Describing the light-to-energy conversion in OSCs requires a multiscale understanding of the involved optoelectronic processes, i.e., an understanding from the molecular, intermolecular, and aggregate perspective. This thesis presents such a multiscale description to provide insight into the processes in the vicinity of the organic::organic interface, which are crucial for the overall performance of OSCs. Light absorption, exciton diffusion, photoinduced charge transfer at the donor-acceptor interface, and charge separation are included. In order to establish structure-property relationships, a variety of different molecular p-type semiconductors are combined at the organic donor-acceptor heterojunction with fullerene C60, one of the most common acceptors in OSCs. Starting with a comprehensive analysis of the accuracy of diverse ab initio, DFT, and semiempiric methods for the properties of the individual molecules, the intermolecular, and aggregate/device stage are subsequently addressed. At all stages, both methodological concepts and physical aspects in OSCs are discussed to extend the microscopic understanding of the charge generation processes.

Two thematic complexes were addressed within this work. One part is related to improvements and new implementations into the CAST program package. Thereby the main focus laid on the delivery of a tool which can be used to characterize complex reactions and their mechanisms. But also within the new force field (FF) method (SAPT-FF) within the CAST program, several improvements were made. The second topic is related to the description of dye molecules and their spectral properties. The main focus within these studies was set on the influence of the environment on these properties. In the first topic improvements of the local acting NEB (nudged elastic band) methods were included and the number of available methods was extended. The initial pathway generation was improved by implementing the IDPP (image dependent pair potential) method and a new method was implemented for describing temperature dependent pathways. Additionally, improvements have been made to the optimization routines (global NEB). As a second part the Pathopt (PO) method was considerably improved. In the beginning of the work the original PO idea was used. In this approach one starts with a global optimization on one n-1 dimensional hyperplane which divides the reaction into two sub-areas for obtaining guesses of TSs (transition states). These found TS guesses were used to optimize to the ”true” TS. Starting from the optimized ones a relaxation to the next connected minima is done. This idea has been automatically implemented and extended to several number of hyperplanes. In this manner a group of pathsegments is obtained which needs to be connected, but within this work it was realized that such a procedure might be not very efficient. Therefore, a new strategy was implemented which is founded on the same constrained global optimization scheme (MCM) for which the user defines the number of hyperplanes generated. The number of such generated hyperplanes should be large enough 134 to describe the space between the concerning reactants in a sufficient way. The found minima are directly used to built up the reaction pathway. For this purpose a RMSD (root mean square deviation) criterion is used to walk along ways of minimal change from one to another hyperplane. To prove the implementations various test calculations were carried out and extensions included to prove the capabilities of the new strategy. Related to these tests a new strategy for applying the move steps in MCM (Monte Carlo with minimization) was realized which is also related to the question of the coordinates representation. We were able to show that the hopping steps in MCM can be improved by applying Cartesian steps in combination of random dihedral moves with respect to the constraint. In this way it was possible to show that a large variety of systems can be treated. An additional chapter shows the improvements of the SAPT-FF implementation and related test cases. It was possible to treat benzene dimer and cluster systems of different sizes consistently also in accordance with high level ab initio based approaches. Furthermore, we showed that the SAPT-FF with the right parameters outperforms the standard AMOEBA implementation which is the basis of the SAPT-FF implementation. In the last three chapters deal with the description of perlyene-based dyes. In the first smaller chapter ground state chemistry description of macro cycles of PBI (perylene bisimide) derivatives were investigated. Therefore, AFM (atomic force microscopy) based pictures were explained within our study. The methods to explain aggregation behavior in dependency of the ring size were MD simulations and configuration studies. The last two chapters deal with opto-electronic or photo-physical properties of PBI and PTCDA (perylene-3,4,9,10-tetracarboxylic dianhydride). In detail, we investigated the role of the environment and the aggregate or crystal surrounding by applying different models. In that way implicit and explicit solvation models, the size of aggregates and vibration motions were used. In the case of PBI the recent work is found on preliminary studies related to my bachelor thesis and extends it. It was shown that the direct influence of a polarizable surrounding, as well as explicit inclusion of solvent molecules on the overall description of the excitations and nature of the excited states is weaker as one might expect. However the inclusion of intra-molecular degrees of freedom showed a stronger influence on the state characteristics and can induce a change of the order of states within the dimer picture. For the PTCDA molecule the main focus was set on the description of the absorption spectrum of crystalline thin films. Related to this older works exist which already gave a description and assignment of the absorption band, but are based on different approaches compared to the one used in this work. We used the supermolecule ansatz, whereas the environment and different aggregate sizes were investigated. Within the dimer based approach we were able to show that using continuum solvation (IEFPCM/COSMO) based description for the environment the relative order of states remains unchanged. Similar to the PBI calculations the influence of the vibrational motions /distortions is larger. The simulation of the crystal environment by using QM/MM (quantum mechanics/molecular mechanics) approaches delivered that an asymmetric charge distribution might induce a localization of the excitation and a stronger mixing of states. For obtaining further insights we go beyond the dimer picture and aggregates of different sizes were used, whereas the simulations up to the octadecamer mono- and even dual-layer stack were carried out. Within these calculations it was shown that the H-coupling is dominating over a weaker J-coupling between different stacks. Additionally the calculations based on DFT (density functional theory) and semi-empirics showed that the lowest state in terms of energy are mostly of Frenkel type, whereas the higher lying states are CT ones which mix with embedded Frenkel type states. The first band of the absorption spectrum was explained by inclusion of vibrational motions within the stacks which induce an intensity gain of the first excited state. This intensity was not explainable by using the undistorted stacks. Also relaxations at the crystal surface might play a role, but are experimentally not explainable.

In this work, biradical boron containing systems with various structures are investigated to reveal the dependency of the biradical character on the ligated carbene (NHC, CAAC) and the related steric demands of the substituents.

The scope of computational chemistry can be broadened by developing new methods and more efficient algorithms. However, the evaluation of the applicability of the methods for the different fields of chemistry is equally important. In this thesis systems with an unusual and complex electronic structure, such as excitonic states in organic semiconductors, a boron-containing bipolaron and the excited states of pyracene were studied and the applicability of the toolkit of computational chemistry was investigated. Concerning the organic semiconductors the focus was laid on organic solar cells, which are one of the most promising technologies with regard to satisfying the world's need for cheap and environmentally sustainable energy. This is due to the low production and material costs and the possibility of using flexible and transparent devices. However, their efficiency does still not live up to the expectations. Especially the exciton diffusion lengths seem to be significantly too short. In order to arrive at improved modules, a fundamental understanding of the elementary processes occurring in the cell on the molecular and supramolecular level is needed. Computational chemistry can provide insight by separating the different effects and providing models for predictions and prescreenings. In this thesis, the focus was laid on the description of excitonic states in merocyanines and perylene-based dyes taking the influence of the environment into account. At first, the photochemical isomerization between two configurations of 6-nitro BIPS observed experimentally was studied by first benchmarking several functionals against SCS-ADC(2) in the gas phase and subsequently calculating the excited-state potential energy surface. The geometries obtained from a relaxed scan in the ground state as well as from a scan in the excited state were used. The environment was included using different polarizable continuum models. It was shown that the choice of the model and especially the question of the state specificity of the approach is of vital importance. Using the results of the calculations, a two-dimensional potential energy surface could be constructed that could be used to explain the experimental findings. Furthermore, the importance of the excited-state isomerization as a potential deactivation channel in the exciton transport was pointed out. Then the assessment of the suitability of different merocyanines for optoelectronic applications with quantum-chemical methods was discussed. At first, the effect of the environment on the geometry, especially on the bond length alternation pattern, was investigated. It was shown that the environment changes the character of the ground-state wave function of several merocyanines qualitatively, which means that the results of gas-phase calculations are meaningless - at least when a comparison with solution or device data is desired. It was demonstrated that using a polarizable continuum model with an effective epsilon, a qualitative agreement between the calculated geometry and the geometry in the crystal structure can be obtained. Therefore, by comparing the bond length alternation in solution and in the crystal, a rough estimate of the effect of the crystal environment can be made. It was further shown that the connection between the HOMO energy and the open-circuit voltage is not as simple as it is often implied in the literature. It was discussed that it is not clear whether the HOMO of a single molecule or a $\pi$-stack containing several monomers should be used and if the environmental charges of the bulk phase or the interface should be included. Investigating the dependence of the HOMO energy on the stack size yielded no definitive trend. Furthermore, it was discussed that the effect due the optimization of the modules (solvent, bulk heterojunction) during the production masks any potential correlation between the HOMO energy and measured open-circuit values. Therefore, a trend can only be expected for unoptimized bilayer cells. It was concluded that ultimately, the importance of the HOMO energy should not be overestimated. The correlation between the exciton reorganization energy and the so-called cyanine limit, which is predicted by a simple two-state model, was also discussed. By referring to the results of VB calculations, it was discussed that the correlation indeed exists and is non-negligible, although the effect is not as strong as one might have expected. In this context, a potential application of a VB/MM approach was covered briefly. The importance of the molecular reorganization energy and the device morphology was also discussed. It was concluded that the optimization of merocyanines for organic optoelectronic devices is inherently a multiparameter problem and one cannot expect to find one particular parameter, which solely controls the efficiency. The perylene-based dyes were studied with a focus on the description of a potential trapping mechanism involving an intermolecular motion in a dimer. The aim was to find methods which can be applied to larger model systems than a dimer and take the effect of the environment into account. As a test coordinate the longitudinal shift of two monomers against each other was used. At first, it was demonstrated how the character of an excited state in a dimer can be defined and how it can be extracted from a standard quantum-chemical calculation. Then several functionals were benchmarked and their applicability or failure was rationalized using the character analysis. Two recipes could be proposed, which were applied to a constraint optimization (only intermolecular degrees of freedom) in the excited states of the PBI dimer and to the description of the potential energy surfaces of ground and excited states along a longitudinal displacement in the perylene tetramer, respectively. It was further demonstrated that the semi-empirical OMx methods fail to give an accurate description of the excited-state potential energy surfaces as well as the ground-state surface along the test coordinate. This failure could be attributed to an underestimation of overlap-dependent terms. Consequently, it could be shown that the methods are applicable to large intermolecular distances, where the overlap is negligible. The results of DFT calculations with differently composed basis sets suggested that adding an additional single p-function for each atom should significantly improve the performance. QM/MM methods are ideally suited to take the effect of the environment on a a dimer model system into account. However, it was shown that standard force fields also give an incorrect description of the interaction between the monomers along the intermolecular coordinate. This failure was attributed to the isotropic atom-atom interaction in the repulsion term of the Lennard-Jones potential. This was corroborated using two simple proof-of-principle anisotropy models. Therefore, a novel force field called OPLS-AA_O was presented that is based on OPLS-AA, but uses an anisotropic model for the repulsion. The model involves the overlap integral between the molecular densities, which are modeled as a sum of atom-centered p-type Gaussian functions. It was shown that using this force field an excellent agreement with the DFT results can be obtained when the correct parameters are used. These parameters, however, are not very generalizable, which was attributed to the simplicity of the model in its current state (using the same exponential parameter for all atoms). As a short excursion, the applicability of an MO-based overlap model was discussed. It was demonstrated that the repulsion term based on the density overlap can be used to correct the failure of the OMx methods for the ground states. This is in accord with the assumption that an underestimation of the overlap terms is responsible for the failure. It was shown that OPLS-AA_O also gives an excellent description of the longitudinal shift in a PBI tetramer. Using the tetramer as a test system and applying the recipe obtained in the TDDFT benchmark for the QM-part and OPLS-AA_O for the MM-part in conjunction with an electrostatic embedding scheme, a QM/MM description of the excited states of the PBI dimer including the effect of the environment could be obtained. In the last chapter the theoretical description of the Bis(borolyl)thiophene dianion and the excited states of pyracene were discussed. The electronic structure of the Bis(borolyl)thiophene dianion - a negative bipolaron - was elucidated using DFT and CASPT2 methods. Furthermore, an estimation of the extent of triplet admixture to the ground state due to spin-orbit coupling was given. In the second project the S1 and S2 states of pyracene were computed using SCS-CC2 and SCS-ADC(2) and an estimation for the balance between aromaticity and ring strain was given. This also involved computing the vibrational frequencies in the excited states. In both studies the results of the computations were able to rationalize and complete experimental results.

The aim of the present work is the development and implementation of new simulation possibilities for the CAST program package. Development included, among other things, the partial parallelization of the already existing force fields, extension of the treatment of electrostatic interactions and implementation of molecular dynamics and free energy algorithms. The most time consuming part of force field calculations is the evaluation of the nonbonded interactions. The calculation of these interactions has been parallelized and it could be shown to yield a significant speed up for multi-core calculations compared to the serial execution on only one CPU. For both, simple energy/gradient as well as molecular dynamics simulations the computational time could be significantly reduced. To further increase the performance of calculations employing a cutoff radius, a linkedcell algorithm was implemented which is able to build up the non-bonded interaction list up to 7 times faster than the original algorithm. To provide access to dynamic properties based on the natural time evolution of a system, a molecular dynamics code has been implemented. The MD implementation features two integration schemes for the equations of motion which are able to generate stable trajectories. The basic MD algorithm as described in Section 1.2 leads to the sampling in the microcanonical (NVE) ensemble. The practical use of NVE simulations is limited though because it does not correspond to any experimentally realistic situation. More realistic simulation conditions are found in the isothermal (NVT) and isothermalisobaric (NPT) ensembles. To generate those ensembles, temperature and pressure control has been implemented. The temperature can be controlled in two ways: by direct velocity scaling and by a Nose-Hoover thermostat which produces a real canonical ensemble. The pressure coupling is realized by implementation of a Berendsen barostat. The pressure coupling can be used for isotropic or anisotropic box dimensions with the restriction that the angles of the box need to be 90. A crucial simulation parameter in MD simulations is the length of the timestep. The timestep is usually in the rang of 1fs. Increasing the timestep beyond 1fs can lead to unstable trajectories since the fastest motion in the system, usually the H-X stretch vibration can not be sampled anymore. A way to allow for bigger timesteps is the use of a constraint algorithm which constrains the H-X bonds to the equilibrium distance. For this the RATTLE algorithm has been implemented in the CAST program. The velocity Verlet algorithm in combination with the RATTLE algorithm has been shown to yield stable trajectories for an arbitrary length of simulation time. In a first application the MD implementation is used in conjunction with the MOPAC interface for the investigation of PBI sidechains and their rigidity. The theoretical investigations show a nice agreement with experimentally obtained results. Based on the MD techniques two algorithms for the determination of free energy differences have been implemented. The umbrella sampling algorithm can be used to determine the free energy change along a reaction coordinate based on distances or dihedral angles. The implementation was tested on the stretching of a deca-L-alanine and the rotation barrier of butane in vacuum. The results are in nearly perfect agreement with literature values. For the FEP implementation calculations were performed for a zero-sum transformation of ethane in explicit solvent, the charging of a sodium ion in explicit solvent and the transformations of a tripeptide in explicit solvent. All results are in agreement with benchmark calculations of the NAMD program as well as literature values. The FEP formalism was then applied to determine the relative binding free energies between two inhibitors in an inhibitor-protein complex. Next to force fields, ab-initio methods can be used for simulations and global optimizations. Since the performance of such methods is usually significantly poorer than force field applications, the use for global optimizations is limited. Nevertheless significant progress has been made by porting these codes to GPUs. In order to make use of these developments a MPI interface has been implemented into CAST for communication with the DFT code TeraChem. The CAST/TeraChem combination has been tested on the $H_2 O_{10}$ cluster as well as the polypeptide met-Enkephalin. The pure ab-initio calculations showed a superior behavior compared to the standard procedure where the force field results are usually refined using quantum chemical methods.