How to get real part of Sigma

GW, Bethe-Salpeter …

Moderators: maryam.azizi, bruneval

Locked
dgontier
Posts: 8
Joined: Fri Jun 08, 2012 4:31 pm

How to get real part of Sigma

Post by dgontier » Tue Mar 11, 2014 2:36 pm

Dear GW team,

I am trying to perform some calculations with Abinit to check some mathematical results.
For instance, I would like to plot the function \omega \to Re ( \Sigma_{ii} (\omega) ) for omega in the real axis.

However, if I am correct, the only thing I can get related to those quantities is the evaluation of the spectral function (in the file out.SIG).
However :
1/ I don't understand everything in the .SIG file.
I have something like
# k = 0.000000 0.000000 0.000000
# b = 4 5
-0.54423E+02 -0.18931E+02 0.32133E-08 0.36350E-12 -0.57501E+01 0.36375E+00 0.25944E-04
-0.53876E+02 -0.19173E+02 0.00000E+00 0.00000E+00 -0.56379E+01 0.00000E+00 0.00000E+00
-0.53329E+02 -0.19337E+02 0.00000E+00 0.00000E+00 -0.57016E+01 0.00000E+00 0.00000E+00
...

Ok, so the first number is the value of omega in the real axis, and for each band, I have 3 values... what are they exactly ? I can not find any documentation about that...

2/ I want to use the ppmodel=4 (Engel and Farid), which gives real-valued Sigma_ii(\omega), so that (I guess) the spectral function is highly degenerate, and I won't be able to work with it. How can I get just the real part of the self-energy ?

3/ How did you get the very nice picture in the 4th chapter of this page : http://www.abinit.org/documentation/hel ... y_mbt.html

Thanks a lot for your answer,

David Gontier

User avatar
gmatteo
Posts: 291
Joined: Sun Aug 16, 2009 5:40 pm

Re: How to get real part of Sigma

Post by gmatteo » Wed Apr 02, 2014 12:05 am

Ok, so the first number is the value of omega in the real axis, and for each band, I have 3 values... what are they exactly ? I can not find any documentation about that...


Real_Sigma, Imag_Sigma, Spectral_function

see 70_gw/m_sigma_results.F90 (routine write_sigma_results, file unit unt_sig)

2/ I want to use the ppmodel=4 (Engel and Farid), which gives real-valued Sigma_ii(\omega), so that (I guess) the spectral function is highly degenerate, and I won't be able to work with it. How can I get just the real part of the self-energy ?


Computing the spectral function with a plasmon-pole model would not give reasonable results, one should use the contour-deformation technique for this (expensive) calculation.
The real part of the self-energy is reported in the _SIG file, If I remember correctly the second GW tutorial discusses how to get it.


3/ How did you get the very nice picture in the 4th chapter of this page : http://www.abinit.org/documentation/hel ... y_mbt.html


The pictures was produced by Martin Stankovski If I rememeber well, the results was obtained with a code he wrote for the homogeneous electron-gas

Locked