Dear All,
I've been practicing some basic calculations with an external electric
field added. I was attempting to calculate the relaxation of 6% strained
silicene for amplitudes between -1 and 1 V/A (eamp). Runs with eamp
between -1 and 0.3 V/A were successful, however, anything above did not
converge. I've tried higher ecutwfc and ecutrho (up to 100 and 100 Ry,
respectively), I've changed the k-point mesh, I've used mixing_beta
between 0.1-0.7, I've changed the diagonalization method and the
mixing_mode, I've tried different pseudopotentials. Still, nothing
helped. Reducing london_rcut allowed the conversion for 0.4 V/A,
however, it had no effect for 0.5 V/A and above.
I don't understand why, but the value of eamp seems to be the problem.
If I reduce eamp, scf will converge, and more perplexingly if I change
the sign of the amplitude, scf will also converge. My input for 0.5 V/A
is given below.
Regards,
Maciej Szary
PhD student,
Institute of Physics,
Poznan University of Technology
&control
calculation='relax'
prefix='T1',
pseudo_dir='$pseudo_dir'
outdir = 'files/',
wf_collect=.true.
etot_conv_thr=1.0D-5
forc_conv_thr=1.0D-4
tefield=.true.
dipfield=.true.
/
&system
ibrav= 4,
a=3.8448
c=30
nat= 2,
ntyp= 1,
ecutwfc = 55,
ecutrho=550,
occupations= 'smearing',
smearing='gauss',
degauss= 0.0001
assume_isolated = '2D'
lspinorb=.true.,
noncolin=.true.,
starting_spin_angle=.true.
vdw_corr='DFT-D'
london_s6=0.75
london_rcut=15.00
edir=3
eamp=0.009723453
eopreg=0.05
emaxpos=0.85
/
&electrons
electron_maxstep = 150
conv_thr= 1.0d-7
diagonalization='cg'
mixing_beta=0.5
mixing_mode='local-TF'
scf_must_converge=.false.
/
&ions
ion_dynamics='bfgs'
/
ATOMIC_SPECIES
Si 28.0855 Si.rel-pbe-n-nc.UPF
ATOMIC_POSITIONS crystal
Si -0.333333300 0.333333330 0.579899915 0 0 1
Si 0.000000000 0.000000000 0.562100085 0 0 1
K_POINTS automatic
12 12 1 0 0 0
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
