Page 1 of 1

Very large lambda and gamma in HfB2

Posted: Mon Jun 17, 2019 8:46 am
by amosyang
Dear all,

Recently, I am trying to calculate the el-ph coupling in high-temperature HfB2. The scf and nscf calculations are all right but there is something wrong with the calculated lambda and gamma at specific q points. The lambda can be large as well as 50 and the gamma is almost 30 meV at some point. I try to use large q mesh, but nothing changes. So can you instruct me how to improve this?

In the following, it is the input file for my case:
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
scf.in

&control
calculation='scf',
prefix='hfb',
pseudo_dir = '../pp',
outdir='./',
tprnfor = .true.,
tstress = .true.,
/
&system
!noncolin = .true.,
!lspinorb = .true.,
ibrav = 4,
celldm(1) = 5.943973377527746,
celldm(3) = 1.109350247693944,
nat= 3 ,
ntyp = 2 ,
ecutwfc = 60,
smearing = 'gaussian'
occupations = 'smearing',
degauss = 0.02
/
&electrons
diagonalization = 'david'
mixing_mode = 'plain'
mixing_beta = 0.7,
conv_thr = 1.0d-12
/
ATOMIC_SPECIES
Hf 178.49 Hf_ONCV_PBE-1.1.upf
B 10.811 B_ONCV_PBE-1.0.upf

ATOMIC_POSITIONS crystal
Hf 0.0000000000000000 -0.0000000000000000 -0.0000000000000000
B 0.3333333333333333 0.6666666666666667 0.5000000000000000
B 0.6666666666666667 0.3333333333333333 0.5000000000000000

K_POINTS automatic
12 12 12 0 0 0

nscf.in

&control
calculation='nscf',
prefix='hfb',
pseudo_dir = '../pp',
outdir='./',
tprnfor = .true.,
tstress = .true.,
/
&system
!noncolin = .true.,
!lspinorb = .true.,
ibrav = 4,
celldm(1) = 5.943973377527746,
celldm(3) = 1.109350247693944,
nat= 3 ,
ntyp = 2 ,
ecutwfc = 60,
smearing = 'gaussian'
occupations = 'smearing',
degauss = 0.01
nbnd= 15
nosym=.true.
/
&electrons
diagonalization = 'david'
mixing_mode = 'plain'
mixing_beta = 0.7,
conv_thr = 1.0d-12
/
ATOMIC_SPECIES
Hf 178.49 Hf_ONCV_PBE-1.1.upf
B 10.811 B_ONCV_PBE-1.0.upf

ATOMIC_POSITIONS crystal
Hf 0.0000000000000000 -0.0000000000000000 -0.0000000000000000
B 0.3333333333333333 0.6666666666666667 0.5000000000000000
B 0.6666666666666667 0.3333333333333333 0.5000000000000000

K_POINTS crystal
1728
0.000000000000000 0.000000000000000 0.000000000000000 0.000578703703704
0.000000000000000 0.000000000000000 0.083333333333333 0.000578703703704
.....

epw.in

--
&inputepw
prefix = 'hfb',
amass(1) = 178.49,
amass(2) = 10.811,
outdir = './'

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

epwwrite = .true.
epwread = .false.

nbndsub = 12
nbndskip = 0

wannierize = .true.
num_iter = 500
dis_win_max = 30
dis_win_min = -25
dis_froz_min= -12
dis_froz_max= 16
proj(1) = 'random'


wdata(1) = 'bands_plot = .true.'
wdata(2) = 'begin kpoint_path'
wdata(3) = 'G 0.0000000 0.0000000 0.0000000 L 0.5000000 0.5000000 0.5000000'
wdata(4) = 'L 0.5000000 0.5000000 0.5000000 M 0.0000000 0.5000000 0.5000000'
wdata(5) = 'M 0.0000000 0.5000000 0.5000000 X 0.0000000 0.5000000 0.0000000'
wdata(6) = 'X 0.0000000 0.5000000 0.0000000 G 0.0000000 0.0000000 0.0000000'
wdata(7) = 'end kpoint_path'
wdata(8) = 'bands_plot_format = gnuplot'

iverbosity = 0


elecselfen = .false.
phonselfen = .true.

!parallel_k = .true.
!parallel_q = .false.

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

a2f = .true.

dvscf_dir = '../phonons/save'

nkf1 = 20
nkf2 = 20
nkf3 = 20

!rand_q = .true.
!rand_nq = 10000
filqf ='qpath.dat'

nk1 = 12
nk2 = 12
nk3 = 12

nq1 = 4
nq2 = 4
nq3 = 4

/
12 cartesian
0.000000000000000E+00 0.000000000000000E+00 0.000000000000000E+00
0.000000000000000E+00 0.000000000000000E+00 0.225357140830587E+00
0.000000000000000E+00 0.000000000000000E+00 -0.450714281661174E+00
0.000000000000000E+00 0.288675134594805E+00 0.000000000000000E+00
0.000000000000000E+00 0.288675134594805E+00 0.225357140830587E+00
0.000000000000000E+00 0.288675134594805E+00 -0.450714281661174E+00
0.000000000000000E+00 -0.577350269189610E+00 0.000000000000000E+00
0.000000000000000E+00 -0.577350269189610E+00 0.225357140830587E+00
0.000000000000000E+00 -0.577350269189610E+00 -0.450714281661174E+00
0.250000000000011E+00 0.433012701892207E+00 0.000000000000000E+00
0.250000000000011E+00 0.433012701892207E+00 0.225357140830587E+00
0.250000000000011E+00 0.433012701892207E+00 -0.450714281661174E+00

.....>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

During the epw calculation, the qpoints file contains 2275 qpoints.

Best regards,

amosyang
Shandong University
jy_yang@sdu.edu.cn

Re: Very large lambda and gamma in HfB2

Posted: Mon Jun 17, 2019 10:07 am
by ilias-serifi
Hi i have the same problem for aluminene



best