Possible bug when set extended_kernel for finite Q kernel calculation

[白云飞]

Dear developer:

I am doing a calculation with both  extended_kernel and exciton_Q_shift turned on.  I found some  indications that BerkeleyGW  may not did it correctly.

When comparing besmat.h5 generated for Q=0 and  a finite Q1. Something strange happened when I check the results in bsemat.h5 ,specifically the kernel matrix.

 I found when the four band indices are all conduction bands. The head matrix ,wing matrix and body matrix elements of Q=0 calculation are exactly the same as those of a Q=Q1 calculation. Otherwise, those matrix are different as expected(when at least one band indices is valance bands, or for exchange matrix)

I suspect, when writing the code, finite Q shift is only applied to valance band state. Which works fine without  exciton_Q_shift but fails in finite Q calculation . 

I wonder whether BerkeleyGW correctly treat extended kernel with finite Q, especially the extended kernel  involving four conduction bands states

[Zhenglu Li]

Dear Yunfei,

[Somehow your email comes in twice.] 

In BerkeleyGW implementation, extended kernel only works with Q = 0.  On the other hand, non-zero finite momentum Q only works with TDA (Tamm-Dancoff approximation, i.e. only cv -> cv transitions).  The two features are not compatible.  The four conduction bands finite-Q falls into the incompatible regime.

The conclusion is: The current BerkeleyGW should not be used for finite-Q extended kernel.  The matrix elements are not useful.  This feature has not been implemented. 

Hope this information is helpful.

Best,

Zhenglu

---

From: 白云飞 <baiyun...@gmail.com>
Sent: Monday, December 16, 2024 10:29 PM
To: BerkeleyGW Help <he...@berkeleygw.org>
Subject: Possible bug when set extended_kernel for finite Q kernel calculation

 

Dear developer: I am doing a calculation with both  extended_kernel and exciton_Q_shift turned on.  I found some  indications that BerkeleyGW  may not did it correctly. When comparing besmat. h5 generated for Q=0 and  a finite Q1. Something strange

--
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/9da0a466-1a24-4f92-b2e7-3cad68961bd0n%40berkeleygw.org.

[白云飞]

Thanks a lot for your reply.

In order to correct grammar, I  deleted the previous post. But unexpectedly, it found you anyway

Sorry for the bothering

[Zhenglu Li]

Never mind.  You are very welcome.

---

From: 白云飞 <baiyun...@gmail.com>
Sent: Monday, December 16, 2024 10:48 PM
To: BerkeleyGW Help <he...@berkeleygw.org>
Cc: Zhenglu Li <zhen...@usc.edu>; 白云飞 <baiyun...@gmail.com>
Subject: Re: Possible bug when set extended_kernel for finite Q kernel calculation

 

Thanks a lot for your reply. In order to correct grammar, I deleted the previous post. But unexpectedly, it found you anyway Sorry for the bothering 在2024年12月17日星期二 UTC+8 14: 34: 09<zhen. . . @ usc. edu 写道： Dear Yunfei, [Somehow your email comes