The ABINIT Users Mailing List ( CLOSED )

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

Good news from Don,

may I add that you should not be using the single 1/2 1/2 1/2 shift in FCC structures, as the resulting grid is not compatible with the full set of symmetries. As detailed in one of the tutorials, you should be using 4 shifts
0.5 0.5 0.5
0.5 0 0
0 0.5 0
0 0 0.5

to get a grid which is shifted but also compatible with all symmetries (ie applying a symmetry will always move a kpt onto another kpt in the grid).

The corresponding shifts for hcp bcc etc.. are different, of course. In most cases it doesn't matter too much if you just integrate over the BZ (just need to converge the kpt density) but if you start using symmetries to complete perturbations etc... it does matter, as you have noticed. 

Matthieu

On Thu, Nov 6, 2008 at 11:19 PM, D. R. Hamann <drhamann@mat-simresearch.com> wrote:

Dear Ganesh,

The strain perturbation depends on being able to symmetrize by sums over subsets of the full BZ k-point mesh in ways that the phonon perturbation does not.  The default kshift of 3*0.5 breaks symmetry in most cases.  If the calculation is organized so that phonon and strain calculations are done in different datasets in the same run, this check should be performed before you start any serious calculation, if I recall correctly.  However, the DDB files just give total energy derivatives with respect to perturbations, so  I think you should be able to merge calculations done with different sets of k-points.  The elastic properties tutorial discusses the k-point symmetry issue for one of the cases treated there, but I guess that is easy to overlook.

Good luck,
Don Hamann

PGanesh wrote:

Hi,

I am computing electric-field and strain perturbations for a 20 atom cell with space group Fm-3m at q=(0,0,0)  ( I am using version 5.4.4).  I have already computed the phonon perturbations successfully.  I get the following error :

symkchk : found identity, with number  1
symkchk : ERROR -
 k-point set must have full space-group symmetry
 there is no match for kpt   1 transformed by symmetry   3
 Action : change kptopt to 2 or 3 and/or change or use shiftk
          shiftk = 0 0 0 is always a safe choice.


My kptopt is 2.  I changed it to 3, but still get the same error.  I am afraid that if I use the other option of using shiftk=0 0 0, then the I won't be able to merge it with the database of phonon calculations which were done with shift 0.5 0.5 0.5.  I have two questions:

1> Why am I getting this error for electric-field and strain perturbations, while I didnt get it for phonon pertirbations (i.e. when rfphon =1 )
2> Is there a way to get around the problem, so that I don't have to recalculate the phonon perturbations at q=0 0 0 ?

Thanks,
Ganesh

--
D. R. Hamann
Mat-Sim Research LLC    | Deptartment of Physics
P.O. Box 742            |  and Astronomy
Murray Hill, NJ 07974   | Rutgers University
phone: 908-370-8079     | 732-445-4381

email: drhamann@mat-simresearch.com

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
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