The ABINIT Users Mailing List ( CLOSED )

[zheng@chips.ncsu.edu]

Hello, Dear Abinitors,

In the second tutorial on GW, when we calculate the GW correction to the band 
structure of Aluminum and do the numerical frequency convolution, the 
integration path is deformed from the real axis to the imaginary axis. We are 
told that the method is described in the paper Lebegue[2003]. 

I checked that paper and I still have something that I do not understand. In 
that paper, the integration along the real axis is replaced by the 
contribution of the integration along the imaginary axis and the contribution 
from the residues. So, there should be no contribution from the real axis 
now. But in the ABINIT code, we have two contributions from both the real 
axis and the imaginary axis specified by two parameters "nfreqim" and 
"nfreqre". 

So, how to understand this, please? Could you give more technical details 
about this integration?

Thank you very much!!

Sincerely,
Zheng