[DFTB-Plus-User] DFTB (Hbonding damping function + 3rd Order energy term)

[Bálint Aradi]

Dear  Mark,

  Sorry for the long response delay, but I have been practically
continously out of office for the last 5 weeks (and will still be for
this week) due to different cooperation visits at various place.

>   I have successfully simulated water using the DFTB+ code with and
> without the hydrogen bonding damping term (DampXHExponent = 4.88) and
> can closely reproduce the oxygen-oxygen RDF found in J Phys Chem A.,
> 2007, 111(26): 5685–5691. When I attempt to also use the 3rd order
> energy term (using DampXHExponent = 4.88, HubbardDerivs O=-0.14 H=-0.05,
> these values were taken from J. Phys. Chem. A., 2007, 111(42):
> 10861-10873) the simulation box forms voids.  The configuration that I
> start with is the final configuration from the previously described
> simulations (128 water for ~100 ps with a 1fs time step at ~300K).  It
> appears that the water density is too high (the waters are overly
> coordinated) and that causes the formation of voids or pocket in the
> simulation cell.  I tested the code on clusters and was able to
> reproduce the energies reported by J. Phys. Chem. A., 2007, 111(42):
> 10861-10873. 

  Which energies did you manage to reproduce? Those with standard DFTB
or those with 3rd order and damping? I was only able to reproduce the
2H2O case, for all other cases I got deviances (also for the standard
DFTB results!). The more than sluggish discussion with the authors at
the moment concluded, that the differences for the standard DFTB results
come from differences in the geometry used. Maybe also the differences
in the 3rd order results are due to this, but I was not able to clarify
that yet.

  But as I get at least for the H2O similar results to those of the
authors, I'm more or less confident, that the implementation is OK (at
least as long as you're calculating water clusters only).

> I am unsure if these results are real (i.e. no bug in code
> or input file) or if there is another problem.  As I stated before the
> basic simulation parameters that I used for the original and hydrogen
> bonding damping simulations were used for the hydrogen bonding damping +
> 3rd order energy term simulations.  So, I think the input file is good. 

  I have no experience with the third order stuff, it had been
implemented in DFTB+ due to public demand rather then my conviction
about the usefulness of this highly empirical extension. I never applied
it to other systems as those referred in the method paper.

 Also, has anyone ever run simulations utilizing both the
> hbonding damping term and the 3rd order energy term?  The above divide
> and conquer simulations (J Phys Chem A., 2007, 111(26): 5685–5691)only
> compared the original parameters and the Hbonding damping function
> improvement.

 I'd suggest to contact the authors of the 3rd order paper and ask about
any possible problems using their extension for the systems you
describe, or just asking them, why they did not use it for the water
simulation, which apparently appeared in the same year as the method paper.

  The only thing I can contribut to you problem, is letting you
immediatley know, when something useful comes out from the comparison of
the 3rd order in DFTB+ with the authors version.


  Best regards

     Bálint


-- 
Dr. Bálint Aradi
Bremen Center for Computational Materials Science, University of Bremen
http://www.bccms.uni-bremen.de/en/~b_aradi


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 260 bytes
Desc: OpenPGP digital signature
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20090316/a6e4bb69/attachment.sig>