Fundamentally one must compute the delta E at a series of points r1, r2, r3, ... where delta E = E(rn) - E(infinite separation). If you want the LJ potential, then you fit the classic LJ function to your delta E data to determine the sigma, epsilon and req values for the potential function.
These calculations must be done at a high enough level of theory to ensure dispersion is properly calculated in addition to dipole-dipole, dipole-induced dipole and repulsive forces.
One can use DFT with DFT-D3 and a large basis (triple or quad zeta with polarization and diffuse functions) to get the delta E information.
On Mon Jan 5 '15 12:59pm, luca wrote
>I am trying to calculate the lennard jones potential of a zinc2+ ion.
> in order to replicate the work that prof Karplus already done (http://onlinelibrary.wiley.com/doi/10.1002/prot.340230104/abstracthttp://onlinelibrary.wiley.com/doi/10.1002/prot.340230104/abstract) and use it as a benchmark for new calculation.
>Roughly, my idea is to move the water molecule (or the zinc ion) from near to 0.1 angstrom to 10 angstrom. This process results in several jobs (as many as the number of point I want to have). I guess I could do the same, with more accurate results, with a surface scan.
>While it seems that the shape of the Energy vs Distance is near to that showed in the so far discussed paper the values are different, in particular for the energy, that is low and low, while the distance at which the wheel is found is really near.
>To do that I am using the total energy the FF outputs in the log file.
>Do you have some experience with this issue? Could you suggest a strategy to work with?
>Thank you and happy new year!