| Three issues. Two answered as an AMSOL
developer, one as a sometime solvation meddler... 1) AMSOL's Z-matrix format is
that used by AMPAC and MOPAC. It's described in lots of places, e.g., http://www.chem.cmu.edu/courses/09-560/docs/msi/modenv/D_Files.html#944883
(under subheading MOPAC). 2) One does not need to specify
atomic partial charges in AMSOL unless one is using the EXTCM (external charge
model) keyword to calculate generalized Born solvation free energies with user
supplied charges. It is quite possible, however, that example input decks
contain such charges. That is because a convenient way to generate a new input
geometry is to cut and paste from the .log file of a previously completed
calculation. In that case, the partial atomic charges from that prior
calculation are indeed found at the end of each atom's line, but that
information is not read in the absence of EXTCM. 3) Jim Kress suggested that
Mark Gordon's effective fragment potential (EFP) as an option for a
supermolecular solute, where the supermolecule is a cluster of solvent molecules
about a central non-solvent-molecule solute. EFP is certainly designed
specifically for the purpose of modeling such clusters, but it does NOT include
charge transfer to the solute molecules when they are represented by the EFP.
Only fully quantum solvent molecules can transfer charge to/from the solute. The
EFP may be thought of as a tremendously sophisticated molecular mechanics
description (including exchange repulsion, polarizability, etc.) of the explicit
solvent molecules, but they do not carry partial charges (unless Mark and his
co-workers have an extension that I haven't yet seen). It could certainly be
added, in principle, but it introduces some trickiness associated with partial
occupation numbers in the quantum MOs. Chris Cramer On Jan 28, 2006, at
6:59 PM, Jim Kress ccl_nospam[A]kressworks.com wrote:
|