The Gaussian Interface
Introduction
The Gaussian interface permits the use of Gaussian 09
for the generation of energies and gradients. Most functions follow the
generic quantum interface; only Gaussian-specific
details are given here.
In order to use the interface, the environment must be set up as required for
Gaussian, with the g09 command in the user's path.
This interface is designed to access the basic functionality of the Gaussian
program package. It is capable of calculating energies and gradients on the HF
and DFT ( blyp, bp86, b3lyp, b3p86, b3pw91, bpw91, svwn) level of theory.
Serial as well as SMP-parallel versions (using nproc=#) are supported.
The following arguments to the gaussian module
are additional to those described in the section on
generic QM interfaces
| Argument |
Argument type |
Mandatory |
Default |
To specify |
| exe= |
keyword |
no |
g98 |
Name of the Gaussian executable. Has to be in
the path. |
| jobname= |
string |
no |
gaussian |
Name of the checkpoint and related files. |
| read_chk= |
filename |
no |
jobname |
A user defined checkpoint file (without extension) which is read by user request.
The file should be in the binary format with the *.chk extension. |
| alter_mos= |
Tcl List
|
no |
undefined |
Used to alter the order of MOs from a checkpoint file, e.g. for the specification of the active space in CASSCF methods.
Equivalent to a Gaussian Guess=Alter job.
The list should be of the form { {mo1 mo2} {mo3 mo4} {..} }, where mo1 will be swapped with mo2, mo3 with mo4, etc. |
| g98_mem= |
integer |
no |
2000000 |
The memory is interpreted as words (8N bytes). The default provides 16 MB of memory.
N may also contain a units specification (e.g. g98_mem= "2 GB") |
| nproc= |
integer |
no |
1 |
The "%nproc=N" line is added to the route 0 section of the gaussian input file. |
| hamiltonian= |
keyword |
no |
hf |
Choice of QM Hamiltonian: hf, blyp, bp86,
b3lyp, b3p86, b3pw91, bpw91, s-vwn, lda, casscf. |
| cas_opts= |
keyword |
no |
empty |
Specification of CASSCF options, e.g. {6,6} will specify a CASSCF(6,6) job.
Mandatory if hamiltonian=casscf.
NB: NRoot should be specified separately using the cas_nroot keyword. |
| cas_nroot= |
keyword |
no |
1 |
CASSCF state of interest for single state calculations (ignored for multiple state calculations). |
| basis= |
keyword |
no |
sto3g |
The following keywords: sto3g, 321g, 431g, 631g, 631gd, 631gdp, lanl2dz, sddall, dz, and dzp; respectively set the basis sets: STO-3G, 3-21G, 4-31G, 6-31G, 6-31G(d), 6-31G(d,p),
LANL2DZ, SDDALL, D95, and D95(d,p). |
| basisspec= |
Tcl List
|
no |
undefined |
The basis set can be read in as a tcl list, which
is appended to the gaussian input file.
The gen basis set is added to the gaussian input file route line. See Basis Manipulation. |
| ecpspec= |
Tcl List
|
no |
undefined |
An ecp can only be used when basisspec
is provided. The ecp data is appended to the basis set data and the keyword
Pseudo=Read is added to the route line of the gaussian input file. See Basis Manipulation. |
| optstr= |
string |
no |
empty |
A list for optional Gaussian keywords which the user can add directly |
| mode= |
string |
no |
undefined |
The mode is either undefined or alternatively can be set to "energy". |
| scfconv= |
integer |
no |
undefined |
Choice of the SCF convergence criteria. The integer (N) sets the convergence to 10-N.
If the default undefined value is used, the scf convergence is controlled by the argument mode. If mode is undefined the scf convergence defaults to tight,
if mode is set to energy the scf convergence will be set to SP. |
| maxcyc= |
integer |
no |
64 |
The maximum number of scf cycles to be performed. |
QM/MM electrostatic embedding calculations with Gaussian
Unlike most other QM programs, Gaussian calculations with point charges
include a contribution from point charge/point charge electrostatic interactions.
In ChemShell these terms are calculated by the MM theory and so they must
be subtracted from the Gaussian result to avoid double counting.
The following is a typical ChemShell output showing the corrections:
[1] Contribution to energy from gaussian: -117.766679 (a.u.)
[2] Contribution to energy from dl_poly: 0.009479 (a.u.)
[3] Contribution to energy from additional MM energy terms: 0.043490 (a.u.)
[4] Contribution to energy from additional charges in gaussian: 2.727874 (a.u.)
-----------------------------------------------------------------------------------
[5] QM/MM Energy: -114.985836 (a.u.)
-----------------------------------------------------------------------------------
Term [1] is the uncorrected Gaussian energy. Term [3] is the correction
applied for point charges corresponding to MM atoms and term [4] is the correction
due to additional charges created by the link atom method. In general term [1]
will not agree with other QM codes but the total corrected energy [5] should be in
agreement.
|
