Hhelo,
increase nmix_ph and reduce alpha_mix, in a difficult case recently I
have used:
nmix_ph=10
alpha_mix=.25
Also, the convergence of the scf process in pw.x should be tighter,
something like
conv_thr = 1.0d-12
Because the total energy is variational, but the phonons are not, they
converge much slower with this threshold.
hth
On 06/06/2025 21:48, Ivan Dario Arellano Ramirez wrote:
I QE users
I want to calculate the phonon linewidth for InN. I am following two
tutorials. However, in both cases, I encounter the same issue: no
convergence when using the ph.x module.
Pert. # 1: Fermi energy shift (Ry) = -5.5393E+02 0.0000E+00
iter # 150 total cpu time : 13815.8 secs av.it <http://av.it>.:
5.3
thresh= 4.679E-04 alpha_mix = 0.700 |ddv_scf|^2 = 4.318E-06
End of self-consistent calculation
No convergence has been achieved
This is an overview of the two approaches I followed:
1) perform very dense scf calculation with la2F=.true. in &systems
(Likely, I am exceeding the maximum of 40000 k points)
2) normal scf on coarser k-mesh
3) phonon calculation using ph.x
4) Inverse Fourier transform of dynamical matrix and the phonon
linewidth using q2r.x
5) Fourier transform to obtain dynamical matrix and the phonon
linewidth using matdyn.x
The other tutorial proposed an initial scf calculation followed by an
nscf calculation with la2F=.true. the rest of steps are the same.
Here are all the inputs.
Any suggestions or insights would be greatly appreciated.
1) very dense k-mesh scf
&control
calculation = 'scf'
restart_mode='from_scratch',
prefix='inn_ep',
tstress = .true.
tprnfor = .true.
pseudo_dir = '/home/pps/pbe.0.3.1/PSEUDOPOTENTIALS/',
outdir='./tmp'
verbosity = 'high'
etot_conv_thr = 1.d-6
forc_conv_thr = 1.d-5
/
&system
ibrav = 0,
nat= 4,
ntyp= 2,
ecutwfc = 90,
ecutrho = 900,
occupations='smearing',
smearing='m-v',
degauss=0.02,
la2F = .true.
/
&electrons
conv_thr = 1.0d-8
electron_maxstep = 200,
mixing_mode = 'plain'
mixing_beta = 0.3D0
diagonalization = 'david'
/
ATOMIC_SPECIES
In 114.818 In.pbe-dn-rrkjus_psl.1.0.0.UPF
N 14.007 N.pbe-n-rrkjus_psl.1.0.0.UPF
CELL_PARAMETERS angstrom
1.793358573 -3.106188164 0.000000000
1.793358573 3.106188164 0.000000000
0.000000000 0.000000000 5.799337066
ATOMIC_POSITIONS crystal
In 0.3333330000 0.6666670000 -0.0019331848
In 0.6666670000 0.3333330000 0.4980668152
N 0.3333330000 0.6666670000 0.3767331848
N 0.6666670000 0.3333330000 0.8767331848
K_POINTS automatic
45 45 36 0 0 0
2) normal scf
&control
calculation = 'scf'
restart_mode='from_scratch',
prefix='inn_ep',
tstress = .true.
tprnfor = .true.
pseudo_dir = '/home/pps/pbe.0.3.1/PSEUDOPOTENTIALS/',
outdir='./tmp'
verbosity = 'high'
etot_conv_thr = 1.d-6
forc_conv_thr = 1.d-5
/
&system
ibrav = 0,
nat= 4,
ntyp= 2,
ecutwfc = 90,
ecutrho = 900,
occupations='smearing',
smearing='m-v',
degauss=0.02,
/
&electrons
conv_thr = 1.0d-8
electron_maxstep = 200,
mixing_mode = 'plain'
mixing_beta = 0.3D0
diagonalization = 'david'
/
ATOMIC_SPECIES
In 114.818 In.pbe-dn-rrkjus_psl.1.0.0.UPF
N 14.007 N.pbe-n-rrkjus_psl.1.0.0.UPF
CELL_PARAMETERS angstrom
1.793358573 -3.106188164 0.000000000
1.793358573 3.106188164 0.000000000
0.000000000 0.000000000 5.799337066
ATOMIC_POSITIONS crystal
In 0.3333330000 0.6666670000 -0.0019331848
In 0.6666670000 0.3333330000 0.4980668152
N 0.3333330000 0.6666670000 0.3767331848
N 0.6666670000 0.3333330000 0.8767331848
K_POINTS automatic
15 15 12 0 0 0
3) phonon calculation
Phonons of InN at G!=0
&inputph
tr2_ph = 1.0d-14,
ldisp = .true.,
prefix = 'inn_ep',
recover = .true.
fildyn = 'inn-ph.dyn',
fildvscf = 'inndv'
nq1 = 5,
nq2 = 5,
nq3 = 3,
electron_phonon = 'interpolated'
el_ph_nsigma = 10
el_ph_sigma = 0.02
amass(1) = 114.818,
amass(2) = 14.007,
outdir = './tmp'
/
4) Inverse Fourier Transform
&INPUT
fildyn = 'inn-ph.dyn',
zasr = 'crystal',
flfrc = 'inn-ph.fc',
la2F = .true.
/
5) Fourier Transform
&input
asr = 'crystal',
flfrc = 'inn-ph.fc',
flfrq = 'inn.band.freq'
flvec = 'inn.band.modes'
amass(1) = 114.818,
amass(2) = 14.007,
q_in_band_form = .true.
la2F = .true.
/
8
0.0000000000 0.0000000000 0.0000000000 40 ! Gamma
0.5000000000 0.0000000000 0.0000000000 25 ! M
0.3333333333 0.3333333333 0.0000000000 40 ! K
0.0000000000 0.0000000000 0.0000000000 25 ! Gamma
0.0000000000 0.0000000000 0.5000000000 40 ! A
0.5000000000 0.0000000000 0.5000000000 25 ! L
0.3333333333 0.3333333333 0.5000000000 40 ! H
0.0000000000 0.0000000000 0.5000000000 1 ! A
Best regards,
--
Iván Darío Arellano Ramírez
Profesor Titular Programa de Ingeniería Física
Director Grupo de Investigación en Modelado y Simulación Computacional
Universidad Tecnológica de Pereira
El contenido de este mensaje y sus anexos son únicamente para el uso
del destinatario y pueden contener información clasificada o
reservada. Si usted no es el destinatario intencional, absténgase de
cualquier uso, difusión, distribución o copia de esta comunicación.
_______________________________________________________________________________
The Quantum ESPRESSO Foundation stands in solidarity with all civilians
worldwide who are victims of terrorism, military aggression, and indiscriminate
warfare.
--------------------------------------------------------------------------------
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
--
Dr. Lorenzo Paulatto
IR IMPMC - CNRS UMR 7590 / Sorbonne Université / MNHN
phone: +33 (0)1 442 79822 / telegram: lpaulatto
http://www.impmc.upmc.fr/~paulatto/ - https://anharmonic.github.io/
23-24/423 B115, 4 place Jussieu 75252 Paris CX 05
_______________________________________________________________________________
The Quantum ESPRESSO Foundation stands in solidarity with all civilians
worldwide who are victims of terrorism, military aggression, and indiscriminate
warfare.
--------------------------------------------------------------------------------
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
