Selective hydrogenation of α,β-unsaturated aldehydes into unsaturated alcohols is a process in high demand in organic synthesis, pharmaceuticals, and food production. This process requires the precise hydrogenation of C═O bonds, a challenge that requires a tailored catalyst. Single-atom alloys (SAAs), where individual atoms of one metal are distributed in a host metal matrix, offer a potential solution to this challenge. Nevertheless, identifying the appropriate SAA capable of targeted adsorption and the efficient activation of C═O bonds remains a substantial hurdle. In this work, we synergistically combine density functional theory (DFT) calculations, active learning, and microkinetic simulations to design SAAs for the efficient and selective hydrogenation of α,β-unsaturated aldehydes. We first comprehensively assessed the potential of 66 SAAs across 264 surfaces (including (100), (110), (111), and (320) crystal planes), to gauge their potential in activating C═C and C═O bonds. Our assessment unveiled the excellent selectivity of the Ti1Au SAA in activating C═O bonds. Moreover, our detailed DFT calculations further demonstrated the high catalytic activity of Ti1Au(320) and Ti1Au(111) surfaces with a low activation energy barrier of only 0.60 eV. Subsequently, we conducted microkinetic simulations on the selective hydrogenation process of crotonaldehyde, by selecting Ti1Au (320) and (111) surfaces as the catalysts and demonstrated that they exhibited a remarkable selectivity and nearly 100% conversion toward crotyl alcohol in the temperature range from 373 to 553 K. The present study not only reveals novel SAAs for targeted hydrogenation of α,β-unsaturated aldehydes but also establishes a promising path toward efficient design of selective hydrogenation catalysts more broadly.
Keywords: C═O selective hydrogenation; active learning; density functional theory; microkinetic simulations; single-atom alloy catalysts.