Development of Accurate and Efficient Electronic Structure Approaches for the Description of Organic Systems

The overarching objective of this thesis is extending and adapting the set of computational tools available for describing molecular precursors of organic semiconductors. The research presented within develops adhering to three principle goals: (1) provide accurate energies and geometries for large-scale assemblies at an affordable computational cost; (2) provide tools that thoroughly explore and map free energy landscapes and emphasize the importance of this comprehensive mapping; (3) improve density functional theory (DFT) descriptions of precursors to charge carrier molecules which are currently quite challenging. The Self Consistent-Charge Density Functional Tight Binding (SCC-DFTB), particularly the most recent variant (DFTB3), provides an excellent balance between accuracy and efficiency needed to study large molecules and to perform molecular dynamic simulations. As DFTB approximates the DFT Hamil- tonian it suffers from the same principle shortcomings, including neglecting dispersion interactions. Such interactions are crucial for establishing the proper relative orientations of organic molecules in aggregate, a property of fundamental importance for elucidating conduction properties. Building upon our laboratories experience with the dDsC dispersion correction, an a posteriori pairwise dispersion correction that depends upon Mulliken charges, dDMC, was developed specifically for DFTB. During the course of this work, a caveat in the DFTB parameterization of sulfur was identified that caused sulfur containing molecules to exhibit strong non-covalent binding, even in the absence of a dispersion correction. From a computational perspective, detecting all conformations present in a chemical space and recognizing those structures that are the most chemically relevant represents a significant challenge. To address this problem, we have borrowed a technique usually used for molecular mechanics simulation and combined it with DFTB (REMD@DFTB3). Using this technique allows through exploration of various chemical systems at a quantum level in which bond breaking and formation is possible, thereby allowing description that surpass the typically employed static picture. Although the DFTB approach is unquestionable practical, certain situations require more sophisticated and accurate treatments. Radical cation dimers are, for instance, illustrative as models for organic charge carrier molecules, yet their computational description possesses inherent challenges. Although accurate, post-HF methods are too expensive for routine use, while DFT approaches suffer from their typical shortcomings (self-interaction error, lack of dispersion interactions, missing static correlation). To overcome these problems, a new density functional, ωB97X-dDsC, was developed in which the parameters of the dDsC correction and those of ωB97X were fitted together. Preliminary examination of the overall performance of ωB97X-dDsC is very promising.