Catalyst discovery can be seen as a multiobjective optimization problem: there are a number of properties that we want the perfect catalyst to have. Unfortunately, such properties are often in conflict and a trade-off exists between them. A typical example is the so-called “activity-selectivity conundrum”, where a very active catalyst is too promiscuous to be selective. Thus, solutions must be found in a Pareto front of different trade-offs, which requires minute control over the optimization. To make things worse, the optimization must often run over a mix of continuous and discrete variables, such as modifications in the chemical structure or composition of the catalyst. This means that significant effort must be devoted to find robust ways to accomplish this daunting task. If a reliable pipeline to catalyst discovery could be established, we would be able to make many widespread industrial processes both greener and more energy and cost efficient.
The way we operate is chiefly through interaction with existing catalyst discovery, optimization and scale-up processes, and then establishing and improving data integration. We then use that data to push the boundaries of design-make-test cycles.