High Throughput Collection of Reaction Data with Digital Image Analysis

As Chemistry adopt Artificial Intelligence and Machine Learning technologies more widely, the need for high quality chemical data has become a key obstacle which limits the progress of the field.

High quality reaction data, which are critical in evaluating reactivity and predicting reaction outcome, are particularly scarce due to the cost and labour associated with traditional reaction profiling techniques, i.e. sampling and analysing 5-10 samples with GC or HPLC over the course of each reaction.

An alternative approach will be developed in this project.

Digital photography and image processing will be used to collect and translate reaction data for 96 reactions at once, with the help of a robotic platform. Thus, 5-10 images may contain the reaction data equivalent to 480-960 HPLC samples.

The student will play the leading role in developing the imaging platform and the colorimetry techniques required to link images with reaction data