Research

Cat-Tail comic: a stick figure marveling at a cat with an unusually long tail catalyzing a CO2 reduction reaction.

Cat-Tail: Statistical Catalysis

Resolving catalytic species in the distribution tail.

A cat is dormant most of the day. But when it's on the job, it's on the move non-stop. Catalysts work the same way. The majority of sites sit in ordered, uninteresting structures throughout catalysis, whilst the disordered minority hastens to catch, break, and make molecules. Those irregularities are metastable and transient, out on the long tail of the distribution of catalyst states. We choose to go after this long, long tail of the cat(alyst).

Watt-If comic: stick figures running from an unstable, erupting volcano while another welcomes the new lands it creates.

Watt-If: Materials Off-Equilibrium

Harnessing the operando instability of energy materials.

It is heartbreaking to see batteries lose capacity, nanoparticles sinter, and electrodes corrode. Some materials last longer, but in the long run, they all die. These highly energetic non-equilibrium processes have been a source of fear. But what if we can leverage them to build disruptive energy technologies, out of the disruption? We constantly ask ourselves "what-if"s and computationally stress-test energy materials in extreme conditions.

Interpreter comic: a stick figure stands before a black box labeled ML and asks 'Who's there?'; an interpreter answers from behind.

Interrogator: Interpretable AI

Teaching ML models to speak quantum chemistry.

It's easy to ship models and max out performance, but it's hard to translate them into the language of chemistry and physics. We strive to understand the inner workings of the omnipotent black box that is called machine learning, and we transcribe them into practical design guidance or elegant analytical expressions. Babel wasn't built overnight, but we are working to train the polyglots of the AI era and lay down the first layer of bricks.

Optimist comic: a stick figure climbing across scattered points on a Pareto-like plot, gesturing 'we can go here, and here, and here'.

Optimist: Multi-Objective Design

Multi-objective global optimization in vast latent spaces.

Some say engineering is about making compromises and concessions. We do not agree with this claim. The chemical space is vast enough, and there is supposed to be a way to maximize everything we love and minimize everything we hate. While the pessimists work under constraints and live with the trade-offs, the optimists take the red pill, venture into the wild, flick around, and find out.