Page 1 of 1

triplets calculation in BSE

Posted: Mon Aug 05, 2013 10:52 pm
by arlonne
Hello all,

The BSE in the TDA is H = E + cV - W, where V and W are the bare Coulomb interactions and the statically screened Coulomb interactions, respectively. Therefore, in principle, we can calculate the singlets exciton energy by setting c = 2, and the triplets exciton energy by setting c = 0. In the manual of Yambo, I don't find how to set the parameter c. So, Yambo calculate the singlet exicton energy in default. Thereby, my question is can I calculate the triplet exciton energy by dropping V (V=0)? and which subroutines should I modified? Although it is interesting to find out the way by myself, asking question is the fastest way to get the answer.

Thanks you,
arlonne

Re: triplets calculation in BSE

Posted: Mon Aug 12, 2013 5:18 pm
by Davide Sangalli
Dear arlonne,
you can:
1) Perform a ground state calculation with spin and then run yambo. In this case the code will take care of the diagonalization in spin space and you should obtain both signlets and triplets poles.
2) set BSENGexx=1 (http://www.yambo-code.org/input_file/va ... ENGexx.php) in the input file. Indeed this means to use only the g=0 vector for the V term, which is zero by definition as yambo is using the "overlined" V. In this case I'm not sure the oscillator strenghts are properly defined.

In any case pay attention that in the BSE approach the dielectric function is computed via the "overlined" polarization, thus you get the poles of this quantity. Check if this is what you need. (http://www.yambo-code.org/theory/docs/doc_BSS.php).

Best regards,
Davide

Re: triplets calculation in BSE

Posted: Mon Aug 12, 2013 8:39 pm
by arlonne
Dear Davide,
It seems helpful. I'll try.

Thanks,
arlonne