Molecular interfaces are ubiquitous and perform critical functions in organic electronic and photovoltaic devices. In typical organic solar cells sunlight is harvested by organic species and charge separation is driven by the potential energy difference at an interface between donor and acceptor species. In any organic device, charge transport to external circuits depends on interfaces between active molecular layers and electrode materials. The properties and functionality of these critical interfaces cannot be deduced directly from those of their isolated constituents. Rather, they emerge from quantum mechanical interactions at the atomistic scale. Predicting the properties of molecular interfaces thus requires a fully quantum mechanical first principles approach.
The configuration space of molecular interfaces is infinitely vast, owing to the endless possibilities of combining layers of one or more molecular species with different substrates. Layers of the same molecular species may adopt different structures on different substrates and therefore exhibit different electronic and optical properties. Furthermore, epitaxial templating may enable stabilizing meta-stable crystal structures with desirable properties in thin film form. Efficient algorithms may significantly accelerate the discovery and design of molecular interfaces with enhanced properties.
We are developing a first-principles framework for structure prediction and design of molecular interfaces by integrating first principles quantum mechanical simulations with genetic algorithms and machine learning. The structure generation code, Genarris , and the genetic algorithm (GA) code, GAtor , which currently work for Molecular Crystals , are being adapted to perform structure prediction and property-based optimization of interfaces.