Anisotropic FBW ME equations using IR basis functions (EPW Summer School 2024 exercise)

[krishnendu]

Hello EPW Forum,

I am trying to run the Exercise 3 of the Hands-on session (Wed. 9) of the "Calculations of superconducting Tc" (EPW Summer School 2024)

I ran upto 4 th step and successfully ran the epw1.in

In the 5th step when I give the command
$ ln -s /epw-lambda/*fmt

I am getting the following:

ln: target '../epw1-lambda/selecq.fmt' is not a directory

So, I guess I am not able to create the symbolic link for all the fmt files.

Kindly let me know what should I do so that I can run the epw2.in

Thank you,
Regards,
Krishnendu

[Shashi]

Dear Krishnedu,

You might have missed the destination path while creating a soft link.

As shown in Exercise -3, the command is "ln -s ../epw1-lambda/*fmt ." You might have skipped . at the end.

Please check again.

Regards,
Shashi

[krishnendu]

Dear Shashi,

Thank you for your reply, It worked.
Please find the epw2.in script below which I made by following the steps given in the Exercise-3, 5th step. When I try to run the script I get the error:

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine epw_readin (1):
Bad line in namelist &inputepw: "" (error could be in the previous line)
%%%%%%%%%%%%%%%%%%%%%%%

Kindly let me know what I should do so that I can run the epw2.in and get the superconducting gap for Nb for the given temperatures.

Thank you,
Regards,
Krishnendu
--------------------------------------------
--
&inputepw
prefix = 'nb',
outdir = './'
dvscf_dir = '../phonons-krishnendu/save' ! directory where .dyn, .dvscf and prefix.phsave/patterns.xx.yy
! files obtained from phonon calculation are stored

ep_coupling = .false. ! run e-ph coupling calculation

elph = .false. ! calculate e-ph coefficients

epwwrite = .false. ! write e-ph matrices in the Wann representation

epwread = .true. ! read e-ph matrices from the 'prefix.epmatwp' file


etf_mem = 0 ! less I/O but more memory is required


vme = 'dipole'

nbndsub = 9

bands_skipped = 'exclude_bands = 1:4'

wannierize = .false.

num_iter = 0

dis_win_min = 10

dis_froz_min = 11

dis_froz_max= 28

proj(1) = 'Nb:s'

proj(2) = 'Nb:p'

proj(3) = 'Nb:d'

iverbosity = 0 ! if you want to get .frmsf files, change this to 2.


fsthick = 0.5 ! Fermi window thickness [eV]

degaussw = 0.1 ! smearing in the energy-conserving delta functions in [eV]

degaussq = 0.5 ! smearing for sum over q in the e-ph coupling in [meV]



ephwrite = .false. ! write ephmatXX, egnv, freq, and ikmap files in prefix.ephmat directory

eliashberg = .true. ! calculate Eliashberg spectral function

laniso = .true.

limag = .true.

fbw = .true.

muchem = .true. ! update the chemical potential at every temperature

gridsamp = 2 ! If gridsamp = 2, use sparse-ir sampling

nsiter = 500 ! convergence threshold for solving ME eqs. on imaginary axis

filirobj = '../../irobjs/ir_nlambda6_ndigit8.dat'

! the file containing the sparse samping points and the IR functions.

broyden_beta = -0.7 ! If this factor is negative, use the linear mixing instead Broyden mixing.

conv_thr_iaxis = 1.0d-3 ! convergence threshold for solving ME eqs. on imaginary axis

wscut = 1.0

temps = 0.2 6.0 10.0 12.0 13.0 13.5 ! temperature list

! nsiter = 500 ! number of self-consistent iterations when solving ME eqs.

muc = 0.24 ! effective Coulomb potential used in ME eqs.





nk1=4
nk2= 4 ! dimensions of the coarse electronic grid
nk3= 4


nq1 =4
nq2=4 ! dimensions of the coarse phonon grid
nq3=4

mp_mesh_k = .true. ! use irreduciable electronic fine mesh
nkf1 = 20
nkf2 = 20 ! dimensions of the fine electronic grid
nkf3 = 20


nqf1 = 20
nqf2 = 20 ! dimensions of the fine phonon grid
nqf3 = 20

/

[Shashi]

Dear Krishnendu,

I think you might not have the "ir_nlambda6_ndigit8.dat" file in the irobjs folder. I could see that only the file included in irobjs is "
ir_nlambda5_ndigit6.dat". If you want to generate "ir_nlambda6_ndigit8.dat", you need to generate it using the command

make irobjs

as mentioned in q-e/EPWirobjs/README.md.

Please let me know if it resolves your issue.

Regards,
Shashi

[krishnendu]

Dear Shashi,

Thank you for your reply. Yes I already had the "ir_nlambda6_ndigit8.dat" file in the irobjs folder. I obtained it from the docs.epw-code.org/doc/Inputs.html: below the gridsamp.

So, I guess the problem is somewhere else. Kindly help me.

Thank you,
Best regards,
Krishnendu

[Shashi]

Hi Krishnendu,

I could not reproduce the error. I tested the tutorial, and it ran successfully without any issues. I also tested your EPW input for the 5th step, and it worked fine. I am not sure what might be causing the problem on your end.

Are you using the latest version of EPW? If not, I would suggest updating, as the writing .ephmat files have changed slightly.

Regards,
Shashi

[krishnendu]

Dear Shashi,

Thank you for your reply and also for testing the inputs.
I am using EPW-v.5.8.1 which is compatible with my existing Quantum Espresso v.7.3.1

So, would you recommend installing EPW-6.0.0 ?

Thank you,
Best regards,
Krishnendu

[Shashi]

Hi Krishnedu,

Please use EPW-6.0.0

wget https://gitlab.com/epw/q-e/-/archive/EP ... 6.0.tar.gz

Regards,
Shashi

[krishnendu]

Hi Shashi,

I am able to run the 5 th step in Exercise 3. Please find below the graph by plotting nb.imag_aniso_gap0_013.50

nb.imag_aniso_gap0_013.50

nb.imag_aniso_gap0_013.50_pic.png (26.72 KiB) Viewed 18899 times

I am little worried that the graph is not matching with that given in the tutorial (Figure 15, of Hands-on session of Calculation of superconducting Tc, EPW School 2024).

There the T[K] + dist. (scaled to 1) is extended till 16 K. While in the attached plot I got it extended till 14.5 K.

And the delta (meV) is between 0.5 and 0.625 in the Figure 15 of tutorial, where I got it between 0.675 and 0.85.

Did I do something wrong ?

Thanking you,
Regards,
Krishnendu

[krishnendu]

Hi Shashi,

In continuation to my earlier post today, I found that in the tutorial in Figure 15, mu-star-c is mentioned to be 0.25. So, I again ran epw2.in with muc=0.25
The plot is as follows:

nb.imag_aniso_gap0_013.50-mustar-0.25.png (47.43 KiB) Viewed 18850 times

So, there is better agreement with the delta (meV) values.
However, the temperature range is not matching. Is that something to worry about ?

I put the values of different input variables exactly same as mentioned in the Exercise 3 of the tutorial that we are discussing.
And I used EPW v.6.0

Thank you,
Regards,
Krishnendu