[CPMD-list] General questions about using CPMD

[Axel Kohlmeyer]

On 6/15/06, Donghee Lim <limkr at umich.edu> wrote:
> Dear all
>
> I would appreciate if you give me any advice on my research topic using
> CPMD program. I am a beginner in using CPMD and do not know what I
> should do first.

first learn to use CPMD on 'simple' compounds (e.g. isolated molecules,
simple bulk materials like Si, Water, ...). there is some material on the
web available (see my previous mail to this list in the archives and others
before that) and also you should familiarize yourself with the underlying
theory (for solid state electron structure calculations _and_ MD).

it is probably useful to participate at a tutorial in those kind of
calculations,
e.g. the upcoming one at UIUC: http://www.mcc.uiuc.edu/summerschool/2006/
this will introduce the quantum-espresso package (including the bundled
car-parrinello MD code), but everything you would learn about plane-wave
pseudopotential calculations there, applies to CPMD just as well.

> I am planning to calculate the interactions between small molecules
> containing C,H,Cl and Fe surface using a periodic DFT with a plane wave
> basis set and pseudopotential.
> My questions are as follows;
>
> 1) For the use of the plane wave basis set, do I have to specify any
> commands in the CPMD input files? How can I use the plane wave basis
> set in CPMD?

with plane waves you have no specially 'massaged' and balanced basis
sets as with local basis set based programs, you just say what is the highest
frequency of a (commensurate) plane wave that you consider by specifying
a maximum energy cutoff. you can enlarge the basis set by increasing this
cutoff. due to the nature of plane waves, the basis set is 'covering' the whole
supercell with the same accuracy, so there is no need for diffuse functions
and alike and not BSSE.

> 2) Where can I obtain ultrasoft pseudopotential files?
> They say I should search the files in the vanderbilt library. What is
> the vanderbilt library?
see:
http://www.cpmd.org/manual/node3.html
http://www.physics.rutgers.edu/~dhv/uspp/
apart from the cpmd.org server, there is additional material
on http://www.pci.unizh.ch/gruppe.hutter/e/information.html

> I found some pseudopotential files in the CPMD homepages, but I can't
> figure out what the names stand for: - Fe_VDB_BLYP(or
> BP,GGA,LDA,PBE)_SEMI.psp   - Fe_GO_BP(or PBE).psp
> - Fe_MT_BLYP(or BP)_NLCC.psp
> - Fe_MT_LDA.psp

the first part is (obviously) the element, the second the type:
VDB stands for vanderbilt ultrasoft. MT for Martins-Troullier norm-conserving,
GO for Goedecker (or Goedecker/Hutter/Hartwigsen to be more precise),
the third is the functional and a fourth segment of the name (NLCC or SEMI)
indicates a pseudopotential with non-linear core-correction or a semi-core
pseudopotential, respectively.

> 3) In order to perform my research, I am curious that if I can use the
> different ultrasoft psedupotentials of C,H,Cl and Fe, for example like
> this;
> - For carbon: C_VDB_BP
> - For Fe    : Fe_MT_BP_NLCC

yes, you can mix-and-match, as long as you use the same functional.
but in this case, it would make very little sense, as the 'hardest'
pseudopotential
(here the iron) determines the required plane wave cutoff for reasonably
converged forces.
>
> 4) I guess I should fix the Fe surface and relax the small molecules to
> figure out the most stable configuration of the small molecule on the
> Fe surface. Can I fix the Fe surface in CPMD code?

yes. actually you need to do a slab calculation, where you fix the bottom
layer(s) of atoms at the value of the bulk lattice constant but let the upper
layer(s) move freely to reconstruct. best you have a look at some of the
multitude of papers using slab calculations for molecules on surfaces
that are listed on the cpmd.org homepage.

> 5) I would like to use a periodic boundary condition for my research.
> Could you give me any example files of the PBC?

there are tons of examples around already.

regards,
     axel.

>
> Thank you in advance,
> Donghee Lim
>
>
> _______________________________________________
> CPMD-list mailing list
> CPMD-list at cpmd.org
> http://cpmd.org/mailman/listinfo/cpmd-list
>
>


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