ABINIT Discussion Forums

Here are 3 questions :

(1) We need all the wave functions of a converged plane wave abinit calculation for crystals of small unit cells. According to the link

https://docs.abinit.org/guide/abinit/#wfkfile

early versions of abinit wrote the wave functions to a file in a simple way as …. cg(ii+index),ii=1,2*npw*nspinor) ...
It should then be possible to read this unformatted file and reconstruct cg as a function of the indices of kpoints, bands, and g or
as coefficient(k_index,g_index, ,band_index)

Question: Where can we find an old non parallel version of abinit such as 4.0 where one may read the wave function correctly in such a simple way ?

(2) To use the information of (1) after having found an old version of abinit we still need to know the
correspondence between k_index and (k1,k2,k3) and between g_index and (g1,g2,g3).

Question: Which is this correspondence ?

(3) We want to judge the success of our exportation of wave function from abinit by the relation
c(k,g,b) = c*(-k,-g,b) since the c(k,g,b) are – presumably – Fourier transforms of psi(k,r,b) which should satisfy time reversal conjugation.

Question: Should the relation c(k,g,b) = c*(-k,-g,b) be satisfied in abinit ?

Thank you in advance for answering the above 3 questions !

***********************************************

If there is a simpler and more appropriate way to export wave functions from more recent versions of abinit, please let me know.

Thank you in advance for your feed back and assistance !

Best regards,

Dietrich Foerster 

Bordeaux