We present a comprehensive
analysis of the interplay between the
choice of an electronic structure method and the effect of using polarizable
force fields vs. nonpolarizable force fields when calculating solution-phase
charge-transfer (CT) rates. The analysis is based on an integrative
approach that combines inputs from electronic structure calculations
and molecular dynamics simulations and is performed in the context
of the carotenoid–porphyrin–C60 molecular
triad dissolved in an explicit tetrahydrofuran (THF) liquid solvent.
Marcus theory rate constants are calculated for the multiple CT processes
that occur in this system based on either polarizable or nonpolarizable
force fields, parameterized using density functional theory (DFT)
with either the B3LYP or the Baer–Neuhauser–Livshits
(BNL) density functionals. We find that the effect of switching from
nonpolarizable to polarizable force fields on the CT rates is strongly
dependent on the choice of the density functional. More specifically,
the rate constants obtained using polarizable and nonpolarizable force
fields differ significantly when B3LYP is used, while much smaller
changes are observed when BNL is used. It is shown that this behavior
can be traced back to the tendency of B3LYP to overstabilize CT states,
thereby pushing the underlying electronic transitions to the deep
inverted region, where even small changes in the force fields can
lead to significant changes in the CT rate constants. Our results
demonstrate the importance of combining polarizable force fields with
an electronic structure method that can accurately capture the energies
of excited CT states when calculating charge-transfer rates.