g_rhf: Program for the calculation of G tensors from RHF data resulting from an AM1, PM3 or MNDO RHF MO calculation using AMPAC or MOPAC or a ZINDO/S RHF MO calculation using Hyperchem Results are the different parts of the G tensor in the axis system the LCAO coefficients were calculated in as well as the eigenvalues and normalized eigenvectors. If you should use this program to create material that you publish I expect you to include proper attributions and also would like to ask you to inform me about where it has been published. Introduction ============ Unfortunately, there is no real documentation for this program. All you will find in this file is a short description of the command line options and how to install the program (UNIX only). If you want to know more about the theoretical background please have a look at J. T. Toerring, S. Un, M. Knuepling, M. Plato, K. Moebius, "On the calculation of G-tensors of organic radicals", J. Chem. Phys. 107, 3905 (1997) Some more detailed information about the algorithms and the literatur used can be found in my PhD thesis - a PDF or Postscript version can be downloaded from http://www.physik.fu-berlin.de/~toerring (but it's in German, sorry). Input Files and Requirements ============================ The program doesn't do a MO calculation by itself but relies on the results of a preceeding semiempirical AM1, MNDO, PM3 or ZINDO/S RHF calculation. Currently, the program does only understand the output files of the AMPAC, MOPAC or Hyperchem (for ZINDO/S) package - other packages (e.g. Gaussian) can not be used! Also problems with different versions of AMPAC or MOPAC may occur - the layout of the output files seem to change quite frequently. If the program does not work and complains about its input please send me an email with the input file! To make sure all the information needed by the program can be found in the output file of the MO calculation the calculation has to be run with the 'ESR' and 'VECTORS' keywords (otherwise only a few of the orbital energies and LCAO coefficients are printed into the output file). Do *NOT* use the 'NOXYZ' keyword! Command line Options ==================== 1. Default is to neglect of two- and three-center terms in the calculation of expectation value of the orbital momentum. With the following flags you may override the default. I would strongly recommended using one of these flags (preferable '-m') because Stone's approximation seems to be rather poor in most cases. -t use two-center contributions in the calculation of expectation value of the orbital momentum -m use two- and three-center terms in the calculation of expectation value of the orbital momentum 2. To reduce the computation time the maximum distance between atoms included in the calculation of two- and three-center terms can be reduced via the '-D' flag. The flag has to be followed by a positive number, i.e. the maximum distance in Angstroem. I found empirically that reducing the maximum distance to 3 Angstroem doesn't lead to any appreciable changes in the final results. -D x reduce maximum distance between atoms included in the calculation of two- and three-center terms to x Angstroem 3. Following Stone, as the default differences of orbital energies are used in the denominator of the perturbation terms. Actually, the correct excitation energies differ from the differences of orbital energies and I got much better results with using corrections to the energy terms. -e use corrections to the differences of orbital energies to get better approximations for the excitation energies 4. Stone used just the second order terms but no relativistic or diamagnetic terms. To get the results from the traditional Stone formula just don't use the '-t' or '-m' flags (i.e. use just the one-center terms) and also don't use the '-e' flag but use '-S' instead (i.e. also discard relativistic and diamagnetic contributions). -S discard relativistic and diamagnetic contributions as in Stones paper 5. I applied some minor changes to the way the diamagnetic contribution is calculated (as compared to Angstl's method). To use Angstl's formula the '-A' flag has to be given. -A use Angstl's formula in the calculation of the diamagnetic contribution 6. For comparison purposes the paramagnetic contribution can be reduced by a factor of 2 as in Stones formulas. -w use only half of paramagnetic contribution 7. If you want a listing of all excitation energies and paramagnetic contributions to the G tensor you can use the '-le' and '-lp' flags. The data will be written to a files with the same name as the input data file but with the extension changed to '.energy' and '.paramag', respectively. -le write excitation energies into file with extension '.energy' -lp write paramagnetic contributions into file with extension '.paramag' 8. Following the flags the name of the input file has to be specified, optionally followed by the name of the output file (if no output file is specified the results are written to standard output). Installation ============ This program will only run under a UNIX operating system. To compile it you need a working standard C compiler, e.g. gcc. To run the program you will also need the a newer Perl interpreter (the scripts expect it to be installed in /usr/local/bin, if it is installed somewhere else you will have to change the first line of all perl scripts). I tested this program only under Linux and Digital Unix (or whatever it's name is today) but I hope that I didn't use any non-standard features and other variants of UNIX will also do. Before you start the compilation please have a look at the first lines in the Makefile. Here you can set the final location for the program and the auxilary files needed (it wont run if you don't install it in this location, sorry). You may also have to change the settings for the compiler and it's flags. When you're satisfied with the settings type 'make' and, when the compilation eneded succesfully, type 'make install' - this will copy the program and the auxilary files to their final locations. Now you may try to run the program, for example by typing g_rhf -m -D 3 -e quinone.out quinone.res ('quinone.out' is the output of a MOPAC AM1 calculation for a quinone radical and this call will (hopefully) create an output file 'quinone.res' with the calculated G tensor of this molecule.) 19-8-2001 Jens Thoms Törring AG Möbius Institut für Molekülphysik Freie Universität Berlin Arnimallee 14 D-14195 Berlin Germany Jens.Toerring@physik.fu-berlin.de