Electron relaxation time calculation

[jir520]

Dear EPW users

I am new to EPW code and I was trying to calculate the electron relaxation time for Al. I referred the Graphene example in https://arxiv.org/pdf/1005.4418v1.pdf (EPW paper by Jesse Noffsinger) related with this code. I have few questions regarding Graphene example in it.

01) Is the figure 05 (Imaginary part of Self E. Vs Electron Energy) in the paper drawn with the data in linewidth.elself output file? If so, how its x-axis (electron energy) comes.

02). Why the Fermi energy is set to zero in the graph instead of the actual value for Graphene ?

03) I am trying to get electron relaxation time for my aluminum by following the same way as in the paper and taking the tau= ?/Ei (Ei= the imaginary part of the self energy at fermi energy from the above graph)
Is that correct or is there an alternative way to get it?

The Graphene example is not in the EPW tutorials now. Do we have access to that tutorial somewhere else? (Because we do not have relaxation time calculation in the current tutorial examples )

I appreciate your comments very much.

Thank you
Jayangani

[carla.verdi]

Dear Jayangani

The graphene figure you mentioned was not drawn directly with the data in linewidth.elself because the previous EPW version did not create such file; instead, the electron self-energy was only written at the end of epw.out. Now you can extract the data you wish to plot either from linewidth.elself or from the output file.
The Fermi energy is set to zero just because the electron energies are usually referred to the Fermi level.
The electron relaxation time has a factor of two, tau=hbar/(2ImS).
I am afraid the graphene example is not included, but you can extract the relaxation times from any other electron self-energy example.

Best
Carla

[Khaled]

Dear developers,

First of all, I'd like to thank you for this nice code. My question is more related to physics and I will be appreciated if you help me. I am calculating the mobility of electron and hole, which is related to the relaxation time. my questions is :

1) Is physically right If I calculate the relaxation time that corresponding to the minimum conduction band (electron relaxation time for electron mobility) using EPW. and the relaxation time that corresponding to the maximum valence band (hole relaxation time for hole mobility). I am confused because I didn't read about the hole relaxation time. Thank you in advance.


Best,

[sponce]

Dear Khaled,

Yes, both electron and hole relaxation time make sense. Actually, one usually speaks about carrier relaxation time (could be electron or hole).

Just to let you know, I'm working on mobility inside EPW by solving the Boltzmann transport equation.

I'm not sure when the code will be ready for public release but hopefully in a couple of months.
If you can wait until then, you should have a user-friendly and optimised code to do carrier mobilities.

Best,

Samuel.

[Khaled]

Dear Dr. Samuel,

Thanks a lot for your respond. I will be of course waiting for the mobility program. I have few questions that related to some input parameters
I couldn't understand them (not clear in the documentation for me) even when I read about them.

1) Proj (1) = ' random' . what I know, It is related to the orbital of the specified materils. How can I get this information or from which source ?

2) nbndskip I did change the value of this parameter, but nothing change in the output results. What this parameter doing ?.

3) degaussq . Is this parameter related to include the spin orbital coupling. You are including it for Pb tutorial. What it is the reason ?.

Thank you in advance.

Best regards,

[carla.verdi]

Dear Khaled,

To answer your questions:
1) For example you can plot the PDOS using projwfc.x of QE. Also, you can refer to the wannier90 user guide for an explanation of the initial projections: http://wannier.org/doc/user_guide.pdf

2) nbndskip needs to be used only if you are skipping some of the lowest valence bands when you wannierize.

3) degaussq is the smearing parameter for the delta in the calculation of a2F; it is not related to the spin-orbit coupling. Don't worry if it is explicitely included only in the SO case of Pb - that value is the default value indeed.

Best
Carla