Seite - 95 - in Charge Transport in DNA - Insights from Simulations

7.2Modeling the IonizationPotentials Table7.3: Statistics of the IP of adenine nucleobases in a polyA oligomer. Left: Sim- ulation without QM/MM coupling; only internal modes of vibration con- tribute to the fluctuations. Right: Simulation with QM/MM coupling to the environment; fluctuations of solvent and DNA backbone contribute to thefluctuations of IP. invacuo nucleobase mean± std. dev. [eV] A1 5.354± 0.099 A2 5.356± 0.090 A3 5.357± 0.101 A4 5.357± 0.095 A5 5.357± 0.095 A6 5.359± 0.096 QM/MM nucleobase mean± std. dev. [eV] A1 5.565± 0.304 A2 5.572± 0.303 A3 5.563± 0.301 A4 5.554± 0.302 A5 5.549± 0.304 A6 5.551± 0.305 an additional shift of IP had to be introduced to account for the environmental effect. Then theamplitudesAn of thecosine functionsof theenvironmentalfluctu- ationswerefitted in combinationwith theamplitudes for the internalfluctuations, obtainedpreviously, so that the largerfluctuations of about 0.3 eVobserved in the QM/MMenvironment are reproduced. Additionally, the amplitudes of the cosine functions that correspond to the nine lowest environmentalfluctuationswere scaleddownbya factorof0.6. These slow fluctuations correspond to movements of the whole molecular system, therefore theyhavevery longperiodsand rather lowamplitudes. TheMDsimulationsperformed toobtain the time series of IPwere conductedus- ing anon-polarizable forcefield. In such simulations, the electric field inducedby the environment is overestimated by 26-34%[80], thereforeQM/MMinteractions have to be scaled down by a factor of 1.5 for charge transfer calculations[73]. To model this scaling,MDsimulationswith andwithout this scaling factorwereper- formedand theCTparameters recorded. The time serieswere analyzed formean values and standard deviations. A new set of amplitudes for the environmental fluctuationswas introduced to reproduce these scaledfluctuations. 95