The ABINIT Users Mailing List ( CLOSED )

[  <mz24cn@hotmail.com>]

Dear Fiorentini,
  Yes, just as what you said, I had noticed that a second order polynomial 
may lead to unstable results sometimes. In fact, from the theoretical point 
of view, second order polynomial is always true to extract the effective 
masses when delta k vector in BZ is so small that higher order components 
of the polynomial can be omitted. But as I had described in last letter, 
'the ABINIT-output eigenvalues is fixed to the format of five-digit decimal 
fraction', when delta k is too small so that you can only acquire the 
'identical' eigenvalues for different k vectors. So, as for ABINIT package, 
kptbounds should be carefully choosed. And use unit eV instead of unit 
hatree to get as possible as many effective digits. I found delta k=0.02 
for ndivk=40 is enough to get ideal results for silicon. For other 
materials, you should carefully try kptbounds and ndivk value OR, use 
higher order polynomial.

Good continuation,
Mike  


> From: Vincenzo Fiorentini <vincenzo.fiorentini@dsf.unica.it>
> Reply-To: forum@abinit.org
> To: forum@abinit.org
> Subject: Re: [abinit-forum] Using ABINIT to find effective masses of 
electrons and holes of semiconductors &
> Date: Sat, 14 Sep 2002 17:39:03 +0200
>
> dr.horsfield@physics.org wrote:
>
>
> > For my thesis (many years ago) I did something similar with the 
> > Corning
> > Code (Abinit's precursor). I found the results were sensitive to 
> > the
> > fitting polynomial. Have you tried anything other than a quadratic?
>
>
> (about effective masses) I would agree a second order polynomial may 
> be poor in some instances. I did something too a while ago (PRB 46, 
> 2086 (92)) with a fourth order polynomial and it seemed fairly 
> stable.
>
> best - Vincenzo
> --
>
>     Prof. Dr. Vincenzo Fiorentini
>
>     vincenzo.fiorentini@dsf.unica.it
>     INFM - Dipartimento di Fisica
>     Universita' di Cagliari
>     Cittadella Universitaria
>     S.P. Monserrato-Sestu Km 0.700
>     I-09042 Monserrato (CA), Italy
>     tel +39 070 6754912
>     fax +39 070 510171




_________________________________________________________________
与联机的朋友进行交流，请使用 MSN Messenger: 
http://messenger.microsoft.com/cn/