Creative use of time, space, and energy for accelerated green chemical discovery

With the rising demand for evermore bespoke and complex molecules and materials, how can chemistry rise to the challenge without an unsustainable environmental cost? Automation has gained tremendous momentum in recent years, but simply staffing labs with robot chemists would only multiply the waste generated without bringing about a qualitative boost to the variety and complexity of molecules accessible. Concurrently, there have been suggestions that chemistry should look to a laboratory-free future, where data mining and computation would supplant most experimental work, at the cost of disembodying the field and massively biasing discoveries towards what’s already known.

My group has been pursuing an alternative deployment of automation, computational modelling and AI in chemistry that expands rather than curtails the experimental possibilities in the lab, leading to radically greener and more efficient discovery paradigms. Instead of simplifying and adapting chemical systems to fit our physical and cognitive limitations, we use reaction media that amplifies the fundamentally inhomogeneous and stochastic nature of chemical reactions, using online analytics and probabilistic models to distil and reason about their behaviour. For example, using picolitre-sized microdroplets we have built massively parallel reactors for discovery using miniscule amounts of reagents and energy compared to traditional workflows. In my talk, I will showcase our progress towards unleashing the power of this paradigm in the context of discovering new molecules and reactions, elucidating mechanisms, and finding innovative and sustainable pathways to accessing new materials.