8.3 Cartesian z coordinate (Angstrom). elem 77-78 a2 Element symbol, right-justified. One END record must be present at the end for proper termination. The PDB input provides data for sections 3.2 and 3.5 (see below). When using the PDB input option, it is currently not possible to read symmetry data (3.3) or reaction path data (3.4). 3.3 ***** Symmetry data ***** ksym=1 ***************************** The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. Symmetry may be imposed by specifying a reference atom L1, a symmetry relation number L2, and up to 10 dependent atoms L3. The symmetry relation number L2 defines the coordinates a(j,L3) of the dependent atoms in terms of the coordinate a(k,L1) of the reference atom. The type of the coordinates involved is defined implicitly by the symmetry relation number L2 (see below). Notation: j internal coordinate cartesian coordinate 1 bond length x-coordinate 2 bond angle y-coordinate 3 dihedral angle z-coordinate There are 33 predefined symmetry relations available, for L2 values between 1 and 33, which are listed below. Variable Columns Format Description L1 1-2 i2 Number of the reference atom. L2 3-5 i3 Number of symmetry relation. = 1 implies a(1,L3)= a(1,L1) = 2 implies a(2,L3)= a(2,L1) = 3 implies a(3,L3)= a(3,L1) = 4 implies a(3,L3)= 90-a(3,L1) = 5 implies a(3,L3)= 90+a(3,L1) = 6 implies a(3,L3)= 120-a(3,L1) = 7 implies a(3,L3)= 120+a(3,L1) = 8 implies a(3,L3)= 180-a(3,L1) = 9 implies a(3,L3)= 180+a(3,L1) =10 implies a(3,L3)= 240-a(3,L1) =11 implies a(3,L3)= 240+a(3,L1) =12 implies a(3,L3)= 270-a(3,L1) =13 implies a(3,L3)= 270+a(3,L1) =14 implies a(3,L3)=-a(3,L1) =15 implies a(1,L3)= a(1,L1)*0.5 =16 implies a(2,L3)= a(2,L1)*0.5 =17 implies a(2,L3)= 180-a(2,L1) =18 implies a(1,L3)= a(1,L1)*depfac =19 implies a(1,L3)= a(1,L1)*0.763932022 =20 implies a(1,L3)= a(1,L1)/sqrt(2.0) =21 implies a(1,L3)=-a(1,L1) =22 implies a(2,L3)=-a(2,L1) =23 implies a(3,L3)=-a(3,L1) =24 implies a(2,L3)= a(1,L1) =25 implies a(1,L3)= a(2,L1) =26 implies a(3,L3)= a(1,L1) =27 implies a(1,L3)= a(3,L1) =28 implies a(3,L3)= a(2,L1) =29 implies a(2,L3)= a(3,L1) =30 implies a(2,L3)= 180+a(2,L1) =31 implies a(3,L3)= a(3,L1)*0.5 =32 implies a(3,L3)= a(3,L1)*2.0 =33 implies a(3,L3)= 120-a(3,L1)*2.0 The most useful symmetry relations are L2=1,2,3,14. The relations L2=4-13 are used only for internal coordinates, and L2=24-29 only for cartesian coordinates. L3(1) 20-22 i3 Number of dependent atom. L3(2) 25-27 i3 Number of dependent atom. L3(3) 30-32 i3 Number of dependent atom. L3(4) 35-37 i3 Number of dependent atom. L3(5) 40-42 i3 Number of dependent atom. L3(6) 45-47 i3 Number of dependent atom. L3(7) 50-52 i3 Number of dependent atom. L3(8) 55-57 i3 Number of dependent atom. L3(9) 60-62 i3 Number of dependent atom. L3(10) 65-67 i3 Number of dependent atom. For definition of n dependent atoms, L3(1) up to L3(n) must be nonzero. End input of symmetry by L1.le.0 or by a blank line. The symmetry relation L2=18 requires additional input of the factor depfac, immediately after the line containing L2=18. The input is in free format for iform.ne.0. depfac 1-10 f10.5 Factor to be used with L2=18 (see above). Only one such factor can be defined for a given molecule. 3.4 ***** Reaction path or grid ***** kgeom=1-4 ****************** The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. Option kgeom=1 selects a one-dimensional reaction path. Option kgeom=2 selects a two-dimensional reaction grid. Option kgeom=3 selects a two-dimensional reaction grid. Option kgeom=4 selects a linear interpolation between geometries. Comments on kgeom=2 and kgeom=3: The first grid variable is incremented in the outer loop, and the second one in the inner loop. The first grid variable always changes from its lower limit to its upper limit. For kgeom=2, this also holds for the second grid variable. For kgeom=3, the second grid variable changes from its lower to its upper limit for odd values of the outer loop counter, and in the opposite direction for even values of this counter. The transition from one outer loop to the next should be smoother for kgeom=3. Comments on kgeom=4: Two consecutive complete input files are required to define the two geometries that serve as the end points for the interpolation. Input files for optimized geometries can be generated by using the option nsav7 in geometry optimization runs. Input files for optimized Cartesian coordinates in the principical axis system are available from force constant runs via nsav7=7. Linear interpolation is applied between the two input structures to generate the coordinates for the in-between points. Single-point calculations are performed at these points. All options require additional input for the path/grid variables. *** First line for kgeom=1-3 *** Variable Columns Format Description l1 1-5 i5 Atom on which the reaction path variable or the first grid variable is located. l2 6-10 i5 Type of path/grid variable. l2 internal coord cartesian coord 1 bond length x-coordinate 2 bond angle y-coordinate 3 dihedral angle z-coordinate l3 11-15 i5 Number of points on the reaction path or along the first grid direction, in addition to the initial point. Maximum 199 for a path, kgeom=1. Maximum 24 for a grid, kgeom=2. step 16-24 f10.5 Step size for path/grid variable. The initial point rc(1) is available from the geometry input for a(l2,l1). The other points rc(i) with i=2,..,l3+1 are computed from rc(1)+(i-1)*step. In the case of step=0, explicit input for rc(i) with i=2,..,l3+1 is required on the following line(s). *** Second line for kgeom=1-3 (only for step=0) *** rc(i) 1-80 8f10.5 l3 values for the path/grid variable. Use more than one line for l3.gt.8. At this point, the input for a reaction path calculation is done. In the case of a two-dimensional grid calculation, an analogous input for the second grid variable follows. *** Only line for kgeom=4 *** l1 1-5 i5 Total number of points including the two end points. Default 11 (corresponding to ten steps of interpolation) lx 6-10 i5 Phase factor for x coordinate. Default 1. Only relevant when input files for the end points are optimized Cartesians in the principal axis system obtained from diagonalization (arbitrary phase). ly 10-15 i5 Phase factor for y coordinate. Default 1. lz 15-20 i5 Phase factor for z coordinate. Default 1. 3.5 **** Definition of fragments ***** ktrial=30-39 ************** The input in this section is formatted. It is used to generate a block-diagonal initial density matrix for large molecules, especially for CG-DMS calculations. It is needed if the following three conditions are satified: - linfrg.gt.0 or ktrial=30-39 (requests use of fragments) - inpfrg.ge.0 (requests input of fragments) - ingeom.ne.1 (input not available from PDB file) This input section is skipped if any of these conditions is not satisfied. *** First and following lines *** nfrags(i) 1-50 10i5 Number of fragment containing atom i. Use as many lines as necessary. *** Subsequent lines *** i 1-5 i5 Number of atom bearing a formal charge. = 0 End of this section of input. ndum 6-10 i5 Formal charge of atom i, nchrgs(i)=ndum. The array nchrgs is initialized to zero. Only nonzero values are needed from input. 3.6 ***** MO occupations ***** abs(iuhf)=3 *********************** This option has not yet been tested extensively and should only be used with caution (see the remarks in the documentation). The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. Occupation numbers can be defined for the following seven Abelian point groups (msub=1-7), with irreducible representations ordered as shown below. msub group order of irreducible representations 1 2 3 4 5 6 7 8 1 Cs A' A'' 2 C2 A B 3 C2v A1 A2 B1 B2 4 D2h Ag Au B1g B1u B2g B2u B3g B3u 5 C2h Ag Bg Au Bu 6 D2 A B1 B2 B3 7 Ci Ag Au *** First line *** msub 1-5 i5 Point group (see above). *** Second line *** mocca(i) 1-40 8i5 Occupation numbers for all irreducible (i=1,8) representations (ordered as above). RHF: Values for doubly occupied MOs. UHF: Values for alpha-spin MOs. *** Third line *** Only for cases that are not closed-shell *** moccb(i) 1-40 8i5 Occupation numbers for all irreducible (i=1,8) representations (ordered as above). RHF: Values for singly occupied MOs. UHF: Values for beta-spin MOs. Omit this line for closed-shell RHF. 3.7 ***** Configuration interaction ***** abs(kci)=1 ************* The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. Variable Columns Format Description k 1-5 i5 Number of MO involved in CI (see below). l 6-10 i5 Number of MO involved in CI (see below). nc 11-15 i5 Number of configurations involved in CI. Default values are defined below. Explicit input is possible only for singlets with imult=0 (see below). lroot 16-20 i5 CI state whose geometry is optimized. Default 1. Three types of minimal configuration interaction are possible which, in each case, involve two RHF MOs k and l. imult=0, singlet, closed-shell RHF MOs, 2*2 CI or 3*3 CI. k is an occupied MO (default HOMO). l is an unoccupied MO (default LUMO). Default. nc=3, 3*3 CI with configurations kk,ll,kl. Input of nc=2 leads to a 2*2 CI with configurations kk,ll. imult=1, singlet, open-shell half-electron RHF MOs, 3*3 CI. k and l are the two singly occupied RHF MOs. nc=3, 3*3 CI with configurations kk,kl,ll. These default values cannot be changed via input. imult=2, doublet, open-shell half-electron RHF MOs, 2*2 CI. Case a. k singly occupied MO, l unoccupied MO (default LUMO). Case b. l singly occupied MO, k occupied MO (default HOMO-1). Case a is the default case. 3.8 ***** Perturbation treatment ***** abs(kci)=2-4 ************** The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. *** First line *** Option No. Columns Format Short description ici1 131 1-4 i4 Number of active occupied orbitals. ici2 132 5-8 i4 Number of active unoccupied orbitals. ioutci 133 9-12 i4 Printing flag for perturbation section. movo 134 13-16 i4 Explicit definition of active orbitals. mpert 135 17-20 i4 Selection of perturbation treatment. jci1 151 21-24 i4 Total number of occupied pi orbitals. jci2 152 25-28 i4 Total number of unoccupied pi orbitals. pipop 153 29-32 i4 Population threshold to identify pi-MOs. Option Full description ici1 Total number of occupied orbitals in the active space. Default all, up to a maximum of 20. Higher ici1 values (ici1.gt.20) are not selected by default and must be read in. ici2 Total number of unoccupied orbitals in the active space. Default all, up to a maximum of 20. Higher ici2 values (ici2.gt.20) are not selected by default and must be read in. ioutci Printing flag for perturbation section. =-5 No output. = 0 Standard output. = 5 Debug print. movo Definition of orbitals involved in the active space. Default 0. = 0 Use ici1 highest occupied orbitals and ici2 lowest unoccupied orbitals. = 1 Read orbital numbers on extra lines (see below). *** Options movo=-1,-2,-3 are designed for correlating the *** pi electrons, with automatic definition of the pi-MOs. =-1 Include pi-MOs in active space, px corresponds to pi-AO. For details see description in GUGA section 3.9. =-2 Include pi-MOs in active space, py corresponds to pi-AO. For details see description in GUGA section 3.9. =-3 Include pi-MOs in active space, pz corresponds to pi-AO. For details see description in GUGA section 3.9. =-4 Include MOs with highest d-population in active space. For details see description in GUGA section 3.9. *** Option movo=-4 is designed for correlating the d electrons *** in transition metals, with automatic definition of d-MOs. mpert Selection of second-order perturbation treatment. Default 0 corresponds to mpert=1. = 0 BWEN treatment (see below). > 0 Four-digit option to choose perturbation treatment. Four perturbation treatments are available: i=1 RSMP, i=2 RSEN, i=3 BWMP, i=4 BWEN. RS Rayleigh-Schroedinger treatment. BW Brillouin-Wigner treatment. MP Moller-Plesset denominators. EN Epstein-Nesbet denominators. The digits 1-4 of mpert define the array ipert(i): ipert(1) = mod(mpert/1000,10) - column 17. ipert(2) = mod(mpert/100,10) - column 18. ipert(3) = mod(mpert/10,10) - column 19. ipert(4) = mod(mpert,10) - column 20. Conventions: ipert(i) > 0 : Perturbation treatment (i) performed. ipert(i) = 1 : Perturbation energy (i) added to total energy. Examples: = 2222 Evaluate all four perturbation energies. = 1000 Evaluate RSMP = MP2 energy and add to total energy. = 0001 Evaluate BWEN energy and add to total energy. *** Note: MNDOC calculations are normally performed with the *** default BWEN treatment (no need for explicit mpert input). jci1 Total number of occupied pi-MOs or d-MOs. For details see description in GUGA section 3.9. jci2 Total number of unoccupied pi-MOs or d-MOs. For details see description in GUGA section 3.9. pipop Population threshold to identify relevant MOs (movo < 0). For details see description in GUGA section 3.9. *** Second line (omit if movo.le.0) *** imoci(i) 1-80 20i4 Numbers of the active occupied orbitals. (i=1,ici1) The numbering refers to the SCF output. Use more than one line for input, if necessary. *** Third line (omit if movo.le.0) *** imoci(i) 1-80 20i4 Numbers of the active unoccupied orbitals. (i=ici1+1, The numbering refers to the SCF output. ici1+ici2) Use more than one line for input, if necessary. 3.9 ***** GUGA configuration interaction ***** abs(kci)=5 ******** The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. *** First line *** General options *** Option No. Columns Format Short description ici1 131 1-4 i4 Number of active occupied orbitals. ici2 132 5-8 i4 Number of active unoccupied orbitals. ioutci 133 9-12 i4 Printing flag for GUGA-CI. movo 134 13-16 i4 Explicit definition of active orbitals. mpert 135 17-20 i4 Not used here. Perturbation treatment. nciref 136 21-24 i4 Number of reference occupations. mciref 137 25-28 i4 Definition of reference occupations. levexc 138 29-32 i4 Maximum excitation level wrt any reference. iroot 139 33-36 i4 Total number of lowest CI states computed. lroot 140 37-40 i4 Defines the CI state of interest. cichg 141 41-44 i4 Total charge of CI state. multci 142 45-48 i4 Spin multiplicity of CI states. ncisym 143 49-52 i4 Symmetry of CI state treated. cidir 144 53-56 i4 Direct CI and algorithmic features. cidiag 145 57-60 i4 Diagonalization of CI Hamiltonian matrix. iuvcd 146 61-64 i4 Spectroscopic properties: UV, CD. imcd 147 65-68 i4 Spectroscopic properties: MCD. ipop 148 69-72 i4 Population analysis for GUGA-CI. ciplot 149 73-76 i4 Control plotting of Shavitt graphs. cilead 150 77-80 i4 Identify leading configurations for output. Option Full description ici1 Total number of occupied orbitals in the active CI space. Default 1 for imult=0 and imult=2, default 2 for imult=1 and imult=3. ici2 Total number of unoccupied orbitals in the active CI space. Default 1 for imult=0 and imult=2, default 0 for imult=1 and imult=3. ioutci Printing flag. =-10 No output. = 0 Small standard output. = 1 Also print energies and leading CSFs of all states. = 2 Also print execution times for the different GUGACI stages and some information on the diagonalization, e.g. Davidson iterations. = 3 Also print CSFs as linear combinations of Slater determinants and complete eigenvectors of all states. = 4 Also print CI Hamiltonian. = 5 Also print some debug information. = 6 Huge debug print including details of integral evaluation. movo Definition of orbitals involved in the active CI space. Default 0. = 0 Use ici1 highest occupied orbitals and ici2 lowest unoccupied orbitals. = 1 Read numbers of active orbitals. Details see below (third line). = 2 This option is no longer supported. Read pairs of orbital numbers to be interchanged for definition of the active CI space. = 3 Read numbers of active orbitals (third line) with a given symmetry (fourth line). *** Special options for specific systems. *** Options movo=-1,-2,-3 are designed for correlating the *** pi electrons, with automatic definition of the pi-MOs. =-1 Include pi-MOs in active space, px corresponds to pi-AO. First calculate sums of px populations to select pi-MOs. Use pipop, jci1, and jci2 to determine the relevant pi-MOs. Use ici1 highest occupied and ici2 lowest unoccupied pi-MOs. If ici1 or ici2 are greater than the number of occupied or unoccupied pi-MOs found, add the relevant number of highest occupied or lowest unoccupied sigma-MOs to the active space. =-2 Include pi-MOs in active space, py corresponds to pi-AO. First calculate sums of py populations to select pi-MOs. Use pipop, jci1, and jci2 to determine the relevant pi-MOs. Use ici1 highest occupied and ici2 lowest unoccupied pi-MOs. If ici1 or ici2 are greater than the number of occupied or unoccupied pi-MOs found, add the relevant number of highest occupied or lowest unoccupied sigma-MOs to the active space. =-3 Include pi-MOs in active space, pz corresponds to pi-AO. First calculate sums of pz populations to select pi-MOs. Use pipop, jci1, and jci2 to determine the relevant pi-MOs. Use ici1 highest occupied and ici2 lowest unoccupied pi-MOs. If ici1 or ici2 are greater than the number of occupied or unoccupied pi-MOs found, add the relevant number of highest occupied or lowest unoccupied-sigma MOs to the active space. *** Option movo=-4 is designed for correlating the d electrons *** in transition metals, with automatic definition of d-MOs. =-4 Include MOs with highest d-population in active space. First calculate sums of d populations to select the d-MOs. Use pipop, jci1, and jci2 to determine the relevant d-MOs. Use ici1 highest occupied and ici2 lowest unoccupied d-MOs. If ici1 or ici2 are greater than the number of occupied or unoccupied d-MOs found, add the relevant number of highest occupied or lowest unoccupied other MOs to the active space. *** Option movo=-5 is designed for selecting the active space *** in non-planar conjugated systems, by identifying the *** orbitals that most closely correspond to the pi-MOs of the *** system in its planar form. =-5 Include "pi-like" MOs in the active space. All atoms with p orbitals are assumed to form part of the conjugated system (unless the conjugated system is defined explicitly using the nconj keyword). For each atom, a local "pi plane" is generated based on the positions of its two nearest conjugated neighbours. The pi population for each atom is calculated along the normal vector to this plane. The relevant MOs are then determined using pipop, jci1 and jci2 in the same way as for movo=-1/-2/-3. mpert Option for perturbation treatment (see chapter 3.8). Not used here. nciref Number of reference occupations. = 0 None. Full CI in the active space. = n Chosen number, maximum 20. Reference configurations are generated automatically for a given occupation. There may be more than one configuration for open-shell reference occupations. mciref Definition of reference occupations. = 0 Chosen by default, no further input. This automatic selection is available only for: * nciref=1: SCF configuration, provided that the spin multiplicity is the same for the SCF and CI calculations. * nciref=2: SCF configuration and doubly excited HOMO-LUMO configuration, after closed-shell SCF treatment. * nciref=3: closed-shell, singly and doubly excited configurations for two active orbitals (kk, kl, ll), after closed-shell or open-shell singlet SCF treatment (k,l=HOMO,LUMO or singly occupied, respectively). = 1 Read occupancies of the orbitals in the active space. Details see below (fourth line). = 2 Read excitation indices relative to the SCF configuration and define the reference occupations accordingly. Numbering according to the SCF output. Details see below (fourth line). = 3 Starting from the reference occupations corresponding to mciref=0, add further references so that their fraction in all CI roots is at least 85%, and repeat the CI calculation once. = 4 Starting from the reference occupations read from the input (see mciref=1), add further references so that their fraction in all CI roots is at least 85%, and repeat the CI calculation once. levexc Maximum excitation level relative to any of the reference configurations. = 1 CIS, only single excitations. = 2 CISD, up to double excitations. = 3 CISDT, up to triple excitations. = 4 CISDTQ, up to quadruple excitations. = n Up to n-fold excitations. *** Default 2 for nciref.gt.0. *** Not used for nciref.eq.0. iroot Total number of lowest CI states computed. Default 1 (or iroot=lroot for lroot.gt.1). For iroot=-1 an additional input line is read with the requested number of states for each symmetry. lroot If positive, lroot is the number of the CI state of interest, e.g. during a geometry optimization. If negative, lroot labels the state in which the CSFs corresponding to reference (-lroot) have the largest coefficients (norm) compared to the other states. Default 1. cichg Total charge of CI state. = 0 Same charge as in SCF calculation. = n Value of charge. =9999 Zero charge. multci Spin multiplicity of CI states. =-1 Varying multiplicity (requires icross=1-5, ncigrd=2) = 0 Same multiplicity as specified by imult (default). = 1 Singlet. = 2 Doublet. = 3 Triplet. = n State with n-1 unpaired electrons. ncisym Symmetry of CI state that is treated. =-1 Do not use any symmetry during CI. = 0 Symmetry not specified. = n Irreducible representation of CI state. Conventions for these numerical labels are the same as in section 3.6. Restricted to Abelian point groups. n=1-2 for Cs , C2 , Ci. n=1-4 for C2v, C2h, D2. n=1-8 for D2h. *** WARNING: If derivatives are calculated numerically, the symmetry may temporarily be lowered and the selection of a state by symmetry may not work. cidir Direct CI and algorithmic features. = 0 Default. Treated as cidir=1. = 1 Keep coupling coefficients and CI Hamiltonian in memory. Coupling coefficients for all requested irreducible representations are computed by the shape-driven algorithm and stored in memory at the same time. The CI Hamiltonian is calculated and stored blockwise for negative iroot. = 2 Keep coupling coefficients and CI Hamiltonian in memory during CI for a requested irreducible representation. Coupling coefficients are computed by the shape-driven algorithm. For the computation of the density matrices for the analytic gradient and the spectroscopic properties, use the direct algorithm (cidir=3). For a non-negative iroot, cidir=2 is treated like cidir=1. = 3 Recompute coupling coefficients by the shape-driven algorithm as needed. =-N Like cidir=N, but coupling coefficients are computed by the loop-driven algorithm. cidiag Diagonalization for CI Hamiltonian matrix. The same choices are available as in the case of option idiag, see section 3.1. In addition, the Davidson diagonalizer can be selected (cidiag=10-15). = 0 Default, treated as cidiag=4 (dspevx) for in-core calculation and as cidiag=10 (Davidson) for direct CI. = 1 Using subroutine tdiag for linear CI matrix (eispack), calls to tred3,tql2,trbak3. = 2 Using subroutines tred2 and tql2 for square CI matrix (eispack). = 3 Using subroutine dspev for linear CI matrix (lapack). = 4 Using subroutine dspevx for linear CI matrix (lapack). = 5 Using subroutine dsyev for square CI matrix (lapack). = 6 Using subroutine dsyevx for square CI matrix (lapack). = 7 Using subroutine dspevd for linear CI matrix (lapack). = 8 Using subroutine dsyevd for packed CI matrix (lapack). = 9 Using subroutine evvrsp for square CI matrix (eispack-based), calls to tred1,eqlrat,einvit,trbak1. =10 Using original Davidson diagonalizer for sparse CI matrix: E. R. Davidson, J. Comp. Phys. 17, 87-94 (1975); P. D. Dacre, Theor. Chim. Acta 43, 197 (1976). The iroot lowest eigenstates of the chosen symmetry (ncisym) are computed. =11 Using Davidson diagonalizer for sparse CI matrix with the modifications by W. Butscher and W. E. Kammer, see J. Comp. Phys. 20, 313-325 (1976). *** For positive lroot: Compute only the lroot-th CI state of the chosen symmetry. *** For negative lroot: Choose the initial-guess vector among the initial eigenvectors such that the abs(lroot)-th reference configuration has the largest coefficient, and ignore the value of iroot. =12 Using an appropriate implementation of Davidson diagonalizer for sparse CI matrix with the modifications by B. Liu (see B. Liu, in: Numerical Algorithms in Chemistry: Algebraic Methods; C. Moler, I. Shavitt, Eds.; Lawrence Berkeley Laboratory, Berkeley (1978), 49-53). The iroot lowest eigenstates of the chosen symmetry (ncisym) are computed. *** CONVENTION FOR GPU COMPUTING AND IN-CORE CI: cidiag=12 is reset to cidiag=14 if GPU computing is supported by the distribution and a CUDA-capable GPU with a compute capability of at least 2.0 is detected. Otherwise, cidiag=12 is reset to cidiag=13. =13 Using Fortran implementation of Davidson diagonalizer with the modifications by B. Liu (see cidiag=12). This variant may be used with in-core or direct CI. =14 Using C++ implementation for GPU of Davidson diagonalizer with the modifications by B. Liu (see cidiag=12). This variant may be used with in-core CI only. =15 Using C++ implementation for CPU of Davidson diagonalizer with the modifications by B. Liu (see cidiag=12). This variant may be used with in-core CI only. *** GENERAL WARNING for cidiag=10-15: See comments for ncisym on symmetry lowering. iuvcd Evaluation of spectroscopic properties for UV and CD spectra. = 0 No such evaluation (default). = 1 Evaluate spectroscopic properties. Print permanent dipole moments of all states calculated. = 2 Also print oscillator strengths, rotational strengths, and transition moments for transitions originating from ground state. = 3 Also print electric and magnetic transition moments between excited states. = 4 Debug print. imcd Evaluation of spectroscopic properties for MCD spectra. = 0 No such evaluation (default). *** MCD not implemented in current version. ipop Population analysis for GUGA-CI wavefunction. = 0 No such evaluation (default). = 1 Population analysis for state lroot. = 2 Population analysis for all states calculated previously. ciplot Produce postscript graphics as output. = 0 No graphics. = 1 Plot Shavitt graph. = 2 As ciplot=1, in addition plot all configurations as directed walks. = 3 As ciplot=2, in addition plot all off-diagonal matrix elements as loops. *** Use the debug options ciplot=2,3 with care. Although individual *.ps-files are small, there are many of them. cilead Define threshold for printing coefficients c(i) of the leading configurations, in units of 0.0001. = 0 Use default value of 1000 (print c(i) if abs(c(i)).ge.0.1). *** Second line *** General options *** Option No. Columns Format Short description jci1 151 1-4 i4 Total number of occupied pi orbitals. jci2 152 5-8 i4 Total number of unoccupied pi orbitals. pipop 153 9-12 i4 Population threshold to identify pi-MOs. inatur 154 13-16 i4 Natural orbital analysis for GUGA-CI. ciselt 155 17-20 i4 Selection threshold for CI references. imomap 156 21-24 i4 Active orbitals in consecutive CI runs. icimap 157 25-28 i4 Reference CSFs in consecutive CI runs. keepci 158 29-32 i4 Keeping CI options for next molecule. ncigrd 159 33-36 i4 Number of CI gradients to be computed. icross 160 27-40 i4 Choice of multi-surface treatment. mindav 161 41-44 i4 Minimum dimension of Davidson subspace. maxdav 162 45-48 i4 Maximum dimension of Davidson subspace. kitdav 163 49-52 i4 Maximum number of Davidson iterations. nrmdav 164 53-56 i4 Davidson convergence criterion: norm of q. maxmap 165 57-60 i4 Maximum number of attempts to map MOs. mapthr 166 61-64 i4 Threshold for successful MO mapping. nconj 167 65-68 i4 Number of conjugated atoms for movo=-5. iciocc 168 69-72 i4 Number of occupied MOs for movo=4-6. icivir 169 73-76 i4 Number of unoccupied MOs for movo=4-6. ldroot 170 77-80 i4 CI state considered for movo=4,6. Option Full description jci1 Total number of occupied pi-MOs or d-MOs. = 0 Include all relevant MOs in the active space, up to a limit of ici1 (default). = n Include n relevant MOs in the active space (n.le.ici1) An MO is relevant if it has a pi population (movo=-1,-2,-3,-5) or d population (movo=-4) above the threshold defined by pipop. For jci1=n, the threshold may be reduced to ensure that there are n relevant MOs. jci2 Total number of unoccupied pi-MOs or d-MOs. = 0 Include all relevant MOs in the active space, up to a limit of ici2 (default). = n Include n relevant MOs in the active space (n.le.ici2) An MO is relevant if it has a pi population (movo=-1,-2,-3,-5) or d population (movo=-4) above the threshold defined by pipop. For jci2=n, the threshold may be reduced to ensure that there are n relevant MOs. pipop Population threshold to identify relevant MOs (movo < 0). = 0 Equivalent to n=4000 (default). = n Every orbital with a target population greater than n/10000 is considered to be relevant. *** If the number of relevant MOs is greater than the number requested according to jci1/jci2, the MOs are selected counting outwards from the HOMO/LUMO. *** If jci1/jci2 is greater than the number of relevant MOs, the pipop threshold is reduced until enough relevant MOs are found. If there are not enough MOs with non-negligible pi character, jci1/jci2 will be reset to the number available. inatur Natural orbital analysis for GUGA-CI wavefunction. = 0 No such evaluation (default). = 1 Natural orbital analysis for state lroot. = 2 Natural orbital analysis for all states calculated. = 3 Same as inatur=2, with debug print. ciselt Threshold for the automatic selection of CI reference configurations, in percent. Default 85, corresponding to a combined weight of the reference configurations of at least 85%. imomap Definition of active orbitals in consecutive CI runs, for example during geometry optimization or dynamics. =-1 Only track MO phase. = 0 Keep original definition of MO labels according to movo. = 1 Track MO character through an overlap criterion and adapt MO labels such that the orbitals retain their character as much as possible. *** Both options are accepted for numerical computation *** of derivatives. = 2 Track MO character using the same criterion as imomap=1, but reject optimization step if the overlap for any active orbital falls below a threshold defined by mapthr. The optimization routine will attempt a smaller step or halt as appropriate. *** imomap=2 is currently implemented for the optimizers defined by jop=0, ief.ge.0 (lsub=0 only for ief=0), including conical intersection searches using those optimizers. It is also implemented for excited state dynamics (icross=6), please see the ADAPT_MAP keyword in section K. = 3 Track MO character via file I/O (imomap.dat) for use with single-point calculations, usually when driven by external programs (e.g. ChemShell). The single-point job will fail if the overlap for any active orbital falls below a threshold defined by mapthr. If imomap.dat is not present or is inconsistent with the input file, the MOs will not be mapped, but a new file imomap.dat will still be saved at the end of the calculation. *** imomap=3 is available for jop<0. When using imomap=3, the options ici1, ici2, movo, nciref, mciref and ncisym should be consistent between runs. The active space definition should be left unchanged and will be overwritten by information from imomap.dat. icimap Definition of reference configurations in consecutive CI runs, for example during geometry optimization or dynamics. = 0 Keep original definition of CSFs according to movo. = n Update definition of reference CSFs gurations in each new CI run by applying the procedure specified for movo=n (n=4-6, see description of option movo). *** The current definition of CSFs is never changed during *** the numerical computation of derivatives. keepci Continued use of current CI options for the next input, for example in multiple calculations on the same molecule. =-1 Not done, define everything from scratch. = 0 Not done, define everything from scratch. = 1 Use current CI options again, especially with regard to the active orbitals and reference configurations. *** Experimental option, use at own risk. ncigrd Number of CI gradients to be computed. = 0 Calculate at most one CI gradient, for CI state lroot. Default case, no further input required. = n Compute CI gradients for n CI states which are specified on an extra line of input (seventh line). Minimum n=2, maximum n=4. *** Needed for surface crossings and conical intersections *** and ignored if icross=0. icross Choice of multi-surface treatment. = 0 No such treatment. = 1 Calculate energies and gradients for the states specified by ncigrd. = 2 As icross=1, but also calculate all interstate coupling gradients between the specified states. *** Options 1-2 are intended for single-point multi-surface *** calculations, in particular for interfacing with external *** optimisation/dynamics drivers. *** If desired they can also be used in conjuction with a *** single surface optimisation. The state to be optimised *** (specified by lroot) must be in the list of ncigrd gradients. = 3 Determine conical intersections between two CI states using the penalty function algorithm of Ciminelli et al. = 4 Determine conical intersections between two CI states using the gradient projection algorithm of Bearpark et al. = 5 Determine conical intersections between two CI states using the Lagrange-Newton algorithm of Yarkony. *** Options 3-5 are intended for use with the internal (MNDO99) *** optimization routines to locate conical intersections. *** For icross=3, the options ief=0, nrst=-1 and ldrop=100 *** are recommended. *** For icross=4, the default optimizer must be used (ief=0). *** For icross=5, ief>0 must be specified because the Yarkony *** optimizer is based on the pure Newton-Raphson routine *** available through the Eigenvector Following option. *** If icross=5 and ief=1, the options llamda=2 and *** lnoupd=1 are automatically chosen. = 6 Perform excited state dynamics with surface hopping using the semiclassical aproach of Tully and the included dynamics driver. *** For icross=6, further input is necessary in the form of *** a namelist "dynvar.in". = 7 As icross=2, but use the full expression to calculate the non-adiabatic coupling matrix elements between the specified states. mindav Minimum dimension of subspace for Davidson diagonalization. For cidiag=10 and cidiag=12-15, the mindav lowest energy configuration state functions are included in the initial set of trial vectors in addition to the reference configuration state functions. Default twice the number of requested CI roots, at most 30. *** WARNING: If symmetry is not used (ncisym.le.0) and mindav is chosen too small, some states may be missed. maxdav Maximum dimension of subspace for Davidson diagonalization. Default mindav+30 for cidiag=10, the number of reference CSFs plus 30 for cidiag=11, mindav+15 plus 15 times the number of requested CI roots for cidiag=12-15. kitdav Maximum number of iterations for Davidson diagonalization. Default 40 times the number of requested CI roots. nrmdav Convergence criterion for norm of q vector in Davidson diagonalization: 10**(-nrmdav). Default 7. nconj Number of atoms in the conjugated part of the system (primarily for movo=-5, but can be used with any movo<0). = 0 Assume all atoms are conjugated (default). = n Read n atom numbers to specify conjugated atoms. Details see below (third line). *** For movo=-5, the conjugated part of the system must contain at least 3 atoms. maxmap Maximum number of failed attempts to map active MOs for the case imomap=2. If MO mapping fails during an optimization step (as defined by mapthr), the optimizer will halve the step size taken and attempt to map the MOs again. After maxmap failed attempts, the optimization is abandoned. Default 5. mapthr Threshold for MO mapping overlap criterion to be considered a success, in percent. Only applicable for the case imomap=2. Default 90. iciocc Number of occupied MOs used with movo=4-6 for automatic selection of active orbitals (see description for movo). Default: use all occupied MOs up to a maximum of 60. icivir Number of unoccupied MOs used with movo=4-6 for automatic selection of active orbitals (see description for movo). Default: use all unoccupied MOs up to a maximum of 60. ldroot Target state used with movo=4,6 in minimal 3*3 singlet CI calculations during automatic selection of active orbitals (see description for movo). Default 2 (S1 state). *** Third line (for movo=1 or movo=3) *** jmoci(i) 1-80 20i4 Numbers of orbitals in the active space. (i=1,ici1+ici2) The numbering refers to the SCF output. If present, the order must be consistent with mocisy (fourth line) and iciref (fifth line). Otherwise the order is irrelevant since the orbitals are sorted before the CI calculation. Use more than one line for input, if there are more than 20 active orbitals. *** Fourth line (omit if movo.ne.3) *** mocisy(i) 1-80 20i4 Symmetry of orbitals in the active space. (i=1,ici1+ici2) Use more than one line for input, if there are more than 20 active orbitals. *** Fifth line (for nciref.gt.0 .and. mciref.eq.1 or 4) *** * Definition of the reference configurations: Option to define reference occupations which are then used to generate the corresponding reference configuration state functions. The dimensions in the current version allow for up to 10 reference occupations and up to 60 active orbitals. Typically there is one line of input for each reference. * Option mciref=1: iciref(i) 1-80 20i4 For each reference, read occupancies of the orbitals in the active CI space. * Default convention: The occupancies are assigned to the active orbitals in the order of increasing energy. * Case movo.ne.0: The occupancies are assigned to the active orbitals in the order as they appear in the input array imoci(i), see above. Use more than one line for input, if there are more than 20 active orbitals. *** Sixth line (omit if iroot.ge.0) *** iroota(i) 1-80 8i4 Requested number of CI roots of symmetry i. If the input is specified in free format, any missing entries have to be filled with zeros so that there are always eight numbers on this line. *** Seventh line (for movo.lt.0, omit if nconj.eq.0) *** iconj(i) 1-80 20i4 List of atom numbers that together form (i=1,nconj) the conjugated part of the system. Use more than one line for input if there are more than 20 conjugated atoms. 3.9.1 *** Multi-state treatment *** abs(kci)=5 and icross>0 ****** *** First line (only for icross=1-7) *** This line is read only for ncigrd.gt.0. Suitable default values are used for ncigrd=0 (see below). igrst(i) 1-80 20i4 CI gradients are computed for the ncigrd CI states identified by their labels igrst(i) on the energy scale. For icross=3-5, this line contains ncigrd=2 entries. The corresponding states are involved in a conical intersection. For ncigrd=0, the following default values are used (no input): ncigrd=2, igrst(1)=1, igrst(2)=2. For icross=1-2 and icross=7, no further input is required. For icross=3-5, further input is required on subsequent lines (see below). For icross=6, further input is specified in the namelist "dynvar.in" (see below). *** Second line (only for multci=-1) *** istmul(i) 1-8 2i4 Spin multiplicity of CI state i Note that multci=-1 implies icross=1-5, ncigrd=2. Multiplicities are specified in the same way as multci. If istmul(1) = istmul(2), multci is reset equal to istmul(1) and a standard calculation is performed. *** Third and subsequent lines (only for icross=3-5) *** This line is read only for ncigrd.gt.0. Suitable default values are used for ncigrd=0 (see below). See TCA paper: Theor.Chem.Acc. 118, 837-844 (2007) for further details on algorithms and options. * For icross=3, penalty function search for conical intersection: c1 1-10 f10.5 Factor c1 in penalty function. See eq.(4) in TCA paper. Default 5.0 (kcal/mol)**(-1). c2 11-20 f10.5 Factor c2 in penalty function. See eq.(4) in TCA paper. Default 5.0 (kcal/mol). * For icross=4, gradient projection search for conical intersection: c3 1-10 f10.5 Prefactor c3 in target gradient. See eq.(8) in TCA paper. Effectively of no significance if using the default initial Hessian guess. Default 1.0. c4 11-20 f10.5 Factor c4 in target gradient. See eq.(8) in TCA paper. Default 0.9. trad 21-30 f10.5 Static trust radius for Newton step. (applied after 'scale' below) See section 2.2 of TCA paper. Default is switched off (0.0). Recommended value: 0.1 scale 31-40 f10.5 Scaling of Newton step. See section 2.2 of TCA paper. Default is switched off (0.0). Recommended value: off (0.0). * For icross=5, Lagrange-Newton search for conical intersection: t1 1-10 f10.5 Threshold (in kcal/mol) below which orthogonalisation of the gradient difference and coupling vectors is switched on. See section 2.3 of TCA paper. Default 0.0001. t2 11-20 f10.5 Threshold (in kcal/mol) above which orthogonalisation is switched off again. See section 2.3 of TCA paper. Default 1.0. * For icross=5 only, subsequent lines specify geometrical constraints: Next line: nygeom 1-4 i4 Number of geometrical constraints. If nygeom=0, no further input is necessary. Each constraint is specified with a two-line input as follows: --- First line of constraint input: iylgc(1,i) 1-2 i2 Type of gradient constraint: 1 = bond length 2 = bond angle 3 = dihedral angle iylgc(j,i) 3-14 4i3 List of atom numbers defining the constraint. 2 atom numbers (j=2-3) for bond length, 3 atom numbers (j=2-4) for bond angle, 4 atom numbers (j=2-5) for dihedral angle. --- Second line of constraint input: ylgct(i) 1-10 f10.5 Target value for constraint (in Angstroms or degrees). * For icross=6, molecular dynamics/surface hopping runs: The dynamics options are specified using the auxiliary input file "dynvar.in", in the form of a Fortran namelist, which is described separately in section K (see below). 3.10 ***** CIS and RPA excited-state module ***** abs(kci)=6-8 ******** The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. *** First line *** Option No. Columns Format Short description ici1 131 1-4 i4 Number of active occupied orbitals. ici2 132 5-8 i4 Number of active unoccupied orbitals. ioutci 133 9-12 i4 Printing flag. nciref 136 21-24 i4 Selects between SASFCIS and SF-XCIS. iroot 139 33-36 i4 Total number of lowest CI states computed. lroot 140 37-40 i4 Defines the CI state of interest. multci 142 45-48 i4 Spin multiplicity of CI states. ncisym 143 49-52 i4 Suppress printing symmetry of states. cidiag 145 57-60 i4 Treatment of overlap matrix for NAC. iuvcd 146 61-64 i4 Spectroscopic properties: UV, CD. Option Full description ici1 Total number of occupied orbitals in the active CI space. Default is the number of alpha electrons. ici2 Total number of unoccupied orbitals in the active CI space. Default is the total number of orbitals minus the number of alpha electrons. ioutci Printing flag. =-10 No output. = 0 Small standard output. = 1 ... = 2 Increasing output size. = 3 ... = 5 Debug print. = 6 More debug print. nciref Selection between spin-flip methods for abs(kci)=7. = 1 Spin-adapted spin-flip CIS (SASFCIS). = 3 Spin-flip extended CIS (SF-XCIS, default). SF-XCIS is actually done for any nciref.ne.1. iroot Total number of lowest CI states computed. Default 6. lroot Index of the CI state of interest (1 for the ground state, 2 for the first excited state, etc.), e.g. during a geometry optimization. Default 1. multci Spin multiplicity of CI states. =-1 Varying multiplicity (requires icross=1-5, ncigrd=2) = 0 Same multiplicity as specified by imult (default). = 1 Singlet. = 3 Triplet. ncisym Flag whether to print symmetry of states. = 0 Do not print (actually for any ncisym.le.0). = 1 Print (actually for any ncisym.gt.0). cidiag Treatment of the derivative of the negative one halfth power of the overlap matrix during calculation of the non-adiabatic coupling matrix elements. = 0 Default: Approximation using Eq. (46) in Liu et al., J. Chem. Phys. 148, 244108 (2018). = 1 Numerical computation using finite differences, actually for any cidiag.gt.0. iuvcd Evaluation of spectroscopic properties for UV and CD spectra. = 0 No such evaluation (default). = 1 Evaluate spectroscopic properties. Print permanent dipole moments of all states calculated. = 2 Also print oscillator strengths, rotational strengths, and transition moments for transitions originating from ground state. *** Second line *** imomap 156 21-24 i4 Indicate active MO mapping failure. ncigrd 159 33-36 i4 Number of CI gradients to be computed. icross 160 27-40 i4 Choice of multi-surface treatment. maxdav 162 57-60 i4 Maximum number of attempts to map MOs. kitdav 163 49-52 i4 Maximum number of Davidson iterations. nrmdav 164 53-56 i4 Davidson convergence criterion: norm of q. mapthr 166 61-64 i4 Threshold for successful MO mapping. Option Full description imomap Indicate active MO mapping failure setting ICALL to -1. = 0 Do not indicate active MO mapping failure (default). = 1 Indicate active MO mapping failure. ncigrd Number of CI gradients to be computed. = 0 Calculate at most one CI gradient, for CI state lroot. Default case, no further input required. = n Compute CI gradients for n CI states which are specified on an extra line of input (seventh line). Minimum n=2, maximum n=4. *** Needed for surface crossings and conical intersections *** and ignored if icross=0. icross Choice of multi-surface treatment. = 0 No such treatment. = 1 Calculate energies and gradients for the states specified by ncigrd. = 2 As icross=1, but also calculate all interstate coupling gradients between the specified states. *** Options 1-2 are intended for single-point multi-surface *** calculations, in particular for interfacing with external *** optimisation/dynamics drivers. *** If desired they can also be used in conjuction with a *** single surface optimisation. The state to be optimised *** (specified by lroot) must be in the list of ncigrd gradients. = 3 Determine conical intersections between two CI states using the penalty function algorithm of Ciminelli et al. = 4 Determine conical intersections between two CI states using the gradient projection algorithm of Bearpark et al. = 5 Determine conical intersections between two CI states using the Lagrange-Newton algorithm of Yarkony. *** Options 3-5 are intended for use with the internal (MNDO99) *** optimization routines to locate conical intersections. *** For icross=3, the options ief=0, nrst=-1 and ldrop=100 *** are recommended. *** For icross=4, the default optimizer must be used (ief=0). *** For icross=5, ief>0 must be specified because the Yarkony *** optimizer is based on the pure Newton-Raphson routine *** available through the Eigenvector Following option. *** If icross=5 and ief=1, the options llamda=2 and *** lnoupd=1 are automatically chosen. = 6 Perform excited state dynamics with surface hopping using the semiclassical aproach of Tully and the included dynamics driver. *** For icross=6, further input is necessary in the form of *** a namelist "dynvar.in". = 7 As icross=2, but use the full expression to calculate the non-adiabatic coupling matrix elements between the specified states. maxdav Maximum dimension of subspace for Davidson diagonalization. Default iroot * kitdav. kitdav Maximum number of iterations for Davidson diagonalization. Default 50. nrmdav Convergence criterion for norm of q vector in Davidson diagonalization. Default 7. mapthr Threshold for MO mapping overlap criterion to be considered a success, in percent. Default 80. 3.10.1 *** Multi-state treatment *** abs(kci)=6-8 and icross>0 *** Options for icross are identical to those for GUGACI (abs(kci)=5). For icross=1-7, additional input is needed (see section 3.9.1). The input description is not repeated here (see section 3.9.1). 3.11 ***** Input of reference data **** inrefd nonzero *********** The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. *** Warning: Input in free format not yet implemented. There is one line of input for each reference datum. Variable Columns Format Description id1 1-2 i2 Definition of reference property. = 0 End of this input section. = 1 Heat of formation (kcal/mol) = 2 Bond length (A) = 3 Bond angle (deg) = 4 Dihedral angle (deg) = 5 Ionization potential (eV) = 6 Excitation energy (eV) = 7 Dipole moment (D) = 8 Polarisability: alpha (A**3) = 9 Polarisability: beta (10D-30 esu) =10 Polarisability: gamma (10D-36 esu) =11 Enthalpy change at 298 K (kcal/mol) =12 Dissociation energy (kcal/mol) =13 Activation enthalpy at 298 K (kcal/mol) =14 Ionization energy (kcal/mol) =15 Electron affinity (kcal/mol) =16 Vibrational wavenumber (cm-1) =17 Atomic charge (e) =18 Population (e) =19 S**2 expectation value =20 Gradient component (kcal/mol*A) =21 Internal gradient norm (kcal/mol*A) =22 Cartesian gradient norm (kcal/mol*A) =23 IPs: Higher ionization (eV) =24 IPs: Difference (eV) =25 Dipole moment: non-ZDO (D) =26 Number of imaginary frequencies =27 Wavenumber difference (cm-1) =28 Solvation energy: adiabatic (kcal/mol) =29 Solvation energy: vertical (kcal/mol) =30 NMR chemical shift: gas phase (ppm) =31 NMR chemical shift (ppm) =32 NICS chemical shifts (ppm) =33 CI transition energy (eV) =34 CI rotational strength (DBM) =35 CI oscillator strength (au) =36 CI dipole or transition moment (D) =37 CI dip/tr moment angle: phi (deg) =38 CI dip/tr moment angle: theta (deg) =39 CI dipole moment components (D) =40 Orbital energy (eV) =41 Interaction energy (kcal/mol) =42 Atomization energy at 0 K w/o ZPVE (kcal/mol) =43 Proton affinity (kcal/mol) =44 Reaction energy (kcal/mol) =45 Adiabatic excitation energy (eV) =46 Overall distance criterion (A) =47 Atomization enthalpy at 298 K (kcal/mol) =48 Relative energy at 0 K w/o ZPVE (kcal/mol) =49 Barrier at 0 K w/o ZPVE (kcal/mol) id2 3-5 i3 Assignment to a user-chosen group. Irrelevant for the present program. Used as input in a separate evaluation program where a negative value excludes the reference datum from the statistics. Note: Negative id2 values below -90 are used internally and are therefore not allowed as input values. id3 6-8 i3 Assignment to a user-chosen subgroup. Irrelevant for the present program. Used as input in a separate evaluation program where a negative value excludes the reference datum from the statistics. id4 9-11 i3 Definition of reference datum (see below). id5 12-14 i3 Definition of reference datum (see below). id6 15-17 i3 Definition of reference datum (see below). id7 18-20 i3 Definition of reference datum (see below). xi 21-30 f10.5 Reference value (usually from experiment). erri 31-40 f10.5 Quoted error of this reference value. Irrelevant for the present program. Used here only for printing purposes. wi 41-45 f5.1 Weight of this reference datum. Irrelevant for the present program. Used here only for printing purposes. liti 46-50 i5 User-chosen label of literature reference. Irrelevant for the present program. Used here only for printing purposes. komi 51-80 a30 Comment on reference datum. Used only for printing purposes. For a given property (id1) the reference data can be defined more precisely by specifying the input variables id4-id7. Conventions: id4 id5 id6 id7 id1= 1 0 0 0 0 Heat of formation. N4 is not allowed to be non-zero for the ODM2 and ODM3 methods. N4 0 0 0 Heat of formation of the current molecule (MOL) relative to a previously calculated molecule (INDEX), with INDEX=N4 for positive N4 ( N4.LT.MOL), INDEX=MOL+N4 for negative N4 (-N4.LT.MOL). N4 N5 0 0 Heat of formation of the current molecule relative to two other molecules. Analogous conventions as above (N4,N5). Heat of dimerization for N4=N5. N4 N5 N6 0 Heat of formation of the current molecule relative to three other molecules. Analogous conventions as above (N4-N6). N4 N5 N6 N7 Heat of formation of the current molecule relative to four other molecules. Analogous conventions as above (N4-N7). 0 N5 N6 N7 Special option only for atoms. Heat of formation of a configuration with occupation numbers N5, N6, and N7 for the s, p, and d orbitals, respectively. Evaluated from the one-center parameters. Special conventions: N5=N6=N7=0 refers to the configuration used in the SCF calculation (default). This input can thus not be employed for the s0p0d0 configuration (no electrons) which is selected by N5=3, N6=N7=0. id1= 2 N4 N5 0 0 Bond length between atoms N4-N5 in the full list (including dummy atoms). Allowed range of values: 1 ... NATOMS id1= 3 N4 N5 N6 0 Bond angle between atoms N4-N5-N6 in the full list (including dummy atoms). Allowed range of values: 1 ... NATOMS id1= 4 N4 N5 N6 N7 Dihedral angle between atoms N4-N5-N6-N7 in the full list (including dummy atoms). Allowed range of values: 1 ... NATOMS id1= 5 0 0 0 0 Lowest vertical ionization potential. Evaluated from Koopmans' theorem. Any input for id4-id7 is ignored. id1= 6 N4 N5 N6 0 Vertical excitation energy. Evaluated from the Hartree-Fock energy of a singly excited configuration built from ground-state closed-shell MOs. N4: Initial MO (default HOMO). N5: Final MO (default LUMO). N6: Multiplicity (default singlet, N6=1). id7 is used internally for counting the excited states treated. id1= 7 0 0 0 0 Total dipole moment. Evaluated in ZDO approximation. N4 0 0 0 Principal-axis dipole moment component. N4: 1,2,3 for A,B,C axis. Any input for id5-id7 is ignored. id1= 8 N4 0 0 0 Polarisability: alpha. N4: 0 from energy derivatives. N4: 1 from dipole moment derivatives. Any input for id5-id7 is ignored. id1= 9 N4 0 0 0 First hyperpolarisability: beta. N4: 0 from energy derivatives. N4: 1 from dipole moment derivatives. Any input for id5-id7 is ignored. id1=10 N4 0 0 0 Second hyperpolarisability: gamma. N4: 0 from energy derivatives. N4: 1 from dipole moment derivatives. Any input for id5-id7 is ignored. id1=11 N4 N5 N6 N7 Enthalpy change at 298 K. Identical conventions as for id1=1. Only difference: Property name (output). 0 N5 N6 N7 Special option only for atoms. Energy of an atomic configuration with occupation numbers N5, N6, and N7 for the s, p, and d orbitals, respectively, relative to the neutral ground state. id1=12 N4 N5 N6 N7 Dissociation energy. Identical conventions as for id1=1. Only differences from id1=11: Opposite sign and property name (output). id1=13 N4 N5 N6 N7 Activation enthalpy at 298 K (conformational or reactive). Identical conventions as for id1=1. Main difference: Property name (output). id1=14 N4 0 0 0 Adiabatic ionization energy. Evaluated as a relative energy. There are no checks whether the partners are matching (molecule and its cation). Analogous conventions as for id1=1. id1=15 N4 0 0 0 Adiabatic electron affinity. Evaluated as a relative energy. There are no checks whether the partners are matching (molecule and its anion). Analogous conventions as for id1=1. id1=16 N4 0 0 0 Vibrational wavenumber. N4: Index of vibration for wavenumbers in ascending order (1 ... 3*N). N4 -1 0 0 N4: Index of vibration for wavenumbers in descending order (highest: N4=1). N4 N5 0 0 N4: Index of vibration for wavenumbers in descending order, for a given symmetry or degeneracy. *** For nondegenerate point groups. N5: Label for irreducible representation. See chapter 3.6 for possible N5 values in this version of the program. Example: N4=N5=1 specifies the highest wavenumber of the totally symmetric modes. *** For degenerate point groups. N5: Degeneracy number (1 or 2 or 3). Example: N4=N5=1 specifies the highest wavenumber of the nondegenerate modes. Note: For the doubly (triply) degenerate modes, the highest wavenumber refers to N4=1-2 (1-3), the second highest one to N4=3-4 (4-6), etc. id1=17 N4 0 0 0 Atomic charges. N4: Number of atom. id1=18 N4 N5 0 0 Subshell populations. N4: Number of atom. N5: 1 for s, 2 for p, 3 for d. id1=19 0 0 0 0 Spin expectation value S**2 (UHF only). Any input for id4-id7 is ignored. Generated by default for UHF. id1=20 N4 0 0 0 Gradient norm. Irrelevant for the present program. Used here only for printing purposes. Used elsewhere to define reference data for parametrizations with fixed geometry. 0 0 0 0 Internal gradient norm of all optimized variables included in the reference data, -1 0 0 0 Cartesian gradient norm of all (3*NUMAT) variables included in the reference data. N4 0 0 0 Gradient component of optimized variable N4 (N4>0) included in the reference data. id1=21 0 0 0 0 Internal gradient norm. Irrelevant for the present program. Used here only for printing purposes. Any input for id4-id7 is ignored. id1=22 0 0 0 0 Cartesian gradient norm. Irrelevant for the present program. Used here only for printing purposes. Any input for id4-id7 is ignored. Generated by default. id1=23 N4 0 0 0 Higher vertical ionization potential. Evaluated from Koopmans' theorem. N4: Label for the ionized orbital. Standard ascending order. Default HOMO. N4 -1 0 0 N4: Label for the ionized orbital. Descending order of occupied orbitals, with N4=1 for HOMO, N4=2 for HOMO-1, etc. Default HOMO. N4 N5 0 0 N4: Label for the ionized orbital. Descending order of occupied orbitals, for a given symmetry or degeneracy. *** For nondegenerate point groups. N5: Label for irreducible representation. See chapter 3.6 for possible N5 values in this version of the program. Example: N4=N5=1 specifies the highest totally symmetric occupied orbital. *** For degenerate point groups. N5: Degeneracy number (1 or 2 or 3). Example: N4=N5=1 specifies the highest nondegenerate occupied orbital. Note: For the doubly (triply) degenerate orbitals, the highest eigenvalue refers to N4=1-2 (1-3), the second highest one to N4=3-4 (4-6), etc. Default: Highest orbital of a given type. id1=24 N4 N5 0 0 Difference of ionization potentials. Evaluated from Koopmans' theorem. Ionizations from orbitals N4 and N5. N4 must be greater than N5. The resulting difference is defined to be positive. Default N4=HOMO and N5=HOMO-1. id1=25 0 0 0 0 Total dipole moment. Evaluated after Loewdin transformation. N4 0 0 0 Principal-axis dipole moment component. N4: 1,2,3 for A,B,C axis. Any input for id4-id7 is ignored. id1=26 0 0 0 0 Number of imaginary frequencies. Any input for id4-id7 is ignored. id1=27 N4 N5 N6 N7 Difference of vibrational wavenumbers. N4 and N5 define the first wavenumber in complete analogy to the conventions for id1=16 (see above). N6 and N7 define the second wavenumber in complete analogy to the conventions for id1=16 (see above). Default. Difference between the two highest wavenumbers. id1=28 N4 N5 0 0 Adiabatic solvation energy. Evaluated as energy difference between solution (COSMO treatment) and gas phase, with separate geometry optimizations. There are no checks whether the partners are matching (in the input file). N4: Analogous conventions as with id1=1. N5: Value for option icosmo in COSMO. See description in chapter 1 above. Values between 0 and 4 are allowed. Default (N5=0): N5 is set equal to the global input value for icosmo (see chapter 1) if icosmo = 1..4, otherwise N5 is set equal to 1. Any input for N6 and N7 is ignored. id1=29 0 N5 0 0 Vertical solvation energy. Evaluated as energy difference between solution (COSMO treatment) and gas phase, at the optimized gas phase geometry (or at the input geometry for jop.lt.0). N5: Value for option icosmo in COSMO. See description in chapter 1 above. Values between 0 and -4 are allowed. Default (N5=0): N5 is set equal to the global input value for icosmo (see chapter 1) if icosmo = -1..-4, otherwise N5 is set equal to -1. Any input for N4, N6, and N7 is ignored. id1=30 N4 N5 0 0 NMR chemical shifts with respect to gas-phase reference data. N4: Number of the atom for which the NMR chemical shift is given. The numbering refers to the list of real atoms (without dummy atoms). N5: In the experimental reference data, atom N4 is equivalent to atom N5. The theoretical shifts are averaged. id1=31 N4 N5 0 0 NMR chemical shifts with respect to liquid-phase reference data. N4: Number of the atom for which the NMR chemical shift is given. The numbering refers to the list of real atoms (without dummy atoms). N5: In the experimental reference data, atom N4 is equivalent to atom N5. The theoretical shifts are averaged. id1=32 N4 0 0 0 NICS chemical shifts. N4: Number of the NICS point for which the chemical shift is given. The numbering refers to the input list of NICS points. id1=33 N4 N5 0 0 CI energy for transition originating from the ground state. N4: If N5.EQ.0, number of excited state, i.e.: 2 for the first excited state. If N5.GT.0, only states with this symmetry are counted. N5: Symmetry number of excited state or 0 if symmetry does not matter. id1=34 N4 N5 0 0 CI dipole-velocity rotational strength for transition originating from ground state. N4: Analogous to id1=33. N5: Analogous to id1=33. id1=35 N4 N5 N6 0 CI oscillator strength for transition originating from the ground state. N4: Analogous to id1=33. N5: Analogous to id1=33. N6: 0 default (treated as N6=1), 1 dipole-length formalism, 2 dipole-velocity formalism, 3 mixed formalism. id1=36 N4 N5 N6 0 CI dipole or transition moment. N4: If N5.EQ.0, number of state, i.e.: 1 for ground state, 2 for first excited state, etc. If N5.GT.0, only states with this symmetry are counted. N5: Symmetry number of excited state or 0 if symmetry does not matter. N6: 0 permanent dipole moment, 1 transition moment for transitions originating from the ground state (dipole-length formalism), 2 transition moment for transitions originating from the ground state (dipole-velocity formalism), 3 difference (N6=1) minus (N6=2). id1=37 N4 N5 N6 0 CI dipole or transition moment angle phi. N4: Analogous to id1=36. N5: Analogous to id1=36. N6: Analogous to id1=36. *** Convention: If the dominant component of the dipole-length dipole transition moment is in the xy plane, the phase factor of the excited state is adjusted such that the angle with the x axis (phi) is between -90 and +90 deg. id1=38 N4 N5 N6 0 CI dipole or transition moment angle theta. N4: Analogous to id1=36. N5: Analogous to id1=36. N6: Analogous to id1=36. *** Convention: If the dominant component of the dipole-length dipole transition moment is on the z axis, the phase factor of the excited state is adjusted such that the angle with the z axis (theta) is between 0 and +45 deg. id1=39 N4 N5 N6 0 CI dipole moment components. id1=40 N4 0 0 0 Orbital energy. N4: Label for the orbital. Standard ascending order. Default HOMO. *** Comparison with id1=23: both occupied *** and unoccupied orbitals are covered, *** opposite sign for occupied levels. *** The following conventions generalize *** the options offered by id1=23,24. N4 -1 0 0 N4: Label for an occupied orbital. Descending order of occupied orbitals, with N4=1 for HOMO, N4=2 for HOMO-1, etc. Default HOMO. N4 N5 0 0 N4: Label for an occupied orbital. Descending order of occupied orbitals, for a given symmetry or degeneracy. Default: Highest orbital of a given type. *** For nondegenerate point groups. N5: Label for irreducible representation. See chapter 3.6 for possible N5 values in this version of the program. Example: N4=N5=1 specifies the highest totally symmetric occupied orbital. *** For degenerate point groups. N5: Degeneracy number (1 or 2 or 3). Example: N4=N5=1 specifies the highest nondegenerate occupied orbital. Note: For the doubly (triply) degenerate orbitals, the highest eigenvalue refers to N4=1-2 (1-3), the second highest one to N4=3-4 (4-6), etc. 0 0 N6 -1 N6: Label for an unoccupied orbital. Ascending order of unoccupied orbitals, with N6=1 for LUMO, N6=2 for LUMO+1, etc. No default defined, explicit input needed. 0 0 N6 N7 N6: Label for an unoccupied orbital. Ascending order of unoccupied orbitals, for a given symmetry or degeneracy. No default defined, explicit input needed. *** For nondegenerate point groups. N6: Label for irreducible representation. See chapter 3.6 for possible N5 values in this version of the program. Example: N6=N7=1 specifies the lowest totally symmetric unoccupied orbital. *** For degenerate point groups. N6: Degeneracy number (1 or 2 or 3). Example: N6=N7=1 specifies the lowest nondegenerate unoccupied orbital. Note: For the doubly (triply) degenerate orbitals, the lowest eigenvalue refers to N6=1-2 (1-3), the second highest one to N6=3-4 (4-6), etc. N4 N5 N6 N7 Energy difference between an unoccupied (N6,N7) and an occupied (N4,N5) orbital. This difference is defined to be positive. The two orbitals are specified according to the given conventions (N4,N5,N6,N7), N5=-1 and N7=-1 may be used. id1=41 N4 0 0 0 Interaction energy of two fragments at the given input geometry. The first fragment consists of the first N4 atoms, the second one of the remaining atoms (N4+1...NUMAT). This input must be followed by two inputs for the fragments specifying single-point calculations; input coordinates of the fragments will be overwritten by those from the geometry of the whole system. Any input for id5-id7 is ignored. id1=42 0 0 0 0 Atomization energy at 0 K without ZPVE. Any input for id4-id7 is ignored. id1=43 N4 0 0 0 Proton affinity. N4 is reference to non-protonated form. ** Traditional conventions (iaterg=1) No input for calculating the heat of formation of the proton is necessary, because the experimental value of 367.171 kcal/mol is used (HoF at 298 K), see JANAF Thermochemical Tables, 1982 Supplement, J. Phys. Chem. Ref. Data 11, 695 (1982); doi: 10.1063/1.555666. ** Conventions ODM2 and ODM3 (iaterg=-1) Computed from difference of total energies and corrected for the difference between experimental and semiempirical ionization potential of the hydrogen atom. ** Identical conventions as for id1=1. ** Any input for id5-id7 is ignored. id1=44 N4 0 0 0 Energy change of any generic reaction. N4 is the stoichiometric coefficient of the current molecule (integer only). Coefficients are positive for products and negative for reactants. Subsequent extra lines are used for input of other chemical species involved in the reaction (see section 3.11.1). id1=45 N4 0 0 0 Adiabatic excitation energy. Identical conventions as for id1=1. Differences: Property name (output), units (eV rather than kcal/mol). Any input for id5-id7 is ignored. id1=46 0 0 0 0 Square root of the normalized squared sum of all interatomic distances. ** Experimental option. ** Reference value determined from ** experimental or theoretical geometry. ** May be used to assess deviation of ** computed optimized geometry from ** the reference geometry. id1=47 0 0 0 0 Atomization enthalpy at 298 K. Any input for id4-id7 is ignored. id1=48 N4 N5 N6 N7 Relative energy at 0 K without ZPVE. Identical conventions as for id1=1. 0 N5 N6 N7 Special option only for atoms. Energy of an atomic configuration with occupation numbers N5, N6, and N7 for the s, p, and d orbitals, respectively, relative to the neutral ground state. id1=49 N4 N5 N6 N7 Barrier at 0 K (conformational or reactive). Identical conventions as for id1=1. End input of reference data with id1=0. In the case of formatted input a blank line may be used. 3.11.1 ***** Input of reaction partners **** id1=44 ************** Warning: Input in free format not yet implemented. There is one line of input for each reaction partner. Input is possible for up to 20 reaction partners. Variable Columns Format Description icoefs 9-11 i3 Stoichiometric coefficient. imol 12-14 i3 The current molecule index. Option Full description icoefs Stoichiometric coefficient. Positive for products, negative for reactants. imol The current molecule index (MOL) relative to a previously calculated molecule (INDEX), with INDEX=IMOL for positive IMOL ( IMOL.LT.MOL), INDEX=MOL+IMOL for negative IMOL (-IMOL.LT.MOL). End input of this section with icoefs=0. 3.12 ***** Input of COSMO data ******** abs(icosmo).gt.2 ********* The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. *** First line *** Option No. Columns Format Short description nspa 231 1-5 i5 Number of segments on the unit sphere. nvdw 232 6-10 i5 Choice of van-der Waals radii. ipot 233 11-15 i5 Choice of electrostatic potential. nitro 234 16-20 i5 Empirical correction for nitro group. modcsm 235 21-25 i5 Memory allocation for COSMO calculation. epsi 16 31-40 f10.5 Dielectricity constant. rsolv 17 41-50 f10.5 Effective radius of solvens molecule. delsc 18 51-60 f10.5 Distance between surfaces (SAS-charges). disex 19 61-70 f10.5 COSMO cutoff parameter. Option Full description nspa Number of segments on the unit sphere. Default 42. Minimum 12, maximum 1082. nvdw Choice of van-der Waals radii (Angstrom). = 0 Use Bondi values by default. A. Bondi, J.Phys.Chem. 68, 441 (1964). =-1 Use Emsley values for radii. =-2 Use same values as in MOPAC. = n Number of additional lines of input to read non-standard van-der-Waals radii, with one line per element (see below). ipot Choice of electrostatic potential used to compute the electrostatic interactions. = 0 Use default value (ipot=2). = 1 Not implemented. Currently treated as ipot=2. = 2 Classical Coulomb potential generated by the atomic monopoles, dipoles, and quadrupoles (no additive terms). Used in the original COSMO code. = 3 Semiempirical potential. = 4 Parametrized semiempirical potential. = 5 Parametrized semiempirical potential. = 6 Parametrized semiempirical potential. = 6 Parametrized semiempirical potential. *** The semiempirical potentials ipot=3-6 are described in *** section 3.13 for mmpot=3-6. The conventions for ipot *** are analogous to those for mmpot. nitro Empirical additive correction for nitro compounds, see A. Gelessus, Ph.D. Thesis, University of Zurich, 1997. = 0 No such correction. = 1 Apply the correction. *** Experimental option, not recommended. modcsm Memory allocation for COSMO calculation. = 0 Determined automatically. = 1 Reduce memory requirements by the use of linearly packed matrices. = 2 Reduce memory requirements further by storing intermediate data on disk. NOT SUPPORTED IN THIS VERSION. The default is recommended (modcsm=0). epsi Dielectricity constant. Default value: 78.4 (for water). rsolv Effective radius of solvens molecule (in Angstrom) to define the distance between the van-der-Waals surface and the solvent accessible surface. Default value: 1.0 delsc Distance between the solvent accessible surface and the surface containing the screening charges (in Angstrom). Default: delsc=rsolv (screening charges located on the van-der-Waals surface). disex Parameter to determine the square (disex2, in Angstrom**2) of the distance beyond which the interactions are evaluated using a single charge per segment. Default for disex: 2.0 Definition of disex2 in the code: rds = max(delsc,0.1) disex2 = (4*(1.5+rsolv-rds)*disex)**2/nspa By default (see above): rds = max(rsolv,0.1) = rsolv disex2 = (6*disex)**2/nspa disex2 = 3.4 Angstrom**2 This yields solvation energies which are usually within 1 % of the values obtained from a full (and much more expensive) calculation with a large input value for disex (e.g. 100.0). Note that any redefinition of nspa affects disex2 (see above). *** Following (nvdw) lines *** one line per element *** ivdw 1-5 i5 Atomic number of the element for which the van-der-Waals radius is redefined. rivdw 6-15 f10.5 New van-der-Waals radius (Angstrom). Default values are taken from A. Bondi, J.Phys.Chem. 68, 441 (1964). 3.13 ***** Input of external points ***** mminp=1-2 ************** The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. Option mminp=1: Properties are evaluated at the external points. In the current version, the electrostatic potential is always calculated (according to the value of option mmpot, see below). The electric field is the negative gradient of the electrostatic potential and is therefore also governed by option mmpot. The electric field is currently evaluated only for mmcoup>2. For a discussion of semiempirical electrostatic potentials see: (1) D. Bakowies and W.Thiel, J.Comp.Chem. 17, 87 (1996). (2) G.P. Ford and B. Wang, J.Comp.Chem. 14, 1101 (1993). (3) P.L. Cummins and J.E. Gready, Chem.Phys.Lett. 174, 355 (1990). (4) M.J. Field et al, J.Comp.Chem. 11, 700 (1990). Option mminp=2: A quantum-chemical (QM) calculation is carried out for a combined QM/MM hybrid treatment. The actual computation depends on the coupling model (as determined by option mmcoup). mmcoup=1 : Standard QM calculation. mmcoup>1 : QM calculation with external point charges. mmcoup>2 : The electric field at the MM atoms is also computed. For a discussion of the relevant coupling models see: D. Bakowies and W.Thiel, J.Phys.Chem. 100, 10580 (1996). *** First line *** Option No. Columns Format Short description numatm 120 1-5 i5 Number of external points. mmcoup 121 6-10 i5 Coupling model in QM/MM calculations. mmpot 122 11-15 i5 Definition of the electrostatic potential. mmlink 123 16-20 i5 Link atom treatment. nlink 124 21-25 i5 Number of link atoms. mmfile 125 26-30 i5 Input file for external points and charges. mcharg 126 31-35 i5 Definition of external charges. mmskip 127 36-40 i5 Skip gradient for fixed MM atoms. Option Full description numatm Number of external points. mmcoup Coupling model in QM/MM calculations. = 0 No QM/MM treatment. = 1 Model A, mechanical embedding. = 2 Model B, electronic embedding. = 3 Model C, electronic embedding with MM polarization. mmpot Definition of the electrostatic potential and the electric field (by implication). = 0 Use default value (mmpot=4). = 1 Obtained classically from the net atomic charges in the QM region (Coulomb law, monopoles only). = 2 Obtained classically from the monopoles, dipoles, and quadrupoles at the QM atoms (no additive terms). This corresponds to the default in the COSMO treatment. = 3 Obtained semiempirically from the standard integral approximations and standard parameters. = 4 Obtained semiempirically from a special parametrization with a one-parametric core-core function. Omega parameters from ref.1 above. = 5 Obtained semiempirically from a special parametrization with a two-parametric core-core function. Omega and delta parameters from ref.1. = 6 Obtained semiempirically from a special parametrization with a two-parametric core-core function. Omega and delta parameters from ref.2. = 7 Obtained semiempirically, ref. 4. = 8 Obtained semiempirically, ref. 3. *** The external test charge is treated classically for mmpot=3-6, and as a hydrogen atom for mmpot=8. *** Standard parameters (omega=alpha, delta=0) are used for mmpot=4-6, whenever special parameters are not available from ref. 1 or 2. mmlink Link atom treatment. Normally only relevant for QM/MM studies. = 0 Use default value (mmlink=1). = 1 Standard semiempirical approach: Link atoms do not see the MM atoms. = 2 Alternative semiempirical approach: Link atoms do see all MM atoms. This leads to excessive interactions unless charges at neighbor atoms are set to zero (responsibility of user). Recommendation: Use zero MM charges in the nearest "charge group". See program documentation for details. nlink Number of link atoms. Normally only relevant for QM/MM studies. mmfile Definition of input file for external points and charges. =-1 Do not read such data from file. These data are defined elsewhere. = 0 Read data from standard input. = 1 Read data from file nb20. mcharg Definition of external charges. = 0 Read charges from input. = n Other options not yet implemented. mmskip Gradient calculation for fixed MM atoms. = 0 Full calculation for all atoms. = 1 Skip gradient calculation for fixed MM atoms which are identified by positive values of the flags ISELCT(I) in common block QMMM6. *** Second line *** only for nlink.gt.0 *** link(i) 1-80 16i5 Numbers of link atoms in the list of all atoms (QM atoms including dummy atoms). There are nlink input data. Use more than one line, if necessary. *** Third and following lines *** only for mmfile.ge.0 *** One line per external point to define its Cartesian coordinates (in Angstrom) and its charge (in e). cm(1,m) 1-12 f12.7 x coordinate. cm(2,m) 13-24 f12.7 y coordinate. cm(3,m) 25-36 f12.7 z coordinate. qm(m) 37-44 f8.4 External point charge. iselct(i) 45-47 i3 Flag (see mmskip), read only for mmskip.eq.1. For mmskip.eq.0, any item after qm(m) on the same line will be ignored. 3.14 **** Data for NMR ***** nmrnuc=6 **************************** Explicit input for NMR computations consists of a set of control records which have the following format: icmd 1-3 i3 Control flag. parms 4-80 a77 Internal file for associated data. Depending on the value of the control flag the input for parms is interpreted as: 15i5 icmd = 2, 3, 12, 13, 102, 103, 104, or 3f10.5 icmd = 101. The following table specifies the available control records. icmd Description of the control record. 0 End of NMR control records. 1 Exclude all atoms from NMR computation. 2 Exclude specified centers from NMR computation: parms contains the list of the atoms to be excluded. The numbering in this list includes any dummy atoms. 3 Exclude specified elements from NMR computation: parms contains atomic numbers of the elements to be excluded. 11 Include all atoms in NMR computation. 12 Include specified centers in NMR computation: parms contains the list of the atoms to be included. The numbering in this list includes any dummy atoms. 13 Include specified elements in NMR computation: parms contains atomic numbers of the elements to be included. 101 Compute chemical shift for points in space (NICS): parms defines these points by specifying their Cartesian coordinates in Angstrom. 102 Compute chemical shift for points in space (NICS): parms defines these points by specifying the numbers of the corresponding dummy atoms in the list of all atoms. 103 Compute chemical shift for points in space (NICS): parms defines these points by specifying the numbers of two atoms (nics1,nics2) which limit a range of reference atoms (nics1 - nics2). The Cartesian coordinates of the NICS point are computed as the nonweighted mean of the Cartesian coordinates of these reference atoms. 104 Compute chemical shift for points in space (NICS): parms defines these points by specifying the numbers of the reference atoms in the list of all atoms (including dummy atoms). The Cartesian coordinates of the NICS point are computed as the nonweighted mean of the Cartesian coordinates of these reference atoms. >900 Obsolete options for debugging which are not described. Note: Requests from the control records are handled sequentially, as these records appear in the input file. Note: Whenever a center is specified in the list of atoms, the numbering includes dummy atoms (icmd=2,12,102-104). Note: Up to MXNICS points may be defined (icmd=101-104). Note: Input from this section is completed when icmd=0 or an end-of-file is encountered. 3.15 **** Definition of masses ***** jop=2-6 and kmass>0 ********* The input in this section is either formatted (option iform=0) or in free format (option iform=1). The column numbers and the formats given below refer to the option iform=0. imass(i) 1-80 20i4 Mass of atom i in vibrational analysis. = 0 Use mass of principal isotope. = m Use mass of isotope with m nucleons. The following isotopes are available. H-2,H-3,Li-6,B-10,C-13,N-15,O-17,O-18, Mg-25,Mg-26,Si-29,Si-30,S-33,S-34,Cl-37. If imass(i) does not correspond to one of these isotopes, the default mass of the principal isotope is taken. Use more than one line, if necessary. For kmass=n>1, there are n such definitions which are read to define n isotopomers. 3.16 **** Data for HDLC optimizer ***** ief<0 ******************** A detailed input description is given in section J. Special data for the HDLC optimizer can be read either from the standard input file (this section) or from an external file. The same input conventions apply in both cases (see section J). The input mode is determined by input option ihdlc3: ihdlc3 = 0 Read from external input file. ihdlc3 = 1 Read from standard input file nb5. 4. ***** Input for the next molecule ***************************** At this point, the next molecule can be read. In the case of standard input, the options from chapters 1 and 2 remain valid so that only the data from chapter 3 are needed for the next molecule (starting with the title line, chapter 3.1). The job is terminated if kharge=99 is read in columns 1-2 of the title line or if an end-of-file is encountered. ****************************************************************** ****************************************************************** E. Outline of keyword-oriented input. The keyword input and the MOPAC input (see overview in section B) share the same overall input structure and will therefore both be discussed in this section. More detailed descriptions of the MNDO keywords and the MOPAC keywords will be given in sections F and G, respectively. The input of keywords is case-insensitive. Internally, keywords are represented by upper-case characters. Lower-case characters from input are converted automatically. The keyword input consists of the following parts. (a) MNDO keywords to define the input options from chapters 1, 2, 2.1-2.4, 3.1, 3.8 (first line), 3.9 (first two lines), 3.10 (two lines), 3.12 (first line), and 3.13 (first line) of the standard input (see sections C and D). The order of the keywords is arbitrary. There is at least one line of input with keywords. An additional line of input with keywords is read when there is a continuation keyword (' &' or ' +') on the current line. Up to ten lines of input with keywords are allowed. (b) Two lines with text to identify the calculation. This text is printed as a header in several output sections. (c) Standard input as specified in chapters 3.2 - 3.16 (except lines covered by keyword input, see above) of section D, in this order. The MOPAC input consists of the following parts. (a) MOPAC and MNDO keywords to define the input options from chapters 1, 2, 2.1-2.4, 3.1, 3.8 (first line), 3.9 (first two lines), 3.10 (two lines), 3.12 (first line), and 3.13 (first line) of the standard input (see sections C and D). The order of the keywords is arbitrary. A second line of input with keywords is read when there is a continuation keyword ( & or + ) on the first line. A third line of input with keywords is read when there is a continuation keyword ( & or + ) on the second line. Hence, up to three lines of input with keywords are allowed. (b) Two lines with text to identify the calculation. This text is printed as a header in several output sections. (c) MOPAC-type input to specify the data from chapters 3.2, 3.3, 3.4, 3.7, and 3.15 of the standard input (i.e., for geometry, symmetry data, reaction paths, configuration interaction, and atomic masses, respectively). This MOPAC-type input is defined in the MOPAC(6.0) manual. The current implementation follows the MOPAC(6.0) conventions in order to allow the direct use of MOPAC input files. (d) Standard input as specified in chapters 3.6, 3.8, 3.9, 3.11, 3.12, 3.13 and 3.14 (except lines covered by keyword input, see above) of the standard input (i.e. for occupation numbers, perturbative correlation treatments, GUGACI options, reference data, COSMO solvation treatment, external points charges, and NMR data, respectively), in this order. Standard input is used for these data because there is no equivalent in MOPAC(6.0). This is done to make all options of MNDO accessible also when using MOPAC-type input. Input for the next molecule: In the case of keyword input or MOPAC-type input, a complete set of data must be read for each molecule, starting with the keywords (a). The job is terminated if an end-of-file is encountered when reading the new keywords or if kharge=99 is used as input. ****************************************************************** ****************************************************************** F. Description of MNDO input keywords. The general format of MNDO keywords is = . Name of any input variable in the standard input, as defined in this input description (section D). See chapters 1, 2, 2.1-2.4, 3.1, 3.8-3.10, 3.12, 3.13. Actual value of this input variable. Input through MNDO keywords is equivalent to standard input. MNDO keywords can be employed in two modes of input: Keyword input: There are only MNDO keywords (no MOPAC keywords) and possibly continuation keywords ( + or &). There is at least one line of keywords and at most ten such lines, any additional line requiring a continuation keyword on the preceding line. MOPAC input: MNDO keywords can be combined with MOPAC keywords on the keyword line(s) of MOPAC input. Note: A list of all available keywords is printed for jprint=7. This list covers both MNDO and MOPAC keywords. ****************************************************************** ****************************************************************** G. Description of MOPAC input keywords in this program. This section gives a complete list of the MOPAC(6.0) keywords and specifies the response of the program when encountering a given keyword. This specification includes the following information. Availability yes Keyword fully implemented. no Keyword not implemented. partly Similar option is available. Basic action okay Use equivalent standard input option. stop Stop the calculation. ignore Continue and ignore the keyword. similar Use a similar standard input option. Translation iop=.. Definition of standard input options used to implement the keyword. See detailed descriptions above. Additional remarks are usually given for each keyword. The list of keywords below is ordered alphabetically. Keyword Avail- Basic Trans- Remarks ability action lation & yes okay next line has keywords + yes okay extra line of keywords 1electron partly similar nprint= 2 print final hcore matrix 0scf yes okay kgeom =-1 for checking input data 1scf yes okay jop =-1 do one SCF and stop aider no stop read ab initio gradients aigin yes okay read ab initio geometry as Gaussian Z-matrix aigout no stop print ab initio geometry in Gaussian format am1 yes okay iop =-2 use AM1 method analyt yes okay ipsana= 1 analytic gradients author yes okay print author of program bar=n.n no stop option for *saddle* biradical yes okay imult = 1 half-electron 3*3 CI kci = 1 for singlet states bonds yes okay nsav16= 2 MOPAC bond orders nprint= 2 print bond orders c.i.=n partly okay kci = 1 if n=2, minimal CI stop if n.ne.2 camp no stop use camp-king method for SCF convergence charge=n yes okay kharge= n molecular charge compfg no ignore printing in compfg connolly no stop option for *esp* cycles=n yes okay maxrtl= n no. of cycles in nllsq dcart no ignore printing in dcart debug no ignore allow debug keywords debugpulay partly similar nprint= 2 option for *pulay* denout yes okay ipubo = 1 save p matrix on file 1 density partly similar nprint= 1 print final p matrix dep no stop generate new code for blockdata section depvar=n yes okay depfac= n special symmetry input for relation L2=18 deriv no ignore printing in deriv dforce yes okay jop = 5 do force constants jop = 6 if *nllsq* is specified lprint= 1 printing in force dfp yes okay idfp = 1 DFP update in flepo dipole no stop fit ESP to total dipole dipx no stop fit ESP to dipole (in x) dipy no stop fit ESP to dipole (in y) dipz no stop fit ESP to dipole (in z) dmax=n.nn yes okay dmax = .. trust radius for *ef* doublet yes okay imult = 2 doublet, RHF or UHF drc no stop dynamic reaction coord. dump=n no ignore restart file is saved automatically echo yes okay echo input data to user ef yes okay ief = 1 eigenvector following eiginv no ignore old option for *ef* eigs no ignore printing in iter enpart no stop energy partitioning esp no stop electrostatic potential esprst no stop option for *esp* esr no stop unpaired spin density RHF - not available UHF - *esr* not needed excited yes okay lroot = 2 optimize geometry for kci = 1 second-lowest singlet imult = 1 CI state external yes okay iparok= 1 read external parameters fill=n no stop force population of MOs flepo partly similar iprint= 1 printing in flepo fmat partly similar kprint= 1 printing in fmat fock partly similar nprint= 2 print final fock matrix force yes okay jop = 5 do force constants jop = 6 if *nllsq* is specified fulscf yes okay igrad = 1 full finite-difference gradients always geo-ok no ignore check on distances gnorm=n.n yes okay iconv =3 convergence criterion iprec =-10 on gradient norm gradients yes okay jop =-2 do one SCF + gradients graph yes okay nsav13= 1 save file for graphics hcore no ignore printing in hcore hess=n yes okay igthes= n hessian for *ef* h-prio no stop option for *drc* hyperf no stop hyperfine couplings interp no stop option for *camp*,*king* irc no stop intrinsic reaction coord isotope yes okay middle= 0 save force constants on file for restart iter no ignore printing in iter itry=n yes okay kitscf= n no. of SCF iterations iupd=n yes okay iupd=n hessian update for *ef* k=(n.nn,n) no stop brillouin zone structure kinetic no stop option for *drc* king no stop use camp-king method for SCF convergence large no ignore printing option let yes okay override safety checks jop = 2 for force constants and for input gradient norm linmin partly similar iprint= 1 printing in linmin localize no stop find localized MOs locmin partly similar iprint= 1 same as *linmin* max yes okay grid size 23*23 mindo/3 yes okay iop = 1 use MINDO/3 method meci no ignore printing in CI treatment micros no stop special CI input mmok yes okay immok = 1 mm correction, CONH mndo yes okay iop = 0 use MNDO method mode yes okay mode = n mode to follow in *ef* moldat no ignore printing in input part ms=n no stop spin component in CI mullik yes okay nsav16= 2 mulliken populations nprint= 2 print these populations nllsq yes okay jop = 1 minimize gradient norm jop = 4 when used with *optfor* jop = 6 when used with *force* noanci yes okay ipsana=-1 non-analytic CI derivs nodiis no ignore no DIIS in optimization nointer no ignore do not print distances nolog no ignore no entry in LOG file nomm yes okay no mm correction, CONH nonr yes okay lnonr = 1 no NR in *ef* nothiel yes okay lsub = 1 no FSTMIN line search noxyz no ignore do not print x,y,z coord nsurf no stop option for *esp* oldens yes okay ktrial= 1 read initial p matrix oldgeo yes okay keep previous geometry opci no ignore printing in CI part open partly okay imult =1-3 RHF open-shell input for ielec=ilevel=1-2, partly stop other more complicated RHF open-shell input oride no ignore old option for *ef* parasok yes okay iparok=-1 use some MNDO parameters in AM1 or PM3 pi no ignore sigma/pi analysis for p pl partly similar nprint= 2 print pl in iter pm3 yes okay iop =-7 use PM3 method point=n yes okay number of points on path point1=n yes okay number of points on grid point2=n yes okay number of points on grid polar yes okay polarizabilities potwrt no stop option for *esp* powsq no stop print option for *sigma* precise partly similar iprec =100 increase precision of various criteria pulay yes okay idiis = 1 use pulay method for SCF convergence quartet partly similar imult = 4 quartet, UHF only quintet partly similar imult = 5 quintet, UHF only recalc=n yes okay ireclc= n hessian recalc for *ef* restart yes okay middle= 1 continue previous job root=n partly okay lroot = n if n.le.3, minimal CI stop if n.gt.3 rot=n yes okay numsym= n symmetry number s1978 yes okay iparok= 4 use 1978 S parameters saddle no stop special ts search scale=n.nn no stop option for *esp* scfcrt=.n yes okay iscf = n SCF criterion on energy scincr=n.nn no stop option for *esp* search partly similar iprint= 1 same as *linmin* setup yes okay read keywords from file sextet partly similar imult = 6 sextet, UHF only shift=n no stop level shifting in SCF si1978 yes okay iparok= 4 use 1978 Si parameters sigma no stop use *nllsq* instead singlet yes okay imult = 0 closed-shell singlet slope no stop option for *esp* spin partly similar nprint= 1 print UHF spin matrix step=n.nn yes okay step size in path step1=n yes okay two-dimensional grid step2=n yes okay two-dimensional grid sto3g no stop option for *esp* symavg no stop option for *esp* symmetry yes okay ksym = 1 impose symmetry t=n yes okay limit = n time limit for job thermo yes okay ntemp = 0 thermodynamics after force calculation at default temperatures thermo(nnn) yes okay ntemp1= n *thermo* at temperatures ntemp2= k specified by (nnn) = (n) or (n,m) or (n,m,k) times no ignore print internal timings t-prio no stop option for *drc* trans no ignore not needed, only real vibrational modes are included by default triplet yes okay imult = 3 triplet, RHF or UHF ts yes okay ief = 1 locate transition state by *ef* procedure uhf yes okay iuhf = 1 do UHF calculation vectors partly similar nprint= 0 print final eigenvectors velo no stop option for *drc* williams no stop option for *esp* x-prio no stop option for *drc* xyz yes okay use x,y,z coordinates with MOPAC input in internal coordinates Further information on the treatment of the MOPAC keywords may be obtained from subroutine keydef which governs the response of this program to any given MOPAC keyword. When using at least one of the above MOPAC keywords (IMOPAC=1), the default values for certain variables differ from the standard values (defined for the standard input), in order to stay as close as possible to the conventions in MOPAC(6.0), e.g. limit=3600, ipl=4, iuhf=-1, and nprint=-1 (see subroutines zlimit, method, and start). Note, in particular, that odd-electron systems are treated as doublets (imult=2) by default when using MOPAC input (IMOPAC=1), whereas imult=2 must be specified when using standard input (to avoid an input error). Some additional keywords have been defined which are listed below. In the notation of this input description, the following keywords are regarded as MOPAC-type keywords although they are not defined in MOPAC(6.0). cndo iop= 2 use CNDO/2 method cosmo icosmo=1 use standard COSMO solvation treatment eofile - to indicate end of input for MNDO mndoc iop=-1 use MNDOC method with BWEN (kci=2) mndod iop=-10 use MNDO/d method with d orbitals mndodh iop=-13 use combination of MNDO/d and MNDO/H mndoh iop=-3 use MNDO/H method for hydrogen bonds molden nsav13=2 generate molden output file molden.dat om1 iop=-5 use OM1 method om2 iop=-6 use OM2 method om3 iop=-8 use OM3 method odm2 iop=-22 use ODM2 method odm3 iop=-23 use ODM3 method optfor jop= 3 optimization followed by force calculation jop= 4 when used with *nllsq*, ts search followed by force calculation sybyl nsav16=1 generate sybyl output file nb16, see nsav16 The preceding MOPAC-type keywords can be replaced by MNDO keywords (see section F). They are kept for convenience and for compatibility with previous program versions (MNDO94). ****************************************************************** ****************************************************************** H. File dictionary. 1 Sequential disk file. Used to store density matrices during job, if inout.gt.0. 2 Sequential disk file. Used to store the two-electron AO integrals in SCF section. Used to store the two-electron MO integrals in pert section. 3 Sequential disk file. Used to store the h,f matrices in SCF section, if inout.gt.1. Used to store ordered AO integrals in pert section. Used to store data for subsequent VB treatment, if ivbse.lt.0. 4 Sequential disk file. Used to store information for job continuation in geometry optimizations and force constant calculations. This file must be permanent if job continuation is desired. 5 Standard input file. 6 Standard output file. 7 Sequential disk file. Used for generating a new input file which contains either the current input geometry in a different format or the final optimized geometry from the current job. 8 Sequential disk file. Used for generating an output file which contains the final geometry of the current job and the atomic charges. 9 Sequential disk file. Used for generating an output file which can be used as input for evaluation programs that accept the pdb format. 10 Sequential disk file. Used to store the density matrix at the reference geometry in finite-difference gradient calculations. 11 Sequential disk file. Used for input of density matrix, if ktrial=1. Used for output of density matrix, if ipubo =1,3. Used for saving the density matrix, for job continuation. 12 Sequential disk file. Used for input of eigenvectors, if ktrial=2. Used for output of eigenvectors, if ipubo =2,3. 13 Sequential disk file. Used to store data for graphical evaluations according to MOPAC conventions. 14 Sequential disk file. Used for input of preliminary or alternative parameters. Available for all semiempirical methods implemented. Conventions and formats see following section. 15 Sequential disk file. Used for saving data after energy and gradient evaluations. Suitable as an interface to other programs and for debugging. 16 Sequential disk file. Used to store data for SYBYL postprocessing according to MOPAC conventions. 17 Sequential disk file. Used to store the inverse A matrix and the B matrix during the COSMO treatment (for later use in the gradient section) if the DIIS converger is applied simultaneously. 18 Sequential disk file. Used to store the OM2 arrays S, B, COR, HO evaluated at the reference geometry (for later use in the gradient section). 19 Sequential disk file. Used to store the reference data. 20 Sequential disk file. Used for input of data for calculations with external points, if mmfile=1. ------------------------------------------------------------------ File 1 will be used only if the available buffer is insufficient (as determined by the program) or if requested by option inout. File 2 will be used only if the available buffer is insufficient (as determined by the program). File 3 will be used only if the available buffer is insufficient (as determined by the program) or if requested by option inout or by option ivbse. File 4 will be used in geometry optimizations and force constant calculations (for middle.ge.0). File 7 will be used if requested by input option nsav7. File 8 will be used if requested by input option nsav8. File 9 will be used if requested by input option nsav9. File 10 will be used in finite-difference calculations for the gradient and electric properties (always). File 11 will be used in geometry optimizations and force constant calculations (for middle.ge.0). File 11 will be used if requested by input option ktrial/ipubo. File 12 will be used if requested by input option ktrial/ipubo. File 13 will be used if requested by input option nsav13. File 14 will be used if requested by input option iparok. File 15 will be used if requested by input option nsav15. File 16 will be used if requested by input option nsav16. File 17 will be used if requested by input option modcsm (COSMO). File 18 will be used if requested by input option inout (OM2). File 19 will be used if requested by input option inrefd. File 20 will be used if requested by input option mmfile. When using MOPAC input, output on files can be requested through the following keywords: geosave for file 7, pdbsave for file 9, graph for file 13, and sybyl for file 16. The numbers of the files can be redefined by changing the array nbf(20) in common block NBFILE, default values see BLOCKDATA BLOCK0: 1-20 for nbf(1)-nbf(20), convention nbi=nbf(i). ------------------------------------------------------------------ Internally the program may use several other files whose numbers are not governed by common block NBFILE. Standard conventions: File Standard Related Subroutine/program section unit number option where the file is accessed nb90 90 inrefd REFSAV, parameterizaton nb91 91 nsav16 GRAPH, SCF postprocessing nb92 92 setup GETTXT, MOPAC input section nb93 93 nsav13 MOLDIN, MOLDEN input file iumix 95 ipsana Analytical derivative code iurhs 96 ipsana Analytical derivative code iuk 97 ipsana Analytical derivative code iures 98 ipsana Analytical derivative code iures+1 99 ipsana Analytical derivative code The numbers of files 91, 92, and 93 cannot be changed via input. The input options iumix, iurhs, and iuk (see section 2.2) determine the corresponding file numbers (default values of 95, 96, and 97, respectively). The remaining two files in the analytical derivative section are used only if the input value of iures is positive (recommended: iures=98, see section 2.2). ****************************************************************** ****************************************************************** I. Input of external parameters. External parameters can be read from file nb14. This allows the definition of preliminary parameters for elements which do not yet have final parameters listed in the blockdata part and which are encountered in the input. File nb14 is read for iparok=1-3. Contents of file nb14: iparok=1 : Keyword-oriented MOPAC(6.0) conventions. iparok=2,3: Numerical input as described below. The options differ only in details of the numbering scheme. ------------------------------------------------------------------ iparok=1 : One line of input for each parameter to be defined. Each line is parsed for three items which must occur in the order (partyp,elemnt,param). partyp Parameter type which must match one of the predefined character strings. The currently available choices are given below under iparok=2 (95 parameter types). Compared with MOPAC(6.0) the list of available parameters types has been extended considerably. elemnt Element symbol which must match one of the standard chemical symbols (preceded by a blank character). param Numerical input value. End input with a blank line, with end-of-file, or with the string END (case-insensitive) for partyp. iparok=1 : Automatically selected for MOPAC-type input with keyword EXTERNAL (sets iparok=1). See MOPAC(6.0) manual for further details. ------------------------------------------------------------------ iparok=2,3: One line of input for each parameter to be defined. 1-3 ipar i3 Type of parameter. Conventions see below. = 0 End of input. 4-6 ielmnt i3 Atomic number of element to which the current parameter belongs. 10-24 param f15.8 Value of the parameter. iparok=2: Conventions for ipar (notation see below). iparok=2: First numbering scheme. 1 USS , 2 UPP , 3 ZS , 4 ZP , 5 BETAS, 6 BETAP, 7 ALPHA, 8 BETPI, 9 BETSH, 10 BETPH, 11 ALPS , 12 ALPP , 13 ALPPI, 14 ALPSH, 15 ALPPH, 16 FVAL1, 17 FVAL2, 18 GVAL1, 19 GVAL2, 20 FG , 21 UDD , 22 ZD , 23 BETAD, 24 ZSN , 25 ZPN , 26 ZDN , 27 POCOR, 28 GSCAL, 29 NUMAO, 30 XX30 , 31 FN11 , 32 FN21 , 33 FN31 , 34 FN12 , 35 FN22 , 36 FN32 , 37 FN13 , 38 FN23 , 39 FN33 , 40 FN14 , 41 FN24 , 42 FN34 , 43 SOLV1, 44 SOLV2, 45 SOLV3, 46 SOLV4, 47 SOLV5, 48 SOLV6, 49 ZSCOR, 50 FSCOR, 51 BSCOR, 52 ASCOR, 53 DELTA, 54 OMEGA, 55 XSCAL, 56 XOFFL, 57 XOFFG, 58 ZSSCF, 59 ZPSCF, 60 ZDSCF, 61 BSSCF, 62 BPSCF, 63 BDSCF, 64 XUSS , 65 XUPP , 66 XUDD , 67 ZSNMR, 68 ZPNMR, 69 ZDNMR, 70 BSNMR, 71 BPNMR, 72 BDNMR, 73 GSS , 74 GPP , 75 GSP , 76 GP2 , 77 HSP , 78 HPP , 79 EHEAT, 80 F0DD , 81 F2DD , 82 F4DD , 83 F0SD , 84 G2SD , 85 ALP01, 86 ALP06, 87 ALP07, 88 ALP08, 89 ALP09, 90 ALP14, 91 ALP15, 92 ALP16, 93 ALP17, 94 ALP35, 95 ALP53, 96 PDDG1, 97 PDDG2, 98 PDDG3, 99 PDDG4. iparok=3: Conventions for ipar (notation see below). iparok=3: Second numbering scheme (from parametrization program). iparok=3: Same list as for iparok=2 up to entry 72. 1 USS , 2 UPP , 3 ZS , 4 ZP , 5 BETAS, 6 BETAP, 7 ALPHA, 8 BETPI, 9 BETSH, 10 BETPH, 11 ALPS , 12 ALPP , 13 ALPPI, 14 ALPSH, 15 ALPPH, 16 FVAL1, 17 FVAL2, 18 GVAL1, 19 GVAL2, 20 FG , 21 UDD , 22 ZD , 23 BETAD, 24 ZSN , 25 ZPN , 26 ZDN , 27 POCOR, 28 GSCAL, 29 NUMAO, 30 XX30 , 31 FN11 , 32 FN21 , 33 FN31 , 34 FN12 , 35 FN22 , 36 FN32 , 37 FN13 , 38 FN23 , 39 FN33 , 40 FN14 , 41 FN24 , 42 FN34 , 43 SOLV1, 44 SOLV2, 45 SOLV3, 46 SOLV4, 47 SOLV5, 48 SOLV6, 49 ZSCOR, 50 FSCOR, 51 BSCOR, 52 ASCOR, 53 DELTA, 54 OMEGA, 55 XSCAL, 56 XOFFL, 57 XOFFG, 58 ZSSCF, 59 ZPSCF, 60 ZDSCF, 61 BSSCF, 62 BPSCF, 63 BDSCF, 64 XUSS , 65 XUPP , 66 XUDD , 67 ZSNMR, 68 ZPNMR, 69 ZDNMR, 70 BSNMR, 71 BPNMR, 72 BDNMR, 73 PDDG1, 74 PDDG2, 75 PDDG3, 76 PDDG4, 77 C6 , 78 R0 , 79 D3S6 , 80 D3S8 , 81 D3A1 , 82 D3A2 , 83 , 84 , 91 GSS , 92 GPP , 93 GSP , 94 GP2 , 95 HSP , 96 HPP , 97 , 98 , 99 EHEAT, 100 CORE , 101 , 102 , 121 F0DD , 122 F2DD , 123 F4DD , 124 F0SD , 125 G2SD , 126 F0PD , 127 F2PD , 128 G1PD , 129 G3PD , 130 , 131 , 132 . 201 ALP01, 202 ALP02, 203 ALP03, 204 ALP04, 205 ALP05, 206 ALP06, 207 ALP07, 208 ALP08, 209 ALP09, 210 ALP10, etc up to 286 ALP86. Comments on notation: USS,UPP,UDD One-center energies for s,p,d (eV) ZS,ZP,ZD Orbital exponents for s,p,d (au) BETAS,BETAP,BETAD Beta parameters for s,p,d (eV) ALP Alpha core-core parameter (1/Angstrom) BETPI,BETSH,BETPH Additional beta parameters, OM1/OM2 (eV) ALPS,ALPP..ALPPH Resonance integral parameters, OM1/OM2 (au) FVAL1,FVAL2 Prefactors for 2-center orthogonalization terms in OM1/OM2 (dimensionless) GVAL1,GVAL2 Prefactors for 3-center orthogonalization terms in OM2 (dimensionless) ZSN,ZPN,ZDN Auxiliary MNDO/d exponents for s,p,d (au) POCOR Additive term for core (au), set equal to additive term for ss if not read in GSCAL Scaling factor for the core-core repulsion used in AM1/d-PhoT and AM1/d-CB1 NUMAO Number of basis functions for given atom Needed for spd-basis, use 9.0 in this case XX30 Superseded by NUMAO, no longer supported FN11..FN34 AM1(PM3)-type core repulsion terms SOLV1..SOLV6 Parameters for solvation models ZSCOR Orbital exponent for 1s-core, OM2 (au) FSCOR One-center energy for 1s-core, OM2 (eV) BSCOR Beta parameter for 1s-core, OM2 (eV) ASCOR Resonance integral parameter (1s), OM2 (au) DELTA,OMEGA Parameters for electrostatic potential (Angstrom,1/Angstrom) XSCAL Scale factor for absolute NMR shieldings XOFFL,XOFFG Offsets (ppm) for NMR shifts: Liquid, gas ZSSCF,ZPSCF,ZDSCF Orbital exponents for s,p,d (au), NMR-SCF BSSCF,BPSCF,BDSCF Beta parameters for s,p,d (eV), NMR-SCF XUSS,XUPP,XUDD One-center energies for s,p,d (eV), NMR-SCF ZSNMR,ZPNMR,ZDNMR Orbital exponents for s,p,d (au), NMR integs BSNMR,BPNMR,BDNMR Beta parameters for s,p,d (eV), NMR integs GSS,GSP,GPP,GP2,HSP One-center two-electron integrals (eV) EHEAT Exp heat of formation of the atom (kcal/mol) CORE Core charge of the atom (e) F0SD,G2SD,F0DD..G3PD Slater-Condon parameters (eV), the current code processes input only for F0SD and G2SD ALP01..ALPxy..ALP86 MNDO/d alpha bond parameters (1/Angstrom) involving partner with atomic number xy PDDG1,PDDG2 PDDG-type core repulsion terms: prefactors PDDG3,PDDG4 PDDG-type core repulsion terms: distances D3S6,D3S8,D3A1,D3A2 Parameters for D3 dispersion correction ****************************************************************** ****************************************************************** J. Input for HDLC optimizer. The following input options are relevant for HDLC optimizations: section D.1 - jop, igeom, ief; section D.2 - maxend, middle, iprint, iprec, iconv, nrst; section D.2 - ihdlc1, ihdlc2, ihdlc3; section D.2.1 - mode, ireclc, iupd; section D.2.1 - dmax, ddmin, ddmax, rmin, rmax, omin. These options are described as part of the standard input. Additional data are read whenever the HDLC optimizer is accessed (options jop=0,1,3,4 and ief.lt.0). These data are provided on an external file nb42 or as an appended input block in the standard input file nb5 (last section for a given molecule) depending on input option ihdlc3: ihdlc3 = 0 Read from external input file nb42. ihdlc3 = 1 Read from standard input file nb5. If the input file cannot be opened, default HDLC options are used, and normally an unconstrained geometry optimization is performed. If there is no symmetry and no dummy atom, an attempt is made to satisfy the constraints implied in the initial input geometry. The input of these additional data is keyword-driven, with the keyword aligned to the left, followed by an equal sign and a value, e.g.: NREMST=81. The order of keywords is arbitrary. Some keywords require additional input on the following line(s) which is always unformatted. If atom sequence numbers need to be provided as additional input, they always refer to the sequence of atoms in the input geometry excluding dummy atoms. Convention for the remainder of this section: n,i,j,k,l,irad denote integer input data. f,target,vrad denote real input data. The following keywords are defined: NFCART=n Specify n frozen atoms. The following line(s) must provide n integer sequence numbers of the frozen atoms. NFCOMP=n Specify n constrained Cartesian coordinates. Each of the following n lines must provide one constraint in the form (i,j,target): i Sequence number of the atom involved. j Type of coordinate (x 1, y 2, z 3) target Corresponding target value (Angstrom). Special remark concerning target: The input for target is optional and may be omitted in the current program version which ignores any input of target and employs the corresponding value from the initial input geometry. Later program versions may use the input for target (upward compatibility). NFBOND=n Specify n constrained bond lengths. Each of the following n lines must provide one constraint in the form (i,j,target): i,j Sequence numbers of the two atoms involved. target Corresponding target value (Angstrom). Special remark concerning target: See above (NFCOMP). NFANGL=n Specify n constrained bond angles. Each of the following n lines must provide one constraint in the form (i,j,k,target): i,j,k Sequence numbers of the three atoms involved. target Corresponding target value (degree). Special remark concerning target: See above (NFCOMP). NFDIHE=n Specify n constrained dihedral angles. Each of the following n lines must provide one constraint in the form (i,j,k,l,target): i,j,k,l Sequence numbers of the four atoms involved. target Corresponding target value (degree). Special remark concerning target: See above (NFCOMP). NMRESI=n Define a residue/fragment/molecule with n atoms. The following line(s) must provide n integer sequence numbers to assign these atoms. Coordinates are delocalized within each residue only. The maximum number of residues is unlimited. Each residue must consist of at least two atoms. NRCORE=n Specify the n-th residue as the reaction core. A microiterative transition state search is performed. The default number of geometrical variables in the reaction core with nrc atoms is set to 3*nrc. This default may be changed using NDFCOR (see below). Internally the atoms are renumbered such that the reaction core becomes the first residue. NDFCOR=n Specify the number of geometrical variables in the reaction core and thereby implicitly select the atoms carrying the first n variables as the reaction core. If NRCORE is not used, all atoms of the reaction core must belong to the first residue of the input geometry. A microiterative transition state search is performed. =0 Default: all variables (ndf=3*nrc) in the case of a transition state search (selected by jop=1,4), and 0 otherwise (no reaction core defined). XYZFAC=f Weighting factor of the Cartesian coordinates for the delocalization to hybrid delocalized coordinates HDLC. =0 Weighting factor = 1.0/real(ndf), default value. >0 Weighting factor = f <0 Weighting factor = abs(f)/real(ndf) ndf: number of degrees of freedom in given residue. MDCRAD=n Type of covalent radii of the elements used for the calculation of the connectivity. =0 Default: all atoms are treated as hydrogen. This produces a small set of redundant primitives based on a connectivity dominated by the shortest distance branched path. =1 Use values for the covalent radii in the HDLC code: one average value for each row in the periodic table. =2 Use the covalent radii provided as input (see NCRAD). NCRAD=n Read input values of n covalent radii. Each of the following n lines must provide one such radius in the form (irad,vrad): irad nuclear charge vrad covalent radius (Angstrom) NREMST=n Number of remembered L-BFGS steps. =0 Default value = min(40,ndf) ndf: number of degrees of freedom in given residue. NRSHDL=n Option to force a periodic restart of the optimizer. After every n energy and gradient evaluations, the optimizer is restarted, and the Hessian of the reaction core is recalculated if required. =0 Default: never force a restart. Note: Restarts can still occur if the HDLC of a given residue break down or if the L-BFGS history does not ensure an efficient optimization anymore. Input is terminated when an end-of-file or an empty input line is encountered. ****************************************************************** ****************************************************************** K. Input for molecular dynamics and surface hopping driver For molecular dynamics runs with the option icross=6, the dynamics options are specified using the auxiliary input file "dynvar.in", in the form of a Fortran namelist with the following general format: &DYNVAR option1 = value1, option2 = value2, / The option names are case-insensitive. If no input file "dynvar.in" is provided, the default options given below are used, and a default file "dynvar.in" is generated. The dynamics options are as follows: Variable Type Default Description iout Int 0 Print level (0-3) nstep Int 10 Number of MD steps dt Double 5.0D-5 Time step (in picoseconds) temp0 Double 300.0D0 Initial temperature (K) norot Logical TRUE Flag to adjust initial velocities to remove overall angular momentum of the system notra Logical TRUE Flag to adjust initial velocities to remove overall linear momentum of the system init_stat Int 1 Initial state write_stats Logical TRUE Flag for writing of statistics sunit Int 30 Fortran unit for statistics file sfile Char(20) stat.out File name for statistics write_traj Logical TRUE Flag for writing of trajectory xunit Int 31 Fortran unit for trajectory file xfile Char(20) traj.out File name for trajectory write_vel Logical FALSE Flag for writing of velocities vunit Int 32 Fortran unit for velocity file vfile Char(20) vel.out File name for velocities ene_tol Double 1.0D-2 Obsolete option fstat Int 1 Frequency for saving statistics to file = 1 every step = n every n steps avstat Logical See text Flag for turning on statistics averaging over fstat = n steps If n > 1, the default is TRUE, else FALSE fsav Int 1 Frequency for saving trajectory and velocities to file = 1 every step = n every n steps restart Logical FALSE Flag to indicate if restarting dynamics from a restart file write_rest Logical TRUE Flag for writing a restart file rfile Char(20) dynam.restart Restart file name ehrenfest Logical FALSE Flag to enable Ehrenfest dynamics tully_hop Logical FALSE Flag to enable Tully surface hopping simhop Int 0 Flag to enable simple hopping = 0 No simple hopping = 1 diabatic surface hopping = 2 diabatic surface hopping with average value of coupling (AS1) = 3 average coupling compared with random number = 4 average coupling compared with random number + velocity adjustment (AS2) = 5 average coupling compared with random number, hop only with negative coupling + velocity adjustment = 6 average coupling correction; average coupling compared with random number + velocity adjustment (AS3) *** Note: ehrenfest,tully_hop and simhop are mutually exclusive. einteg Char(5) UP3 Integration algorithm for use with the Tully electronic propagation UP1 = Unitary exponential propagator computed at midpoint. Eigenvalue expansion used to build the exponetial UP2 = Unitary exponential propagator computed at the beginning of the interval. Eigenvalue expansion used to build the exponetial UP3 = Unitary exponential propagator computed at midpoint. Taylor expansion used to build the exponential ABM5 = Adams-Bashforth-Moulton 5-th order predictor-corrector with local truncation error evaluation. Also iterative SCF cycles are performed if error is greater than 1.D-10 ABM4 = Adams-Bashforth-Moulton 4-th order predictor-corrector with local truncation error evaluation. Also iterative SCF cycles are performed if error is greater than 1.D-10 ABM2 = Adams-Bashforth-Moulton 5-th order predictor-corrector RK2 = second order Runge-Kutta RK4 = fourth order Runge-Kutta euler= Euler method (only for testing purposes) ne Int 200 Number of steps for integration of the Tully electronic equation dec_cor Double 0.1D0 Decoherence correction 0.1 = Granucci & Persico empirical decoherence correction, standard value (hartree) -1 = Shenvi, Subotnik & Yang parameterless decoherence correction 0.0 = switch off decoherence correction cuthop Double 0.0D0 Do not allow hops at energy differences larger than cuthop (in kcal/mol) to avoid unphysical hops =0 : (standard) do not use this feature >0 : simple cut at cuthop value <0 : cut by gaussian scale of probability centered at cuthop value num_cc Logical FALSE Flag for calculation of numerical non-adiabatic couplings an_cc Logical TRUE Flag for calculation of analytical non-adiabatic couplings *** Note: num_cc and an_cc are not mutually exclusive, but if both are selected the analytical couplings are used in the Tully procedure. If only numerical couplings are selected, analytical couplings must still be calculated each time a hop occurs. nac_ph Double 0.1D0 Obsolete option. rnd_gen Char(8) standard Algorithm for generating random numbers for initial velocity assignment and for the Tully surface hopping algorithm standard = intrinsic Fortran subroutine (RANDOM_NUMBER) NB: compiler-dependent PM_BD = Park and Miller algorithm with Bays-Durham shuffle knuth = Knuth subtractive algorithm rnd_seed Int -1 Integer to seed the random number generator -1 = random seed (e.g. based on the system clock) NB: For rnd_gen = standard, each member of the seed array is set to rnd_seed fol_stat Logical TRUE Flag to enable state following write_hop Logical See text Flag for writing of hopping data Default is TRUE if Ehrenfest or Tully dynamics are selected, FALSE otherwise hopunit Int 33 Fortran unit for hopping data file hopfile Char(20) hopping.out File name for hopping data fhop Int 1 Frequency for saving hopping data to file = 1 every step = n every n steps vs Logical FALSE DEPRECATED ene_cons Logical FALSE Flag for total energy conservation. Deviations are handled by an exponential velocity scaling. ene_scal Double 1.0D0 Scaling factor for energy conservation. The velocities are scaled, so that the energy difference between total energy and initial total energy is reduced by factor ene_scal in 1 fs. thermostat Int 0 Choice of thermostat. = 0 Do not use a thermostat = 1 Nose-Hoover (chain) thermostat thermo_T Double 300.0D0 Target temperature for thermostat thermo_len Int 1 Length of Nose-Hoover chain = 1 Nose-Hoover thermostat = n Nose-Hoover chain thermostat with chain length n. thermo_tau Double 1.0D0 Characteristic time of Nose-Hoover thermostat in ps. thermo_equi Int 0 Number of equilibration steps. The writing of the trajectory file will start after the equilibration ends. g0 Double 0.0D0 Damping coefficient. A positive value selects Langevin dynamics. Unit: 1/ps. ADAPTIVE TIME STEP REDUCTIONS: If one of the criteria (adapt_ene or adapt_map) is activated, the MD simulaton will revert a calculation step and try a smaller time step, if these criteria are not fulfilled. adapt_map Int 0 Threshold for orbital mapping procedure. If one mapping of active orbitals is smaller than threshold, the step size will be reduced. High values (95 or higher) can usually be selected. = 0 turn off this criterion adapt_ene Double 0.0D0 Absolute (kcal/mol) or relative threshold for total energy conservation (see the following option adapt_ene_mode). If change is bigger than threshold, step size gets reduced. Low values can usually be selected, e.g. 0.0001 or lower for relative threshold, 0.01 or lower for absolute threshold; = 0.0 turn off this criterion adapt_ene_mode Int 0 Select if the energy threshold value in adapt_ene is relative or absolute. = 0 adapt_ene is relative threshold (e.g. 0.01 corresponds to a maximum change of 1% between two steps). = 1 adapt_ene is absolute threshold (kcal/mol). adapt_tries Int 7 Maximum number of time step reduction attempts. adapt_proceed Logical FALSE Program behaviour if a time step reduction is not successful after adapt_tries attempts. = T The calculation will proceed even if one of the criteria is not fulfilled after adapt_tries reductions. mapthr is then internally set to 1. = F The calculation will stop after adapt_tries reductions. File dictionary (with default Fortran units). 30 Sequential disk file (default name: stat.out) Used to store statistics of the dynamics run. 31 Sequential disk file (default name: traj.out) Used to store trajectory. 32 Sequential disk file (default name: vel.out) Used to store velocities. 33 Sequential disk file (default name: hopping.out) Used to store hopping data. In addition, there is a restart file (default name: dynam.restart) ****************************************************************** ****************************************************************** L. Input for Born-Oppenheimer ground-state molecular dynamics For molecular dynamics runs with the option jop=-3, the dynamics options are specified using the auxiliary input file "dynvar.in", in the form of a Fortran namelist with the following general format: &DYNVAR option1 = value1, option2 = value2, / The option names are case-insensitive. If no input file "dynvar.in" is provided, the default options given below are used, and a default file "dynvar.in" is generated. The Born-Oppenheimer ground-state MD code has been generated by simplifying the general multi-state surface-hopping MD code. *** Nose-Hoover thermostat not available in the simplified code. *** Instead, the code only offers simple velocity scaling. *** Most options from section K kept, same meaning and defaults; their description is copied below for the sake of covenience. *** Options from section K removed that are no longer supported, including multi-state surface-hopping and thermostat options. *** New options for velocity scaling (vs, Tfix, fvs, vs_Emax). The dynamics options are as follows: Variable Type Default Description iout Int 0 Print level (0-3) nstep Int 10 Number of MD steps dt Double 5.0D-5 Time step (in picoseconds) temp0 Double 300.0D0 Initial temperature (K) norot Logical TRUE Flag to adjust initial velocities to remove overall angular momentum of the system notra Logical TRUE Flag to adjust initial velocities to remove overall linear momentum of the system init_stat Int 1 Initial state write_stats Logical TRUE Flag for writing of statistics sunit Int 30 Fortran unit for statistics file sfile Char(20) stat.out File name for statistics write_traj Logical TRUE Flag for writing of trajectory xunit Int 31 Fortran unit for trajectory file xfile Char(20) traj.out File name for trajectory write_vel Logical FALSE Flag for writing of velocities vunit Int 32 Fortran unit for velocity file vfile Char(20) vel.out File name for velocities ene_tol Double 1.0D-2 Obsolete option fstat Int 1 Frequency for saving statistics to file = 1 every step = n every n steps avstat Logical See text Flag for turning on statistics averaging over fstat = n steps If n > 1, the default is TRUE, else FALSE fsav Int 1 Frequency for saving trajectory and velocities to file = 1 every step = n every n steps restart Logical FALSE Flag to indicate if restarting dynamics from a restart file write_rest Logical TRUE Flag for writing a restart file rfile Char(20) dynam.restart Restart file name rnd_gen Char(8) standard Algorithm for generating random numbers standard = intrinsic Fortran subroutine (RANDOM_NUMBER) NB: compiler-dependent PM_BD = Park and Miller algorithm with Bays-Durham shuffle knuth = Knuth subtractive algorithm rnd_seed Int -1 Integer to seed the random number generator -1 = random seed (e.g. based on the system clock) NB: For rnd_gen = standard, each member of the seed array is set to rnd_seed vs Logical FALSE DEPRECATED ene_cons Logical FALSE Flag for total energy conservation. Deviations are handled by an exponential velocity scaling. ene_scal Double 1.0D0 Scaling factor for energy conservation. The velocities are scaled, so that the energy difference between total energy and initial total energy is reduced by factor ene_scal in 1 fs. Tfix Double temp0 Temperature to be reached by velocity scaling. Default temp0 is input via parameter list. fvs Int 1 Frequency for performing velocity scaling = 1 every step = n every n steps vs_Emax Double -1.0D0 Maximum kinetic energy correction allowed in velocity scaling Default of -1.0D0 implies: no limit. g0 Double 0.0D0 Damping coefficient. A positive value selects Langevin dynamics. Unit: 1/ps. File dictionary (with default Fortran units) - same as in section K: 30 Sequential disk file (default name: stat.out) Used to store statistics of the dynamics run. 31 Sequential disk file (default name: traj.out) Used to store trajectory. 32 Sequential disk file (default name: vel.out) Used to store velocities. 33 Sequential disk file (default name: hopping.out) Used to store hopping data. In addition, there is a restart file (default name: dynam.restart)

♦ Imprint ♦ Privacy policy ♦