Since one of the main interests of our group is to understand how the motions of solvent molecules control the dynamics of electron transfer reactions, it makes sense to build computer simulations of the charge transfer process, allowing us to determine the role of every solvent molecule in the reaction.  Although it is generally considered acceptable to treat the solvent classically (there are too many solvent molecules involved to treat the entire problem quantum mechanically), the problem is still difficult because the electron must be treated quantum mechanically:  electron transfer is inherently a quantum phenomenon that cannot be described classically.  What makes this problem really challenging is the fact that the electron transfer processes we study not only take place from electronic excited states, but also are non-adiabatic.  This means that we can't attack this problem using standard quantum chemistry techniques such as density functional theory (DFT) because standard techniques do not provide information about excited-state wave functions.  Moreover, the fact that the reactions we're interested in are non-adiabatic means that we cannot take advantage of the Born-Oppenheimer approximation:  we cannot assume that nuclear motions of the solvent molecules are slow relative to the electronic dynamics.  In fact, it is the nuclear motions of the solvent that mix the electronic energy levels together and induce radationless transitions that allow the system to relax from the electronic excited state.  Fortunately, there are prescriptions in the literature for dealing with the breakdown of the Born-Oppenheimer approximation.  All of these prescriptions make uncontrolled assumptions about the nature of quantum decoherence -- as of the current time, there is no rigorous way to treat non-adiabatic dynamics in mixed quantum/classical systems (for more on decoherence in quantum non-adiabatic molecular dynamics simulations, see B. J. Schwartz et al., J. Chem. Phys. 104(15), 5942 (1996)). In our group, we both take advantage of existing approaches for non-adiabatic dynamics, and we develop new methods. For example, we recently have extended an algorithm developed by others to be capable of calculating non-adiabatic dynamics for multi-electron systems (before our method, such calculations were limited to systems with only a single quantum degree of freedom).  We are now working to test our new method and apply it to study systems such as solvated dielectrons (see Figure above) as well as the alkali metal anions being studied experimentally in our group via femtosecond spectroscopy.  See, e.g.,

R. E. Larsen and B. J. Schwartz, "An Efficient Real-Space Configuration-Interaction Method for Simulation of Multi-Electron Mixed Quantum/Classical Non-Adiabatic Dynamics in the Condensed Phase," J. Chem. Phys. 119(15), 7672-84 (2003).

C. J. Smallwood, W. B. Bosma, R. E. Larsen and B. J. Schwartz, "The Role of Symmetry in Charge-Transfer-to-Solvent Reactions:  Quantum Non-Adiabatic Computer Simulation of Photoexcited Sodium Anions," J. Chem. Phys. 119(21), 11263-77 (2003). 

This work is supported in part by the National Science Foundation under award CHE-0204776.