Machine Learning for Orbital Energies of Organic Molecules Upwards of 100 Atoms

Organic semiconductors are promising materials for cheap, scalable, and sustainable electronics, light-emitting diodes, and photovoltaics. For organic photovoltaic cells, it is a challenge to find compounds with suitable properties in the vast chemical compound space. For example, the ionization energy should fit to the optical spectrum of sunlight, and the energy levels must allow efficient charge transport. Herein, a machine learning model is developed for rapidly and accurately estimating the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energies of a given molecular structure. It is built upon the SchNet model [Schutt et al. (2018)] and augmented with a Set2Set readout module [Vinyals et al. (2016)]. The Set2Set module has more expressive power than sum and average aggregation and is more suitable for the complex quantities under consideration. Most previous models are trained and evaluated on rather small molecules. Therefore, the second contribution is extending the scope of machine learning methods by adding also larger molecules from other sources and establishing a consistent train/validation/test split. As a third contribution, a multitask ansatz is made to resolve the problem of different sources coming at different levels of theory. All three contributions in conjunction bring the accuracy of the model close to chemical accuracy.

Automated summary

Organic semiconductors are ideal for affordable, scalable, and sustainable electronics, LEDs, and solar cells. A machine learning model is developed to accurately and quickly estimate the HOMO and LUMO energies of a given molecular structure, addressing a challenge in finding suitable compounds for organic photovoltaic cells.