Meet the future autonomous laboratory

April 17, 2023

(Nanowerk News) To accelerate the development of useful new materials, researchers are building a new kind of automated laboratory that uses robots guided by artificial intelligence.

“Our vision is to use AI to discover the materials of tomorrow,” said Yan Zeng, staff scientist who leads A-Lab at the Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab). The “A” in A-Lab is intentionally ambiguous, standing for artificial intelligence (AI), automated, accelerated, and abstracted among other things.

Scientist has computationally predicted hundreds of thousands of new materials which may hold promise for new technologies – but testing to see if those materials can be made in reality is a slow process. Enter A-Lab, which can process 50 to 100 times more samples than a human every day and uses AI to rapidly pursue promising discoveries.

A-Lab can help identify and quickly track materials for several research areas, such as solar cells, fuel cells, thermoelectrics (materials that generate energy from a difference in temperature), and other clean energy technologies. For starters, researchers will focus on discovering new materials for batteries and energy storage, addressing the critical need for an affordable, equitable and sustainable energy supply. Berkeley Lab researcher Yan Zeng sees a starting point at A-Lab. The new lab combines automation and artificial intelligence to accelerate materials science discoveries. (Image: Marilyn Sargent, Berkeley Lab)

Once a target material is selected – by human researchers or their AI agents – a series of robots perform the steps in A-Lab to synthesize it:

The first robots weighed and mixed various combinations of starting materials known as powder precursors. The robot can choose from nearly 200 precursors, including various metal oxides containing elements such as lithium, iron, copper, manganese and nickel. After mixing the powder with the solvent to distribute it evenly, the robot transfers the slurry into the crucible.

The next robotic arm loads the crucible into a furnace that can reach 2200 degrees Fahrenheit and injects various mixtures of gases, such as nitrogen, hydrogen, oxygen and air. This allows ingredients to bake in different environments and have different properties. The AI ​​system determines at what temperature the samples should be baked, and for how long.

After the robot removes the baked crucible, it must extract a new ingredient. An automatic machine modeled on the gumball dispenser adds ball bearings to the cup. The intense shaking grinds the new substance into a fine powder that the robot loads onto the slide.

The final robotic arm transfers the samples into two automated machines for analysis. The X-ray diffractometer determines whether one or more new chemicals have formed, and how much of the starting materials remain. The automated electron microscope performs further shaping 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 the new starting combination and precursor count and instructions for the furnace. Researchers monitor the system via video feeds and alerts that can signal success, such as if a sample returns with the desired result, or if the robot finds an error.

“Some people might compare our setup to manufacturing, where automation has been in use for a long time,” says Zeng. “What I find interesting here is that we have adapted to a research environment, where we never know the outcome until the material is produced. The entire setup is adaptive, so it can handle a changing research environment, not always do the same thing.”

The systems at A-Lab are designed as “closed loop”, where decision making is handled without human intervention. Robots operate around the clock, freeing up researchers to spend more time designing experiments.

“We saw this as a new way of doing research,” said Gerd Ceder, A-Lab principal investigator. In many ways, says Ceder, laboratory research has remained much the same over the past 70 years: the equipment may have gotten better, but ultimately it takes someone to take the measurements, analyze the results, and decide what to do next.

“We need material solutions to things like the climate crisis that we can create and implement now, because we can’t wait – so we are trying to break this very slow cycle by having a machine that repairs itself,” said Ceder. “What’s important is not working in parallel, but switching 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 considered to be the first automated lab to use inorganic powders as starting materials. This “solid-state synthesis” is a more difficult task than automating processes that use liquids, which can easily be delivered with pumps and valves. But the extra effort comes with big payoffs.

“Our solid-state synthesis is more realistic, can incorporate a wider variety of materials, and can manufacture larger quantities of materials,” said Ceder. “You can generate quantities that are ready to apply, not just science exploration. It is ready to scale.”

A-Lab researchers must adapt hardware and software to robots, furnaces and analysis tools, making them perform certain actions and talk to a central hub controlled by AI. In some cases, such as a whisk to dispense a freshly baked ingredient, they have to create a new solution entirely from scratch.

When automated systems create and analyze samples, data flows back to A-Lab researchers as well as data stores such as Project Materials. The scientists are also building integrations with other projects, such as, and leveraging the x-rays of Berkeley Lab’s powerful synchrotron, Advanced Light Source.

“You can imagine the power of a laboratory that independently starts with predictions, requests data, and calculations to get the information it needs, and then continues,” said Zeng. “When A-Lab tests materials, we will study gaps between our calculations and reality. It not only gives us some useful new material, but also trains 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 DOE’s Office of Science and Laboratory Directed (LDRD) Research & Development Program (LDRD), which encourages innovative ideas and experimentation. Zeng and a team of 10 students and postdoctorals began building the lab in earnest in early 2022 and installed the final pieces more than a year later.

A-Lab started operations in February and has synthesized several new materials in collaboration with the Materials Project. Researchers are currently refining the system while continuing to add features. These include robots that can replenish stocks and change precursors, synthesis instruments that allow them to mix and heat liquids, and additional equipment to analyze newly created materials.

Source link

Related Articles

Back to top button