Water molecules at the intersection of implicit and explicit solvents behave abnormally

[343333333]

Dear developer, I recently use VASPsol to simulate the MD of the interface system. At first the water molecules are behaving normally, but then they start to move towards the vacuum layer(pic below). Is this a normal phenomenon? If that's not the case, then how do I set up the explicit solvent/implicit solvent interaction layer to be more stable. Any suggestions are welcome.

pic: The beginning of the MD

The end of the MD.

And here is my INCAR:

ISTART = 0
ICHARG = 2
INIWAV = 1

ISPIN = 2

ENCUT = 400
NELM = 200
NELMIN = 4
EDIFF = 1E-05
PREC = Normal
LREAL = Auto

for vaspsol

LSOL = .TRUE. EB_k = 78.4 LAMBDA_D_K = 3.04

EDIFFG = -0.02
NSW = 1000
IBRION = 0
ISIF = 2
ISYM = 0
POTIM = 1
TEBEG = 300 TEEND = 300 SMASS = 2 NBLOCK = 1 KBLOCK= 100 MDALGO = 2

ISMEAR = 0
SIGMA = 0.05

ALGO = VERY FAST
LPLANE = .TRUE. LSCALU = .FALSE. NCORE = 4 NSIM = 4
POMASS = 55.847 12.011 14.001 16.000 2.000 IWAVPR=12

LWAVE = .FALSE.
LCHARG = .FALSE.
LVTOT = .FALSE.
LELF = .FALSE.
IVDW= 12

[nkuhuxu]

Dear 欣懋, I have encountered a similar problem. After adding the VASPsol parameter, the water molecules run towards the vacuum layer. Do you know how to fix it or do you have other suggestions? Best wishes Xu Hu

[343333333]

Dear 欣懋, I have encountered a similar problem. After adding the VASPsol parameter, the water molecules run towards the vacuum layer. Do you know how to fix it or do you have other suggestions? Best wishes Xu Hu

Hello Xu Hu, This is because the background charges are added to the explicit aqueous solvent, causing the repulsion between the water molecules to increase. You can look at Figure 1 in the original paper for vaspsol (10.1063/1.4865107), which defines the range at which the background charge is added. Therefore, the solution needs to change this charge density error function, which involves vaspsol source code, so there is no good solution at present. Best wishes, Ben

[nkuhuxu]

Dear 欣懋, I have encountered a similar problem. After adding the VASPsol parameter, the water molecules run towards the vacuum layer. Do you know how to fix it or do you have other suggestions? Best wishes Xu Hu

Hello Xu Hu, This is because the background charges are added to the explicit aqueous solvent, causing the repulsion between the water molecules to increase. You can look at Figure 1 in the original paper for vaspsol (10.1063/1.4865107), which defines the range at which the background charge is added. Therefore, the solution needs to change this charge density error function, which involves vaspsol source code, so there is no good solution at present. Best wishes, Ben

Thanks for your help and explanations, it helps me a lot！