Dear VASP developers,
I am trying to calculate the standard redox potential of the I3-/I- couple in aqueous solution. My cell contains the iodine species plus 64 water molecules, and I run MD at 300 K with an on-the-fly machine learning force field.
I have two questions about setting up the thermodynamic integration.
First, is VCAIMAGES the right choice here? My two states are the oxidized and reduced forms of a redox couple in a liquid. SCALEE and TILAMBDA both seem to need a harmonic solid or an ideal gas as the reference state, which does not fit a solution, so VCAIMAGES looked like the only option. I would like to confirm that this is correct.
Second, I am confused about the POSCAR files. The wiki says the subdirectories 01 and 02 must contain identical POSCAR files. But I3- and 3I- do not have the same structure - the I-I bonds are present in one state and not in the other. If both subdirectories hold the same coordinates, VASP only ever sees one geometry. Where is the structure of the other state supposed to come in?
I tried putting the reduced-state POSCAR in the parent directory, but the results were exactly the same with and without it, so it seems VASP reads only the subdirectories.
My current setup uses the same oxidized-state POSCAR in both 01 and 02, and the two images differ only in NELECT (534 for I3-, 536 for 3I-). Is this the intended way to do it? Or should 01 hold the oxidized structure and 02 the reduced one?
I have attached the INCAR I use for the TI run, and the paper my setup is based on is: Jinnouchi, R., Karsai, F. & Kresse, G. Machine learning-aided first-principles calculations of redox potentials. npj Comput Mater 10, 107 (2024). https://doi.org/10.1038/s41524-024-01295-6
# Thermodynamic Integration
VCAIMAGES = 0.2
NCORE_IN_IMAGE1 = 32
PREC = Normal
ALGO = Normal
LREAL = Auto
ISYM = 0
GGA = RP
IVDW = 12
IBRION = 0
ISIF = 2
NSW = 5000
POTIM = 0.5
TEBEG = 300
TEBEG = 300
MDALGO = 3
LANGEVIN_GAMMA = 10.0 10.0 10.0
NCORE = 4
LWAVE = .FALSE.
LCHARG = .FALSE.
# IMAGE_1: I3- (oxidized, 01/)
IMAGE_1 {
ENCUT = 500
ISMEAR = 0
SIGMA = 0.05
EDIFF = 1E-5
NELECT = 534
ML_LMLFF = .TRUE.
ML_ISTART= 0
}
# IMAGE_2: 3I- (reduced, 02/)
IMAGE_2 {
ENCUT = 500
ISMEAR = 0
SIGMA = 0.05
EDIFF = 1E-5
NELECT = 536
ML_LMLFF = .TRUE.
ML_ISTART= 0
}
Thanks very much for any advice.