[molpro-user] non-integer atomic charges

[Benjamin Levine]

Dear Molpro-users,
I'm working on a problem that would greatly benefit from the assignment 
of non-integer charges to specific atoms.  In a previous Molpro-user 
discussion 
(http://www.molpro.net/pipermail/molpro-user/2011-May/004406.html) 
another user raised a similar question, and it was suggested to set the 
charges of the atoms to zero and use lattice charges right on top of the 
dummy centers to represent the non-integer charge.  This would allow 
energy calculations, but if I am correct, this solution wouldn't allow 
us to compute energy gradients with respect to the motion of the 
non-integer-charged atoms, which is important in our application.  Might 
there be a way to truly create custom atom types?  Thanks for your time 
and help!

Ben

-- 
Benjamin G. Levine
Assistant Professor
Department of Chemistry
Michigan State University
East Lansing, MI

517-355-9715 x169
http://levinegroup.chemistry.msu.edu/