Automated chemistry has been a priority for many companies since the outbreak of the pandemic. Paul Whittles is the sales director at Deepmatter Group (a company based in Glasgow, UK that develops tools for automated chemical synthesis).
Many companies currently work at 25% capacity. However, with fully automated synthesis, chemists could sit at home and program a robot to maintain the lab’s functioning. Whittles says that although it’s something people have long aspired to, the idea is being taken seriously by more people.
Whittles says that “Automating Chemistry isn’t really that new.” “It’s something people have been trying since the early 1990s. But, the last decade has seen an explosion of its use and the commercially available tools to accomplish it. Michael Schneider, chief executive of automation specialists Chemspeed, is stating that it is crucial for everyone to be competitive in research, development, and quality control. Richard Bourne, a UK University of Leeds professor, believes that the industry is now on board. He has worked with GlaxoSmithKline and AstraZeneca on automating their process chemicals, and both companies are looking forward to a future of large-scale robotic labs.
Automation can be used to automate almost any type or scale of chemistry, from automating reaction conditions to using robotics to manipulate reagents or carry out complete reactions. Bourne is currently developing automated flow systems to provide online control for pharmaceutical development in the late stages. Automation is also being used at the Materials Innovation Factory (MIF), University of Liverpool, UK. Anna Slater, a Liverpool synthetic materials chemist, says that this is where the real excitement lies. “You can use automation within the materials space to give yourself a benefit that’s not been seen before,” says Anna Slater, Liverpool synthetic materials chemist. The MIF opened in 2018 and already has spun off CageCapture, a company that uses nanocage molecules for removing toxic pollutants from the atmosphere.
The ultimate goal for medicinal chemists is to automate a complete workflow, from design and synthesis to assaying, and analysis. This process is sometimes known as the DMTA cycle (design-make-test-analyse). Steve Coles chief technical officer at Deepmatter says that although it’s a difficult task, this is what they are trying to accomplish. “If you can reduce the process from 90 to 20 days, that will result in a fourfold increase in the number of candidates you can screen.
Automation offers more than speed. Automation offers many benefits, including increased productivity, higher yields, product purity, reduced waste, and lower disposal costs. Radleys in Essex, UK is a manufacturer of glassware and instruments that makes automated reaction stations. Taylor says that automation can improve safety because it allows for greater control over reaction conditions. Bourne also says that automated processes are more repeatable. “That’s really important for the pharmaceutical industry. This makes it easier to scale up their processes.
Slater says that the most exciting part of the process is the amount and depth of information that can be gathered. “That might point you in the direction that’s really unexpected.” Automated systems can gather massive amounts of information, such as temperature and pH at millisecond intervals. Some of the most recent advances allow for this data to be fed back. Coles explains that by collecting data, it is possible to determine whether your reaction has ended. This will allow you to stop boiling for the next two days.
However, automation is not likely to work in all chemistry. Coles says that handling solids is still difficult and that some solids are more dangerous than others. For example, a Grignard reaction uses powdered magnesia, which can be dangerous.
Automation is not for everyone. Oliver de Peyer is an application specialist at Peak Analysis and Automation. He’s based in Farnborough in the UK and has experience working in lab automation in academia and industry. “People have said to my, “If your were a scientist, you would do it by hand”, and it’s fundamentally lazy to use a robot to do it.”
It seems that there is a concern that automation will replace the bench chemist and change the culture of lab science. Schneider admits that changes mean leaving your comfort zone. However, he believes automation will quickly become mainstream as more chemists realize the benefits.
Automation will undoubtedly change the role and responsibilities of chemists, but it may be a good thing. Bourne says that automation will allow chemists, who often do repetitive and laborious work, to think more about the science. It is possible that the balance of time spent in the lab and at the office will shift more towards the office. Cole calls this ‘information workers’. This could allow for greater intellectual curiosity. De Peyer said that scientists often lose their intellectual curiosity due to the repetitive nature and tediousness of scientific research. De Peyer hopes that automation will have a positive effect on the PhD experience. “It won’t be down to you working 80 hour weeks in the lab. You won’t feel that level of physical stress.”
However, most people in the field aren’t convinced that automation will decrease the need for chemists. Coles says that the shift will be to optimize the time of the synthesis robot by not asking it for things that won’t work. Slater believes that there will be some ‘tinkering around in the lab’. “Part of formulating the right problem to automate is understanding the chemical process. And part of understanding the chemistry is doing it.” Slater still does exploratory experiments in round-bottomed flasks before moving on with more automated methods.
The role of the chemist will likely change with automation, but that could be a good thing. There may be fewer technician positions that are heavily dependent on repetitive laboratory work. However, most technicians are highly skilled specialists. Automation will likely create many new roles to maintain and run automated systems.
Lab automation is becoming more important, which has implications for the skills that undergraduates should acquire. Slater says that chemists need to have practical skills. However, ‘if anyone asks anybody what we should be teaching our students about chemistry, coding would be the first thing they answer’. For coding current automated systems, it is important to have a working knowledge of Python and other programming languages. The University of Bristol in the UK offers a degree program in chemistry and scientific computing. There are more likely to be offered.
However, what is most important for undergraduates is lab automation exposure, Schneider says. Many companies claim that their platforms will soon offer user-friendly interfaces that allow anyone to set up experiments.
Although robots are not yet coming to your job, they may in the future. However, advanced artificial intelligence and automation could make it possible. “In about 10-20 years, I can see that models with AI will be available. This is where you send a request and the web server executes the chemistry you have specified. You then get a response. Bourne says that it all happens in a few hours and is fully automated.
It is possible that we are already close to being there. IBM Research Europe and Chemspeed created RoboRXN in 2020. This automated chemical synthesis platform combines IBM’s predictive retrosynthesis model with an automated robotic system to create molecules remotely. Simply type the chemical you want to synthesize and the combination of AI, cloud technology, and chemistry automation will do the rest. AI models have been trained to recognize the right sequence of operations for a particular target molecule.
Others labs use a combination deep learning and automation to decide which experiments to conduct. An AI-powered robot called Eve was developed by scientists from the Universities of Manchester and Aberystwyth in the UK. It identified triclosan, a common toothpaste ingredient, as an anti-malarial drug potential. Eve used a yeast-based method to generate and test hypotheses, and then ran follow-up high-throughput experiments in order to identify molecules that could target a malaria parasite enzyme.
Andrew Cooper, University of Liverpool, published recent work by his mobile robot scientist, who can also beat the best PhD students, working 21 hours per day. A search algorithm is used to plan experiments, based on previous experiments. The robot discovered a new catalyst from 688 experiments over 8 days. The robot, weighing 400 kg, can be used in the laboratory alongside humans to access samples vials and use the same equipment and instruments as human scientists. It remains to be seen if such robot scientists are able to keep up with lab gossip.
“It feels like a turning moment where [automation]will be used more often,” says Slater. “I think these tools can be used in new, unexpected ways and will impact both the quality of science done and the kinds of advances that are made. But it’s not going solve all the problems that chemists face!”
