|
Many of the parameters in Wingam can be set by the user in the configuration file wingam.ini. A template for this is presented when Wingam is started for the first time. It is opened in the editor for you to adapt it to your system and save it. The most important items are
1) the path name of the directory where you have installed the WinGamess distribution, 'C:\WinGAMESS'.
2) the correct version of the WinGAMESS binary to be used, perhaps '08'.
The configuration file can later be edited by clicking on 'Configure/Conf File' in the Master Menu. Save changes. You might have to exit and restart wingam for the changes to take effect.
The directories of the file editor, the browser, the pdf-Reader, the builders, and the structure viewers must be specified in the Configuration file. Four builders may be defined. Set their call specifications, and the name to be associated with those builders. Similarly, four structure viewers can be specified. The calls for them are to be set and their names specified. If the viewer is capable of reading an output file upon loading, the value of the file extension may be set to, e.g. “pdb” or “out” (for 'Molekel' and 'Molden'). If it is set to “”, no file will automatically be loaded when the viewer is opened. It has to be specified by the usual 'File/Open' dialogue.
Some of the programs mentioned in this manual have licensing restrictions. Be sure to consult the license agreements before using the programs.
Although any text file editor that reads large files can be used, I have chosen 'Notepad.exe'. You can choose another, e.g. 'Wordpad.exe', and give the correct calling string. Only for the Configuration file 'Notepad.exe' from Windows has to be used.
Options in Wingam
BUILDER
The program used to draw and build the molecule is set in Wingam.ini. Four builders may be set in the program. For now only one builder is actually installed: ArgusLab.exe4.
VIEWER
Four viewers may be specified in Wingam.ini. If the job is saved, the output structures will be converted to .pdb format with the name ‘jobname’.pdb and in .xyz format with the name ‘jobname’.xyz. Wingam has been tested with gOpenMol10, RasWin12, Molekel13, Jmol15, and Molden16, but other viewers may be used as well. Electron densities and molecular vibrations can be generated in Molekel, gOpenMol, Jmol, and Molden. Files for creating high quality Pov-Ray™14 ray-traced images can be generated in VMD©11. One of the best and time-proven viewers is Molden16 (for Linux and Windows).
JOB Name
The name of the JOB must be entered here. The files ‘jobname’.out, ‘jobname’.pdb, ‘jobname’.pun, and ‘jobname’.xyz will be generated in the output directory if the job is saved. The input coordinate file must also have the name ‘jobname’.xxx as specified below. In order to prevent unwanted overwriting of output files, ‘jobname’ here always means the given 'Job Name' plus a four digit number representing the time in hr:min of the start of the computation.- If you [Outputdata/Read *.out] (Master Menu) an earlier run its jobname is automatically placed into the 'Job Name' box.
BUILD
The BUILD button calls up the builder specified above. The molecule must be converted to 3D and saved as ‘jobname’.mol, ‘jobname’.ent, ‘jobname’.pdb, or ‘jobname’.xyz in the input directory (specified automatically). A previous GAMESS output file, named ‘jobname’.out, may also be used for coordinate input. In the Master Menu click on [Properties, Show/Print], click on the output file desired and then choose 'Coordinates'. This generates a file 'gamess.coo' which can be read in when composing the input file, choosing 'COO' as Coordinate Type. In Wingam input/output files are in separate directories called ./data and ./output.
GAMESS INPUT
The Gamess Input button allows one to set many of the input parameters for gamess, thus defining the nature of the quantumchemical calculation to be performed. The default parameters are set for a simple structure optimization job of a small to medium size, uncharged, singlet molecule. At a minimum the RUNTYP, BASIS, and COORDINATE TYPE will have to be set for each run. A coordinate input file must be in the ./data directory unless it is saved from a previous run in 'gamess.coo' or 'gamess.equ' in the Wingam directory. Coordinates can also be entered manually, see below.
Some of the GAMESS options require additional input that cannot (yet) be entered through Wingam. It is the responsibility of the user to review the input file generated by Wingam and make any additions or changes that are necessary for a successful run. I cannot guarantee that the settings that are placed in the input file will necessarily give a meaningful run. Be sure to read the GAMESS documentation files, particularly INPUT.DOC. A copy of INPUT.DOC is placed in the Wingam directory, and it can be accessed from HELP of the Master Menu. Note that this is the most recent INPUT.DOC from the Gamess-US distribution.
Summary of GAMESS INPUT Options
Items in bold - or a red bar on the panel - are the ones that are changed most often.
Gamess input orders the keywords into "groups", beginning with a "$" in the second
position of a new line and ending with " $END". Groups may be given in any order (and even repeated on another line), keywords within groups, too. The following list of groups and their keywords is implemented in the listboxes of Wingam. Keywords - parameters of the calculation - may be set by selecting appropriate values in those boxes. There are many more groups and keywords available for more sophisticated jobs, see INPUT.DOC. Names of groups and keywords are not case sensitive. Keywords within a group must be separated by space or newline.
$DATA
|
INPUT COORDINATE TYPE
|
This is by far the most important entry for any type of quantumchemical computation: The structure of a molecule, i.e. the spacial arrangement of its constituent atoms, determines its energy and most other properties. If not given carefully the whole effort may be meaningless.
|
|
|
NONE
|
The coordinates are to be entered manually by the user when 'View(ing) Input File'. This is especially necessary if you want to give internal coordinates as Gaussian Z-matrix 'ZMT' or Mopac Z-matrix 'ZMTMPC' format which the script is not able to provide, see 'Input Description' under [Help] in the Master Menu. Notice: I have not provided for the wide variety of Gamess-US input for the names of atoms. The script and its various file manipulations only understand IUPAC atomic symbols. With cartesian coordinates Gamess allows any name for the atoms because their identity is defined by the nuclear charge. Wingam insists on using IUPAC symbols because these are used later for the .pdb and .xyz files to be produced at the end of a job. Furthermore, with ZMT or ZMTMPC coordinates Gamess allows to number the atoms, e.g. C1, C12 or Cu1, Cu19, etc. Please do not use this numbering. In ZMT(MPC) the numbers of the atoms are defined by the (implicit) number of the row in the matrix of structure data on which they appear. However, atomic symbols may be given as 'Se', 'sE', 'SE' or 'se', and ' C', 'C ', ' c', 'c ', 'C' or 'c'.
|
|
|
EQU
|
'Equilibrium' (Stationary State) Coordinates are to be read from a file named gamess.equ automatically created with the [SAVE] button or by clicking [Properties, Show/Print] 'Coordinates' in the Master Menu. This file, if it exists, is in the gamessdir directory. It is only created after a successful Optimize run and can be used for subsequent Hessian, and Raman runs. It has the full 10-digit precision of Gamess output. It is not overwritten by a subsequent run, unless this is a successful 'Optimize' run.
|
|
|
UNQ
|
Unique 'Equilibrium' (Stationary State) Coordinates are to be read from a file named gamess.unq automatically created with the [SAVE] button or by clicking [Properties, Show/Print] 'Coordinates' in the Master Menu. This file, if it exists, is in the gamessdir directory. It is only created after a successful Optimize run and can be used for subsequent runs of the same molecule with the same pointgroup symmetry. It has the full 10-digit precision of Gamess output. It is not overwritten by a subsequent run, unless this is a successful 'Optimize' run.
|
|
|
COO
|
Coordinates are to be read from a file named gamess.coo automatically created with the [SAVE] button from any run or by clicking [Properties, Show/Print] 'Coordinates' in the Master Menu. This file, if it exists, is in the gamessdir directory. The precision of coordinates is given in the 10-digit format of Gamess, if they are from an 'Optimize' run. In the case of Gamess runs other than 'Optimize' the atomic coordinates are taken from the input (in Bohr, but converted to Angstrom). The precision is given to 8-digits. They may be less precise than what the number of digits indicates, depending on the source of the input.
|
|
|
MOL
|
Coordinates are to be read from a file in “mol” format named ‘jobname’.mol and is usually produced by one of the molecule builders or read from a chemical databank and saved in the gamessdir\data directory.
|
|
|
PDB
|
Coordinates are to be read from a file in “pdb” format named ‘jobname’.pdb. PDB files may be exported from several molecule builder programs but also exist in vast numbers in chemical databanks (26319 structures by July 2004). Before using them they have to be deposited in the gamessdir\data directory. With the [SAVE] button the structural data of a successful Gamess run are saved as ‘jobname’.pdb in the .\output directory. If you request coordinates from a PDB file with the jobname given in the main panel of Wingam and Wingam does not find it in the \data directory, the \output directory is opened for you to choose a jobxxxx.pdb file, 'xxxx' being the time stamp of a previous run. Please note: PDB files are in a certain standard PDB format and contain coordinates only to 3-digit precision. Gamesss output coordinates have to be rounded to three digits while being written into a 'jobXXXX.pdb' file.
|
|
|
ENT
|
These structure files have the identical format of pdb files. They are produced and read by HyperChem named ‘jobname’.ent
|
|
|
XYZ
|
Coordinates are to be read from a file in “xyz” format named ‘jobname’.xyz. This is a format used by Rasmol (RasWin), Chime, and other viewers and is produced with [SAVE] at the end of a Gamess run and saved as ‘jobname’.xyz in the .\output directory. If you request coordinates from an XYZ file to read into an input file and Wingam does not find it as 'jobname.xyz' in the \data directory, the \output directory is opened for you to choose a jobxxxx.xyz file, 'xxxx' being the time stamp of a previous run. The precision of coordinates is given to 8-digits, derived from the 10-digit format of Gamess. In the case of Gamess runs other than 'Optimize' the atomic coordinates are taken from the input (in Bohr, but converted to Angstrom). They may be less precise than what the number of digits indicates.
The format of XYZ files is not standardized (as is the case for MOL and PDB). The script can read XYZ files form ChemCraft and Ghemical builders, too.
|
|
GROUP
|
|
Indicates the point group symbol that is to be used to build the molecule from symmetry unique coordinates. If the coordinates of all atoms are specified as input cart, choose C1 as the point group. This always works, even if the molecule has a higher symmetry. One disadvantage is, that an optimization run in the C1 group will not produce orbitals (SALC's) and (exact) coordinates reflecting the possibly existing higher symmetry. You may also enter the actual group with all cart coordinates if these accurately transform under the group's symmetry operations. However, a quirk of Gamess, the input generated by Gamess from your input coordinates may break the symmetry and e.g. produce IR/Raman lines violating the selection rules under the full symmetry group. If this should happen, use the equilibrium coordinates of the full group, but choose the group C1 and coord=cart when computing Raman transitions.
Do not forget to enter a blank line between the point group symbol and the first atom coordinates in the $DATA group unless the symbol is C1 when you must skip that blank line. Wingam does this correctly for you!
|
|
NAXIS
|
|
Indicates the order of the principal axis when the point group specification includes an “N”. E.g. for C3v NH3 molecule, select 'Cnv' for GROUP and '3' for NAXIS.
|
$CONTRL
|
EXETYP
|
CHECK
|
Indicates that the input file is to be checked for errors.
|
|
|
RUN
|
Indicates that a full GAMESS run is to be done.
|
|
RUNTYP
|
ENERGY
|
A single point calculation is to be done at the geometry specified in the input file.
|
|
|
OPTIMIZE
|
The geometry of the molecule is to be optimized. If you set HSSEND=.true. in the $STATPT group, see below, the Hessian is computed with the converged coordinates and a normal mode analysis performed as in the next case.
|
|
|
HESSIAN
|
The force constants and vibrational frequencies in the harmonic approximation are to be calculated at the equilibrium geometry specified in the input file. The intensities for infrared transitions are determined.
|
|
|
RAMAN
|
The force constants and vibrational frequencies in the harmonic approximation are to be calculated at the equilibrium geometry specified in the input file. IR- and Raman-intensities are computed. The inclusion of a $HESS group from a previous Hessian run into the input file is mandatory. If the 'RAMAN' keyword is set you are prompted to select a previous output .dat file for extraction of a $HESS group. In case there is none you are alerted. If there are several, the last one with (hopefully) converged force constants is chosen.
|
|
|
VSCF
|
The force constants and vibrational frequencies in an anharmonic approximation are to be calculated at the equilibrium geometry specified in the input file. IR intensities are computed and the coupling between oscillators is treated as well. The inclusion of a $HESS group from a previous Hessian run into the input file is mandatory. If the 'VSCF' keyword is set you are prompted to select a previous output .dat file for extraction of a $HESS group. In case there is none you are alerted. If there are several within the same .dat file, the last one with (hopefully) converged force constants is chosen.- There are more input options, see the keyword $VSCF in [Help/Input Description].
|
|
|
G3(MP2)
|
This new runtype (starting with Gamess 11 APR 2008 (R1)) offers a
combination of calculations similar to the same method in Gaussian03
with the aim of obtaining accurate thermochemical quantities as
described in L.A.Curtiss et al. JCP 110,4703-4709(1999). Gamess uses a
step with CCSD(T), as in L.A.Curtiss et al. CPL 314,101-107(1999),
whereas the same step in Gaussian has QCSID(T) model chemistry as in
the first publication. Gamess carries the evaluation one step further
and gives the expected final result as Std. Enthalpy of Formation ΔHf°(0 and 298)
in kcal/mol. Gameix extracts all relevant G3MP2 quantities from the
output file and offers the runtype G3MP2 when composing the job such
that the requirements of Gamess are fulfilled in the input file. Do not
forget to provide adequate memddi by becoming root and typing /sbin/sysctl -w kernel.shmmax=100000000
or more, depending on the size of the job. If you are using WinGamess
instead of Gamess-US do not declare memddi because that is not
supported in WinGamess and you obtain an abort.
|
|
|
SADPOINT
|
A saddle point location calculation is to be done, see $STATPT for options and Master Menu [Help/Input Description] for details.
|
|
|
IRC
|
An Intrinsic Reaction Coordinate calculation is to be run, see $IRC in the 'Advanced Options' for fine tuning this calculation type and Master Menu [Help/Input Description] for details.
|
|
|
NMR
|
NMR shielding tensors (chemical shifts) for closed shell molecules by the GIAO method. The standard defaults are used. These can be finetuned with the help of a $NMR group. See [Help/Input Description] for details.
|
|
|
PROP
|
Certain specified properties of the molecule at the geometry specified in the input file are to be calculated. This requires some manual editing of the file.
|
|
SCFTYP
|
RHF
|
A restricted Hartree Fock calculation is to be carried out. This option is used for closed shell systems.
|
|
|
UHF
|
An unrestricted Hartree Fock calculation is to be carried out. This option is generally used for systems containing unpaired electrons and produces separated alpha- and beta-spin orbitals.
|
|
|
ROHF
|
A restricted open shell Hartree Fock calculation is to be carried out. This option is sometimes employed for systems containing unpaired electrons and produces orbitals with paired spins as much as possible.
|
|
MPLEVL
|
0
|
No Møller-Plesset perturbation calculation is to be carried out.
|
|
|
2 [,3,4]
|
Electron correlation is included through an MP2, [MP3, MP4] perturbation theory calculation following the Hartree Fock calculation. Please note: Gamess-US only offers MP2! If you need MP3, MP4 (up to the full set) you have to change to PC Gamess, V. 6.4 or later, script RUNpcg.
|
|
ISPHER
|
-1
|
Use Cartesian basis functions to construct symmetry-adapted linear combinations (SALC) of basis functions. The SALC space is the linear variation space used. (default)
|
|
|
0
|
Use spherical harmonic functions to create SALC functions, which are then expressed in terms of Cartesian functions. The contaminants are not dropped, hence this option has EXACTLY the same variational space as ISPHER = -1. The only benefit to obtain from this is a population analysis in terms of pure s,p,d,f,g functions which is valuable for teaching purposes.
|
|
|
1
|
Same as ISPHER=0, but the function space is truncated to eliminate all contaminant Cartesian functions [3S(D), 3P(F), 4S(G), and 3D(G)] before constructing the SALC functions. The computation corresponds to the use of a spherical harmonic basis.
|
|
ICHARG
|
Specifies the overall charge on the system.
|
|
MULT
|
Specifies the multiplicity of the system. This equals n+1, where n is the number of unpaired electrons.
|
|
ECP
|
NONE
|
Effective core potentials (pseudopotentials) are not being used.
|
|
|
SBKJC
|
Use the Stevens, Basch, Krauss, Jasien, Cundari ECP’s for the core electrons.
|
|
|
HW
|
Use the Hayes-Wadt ECP’s for the core electrons.
|
|
|
READ
|
The ECP’s are to be specified in the input file. This option requires manual editing of the input file.
|
|
MAXIT
|
The maximum number of iterations that are permitted to achieve SCF convergence, default = 30.
|
|
COORD
|
CART
|
The atom positions are expressed in Cartesian coordinates. This option must be used if the molecule is built with the BUILDERs in Wingam.
|
|
|
UNIQUE
|
Give the coordinates of the symmetry unique atoms only when the point group is specified. The coordinate frame has to be defined in a certain way, see Input Description in the Master Menu [Help]. Cart = unique is Gamess default.
|
|
|
ZMT
|
The coordinates are expressed in the form of a Gaussian™ Z-matrix. The coordinates must be supplied manually if this option is selected. (Note that this option is not correctly implemented in pcgRUN1.0, since 'ZMAT' instead of 'ZMT' is used as keyword. 'ZMAT' is used as $ZMAT group to define internal coordinates if NZVAR > 0, see Input Description)
|
|
|
ZMTMPC
|
The coordinates are expressed in the form of MOPAC type internals. The coordinates must be supplied manually if this option is selected, see Input Description or any MOPAC manual.
|
|
DFTTYP
|
NONE
|
No Density Functional Calculation is done (default)
|
|
|
B3LYP
|
or 30 other DFT functionals may be selected. When right-clicking into the edit box, the list below with a short description of the available functionals pops up. Click on your favorite functional to get it written into the editbox. If you plan to often use DFT in your computations you probably should prefer PC Gamess over Gamess-US. A. A. Granovsky has optimized numerical integration and convergence such that DFT runs about four times faster than with Gamess-US, see the RUNpcg script.
DFT does not work with the semiempirical Hamiltonians MNDO, AM1, and PM3. If you select one of those, DFTtyp is automatically set to "NONE".
This is the list of available functionals for grid integration (default); we have not implemented the gridfree method here, see the Input Manual:
exchange functionals:
SLATER Slater exchange
BECKE Becke 1988 exchange
GILL Gill 1996 exchange
OPTX Handy-Cohen exchange
PW91X Perdew-Wang 1991 exchange
PBEX Perdew-Burke-Ernzerhof exchange
pure correlation functionals:
VWN Vosko-Wilk-Nusair correlation (VWN5)
VWN1 Vosko-Wilke-Nusair correlation, RPA params.
PZ81 Perdew-Zener 1981 correlation
P86 Perdew 1986 correlation
LYP Lee-Yang-Parr correlation
PW91 Perdew-Wang 1991 correlation
PBEC Perdew-Burke-Ernzerhof correlation
OP One-parameter Progressive correlation
exchange/correlation combination functionals:
SVWN SLATER exch. + VWN5 corr.= LDA/LSDA in physics
BLYP BECKE exchange + LYP correlation
BOP BECKE exchange + OP correlation
BP86 BECKE exchange + P86 correlation
GVWN GILL exchange + VWN5 correlation
GPW91 GILL exchange + PW91 correlation
PBEVWN PBE exchange + VWN5 correlation
PBEOP PBE exchange + OP correlation
PW91PBE PW91 exchange + PBE correlation
OLYP OPTX exchange + LYP correlation
PW91 PW91 exchange + PW91 correlation
PBE PBE exchange + PBE correlation
hybrid functionals:
BHHLYP HF and BECKE exchange + LYP correlation
B3LYP Becke + Slater + HF exch. and LYP + VWN5 corr.
B3LYP1 use VWN1 in place of VWN5 (as in Gaussian)
PBE0 a hybrid made from PBE
X3LYP HF+Slater+Becke88+PW91 exch., LYP+VWN1 corr.
|
$SYSTEM
|
TIMLIM
|
This gives the maximum time in minutes allowed for the run.
|
|
MWORDS
|
This selects the maximum number of megawords of memory allowed for the run. One word = 8 bytes. Large jobs will require a larger number than the default. Instead of MWORDS the keyword MEMORY may be given. The unit is word. MEMORY = 1000000 is equivalent to MWORDS = 1.
|
|
[ MEMDDI
|
Not provided on the panel, see e.g. the testinput exam36.inp: Some runtypes, especially MPLEVL = 2, require a larger block of memory for communication between the parallel processes than offered by default. This is handled by adding MEMDDI=1 (megaword) in the $SYSTEM group. If you do not include this when the program needs it, you get an abort. With [View Output File] the error is explained and you are advised of the size for MEMDDI to add. Just [View Input File], make the necessary change to the $SYSTEM group by adding a line 'MEMDDI = 1' (or a larger number) and restart the computation. It may be necessary to click on [SAVE] before, to empty the \scratch and \temp directories from intermediate files of the abortive run. ]
|
$BASIS
|
GBASIS:
NGAUSS
|
These options are used to specify the basis set. To indicate STO-3G, set GBASIS to STO and NGAUSS to 3. For 3-21G, set GBASIS to N21 and NGAUSS to 3. For 6-31G, set GBASIS to N31 and NGAUSS to 6. Several additional basis set options, including those for use with ECP’s, are given as well. See the 'Input Description'.
|
|
NDFUNC
|
Gives the number of sets of d polarization functions to be added to the heavy atoms. For 6-31G(d), which is often designated as 6-31G*, and for 6-31G(d,p), which is often called 6-31G**, NDFUNC=1 (max. 3).
|
|
NFFUNC
|
Gives the number of sets of f polarization functions to be added to the heavy atoms. NFFUNC=0 or 1.
|
|
NPFUNC
|
The number of sets of p polarization functions to be added to hydrogen atoms. For 6-31G(d,p), NDFUNC=1 and NPFUNC=1 (max. 3).
|
|
DIFFSP
|
.TRUE.
.FALSE.
|
A diffuse sp function is included on non-hydrogen atoms in the basis set. This is often used with anions and is designated with a + in the basis set specification. For 6-31+G(d,p), DIFFSP=.TRUE..
|
|
DIFFS
|
checked
|
A diffuse s function is included on hydrogen atoms in the basis set. This is often used with anions and is designated with a + in the basis set specification. For 6-31++G(d,p), DIFFSP=.TRUE. and DIFFS=.TRUE..
|
$SCF
|
DAMP
|
.TRUE.
.FALSE.
|
Check DAMP to damp oscillations in the energy during SCF iterations. This is often necessary with transition metal complexes.
|
|
SHIFT
|
.TRUE.
.FALSE.
|
Check SHIFT to shift the energies of the virtual orbitals to assist convergence during SCF iterations. This is often necessary with transition metal complexes.
|
|
DIRSCF
|
checked
|
Direct SCF calculation in RAM. The default (unchecked) is 'conventional' SCF with integrals stored on disk. DIRSCF uses much less disk space and is faster for large numbers of basis functions. For smaller systems conventional SCF is faster. The 'crossover' point is dependent on the kind of computer system and parallelization, if any.
|
|
CONV
|
Gives the SCF convergence limited as 10**(-conv).
|
$STATPT
|
OPTTOL
|
Gives the gradient convergence tolerance in Hartree/Bohr. If this value is changed, the value of NCONV will also have to be adjusted.
|
|
HESS
|
GUESS
|
Chooses a positive definite diagonal Hessian as an initial guess
|
|
|
READ
|
Reads the Hessian from $HESS. You are prompted to load a $HESS group from a previous Hessian run in a .dat file.
|
|
|
CALC
|
The initial Hessian is computed. See $FORCE. Additional input may be required.
|
|
HSSEND
|
checked
|
If checked, the force constants and vibrational frequencies are calculated at the end of a geometry optimization, if converged. You can then [Show/properties] with added 'Normal Coordinates' and 'Thermochemistry', and generate an [IR Spectrum], similar to a Hessian run.
|
|
NSTEP
|
Indicates the maximum number of cycles allowed in a geometry optimization, default = 20.
|
|
METHOD
|
Selects the optimization algorithm. The default is QA = Quadratic Approximation. You can select NR = straight Newton-Raphson iteration, RFO = Rational Function Optimization, or CONOPT = CONstrained OPTimization. The latter must start from an energy minimum and is used for locating transition states by trying to push the geometry uphill along the mode chosen with IFOLOW, see below. For details see Input Description under 'METHOD'.
|
|
IFOLOW
|
Mode selection switch for RUNTYPE = SADPOINT. The default is 1, meaning that the first, lowest, vibrational 'mode' (rotational and translational degrees removed!) is very likely the reaction coordinate along which the potential energy has a negative curvature. Check the result to be sure that the selection was correct. After a saddle point location run it is recommended to run a Hessian job with the saddle point coordinates. The chosen, and only the chosen mode, usually mode 1, should then possess an imaginary frequency.
|
Advanced Options
Generally these parameters do not have to be changed. Those indicated are set by default without the necessity to click the button for 'Advanced Options'. However, they permit additional control over the calculation, see 'Input Description'. If you set 'Advanced options' different from the default values shown in the boxes, they are made available to the input file after clicking 'Save' on the 'Advanced Options' panel. They do not survive to the input file of your next job, however.
$CONTRL
|
MOLPLT
|
checked
|
Specifies whether the final coordinates are to be saved in Molplot format in ‘jobname’.dat. This can be processed by the routines in the GRAPHICS directory of a Gamess Cygwin- or Linux-Installation.
|
|
PLTORB
|
checked
|
Specifies whether the final coordinates and wavefunction informations are to be saved in PlotOrb format in ‘jobname’.dat. This can be processed by the routines in the GRAPHICS directory of a Gamess Cygwin- or Linux-Installations.
|
|
AIMPAC
|
checked
|
Specifies whether information for a Bader Atoms in Molecules input file is saved.
|
|
NOSYM
|
checked
|
The default is unchecked and means that the symmetry specified in $DATA is to be used as much as possible in integrals, SCF, gradients etc. If checked, the symmetry in $DATA is only used to build the molecule from unique coordinates, then not used anymore.
|
|
INTTYP
|
POPLE
HONDO
|
Indicates whether Pople or Hondo integrals are used. See GAMESS documentation.
|
|
IREST
|
0
|
Restart control options. "0" defaults to no restart planned. At the end of a Gamess run all files are erased except job.out, job.dat (=Gamess PUNCH file), which are saved to \output, and the input file saved as job.inp to \data.
|
|
|
2
|
Setting restart to "2" prevents some data files to be erased. This allows for SCF restart with 1-, 2-e integrals and MO's saved. There are more options to IREST, see Input Description.
|
|
LOCAL
|
NONE
|
Controls orbital localization. The default is 'none', skipping localization. A large number of options for finetuning localization is offered when including a $LOCAL group. Wingam does not write a $LOCAL group. You have to compose it along the details given in Input Description.
|
|
|
BOYS
|
Do Foster-Boys localization.
|
|
|
RUEDNBRG
|
Do Edmiston-Ruedenberg localization.
|
|
|
POP
|
Do Pipek-Mezey population localization.
|
$DFT
|
NRAD
|
96
|
number of radial grid points in Euler-Maclaurin quadrature.
|
|
NTHE
|
12
|
number of angle theta grid points in Gauss-Legendre quadrature.
|
|
[ NPHI
|
24
|
number of angle phi grid points in Gauss-Legendre quadrature. Automatically assigned as twice NTHE.]
|
$GUESS
|
GUESS
|
HUCKEL
|
an (extended) Huckel approximation is to be used to generate the initial MO wavefunctions
|
|
|
MOREAD
|
the MO’s are to be read from the $VEC group of a previous calculation. When you are settting 'GUESS=MOREAD' you are asked for an outputfile name from where to append a $VEC group to the input file. In case there is no $VEC group in the selected file you are alerted. If there are several $VEC groups, as usual for Optmize jobs, the last one with (hopefully) converged orbitals is read in. Generally, using the $VEC group from a previous run to start from, is to be preferred compared to the default option 'GUESS=HUCKEL' since the SCF orbitals are of better quality than those from an extended Huckel computation. However, a real gain in computertime is only observable with large jobs.- It is mandatory to give the number of orbitals in NORB, see next item.
|
|
|
NORB
|
the number of MO’s to be read from a $VEC group when GUESS=MOREAD. You can look at the appended $VEC group when [View(ing) Input File] and write the largest number leftmost of the $VEC table into the NORB variable. There are other choices depending on your job, see the GAMESS Input Description in [Help] in the Master Menu.
|
|
HESS
|
GUESS
|
Chooses a positive definite diagonal Hessian as an initial guess
|
|
|
READ
|
Reads the Hessian from $HESS. Additional data must be supplied manually.
|
|
|
CALC
|
The initial Hessian is computed. See $FORCE. Additional input may be required.
|
$IRC
|
IRC
|
If RUNTYP=IRC this group governs the location of the intrinsic reaction coordinate, a steepest descent path connecting a saddle point to reactants and products. Therefore the prerequisite is a successful saddlepoint location run with RUNTYP=SADPOINT.
|
|
|
PACE
|
There are five integration methods: The default is GS2, the Gonzalez-Schlegel second order method using BFGS for updating the Hessian. There are more keywords for finetuning GS2, see Input Description. The other four choices for PACE are 'LINEAR', 'QUAD', 'AMPC4', and 'RK4', see Input Description, again.
|
|
|
SADDLE
|
If checked the IRC assumes starting from a precise saddle point. In this case the $HESS group of a SADPOINT run has to be attached to the input file. If unchecked, IRC starts from some other point _on_ the IRC path. The safest way is to start IRC from a converged SADPOINT run, check SADDLE, and read the $HESS group by setting HESS='READ' in the $STATPT group.
|
|
|
FORWRD
|
This defines in which direction the IRC starts from a saddle point. Default is FORWRD=checked, meaning that the IRC starts in the direction where the largest magnitude component of the imaginary normal mode is positive. You can identify this, if you look up the vibrational amplitudes of the imaginary frequency (normal mode table of the preceding SADPOINT run). If you pick the wrong direction you can always correct this in a second run with the advantage of thus getting an overview of both reaction directions, back to the reactants and forward to the products!
|
|
|
NPOINT
|
The number of IRC points to be located in this run, separated by STRIDE.
|
|
|
STRIDE
|
Determines how far apart points on the reaction path will be. STRIDE is used to calculate the step taken, according to the PACE method you selected. If you choose the robust method GS2 it can be 0.30 sqrt(amu)-Bohr, for the other methods it should be smaller, 0.1 or even 0.05.
|
|
|
MXOPT
|
Maximum number of constrained optimization steps for each IRC point. The default=20 is similar to NSTEP (in $STATPT) pertaining to optimization runs. If an IRC point does not converge, select a larger MXOPT and repeat the run.
|
$CIS/TDDFT
|
CIS/TDDFT
|
If CITYP=CIS or CITYP=TDDFT this group defines
the details. Please consult the instructions, CIS and
TDDFT computations are not entirely black-box for
significant results and there are more parameters to select
differently from the defaults than those offered on the panel.
Note that TDDFT (time dependent DFT) often
gives better results than CIS with the same basis.
|
|
|
NSTATE
|
Number of states to be found (excluding the ground
state, so '1' means one excited state). For this number of singly excited states excitation energy
and oscillator strength is computed.
|
|
|
IROOT
|
State for which properties and/or gradient (only for CIS) will
be calculated. Only one state can be chosen.
|
|
|
MULT
|
Multiplicity (1 or 3) of the singly excited SAPS
(the reference is necessarily single RHF). Only relevant for
SAPS based run. SAPS are spin-adapted antisymmetrized products of
the desired MULT.
|
|
|
HAMTYP
|
For CIS only: Type of CI Hamiltonian to use:
SAPS: spin-adapted antisymmetrized product of the desired MULT will be used (default)
DETS: determinant based, so both singlets and triplets will be obtained
|
|
|
RDVEC
|
For CIS only: Read CIS vectors from a previous computation
of the same system, if you want to get other states
(default = .false.)
|
|
|
NACORE
|
For CIS only: Omits the first n occupied alpha and beta orbitals from
the calculation. The default for n is the number of
chemical core orbitals.
|
APPLY
The input file is written with the specified parameters. Until this button is pushed, nothing is written to your hard disk. The file “input” of a previous run is overwritten, but that file has already been saved by 'SAVE' as jobname.inp (including a time stamp) into the \data directory, so nothing is lost (unless you forgot to click 'SAVE', see below)!
VIEW INPUT FILE
The VIEW INPUT FILE button calls up the required Gamess input file “input” in the editor for checking and, if necessary, editing. The “title” of the run and any special parameters can then be set manually and the file saved before starting a computation.
RUN
The RUN button calls rungms.sh which has all required parameters
to call ddikick.x and gamess.XX.x. The 'input' file is presented
to gamess for computation. A shell window shows the start and
calling string of the run, and announces the termination of the
program. If there is an error condition announced by ddIkick there
is, usually a comment to guide you to the next steps. If all's
well – normal termination – you are prompted to look into the
summary.
Summary
The SUMMARY button shows a summary of the run in the Dialect worksheet. It announces whether the run was successful or unsuccessful. If a structure optimization has been run, convergence to a stationary state or failure to do so is announced as well as the final total energy. The output is written to the file .\output\‘jobname’.out.
SAVE
The SAVE button generates ‘gamess.coo’ and the two files ‘jobname’.pdb, and ‘jobname’.xyz from either the input coordinates (runtype 'Energy', 'Gradient', Hessian', 'Raman', or an unsuccessful 'Optimize') or the last set of coordinates (successful 'Optimize') of a gamess run. The last two files are saved in the output directory. The equilibrium coordinates of a successful 'Optimize' run are additionally written into a file 'gamess.equ'. If you want to reuse coordinates in a new job with the same name, you can get them from three locations: 'COO', 'PDB', and 'XYZ' see Input Coordinate Type or from 'EQU' and 'UNQ', if you have deposited them from a converged 'Optimize' run. Note that ‘jobname’ contains a four digit time stamp to prevent overwriting files when you use the same job name in a new run. In addition the GAMESS 'PUNCH' file is moved to the output directory as ‘jobname’.dat. In \data a copy of the inputfile is saved as ‘jobname’.inp.
Essential: Click on [Summary] and [Save] to save the output and clean the system for a new run.
VIEW
The VIEW button opens the output file in the editor to study the detailed results. If the run was unsuccessful you find hints on what went wrong. Correct your input accordingly for a new try.
VIEW STRUCTURE
The VIEW STRUCTURE button calls up the specified viewer. If vfile is set to “pdb” for that viewer, the output structure ‘jobname’.pdb is read into the viewer immediately. If vfile is set to "" in Wingam.ini, the viewer is called up, but the user must select the file to be rendered manually. Molekel and Molden are different: They read all the pertinent parts of the output file, render the structure in the opening window and then let you choose any of their features in control menus.
IR/RAMAN SPECTRUM
In order to simulate a spectrum of the fundamental vibrations, the vibrational frequencies and their intensities have to be extracted from the output: From a 'Hessian' run or an 'Optimize' run with HSSEND = true (both are called 'Hessian', here) the Infrared intensities, from a 'Raman' run the IR and Raman intensities (and the depolarization ratios for the latter) are exported. This is done with the help of the Master Menu. There are three cases:
- If you have a Hessian or Raman run 'on-line', i.e. the job has just been terminated, and Wingam not yet exited, click on [Properties, Show/Print]: You see the drop-down menu of selectable properties. Choose 'Normal Coordinate Analysis'.
- If you want to look at the vibrations of a previous run [Outputdata/Load *.out] opens the \output directory to select a Hessian or Raman output file. This done click on [Properties, Show/Print]. The drop-down menu of properties opens. Choose 'Normal Coordinate Analysis'.
- You can click on 'Vibration Spectrum' whereby a drop-down
list of all jobname.res files in the \output directory is shown for
choosing the Normal Coordinates of a previously saved Hessian or Raman run.
The normal coordinates are not shown in this case but the spectrum panel opens and
lets you proceed as follows.
In the first two cases the frequencies, IR-, and, possibly,
Raman-intensities are shown in an editor window. This is automatically
saved as output\jobname.res for later reuse.
Click [Vibrational Spectrum] on the Wingam panel. A list of saved
IR/Raman spectra - including the most recent - is opened for you to select one.
The spectrum panel
opens for you to click on either one of the next two
buttons:
|
Lin (1/cm)
|
this produces a lineplot with an overlay of Lorentzian line shapes on a linear wavenumber scale. It resembles a measured IR spectrum, probably from an FTIR machine, as change in 'transmittance' (blue trace). The Raman bands, if they have been computed, are shown in 'emission' (red trace). The spectral range goes linearly from about 20 to 4200 cm-1.
|
|
Lin (µ)
|
this produces a lineplot with an overlay of Lorentzian line shapes in a linear wavelength scale. It resembles a measured IR spectrum from a Rock Salt prism spectrometer as change in 'transmittance' (blue trace). The 'fingerprint' region is better visible than in the first plot but has less detail in the C-H stretch region and the 'skeletal' motions below 666/cm are missing. The Raman bands, from a Raman run, are shown in 'emission' (red trace). The spectral window shown goes from 2.4 - 15 µ (with constant transmission!). Vibrations below about 650 cm-1 are not shown!
|
|
Half Intensity Width (1/cm)
|
You may adjust the 'Linewidth' (width of the spectral 'line' at half intensity) to approximate an experiment with varying resolution. Furthermore, the rotational part of a vibration-rotation band is not explicitly simulated. This can be approximately taken care of by adjusting the linewidth.
|
|
Intensity Scale
|
This button allows to make weaker IR transitions or Raman emissions visible, or to scale overshooting transitions down. There is no simulation of the transmission behavior of your spectrometer. Assume, that the simulated spectrum has been corrected for experimental shortcomings!
|
|
Scale Frequencies
|
It is a sad fact that even the best 'ab initio' computations have problems with the vibrational frequencies in the harmonic approximation. Most calculated frequencies are up to 10% too large, depending on the model chemistry used. This error has been determined over a large sample of calculated versus observed frequencies. It is fairly constant, hence it has become customary to correct calculated spectra by this 'fudge factor'. You can check 'Scale Frequencies' and then select a factor corresponding to your model chemistry. The factors used are published in many locations, e.g. on page 64 of the book 'Exploring Chemistry with Electronic Structure Methods', 2nd ed., by James B. Forseman & Æleen Frisch, ISBN 0-9636768-3-8.
You can get a better prediction of vibrational frequencies, if you run an anharmonic vibrational analysis using RUNTYP = VSCF and the keyword $VSCF (search for 'VSCF' in Master Menu [Help/Input Description]). However, this takes much more computer time than the harmonic approximation (determining many structure points along each normal coordinate to SCF precision, computing the vibrational eigenfunctions, and taking account of coupling between them). You are prompted to add a $HESS group from a previous Hessian run for starting VSCF. It is recommended to run VSCF only with a good basis set, about 6-31g(d) or better, since minimal basis sets, like STO-3G, often do not produce results that make sense.
Note: Molekel (version 4.3), Molden and ChemCraft can animate vibrational modes from the "View Structure" button and their own menus. However, Molekel cannot use output from Raman runs whereas Molden and ChemCraft can use IR- and Raman data and also draw a spectrum of both.
|
References
1. Dialect®
http://www.downlinx.com/proghtml/104/10441.htm
2. Gamess
| Citations: GAMESS-US: Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S. J.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993, 14, 13471363.
MacGAMESS: Bode, B. M., Ames Laboratory, Iowa State University, Ames, IA 50011.
PC GAMESS: Granovsky, A. A., Moscow State University, Moscow, Russia. |
WinGamess has been compiled with gcc-gfortran from the current version of GAMESS-US ( 11 APR 2008 (R1)), an ab initio quantum chemical package, in a Cygwin environment on Windows.- If you prefer PC GAMESS, another Gamess version for running in the Windows platform, an equivalent script RUNpcg is available. PC-Gamess is several times faster than WinGamess and runs on multiple CPU's/Nodes. However, it lacks several newer features of current (Win-)Gamess but adds some unique new ones as well.
http://www.msg.ameslab.gov/GAMESS/gamess.html
http://www.chemsoft.ch/qc/binaries.htm
3. NoteTab Lite®
http://www.notetab.com
4. ArgusLab 4.01© Freeware
http://www.arguslab.com/
5. ACD/ChemSketch© Freeware
http://www.acdlabs.com/download/chemsk.html
6. ISIS/Draw™
http://www.mdli.com
7. ViewerLite™
Discovery Studio Visualizer/
8. HyperChem®
http://www.hyper.com/
9. PCModel®
http://serenasoft.com/index.html
10. gOpenMol
gOpenMol is maintained by Leif Laaksonen, Center for Scientific Computing, Espoo, Finland.
http://www.csc.fi/gopenmol/
11. VMD©
VMD was developed by the Theoretical Biophysics Group in the Beckman Institute for Advanced Science and Technology at the University of Illinois at Urbana-Champaign.
See Humphrey, W., Dalke, A. and Schulten, K., VMD - Visual Molecular Dynamics, J. Molec. Graphics 1996, 14.1, 33-38.
http://www.ks.uiuc.edu
12. RasWin
http://www.umass.edu/microbio/rasmol/getras.htm
13. Molekel, version 4.3
http://www.cscs.ch/molekel
14. Pov-Ray™
http://www.povray.org/
15. ChemCraft™ version 1.4 beta, build 237
Download a 90 day trial version/
16. (g)Molden, version 4.6
download 4.6 for Windows; you need an Xwindows driver
17. Ghemical, version 1.02(stable) or 1.90 (dev)
http://www.uku.fi/~thassine/ghemical/
18. VB2000, Version 1.8 Rev. 3, now version 2.0 for WinGamess
VB2000 is developed by Jiabo Li, Brian Duke, and Roy McWeeny, Scinet Technologies, San Diego, CA, 92127, USA: Jiabo Li, and Roy McWeeny, "VB2000: Pushing Valence Bond Theory to New Limits", Int. J. Quantum Chem., 89(2002)208-216
http://www.vb2000.net
|
|
1997-2008 Ernst Schumacher, last update 18 July 2008 |
|
| |