This is just a follow-up on this issue for the sake of other users:
William did have G09 Rev B.01. The weird charges he got when he
used the fixreadinesp.sh script are the result of the automatic
reorientation that Gaussian performs. The Gaussian customer
support gave me a nice description of the problem and an easy
solution. See below ....
--Niel
(From Gaussian support)
This issue likely has to do with a reorientation of the molecule. By
default, Gaussian would rotate and translate the input geometry to
the so called "Standard orientation". The only way to prevent this
translation and rotation is by means of the "NoSymm" keyword, which
keeps the molecule in the orientation given as input ("Input orientation").
Note that the script to reformat the output ("fixreadinesp.sh") takes
the list of grid points from the normal run, generates a new Gaussian
input file where the list of grid points is given explicitly and takes the
rest of the information from the checkpoint file. One then needs to
rerun this second quick Gaussian job to get an output with the proper
format. One needs to make sure that the orientation of the molecule
read from the checkpoint file matches the orientation for which the
grid was created. If the molecule is given in a different orientation
than that for which the grid was created then the values of the
electrostatic potential would not be those at the expected surface
around the molecule.
The Gaussian input file generated by the "fixreadinesp.sh" script will
make use of the "NoSymm" keyword to keep the molecule in the
orientation saved in the checkpoint file. This assumes that one used the
"NoSymm" keyword in the initial "pop=mk" that one run with G09 rev.
B.01, so the molecule was never reoriented and all the orientations match.
My guess is that in the problematic case you mentioned, the "NoSymm"
keyword was not used in the initial G09 rev. B.01 job. This means that the
molecule was rotated to the "Standard orientation" and that the grid points
match such orientation. Since the "fixreadinesp.sh" script generates a new
Gaussian input file with the "NoSymm" keyword, the orientation of the
molecules in the two G09 runs would not match and the problem will appear.
In short, there are two ways to proceed to ensure the correct orientation:
1) Use the "NoSymm" keyword in the original "pop=mk"; or
2) Remove the "NoSymm" keyword from the Gaussian input file that is
generated by the "fixreadinesp.sh" script, before running the second quick
Gaussian job.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Aug 25 2011 - 15:00:03 PDT
