the interpolated phonon from EPW doesn't match DFPT result

[physics.yang]

Hi,

I want to study the electron-phonon coupling in MoS2. But I found that the interpolated phonon spectrum from EPW greatly differs from the DFPT spectrum (https://nusu-my.sharepoint.com/:i:/g/personal/e0001020_u_nus_edu/EdGxph3uDyVFixgJhaIJGpMBRi-VFsypzlntB4PtlZ9MtA).

ph.x:
----------------------
--
&INPUTPH
amass(1) = 32.065
amass(2) = 95.94
prefix = 'mos2'
fildyn = "mos2.dyn"
fildvscf = 'dvscf'
ldisp = .true.
nq1 = 9,
nq2 = 9,
nq3 = 1,
tr2_ph = 1.0d-15
recover=.true.
/
-----------------------------

EPW:
-----------------------------
&inputepw
outdir = 'temp/',
prefix = 'mos2'
amass(1) = 32.065
amass(2) = 95.94

iverbosity = 1

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

epwwrite = .true.
epwread = .false.

lpolar = .true.
lifc = .true.

efermi_read= .false.
fermi_energy= -1.128

nbndsub = 11
nbndskip = 6

wannierize = .true.
num_iter = 2000
iprint = 2
dis_win_max= 3
dis_win_min= -8
proj(1) = 'S:l=1'
proj(2) = 'Mo:l=2'
wdata(1) = 'guiding_centres=.true.'
wdata(2) = 'use_ws_distance=.true.'
wdata(3) = 'write_xyz=.true.'
wdata(4) = 'bands_plot=.true.'
wdata(5) = 'begin kpoint_path'
wdata(6) = 'G 0.0 0.0 0.0 K -0.3333 0.6667 0.0'
wdata(7) = 'K -0.3333 0.6667 0.0 M 0.0 0.5 0.0'
wdata(8) = 'M 0.0 0.5 0.0 G 0.0 0.0 0.0'
wdata(9) = 'G 0.0 0.0 0.0 k 0.3333 0.3333 0.0'
wdata(10)= 'k 0.3333 0.3333 0.0 M 0.0 0.5 0.0'
wdata(11)= 'end kpoint_path'
wdata(12)='dis_num_iter=50'
!wdata(13)='wannier_plot = .true.'

elecselfen = .false.
nest_fn = .false.
phonselfen = .false.
a2f = .false.

specfun_el = .false.
wmin_specfun = -5
wmax_specfun = 3
nw_specfun = 800

parallel_k = .true.
parallel_q = .false.

fsthick = 8
eptemp = 0.01
degaussw = 0.025
!degaussq = 0.05

dvscf_dir = '../../7_phonon/save/'
filukk = './mos2.ukk'

filkf = 'meshes/path.dat'

!nkf1 = 108
!nkf2 = 108
!nkf3 = 1

nqf1 = 108
nqf2 = 108
nqf3 = 1

nk1 = 9
nk2 = 9
nk3 = 1

nq1 = 9
nq2 = 9
nq3 = 1
/
12 cartesian
0.000000000000000E+00 0.000000000000000E+00 0.000000000000000E+00 0.08333333333333333
0.000000000000000E+00 0.128300059819916E+00 0.000000000000000E+00 0.08333333333333333
0.000000000000000E+00 0.256600119639831E+00 0.000000000000000E+00 0.08333333333333333
0.000000000000000E+00 0.384900179459747E+00 0.000000000000000E+00 0.08333333333333333
0.000000000000000E+00 0.513200239279662E+00 0.000000000000000E+00 0.08333333333333333
0.111111111111118E+00 0.192450089729873E+00 0.000000000000000E+00 0.08333333333333333
0.111111111111118E+00 0.320750149549789E+00 0.000000000000000E+00 0.08333333333333333
0.111111111111118E+00 0.449050209369704E+00 0.000000000000000E+00 0.08333333333333333
0.111111111111118E+00 0.577350269189620E+00 0.000000000000000E+00 0.08333333333333333
0.222222222222236E+00 0.384900179459747E+00 0.000000000000000E+00 0.08333333333333333
0.222222222222236E+00 0.513200239279662E+00 0.000000000000000E+00 0.08333333333333333
0.333333333333354E+00 0.577350269189620E+00 0.000000000000000E+00 0.08333333333333333
-------------------------------------------------------------------------------------------------------------------------------

Anybody knows the mismatch origin?

Cheers,
Yang Tong
Department of Physics,
NUS

[sponce]

Dear Yang,

EPW does not have specific support for 2D materials.
It is still possible to compute them but you have to treat them as 3D with vacuum.

This means you need to have some k/q-points along the z direction.

Try a 9x9x2 or 9x9x3 coarse grid.

Best wishes,
Samuel

[physics.yang]

Dear Dr. Samuel,

I tried 9x9x2 k/q grids. The interpolated phonon spectrum is the same as the one obtained with the 9x9x1 coarse k/q grids, dramatically different from the DFT one.

Best regards,
Yang Tong