[CPMD-list] a small biplatinum coordination compound

[Axel Kohlmeyer]

On Mon, 18 Feb 2008, Mustafa Hussein wrote:

MH> Dear All,

dear mustafa,

MH>   I am doing a CPMD study on a biplatinum complex. I am now testing 
MH> my parameters and pseudopotentials. I am looking for a small 
MH> biplatinum molecule in which the platinum atoms are close in space, 
MH> either actually bonded or nonbonded, to test for my PW cutoff value 
MH> and for other to-be-set parameters. My molecule is large and 
MH> computationally demanding for such kind of testing. I have already 

why don't you try what every quantum chemist does (when molecules
get too large): cut of pieces of the molecule and saturate the
dangling bonds with hydrogen. that may not be perfect, but it will
get you into the ballpark and of course you can check for "convergence"
regardless, since convergence is reached when the result does not
change, whether you know the result a priori or not. whether that
converged result is _accurate_ depends on a lot of other parameters, 
too. BTW: if you are feeling adventurous, you can also try the
density optimized capping pseudopotentials, that have been pioneered
in your (current) group. ;-)

MH> some suggested values but I need to choose the best. (My suggestion 
MH> is 90 Ry for the PW cutoff value, for example). You can think of 
MH> what I want as finding the geometry convergence (in terms of bond 
MH> length change) as a function of the applied plane wave cutoff. I am 

the simplest suitable platinum compound that i can think off,
would be cis-platinum, for which a plethora of literature exists
(including, again, papers from former group members of the group
that you are visiting).
http://en.wikipedia.org/wiki/Cisplatin

MH> working with BLYP or PBE XC functionals. Basis set for core 
MH> electrons is minimal STO basis set (the default in the CPMD 
MH> package).

please note, that CPMD does _not_ use an STO basis set for core
electrons, but does not treat them at all. you use pseudopotentials.
the minimal slater basis is only used (by default, you can use other 
atomic wavefunctions as well) to construct an initial guess, and
of course only valence states are considered.

MH>   Any way, the test molecule should be previously reported in the 
MH> literature, just to enable me to try to reproduce the measured bond 
MH> lenghts by plane wave DFT calculations.

please be careful here. there are two issues to consider:
correctness within the method and absolute correctness.
to verify correctness within the method (plain DFT will give you
different results than, say, hartree-fock, MP2, coupled cluster,
hybrid DFT) you can run the same fragment with another
(local basis set) quantum chemistry program and see how
well you can reproduce the results using the exact same 
functional and an adequate basis set. 

how good DFT will be able to reproduce experimental results
is a completely different issue.

cheers,
   axel.

MH>   You can refer me to a paper or some article



MH>    
MH>   Sincerely,
MH>   Mustafa
MH> 
MH> 
MH> Mustafa Hussein Muhammad 
MH> Independent Research Student
MH>   Quantum Chemistry
MH>    
MH>   Senior Undergraduate
MH> Faculty of Pharmacy 
MH> Minia University 
MH> Minia 61511, Egypt 
MH> Cell: + (20) 10 187 6193
MH> 
MH>        
MH> ---------------------------------
MH> Be a better friend, newshound, and know-it-all with Yahoo! Mobile.  Try it now.

-- 
=======================================================================
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.