High pressure in x and y axis in a buble in a nano channel

User discussions
, ,
1

GROMACS version: gromacs/2024.4
GROMACS modification: No
Hi everyone,

I am simulating a mixed H₂O/H₂ system confined in an Al₂O₃ nanochannel. The channel walls lie in the X–Y plane, and the confinement direction is along Z.

I would like to apply anisotropic pressure coupling such that the pressure inside the nanochannel is 25 bar in the Z direction and 1 bar in the X and Y directions.

; Pressure coupling
pcoupl = Parrinello-Rahman ; Pressure coupling method
pcoupltype = semiisotropic ; Semiisotropic pressure coupling
tau_p = 2 ; Pressure coupling time constant (ps)
ref_p = 1 25.0 ; Reference pressures: 1 bar in x/y, 25.0 bar in z
compressibility = 4.5e-20 4.5e-5 ; Compressibility: near-zero in x/y, standard in z
nstpcouple = 20 ; Frequency for pressure
refcoord_scaling = COM

After equilibration, the pressure along the Z direction converges to 25 bar, which is the expected behavior.
However, the pressures along the X and Y directions increase to very large values, as shown below:

Pressure (bar)
3.02652e+04 -6.38528e+00 -8.65890e-01
-6.38820e+00 3.01417e+04 1.50278e+01
-8.57413e-01 1.50357e+01 2.37052e+01

Can someone please help me understand the origin of these high pressures in the X and Y directions?
Is this behavior expected for a nanochannel system under periodic boundary conditions?

Thank you in advance for your time and help.

2

Hi,

compressibility = 4.5e-20 4.5e-5 ; Compressibility: near-zero in x/y, standard in z

I am not really sure how one could impose any meaningful pressure in with such a small compressibility. It’s basically a NPTA (constant area) ensemble. Have you tried increasing the compressibility in x/y (assuming it’s what you actually want to do)?

P.S. the high values can be also due to the fact that walls are frozen/restrained (are they)?

3

You can not use pressure coupling, which acts on the whole system, to enforce a pressure in part of the system, in particular when there are solid structures involved. Furthermore, I don’t understand what you want to achieve. A gas, or a liquid, can not have different pressures along different dimensions.