HCP has two parameters to optimize namely a and c/a; the process detailed here gives pretty accurate result to determine those parameters. Zr is taken as an example here, where positions are special, high symmetry positions, i.e. (1/3, 2/3,1/4) and (2/3, 1/3,3/4)
i. Take 7 to 8 different volumes across the estimated equilibrium volume (from published XRD data).
ii. Range of volume scanned should be typically within +/- 10%
iii. If atom positions are in high symmetry positions then atoms should not be moved. However, in the POSCAR file set 'selected dynamics' and motion should be set to T, to get force info in OUTCAR, but set ISIF=5 in INCAR to change only the shape of the unit cell.
An example of POSCAR -
HCP Zr parama (vary this to get different volume) 1.000000000 0.000000000 0.00000000000 -0.500000000 0.866025404 0.00000000000 0.000000000 0.000000000 1.60000000000 Zr 2 Selective dynamics Direct 0.6666666666666667 0.3333333333333333 0.7500000000000000 T T T 0.3333333333333333 0.6666666666666667 0.2500000000000000 T T T
iv. INCAR setting
ISTART=0 LREAL=.FALSE. PREC=Accurate ADDGRID=.TRUE. ENCUT=550 ISPIN=2 NPAR=6 MAGMOM=2*2.0
IBRION=2 ISIF=4 POTIM=0.15 ISMEAR=1 SIGMA=0.01
NSW=100 NELM=50 EDIFF=1.0E-6
v. Fit 3rd order BM EOS to determine the equilibrium energy, volume and bulk modulus (E0, V0 and B)
vi. For each run also find the final a and c/a
vii. Fit 3rd order BM EOS between E vs. a^3 to determine the equilibrium a0^3, during this fitting keep E0 constant (found from step v)
viii. Now we have both V0 and a0, from which c0/a0 can be calculated. Use these final cell parameters to determine the energy accurately using ISMEAR = -5 in INCAR.
**** Modified version ****
HCP has two parameters to optimize namely a and c/a; the process detailed here gives pretty accurate result to determine those parameters.
i. Take 7 to 8 different volumes across the estimated equilibrium volume (from published XRD data) using ISMEAR = 0 or 1
ii. Range of volume scanned should be typically within +/- 10%
iii. If atom positions are in high symmetry positions then atoms should not be moved; accordingly, use ISIF = 4 or 5
v. Fit 3rd order BM EOS to determine the equilibrium energy, volume and bulk modulus (E0, V0 and B), the useful quantity from this fitting is V0 and B
vi. With this fixed V0 run again with ISIF=4 or 5 to determine a0 and c0
vii. Keeping this a0 and c0 fixed (i.e. fixed V0 ) take another run using ISMEAR=-5 for accurate ground state energy (E0) calculation
Hi Chandra Bhanu Basak
I never tried a particular HCP structure, but i am wondering why you can't follow the regular procedure,
i.e., change 'a' first by putting some optimal guess on 'c', plot its Ground state energy as a function of 'a' to get the optimized 'a'. And then keeping this optimized 'a' fixed, change 'c' to get its optimized value. (Keep ISIF = 2, NSW = 0, and ISMEAR = 0 here)
and at the end do a final relaxation with ISIF = 2, NSW = 200, and ISMEAR = 0.
I hope this should work.
Another way can be keeping ISIF = 3, ISMEAR = 0, NSW =200 to relax the system in one single step for 3D system. But VASP developers recommend the first one.
Hope this helps!
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India