VASP Force Correction Patch: Tutorial
In this post, I'll explain how to use the constrained forces method in VASP to simulate constant electric field using the force correction patch from my github. Explanation on the background of this method can be found in my recent manuscript.
First, you need to calculate the Born effective charge tensor of your non-polar structure. For example, the non-polar structure of lead titanate would have following CONTCAR: (space group #123, P4/mmm)
PbTiO3
1.00000000000000
3.9064260257101253 0.0000000000000000 0.0000000000000000
0.0000000000000000 3.9064260257101253 0.0000000000000002
0.0000000000000000 0.0000000000000001 3.9719819950307556
Pb Ti O
1 1 3
Direct
0.0000000000000000 0.0000000000000000 0.0000000000000000
0.5000000000000000 0.5000000000000000 0.5000000000000000
0.5000000000000000 0.5000000000000000 0.0000000000000000
0.0000000000000000 0.5000000000000000 0.5000000000000000
0.5000000000000000 0.0000000000000000 0.5000000000000000
Here, I've imposed constant strain condition in the lateral direction, so that it has same lattice parameter a as its ground state (space group 99, P4mm). After you obtain the optimized structure of your non-polar system, you need to calculate the Born effective charge tensor using either LEPSILON or LCALCEPS tag. It is ideal to pair this calculation with second-derivative calculations using IBRION=6 or IBRION=8, as the phonon calculation at the Gamma point provides useful informations.
INCAR:
#K-grid
KSPACING = 0.15 #0.15=very fine 0.2=fine 0.3=normal
#start parameters
NWRITE = 1
PREC = Accurate #precision mode
#electronic optimization
ENCUT = 520 #cutoff energy
EDIFF = 1.0e-9 #breakout condition for SC loop
NELMIN = 6 #minimum number of electronic SCF steps
NELM = 100
ALGO = Normal
#ionic relaxation
IBRION = 6 #relaxation method
LEPSILON =.TRUE. #linear response theory
#DOS-related
ISMEAR = 0 #determines how the partial occupancies are set for each orbial
SIGMA = 0.05
#Exchange correlation treatment
GGA = MK
PARAM1 = 0.1234
PARAM2 = 1.0
LUSE_VDW = .TRUE.
AGGAC = 0.0
LASPH = .TRUE.
The Born effective charge tensor appears in OUTCAR file as following:
BORN EFFECTIVE CHARGES (in e, cummulative output)
-------------------------------------------------
ion 1
1 3.89297 0.00000 0.00000
2 0.00000 3.89297 0.00000
3 0.00000 0.00000 3.83680
ion 2
1 7.37749 -0.00000 0.00000
2 0.00000 7.37749 0.00000
3 -0.00000 0.00000 7.03898
ion 3
1 -2.65986 0.00000 0.00000
2 0.00000 -2.65986 -0.00000
3 0.00000 -0.00000 -5.75317
ion 4
1 -6.07061 -0.00000 0.00000
2 0.00000 -2.54000 -0.00000
3 0.00000 -0.00000 -2.56131
ion 5
1 -2.53999 -0.00000 0.00000
2 0.00000 -6.07061 -0.00000
3 0.00000 -0.00000 -2.56131
Using this, we can now set the FORCES_Z tag in the INCAR for contrained-forces calculations. To apply electric field in the z direction, you only need the FORCES_Z tag as only the diagonal componenets in the Born effective charge tensor is present.
We'll start by inducing small polarization to the non-polar structure.
INCAR:
#K-grid
KSPACING = 0.15 #0.15=very fine 0.2=fine 0.3=normal
#start parameters
NWRITE = 1
PREC = Accurate #precision mode
#electronic optimization
ENCUT = 520 #cutoff energy
EDIFF = 1.0e-8 #breakout condition for SC loop
NELMIN = 6 #minimum number of electronic SCF steps
NELM = 100
ALGO = Normal
#ionic relaxation
IBRION = 1 #relaxation method
ISIF = 3 #relax dof
NSW = 500 #maximum number of ionic steps
EDIFFG = -0.001 #break condition for ionic relaxation
POTIM = 0.2
NFREE = 20 #MUST BE PROVIDED for BRION=1, mute for BRION=2
#DOS-related
ISMEAR = 0 #determines how the partial occupancies are set for each orbial
SIGMA = 0.05
#Write flags
#Exchange correlation treatment
GGA = MK
PARAM1 = 0.1234
PARAM2 = 1.0
LUSE_VDW = .TRUE.
AGGAC = 0.0
LASPH = .TRUE.
#Force correction
FORCES_Z = 3.83680 7.03898 -5.75317 2*-2.56131
SCALING = 0.01
LFIX_XY = .TRUE.
#performance optimization
NCORE = 8
Note that positive value of SCALING is used, which will simulate negative electric field. Because PbTiO3 is ferroelectric (and it is in the "negative capacitance" region), it will induce structure with positive polarization and lower energy, compared to your non-polar structure. Always start with small value of SCALING, and use the structure of smaller SCALING as starting point of larger SCALING.
For your POSCAR, you cannot use the CONTCAR of your non-polar structure without setting ISYM=0. But since we know that the space group of the polarized PbTiO3 is P4mm, we'll circumvent this by altering the POSCAR a little bit. In general, it is always good to compare your calculation with a trial calculation of ISYM=0.
POSCAR:
PbTiO3
1.00000000000000
3.9064260257101253 0.0000000000000000 0.0000000000000000
0.0000000000000000 3.9064260257101253 0.0000000000000002
0.0000000000000000 0.0000000000000001 3.9719819950307556
Pb Ti O
1 1 3
Direct
0.0000000000000000 0.0000000000000000 0.0000000000000000
0.5000000000000000 0.5000000000000000 0.5001000000000000
0.5000000000000000 0.5000000000000000 0.0000000000000000
0.0000000000000000 0.5000000000000000 0.4999000000000000
0.5000000000000000 0.0000000000000000 0.4999000000000000
After you converge your calculation, you can check the forces of your structure in the OUTCAR:
POSITION TOTAL-FORCE (eV/Angst)
-----------------------------------------------------------------------------------
0.00000 0.00000 3.97466 0.000000 0.000000 0.003863
1.95321 1.95321 1.98860 0.000000 0.000000 0.007021
1.95321 1.95321 -0.00013 0.000000 -0.000000 -0.005758
0.00000 1.95321 1.98541 0.000000 -0.000000 -0.002563
1.95321 0.00000 1.98541 -0.000000 -0.000000 -0.002563
-----------------------------------------------------------------------------------
total drift: 0.000000 0.000000 0.000034
And you'll see that your forces have converged to the value you set, FORCES_Z*SCALING.