[DFTB-Plus-User] Spin-polarized md simulation using DFTB+

Dhungana, Kamal B kamal-dhungana at uiowa.edu
Fri Mar 24 18:03:41 CET 2017

Dear Dr. Aradi,

Thank you very much for explaining our concerns very clearly. We are now very much clear about the spin-constant parameters for spin-polarized dftb calculations.

Best regards,


From: DFTB-Plus-User <dftb-plus-user-bounces at mailman.zfn.uni-bremen.de> on behalf of Bálint Aradi <aradi at uni-bremen.de>
Sent: Thursday, March 23, 2017 10:30:14 PM
To: User list for DFTB+ related questions
Subject: Re: [DFTB-Plus-User] Spin-polarized md simulation using DFTB+

Dear Kamal,

> 1) Whereas 3ob is clearly a set of parameters for atomic resolved SCC
> it does provide a set of orbital resolved spin parameters. Am I to
> understand that I should never use these orbital resolved spin
> parameters and always stick to atom resolved spin parameters for 3ob
> (i.e. always say ShellResolvedSpin=No)?

Unfortunately, different groups are following different traditions. As
for the spin, we (and also the 3ob people) use always shell resolved
spin. Actually, the ShellResolvedSpin option was only added recently, in
order to be able to accomodate the linear response extension, which
traditionally used atomic spin constants instead of shell resolved ones.

> 2) If orbital resolved spin parameters are OK to use with 3ob
> (ShellResolvedSpin=Yes), then are you saying that I should not use
> the values for Sulfur Wsd, Wds, Wpd, Wdd, Wdp provided with the 3ob
> parameter set because the d orbitals are empty?

No. If sulfur had d orbitals, you would have to provide also W*d and Wd*
spin constants. One should only be careful, what one chooses. Again,
different people use different convention here. I prefer to use highest
occupied value for all virtual ones (e.g. in case of S: Wsp for Wsd, Wpp
for Wdd, etc.), but other people may use different strategies. Anyway,
if the authors of a set officially publish some spin coupling constants
for their parameterization, I would advice to use that one, as probably
this is what they used to test their set.

> I do understand that if I was to use atom resolved spin parameters
> Wpp should be the value to use for the S atom.

That is correct, yes. I you use atomic spin, then it should be the one
for the highest occupied atomic orbital.

Best regards,


Dr. Bálint Aradi
Bremen Center for Computational Materials Science, University of Bremen

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20170324/5676ca8a/attachment.html>

More information about the DFTB-Plus-User mailing list