figure a

A photograph of the ASTRAL robotic laboratory in Cambridge, Mass., which the researchers used to validate their principles of precursor selection. The robot arm obtained precursors from a library, mixed them in the rotary mixer, fired them in the vacuum oven, and then characterized the products with x-ray diffraction. Credit: Nature Synthesis.

Inorganic synthesis seems so simple on paper: combine two or more precursors to create a target product. In practice, however, it is a process riddled with complications. All too often, two of the precursors will waste energy forming an intermediate, leaving the reaction without enough energy to reach the desired target. Jiadong Chen of the University of Michigan and colleagues have studied an alternative approach: find precursors that would result in high-energy intermediates, leaving a reaction with enough energy to reach the end. They successfully validated their approach in a robotic laboratory, allowing them to test far more reactions than possible under a manual approach.

“What the robotic lab really helped us do is demonstrate this concept over a very broad chemical space,” says Wenhao Sun, a materials scientist at the University of Michigan and one of the authors who published this study in a recent issue of Nature Synthesis (https://www.nature.com/articles/s44160-024-00502-y).

The Michigan group and their collaborators at Samsung Semiconductor Inc., Cambridge, Mass., had previously shown that when they combined three solid powders, only two reacted at a time: starting with the pair that reacted with the largest driving force.

With this idea in mind, the researchers laid out their novel approach for choosing reagents. To synthesize a given target compound, they would select three precursors such that two would react first; only then would the team add a third to the resulting intermediate, which would still have enough energy to reach the target compound. The researchers began applying these principles to recipes for lithium-, sodium-, and potassium-based oxides.

They put their approach to the test in a Samsung robotic ceramics synthesis laboratory. Sun says that if the researchers had demonstrated their principles manually, they could have been able to try them on three to five systems. Instead, the robotic laboratory allowed the researchers to validate their theory across 35 oxides—an amount that Sun says would have once been limited to a review paper.

“I believe the work described in this paper is a significant advance,” says Chris Bartel, a materials scientist at the University of Minnesota who was not involved in the work. “This work also reinforces the promise of automated experimental platforms to accelerate our understanding of synthesis science,” he says.

David Johnson, a chemist at the University of Oregon who was also not involved in the work, says that the robotic laboratory’s ability to speed up multiple synthesis runs is a great cost-saving advance. However, he cautions that the fastest reactants may not always be the most practical for use on commercial scales. “In the end, when you’re going to make a solid, if you have to make kilotons, then you’re going to look for the cheapest reactants you can, not necessarily the ones that react fastest,” he says. “There’s lots of other considerations that come into play.”

For the researchers’ part, Sun says they are continuing their work with the robotic laboratory to better understand the precursors, temperatures, and times needed for solid-state reactions in hopes of eventually crafting new materials. This method could one day yield materials for ceramics and electrodes.

Rahul Rao