| Apr 17, 2023 |
Meet the autonomous lab of the future
(Nanowerk News) To accelerate development of useful new materials, researchers are building a new kind of automated lab that uses robots guided by artificial intelligence.
|
|
“Our vision is using AI to discover the materials of the future,” said Yan Zeng, a staff scientist leading the A-Lab at the Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab). The “A” in A-Lab is deliberately ambiguous, standing for artificial intelligence (AI), automated, accelerated, and abstracted, among others.
|
|
Scientists have computationally predicted hundreds of thousands of novel materials that could be promising for new technologies – but testing to see whether any of those materials can be made in reality is a slow process. Enter A-Lab, which can process 50 to 100 times as many samples as a human every day and use AI to quickly pursue promising finds.
|
|
A-Lab could help identify and fast-track materials for several research areas, such as solar cells, fuel cells, thermoelectrics (materials that generate energy from temperature differences), and other clean energy technologies. To start, researchers will focus on finding new materials for batteries and energy storage, addressing critical needs for an affordable, equitable, and sustainable energy supply.
|
 |
| Berkeley Lab researcher Yan Zeng looks over the starting point at A-Lab. The new lab combines automation and artificial intelligence to speed up materials science discovery. (Image: Marilyn Sargent, Berkeley Lab)
|
|
Once a target material is selected – by human researchers or their AI agents – a series of robots carry out the steps at A-Lab to synthesize it:
|
|
The first robot weighs and mixes different combinations of starting ingredients known as powder precursors. The robot can choose from nearly 200 precursors, including different metal oxides containing elements such as lithium, iron, copper, manganese, and nickel. After mixing the powders with solvent to evenly distribute them, the robot moves the slurry into crucibles.
|
|
The next robotic arm loads the crucibles into furnaces that can reach 2200 degrees Fahrenheit and inject various mixtures of gases, such as nitrogen, hydrogen, oxygen, and air. This allows the ingredients to bake in different environments and take on different properties. The AI system determines what temperature the samples should bake at, and for how long.
|
|
After the robot removes the baked crucibles, it must extract the new material. An automated machine modeled on a gumball dispenser adds a ball bearing to the cup. Intense shaking grinds the new substance into a fine powder that the robot loads onto a slide.
|
|
The final robotic arm moves the samples into two automated machines for analysis. The X-ray diffractometer determines whether one or more new chemicals have been formed, and how much of the initial ingredients are left over. The automated electron microscope does further shape and chemical analysis. Both tools send their results back to the AI system.
|
|
Guided by artificial intelligence, the cycle adjusts and begins again. The AI at the heart of the system sets new starting combinations and amounts of precursors and instructions for the furnaces. Researchers keep an eye on the system through video feeds and alerts that can flag successes, like if a sample comes back with a desired result, or if a robot encounters an error.
|
|
“Some people might compare our setup with manufacturing, where automation has been used for a long time,” Zeng said. “What I think is exciting here is we’ve adapted to a research environment, where we never know the outcome until the material is produced. The whole setup is adaptive, so it can handle the changing research environment as opposed to always doing the same thing.”
|
|
The system at A-Lab is designed as a “closed-loop,” where decision making is handled without human interference. The robots operate around the clock, freeing researchers up to spend more time designing experiments.
|
|
“We see this as a new way of doing research,” said Gerd Ceder, the principal investigator for A-Lab. In many ways, Ceder noted, lab research has been the same for the last 70 years: the equipment may have gotten better, but ultimately a person is needed to take measurements, analyze results, and decide what to do next.
|
|
“We need materials solutions for things like the climate crisis that we can build and deploy now, because we can’t wait – so we’re trying to break this cycle that is so slow by having machines that correct themselves,” Ceder said. “The important thing is not working in parallel, but instead to iterate rapidly, the way scientists operate. We want the system to try something, analyze the data, and then decide what to do next to get closer to the goal.”
|
|
A-lab is thought to be the first fully automated lab that uses inorganic powders as the starting ingredients. This “solid-state synthesis” is a more difficult task than automating processes that use liquids, which can be easily dispensed with pumps and valves. But the extra effort comes with a big payoff.
|
|
“Our solid-state synthesis is more realistic, can incorporate a wider variety of materials, and can make larger quantities of materials,” Ceder said. “You can produce quantities that are ready for application, not just science exploration. It’s ready to scale.”
|
|
A-Lab researchers had to adapt both hardware and software for the robots, furnaces, and analysis tools, getting them to perform certain actions and talk to the central hub controlled by the AI. In some cases, such as the shaker to remove the newly baked material, they had to build a new solution entirely from scratch.
|
|
As the automated system creates and analyzes samples, the data will flow back to both A-Lab researchers as well as data repositories such as the Materials Project. Scientists are also building out integrations with other projects, such as MaterialSynthesis.org, and leveraging x-rays from Berkeley Lab’s powerful synchrotron, the Advanced Light Source.
|
|
“You can imagine the power of a lab that autonomously starts with predictions, requests data and computations to get the information it needs, and then proceeds,” Zeng said. “As A-Lab tests materials, we’re going to learn the gap between our computations and reality. That will not only give us a handful of useful new materials, but also train our models to make better predictions that can guide future science.”
|
|
Work on A-Lab began in 2020, and the project later received funding from the DOE’s Office of Science and Laboratory Directed Research & Development (LDRD) Program, which encourages innovative ideas and experiments. Zeng and a team of 10 students and postdocs began building out the lab in earnest at the start of 2022 and installed the final piece a little over one year later.
|
|
A-Lab began operating in February and has already synthesized several novel materials in collaboration with the Materials Project. Researchers are currently fine-tuning the system while continuing to add features. These include robots that can restock supplies and change precursors, synthesis instruments that let them mix and heat liquids, and additional equipment to analyze newly created materials.
|
