Sigma for only a single k point or two k points

[Rohit Rana]

Hi,

Similar to the silicon example, I was able to run Epsilon successfully for a system with a 6x6x6 Gamma-Centered grid for WFN and 0.001 offset for WFNq. I only want to run Sigma at the Gamma point and at k = (0,0,0.005) in crystal coordinates. What I did was two separate WFN_co calculations for a 1 x 1 x 1 grid corresponding to these points. My intent was then to do two separate Sigma calculations  while using these two coarse grids as my WFN_inner's. 

However, when I tested this out for the Gamma point, I get the following error:
ERROR: genwf: rkq mismatch

May you please help me with this? I can provide any input or output files that are needed.

Thank you for your time.

Best,

Rohit

[Zhenglu Li]

Dear Rohit,

You need to still use the WFN that contains the full kgrid.  However, you can simply just set 1 kpoint of your interest in sigma.inp.  This k-point should exist in WFN.  The WFN should also be consistent with the epsilon calculation.

Best,

Zhenglu

---

From: 'Rohit Rana' via BerkeleyGW Help <he...@berkeleygw.org>
Sent: Saturday, June 28, 2025 7:36 PM
To: BerkeleyGW Help <he...@berkeleygw.org>
Subject: Sigma for only a single k point or two k points

 

Hi, Similar to the silicon example, I was able to run Epsilon successfully for a system with a 6x6x6 Gamma-Centered grid for WFN and 0. 001 offset for WFNq. I only want to run Sigma at the Gamma point and at k = (0,0,0. 005) in crystal coordinates. 

--
You received this message because you are subscribed to the Google Groups "BerkeleyGW Help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to help+uns...@berkeleygw.org.
To view this discussion visit https://groups.google.com/a/berkeleygw.org/d/msgid/help/0aa38140-c18e-4cc2-a090-85d8f45a453an%40berkeleygw.org.

[Rohit Rana]

Hi Zhenglu,

Thank you for the quick reply. Just to make sure I understand this well, I just had two more questions. 

1. When you say the WFN should be consistent with epsilon, do you mean that the qpoints in epsilon match with the kpoints that generate the Sigma WFN_inner? 

For example, I used a half-shifted grid for my WFN in epsilon, even though the q-points in the input come from the unshifted grid generated by kgrid.x. As far as I can tell, this seems ok to do for epsilon, but I have to used the unshifted grid’s WFN for WFN_inner in Sigma. 

2. If I wanted to get the value of the energy at a k-point in between two points within my sigma calculation, is it safe to use the inteqp.x executable along with a “fine” WFN grid that is 1x1x1 with dk1,dk2,dk3 set to the value of the k vector?

Thank you again for the help. I really appreciate it. 

Best,

Rohit

[Zhenglu Li]

Dear Rohit,

For the shifted grid, yes, you can use the unshifted grid with the same k-mesh density for Sigma.  As long as all q (from epsilon) can be found in the WFN k-grid, defined as q = k - k', i.e. all q points can be found in the WFN_inner from the difference of two kpoints, that would be fine.

For the second point, you can try use WFN_outer, which can, in principle, get quasiparticle energy of a different kpoint, but this functionality hasn't been maintained/tested for years.  Inteqp should also work, you can give it a try.  Alternatively, Wannier interpolation is also a straightforward process to get band energy of an arbitrary kpoint.  The relevant utility is sig2wan.x.

Best,

Zhenglu

---

From: Rohit Rana <rohit...@berkeley.edu>
Sent: Saturday, June 28, 2025 10:26 PM
To: Zhenglu Li <zhen...@usc.edu>
Cc: BerkeleyGW Help <he...@berkeleygw.org>
Subject: Re: Sigma for only a single k point or two k points