[molpro-user] CCSD Electron Density Calculations?

Gerald Knizia knizia at theochem.uni-stuttgart.de
Tue Mar 3 11:41:11 GMT 2009


On Monday 02 March 2009 22:23, Mark Hutchison wrote:
> I am an undergraduate physics student doing some research on the variable
> decay constant of singly and doubly ionized beryllium 7 atoms.  I need to
> be able to get an accurate calculation (I'm hoping at least on the CCSD
> level) for the charge density near the nucleus of a single atom.  I have
> been using NWChem, but when I asked if they had the capability of
> calculating the charge density at the CCSD level, they said it was still
> under development. My advisor is a plasma physicist and hasn't been able to
> help me very much on this project because this is unfamiliar territory to
> him.  I was wondering if Molpro has the capability of doing this kind of
> calculation or if there are any users that have done similar calculations
> and would be able to steer me in the right direction on how to go about
> doing this kind of calculation I would appreciate any kind of feedback that
> you could give me. Thanks.

Molpro can calculate relaxed density Matrices for closed shell CCSD, but not 
for open-shell systems if I'm not mistaken. However, If you are only 
interested in the electron density at the nucleus, you can get the results 
from finite field calculations using any method Molpro supports (see 
http://www.molpro.net/info/current/doc/manual/node438.html ).

Nevertheless, you might also want to look into the ACES II program package 
(Mains-Austin-Budapest version, www.aces2.de). This package has extensive 
support for all kinds of property calculations.

Note that with singly- and doubly ionized beryllium you also need to be 
careful with the basis sets, as the species changes substantially during the 
process. For data at the nucleus, you must avoid the frozen core 
approximation (use "core,0;" in molpro) and use basis sets which are flexible 
near the nucleus (i have no experience with this, but cc-pwCVnZ (if existing) 
or cc-pCVnZ might work. cc-pVnZ would most likely not, however.)
-- 
Gerald Knizia



More information about the Molpro-user mailing list