The application of machine learning to polymer chemistry offers the opportunity for a step change in speed of innovation. High impact examples have been demonstrated in small molecule chemistry, such as those with pharmaceutical applications. In partnership with online analysis and automated reactor technology, machine learning provides the potential for rapid adjustments in response to observed changes, or the collection of high-resolution data with a much-reduced workload. Online analysis integrates particularly well with flow chemistry; and has been used to great success with polymer chemistry.<br/>Herein, we present an automated polymer synthesis platform, with integrated online NMR spectroscopy and GPC capable of autonomous exploration of polymerisation reaction space.<br/>The online analyses provide real-time molecular weight and conversion information for a range of RAFT polymerisations. High throughput experiments screen reaction conditions in an ordered fashion with little user input. Further complexity is achieved with integration of a multi-objective optimisation algorithm (TS-EMO), which reduces wasted experiments and explores the achievable reaction space for a range of RAFT polymerisations with varied RAFT agents, monomers and initiators. using a multi-objective optimisation algorithm (TS-EMO). The algorithm takes an initial near-random screen of conditions to generate a model for the system and uses this to identify further experiments to achieve the given objectives (in this case increased conversion and reduced dispersity). These results are then assimilated into the model and further optimisation performed. The incorporation of our pressurised flow reactor means that conventionally forbidden conditions (i.e. temperatures above the boiling point of solvents) are accessed and explored.

Application of an autonomous continuous-flow platform for multi-objective optimisation of polymer synthesis

ACS Spring Meeting 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.