|
Many
thanks. My
original question was I am calculating the heats of reactions containing
organo-mercury compounds. I am using G03 and DFT methods. The program, when
mercury is involved, defaults to an unpruned integration grid for this atom.
The manual states that one should use the same grid when comparing energies.
Does this mean that I should be defining an unpruned grid for all of my
calculations i.e. the compounds without mercury in the isodesmic reaction? If
so how do I define an unpruned grid the same as the one being used by gaussian
when mercury is involved. The manual does not talk about unpruned grids. I have
enclosed the relevant output section below. And
the
replies:- Only the mercury atom will use
the unpruned grid. All the
other atoms will use
the default, pruned grid if it is defined. So you need not do anything
special in generating the pieces of the
reaction. Each element will be
done the same way if you use the defaults in
Gaussian. The point of printing the
informational message is in the event that you compare with
another program or a future release of Gaussian were a pruned grid
for mercury might be used. Douglas J.
Fox Technical
Support Gaussian,
Inc. fox|at|gaussian.com I think the
manual is referring to using different grid densities, that is,
INT=ULTRAFINE versus INT=FINE. A pruned grid has some of the
points removed (hence
the word 'pruned') so it is faster, but it should give the same result
as an unpruned grid, for the same integration grid density. To summarise, I
think you don't need to worry about the pruned grid, so long as you use
the same integration grid density (for example, the default
integration grid density) for all calculations. Regards, Noel Ian Hovell -
Ph.D. NUCLEO DE
MODELAGEM MOLECULAR-NMM Centro de Tecnologia Mineral - CETEM Ministerio da Ciência e da Tecnologia- MCT Avenida Ipê, No 900 - Cidade Universitaria Ilha do Fundão Rio de Janeiro RJ Brasil CEP 21941-590 tel 00 55 (xx) 3865 7344 ou 3865 - 7216 Fax 00 55 (xx) 22602837 ou 2290-4286 e-mail hovell|at|cetem.gov.br |