(275b) Lessons and Opportunities in Data-Driven High Throughput Experimentation

The adoption of data-science methods within chemical engineering is primed to revolutionize materials discovery and to accelerate developments in molecular understanding. Yet, the movement is largely dominated by data and information originating from molecular modelling and simulation, or that is acquired from legacy sources (e.g. databases, literature mining) accumulating data from decades of careful experimental work. In order to advance data-driven materials developments well into the future, high-throughput experimentation (HTE) and automation throughout the complete laboratory workflow must also be developed and widely adopted to accelerate rates of experimental data production. In this talk we outline several examples and experiences showcasing how our research group is developing and adapting hardware and software infrastructure to accelerate the pace of molecular discovery in soft-matter systems for applications in health care, clean energy and materials synthesis. The talk will highlight recent research examples related to the implementation of HTE for electrolyte discovery and colloidal formulation/synthesis. We will also highlight significant challenges that emerged while transitioning established ‘wet-laboratory’ practices to HTE. These relate to adapting routine experimental methods to HTE, developing new skills within the research workforce, the adoption of good data stewardship practices, needs for autonomous data sorting/classification, algorithms for automatic modeling and analysis amongst many others. Conversely, we will also highlight the opportunities that emerge for enhancing virtual collaboration, improving information quality, enabling open data/hardware/software sharing, tackling irreducible problems (e.g. optimization of complex formulations), and outlooks for the implementation of self-driving laboratories.