Tobias Osterrieder - Autonomous Optimization of ternary OPV devices on automated research lines

For more information about the NGSE PhD-Postdoc series and register for future talks go to:
https://www.ngse.info/phd-postdoc-ser...

Follow us on Twitter at:
  / imeet_pv  
  / ngse_conf  

Talk by Tobias Osterrieder currently a PhD student at Friedrich-Alexander-Universität Erlangen-Nürnberg on October 26th, 2022
Abstract:
The optimization of a organic photovoltaic device architecture is usually a challenging labor and time-intensive process. The vast parameter space, spanned by the possible material compositions and process parameters, typically leads to a high number of required experiments to find the best conditions. The complexity is further increased due to the fact that typically the optimization is not only in multiple parameters, but also according to multiple objectives, e.g. efficiency and stability. Artificial intelligence approaches hold the potential to drastically reduce the time and required number of samples for such optimizations. In our laboratories we developed an automated research line (AMANDA L1), which is capable of manufacturing and characterizing complete OPV devices. We combine this system with a Bayesian optimizer to create a self-driving laboratory. Our system has the capability to autonomously perform a multiobjective optimization of OPV devices for efficiency and stability. We demonstrate the power of this approach by optimizing the ratios of multi-component blends and process parameters in a self-driving closed-loop approach. The system significantly decreases the number of required samples for an optimization and converges on the optimal conditions in less than 15 iterations. Through the addition of a Gaussian Progression Regressor for the prediction of the efficiency from characteristic features of the absorption spectrum of the active layer we were able to decrease the time of one iteration even further. Our approach is generic and demonstrates the power of artificial intelligence to accelerate and transform materials science altogether.