The ABINIT Users Mailing List ( CLOSED )

[ธนูสิทธิ์ บุรินทร์ประโคน <thanusit@kku.ac.th>]

Dear ENRICO

I think your results is from the mis-interpretating of the "acell" parameter, as mentioned by PMA. For fcc lattice, the lattice parameter "a" of can be relalted to an "acell" parameter by a  = 2 * rprim*acell. (anyone please correct me if this is not so) 

In case of "optcell=1", rprim (any non-zero commopnent) is 0.5 so that we get a = acell. However, for "optcell=2", rprim is 0.7xxx so that "a" is not equal to "acell". This is why you get the big differences. 

Hope this help.

 

Kind Regards,

Thanusit Burinprakhon


On September 26, 2008 11:17:52 PM ICT, "Anglade Pierre-Matthieu" <anglade@gmail.com> wrote:

Hi,

I think it may be related to a change of "rprim" in you input file.
Abinit Cell relaxation algorithm is pretty funny for that. It
imposes that all the vectors of rprim have a norm of 1...
Have a look in your output and please tell us if your problem is not
just linked to misreading of abinit output file ...
Usually you will see that you rprim have change from
rprim 0.5 ....
rprim 0.7xx ....
You have to interpret your acell accordingly. And it will account for
the very large discrepancies with respect to experiment you are
witnessing.


regards

PMA

On Fri, Sep 26, 2008 at 6:10 PM, Enrico wrote:
> Hi,
> i've done some relaxation on ( optcell=2 Volume and atomic position) Ni FCC.
> And i got very big difference in lattice parameter; in lDA 48% of the original
> value!
> In GGA 40,3%.
>
> I did some other FCC like La FCC and LaNi2 in LDA and again i noticed this huge
> variation!
>
> Instead using ALWAYS THE SAME OPTIMIZATION parameters I did LaNi5 ( in LDA) and
> i got only 3,1% on a and 3,3% on c, which is more what LDA should do.
>
> My PSEUDO in LDA are HGH WITHOUT SEMICORE only 10 electroins.
>
> I've just noticed the same variation in GGA with ixc=11 on Ni using
> 28-Ni.GGA.fhi pseudo.
>
> Could tell me where i'm wrong?
>
> Tahnk you
> ENRICO
>
> here follows the INPUT FOR Ni, I use to copy and paste the Relax part of the
> input.
>
> #UNIT CELL Ni
>
> acell 3*3.5232 Angstr
> rprim 0.0 0.5 0.5 # FCC Real space PRIMitive translations
> 0.5 0.0 0.5
> 0.5 0.5 0.0
> ntypat 1
> znucl 28
> natom 1 # atom per cell
> typat 1
> xred
> 0.0 0.0 0.0
>
> #GEOM OPTIMISATION
> optcell 2 # optimisation of volume & ...
> ionmov 3 # conduct structural optimization using the
> Broyden-Fletcher-Goldfarb-Shanno minimization (BFGS),
> # modified to take into account the total energy as well as the
> gradients (as in usual BFGS)
> ntime 30
> dilatmx 1.09 # the maximal permitted scaling of the lattice parameters when
> the cell shape and dimension is varied
> ecutsm 0.5 # Using a non-zero ecutsm, the total energy curves as a function
> of ecut, or acell,
> can be smoothed, keeping consistency with the stress
>
> #RECIPROCAL SPACE INTEGRETION
> kptopt 1
> nshiftk 4 # shiftk=4 particular efficient for FCC
> shiftk 0.5 0.5 0.5
> 0.5 0.0 0.0
> 0.0 0.5 0.0
> 0.0 0.0 0.5
> ngkpt 8 8 8
> nband 19
> tsmear 0.02
> occopt 4
>
> #PLANE WAVES
> ecut 140
>
> #SCF PROCEDURE
> nstep 150
> ixc 1
> toldfe 1.0d-7
>



--
Pierre-Matthieu Anglade