Hi,

I have used GAMESS program package and used Gaussian 09 program package as well. I can write the program for MP2/6-311++G(d,p) for monomer thermochemistry calculations. The program as follows for GAMESS:

$CONTRL SCFTYP=RHF RUNTYP=OPTIMIZE MPLEVL=2 MAXIT=200

MULT=1 NOSYM=1 COORD=UNIQUE $END

$SYSTEM TIMLIM=525600 MWORDS=20 MEMDDI=960 $END

$BASIS GBASIS=N311 NGAUSS=6 NDFUNC=1 NPFUNC=1 DIFFSP=.T.

DIFFS=.T. $END

$SCF DIRSCF=.T. $END

$STATPT NSTEP=1000000 HSSEND=.T. $END

$DATA

You can use Gen keyword for Gaussian...

you can use the MP2/gen keyword and then add the description of the basis set at the end of the input in gaussian format. As far as I know the 6-31G*(0.25) basis is constructed from split valence 6-31G adding a set of diffuse functions having an exponent of 0.25 for second row elements. So you can change the basis set input manualy. most basis sets (e.g. 6-31G) you can download from emsl basis set exchange web.

Many thanks for the suggestion Jakub. I have managed to understand it and implement it.

Is thera any particular reason for using this combination of method and rather restricted basis? System size? No hydrogens? I would guess, that additional polarization functions will improve the results sizably.

Dear Jan, I suppose it is system size. I have a system with 20 non-hydrogen atoms and I would like to estimate the pi-pi interactions binding the monomer units. I tried with MP2/631+G(d,p) and with my computational facility, I am not able to get results in reasonable time. However, I am informed that the addition of diffuse functions (+) has minimal effects on the estimated binding energy if the 6-31G*(0.25) basis set utilized. Since I can afford the calculation with the 6-31G*(0.25), I wanted to try it out and see its performance on estimating binding energy with respect to my system.

If the system that you are using is present without hydrogen atoms hence use the gen keyword at the MP2 level and manually input the description for 6-31G*(0.25). Please use this as citation http://144.206.159.178/ft/162/73204/1254325.pdf. I guess that this is the starting work for the correlated interactions including both H-bonded and stacked structures.

Thanks, Beltran. Actually it is 20 non-hydrogen atoms plus a number of hydrogen atoms. I guess the correct basis set is 6-31G**(0.25, 0.15). Thanks for the reference you have direct me to, I am sure to reference it and other works related to Hobza's group.

what about using local-MP2 instead of cannonical?

Dear Kaminský, thank you for the suggestion. Unfortunately I am still using Gaussian 03, version D1. I am not sure if local-MP2 is implemented in this version. I have local -MP2 in Spartan 10, but in that software I am not sure of how I can change the basis set. May you suggest what kind of instruction I may use to initiate an LMP2 method in Gaussian. Many thanks.

Sorry, I forgot you need to use Gaussian. For local MP2 I would recommend Molpro (proper domain description).