[Help] Seeking MDP files for graphene oxide / phenol / water adsorption system

User discussions
1 


Dear GROMACS community,

I am currently working on a molecular dynamics simulation of phenol 
adsorption on graphene oxide (GO) in aqueous solution using GROMACS 2025.1.

## System Description

- Graphene oxide (GO) sheet: modeled as an infinite periodic sheet 
  in the XY plane, frozen during simulation (freezegrps)
- Adsorbate: phenol molecules
- Solvent: TIP3P/SPC water
- Force field: OPLS-AA (or CHARMM36, depending on suggestions)
- Box type: rectangular, periodic boundary conditions in XYZ
- Temperature: 298.15 K
- Pressure: 1.0 bar

## Simulation Protocol

I am following the standard 4-step protocol:
1. Energy minimization (EM)
2. NVT equilibration (100 ps)
3. NPT equilibration (1 ns)
4. Production MD (100 ns)

## Current Issues

I am experiencing simulation instability during the NPT equilibration 
phase. Specifically:

1. LINCS constraint warnings appear from ~3.6 ps, with bond lengths 
   exploding (e.g., 0.108 nm → 11.95 nm for atoms in the phenol molecule)

2. Repeated pressure scaling warnings:
   "Pressure scaling more than 1%. This may mean your system is 
   not yet equilibrated."

3. The simulation eventually crashes with a CUDA illegal memory 
   access error (CUDA error #700)

I suspect the instability is related to:
- Using Parrinello-Rahman barostat during equilibration (now switched 
  to Berendsen)
- Possible issues with periodic_molecules = yes setting
- Force field compatibility between GO, phenol and water

## Current MDP Settings (NPT, simplified)

integrator       = md
dt               = 0.002
pcoupl           = Berendsen
pcoupltype       = semiisotropic
tau_p            = 5.0
ref_p            = 1.0 1.0
compressibility  = 0 4.5e-5
constraints      = h-bonds
lincs_iter       = 2
lincs_order      = 6
freezegrps       = g
freezedim        = Y Y Y
periodic_molecules = yes

## Questions

1. Could anyone share a working set of MDP files (em / nvt / npt / md) 
   for a similar graphene or graphene oxide + small molecule + water 
   system?

2. Is semiisotropic pressure coupling appropriate for a frozen GO sheet 
   system? Or should pressure coupling be disabled entirely when using 
   freezegrps?

3. Is it recommended to use position restraints instead of freezegrps 
   for the GO sheet? If so, what force constant is typically used?

4. Any recommendations on the appropriate force field combination for 
   GO + phenol + TIP3P water?

## System Details

- Total atoms: ~20,000–30,000
- GO atoms: 1,008 (frozen)
- Water molecules: ~8,000
- Phenol molecules: several tens
- GROMACS version: 2025.1
- GPU: NVIDIA (CUDA enabled build)
- OS: Windows 11 + WSL2

Any advice, shared MDP files, or references to published simulation 
parameters would be greatly appreciated.

Thank you very much for your time and help!

Best regards