absorption spectrum

[xjxiao]

Dear all,
The absorption spectrum calculated by BSE does not rise slowly from zero. Is this an issue? What could be the possible cause of this problem?

absorption.png

Looking forward to your reply.

[Daniele Varsano]

Dear Xiangjun,

this is due to a tail of the Lorentzian broadening. This is not an issue, and you can avoid this by lowering the damping (BDmRange) in the input file.

Best,

Daniele

[xjxiao]

Dear Daniele,
Thanks a lot for your help!
I kept reducing the BDmRange, but it still seems impossible to make the imaginary part of the dielectric function zero at zero energy (the system has no excitons at zero energy). Why does this happen?
o-bse_44_qp.alpha_q1_diago_bse

Code: Select all

% BEnRange
0.000000 | 4.000000 |         eV    # [BSS] Energy range
%
% BDmRange
0.100000 | 0.100000 |         eV    # [BSS] Damping range
%
BEnSteps= 300                    # [BSS] Energy steps

o-bse_44_qp.alpha_q1_diago_bse_01

Code: Select all

% BEnRange
0.000000 | 4.000000 |         eV    # [BSS] Energy range
%
% BDmRange
0.014000 | 0.014000 |         eV    # [BSS] Damping range
%
BEnSteps= 300                    # [BSS] Energy ste

o-bse_44_qp.alpha_q1_diago_bse_02

Code: Select all

% BEnRange
0.000000 | 4.000000 |         eV    # [BSS] Energy range
%
% BDmRange
0.005000 | 0.005000 |         eV    # [BSS] Damping range
%
BEnSteps= 300                    # [BSS] Energy ste

BDmRange_contrast.png

[Daniele Varsano]

Dear Xiangjun,

as you can see from your plot this is just a tail of the Lorentzian broadening.
You have an excitation at very low energy, so it is hard to have the absorption exactly zero at zero energy.

Anyway, this is just aesthetic. For clarity, you can superimpose a vertical line at the excitation energies. You can use ypp to print the excitations:
ypp -e s -b 1

and plot the resulting o*_E_sorted file as:

>gnuplot
p "o*_E_sorted" u 1:2 w i

Best,

Daniele