EPW Forum

Dear Samuel,

I have a general question regarding breakdown of quasi-particle approximation in many-body perturbation theory :

In case of QP breakdown or in other words small values for the Z factor (the renormalization factor), what is the remedy to be able to include the non-adiabatic dynamical Fan contributions ?

I see in this paper http://www.nature.com/articles/ncomms11327 by Andrea that although QP picture breaks down, people keep calculating for instance an energy correction to HOMO!!!

Does it make sense to keep calculating even if the Z factor is small and hence unphysical ?


Many thanks and best wishes
LUCA

Dear LUCA,

If your Z factor is strongly different from 1, you need to be very careful indeed.

For more information on Andrea's paper, you should contact him

Best,

Samuel

Dear Samuel,

Unfortunately the first 3 authors responsible for the nature article do not respond.

Any way my fear is that in many systems the dynamical structure of the Fan self-energy can still dramatically
change the band gap correction, this is something that is consistently neglected in many recently published papers such as this one
http://pubs.acs.org/doi/abs/10.1021/acs.jpclett.6b02560

I have no idea how reliable static calculations are, as the dynamical structure of the real part of the Fan self-energy can be quite complex and non-trivial.

On the other hand if you want to include these dynamical effects, Z factor becomes far away from 1.
Here http://epjb.epj.org/articles/epjb/abs/2 ... 20105.html you see that the dynamical effects in diamond are included, leading to an increase of the
gap correction by 9%, however the Z factor of the conduction band is 0.55 !

It is difficult to understand what is correct and reliable and what is wrong !

Bests
LUCA

Well ... it's not consistently neglected.

See for example https://journals.aps.org/prb/abstract/1 ... .92.085137

If you want the dynamical structure, then you should plot the spectral function (see the paper above).

For the paper you are referencing, you need to keep in mind the computational cost: you can always do fancy theory on simple materials or simple theory on fancy materials.
Usually, both is quite difficult ...

In my option, as long as you mention explicitly what you are doing and it's the best you could do, I don't think it's an issue.

Best,
Samuel

Dear Samuel,

Before your answer, i accidentally came across your paper https://journals.aps.org/prb/abstract/1 ... .92.085137 .

And on my system i also obtain a reduction of the band gap correction at Gamma-point of about 50-60 % if i account for dynamical effects.
Your paper confirms my fear about the reliability of static electron-phonon calculations !

However, on other hand i have no idea how reliable the dynamical calculations are, since the Z factor is far from 1, in your paper about 0.6 ! on bulk diamond 0.55 (calculated by Andrea) !
The breakdown of QPA already sets on ! Moreover, the Z factor can become even smaller if you set the broadening to zero !


Sorry if i am drilling too much.
Best wishes
LUCA

Hello,

Not at all. You should drill. If you feel like there is a gap in the current literature, feel free to fill it and write a paper about it.

I'm sure it will be appreciated by the community.

Best,

Samuel