EPW Forum

Hello,
Recently I am trying to reproduce the scattering rates of Si as mentioned in the Fig. 8 in the new EPW paper 'EPW: Electron-phonon coupling, transport and superconducting properties using maximally localized Wannier functions'.
Do we need to set the frozen window around the Fermi level (say including some valance and conduction bands around Fermi level) to obtain the scattering rates of Si same as Fig. 8?

Thanks in advance.

Best

Shudong Wang

IMU HOHHOT, China

Dear Shudong Wang,

Yes you are correct. You need a large window around the Fermi level ( depending on the energy range you are interested in).

This means that you need to Wannierize both the valence and conduction bands.

Best,

Samuel

Dear Samuel,
Happy new year! Thank you for your kindly reply!
I am calculating the scattering rate of phosphorene as you suggested like in Si. I just want to get the scattering around
the Fermi level (4 valance+4 conduction bands). My idea is: set the frozen window(dis_froz_max) just below the highest 4 valance bands, and the dis_win_max above the lowest 4 conduction bands, and then use 8 ndnbsub to wannierization. My setting is:

nbndsub = 8
nbndskip =0

wannierize = .true.
num_iter = 600
iprint = 2
dis_win_max = 3.6
dis_froz_max= -6.5
proj(1) = 'random'

However, the spread is as large as 26 angstrom. If I set the 'dis_win_max' and 'dis_froz_min' as the default value, the spread is about 6 angstrom. But in this case the 4 nbndsub is the lowest valance bands instead of 4 valance+4 conduction bands which I am interested in.
So how could I set the window or the ndndsub if I want to obtain the scattering rate around the Fermi level?
Thank you!
Best regards,

Shudong Wang

IMU HOHHOT 010021, China

Dear sdwang,

The random initial proj(1) = 'random' does not always work.
If you know the chemical type of bonds that you should have, its usually works much better.

For more info on the way to define initial projection, you can read the user manual at http://www.wannier.org/

For the frozen windows, I guess you should set dis_froz_max above the highest 4th VBM (so inside the bandgap) and the dis_win_max above the highest 8th band (4 valences + 4 conductions bands).

Best,

Samuel

Thank you Samuel!
I finally got the spreads are about 23 angstrom with 16 nbndsub, and each spread is 1.4 angstrom.
My question is what the criteria of the convergence of the spread for each WF and for the Final spread? Could we say it converged
My output of spreads are:
Final State
WF centre and spread 1 ( 1.639838, 11.273117, 5.670956 ) 1.55348521
WF centre and spread 2 ( 1.634521, 11.076183, 4.723929 ) 1.44140174
WF centre and spread 3 ( 1.099530, 10.509948, 5.380350 ) 1.42095700
WF centre and spread 4 ( 2.124875, 10.473632, 5.381048 ) 1.41810243
WF centre and spread 5 ( 1.618750, 9.729127, 7.021076 ) 1.55406823
WF centre and spread 6 ( 1.621587, 9.927023, 7.967210 ) 1.44235149
WF centre and spread 7 ( 2.143466, 10.505086, 7.310030 ) 1.41777199
WF centre and spread 8 ( 1.115382, 10.518238, 7.310719 ) 1.41773193
WF centre and spread 9 ( 0.004732, 9.107657, 7.021136 ) 1.55406380
WF centre and spread 10 ( 0.002331, 8.909626, 7.967211 ) 1.44241385
WF centre and spread 11 ( -0.517714, 8.330075, 7.309983 ) 1.41753979
WF centre and spread 12 ( 0.510533, 8.319952, 7.310698 ) 1.41774171
WF centre and spread 13 ( -0.005927, 11.895503, 5.670223 ) 1.55410104
WF centre and spread 14 ( -0.003074, 12.093433, 4.724100 ) 1.44235535
WF centre and spread 15 ( 0.518608, 12.671603, 5.381308 ) 1.41757517
WF centre and spread 16 ( -0.509490, 12.684509, 5.380589 ) 1.41786659
Sum of centres and spreads ( 12.997949,168.024712,101.530567 ) 23.32952732

Spreads (Ang^2) Omega I = 20.163904211
================ Omega D = 0.041800586
Omega OD = 3.123894299
Final Spread (Ang^2) Omega Total = 23.329599096

Thanks!

Best,
Shudong Wang

IMU Hohhot, China

Dear Shudong and Dr. Ponce,

May I ask a quick question please?
Are you still using your previous settings?


These are different from the 11th example in Wannier90 for Si:


i) As far as I know, setting "proj(1)='Si:sp3'" does not work in EPW for Si, do you have any ideas why and how to resolve this?
ii) Also, why are you setting "dis_froz_max=-6.5"? Is it because the energies in EPW are different from Wannier90 please (e.g. reference to different Fermi levels)?
Actually, from the ouput of EPW, I found

Fermi energy coarse grid = 6.506759 eV
Fermi energy is calculated from the fine k-mesh: Ef = -5.767008 eV

Moreover, when I plot the self-energies from "linewidth.elself", all the electron energies are shifted up by about 6eV. In this file, therefore, the Fermi energy (at mid-gap) is around 12eV. Which one should be considered as the Ef? Did I make mistakes in setting these min/max values please?

Any help or explanation on these min/max settings will be greatly appreciated.

Thanks a lot,
Tai

Dear Tai,

As you can see from Shudong post, he is doing phosphorene (not Silicon). He just use the Si example as a template for his phosphorene one.

Therefore the -6.5 eV is for his phosphorene compounds. At least that's what I understood.

Otherwise, "proj(1)='Si:sp3'" should work for Silicon.

Best,

Samuel