TURBOMOLE Users Forum

TURBOMOLE Modules => Escf and Egrad => Topic started by: olga on April 09, 2013, 10:45:50 AM

Title: Re and Im part for dynamic polarizability
Post by: olga on April 09, 2013, 10:45:50 AM

Dear colleagues,
my current study the dependency of dynamic polarizability on the applied frequency, both for real and imaginary axes and I've faced with the problem that only the real part of resulting polarizability is printed in output but not the imaginary. I wonder if there is a special keyword to force the printing also imaginary polarizability? For sure, if no, I can perform Kramers-Kronig transform and will get the imaginary part of response function using real part. The only problem is that for precise KK-transform one need to have a function in the frequency range from 0 to infinity and in my case I've observe some convergency problems for frequencies higher that 6 eV which is far away from infinity...

Title: Re: Re and Im part for dynamic polarizability
Post by: filipp on April 12, 2013, 05:26:03 AM

Dear Olga,

You can obtain the imaginary part of the polarizability (at real frequencies) simply by computing excitation spectra.

Hope this helps,

Filipp

Title: Re: Re and Im part for dynamic polarizability
Post by: olga on April 15, 2013, 11:01:05 AM

Dear Filipp,
thanks for the idea, but I didn't completely get it. Should I just run escf module and calculate excitation spectra with for example $scfinstab rpas? In this case I have no chances to choose excitation frequencies I need to apply and also there is nothing about polarizabilities in output. Could you please clarify a bit your idea?