Rationalizing the d-Band Model from Theory to Practice in Catalyst Design

Catalyst design remains central to the industrial implementation of catalytic processes, and the descriptor-based approach marks a pivotal milestone toward more rational design strategies. Among various descriptors, the d-band has emerged as one of the most thoroughly investigated and theoretically grounded concepts, underpinning fundamental models of adsorbate-metal interactions. Despite substantial theoretical progress and numerous in-depth reviews, its practical application in catalyst design remains nontrivial due to the inherent complexity and variability of real catalytic systems. Hence, this Perspective is rooted in emerging tools and real-world catalysts, highlighting new insights rather than rehashing classical theory on the modulation, characterization, and interpretation of the metal d-band structure. We begin by revisiting the classical d-band model with constant d-band filling (rigid band) and elucidate how strain and ligand effects modulate the d-band center, thereby tuning catalytic performance via a bottom-up strategy combined with machine learning. We then discuss recent progress in experimental techniques that enable independent characterization of d-band center and filling. Finally, we examine the conflicting interpretations frequently reported in the literature and seek to contextualize them within a flexible-band framework that accounts for variable d-band filling, with the goal of developing a multidescriptor approach capable of effectively guiding catalyst design under realistic conditions. This Perspective aims to clarify recent advances, identify key challenges, and provide a forward-looking outlook for descriptor-based catalyst design.