Solvent effects on electronically excited states

Theoretical modeling of solvation has received strongly increased attention in recent years. Such investigations give a deeper insight into the interactions between solute and solvent molecules and can help in the understanding of solvation phenomena. However, the calculation of solvation effects is still a considerable challenge in quantum chemistry. The simplest way to proceed is to regard the solute-solvent system as a "supermolecule" and to compute solute-solvent clusters with an increasing number of solvent molecules. In this way specific interactions between solute and solvent, like hydrogen bonds, can be taken into account very well. However, this approach reaches very soon its limits due to intractable amounts of computer time needed in such calculations for larger clusters. In order to circumvent these problems, continuum models have been introduced into the quantum chemical methodology. One of the most popular approaches nowadays is the polarizable continuum model (PCM) developed by Tomasi and coworkers. The advantages of this model are: (i) the solute molecule is treated by quantum chemical methods, (ii) the shape of the cavity is modeled quite realistically by a set of interlocking spheres around the nuclear centers, and (iii) the polarized medium is described by a set of image charges located at the molecular surface. Most of the application of PCM (and of related models) have been made for the electronic ground state using SCF or SCF-type methods. The use of continuum methods for electronic excitation processes is more involved. The major complication comes from the fact, that the quantum chemical calculation of excited states is considerably more difficult than that of the ground state. So far, applications have been limited mostly to cases easily accessible to quantum chemical calculations, such as the n-pi* transitions in carbonyl compounds. The aim of the present project is to combine the methodology for continuum solvation with the methods for high- level ab initio computation of excited states as has been developed within the COLUMBUS program system. For the quantum chemical part of the calculation the MR-CISD and MR-AQCC methods shall be used. Especially the latter is of great interest since size-extensivity corrections for multi-reference wave functions are available in this way. Moreover, these methods are very stable and robust and can take into account very well the variable multi- reference character of wave functions and the interactions between different states as they occur e.g. in avoided crossings. It is our plan to investigate solvation effects on various types of electronic excitations beyond the standard n-pi* excitations in carbonyl compounds. Most interesting cases are pi-pi* and Rydberg excitations. Especially the diffuse Rydberg states are strongly affected by solvent interactions due to the confinement of the solute molecule in the solvent cavity. A first solvation shell will be constructed by including water molecules explicitly into the calculation. Additional solvation effects can be computed by means of continuum models. Our calculations should provide valuable insight into the details of the shrinking of the Rydberg orbitals, on the relative ordering of states and on avoided crossings. Protonation- and proton transfer processes in excited states have experienced increasing interest in recent years. Quantum chemical calculations for these processes are very difficult and only at the beginning. We want to start to study protonation processes in excited states and the influence of solvation on various tautomeric forms of protonated formaldehyde, formamide, alanine and related compounds. As a final example we want to investigate solvation effects on the intramolecular proton transfer of malonaldehyde in excited states

Solvation effects and environmental interactions in general play an important role for the properties and the reactivity of molecular and biomolecular systems. In the present project methods for the theoretical calculation of solvation effects in excited molecular states have been developed. The goal is to construct comprehensive modelling schemes for the simulation of photochemical and photobiological processes. Prominent practical examples are proton transfer processes in excited states, which are of utmost importance for the photostability of DNA bases, and molecular rearrangements (cis-trans isomerization) in the excited state for retinal, the chromophore of rhodopsin, which are relevant for the primary process of vision. Whereas the quantum chemical calculation of molecular ground-state properties can be routinely performed nowadays, the calculation of excited states and of chemical reaction profiles is still very challenging. Our work was primarily based on COLUMBUS, a quantum chemical public domain program package, which has been developed in our group in cooperation with colleagues from the USA. The computational methods used in COLUMBUS are particularly well suited for such difficult situations as the afore-mentioned electronic excitations. Solvation effects have been implemented into COLUMBUS via the polarizable continuum solvation model COSMO and applied to several interesting problems. By using a pure continuum model, specific interactions (mainly hydrogen bonds) are not taken into account. Semi-continuum models can be constructed where a few solvent molecules are included explicitly and only the remaining interactions are treated via the continuum model. In this way retinal models are being investigated at present. Since the methods used in COLUMBUS are very time-consuming, we were using other, less expensive methods (e.g. coupled cluster and time-dependent density functional theory) for the description of larger molecular systems. Within this approach the absorption and fluorescence properties of 7-hydroxy-4-methylcoumarin, an important biologically active compound, have been investigated by semi-continuum methods and very interesting aspects of proton transfer along water wires have been observed.