We present a molecular model for studying the prototypical ferric–ferrous electron transfer process in liquid water, and we discuss its structural implications. Treatment of the nonequilibrium dynamics will be the subject of future work. The elementary constituents in the model are classical water molecules, classical ferric ions (i.e., Fe3+ particles), and a quantal electron. Pair potentials and pseudopotentials describing the interactions between these constituents are presented. These interactions lead to ligand structures of the ferric and ferrous ions that are in good agreement with those observed in nature. The validity of the tight binding model is examined. With umbrella sampling, we have computed the diabatic free energy of activation for electron transfer. The number obtained, roughly 20 kcal/mol, is in reasonable accord with the aqueous ferric–ferrous transfer activation energy of about 15 to 20 kcal/mol estimated from experiment. The Marcus relation for intersecting parabolic diabatic free energy surfaces is found to be quantitatively accurate in our model. Due to its significance to future dynamical studies, we have computed the tunnel splitting for our model in the absence of water molecules. Its value is about 1 kB T at room temperature for ferrous–ferric separations around 5.5 Å. This indicates that the dynamics of the electron transfer are complex involving both classical adiabatic dynamics and quantal nonadiabatic transitions. The dynamics may also be complicated due to glassy behavior of tightly bound ligand water molecules. We discuss this glassy behavior and also describe contributions to the solvation energetics from water molecules in different solvation shells. Finally, the energetics associated with truncating long ranged forces is discussed and analyzed.

Path integral Monte Carlo methods are applied to the ST2 model of water to examine both the importance of quantum effects on the structure of liquid water and the differences in the structures of light and heavy water. Significant differences are apparent among the results for classical water, quantum H2O, and quantum D2O. For all quantities considered, the classical liquid is the most structured while quantum H2O is the least structured. The implications of these results for neutron and x-ray diffraction experiments are examined; of the three atom–atom partial structure factors, the oxygen–oxygen function is found to be the most sensitive to quantum effects and the structural differences between light and heavy water appear to be large enough to be measured by x-ray diffraction experiments.

By computer simulation and also by analytical methods we have computed the nuclear tunneling enhancement of the rate for ferrous–ferric exchange in water. The model we have examined is the one studied earlier where we treated water as a classical liquid [R. A. Kuharski, J. S. Bader, D. Chandler, M. Sprik, M. L. Klein, and R. W. Impey, J. Chem. Phys. 89, 3248 (1988)]. But now we treat water quantum mechanically and find that the tunneling enhancement is a factor of 60 in the rate constant. Further, as observed experimentally, we find that the isotope shift on the rate when changing from D2O to H2O is approximately a factor of 2. The computer simulation aspects of these calculations employ path integral methods and a novel partitioning of the free energy associated with electron transfer. Our results show that it is insufficient to quantize only the atoms composing the ligands. The quantum dynamics of water molecules beyond the first solvation shell prove quite significant.

No abstract

No abstract