A high-throughput co-exploration strategy reveals catalyst-reaction combinations that are missed in conventional target-driven screening processes

Conventional catalyst research has mainly focused on incrementally improving catalysts for known reactions or optimizing reaction conditions for known catalysts. A new study demonstrated a new catalyst-catalysis co-exploration strategy that searches for catalysts and reaction conditions simultaneously, rather than optimizing catalysts for a predefined reaction. Using methane conversion as a test case, the team evaluated 200 catalysts, which revealed high-performance catalyst-catalysis combinations.

Catalysts are the hidden engines of modern manufacturing, directly involved in more than 80% of chemical processes. However, catalyst development is highly complex because performance is governed by the interplay of the catalyst, local operating conditions, reactant composition, and product formation.

Conventional catalyst research usually begins by defining a target reaction and then searching for better catalysts to improve that reaction. However, this target-driven approach can miss valuable catalytic processes that appear only under certain combinations of catalyst and reaction conditions.

To address this limitation, researchers at the Japan Advanced Institute of Science and Technology (JAIST), in collaboration with the National Institute for Materials Science, Japan (NIMS), have demonstrated a new strategy of catalyst-catalysis co-exploration. The study was led by Professor Toshiaki Taniike, Professor at JAIST, along with Research Associate Professor Patchanee Chammingkwan from JAIST and Dr. Ryo Tamura, Group Leader at NIMS. The findings of the study were published in ACS Catalysis on July 8, 2026. "Our approach can support the development of low-carbon, efficient routes for producing fuels, plastics, and chemical feedstocks in the future," mentioned Prof. Taniike, talking about the motivation behind the study.

The researchers explored catalysts and reactions simultaneously by considering methane conversion as a broad reaction space for converting methane into useful compounds. The CH₄-O₂-CO₂ ternary system, consisting of methane, oxygen, and carbon dioxide, was specifically targeted for the research as it involved multiple known reactions. The team evaluated a library of 200 catalysts, including single oxides, mixed oxides, and supported catalysts. Using a high-throughput reactor platform, each catalyst was examined under 25 different methane-oxygen-carbon dioxide compositions at 600 °C and 800 °C.

The uniqueness of the system was its non-targeted product analysis. Instead of selected expected products, the system was allowed to detect major products as well as minor and unexpected products, resulting in 1,000,000 data points. The results highlighted why catalyst and reaction exploration should not be separated. The feed composition that produced high performance differed greatly from catalyst to catalyst. A catalyst that appeared weak under one condition could show excellent activity under another. "Evaluating catalysts under only one predefined reaction condition can distort catalyst rankings and overlook promising materials," explained Dr. Chammingkwan.

Importantly, the method also revealed minor products, including 1-butene, 1,3-butadiene, and benzene. Their detection suggests that the catalyst-catalysis co-exploration system can reveal early signs of unknown reaction pathways.

The broader exploration space showed improved performance beyond conventional reaction-centered screening. For hydrocarbons such as ethylene and propylene, the maximum yield was about 27% for known reaction conditions but exceeded 30%, with selectivity exceeding 80%, when the broader reaction space was explored. Similarly, hydrogen yield increased from about 85% near conventional conditions to nearly 100% under the expanded exploration approach.

This new research framework is the first demonstration of 'reaction exploration,' which simultaneously searches for catalysts and reactions. "By combining high-throughput experimentation, broad reaction-space design, comprehensive product analysis, and eventually machine learning, this approach can accelerate the discovery of catalytic systems for sustainable chemical production and carbon-neutral technologies," explained Prof. Taniike.

In the long term, such systems may help shift catalyst research away from relying mainly on prior assumptions about which reactions should be improved. Instead, exploring large catalyst-reaction spaces can be useful to discover combinations beyond human intuition.

Image title: Concept of catalyst-catalysis co-exploration
Image caption: Conventional catalyst development usually fixes a target reaction first and then searches for better catalysts. In this study, catalysts and reactions are explored simultaneously across a broad CH₄-O₂-CO₂ reaction space. By combining high-throughput experimentation and non-targeted mass analysis, a one-million-point reaction dataset was obtained, revealing high-performance catalyst-reaction combinations and hints of previously overlooked reactions.
Credit: Professor Toshiaki Taniike from Japan Advanced Institute of Science and Technology (JAIST), Japan
Source link: N/A
License type: Original content
Usage restrictions: Cannot be used without permission.

Reference

Additional information for EurekAlert

Funding information

This work is funded by the Japan Science and Technology Agency (JST), MIRAI (Grant Number: JPMJMI25G1).

July 9, 2026