On 09/10/2018 11:52, Michal Krompiec wrote:
Hello,
I'm struggling with SCF convergence of something that should be a simple
calculation: InP (zincblende structure taken from a CIF). I tried
ultra-soft pseudopotentials, PAW, changing the k-point grid, increasing
ecutwfc and ecutrho, decreasing mixing_beta and every time the SCF
energy is oscillating instead of converging. I would be grateful for any
advice.
Hello, are you sure you system does not turn out to be metallic? That
would make convergence problematic.
Try this the following input (I have removed all input variables that
you MUST NOT use, unless you really know what you are doing). If it
works remember to check convergence with ecutwfc/wcutrho (your values
look very high to me) and of the k-points w.r.t. the degauss value.
cheers
&CONTROL
calculation = "scf",
pseudo_dir = ".",
outdir = ".",
prefix = "InP",
verbosity='high'
restart_mode = "from_scratch"
/
&SYSTEM
ibrav = 2,
nat = 2,
ntyp = 2,
ecutwfc = 80,
ecutrho = 800,
occupations = "smearing",
smearing="mv"
degauss=0.02
A=5.8687,
/
&ELECTRONS
mixing_beta=0.2,
mixing_mode="plain"
/
ATOMIC_SPECIES
In 49 In.pbe-dn-kjpaw_psl.1.0.0.UPF
P 15 P.pbe-n-kjpaw_psl.1.0.0.UPF
ATOMIC_POSITIONS {alat}
In 0.00000 0.00000 0.00000
P 0.25 0.25 0.25
K_POINTS {automatic}
9 9 9 0 0 0
Thanks in advance,
Michal
*Dr. Michal Krompiec*
Adjunct Professor
School of Chemistry, University of Southampton
Highfield, Southampton SO17 1BJ, UK
and
Head of Computational Modelling | Performance Materials | Early Research
and Business Development
Merck
_______________________________________________
users mailing list
[email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
--
Lorenzo Paulatto - Paris
_______________________________________________
users mailing list
[email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
