[Pw_forum] oscillating scf accuracy in relax run

pulkit rai pulkitrai at gmail.com
Wed Jan 6 13:07:36 CET 2016


Hello everyone,

I am trying to relax the structure of the compound HoMnO3.

The scf run converged in 27 iterations but the relax run does not converge
and keeps on oscillating in scf accuracy.

Why is this happening ? Is it because of the pseudopotentials ?

I am new to quantum espresso. Please give some suggestions.

I have attached my input file below.

Thank you.. !



 &control
    calculation='scf'
    restart_mode='from_scratch',
    tstress = .true.
    tprnfor = .true.
    prefix='HoMn_lda_scf',
    pseudo_dir = '/home/HoMnCr/QEpseudo',
    outdir='/lustre/01/HoMnO3/'
    disk_io='low'

 &system
ecutwfc=  80.00
ibrav=8, A= 5.616838, B= 7.458766, C= 5.242448, cosAB=0,cosBC=0,cosAC=0,
nat=20
ntyp=4,
occupations='smearing'
degauss=0.005

 &electrons
electron_maxstep=200
conv_thr=1.0e-8

ATOMIC_SPECIES
O     15.99     O.pz-hgh.UPF
Mn   51.99     Mn.pz-sp-hgh.UPF
Ho    164.93   Ho.pz-sp-hgh.UPF
ATOMIC_POSITIONS (crystal)
O        0.30916       0.05393      -0.30577
O        0.19084      -0.05393       0.19423
O       -0.30916       0.55393       0.30577
O        0.80916       0.44607       0.80577
O       -0.30916      -0.05393       0.30577
O        0.80916       0.05393       0.80577
O        0.30916       0.44607      -0.30577
O        0.19084       0.55393       0.19423
O        0.46235       0.25000       0.10381
O        0.03765       0.75000       0.60381
O       -0.46235       0.75000      -0.10381
O        0.96235       0.25000       0.39619
Mn      0.50000       0.00000       0.00000
Mn      0.50000       0.50000       0.00000
Mn      0.00000       0.00000       0.50000
Mn      0.00000       0.50000       0.50000
Ho       0.07091       0.25000      -0.01203
Ho       0.42909       0.75000       0.48797
Ho      -0.07091       0.75000       0.12030
Ho       0.57091       0.25000       0.51203
K_POINTS automatic
8 6 8  1  1  1
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20160106/9beaa231/attachment.html>


More information about the users mailing list