GROMACS version: 2024.3
GROMACS modification: No
My system is composed by two ionic surfactants and their counterions soaked in water to form micelles. I’ve performed MDs to generate the micelles, and once in equilibrium, I’m removing the water molecules and switching on BDs. It’s worth mentioning that I won’t use the word “molecule” to refer to the “components” of the system, because strictly neither the surfactant ions nor their counterions are molecules, and water molecules are already neglected.
I’ve carried out several trials to understand how the BD algorithm works in GROMACS. I’ve noticed that it performs a kind of stochastic temperature coupling so that the temperature will fluctuate around the reference temperature (ref_t) as far as the time step (dt) is small enough to reproduce the correct physical trajectory. In contrast, if larger, the temperature will boost up. This also depends on either the friction coefficient (bd-fric) or the time constant (tau_t) chosen.
*bd-fric=constant for all groups of atoms.
*bd-fric=0 then need to choose a proper tau_t.
I calculated manually the friction coefficient for each component as a whole (ionic surfactant and counterion) from the hydrodynamic radius extracted from the .pdb files, or from the literature, and using the Stokes-Einstein equation; but noticed from the trials, and afterwards reading from the GROMACS manual for BDs, that the friction coefficient isn’t for each component as a whole, but more fundamentally for each group of atoms forming the components, isn’t it?
Apart from corroborating the previous statement, I would like to know:
*If choosing bd-fric=constant, will the mass of each atom (in amu units) somehow be taken into account? Or the dynamics for all the particles will be evenly mediated by this magnitude.
*If bd-fric=0, how should I choose tau_t? As I already mentioned, I calculated the friction coefficient for every component as a whole, and thus tau_t=mass/bd-fric. But unless I’m mistaken, this is not how GROMACS works, and the mass isn’t the component’s mass, but each atom’s mass. Besides, I think I should use the atomic radius, instead of the hydrodynamic radius of the components as a whole.
In my view, in case the former option isn’t discriminating between species, the latter should be more accurate, considering that each size and mass should influence in the friction each particle is undergoing due to the solvent. Thus, I also found in the GROMACS manual that tau_t can be estimated from total heat capacity of the system, and a time ratio of 3 for water. However, this formulation is for temperature coupling, and i don’t know if this is valid for this kind of stochastic temperature coupling. Perhaps it would be better if I could choose a friction coefficient for each component, but I don’t think this is possible.