The photovoltaic effect was first observed in organic compounds 60 years ago with magnesium phthalocyanine. Since then, the range of organic solar cells has dramatically expanded to include an immense variety of low band gap polymers and molecular systems. The fascination with the ability to print organic, or ‘plastic’, solar cells (OSCs) as easily as newspapers has grown rapidly over the past 20 years, but recently, the field has undergone substantial turmoil and disruption. OSCs have, however, remarkable features that make them ideal for missions to remote locations on earth, and to space, including their extreme lightweight and flexible nature, the possibility of printing in ambient (on location), and recyclability. In addition, the energy payback times of OSCs are theoretically on the order of a few days, versus 1-3 years for crystalline silicon. These positives are tempered, however, by concerns regarding long-term stability, irreproducibility of manufacturing processes that need to result in well-defined nanoscale morphologies, as well as the competing excitement of new materials (i.e. metal-halide perovskites) that is pulling researchers into other areas. The rate of improvement of OSC performance is slowing, as the number of possible components, from interfacial layers to light-absorbing materials has become exceedingly large – designing an optimal architecture is complicated by the massive number of experimental variables. How does one choose an ideal structure from literally thousands of combinations? To accelerate the production of efficient and stable OSC devices, we show how the principles of experimental design and machine learning can be applied to dramatically decrease the number of experiments required to arrive at a highly optimal solar cell architecture that has high efficiency and stability. Rational screening of multiple variables simultaneously enables much faster optimization of organic solar cells, from choice of composition of the bulk heterojunction and interfacial layers, to processing parameters, and so forth. The potential of lightweight rolls of plastic solar cells for applications on earth, and in space is both compelling and obtainable, and results showing how design of experiment can lead to rapid optimization will be shared.

Optimization of organic solar cells through experimental design and machine learning

256th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.