Accelerating Materials Development via Automation, Machine Learning, and High-Performance Computing

August 2018 - Joule : Cell Press

https://www.sciencedirect.com/science/article/pii/S2542435118302289

In collaboration with Prof. Tonio Buonassisi (MIT), we lay out the pathway for increasing the speed of discovery.

Successful materials innovations can transform society. However, materials research often involves long timelines and low success probabilities, dissuading investors who have expectations of shorter times from bench to business. A combination of emergent technologies could accelerate the pace of novel materials development by ten times or more, aligning the timelines of stakeholders (investors and researchers), markets, and the environment, while increasing return on investment. First, tool automation enables rapid experimental testing of candidate materials. Second, high-performance computing concentrates experimental bandwidth on promising compounds by predicting and inferring bulk, interface, and defect-related properties. Third, machine learning connects the former two, where experimental outputs automatically refine theory and help define next experiments. We describe state-of-the-art attempts to realize this vision and identify resource gaps. We posit that over the coming decade, this combination of tools will transformthe way we perform materials research, with considerable first-mover advantages at stake.

Wu Jing uses vacancy-induced Kondo hybridization to tune the thermoelectric transport in 2D MoS2

June 2020 - Proceedings of National Academy of Sciences

https://www.pnas.org/content/117/25/13929.short

In a careful and seminal study that took 3 years, a very careful study and analysis, we show that new physics such as Kondo scattering can enhance thermoelectric properties substantially!

The study of correlated phenomena in 2D semiconductors opens up new pathways toward understanding and engineering material functionalities (such as thermoelectrics) in easily accessible van der Waals solids. Local structural defects such as vacancies inevitably exist in natural as well as synthetic TMD crystals and have been predicted to serve as magnetic impurities capable of enhancing the strongly correlated effect. Herein we discover unusual thermoelectric behavior in sulfur vacancy-enriched MoS2 by rationally selecting h-BN as the substrate. We demonstrate that the thermoelectric transport properties can be strongly manipulated by vacancy-induced Kondo hybridization. A significant enhancement of thermoelectric power factor by two orders of magnitude is achieved in the MoS2/h-BN device.

Inverse design of inorganic crystals using generalized invertible crystallographic representation

Deep learning has fostered many novel applications in materials informatics. However, the inverse design of inorganic crystals, i.e. generating new crystal structure with targeted properties, remains a grand challenge. An important ingredient for such generative models is an invertible representation that accesses the full periodic table. This is challenging due to limited data availability and the complexity of 3D periodic crystal structures. In this paper, we present a generalized invertible representation that encodes the crystallographic information into the descriptors in both real space and reciprocal space. Combining with a generative variational autoencoder (VAE), a wide range of crystallographic structures and chemistries with desired properties can be inversedesigned. We show that our VAE model predicts novel crystal structures that do not exist in the training and test database (Materials Project) with targeted formation energies and band gaps. We validate those predicted crystals by first-principles calculations. Finally, to design solids with practical applications, we address the sparse label problem by building a semi-supervised VAE and demonstrate its successful prediction of unique