Actually, the first minimum at 1.0 nm is not a well-defined minimum. Often on the rugged free energy surface such a 'well' with a depth of 1-2 kcal/mol shows just local fluctuation but not a meta-stable state.

For your system, what you should do, from my perspective, is run conventional MD rather than constructing a PMF. If you see your system reaches the 2.7 nm position in very short simulations (e.g. several ns), then 2.7 nm is the most stable state. If this is not expected from a theoretical point of view, then something is wrong with your modeling. Here, I mean the Hamiltonian (the force field).

If the conventional simulation does stay at 1.0 nm, then the way you construct the PMF is wrong. You should check the protocol in details.

As far as I see, there is no potential well in your plot. The PMF seems OK.

Thanks a lot for the answer Zhaoxi Sun, but I wasn't expecting the second minima between 2.5-3 nm... I was expecting the curve to flatten out after the first and smaller minima at around 1 nm, similar to a Lennard-Jones type curve. Any ideas on why we see the second minima and the curve doesn't flatten out?

Actually, the first minimum at 1.0 nm is not a well-defined minimum. Often on the rugged free energy surface such a 'well' with a depth of 1-2 kcal/mol shows just local fluctuation but not a meta-stable state.

For your system, what you should do, from my perspective, is run conventional MD rather than constructing a PMF. If you see your system reaches the 2.7 nm position in very short simulations (e.g. several ns), then 2.7 nm is the most stable state. If this is not expected from a theoretical point of view, then something is wrong with your modeling. Here, I mean the Hamiltonian (the force field).

If the conventional simulation does stay at 1.0 nm, then the way you construct the PMF is wrong. You should check the protocol in details.

At least, you should show the uncertainty from your PMF to prove that in 1.0nm is minimum