 # Short range LJ potential of TIP3P water molecules

GROMACS version: 16
GROMACS modification: Yes/No
I simulated a water box with TIP3P water model and run a short production run. Though the whole system energy came (-)ve but the LJ potential came positive as I did the energy calculation using ‘gmx energy’. Why did I get the LJ potential (+)ve?

The largest contributor to water-water interactions is electrostatics, which are overwhelmingly favorable due to hydrogen bonding. Decomposing energies between the individual contributions (Coulomb and LJ) is not useful because no force field parametrizes these quantities to be meaningful in any way.

Thank you sir for your kind reply. Yes, total energy of the system came in right trend. Can you please refer me the methodology, how the total energy (Coulomb + LJ) is distributed to those indivisual contributions?

I don’t understand the question. The potential energy of the system is simply the sum of the bonded and nonbonded energies. My point was that trying to draw some deeper understanding from LJ and Coulombic contributions individually has no basis in physical reality as the individual terms are no parametrized specifically against any empirical or theoretical target data (because it doesn’t exist).

Thank you sir. Yup, my doubt was, does the ‘gmx energy’ use any other algorithm to distribute the total energy in LJ and Coulombic contribution or it just gives the energy calculated for LJ and Coulombic potentials from the FF equation. Now I got your answer that ‘gmx energy’ gives the calulated LJ and Coulombic potentials, calculated from the FF equation, separately. However, we can’t draw any conclussion from LJ and Coulombic potentials separately, as we parametrized the whole energy of the system, not the indivisual contributions.

Is there any way to define a bond as ionic or covalent, in GROMACS. As we are dealing with Molecular Mechanics, instead of Quantum Mechanics, I couldn’t find a way to differenciate the bonding nature.
My next question is, can I calcualte the contribution of H-bonding energy (because we can calculate the number of H-bonds)?

No. The only thing that matters are bonds as defined in the topology, with associated force constants and equilibrium lengths. There is no concept of different bond types, hybridization, etc. in classical molecular mechanics.

There is no such energy term in any modern force field. You can calculate an interaction energy, but you cannot assign that exclusively to the effects of hydrogen bonding.

Thank you sir, for your valuable response.