My Community

Dear VASP users and developers,

I'm calculating the phonon frequency for LaCrO3 at Gamma point by using IBRION = 8. However, the acoustic frequency is a lit bit large ( ~7.5 cm-1). I tried to set EDIFF =1E-10, or ENCUT=700 , but the results become even worse. I thought I might omit some parameters in INCAR file. Is there anyway to improve the accurancy for IBRION = 8 calculations?

Thanks!

My INCAR file: ( I relax the structure first, and then use the WAVECAR to calculate the phonons.

SYSTEM = LaCrO3
EDIFF = 1E-8
PREC= Accurate
ADDGRID = .TRUE.
ENCUT= 650.00 eV
ISTART=1
ICHARG=0
ISMEAR = 0
SIGMA = 0.05
IBRION= 8
LEPSILON=.TRUE.
LRPA=.FALSE.
#NFREE=2
#POTIM=0.01 !displacement of atoms in Ang.
IALGO= 48
#LREAL=Auto !needed for good convergence
NSIM = 4
LCHARG = .False.
LPLANE=.TRUE.

LMAXMIX = 4
LDAU = .TRUE.
LDAUTYPE = 2
LDAUU = 0 4.0 0
LDAUPRINT = 2
LDAUL = -1 2 -1


POSCAR:
LaCrO3
7.78000000000000
0.0000000000000000 0.5040000198716702 0.5040000198716702
0.5040000198716702 0.0000000000000000 0.5040000198716702
0.5040000198716702 0.5040000198716702 0.0000000000000000
La Cr O
2 2 6
Direct
0.0000000000000000 0.0000000000000000 0.0000000000000000
0.5000000000000000 0.5000000000000000 0.5000000000000000
0.2500000000000000 0.2500000000000000 0.2500000000000000
0.7500000000000000 0.7500000000000000 0.7500000000000000
0.5000000000000000 0.0000000000000000 0.0000000000000000
0.0000000000000000 0.5000000000000000 0.0000000000000000
0.0000000000000000 0.0000000000000000 0.5000000000000000
0.5000000000000000 0.5000000000000000 0.0000000000000000
0.0000000000000000 0.5000000000000000 0.5000000000000000
0.5000000000000000 0.0000000000000000 0.5000000000000000

KPOINTS:
Automatic generation
0
Gamma
8 8 8
0.0 0.0 0.0

your parameters provide very high accuracy already. I also suppose that you have made careful convergence checks of the k-point grid (8x8x8 seems reasonable to me)
usually, the experimental frequencies are best reproduced if one uses a XC-type that best reproduces the experimental lattice parameters (see e.g. R. Wahl et al, PRB 78, 104116 (2008) ). Please check how large the discrepancies between the experimental and the calculated lattice constants are; have you tried calculate the frequencies using eg PBEsol (GGA=PS)?