[gmx-users] RF and TI

[Alexey Odinokov]

Dear gmx users,
I calculated the hydration free energy of benzene molecule two times.
First time I used a decoupling method implemented in GROMACS. Second
time I performed thermodynamical integration between different
topologies of the benzene molecule: non-interacting with media and
normal one. All intramolecular interactions were turned off. The
results were different: decoupling method gave energies about 2.8
kcal/mol higher. I investigated the problem and found that the
difference of the <dH/dl> values was due to the Reaction-field term.
When I switched electrostatic to the simple cut-off scheme, decoupling
method gave almost the same result, while for dual topology method the
result increased on 2-3 kcal/mol.
Is the RF method consistent with free energy calculations? Can it be
treated within the frames of decoupling scheme?

Thanks,

Alexey