Page 1 of 1

Dipoles using the non-local part of the pseudo potential

Posted: Tue Oct 22, 2024 2:36 pm
by Franz Fischer
Dear developers,

I know that to correctly compute the single-particle v -> c transition dipoles for extended systems it is necessary to not only use the transition dipole in k-space, but also to add the commutator of the position vector r and the nonlocal part of the pseudo potential V^NL, i.e. i[V^NL, r].
I want to understand how this is handled in practice by yambo and thus I want you to guide me to the corresponding routine in the source code where this is done.

PS: I am using the Quantum Espresso interface to yambo.

Thanks in advance!

Best,
Franz

Re: Dipoles using the non-local part of the pseudo potential

Posted: Fri Oct 25, 2024 9:04 am
by Daniele Varsano
Dear Franz,

the i[V^NL, r] commutator is taken into account unless the ns.kb_pp (produced by p2y) are renamed/deleted.
The commutator is evaluated in src/dipoles/DIPOLE_transverse.F routine.

Note that dipoles can also be calculated with the covariant approach (see DIPOLE_driver.F) (PRB 69, 085106 (2004)). In this case, the commutator is not needed.

Best,
Daniele

Re: Dipoles using the non-local part of the pseudo potential

Posted: Mon Oct 28, 2024 1:31 pm
by Franz Fischer
Cheers Daniele,

that helps a lot!

Best,
Franz