Discrepancy in Fermi Energy Levels from EPW and SCF Calculations for AB-Stacked Bilayer Graphene

[Apricity]

Dear all,
I tried to calculate the Electron phonon coupling matrix of AB-Stacked Bilayer Graphene (have four C atoms). But, I observed a significant discrepancy between the Fermi energy level obtained from EPW calculations and the one derived from the SCF calculations.
The Fermi energy level for AB-Stacked Bilayer Graphene in SCF calculations is -4.2568 eV,but in EPW calculations it is -0.648629 eV.
This is my SCF input：
&CONTROL
calculation = 'scf'
prefix = 'graphene'
verbosity = 'high'
tstress = .true.
tprnfor = .true.
outdir = './tmp'
pseudo_dir = '../pseudo',
/
&SYSTEM
ibrav = 4
celldm(1)=4.660181
celldm(3)=8.515603
nat = 4
ntyp = 1
nbnd = 30
ecutwfc = 120.0
occupations = 'smearing'
smearing = 'mv'
degauss = 0.01
assume_isolated = '2D'
nosym = .true.
noinv = .true.
vdw_corr = 'DFT-D3'
dftd3_version = 4
dftd3_threebody = .false.
/
&ELECTRONS
conv_thr = 1.0D-16
/
ATOMIC_SPECIES
C 12.0107 C.pbe-n-kjpaw_psl.1.0.0.UPF
ATOMIC_POSITIONS (crystal)
C 0.3333319724 0.6666680574 0.0806629285
C 0.9999977350 0.0000022515 0.0806802213
C 0.0000023307 0.9999976754 0.2431297898
C 0.6666679382 0.3333320320 0.2431470603
K_POINTS (automatic)
30 30 1 0 0 0



and here is my EPW input：

epw
&inputepw
prefix = 'Gra'
outdir = '../tmp/'
dvscf_dir = './save/'
system_2d = 'dipole_sp'
asr_typ = 'simple'
vme = 'dipole'

nbndsub = 16
ep_coupling = .true.
elph = .true.
use_ws = .false.
num_iter = 800

dis_froz_min = -25
dis_froz_max = -2

proj(1) = 'f=0.3333319724,0.6666680574,0.0806629285:s;px;py;pz'
proj(2) = 'f=0.9999977350,0.0000022515,0.0806802213:s;px;py;pz'
proj(3) = 'f=0.0000023307,0.9999976754,0.2431297898:s;px;py;pz'
proj(4) = 'f=0.6666679382,0.3333320320,0.2431470603:s;px;py;pz'

fsthick = 5
eps_acustic = 5
degaussw = 0.005
degaussq = 0.01

wdata(1) = 'dis_mix_ratio = 0.7'
wdata(2) = 'dis_num_iter = 800'
wdata(3) = 'conv_window = 3'
wdata(4) = 'trial_step = 1.0'
wdata(5) = 'guiding_centres=.true.'

epwwrite = .true.
epwread = .false.
wannierize = .true.

nk1 = 30, nk2 = 30, nk3 = 1
nq1 = 10, nq2 = 10, nq3 = 1

nkf1 = 150, nkf2 = 150, nkf3 = 1
nqf1 = 150, nqf2 = 150, nqf3 = 1

/

The above file obtained good results in a single layer of graphene，and I think the chosen energy range fully contains the valence bands.
Thank you in advance for your time and assistance. I look forward to your guidance.

Thanks

[Shashi]

Dear Apricity,

You mentioned that the Fermi energy for AB-stacked bilayer graphene is –4.2568 eV in the SCF calculations, but –0.648629 eV in the EPW calculations.

This suggests that your Wannierization may not be accurate. The Fermi level in EPW should be close to the SCF value, typically differing only after the second decimal place, not by this large an amount.

Please share your epw.out files so I can provide better guidance.

Regards,
Shashi

[Apricity]

Dear Shashi,

Here is my nscf.in, nscf.out, epw.in and epw.out files.

I interrupted the calculation of epw because I noticed that the Fermi energy was completely wrong, which led to the final part of the output being truncated.

Regards,
Apricity

[stiwari]

Hi Apricity,

I see that in the output epw writes " Fermi energy coarse grid = -0.648629 eV". Which means this is the Fermi-level obtained after the NSCF calculation. The reason is that in nscf calculation you have used "smearing = mv" which results in Fermi-level being dependent on your "degauss". Therefore, this Fermi level does not reflect the highest-occupied level unlike the SCF calculation.

If you want both the SCF and NSCF Fermi-levels to be the same you either need to remove smearing from NSCF, or use the same smearing for the SCF calculation.

Best,
Sabya.