[Pw_forum] ecutvcut in Hybrid functionals

[Layla Martin-Samos]

Dear Leonardo and Jianpeng, the "issue" is that HSE contain only the
shortrange part of coulomb potential for the fock term (thas why HSE
contain an exponential with a screening lenght). Ecutvcut, spehrical cutoff
and all similar tricks APPLY ONLY for the "REAL" Coulomb potential, the
FULL Coulomb potential. So, one has to be very careful when dealing with
hybrids that do not contain the full coulomb potential.

cheers

Layla

2012/11/23 Jianpeng Liu <jpliu at physics.rutgers.edu>

> Hi Leonardo,
>
> I really do not know about this. For security, I followed the example and
> set ecutvcut to be zero for a HSE hybrid functional.
>
> best,
> Jianpeng
>
> > HSE is "a PBE0 with a cutoff". I don't know if it has been fixed, but
> some
> > time ago ecutvcut interfered with that cutoff, making HSE behave just as
> > PBE0 (i.e., when ecutvcut was on, calculations done with HSE and PBE0
> gave
> > the same results).
> > Maybe that's why in the examples with HSE was used ecutvcut=0.0. I just
> > don't know if this is still happening, or if the bug was fixed and just
> > the
> > example that wasn't updated.
> >
> > Leonardo Jorge
> > PhD Student at USP - Brazil
> > Nanomol Group
> > http://nanomol.if.usp.br
> >
> > On Tue, Nov 20, 2012 at 10:48 AM, Jianpeng Liu
> > <jpliu at physics.rutgers.edu>wrote:
> >
> >> Hi Layla,
> >>
> >> Your comments are really really helpful! Thanks a lot!
> >>
> >> best,
> >> Jianpeng
> >>
> >> > Dear Jianpeng, ecutvcut is a parameter which is not strictly bonded to
> >> > hybrid functional parameters, but to strongly anisotropic supercells
> >> like
> >> > in the case of quasi 1D systems and fock like operators. I will be
> >> more
> >> > precise. If you have an hybrid functional AND you want to have the
> >> > electronic properties of a polymer chain or a nanotube, i.e. if you
> >> > multiply the cell parameter by the number of kpoint in each direction
> >> then
> >> > you obtain an effective supercell which is anisotropic (like 6*6*50),
> >> then
> >> > to avoid divergencies for small q vectors in the integration of the
> >> fock
> >> > operator, one has to turn on exxdiv_treatment="vcut_ws" and
> >> ecutvcut=1.2
> >> > or
> >> > something appropriate (you need to converge with this value). vcut_ws
> >> > applies a real space cutoff in the wigner Seitz cell of your supercell
> >> > until a reciprocal space cutoff of ecutvcut. The more is your cell
> >> > anisotropic bigger has to be ecutvcut. If you are dealing with bulk
> >> > materials, you do not have to care about this!!!
> >> >
> >> > hope it helps cheers
> >> >
> >> > Layla
> >> >
> >> > 2012/11/19 Jianpeng Liu <jpliu at physics.rutgers.edu>
> >> >
> >> >> Dear QE users and developers,
> >> >>
> >> >> I am doing hybrid-functional calculations, and I am confused about
> >> the
> >> >> setting of ecutvcut. In the examples of Si, for pbe0-type hybrid
> >> >> functional, the ecutvcut is set to be 0.7Ry, while for hse-type
> >> hybrid
> >> >> functional with a screened Coulomb potential, ecutvcut is just set as
> >> >> default which is 0.0Ry. Could anyone explain why? For a typical
> >> >> semiconductor with s or p valence orbitals, what kind of ecutvcut
> >> value
> >> >> is
> >> >> recommended?
> >> >>
> >> >> Thanks a lot.
> >> >>
> >> >> best,
> >> >> Jianpeng Liu
> >> >>
> >> >> -------------------------------------
> >> >> Ph.D. candidate,
> >> >> Department of physics and astronomy,
> >> >> Rutgers University
> >> >> -------------------------------------
> >> >>
> >> >> _______________________________________________
> >> >> Pw_forum mailing list
> >> >> Pw_forum at pwscf.org
> >> >> http://pwscf.org/mailman/listinfo/pw_forum
> >> >>
> >> > _______________________________________________
> >> > Pw_forum mailing list
> >> > Pw_forum at pwscf.org
> >> > http://pwscf.org/mailman/listinfo/pw_forum
> >>
> >>
> >> _______________________________________________
> >> Pw_forum mailing list
> >> Pw_forum at pwscf.org
> >> http://pwscf.org/mailman/listinfo/pw_forum
> >>
> > _______________________________________________
> > Pw_forum mailing list
> > Pw_forum at pwscf.org
> > http://pwscf.org/mailman/listinfo/pw_forum
>
>
> _______________________________________________
> 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/20121123/c4a4e274/attachment.html>