Drude term in RPA in metallic systems

[reilya]

Hello,

I have been trying to obtain the optical spectrum of a simple Au fcc strucutre within RPA (the input file is attached bellow), so it is a metallic system. All parameters seemed to have converged (number of bands Gvectors, k points (30) etc). My question relates to the Drude term. If I do not take the Drude term into consideration in the calculation of the absorption and later add it by hand , the absorption matches perfectly experimental ones.

If, however, I explicitly include it in the calculation as seen bellow, the contribution comes out far too small. Is there anything that I am doing wrong?

Best
Alex

#
# __ __ ________ ___ __ __ _______ ______
# /_/\/_/\ /_______/\ /__//_//_/\ /_______/\ /_____/\
# \ \ \ \ \\::: _ \ \\::\| \| \ \\::: _ \ \\:::_ \ \
# \:\_\ \ \\::(_) \ \\:. \ \\::(_) \/_\:\ \ \ \
# \::::_\/ \:: __ \ \\:.\-/\ \ \\:: _ \ \\:\ \ \ \
# \::\ \ \:.\ \ \ \\. \ \ \ \\::(_) \ \\:\_\ \ \
# \__\/ \__\/\__\/ \__\/ \__\/ \_______\/ \_____\/
#
#
# GPL Version 4.2.0 Revision 109. (Based on r.10030 h.c8f2c8
# MPI+SLK+OpenMP Build
# http://www.yambo-code.org
#
optics # [R OPT] Optics
chi # [R CHI] Dyson equation for Chi.
X_all_q_CPU= "4.14.1.1" # [PARALLEL] CPUs for each role
X_all_q_ROLEs= "v.c.k.q" # [PARALLEL] CPUs roles (q,k,c,v)
X_Threads= 1 # [OPENMP/X] Number of threads for response functions
DIP_Threads= 1 # [OPENMP/X] Number of threads for dipoles
SE_CPU= "56.1.1" # [PARALLEL] CPUs for each role
SE_ROLEs= "b.qp.q" # [PARALLEL] CPUs roles (q,qp,b)
SE_Threads= 1 # [OPENMP/GW] Number of threads for self-energy
Chimod= "Hartree" # [X] IP/Hartree/ALDA/LRC/BSfxc
NGsBlkXd= 3500 mHa # [Xd] Response block sizee block size
% QpntsRXd
1 | 1 | # [Xd] Transferred momenta
%
% BndsRnXd
1 | 300 | # [Xd] Polarization function bands
%
% EnRngeXd
0.00000 | 10.00000 | eV # [Xd] Energy range
%
% DmRngeXd
0.10000 | 0.10000 | eV # [Xd] Damping range
%
DrudeWXd = ( 8.55 , 0.070 ) eV # [Xd]
ETStpsXd= 500 # [Xd] Total Energy steps
% LongDrXd
1.000000 | 0.000000 | 0.000000 | # [Xd] [cc] Electric Field
%
% XfnQP_E
0.000000 | 1.200000 | 1.200000 | # [EXTQP Xd] E parameters (c/v) eV|adim|adim
%

[reilya]

On a small update on the calculations, I notice that the real part of the dielectric constant is in very good agreement. The problem seems to be only on the real part.

[Daniele Varsano]

Dear Alexandre,
I can't see anything wrong with your input.
Consider that what Yambo does is to add the Drude term with the empirical parameters assigned in the input to the standard RPA expression, so it should be not different to what you are doing by hand.
Have you tried to see if the Independent Particle spectrum is reasonable?

Best,
Daniele

[reilya]

Dear Daniele,

Thank you for your reply. First, let me correct my previous e-mail

On a small update on the calculations, I notice that the real part of the dielectric constant is in very good agreement. The problem seems to be only on the real part.

The real part is ok, the imaginary part seems to be off by a factor of 30 or so.

Turning to your question. I had not performed an IP calculation for a larger number of k points (I had only converged the parameters for a smaller grid), but comparing everything on a 10x10x10 grid the low frequency part for both imaginary and real terms of the dielectric function look reasonable whereas the RPA looks the same as the case with more k points.