Just to add that plain wave is not a good approximation to reproduce molecular 
orbitals.

Mortaza A

PhD of Physics 
Ulm University, Germany 

> On 28 Nov 2016, at 21:43, Giovanni Cantele <[email protected]> 
> wrote:
> 
> if you try to ask the question: "what the k-point mesh is for?" you can 
> easily understand
> that a k-point mesh is for the Brillouin zone (BZ) sampling, that is, perform 
> sums over k-points
> that approximate, with a required level of accuracy, the integral over the 
> whole BZ. This is
> because in a crystal, the energy E(K) depends on the k vectors, and the same 
> holds for the
> wave functions. However, when you build for example a cubic supercell 
> including vacuum to
> simulate a molecule, you are actually dealing with a fictitious cubic crystal 
> where the repeated
> units are sufficiently far apart from each other to not interact.
> 
> Absence of interaction means flat (atomic-like) energy levels, that do not 
> have dependence
> on the k vector. That is, the E(K) dispersion that you get in the case of a 
> molecule, provided
> you include a sufficient vacuum space in your supercell, does not depend on 
> k. If this is
> the case, you can use only the gamma point (a larger grid means large but 
> useless
> computational time, because you are summing over eigenvalues at different 
> k-points that 
> are all the same!) for self consistency and BZ path is not needed, because if 
> you select any
> path you should get flat, k-independent bands
> 
> Giovanni
> 
> 
>> On 28 Nov 2016, at 18:12, ashkan shekaari <[email protected]> wrote:
>> 
>> Dear experts,
>> 
>> Is it necessary using kpoint mesh when we are dealing with molecules?
>> 
>> In the case of a molecule in cubic vacuum, what path in BZ should be 
>> selected?
>> --
>> Regards,
>> Ashkan Shekaari
>> Plasma Physics Research Center
>> Science and Research Branch
>> I A U, 14778-93855 Tehran, Iran.
>> _______________________________________________
>> Pw_forum mailing list
>> [email protected]
>> http://pwscf.org/mailman/listinfo/pw_forum
> 
> _______________________________________________
> Pw_forum mailing list
> [email protected]
> http://pwscf.org/mailman/listinfo/pw_forum
_______________________________________________
Pw_forum mailing list
[email protected]
http://pwscf.org/mailman/listinfo/pw_forum

Reply via email to