Inquiry about use_velocity and use_momentum in absorption calculation

[Kostas Fykouras]

Hello BGW group,

I am looking through the documentation of absorption.x and some questions rose about the tags use_momentum and use_velocity.

If I use the use_velocity tag, I get the files absorption_eh.dat and absorption_noeh.dat which I can plot using the formula 2(-ε1+(ε1+ε2)**(1/2))**(1/2) to get the absorption spectrum.

With the use_momentum tag I get the  files absorption_b1,2,3_eh.dat and absorption_b1,2,3_noeh.dat. 

Is assumption correct that the b1, b2 and b3 correspond to different directions in reciprocal space ?

If that is the case, does that mean that the Real and Imaginary part of epsilon are decomposed in the three reciprocal directions ?

In addition to that, would the plotting formula for the absorption spectrum change? Instead of just ε1 would I have ε1,b1+ε1,b2+ε1,b3?

Also, In the documentation for use_momentum it states "When you use the momentum operator, you are throwing away the contribution from the non-local part of the pseudopotential to the dipole transition matrix elements". Does that mean that for local (LDA) and semilocal (PBE) pseudopotentials the use_momentum and use_velocity give the same results ?

Thank in advance for your time

[Zhenglu Li]

Dear Kostas,

When you 'use_momemtum', the absorption spectrum is computed based on 'polarization' you specified.  The absorption_b1,2,3 are generated for free as a feature of 'use_momemtum'.  The meaningful interpretation of the absorption spectrum depends on the photon polarization of the light, which can be specified with 'polarization' with 'use_momemtum', or if 'use_velocity', depending on the q_shift direction.

When 'use_momemtum', non-local part of the H (or pseudopotential in practice) is omitted.  This is the (often) case even with LDA or PBE because when the pseudopotentials are generated, there are non-local components in the angular momentum projectors.  It is always safe to 'use_velocity'.  If you want to 'use_momemtum', then tests against 'use_velocity' is needed for each individual set of pseudopotentials.

Best,

Zhenglu

---

From: Kostas Fykouras <kostasf...@gmail.com>
Sent: Monday, August 28, 2023 4:52 AM
To: BerkeleyGW Help <he...@berkeleygw.org>
Subject: Inquiry about use_velocity and use_momentum in absorption calculation

 

--
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 on the web visit https://groups.google.com/a/berkeleygw.org/d/msgid/help/bcf87f60-834f-47e2-9068-3b8722005c95n%40berkeleygw.org.

[David Strubbe]

Dear Kostas,

 

I suggest you have a look at the BerkeleyGW paper in Computer Physics Communications where we provided the formulas for the absorption spectrum and about the non-locality included with the momentum operator. In short, just \epsilon_2 is the absorption spectrum, for both velocity or momentum. The non-local part of the pseudopotential (and additional non-locality from the self-energy operator) are present in general whether you use a non-local exchange-correlation functional or not: the meaning of the term “non-local” is a bit different in DFT from the usage here. So, the momentum operator is more accurate for LDA and PBE as well.

 

Best,

David

 

David A. Strubbe (he/him/his)

Associate Professor of Physics

Chair of the Physics Graduate Group

University of California, Merced

Office: ACS 251

Phone: +1-209-228-4481

http://sail.ucmerced.edu 

 

 

From: Kostas Fykouras <kostasf...@gmail.com>
Date: Monday, August 28, 2023 at 4:52 AM
To: BerkeleyGW Help <he...@berkeleygw.org>
Subject: Inquiry about use_velocity and use_momentum in absorption calculation

--

[Kostas Fykouras]

Thank you both your answer and your time.

Zhenglu you mentioned that, I get the files absorption_b1,2,3 for free with 'use_momentum'. The b's are different polarization directions of light in this case ?

[Zhenglu Li]

That's correct, the b's corresponds to polarizations along the three reciprocal lattice vectors, respectively.

---

From: Kostas Fykouras <kostasf...@gmail.com>
Sent: Tuesday, August 29, 2023 6:59 AM
To: BerkeleyGW Help <he...@berkeleygw.org>
Subject: Re: Inquiry about use_velocity and use_momentum in absorption calculation

 

To view this discussion on the web visit https://groups.google.com/a/berkeleygw.org/d/msgid/help/8a627a27-0195-4f6b-acda-d3a3e5157430n%40berkeleygw.org.

[Renhui Liu]

Dear all, 

Thank you for your insightful discussion and clarification regarding use_velocity and use_momentum in absorption calculation.

I would like to confirm one more thing: In the case of use_velocity, does it mean that we can also calculate polarization of light in real space by setting the corresponding q shift in reciprocal space?

Looking forward to your reply.

Best,

Renhui