Establishing Quantitative Relationships Between Composition, Morphology, and Performance in Polymer Electronics

Abstract

π-Conjugated polymers (CPs) have been the subject of a rapidly growing body of research ever since their discovery in the 1970s. A great deal of work has been conducted to improve device performance and better understand the relationships between monomer design and optoelectronic properties. However, the same effort has not been put into understanding the driving factors for morphological development or the impact of morphology on the final device performance, partially due to the complex and highly dependent nature of the morphology in these materials. This work seeks to 1) understand the role of structural defects on the crystal lattice and morphology of CPs, 2) to develop quantitative methods for studying morphological development to predict device performance. First, self-assembled nanowires of poly(3-hexylthiophene), a well-studied CP, are used to probe the effect of regio-defects and chain end-groups on the crystal. The defects were found to readily incorporate into and disrupt the crystal lattice, and they were used to influence nanowire morphology. Next, an unsupervised machine learning toolkit is developed, which automatically processes and labels the phases and domains in scanning probe microscopy images. Finally, the labels generated by this toolkit are used to examine the benefits of quantitative morphological information using a variety of regression and neural network techniques. This data, and subsequently the models, describe changes in morphology and device performance for two types of CP-based devices during thermal annealing. Current efforts and future work are discussed.