[Pw_forum] graphite cell optimization failed

Wed Mar 23 14:56:15 CET 2011

Dear Giuseppe and Gabriele
I agree with Gabriele that PBE-D2 gives better results than vdw-DF in the
case of graphite.
About CP, some time ago I interfaced the DFT-D2 module of PWscf to cp.x. At
that time, I used the 4.2 release of espresso, which is a little old now. I
can send you that old package if you are interested.
We also tested the DFT-D3 method on a couple of systems and, in my opinion,
it does not give very exciting results. But interestingly, as suggested by
Giuseppe, in some cases we found a good agreement between DFT-D3 and vdw-DF
data. Hopefully the two series vdw-DF(n) and DFT-D(n+1) will converge to the
same result :)

regards

daniel

2011/3/23 Gabriele Sclauzero <sclauzer at sissa.it>

> Dear Giuseppe,
>
>    I think your reply reached also the PW-forum, since I've received it
> twice, so everyone should have been able to read it.
>
> Il giorno 23/mar/2011, alle ore 10.48, giuseppe.mattioli at mlib.ism.cnr.itha scritto:
>
> Dear Gabriele
>
>   is there any concrete evidence that "ab initio" van der Waals
>
> vdW-DF is systematically better than Grimme's vdW?
>
>
> Yes, at least in the case of molecule-surface interactions.
>
>
> Maybe graphite is another story... with PBE-D2 you should get an interlayer
> spacing just 3.4% smaller than the experimental value, which is not so bad
> considering the much larger PBE error.
>
>
> There is a
> recent (2010) Grimme paper (I'm sorry, I cannot be more accurate
> today...)
>
>
> Maybe this one:
> THE JOURNAL OF CHEMICAL PHYSICS *132*, 154104 (2010)
> which is the paper introducing the D3 method (and a comparison with its
> predecessor D2)
>
> that reports an overestimate of the above interactions when
> using DFT-D2, currently implemented in QE. I've performed comparative
> tests (classical MM, DFT-D2, vdW-DF) on several molecules interacting
> with the non-polar 10-10 ZnO surface and I've obtained the best
> results (some of them comparable with experimental data, i.e., the
> thiophene ring) by using the vdW-DF approach. The DFT-D2 setup,
> however, is quite good for molecules in my experience, and can be
> tuned by changing the C6 coefficients by hand
> (QE/Modules/mm_dispersion.f90), following one of the available force
> fields (AMBER, CHARMM, ...).
>
> Maybe DFT-D3 and vdW-DF would converge to
> similar results...
>
>
> I can't tell this, but I know that DFT-D3 should include some notable
> improvements with respect to the D2 version, since it now takes into account
> some kind of local coordination of the atoms, atom-pair specific
> coefficients (from ab initio), and should also separate better the long
> range regime from the short range one. I know there is someone who offered
> to look at implementing the extension to D3, maybe it will be available in a
> few months time.
>
> What about cp?
>
>
> No clear idea here. However I believe it should not be difficult to port
> DFT-D2 to CP, since it requires only atomic positions and the cell setup.
> These should be handled in a similar way in PW and CP.
>
>
> Gabriele
>
>
> Yours
>
> G.
>
> Quoting Gabriele Sclauzero <sclauzer at sissa.it>:
>
> Dear Giuseppe,
>
>
>   is there any concrete evidence that "ab initio" van der Waals
>
> vdW-DF is systematically better than Grimme's vdW?
>
> Are you aware of any paper comparing the latest version of the two
>
> methodologies (or "phisolophies", if you want), i.e. vdW-DF2 vs
>
> DFT-D3?
>
>
> Cheers,
>
>
> GS
>
>
>
> Il giorno 23/mar/2011, alle ore 10.12,
>
> giuseppe.mattioli at mlib.ism.cnr.it ha scritto:
>
>
>
> Dear all
>
> You can do better than the semiempirical "Grimme" correction. There is
>
> a new ab initio van der Waals correlation functional implemented in QE
>
> 4.3a. Try    input_dft='vdW-DF', in the &sysytem list.
>
> By the way, what about a (semiempirical or ab initio) vdW scheme in
>
> cp.x? Is there any project going on?
>
>
> Yours
>
>
> Giuseppe
>
>
> Quoting Masoud Nahali <masoudnahali at gmail.com>:
>
>
> Dear Eric
>
>
> Try london="true" to get improved answer; as you should know DFT is weak
>
> to calculate dispersion forces and by using the keyword one can include
>
> these forces semi-empirically. I hope it helps.
>
>
> Best Wishes
>
>
> Masoud
>
> --------------
>
> Masoud Nahali, Sharif University of Technology
>
>
>
> Eric Wrote:
>
>
> Dear all,
>
>
> I'm trying to optimize unit cell parameter of graphite.
>
> The problem is the c axis gets too long.
>
> Starting from published data my initial parameters are a=4.59203 Bohr
>
> and c=12.513 Bohr.
>
>  The final values are 4.664 and 15.829 respectively.
>
> I need someone to explain to me what I did wrong, I've attached my input
>
> file.
>
> That's a basic calculation so it's very frustrating to not make it done.
>
> I thank you in advance,
>
>
>                                    Eric.
>
>
> --
>
> /Be the change you wish to see in the world
>
> / --- Mahatma Gandhi ---
>
>
> Dr. Éric Germaneau <mailto:germaneau at gucas.ac.cn
>
> <http://www.democritos.it/mailman/listinfo/pw_forum>>
>
>
> College of Physical Sciences
>
> Graduate University of Chinese Academy of Sciences
>
> Yuquan Road 19A
>
> Beijing 100049
>
> China
>
>
>
> Giuseppe Mattioli
>
> ISM-CNR
>
> Italy
>
>
>
>
>
> _______________________________________________
>
> Pw_forum mailing list
>
> Pw_forum at pwscf.org
>
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>
>
> § Gabriele Sclauzero, EPFL SB ITP CSEA
>
>   PH H2 462, Station 3, CH-1015 Lausanne
>
>
>
>
>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>
>
> § Gabriele Sclauzero, EPFL SB ITP CSEA
> *   PH H2 462, Station 3, CH-1015 Lausanne*
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>

-- 
 ----------------------------------

  Daniel Forrer
  Dipartimento di Scienze Chimiche
  Università degli Studi di Padova
  V. Marzolo 1, Padova
  mail   daniel.forrer at unipd.it
  tel.   +39 049 8275166

 ----------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20110323/b931d5d5/attachment.html>