[QE-users] How much effect of el_ph_sigma and el_ph_nsigma flags on the critical temperature Tc?

[Zhibin Gao]

Dear QE users and developers,

Recently, I was trapped by a question about the critical temperature Tc for
a very long time.
In one calculation, I can successfully get the lamda (electron-phonon
coupling) and Tc like:
lambda        omega_log          T_c
   0.15006       335.625              0.000
   0.23881       279.676              0.007
   0.37627       249.217              0.724
   0.55241       226.383              3.971
   0.74555       208.449              8.401
   0.93962       193.607             12.205
   1.12670       180.939             14.944
   ...                 ...                       ....

But when I add two flags:
el_ph_sigma=0.005,
el_ph_nsigma=10,

into the system. elph.in file. I obtained the result:
lambda        omega_log          T_c
   0.05937      1332.783          *********
   0.06470      1324.945          *********
   0.06835      1309.450          *********
   ...                 ...                       ....

I compare the phonon dispersion, both are identical and without imaginary
frequency. I also check

the elph.inp_lambda.* file, no imaginary frequency also. I use the pwscf
6.3 version.


Both calculations are identical except for these two parameters
(without/with).* Lamda changes significantly.*

*I am wondering how much effect of these two flags in pwscf on the
superconductivity?* I checked the manual,

but the related info is limited. In the latest version of 6.4, the official
website has deleted these two flags.

Why these two parameters would lead to the infinite Tc?

Thanks a lot!


Best wishes,

Zhibin

Tongji University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190808/91ba5f41/attachment.html>