Running EPW for 2D semiconductors

[hanwooh]

Dear all,

I want to calculate electron self-energy for scattering rates by electron-phonon interactions for 2D semiconductors having finite gap.

In the paper (S. Ponc et al. / Computer Physics Communications 209 (2016) 116133), since electron scattering rates for undoped silicon were shown, I believe that EPW can calculate electron self-energies for not only metals but semiconductors.

However, when I tried to run EPW v4.1 as implemented in qe 6.0 for 2D semiconductors (I used tag system_2d = .true.), I faced problems that self energies become unrealistic huge values and values in decay.epmat_wanep are strange while values in decay.epwan doesn't have any problem.

The final outputs for electron self-energy, decay.epmat_wanep, and decay.epwan are described as below.

Code: Select all

# Electron lifetime (meV)
#      ik       ibnd                 E(ibnd)      Im(Sgima)(meV)
        1          1   -0.15999555711729E+02    0.21825383185903E+70
        1          2   -0.14280977405867E+02    0.23882330257160E+70
        1          3   -0.14110604428389E+02    0.77447215906861E+68
        1          4   -0.99986280252164E+01    0.49041383819028E+70
        1          5   -0.93109835174232E+01    0.64127010367182E+70
        1          6   -0.72418709172874E+01    0.23224606667463E+70
        1          7   -0.55275232062358E+01    0.40327282641546E+70
        1          8   -0.44959862051001E+01    0.13195538727356E+70
        1          9   -0.38607786125222E+01    0.97470765626990E+68
        1         10   -0.37960743306642E+01    0.97775694174974E+68
        1         11   -0.34783674871345E+01    0.49426565858991E+69
        1         12   -0.31457145941218E+01    0.18289964703863E+68
        1         13   -0.23037737592355E+01    0.12878046165150E+70
        1         14   -0.21821556104848E+01    0.27546726435307E+70
        1         15   -0.91689570772186E+00    0.49849784396679E+69
        1         16    0.35050596335270E+00    0.48505673827446E+69
        1         17    0.15277075491323E+01    0.13310271379821E+70
        1         18    0.18053124124261E+01    0.10333756926782E+70
        1         19    0.22840073069910E+01    0.14850500519154E+70
        1         20    0.23898102667957E+01    0.12670152113566E+70
        1         21    0.29403289568160E+01    0.10128077867223E+70
        1         22    0.33422893155053E+01    0.10512779864113E+70
        1         23    0.40901137616935E+01    0.12343982954155E+70
        1         24    0.50314389154001E+01    0.30163515826647E+70
        2          1   -0.15996087175471E+02    0.32446348472872E+70
        2          2   -0.14280964316454E+02    0.24829855054775E+70
        2          3   -0.14103899569592E+02    0.12092444712278E+70
        2          4   -0.99978475174984E+01    0.51406473617840E+70
        2          5   -0.93098952452556E+01    0.72611846157543E+70
        2          6   -0.72397288644130E+01    0.35127960659679E+70
        2          7   -0.55287121885116E+01    0.40467005042930E+70
        2          8   -0.44973963324532E+01    0.17414760265497E+70
        2          9   -0.38953474499550E+01    0.79431605711224E+69
        2         10   -0.38458557052216E+01    0.60084841736295E+69
        2         11   -0.34480450435216E+01    0.85410126148159E+69
        2         12   -0.31289694291679E+01    0.11496459370621E+70
        2         13   -0.23334809942658E+01    0.22787190371606E+70
        2         14   -0.21805755550129E+01    0.39106222318838E+70
        2         15   -0.92128532448568E+00    0.13882298290173E+70

Code: Select all

#  R_e,    R_p, max_{m,n,nu} |g(m,n,nu;R_e,R_p)| 
   0.0000000000   0.0000000000***************
   3.2802192104   0.0000000000***************
   3.2802192104   0.0000000000***************
   6.5604384209   0.0000000000***************
   6.5604384209   0.0000000000***************
   7.2373226221   0.0000000000***************
   7.2373226221   0.0000000000***************
   7.2373226221   0.0000000000***************
   7.2373226221   0.0000000000***************
   8.5964245401   0.0000000000***************
   8.5964245401   0.0000000000***************
   8.5964245401   0.0000000000***************
   8.5964245401   0.0000000000***************
   9.8406576313   0.0000000000***************
   9.8406576313   0.0000000000***************
  10.8137813529   0.0000000000***************
  10.8137813529   0.0000000000***************
  10.8137813529   0.0000000000***************
  10.8137813529   0.0000000000***************
  14.0980678421   0.0000000000***************
  14.0980678421   0.0000000000***************
  14.4746452443   0.0000000000***************
  14.4746452443   0.0000000000***************
  14.4746452443   0.0000000000***************
  14.4746452443   0.0000000000***************
  15.5497546332   0.0000000000***************
  15.5497546332   0.0000000000***************
  15.5497546332   0.0000000000***************
  15.5497546332   0.0000000000***************

Code: Select all

# Spatial decay of e-p matrix elements in Wannier basis
  0.000000000000000E+000  0.291346855638052
   3.28021921044301       6.490853709214308E-003
   3.28021921044301       6.490853710880915E-003
   6.56043842088602       1.837076326283176E-003
   6.56043842088602       1.837076327293353E-003
   7.23732262214935       3.833923468712698E-002
   7.23732262214935       3.830008207020476E-002
   7.23732262214935       3.830008207136522E-002
   7.23732262214935       3.833923468621558E-002
   8.59642454013257       4.716060827665735E-003
   8.59642454013257       4.703116401033521E-003
   8.59642454013257       4.703116400848962E-003
   8.59642454013257       4.716060827658721E-003
   14.0980678420747       1.782159442384595E-003
   14.0980678420747       1.782159442624803E-003
   14.4746452442987       4.537116385138993E-003
   14.4746452442987       4.540590812053117E-003
   14.4746452442987       4.540590810329914E-003
   14.4746452442987       4.537116384385748E-003
   9.84065763132902       6.625166564662149E-004
   9.84065763132902       6.625166555358239E-004
   10.8137813529048       9.164143016137713E-004
   10.8137813529048       9.132745853651637E-004
   10.8137813529048       9.132745851209546E-004
   10.8137813529048       9.164143019652921E-004
   13.1208768417720       5.103308308118051E-004
   13.1208768417720       5.103308291178507E-004
   13.4720783682321       4.257495095769741E-004
   13.4720783682321       4.307649960071313E-004
   13.4720783682321       4.307649949236289E-004
   13.4720783682321       4.257495090145206E-004
   15.5497546332403       6.730503551253587E-004
   15.5497546332403       6.730474091634435E-004
   15.5497546332403       6.730474091521625E-004
   15.5497546332403       6.730503551355285E-004
   16.3577382332724       2.642142181761415E-004
   16.3577382332724       2.636055808538121E-004
   16.3577382332724       2.636055818951108E-004
   16.3577382332724       2.642142181628109E-004

As I checked wannierization and phonon dispersion curve, I cannot found any problems for the wannier and phonon.

When I saw output file carefully, there was warning message for difference between Fermi energies from coarse grid and fine k-mesh.

Does it give serious problems?

Is there anyone who encounters the same problem or knows how to solve this problem?

Thank you.

Best regards,

[chris]

I am also trying to run epw on 2D semiconductors with the same version of quantum espresso and epw, but I get similar error as you. Some of the values are unphysically large. Did you manage to solve your problem?

[JuniorKanga]

Try to check the spread of each band in the .wout file. The problem must be related to a very large value of the spread. Each spread must be lower than 5. I did it in my case and it worked.

Best

[vinpandey]

Hi users, I was running EPW to calculate electron relaxation time for PbI2 with incorporating SOC. after it writes in output file
===================================================================
irreducible q point # 1
===================================================================

Symmetries of small group of q: 6
in addition sym. q -> -q+G:

Number of q in the star = 1
List of q in the star:
1 0.000000000 0.000000000 0.000000000
, job gets stopped automatically. and writes in error file, At line 231 of file readmat_shuffle2.f90 (unit = 81, file = '/home/EPW/phonon/save/pbi2.dyn_q1').
I have renamed pbi2.dyn_q1.xml to pbi2.dyn_q1. can you help me to remove this error. I am giving my epw input file
--
&inputepw
prefix = 'pbi2'
amass(1)=207.2
amass(2)=126.904477

outdir = './'
iverbosity = 0
elph = .true.
epbwrite = .true.
epbread = .false.
etf_mem = .false.
epwwrite = .true.
epwread = .false.
nbndsub = 12
nbndskip = 7
wannierize = .true.
num_iter = 300
wdata(1) = 'dis_num_iter= 100000'
iprint = 2
dis_win_max = 10
dis_froz_max= 0.0
dis_win_min = -5.0d0
proj(1) = 'Pb:s'
proj(2) = 'I:s'
proj(3) = 'Pb:p'
proj(4) = 'I:p'
wdata(2) = 'bands_plot = .true.'
wdata(3) = 'begin kpoint_path'
wdata(4) = 'K 0.33333 0.33333 0.00 G 0.00 0.00 0.00'
wdata(5) = ' G 0.00 0.00 0.00 M 0.50 0.0 0.00'
wdata(6) = 'M 0.50 0.0 0.00 K 0.33333 0.33333 0.00'
wdata(7) = 'end kpoint_path'

system_2d = .true.
elecselfen = .false.
nest_fn = .false.
phonselfen = .true.
a2f = .false.
parallel_k = .true.
parallel_q = .false.

fsthick = 10 ! eV
eptemp = 300 ! K (same as PRB 76, 165108)
degaussw = .10 ! eV

dvscf_dir = '/home/prasenjit/vineet/pbi2/EPW/phonon/save'
filukk = './pbi2.ukk'
filqf = '/home/prasenjit/vineet/pbi2/EPW/epw/mesh/path.dat'

nkf1 = 150
nkf2 = 150
nkf3 = 1

nk1 = 15
nk2 = 15
nk3 = 1

nq1 = 3
nq2 = 3
nq3 = 1
/
3 cartesian
0.000000000000000 0.000000000000000 0.000000000000000 1.0
0.000000000000000 0.384900179363945 0.000000000000000 1.0
0.333333333333333 0.577350269045918 0.000000000000000 1.0

thanks in advance

[sponce]

Dear vinpandey,

Since you are using SOC, renaming them wont help since the code then expects xml format.
Are you sure you have the two SOC flags activated both in your scf.in and nscf.in calculations?

If you do have SOC, then you should keep .xml.

You need to use the correct pp.py python script that deals with .xml format. For that you can look at the Pb example with SOC and take the pp.py script from there.

Best,
Samuel

[vinpandey]

sir, thanks for your help, the error was there only , I was not using those flags of SOC.
After that I did this calculation to calculate electron relaxation time for 2D PbI2, just after it does wannierisation part, it gives me error "segmentation fault". Basically I did a phonon calculation with supercell3x3x1 and then SCF, nscf for EPW with 15x15x1. here I am giving you my epw input file . can you suggest me to rectify this error ?
--
&inputepw
prefix = 'pbi2'
amass(1)=207.2
amass(2)=126.904477

outdir = './'
iverbosity = 0
elph = .true.
epbwrite = .true.
epbread = .false.
epwwrite = .true.
epwread = .false.
nbndsub = 18
nbndskip = 0
wannierize = .true.
num_iter = 300
wdata(1) = 'dis_num_iter= 10000'
iprint = 2
dis_win_max = 10
dis_froz_max= 0.0
dis_win_min = -5.0d0
proj(1) = 'Pb:p'
proj(2) = 'I:p'
wdata(2) = 'bands_plot = .true.'
wdata(3) = 'begin kpoint_path'
wdata(4) = 'K 0.33333 0.33333 0.00 G 0.00 0.00 0.00'
wdata(5) = ' G 0.00 0.00 0.00 M 0.50 0.0 0.00'
wdata(6) = 'M 0.50 0.0 0.00 K 0.33333 0.33333 0.00'
wdata(7) = 'end kpoint_path'

system_2d = .true.
elecselfen = .false.
nest_fn = .false.
phonselfen = .true.
a2f = .false.
parallel_k = .true.
parallel_q = .false.

fsthick = 10 ! eV
eptemp = 300 ! K (same as PRB 76, 165108)
degaussw = .10 ! eV

dvscf_dir = '/home/vpandey/vineet/pbi2/epw-soc/phonon/save'
filukk = './pbi2.ukk'
filqf = '/home/vpandey/vineet/pbi2/epw-soc/epw/mesh/path.dat'

nkf1 = 100
nkf2 = 100
nkf3 = 1

nk1 = 15
nk2 = 15
nk3 = 1

nq1 = 3
nq2 = 3
nq3 = 1
/
3 cartesian
0.000000000 0.000000000 0.000000000 1.0
0.000000000 0.384900179 0.000000000 1.0
0.333333333 0.577350269 0.000000000 1.0
can you suggest me , what can be a better kpoint fine grid ? or do I need to do different calculation by changing kpoint fine grid
one more thing, I was following tutorial given in qe-6.0. where it first calculates phonon relaxation time and then electron relaxation time. can I do directly electron relaxation time calculation by giving fermi energy from my nscf file(half of sum of VBM and CBM). may be, last doubt is more silly, but I am not understanding.
thank you in advance.