Re: Wannier 90 interface in 4.5.2
Posted: Thu Nov 20, 2025 8:55 am
Dear Deyang,
but it is not entirely clear to me what quantity you are trying to compute.
As you know, solving the BSE provides the electron–hole amplitudes A^lambda_cvk that define each exciton. From these amplitudes, you can construct the excitonic wavefunction as a post-processing step. This is a two-particle wavefunction given by:
\Psi_exc^\lambda(re,rh) = \Sum_cvk A^lambda_cvk \phi_vk(r_v) \phi^*_ck(r_e)
where \phi are the KS states.
The A^lambda_cvk amplitudes are stored in the ndb.BS_diago database and they can be accessed using yambopy.
I hope this clarifies things
Best,
Daniele
but it is not entirely clear to me what quantity you are trying to compute.
As you know, solving the BSE provides the electron–hole amplitudes A^lambda_cvk that define each exciton. From these amplitudes, you can construct the excitonic wavefunction as a post-processing step. This is a two-particle wavefunction given by:
\Psi_exc^\lambda(re,rh) = \Sum_cvk A^lambda_cvk \phi_vk(r_v) \phi^*_ck(r_e)
where \phi are the KS states.
The A^lambda_cvk amplitudes are stored in the ndb.BS_diago database and they can be accessed using yambopy.
I hope this clarifies things
Best,
Daniele