[Pw_forum] intrinsic dipole of (large, anisotropic) molecule

[Marton]

Hi Chris,

If you want to compute the ground state dipole moment of a molecule then
there is no need to apply any electric field. Set

assume_isolated = 'mp'

in the system namelist, which will then give you the dipole moment of the
molecule at the end of the scf, supposing the molecule is isolated. Note
that the dipole moment computed this way converges quite slowly with
respect to the size of the box. I know that your attached input is not
fully converged but in this case, the size of the box is crucial and your
box size is certainly not enough to get any realistic number. I'm actually
quite sure that your finite field calculations are not giving reasonable
numbers because there is basically no vacuum region: in the finite field
approach the region where the sawtooth potential changes sign should be in
the "vacuum" with little or no density:

http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PW.html#idm140629872333184

Also, if you are interested in computing the polarizability using DFPT you
won't need to do 3N perturbations. Just ask QE to compute the
high-frequency dielectric constant (epsil=.true. and trans=.false. in the
ph.x input). There will only be three perturbations corresponding to the
three cartesian components of the external field. If you use the gamma
k-point to sample the BZ zone then QE will extract the molecular
polarizability from the dielectric constant using the Clausius-Mossotti
formula and will report it in the output. Again, you would have to converge
the result with the cell size.

HTH,
Marton Voros

--
Materials Science Division
Argonne National Laboratory

On Sat, Apr 15, 2017 at 10:38 PM, Christoph Wolf(신소재공학과) <
chwolf at postech.ac.kr> wrote:

> Dear all!
>
>
>
> I am currently trying to reproduce calculated dipoles for organic
> molecules previously reported. As an example I will use Alq3 because it is
> well studied in literature.
>
>
>
> I have read a previous discussion on the topic of dipole calculation using
> external fields: http://pw_forum.pwscf.narkive.com/1odbbguH/dipole-moment-
> calculation
>
>
>
> In brief, using lelfield and dipfield Giovanni and Aihua were able to
> calculate the dipole of H2O in good agreement with literature (1.88 D vs
> 1.89 D).
>
>
>
> Alq3 has  been previously reported e.g. https://journals.aps.org/prb/
> pdf/10.1103/PhysRevB.61.15804 there, the authors note:
>
>
>
> “We find that the calculated molecular polarizability accurately describes
> the solid-state polarization corrections and can be used to determine the
> measured static dielectric constant. The calculated molecular dipole moment
> can be used to interpret the electric-field dependence of the electron
> mobility.” Using B3LYP in Gaussian90 they arrive at a dipole moment of
> d=5.3 Debye; unfortunately I have never worked with Gaussian and don’t know
> how this calculation was most likely performed.
>
>
>
> As a “quick and dirty” approach I used the method outlined for H2O in the
> first link and calculate the dipoles for efield=1,2,3 (input structure is
> in the attached image); the results are:
>
>
>
> Computed dipole along edir(1) :
>
>         Elec. dipole         -0.5198 Ry au,         -1.3213 Debye
>
>         Ion. dipole          -0.5375 Ry au,         -1.3662 Debye
>
>         Dipole              -11.2534 Ry au,        -28.6033 Debye
>
>         Dipole field         -0.0177 Ry au,
>
>
>
> Computed dipole along edir(2) :
>
>         Elec. dipole         -0.1980 Ry au,         -0.5033 Debye
>
>         Ion. dipole          -0.2097 Ry au,         -0.5330 Debye
>
>         Dipole               -7.4406 Ry au,        -18.9122 Debye
>
>         Dipole field         -0.0117 Ry au,
>
>
>
>   Computed dipole along edir(3) :
>
>         Elec. dipole          0.0367 Ry au,          0.0932 Debye
>
>         Ion. dipole           0.0346 Ry au,          0.0879 Debye
>
>         Dipole               -1.3285 Ry au,         -3.3768 Debye
>
>         Dipole field         -0.0021 Ry au,
>
>
>
> Now this gives rather large dipoles in x,y and “compatible” dipole
> magnitude in z direction.
>
>
>
> Can someone comment on how to improve (in terms of quality of the
> reproduction of d=5.3 D) these results, ideally I want to avoid ph.x
> polarizability calculations because the molecules are rather large (52
> atoms x 3 modes = long time…)
>
>
>
> Input file is attached, please not I have not converged the calculation
> with respect to position of the molecule in the box and cell size.
>
>
>
> Yours,
>
> Chris
>
>
>
> Christoph Wolf
>
> Postech University, Department of Materials Science and Engineering
>
> Pohang, Republic of Korea
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20170416/a768f4f6/attachment.html>