Controlling an organic synthesis robot with machine learning to search for new reactivity

Abstract

The discovery of chemical reactions is an inherently unpredictable and time-consuming process. An attractive alternative is to predict reactivity, although relevant approaches, such as computer-aided reaction design, are still in their infancy. Reaction prediction based on high-level quantum chemical methods is complex, even for simple molecules. Although machine learning is powerful for data analysis, its applications in chemistry are still being developed. Inspired by strategies based on chemists' intuition, we propose that a reaction system controlled by a machine learning algorithm may be able to explore the space of chemical reactions quickly, especially if trained by an expert. Here we present an organic synthesis robot that can perform chemical reactions and analysis faster than they can be performed manually, as well as predict the reactivity of possible reagent combinations after conducting a small number of experiments, thus effectively navigating chemical reaction space. By using machine learning for decision making, enabled by binary encoding of the chemical inputs, the reactions can be assessed in real time using nuclear magnetic resonance and infrared spectroscopy. The machine learning system was able to predict the reactivity of about 1,000 reaction combinations with accuracy greater than 80 per cent after considering the outcomes of slightly over 10 per cent of the dataset. This approach was also used to calculate the reactivity of published datasets. Further, by using real-time data from our robot, these predictions were followed up manually by a chemist, leading to the discovery of four reactions.

Flow Photochemistry: Shine Some Light on Those Tubes!

“…Many automated strategies for screening and optimization in flow, for both discrete (solvents, catalysts, ligands) and continuous (temperature, loading, time) reaction variables, can now be found in the literature. Similar automated systems can be employed for reaction discovery (by combining stocks of different reactants) and for the rapid synthesis of compound libraries, amply demonstrating the rapidity and versatility of flow systems for such purposes [68][69][70][71]. Inline analytics allow immediate evaluation of the reaction's outcome and combination with smart algorithms [machine learning, artificial intelligence, and design of experiments (DoE)] enables self-optimization protocols [72,73], reducing the number of experiments necessary to optimize complex systems.…”

Kinetics Studies on a Multicomponent Knoevenagel−Michael Domino Reaction by an Automated Flow Reactor

“…[2] Thanks to the rapid technological change combined with a much stronger interplay between chemical synthesis and the engineering disciplines, this drawback was recently overcome, as documented by reports on, e.g., self-optimizing automated flow reactors, [3] algorithm-based optimum catalyst selection, [4] or the discovery of new reactivities by an organic synthesis robot using machine learning. [5] Importantly, all these studies utilize platforms that fullyautomatically collect a large amount of data in a short period of time with the general ambition to transform synthetic chemistry into a more data-driven discipline. [6,7] We have recently presented an automated flow system that relies on an efficient combination of a continuous-flow micro-reactor and an online coupled HPLC analysis system to rapidly collect data on heterogeneous catalytic reactions.…”

Evolving embodied intelligence from materials to machines

“…Accelerating the development cycle of bespoke materials is key as the space of possibilities is so vast (using e.g., robotic materials synthesis 10 , or combinatorial materials libraries 11 ). Semi-or fully-autonomous 'closed loop' systems use robots to efficiently perform experiments with reduced reliance on humans, and naturally couples with techniques that automatically plan optimal sets of experiments 12,13 to reach desired material properties 14 . Robots can perform multiple simultaneous experiments, vastly reducing time and human effort whilst increasing providence of experimental data for computational, AI and machine learning methods, which are now mature enough to reliably predict properties of new materials and allow a vast set of previously-physical experiments to be conducted (cheaper and faster) virtually 8 .…”