[DFTB-Plus-User] MD in dftb+ (SCC=Yes/No?)

Ben Hourahine benjamin.hourahine at strath.ac.uk
Mon Dec 17 13:34:27 CET 2012


Dear 刘雷,

as Bálint explains, the crucial question is whether your SK-set parameters are constructed to be
used for SCC calculations, this is given in the documentation at the end of each SK file.

Most sets do require SCC, matsci and some of the PBC set being exceptions. If you have a homonuclear
system (the pure C example Bálint gives, then non-SCC may be safe to use even for SCC SK parameters
(but as aways, test this before relying on it).

Regards

Ben

On 17/12/12 11:51, Bálint Aradi wrote:
> Dear 刘雷,
>
> On 12/17/2012 03:18 AM, 刘雷 wrote:
>> Dear all,
>> I did a test about "SCC=Yes" or "SCC=No" when I try to do MD simulation in dftb+,the simulation speed of "SCC=No"  is ten times of the speed of "SCC=Yes".
>> Previously,I have noticed there was someone who used "SCC=No" when he did MD simulation.In my opinion,we shoule use "SCC=Yes" though its slow simulation speed.
>> Am I right? If not,I want to know the explanation.
>   Of course non-scc is faster, as you do not make any self consistency
> loop for the charges. In general, you should use SCC = Yes, especially
> if you are using SK-sets which were made for SCC calculations.
>
>    You may get away to make a *few* non-scc steps, if SCC does not
> converge for your geometry. Non-scc could be also an option, if you have
> a homonuclear system without any significant charge transfer between the
> atoms (e.g. carbon nanotubes, fullerenes). But apart of those special
> cases, the general advice is to use SCC, as otherwise your forces and
> consequently your MD-trajectory could be wrong.
>
>    Best regards,
>
>    Bálint
>
>

-- 
       Dr. B. Hourahine, SUPA, Department of Physics,
     University of Strathclyde, John Anderson Building,
             107 Rottenrow, Glasgow G4 0NG, UK.
     +44 141 548 2325, benjamin.hourahine at strath.ac.uk

   Strathclyde 2012 THE Awards UK University of the Year

    The University of Strathclyde is a charitable body,
         registered in Scotland, number SC015263



More information about the DFTB-Plus-User mailing list