[Pw_forum] charged species

Nicola Marzari marzari at MIT.EDU
Wed Mar 12 10:41:52 CET 2008

Igor Pasti wrote:
> Dear QE users,
> I was wondering if it is possible to study charged species with QE. For 
> example, is it possible to define OH- ion adsorbed on Ni surface in such a way 
> that charge is localised on O atom. If this is possible, I would greatly 
> appreciate if anyone can tell me how to do that.
> Regards,
> Igor Pasti

Hi Igor,

1) you can define the number of electrons in your unit cell at will

2) with some care, you can then minimize the errors deriving from having
a unit cell with a number of electrons different from the sum of
ionic valence charges (see e.g. a recent arxiv.org post by Dabo et al.,
and references therein). Since in periodic conditions it you cannot have
a charged unit cell, the electrostatic algorithms in the code 
automatically assume a compensating background - understanding how this
compensation changes your results is important.

3) the charge defined in 1) will go wherever your hamiltonian tells it 
to go - LDA or GGA sometimes perform poorly in this respect.

In your case, if you want to study a negatively charged adsorbed OH^-, 
you need to specify the number of electrons (one more the ones that
PWSCF would choose), and make sure your errors from the compensating 
background are not too large (e.g. change the amount of vacuum, or the 
lateral dimensions). Then, you hope that generically GGA will correctly
assign one more electron around OH - do a charge density difference 
between the calculation done with the charged system, and the neutral 
system, and see where the extra electron has gone.

Last, compare your results with experiments - e.g. the vibrational
frequencies for your group, charged or neutral, and see if it all
makes sense (see e.g. also Dabo et al JACS 2007).


Prof Nicola Marzari   Department of Materials Science and Engineering
13-5066   MIT   77 Massachusetts Avenue   Cambridge MA 02139-4307 USA
tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu

More information about the users mailing list