Our new paper is out in Nature Catalysis! A huge congratulations to Seungchang Han!

Two years ago, I proposed theoretical descriptors for CO₂RR product selectivity—specifically the potential of zero charge (work function) and CO adsorption energy based on electric double layer aware DFT calculations of bare transition metals (https://lnkd.in/gWqFyKVt). The question that remained was: How do they perform in actual catalyst design? This motivated the group of Jihun Oh (KAIST) to design new non-Cu metal alloys. In our collaborative work, the experimental group showed that these descriptors can successfully predict catalysts optimized for CO and formate production. But there’s a twist: none of the materials produced C₂ products, even when matching Cu-like descriptor values. Further theory from my PhD student Seungchang Han revealed that surface heterogeneity introduces adsorption sites that block specific key steps toward C₂ formation. Copper remains unique in its homogeneous distribution of active sites—it binds CO₂ strongly, retains CO on the surface, and enables efficient CO–CO coupling. The takeaway: descriptor-based models must evolve to capture multi-site, multi-element complexity before being deployed in high-throughput and AI-driven catalyst discovery.