Machine Learning for Optimisation of Continuous Chemical Processes

This proposal is to develop an instrument that will use machine learning and flow chemistry to automatically screen a defined chemistry experimental space whilst also scanning a range of orthogonal UPLC methods in order to find all potential process related impurities. Machine learning can then be used to self-optimise the separation between all identified peaks. The use of a small scale flow reactor will enable us to achieve this with a minimal amount of starting material (target 2 to 5g) and access a wider parameter space (super ambient conditions) due to the higher pressures accessible in such systems.