Computational methods to engineer process-structure-property relationships in organic electronics: The case of organic photovoltaics

Ever since the Nobel prize winning work by Heeger and his colleagues, organic electronics enjoyed increasing attention from researchers all over the world. While there is a large potential for organic electronics in areas of transistors, solar cells, diodes, flexible displays, RFIDs, smart textiles, smart tattoos, artificial skin, bio-electronics, medical devices and many more, there have been very few applications that reached the market. Organic photovoltaics especially can utilize large market of untapped solar power -- portable and affordable solar conversion devices. While there are several reasons for their unavailability, a major one is the challenge of controlling device morphology at several scales, simultaneously. The morphology is intricately related to the processing of the device and strongly influences performance. Added to this is the unending development of new polymeric materials in search of high power conversion efficiencies. Fully understanding this intricate relationship between materials, processing conditions and power conversion is highly resource and time intensive. The goal of this work is to provide tightly coupled computational routes to these expensive experiments, and demonstrate process control using in-silico experiments. This goal is achieved in multiple stages and is commonly called the process-structure-property loop in material science community. We leverage recent advances in high performance computing (HPC) and high throughput computing (HTC) towards this end. Two open-source software packages were developed: GRATE and PARyOpt. GRATE provides a means to reliably and repeatably quantify TEM images for identifying transport characteristics. It solves the problem of manually quantifying large number of large images with fine details. PARyOpt is a Gaussian process based optimization library that is especially useful for optimizing expensive phenomena. Both these are highly modular and designed to be easily integrated with existing software. It is anticipated that the organic electronics community will use these tools to accelerate discovery and development of new-age devices