Finding new materials is challenging and needed more than ever. Especially in renewable energy conversion, targeted materials design has a significant impact on efficiency and cost of devices such as batteries and fuel cells. Traditionally, new materials are developed via trial and error in laboratory experiments. The obvious disadvantage of this method is its high cost and slow progress, but it also implies the risk of missing out materials candidates or important considerations regarding their properties and the business case associated with their use. Materials simulation and quantum-chemical calculation methods such as Density Functional Theory (DFT) help overcome these barriers but are difficult to get started with, and do not account for the business case.

We developed a novel Materials Informatics solution to overcome these barriers. Our software platform MatterMine has been launched as a spin-off project at the German Aerospace Center (DLR) and commercialized as an independent company (ExoMatter) in 2022. Its roots go back to research efforts at DLR in collaboration with the Lawrence Berkeley National Laboratory (“Materials Project”), and our own efforts build upon this very successful endeavour.

In this talk, we show how we use DFT calculations, Machine Learning, and cost data to enable a multi-dimensional materials development process. Our scoring and ranking algorithms help compare materials data across chemical and physical, as well as sustainability, and cost related aspects. Our presentation highlights the path from first-principles data to usable, application-related materials properties. This allows a virtual materials screening process without prior knowledge in theoretical methods. We show an exemplary use case at DLR, in which our methods allowed the fast identification of doped ceria materials for solar-thermochemical fuel production, and give a short overview of ongoing and past industrial applications.