[Wannier] different dipole moments due to molecule translation

[Giovanni Pizzi]

Dear Pedro,
in a solid, the polarization is defined modulo a lattice translation 
(times e/cell_volume). See for instance 
http://www.physics.rutgers.edu/~dhv/pubs/local_preprint/dv_fchap.pdf
You can understand this also by noticing that a complete basis set is 
given by each WF translated by a generic real-space vector R; the 
minimization routine in the code can pick up any of these.

In your case of a single molecule and not of a solid, you should try to 
translate all Wannier centres back inside the unit cell to have a 
meaningful result (you may also try the translate_home_cell input flag 
that should do this automatically).

At this point you should get the same result. If you still get slightly 
different results: it seems to me that the two relaxations that you have 
done provide different atomic positions for the H20 molecule (apart for 
the global shift).
Try to simply do the h20.relax step once; then simply rigidly translate 
the molecule to the cell centre, and perform a scf+nscf+wannier 
calculation to check your results.

Best,
Giovanni Pizzi


On 07/17/2012 08:07 PM, Pedro Augusto F. P. Moreira wrote:
>   Dear all.
>
>  I am trying to calculate electric dipole moments of an isolated water 
> molecule through Wannier functions (Wannier90 and QE). Basically, I 
> want to reproduce the results on Silvestrelli & Parrinello, 1999 - J 
> Chem Phys. Their computations were performed at the Gamma point only 
> of the Brillouin zone, with a plane-wave cutoff of 70 Ry and a BLYP 
> density functional. The monomer was optimized in a cubic supercell of 
> side 10.6 Å.  They found a dipole moment of 1.87 D.
>
>  I used the same description in two simulations. However, in one of 
> them, the water molecule were translated by 5.0 Å at z-direction in 
> the beginning of the computations. That is the only difference between 
> the simulations. I provide the inputs and the main results attached to 
> the e-mail. My WF calculations followed steps below:
>
> 1) pw.x < h2o.relax > relax.out
> 2) pw.x < h2o.nscf > nscf.out
> 3) wannier90.x -pp h2o
> 4) pw2wannier90.x < h2o.pw2wan > pw2wan.out
> 5) wannier90.x h2o
>
>  In the centered-molecule simulation, I found 1.8559 D, which I can 
> say this value agrees with the paper one. While, in the 
> translated-molecule simulation, I found 103.4 D.
>  The difference is enormous for two very similar simulations. I cannot 
> explain so great difference due to a translation that is lesser than 
> the supercell side.
>
>  I would be pleased if someone could confirm my results. The attached 
> files with the characters '_tr' are of the translated-molecule 
> simulation.
>
>  Cheers,
>
>  Pedro Moreira
>
> --------------------------------
> IFGW - Unicamp - Brazil
>
>
>
>
>
>
>
> _______________________________________________
> Wannier mailing list
> Wannier at quantum-espresso.org
> http://www.democritos.it/mailman/listinfo/wannier


-- 
Giovanni Pizzi
Post-doctoral Research Scientist
EPFL STI IMX THEOS
MXC 319 (Bâtiment MXC)
Station 12
CH-1015 Lausanne (Switzerland)
Phone: +41 21 69 31159

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/wannier/attachments/20120717/3b2071eb/attachment.html>