Here attached, a bandstructure of a system of fcc nikel (32 atoms) with an element in a interstitial site.
the reciprocal points of the x axis are : Gamma X W K Gamma L U W L K, U X (ref :Setaywan an d Cutarolo, Computational Materials Science (2010))
I used the following set up in VASP to converged the structure :
%============================%
ISTART = 0
ICHARG = 1 (I added a CHGCAR of a previous run)
LPLANE = TRUE
LWAVE = FALSE
ENCUT = 400
ISPIN = 2
PREC = Accurate
ISMEAR = 1
SIGMA = 0.1
NSW = 40
IBRION = 1
ISIF = 3
EDIFFG = -1E-3
EDIFF = 1E-8
POTIM = 0.02
%============================%
The K point grid which has been used is a 5³ , Monkhorst Pack
Then the following set up used for the displacement method (using phonopy)
%============================%
LWAVE=FALSE
LPLANE=TRUE
ENCUT = 400
ADDGRID=.TRUE.
PREC = Accurate
ISPIN = 2
ISMEAR=1
SIGMA=0.1
IBRION = -1
NSW = 0
EDIFF = 1.e-6
%============================%
And the same K points grid of 5³ has been used
The point is, if you look between Gamma and X one of the acoustic modes become slighly complex.
And I don't have any explanation about this "behavior" . The same fact is observed using a supercell of 108 Ni atoms.
Does someone has any explanations or remarks on this ?
Hello Hua Wang,
I'm checking it. with a better ENCUT.
Pre relaxation run has been converged with EDIFF=1E-8 and ENCUT=400. I will try with ENCUT=600 and then recalculate the phonon band structure.
I will update the results.
Matthieu
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science