The ABINIT Users Mailing List ( CLOSED )

["matthieu verstraete" <matthieu.jean.verstraete@gmail.com>]

if your periodic cell is charged the coulomb energy should explode. Abinit can deal with this if a uniform compensating charge is added (http://www.abinit.org/Infos_v5.6/input_variables/vargs.html#charge). In principle you should add corrections to the energy (e.g. Makov and Payne http://prola.aps.org/forward/PRB/v51/i7/p4014_1 although this is far from the best, see also PA Schultz's articles), but these are not present in abinit. This approach is used quite frequently in solid state defect studies, but actually can converge quite badly (even with the corrections) with system size. No real solution has been found yet to get small sizes and converged defect energies etc... Others would tell you to use a finite system cluster code, but then you don't have a very realistic crystal. The jellium background is a better and better approximation as the cell gets bigger of course.

The other option (for finite systems or clusters in abinit) is to use the wavelet Poisson solver (icoultrtmt - note the restrictions on cell size and shape) and remove all interaction with neighboring cells (but if you have a crystal this will not be satisfactory until large cell sizes).

"it's a hard knock life"

Matthieu

On Thu, Sep 11, 2008 at 5:07 PM, Manuel Pérez Jigato <Manuel.PerezJigato@fys.kuleuven.be> wrote:

Dear all

is it possible to calculate charged cells using ABINIT. i mean for defect systems
regards

Manuel Perez Jigato
KU Leuven

Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dr. Matthieu Verstraete

European Theoretical Spectroscopy Facility (ETSF) 
Dpto. Fisica de Materiales, 
U. del Pais Vasco,            
Centro Joxe Mari Korta, Av. de Tolosa, 72, Phone: +34-943018393
E-20018 Donostia-San Sebastian, Spain Fax  : +34-943018390

Mail : matthieu.jean.verstraete@gmail.com
http://www-users.york.ac.uk/~mjv500