GROMACS version:
GROMACS modification: Yes/No
Here post your question
Hi my question may be a bit basic.
I am a little confused with the calculation of PME. I know that the short-range Coulombic interaction is dealt with by the summing paired interaction within the cut-off range, and the long-range interaction is taken care of by the PME algorithm. But if my epsilon-rf is set to infinity, known as the tin-foil boundary condition, does it mean the force beyond the cut-off is essentially zero? In this case, the PME is only taken care of the vdw attraction? It seems a little odd to me, as it would imply that two charges won’t see each other if they are separation by the cut-off distance, say 1nm, which is quite small.
Answer by ChatGPT:
The tin-foil boundary condition only affects the Coulombic interactions between particles. Other types of interactions, such as the Lennard-Jones interaction, are still treated using the PME method or some other method.
In the PME method, the Coulombic interactions are treated differently from other interactions. The PME method uses a Fourier Transform to calculate the Coulombic potential at each grid point in the simulation cell, while other interactions are typically treated using a cutoff distance. The Lennard-Jones interaction, for example, is usually treated using a cutoff distance and a smoothing function to smoothly switch off the interaction beyond the cutoff distance.
In summary, the PME method is used to calculate long-range Coulombic interactions within the simulation cell, while the tin-foil boundary condition is used to set the Coulombic interactions between particles beyond the cut-off distance to zero. Other interactions, such as the Lennard-Jones interaction, are treated using a cutoff distance and a smoothing function.
epsilon-rf sets the dielectric outside the cut-off distance for the reaction-field method. PME is a different method. Note that PME also has dieletric constant for the boundary at infinity. This is set using epsilon-surface.
Thank you Hess. Could you help me to understand, if it is correct to say that two charges won’t see each other beyond the cut-off distance if epsilon-surface is set to infinity when using the PME method?
Then what kind of scenario is represented respectively by setting epsilon-surface to zero, infinity, or “the relative permittivity of the imaginary surface around your infinite system” as described by the manual?
Infinity is the default value and recommended for most simulations that are supposed to represent a small part of a system embedded in a much larger system. This gives fluctuations of the charge distribution, for instance the net dipole of the system, as one would see in a small part of a larger system.
Zero is useful for simulations in an electric field. If one puts an external electric field on a system with epsilon-surface=inf, the field is amplified by the dieletric constant of the liquid/solvent. With epsilon-surface=zero there is a penalty for building up a net dipole and the field is what one expect. But the fluctuations are also suppressed.