Chemical reactions are the universal path to new molecules, and are of central importance in science. The description of reactions is largely obtained once they are completed, with a key role of nuclear magnetic resonance (NMR), through the structural information found in the spectra of purified products. Structural information is, however, elusive for ongoing reactions. While time-resolved NMR provides kinetics for known compounds, compound identification from a reaction mixture is not generally feasible today. Such identification would be transformative for the discovery, understanding and optimisation of reactions.

The UNMIX project aims at providing structural and quantitative information from ongoing reactions in experimentally relevant conditions. Its central concept is the unmixing of compound-specific data by time-resolved diffusion NMR in continuous flow. Diffusion NMR is a method with the potential to separate the spectra of mixture components. Until recently, diffusion NMR was limited to reactions performed in an NMR tube, a setup that often fails to replicate experimental conditions of interest. In contrast, emerging continuous-flow approaches give real-time monitoring in relevant conditions.

Building on the possibility to obtain high-quality diffusion NMR data for flowing samples, which we recently demonstrated, we aim aims to develop original flow NMR pulse sequence and associated processing and analysis programs, and aaply these to the investigation of organic chemical reactions, and the design of autonomous flow reactors. These general methods may have broad impact in chemical science and engineering.

Jean-Nicolas DUMEZ

CNRS research

Project leader