Computational Photochemistry and Photobiology

In the last years, Computational Chemistry has made dramatic progress due to combined developments in computational methods and in computer technology. Computer simulations based on quantum chemical methods have reached an accuracy, which, in many cases, is competitive to experiment. The field of photochemistry and photobiology is fascinating since it combines a set of challenging theoretical questions with interesting chemical problems. The quantum chemical description of excited states is significantly more demanding than that for the ground state. Key features of excited-state energy surfaces are conical intersections between different states. In the region of conical intersections the Born-Oppenheimer approximation breaks down, which is a main fundament of Quantum Chemistry. These intersections are also of great mechanistic importance since they form funnels through which ultra-fast radiationless transfer to the other state occurs. In the last years we have made substantial progress in developing methods for the analytic computation of energy gradients for excited states and for analytic nonadiabatic coupling vectors based on the multireference configuration interaction (MRCI) method. The availability of analytic gradients and coupling vectors is of utmost importance for the efficient determination of stationary points and minima on the crossing seam. The methods we have implemented into our COLUMBUS program system are worldwide unique features, which enhance the capabilities of excited-state calculations drastically. The power of our methods has been demonstrated by a large number of applications. It is the aim of the present project to use the just-mentioned computational possibilities in order to tackle outstanding problems in photochemistry and photobiology and to perform realistic simulations based on reliable excited-state energy surfaces. Two important classes of applications of photochemical/photobiological relevance will be investigated. The first one is the photodynamics of conjugated pi systems with particular emphasis on protonated Schiff bases as models for retinal and the primary process of vision. The second class of applications refers to intra- and intermolecular proton/hydrogen transfer and hydrogen detachment processes in excited states. These processes are of importance for the photostability of organic compounds, particularly for DNA bases. For the construction of excited-state energy surfaces we plan to use a combined approach of creating accurate benchmark results based on our MRCI methods and assessing other, less time-consuming approaches. For the latter, mainly, the TDDFT and RI-CC2 methods will be used. In addition to the accurate calculation of energy surfaces, dynamics simulations based on these surfaces constitute one key issue of our project. We plan to continue our strategy of adjusting parameters of semiempirical methods (AM1) to our MRCI results and use those for efficient surface-hopping dynamics studies. Furthermore, solvation and environmental effects in general will be computed in excited states using various approaches starting from explicit inclusion into the quantum chemical calculation via QM/MM techniques to continuum solvation. In particular, the sequential Monte Carlo Quantum Mechanics (S-MC/QM) approach will be applied in order to combine the advantages of creating statistically relevant ensembles and the accuracy of ab initio calculations.

Photochemical processes represent an extremely fascinating topic since they deviate considerably and many times quite unexpectedly from conventional chemical reactions in the ground state. A particular interesting question refers to the photostability of certain classes of substances, a property which protects molecules from dangerous destruction due to the absorption of the UV light of the sun. The genetic material of the DNA is especially susceptible to such damage since its nucleobases absorb UV light extremely well. Without shielding mechanisms a rapid degradation would be the consequence. How do these protective mechanisms look like? The answer to this question was one of the major topics of the present research project. The goal was to develop and apply theoretical simulation techniques which allow a vivid dynamical picture of the molecular processes involved. In order to achieve this goal several conditions had to be met and many preparations to be performed. The quantum chemical calculation of electronic excitation processes and the interaction of different electronic states are very involved and extremely challenging. Special methods and program packages have to be used. The COLUMBUS program system developed in our group provides excellent conditions for these calculations. For the dynamics simulations a special program package - NEWTON-X - was developed which used the mixed classical-quantum mechanical approach of surface hopping. In combination with COLUMBUS outstanding simulation of the photodynamical evolution of molecular systems could be performed. One of the major objects of these simulations were the afore-mentioned DNA/RNA nucleobases. For the first time, a comprehensive overview of the actually occurring photodynamical decay mechanisms could be given. It was found that there were two groups of nucleobases showing significantly different behavior. One group, the purine bases adenine and guanine, displayed a straightforward decay leading from the excited pi-pi* state back to the ground state in only 300 to 600 femto seconds. The pyrimidine bases thymine, uracil and cytosine showed a more complex behavior with trapping in local minima of excited states. But also in these cases the decay rates were ultrafast with decay times up to 5 pico seconds. This ultrafast decay is the core of the photostability since in this way the electronic excitation energy is converted directly into heat, and the potential damage can be avoided. Beyond simulation of the properties of isolated molecules, interactions with the environment (solvent or biological systems) were investigated. A quantum mechanical/molecular mechanics formalism was developed which allows efficient treatment of DNA and protein environments.