Autonomous Mobile Robot Chemists

The use of autonomous robotic technologies is increasingly common for applications such as manufacturing, warehousing, and driverless vehicles. Automated robots have been used in chemistry research, too, but their widespread application is limited by the cost of the technology, and the need to build a bespoke automated version of each instrument that is required. We have developed a different approach by using mobile 'robotic chemists' that can work within a relatively standard laboratory, replicating the dexterous tasks that are carried out by human researchers. These robots can operate autonomously, 24/7, for extended periods, and they can therefore cover a much larger search space that would usually be possible. Also, the robots are driven by artificial intelligence (AI) and can search highly complex multidimensional experimental spaces, offering the potential to find revolutionary new materials. They can also carry out multiple separate experiments in parallel, if needed, to make optimal use of the available hardware in a highly cost-effective way.

Our proposal is to establish a globally unique user facility in Liverpool that covers a broad range of materials research problems, allowing the discovery of useful products such as clean solar fuels catalysts, catalysts for plastics recycling, medicinal materials, and energy materials. This facility will allow researchers from both academic teams and from industry to access this new technology, which would otherwise be unavailable to them. Because the automation approach is modular, it will be possible for users to bring along specific equipment for their experiments to be 'dropped in' temporarily to create new workflows, greatly expanding the possible user base. The scope here is very broad because we have recently developed methods that give these robots have very high placement precision (+/- 0.12 mm): to a large extent, if a human can use the instrument, then so can the robot.

We have identified, initially, a group of 25 academic users across 12 universities as 'day one' prospective users, as well as 7 industrial organisations with a specific interest in this technology. The potential user base, however, is far broader than this, and we will solicit applications for access throughout the project and beyond. This will be managed by a Strategic Management Team and an Operational Management Team that involves academics as well as permanent technical, administrative, and business development staff in the Materials Innovation Factory in Liverpool. Our overall objective is to build a sustainable AI-driven robotic facility that will provide a unique competitive advantage for the UK to discover new functional materials on a timescale that would be impossible using more conventional research methodology. In addition to focusing on excellent science, we will also consider diversity and career stage when prioritising access; for example, even a short, one-week visit to this autonomous facility might lead to 100's or even 1000's of new materials with associated property measurements, which might radically transform a PhD project or the change the direction of the research programme for an Early Career Researcher. This facility will therefore build the base of the UK research pyramid, as well as supporting activity that is already internationally leading, and our day-one user base includes researchers at all career stages.

The impact of this proposal will be in 4 areas:

(1) Economic Impact. To be achieved by (i) direct acceleration of research programmes for industrial users (e.g., see Example Project 3 and Table 1 in proposal), and; (ii) generation of new materials within academic projects that create future value. We have an excellent collective track record in both respects. The £82 M MIF is the largest academic/industry collocation in the physical sciences in the UK, and approx. 80 industry researchers work in the MIF building. We also have a strong track record in economic impact through academic innovation, and a number of the prospective users have formed successful spin-outs, such as, Cooper (Liverpool, Porous Liquid Technologies), Rannard (Liverpool, Tandem Nano Ltd), Cronin (Glasgow, DeepMatter), Williams (Oxford, Econic), Scherman (Cambridge, Aqdot), and Aspuru-Guzik (Toronto, Zapata Computing and Kebotix). This project has the potential to deliver a broad range of economic impacts for industry users, for established academic innovators (see list above), and for early career academics, who have not yet participated in knowledge exploitation of this kind (e.g., by generating a large dataset across a range of materials to build strong IP).

(2) Societal Impact. The potential impact here is longer term, but this new facility will enable projects in areas that have strong societal importance, such as energy (generation and storage; multiple projects), health (e.g., materials for cancer therapy, Scherman project; Rannard's work on anti-HIV nanomedicines), and sustainable plastics (e.g., projects with Dove, Shaver & Williams).

(3) People Impact. As detailed in the proposal, it is possible in some areas for 1-2 weeks of access to this robotic facility to lead to volumes of data that might otherwise be attainable in an entire PhD programme. This has the potential to revolutionise PhD projects, hence our strong emphasis on PhD access. This unique facility will also 'jump start' the careers of early career academics (Table 1 for day-one list), offering the potential to carry out experiments with a breadth and complexity that would otherwise be impossible. The facility will also make a training impact by familiarizing a range of users from mulitiple organizations with the tools that might underpin the future of digital materials research.

(4) Knowledge Impact. This can operate on two main levels: (i) Discovery Impact - using autonomous robots to find, for example, a much better solar fuels catalyst, a target drug polymorph, a home & personal care product, or an enhanced polymerisation catalyst. (ii) Understanding Impact - this approach does not only 'screen' for optimal materials; it is also possible to extract understanding from the large datasets that arise. Hence, we see the primary knowledge output as the publication of first-in-class papers demonstrating the autonomous discovery of functional materials with unprecedented properties coupled with a degree of understanding of the associated design rules based on the (large number) of failed experiments in these datasets.

We will maximise the impact of this project by:
(1) Exploiting Existing Networks. We have strong connectivity to both academic users and industry users via the MIF and the Leverhulme Research Centre for Functional Materials Design (PI is Director for both), and the Henry Royce Institute (PI is Theme Champion for Chemical Materials Discovery). Dr Ben Slater (Royce / MIF BDM) will assist help to facilitate this.
(2) Engaging with the Centre for Process Innovation to further broaden our industrial impact.
(3) Holding annual user workshops, extending beyond the initial 3-year project duration, to broaden the user base for the facility and disseminate its value.

Dr Ben Slater (MIF / Royce BDM) will maintain an impact register, capturing impact from all of the 4 areas described above (Economic, Societal, People, Knowledge).