High-Throughput Experimentation in Organic Photovoltaics: Solubility, Ink-Formulation & Films

In this doctoral thesis, a complete procedure of ink formulation, film processing, film optimization and film degradation is discussed. First, a high-throughput robot-platform is introduced that is used to determine Hansen parameters and solubility values of organic semiconductors in an automated way. Afterwards, insights about different ink formulations and microstructure manipulations are given. In this context, a novel processing method is developed and successfully applied for organic photovoltaic (OPV) fabrication. At last, a high-throughput film deposition method is introduced by which a 4-dimensional material space is screened. This is performed on the one hand by a classical grid and on the other hand by a closed-loop machine-learning guided discovery. The individual stations are broken down in more detail below. In a first step, a robotic platform is introduced that allows to determine the solubility and Hansen solubility parameters (HSP) of organic semiconductors with outstanding precision. This research contribution is of high interest for the community, since several key aspects are addressed simultaneously: improvement of process accuracy and data quality while reducing process time and material consumption. In addition, the robot-based process reduces manual laboratory time and enables flexible handling of small molecules and polymers with different molecular weights with high accuracy. After optimizing the robotic experimental workflow for HSP determination, a material requirement of only 20-30 mg and a process time of about 45 minutes were achieved. Compared to the manual process, this means a reduction of 70% and 95%, respectively. In the second step, the generated solubility and HSP data is correlated and compared with classical theoretical models. Based on this, a prediction tool is developed which allows to identify and investigate various influencing factors, such as molar volume, temperature, or solvent mixtures. Further, in a wide-ranging virtual screening procedure, the solubility properties of several materials in hundreds of solvents are predicted. Furthermore, the solubility trends in non-solvent mixtures and the impact of a co-material on the solubility of a host material is discussed. Based on the generated knowledge, different approaches for the formulation of semiconductor inks were originated and applied to organic solar cells. The focus was on replacing toxic solvents by friendlier counterparts toward an environmentally friendly fabrication of organic solar cells. This was successfully realized by introducing a novel temperature gradient method based on a single solvent. By making use of the temperature-solubility relationship, chlorobenzene is replaced by o-methyl-anisole resulting in power conversion efficiencies (PCE) of 5.4% for P3HT-oIDTBr based solar cells. This means that the PCE of the reference ink is almost reproduced with a matching factor of 96.4%. Furthermore, the impact of solvent mixtures on the performance of organic solar cells is explored. In this context, several solvent blends with different solubility gradients for donor an acceptor are investigated resulting in in competitive solar cell efficiencies as compared to the reference solvent system. Furthermore, the impact of solubility on the ability on morphological reorganization based on solvent vapor post-annealing (SVA) is studied. Here, an inverse dependency between solubility and SVA treatment time is revealed which means that for solvents with lower solubility an increased treatment time is needed. Based on the variation of combinatorial chemical structures and material compositions, an almost infinite number of inks can be formulated. Theoretical models and virtual screening can drastically reduce the number of combinations, but there still remain several hundred or thousand potential inks that cannot be investigated in a reasonable time by manual testing. Thus, an automated process for accelerated film formation is needed. Therefore, in a last step, a high-throughput coating method based on drop-casting with up to 6048 films per day and a material consumption of only 90 mg is introduced. The coating-approach is successfully tested by recording the photostability of around 2000 films within 7 days. Moreover, by combining the robotic platform with a Bayesian optimizer, a 30-fold accelerated screening velocity and a simplified exploration towards a global optimum is achieved.