Pb superconducting properties calculation (tutorial at 2018 ICTP school)

[blessing1223]

Dear all,

I have tried to calculate on Pb superconducting properties according to tutorial of Prof. Roxana given at 2018 ICTP school.
I have communicated with Prof. Roxana several times and I posted e-mail contents below.
However, problem is still ongoing. Please give some suggestion and comments about this.

Dr. Sungwoo Lee

-----------------------------------------------------------------------------------------------------------------
Sent on Thursday, Apr 30 11:33 PM

Dear Sungwoo,

As I mentioned in my previous email please post all questions regarding EPW on the EPW website https://forum.epw-code.org/.

I am not sure I understand your question. A description of all EPW input variables can be found at http://epw.org.uk/Documentation/Inputs#nkf1

Variables: nk1, nk2, nk3

Type: INTEGER
Default: 0
Description: Dimensions of the coarse electronic grid, corresponds
to the nscf calculation and wfs in the outdir.

Back to Top

Variables: nkf1, nkf2, nqf3

Type: INTEGER
Default: 0
Description: Dimensions of the fine electron grid, if filkf
is not given.

Back to Top

Variables: nq1, nq2, nq3

Type: INTEGER
Default: 0
Description: Dimensions of the coarse phonon grid, corresponds
to the nqs list.

Back to Top

Variables: nqf1, nqf2, nqf3

Type: INTEGER
Default: 0
Description: Dimensions of the fine phonon grid, if filqf
is not given.

Best,
    Roxana

-----------------------------------------------------------------------------------------------------------------
Sent on Thursday, Apr 30 11:17 PM

Dear Roxana,

Are you healthy? I hope the problem of COVID-19 in the United States will quickly settle down.

I did several calculations using the input file you sent me last time.
There are parameters for nkf and nqf in the input file, which seems to probably be corrected with nk and nq. Am I right?

So, I changed nkf and nqf to nk and nq, I entered the corresponding values according to my nscf calculation and ph.x calculation. (nk1 = nk2 = nk3 = 8, nq1 = nq2 = nq3 = 8)
Also, the q-point list at the end of the input file was input accordingly. (28 points as same as prefix.dyn0)

However at this time another error occurred. Please look inside attached epw output file.
I found that this error occurred at efermig.f90 of pwscf code but I have no idea to solve this problem.

Have you ever seen this kind of error? Please let me know any idea might be helpful.
I'm sorry I seem to be taking your time with a simple calculation.

Best,

Sungwoo Lee, Ph.D.

attachement: epw_rand.out
-----------------------------------------------------------------------------------------------------------------
Sent on Saturday, Apr 18 12:54 PM

Dear Roxana

I appreciate your concern and input file.
I’m going to do it and I will post my question and your answer on EPW forum.
Have a nice weekend!

Many thanks,

Sungwoo Lee, Ph.D.

-----------------------------------------------------------------------------------------------------------------
Sent on Friday, Apr 17 11:58 PM

Hi,

To run the isotropic Eliashberg calculations using a random grid you need to perform two steps.

1) In the first step you calculate the Eliashberg spectral function (using an input file epw_rand in  attachment)

At the end of this run you should get a file name.a2f.01 that can be used in the second step.

2) In the second step you can solve the isotropic ME equations starting from a file that contains the Eliashberg spectral function (below the file name is lead.a2f_iso).

  fila2f      = 'lead.a2f_iso'
  eliashberg  = .true.

I hope this helps.

Please in the future post your questions on the EPW forum so others can benefit from our discussions and even answer your questions. The forum is in the process of being moved and it should be fully functional pretty soon.

Best,
   Roxana

attachement: epw_rand
-----------------------------------------------------------------------------------------------------------------
Sent on Monday, Apr 13 10:35 PM

Dear Roxana

I am trying to do calculation again. However I have a question about method described in you PRB paper.

In the last step of calculations(=EPW calculations to solve Eliashberg eq.), you said that random grid was used for the Pb case. How can I switch on this random grid option in EPW input file?
Because I think that input options described in the tutorial are for the uniform Γ-centered grid calculation.
(I guess these are related to mp_mesh_k and mp_mesh_q parameters)

Best,

Sungwoo Lee, Ph.D.

-----------------------------------------------------------------------------------------------------------------
Sent on Friday, Apr 10 5:37 PM

Thanks for your reply and concern.

I will try again with conditions described in you paper.

Best,

Sungwoo Lee, Ph.D.

-----------------------------------------------------------------------------------------------------------------
Sent on Tuesday, Apr 7 9:51 PM

Dear Sungwoo,

Yes, the analytic continuation didn't work because the Pade approximation failed (the gap from the Pade is used as an initial guess in the iterative solutions for analytic continuation). You can  play with the wscut parameter that defines the number of Pade approximants.But keep in mind that all parameters used in the tutorial are not converged, so I would rather advise you changing them to the ones used in my PRB paper on Migdal-Eliashberg formalism.

I hope this helps.

Thank you for asking, my family and myself are fine. I hope you stay healthy as well.

Best wishes,
    Roxana

-----------------------------------------------------------------------------------------------------------------
Sent on Monday, Apr 6 5:39 PM

Dear Prof. Margine

Hello, I'm Sungwoo Lee, a postdoctoral researcher at Seoul National University in Korea.

Nowdays, I am interested in superconductors, so I want to calculate the superconducting properties of certain materials using EPW.

So I studied alone and looked reference materials. I found the reference materials of the electron-phonon physics school held at ICTP in 2018. Thanks for your nice lecture and tutorial materials.

I tried to calculate the superconducting property of Pb according to the tutorial you gave (I attached this tutorial pdf file below). However, despite following the instructions in the tutorial, there was a problem with the calculation.

I attached output of EPW calculation. (epw1.txt) You can find error message at the end of output file, which is error about broyden mixing. However, in my short opinion, several NaN printed at Pade approximate of isotropic Eliashber eq. and Analytic continuation of isotropic Eliashberg eq. parts are also problem.

I have tried to find problem by looking into the code in detail, but it seems a little difficult to solve to me.

Any advice would be appreciated.
I hope you and your family are all healthy because there are many difficulties caused by the coronavirus in the U.S.

Best,

Sungwoo Lee, Ph.D.

[roxana]

Hi,

As shown in all examples on EPW there are two sets of grids for both electrons and phonons:
1) the starting coarse grids that have to be uniform and consistent with nscf and phonon calculations and
k- grid should be the same as the nscf calculation
nk1 = XXX
nk2 = XXX
nk3 = XXX

q-grid should be the same as the mesh on which the phonons were calculated with QE
nq1 = XXX
nq2 = XXX
nq3 = XXX

2) the final fine grids that can be either uniform or random (you cannot set both ways in the same run)

a) for uniform grids you need to use
nkf1 = XXX
nkf2 = XXX
nkf3 = XXX

nqf1 = XXX
nqf2 = XXX
nqf3 = XXX
If you use uniform grids for transport or superconductivity calculations you also need to set mp_mesh_k = true and the k and q grids need to be multiples of each other.

or

b) for random grids you need to use
rand_k = .true.
rand_nk = XXX
rand_q = .true.
rand_nq = XXX

I hope this clarifies your misunderstandings.

Best,
Roxana

[hlee]

Dear Sungwoo Lee:

Are you referring to the EPW v5.2 included in the official Quantum Espresso (QE) v6.5 which can be downloaded at
https://github.com/QEF/q-e/releases or https://gitlab.com/QEF/q-e/-/tags ?

Or, are you referring to the development version of EPW which can be downloaded at https://gitlab.com/QEF/q-e ?

For the former, you should also include "nbndskip=5" in your input since you added the input of "bands_skipped = 'exclude_bands = 1-5" in epw_rand.

For the latter, EPW can automatically calculate nbndskip from the input of "bands_skipped = 'exclude_bands = 1-5" and you don't need to add "nbndskip=5".

In your output, there is no statement saying "Skipping the first 5 bands:" even if you used the input of "bands_skipped = 'exclude_bands = 1-5"
Therefore, I guess that you are using the EPW v5.2 included in the official Quantum Espresso (QE) v6.5
In this case, if you don't include both nbndskip and "bands_skipped = 'exclude_bands = ..." in the EPW input when you want some bands to be excluded in Wannierization step, nelec (number of electrons) is wrong, leading to the error in efermig.

Sincerely,

H. Lee

[blessing1223]

Dear H. Lee

Thank you very much for you comment.
Calculation was finished successfully and lead.a2f.01 file was created.

Now I am trying to solve the isotropic ME equations starting from a file that contains the Eliashberg spectral function (lead.a2f.01)
However it seems that calculation was stopped where ME equations part.

my input is here,

Code: Select all

--
&inputepw
  prefix      = 'lead'
  amass(1)    = 207.2
  outdir      = './'

  elph        = .true.
  kmaps       = .false.
  epbwrite    = .true.
  epbread     = .false.

  epwwrite    = .true.
  epwread     = .false.

  vme         = .false.

  nbndsub     =  4
  bands_skipped = 'exclude_bands = 1-5'
  nbndskip    =  5

  wannierize  = .true.
  num_iter    = 300
  dis_win_max = 21
  dis_win_min = -3
  dis_froz_min= -3
  dis_froz_max= 13.5
  proj(1)     = 'Pb:sp3'

  wdata(1) = 'bands_plot = .true.'
  wdata(2) = 'begin kpoint_path'
  wdata(3) = 'G 0.00 0.00 0.00 X 0.00 0.50 0.50'
  wdata(4) = 'X 0.00 0.50 0.50 W 0.25 0.50 0.75'
  wdata(5) = 'W 0.25 0.50 0.75 L 0.50 0.50 0.50'
  wdata(6) = 'L 0.50 0.50 0.50 K 0.375 0.375 0.75'
  wdata(7) = 'K 0.375 0.375 0.75 G 0.00 0.00 0.00'
  wdata(8) = 'G 0.00 0.00 0.00 L 0.50 0.50 0.50'
  wdata(9) = 'end kpoint_path'
  wdata(10) = 'bands_plot_format = gnuplot'

  iverbosity  = 0

  phonselfen  = .true.
  a2f         = .false.
  delta_approx = .true. 

  fsthick     = 6 ! eV
  eptemp      = 300 ! K
  degaussw    = 0.1  ! eV
  degaussq    = 0.05 ! meV

  dvscf_dir   = './save'

  nk1        = 8
  nk2        = 8
  nk3        = 8

  nq1        = 8
  nq2        = 8
  nq3        = 8

  rand_k = .true.
  rand_nk = 40000

  rand_q = .true. 
  rand_nq = 10000

  fila2f = 'lead.a2f.01'
  eliashberg = .true.
 /
29 cartesian
   0.000000000000000E+00   0.000000000000000E+00   0.000000000000000E+00
  -0.125000000000000E+00   0.125000000000000E+00  -0.125000000000000E+00
  -0.250000000000000E+00   0.250000000000000E+00  -0.250000000000000E+00
  -0.375000000000000E+00   0.375000000000000E+00  -0.375000000000000E+00
   0.500000000000000E+00  -0.500000000000000E+00   0.500000000000000E+00
   0.000000000000000E+00   0.250000000000000E+00   0.000000000000000E+00
  -0.125000000000000E+00   0.375000000000000E+00  -0.125000000000000E+00
  -0.250000000000000E+00   0.500000000000000E+00  -0.250000000000000E+00
   0.625000000000000E+00  -0.375000000000000E+00   0.625000000000000E+00
   0.500000000000000E+00  -0.250000000000000E+00   0.500000000000000E+00
   0.375000000000000E+00  -0.125000000000000E+00   0.375000000000000E+00
   0.250000000000000E+00   0.000000000000000E+00   0.250000000000000E+00
   0.000000000000000E+00   0.500000000000000E+00   0.000000000000000E+00
  -0.125000000000000E+00   0.625000000000000E+00  -0.125000000000000E+00
   0.750000000000000E+00  -0.250000000000000E+00   0.750000000000000E+00
   0.625000000000000E+00  -0.125000000000000E+00   0.625000000000000E+00
   0.500000000000000E+00   0.000000000000000E+00   0.500000000000000E+00
   0.000000000000000E+00   0.750000000000000E+00   0.000000000000000E+00
   0.875000000000000E+00  -0.125000000000000E+00   0.875000000000000E+00
   0.750000000000000E+00   0.000000000000000E+00   0.750000000000000E+00
   0.000000000000000E+00  -0.100000000000000E+01   0.000000000000000E+00
  -0.250000000000000E+00   0.500000000000000E+00   0.000000000000000E+00
   0.625000000000000E+00  -0.375000000000000E+00   0.875000000000000E+00
   0.500000000000000E+00  -0.250000000000000E+00   0.750000000000000E+00
   0.750000000000000E+00  -0.250000000000000E+00   0.100000000000000E+01
   0.625000000000000E+00  -0.125000000000000E+00   0.875000000000000E+00
   0.500000000000000E+00   0.000000000000000E+00   0.750000000000000E+00
  -0.250000000000000E+00  -0.100000000000000E+01   0.000000000000000E+00
  -0.500000000000000E+00  -0.100000000000000E+01   0.000000000000000E+00

and the very last part of output is

Code: Select all

     Number of (k,k+q) pairs on the Fermi surface:    39423 out of    40000


     ismear =     1 iq =   10000 coord.:   0.41985  0.39185  0.10653 wt:   0.00010
     -------------------------------------------------------------------
     lambda___(   1 )=       0.285697   gamma___=       0.006604 meV   omega=      5.7497 meV
     lambda_tr(   1 )=       0.254872   gamma_tr=       0.005892 meV   omega=      5.7497 meV
     lambda___(   2 )=       0.364144   gamma___=       0.009734 meV   omega=      6.1829 meV
     lambda_tr(   2 )=       0.368927   gamma_tr=       0.009862 meV   omega=      6.1829 meV
     lambda___(   3 )=       0.221585   gamma___=       0.017951 meV   omega=     10.7634 meV
     lambda_tr(   3 )=       0.216007   gamma_tr=       0.017499 meV   omega=     10.7634 meV
     lambda___( tot )=       0.871426
     lambda_tr( tot )=       0.839806
     -------------------------------------------------------------------


     Number of (k,k+q) pairs on the Fermi surface:    39380 out of    40000

     ===================================================================
     Memory usage:  VmHWM =       681Mb
                   VmPeak =      1049Mb
     ===================================================================

Now I am looking into the code also. Please comment and give some ideas anybody.

Best,
Sungwoo Lee

[hlee]

Dear all:

I am sorry, but there was a typo in my previous post:
(X) "bands_skipped = 'exclude_bands = 1-5"
(O) bands_skipped = 'exclude_bands = 1-5'

Sincerely,

H. Lee

[roxana]

Hi,

If you plan only to run isotropic ME starting from a file you need to use something like in the following input

--
&inputepw
prefix = 'lead'
amass(1) = 207.2
outdir = './'

ep_coupling = .false.
elph = .false.
kmaps = .false.
epbwrite = .true.
epbread = .false.

epwwrite = .true.
epwread = .false.

vme = .false.

nbndsub = 4
bands_skipped = 'exclude_bands = 1-5'

wannierize = .false.
num_iter = 300
dis_win_max = 21
dis_win_min = -3
dis_froz_min= -3
dis_froz_max= 13.5
proj(1) = 'Pb:sp3'

wdata(1) = 'bands_plot = .true.'
wdata(2) = 'begin kpoint_path'
wdata(3) = 'G 0.00 0.00 0.00 X 0.00 0.50 0.50'
wdata(4) = 'X 0.00 0.50 0.50 W 0.25 0.50 0.75'
wdata(5) = 'W 0.25 0.50 0.75 L 0.50 0.50 0.50'
wdata(6) = 'L 0.50 0.50 0.50 K 0.375 0.375 0.75'
wdata(7) = 'K 0.375 0.375 0.75 G 0.00 0.00 0.00'
wdata(8) = 'G 0.00 0.00 0.00 L 0.50 0.50 0.50'
wdata(9) = 'end kpoint_path'
wdata(10) = 'bands_plot_format = gnuplot'

iverbosity = 0

fsthick = 6 ! eV
eptemp = 300 ! K
degaussw = 0.1 ! eV
degaussq = 0.05 ! meV

dvscf_dir = './save'

fila2f = 'lead.a2f.01'
eliashberg = .true.

wscut = 1

liso = .true.
limag = .true.
nstemp = 1
tempsmin = 5
tempsmax = 6

muc = 0.1

nk1 = 8
nk2 = 8
nk3 = 8

nq1 = 8
nq2 = 8
nq3 = 8

nkf1 = 8
nkf2 = 8
nkf3 = 8

nqf1 = 8
nqf2 = 8
nqf3 = 8
/
29 cartesian
0.000000000000000E+00 0.000000000000000E+00 0.000000000000000E+00
-0.125000000000000E+00 0.125000000000000E+00 -0.125000000000000E+00
-0.250000000000000E+00 0.250000000000000E+00 -0.250000000000000E+00
-0.375000000000000E+00 0.375000000000000E+00 -0.375000000000000E+00
0.500000000000000E+00 -0.500000000000000E+00 0.500000000000000E+00
0.000000000000000E+00 0.250000000000000E+00 0.000000000000000E+00
-0.125000000000000E+00 0.375000000000000E+00 -0.125000000000000E+00
-0.250000000000000E+00 0.500000000000000E+00 -0.250000000000000E+00
0.625000000000000E+00 -0.375000000000000E+00 0.625000000000000E+00
0.500000000000000E+00 -0.250000000000000E+00 0.500000000000000E+00
0.375000000000000E+00 -0.125000000000000E+00 0.375000000000000E+00
0.250000000000000E+00 0.000000000000000E+00 0.250000000000000E+00
0.000000000000000E+00 0.500000000000000E+00 0.000000000000000E+00
-0.125000000000000E+00 0.625000000000000E+00 -0.125000000000000E+00
0.750000000000000E+00 -0.250000000000000E+00 0.750000000000000E+00
0.625000000000000E+00 -0.125000000000000E+00 0.625000000000000E+00
0.500000000000000E+00 0.000000000000000E+00 0.500000000000000E+00
0.000000000000000E+00 0.750000000000000E+00 0.000000000000000E+00
0.875000000000000E+00 -0.125000000000000E+00 0.875000000000000E+00
0.750000000000000E+00 0.000000000000000E+00 0.750000000000000E+00
0.000000000000000E+00 -0.100000000000000E+01 0.000000000000000E+00
-0.250000000000000E+00 0.500000000000000E+00 0.000000000000000E+00
0.625000000000000E+00 -0.375000000000000E+00 0.875000000000000E+00
0.500000000000000E+00 -0.250000000000000E+00 0.750000000000000E+00
0.750000000000000E+00 -0.250000000000000E+00 0.100000000000000E+01
0.625000000000000E+00 -0.125000000000000E+00 0.875000000000000E+00
0.500000000000000E+00 0.000000000000000E+00 0.750000000000000E+00
-0.250000000000000E+00 -0.100000000000000E+01 0.000000000000000E+00
-0.500000000000000E+00 -0.100000000000000E+01 0.000000000000000E+00

Best,
Roxana