[CPMD-list] background charge

[Axel Kohlmeyer]

On Sun, 15 Apr 2007, Weijie HUA wrote:

WH> Dear Dr. Kohlmeyer,

dear weijie,

WH> Thanks for your kind reply. We would like to study such a system. In
WH> a 10*10*10 box, there are 4 water molecules which are represented by
WH> real atoms, and another 4 water molecules which are represented by
WH> their natural charge(They are treated as background point charge for
WH> the first 4 waters). I may write a pseudo input file describing the

ok. so now you have explained _what_ you are planning to do,
but i am still puzzled as to _why_ you want to do this.

WH> system as this:

[...]

WH> The first 12 lines are the coordinates of the four real atoms, the
WH> following 3 lines are the translational vector. The last 12 lines
WH> are the distribution of the point charges in the format of [ x y z
WH> charge].

WH> I really tried to use some softwares using gaussian basis sets, they
WH> can treat background charges for isolated systems, but not the
WH> periodic systems yet. So I turned to try this in CPMD.

i hope you are aware that due to the way CPMD works, you are _always_
doing periodic calculations. so what about doing an all-QM calculation 
right away? that would save you a _lot_ of trouble.
the next best option would be to look into how the CPMD/ego/gromacs 
QM/MM interface works (see egointer.F and related files).

WH> I would like to ask how to write a CPMD input file (e.g. the format
WH> of the potential file) for such a purpose .

as with all package programs, if you want to do something non-standard,
you have to read and understand the relevant parts of the sourcecode,
so you can implement what you need. the important information about
pseudopotentials is in the files ratom.F and recpnew.F. most likely
you'll have to make minor modifications to the sources.

WH> And is there any criterion to judge whether the result I would
WH> obtain is right or wrong due to the 'spillout' of the electrons?

see above, i'd just do a regular calculation with all atoms 
made QM and compare forces etc. also you can write out the density
with the RHOOUT command, convert the resulting DENSITY file with
cpmd2cube.x and then visualize it with your favorate visualization
program for cube files.

cheers,
   axel.


WH> Thanks for your time and effort.
WH> 
WH> Best regards,
WH> Weijie	
WH> 
WH> ----
WH> Weijie HUA, Ph. D. candidate
WH> Institute of Theoretical and Computational Chemistry,
WH> Key Laboratory of Mesoscopic Chemistry,
WH> School of Chemistry and Chemical Engineering,
WH> Nanjing University,Nanjing 210093,
WH> P.R.China
WH> Email:huawj at itcc.nju.edu.cn
WH> 2007-04-15
WH> 
WH> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.