Research into chemical discovery, testing optimisation and analysis can be a labour-intensive and time-consuming process. With many of the stages requiring manual preparation, sampling, and analysis this can lead to increased time scales, higher costs, potential for human error and can limit the scope of exploration.
A team of researchers, led by Professor Nick Warren, Chair in Sustainable Materials in the School of Chemical, Materials and Biological Engineering at the University of Sheffield, have developed a new automated platform, or self-driving laboratory, that acts like a sophisticated chemical assembly line which is powered by artificial intelligence. Instead of traditional flasks, reactants flow through tiny tubes and reactors, allowing for incredibly precise control over the reaction. It's equipped with sensors that constantly monitor the reaction and can simultaneously target multiple product properties, such as reaction conversion, purity, particle size, and uniformity. This real-time data is fed into a machine learning algorithm, which then adjusts the reaction conditions – the amounts of ingredients, the speed, and other factors – without any human intervention.
In a collaborative project, between the University of Sheffield, the University of Leeds and the University of York, researchers developed technology for high-value, low volume nanoparticle based materials which has potential applications in healthcare. Similar materials are used for encapsulating difficult to deliver drugs, and mRNA in new vaccine technologies.
Professor George Panoutsos, Head of the School of Electrical and Electronic Engineering at the University of Sheffield, and a co-investigator in the research grant, said: “Our self-driving lab platform offers unprecedented insights into complex polymer synthesis, enabling days of unsupervised experiments. This work highlights the challenges and diverse approaches – from automated screens to AI-based many-objective optimisation – crucial for effectively supporting discovery as well as practical decision-making .”
Professor Warren has further developed this technology for optimising conditions for making polymers which are used in large volume products such as paints and adhesives. This will allow optimisation of new “greener” products on the faster timescales required to meet sustainability demands.
He said: “This work represents the first instance of a reactor platform capable of closed-loop self-optimisation of emulsion polymers, unlocking the ability to accelerate the development of new polymeric materials."
More recent findings in a collaboration with Karlsruhe Institute of Technology have demonstrated the capability of their self-driving laboratory to create highly functional polymer building blocks suitable for advanced applications. In this newly published study, the automated system was used to precisely synthesise poly(pentafluorophenyl acrylate) (PFPA), a versatile polymer readily amenable to post-polymerisation modification. The self-driving laboratory, equipped with real-time Nuclear Magnetic Resonance (NMR) and Size Exclusion Chromatography (SEC) analysis, autonomously identified the optimal conditions for PFPA production. This enables scientists to create polymers with specific "active" sites that can then be tailored with different chemical components, paving the way for next-generation high-performance materials with precisely controlled properties for diverse applications.
Looking to the future, Professor Warren said: "Moving forward, we now intend to further evolve these technologies in collaboration with academics and industry partners worldwide to accelerate the development of a wider range of polymer materials. We will specifically focus on adapting self-driving laboratories for discovery of polymers and nanomaterials which can meet important societal challenges in the context of sustainability and health. Since moving to Sheffield, we have already started collaborating with experts in the Centre for Machine Intelligence (CMI) and the Grantham Centre for Sustainable Futures to enhance the impact of this research."
This new technology has several advantages over traditional methods: As the process is automated it speeds up the development of new materials; Less waste is generated as the process can be so precisely controlled making it more energy efficient and sustainable; Automation reduces human exposure to potentially hazardous chemicals, making operations safer; The platform can be programmed to produce materials with specific properties, opening up a world of possibilities for customised products.
Read the three recently published three papers demonstrating this shift towards more efficient, data-driven, and autonomous methods in chemical research: