Calculation of the intensity of lambda (ele-phon coupling constant) is too big

[fale li]

Dear all,

When I calculate the EPC strength (lambda) of BCC V according to the example that pb has no soc, the result in epw.out is extremely large, reaching 3799,and the lambda_tr reaching 3357. I found that someone in the forum had encountered this problem before, and he solved it by changing the pseudopotentials files, but I failed.

The energy band diagram obtained by Wannier processing is basically consistent with the energy band diagram calculated by QE near Fermi level.My calculated phonon spectrum is also correct.

The abnormal situation I found was that I set nq point to 4*4*4 when I first calculated EPW, and then I found that there were too few Q points, and the density of states obtained by nscf calculation was discontinuous. So I set nq point to 12*12*12 and calculated epw. As a result, the lambda in epw.out became ************, and the lambda_tr was as high as 6813.What should I do to get the correct ele-phon coupling constant?

The following is the details of my epw.in file：

epw calculation
&inputepw
prefix = 'V',
amass(1) = 50.942
outdir = './'

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

epwwrite = .true.
epwread = .false.
nbndsub = 12
bands_skipped = 'exclude_bands = 1:8'
!efermi_read=.true.
!fermi_energy=16.527850

wannierize = .true.
num_iter = 10000
! dis_win_max = 40
! dis_win_min = -50
dis_froz_min= 12.5
dis_froz_max= 18
proj(1) = 'V:s,d'
!dis_num_iter=1000

wdata(1) = 'bands_plot = .true.'
wdata(2) = 'begin kpoint_path'
wdata(3) = 'G 0.00 0.00 0.00 H 0.50 -0.50 0.50'
wdata(4) = 'H 0.50 -0.50 0.50 N 0.00 0.00 0.50'
wdata(5) = 'N 0.00 0.00 0.50 G 0.00 0.00 0.00'
wdata(6) = 'G 0.00 0.00 0.00 P 0.25 0.25 0.25'
wdata(7) = 'P 0.25 0.25 0.25 H 0.50 -0.50 0.50'
wdata(8) = 'H 0.50 -0.50 0.50 N 0.00 0.00 0.50'
wdata(9) = 'end kpoint_path'
wdata(10) = 'bands_plot_format = gnuplot'

iverbosity = 0

elecselfen = .false.
phonselfen = .true.
!eliashberg = .true.
!liso = .true.
!limag = .true.
wscut = 1

fsthick = 6 ! eV
temps = 0.075 ! K
degaussw = 0.05 ! eV

a2f = .true.

dvscf_dir = '../phonon/save-pp-paw'

nkf1 = 12
nkf2 = 12
nkf3 = 12

nqf1 = 12
nqf2 = 12
nqf3 = 12

nk1 = 4
nk2 = 4
nk3 = 4

nq1 = 4
nq2 = 4
nq3 = 4

/
Best wishes,
fale li

[stiwari]

Hi Fale,

I see that your coarse k,q-grids are a bit too coarse. Are you sure the Wannier interpolation is good? Also, have you checked your phonon dispersion obtained from EPW and direct PH calculation (using matdyn)? Try increasing the coarse grid quantities to see if you see a change in the lambda value.

Also, what pseudopotential are you using?

Best regards,
Sabya.

[fale li]

Hi Sabya

I think the energy band diagram fitted by my Wannier function is still good. I attach the picture.I have tried to increase the coarse grid quantities, and the lambda will change but it is still unusually large.I have used pbe, pbesol and uspp pseudopotentials to calculate, but it has no substantial influence on the results.

Finally, I got a reasonable result by adding the command trans=.true.. In this case, the calculated lambda value is 1.5. But I don't know why this will get the correct result, because I found that the default value of trans is .true..

The following is the input file when I calculated the ph.x
phonon calcuation
&inputph
outdir = './'
prefix = 'V'
fildvscf = 'dvscf'
tr2_ph = 1.0d-15
!epsil = .true.
ldisp = .true.
trans = .true.
!nmix_ph = 10
!alpha_mix = 0.3
amass(1) = 50.942
nq1 = 4, nq2 = 4, nq3 = 4
fildyn = 'V.dyn'
/

Best wishes