Unravelling the Catalytic Activity of Dual-Metal Doped N₆-Graphene for Sulfur Reduction via Machine...

Mitigating polysulfide shuttling and sluggish redox kinetics is crucial for the practical utilization of lithium-sulfur batteries (LSBs). Using first-principles density functional theory calculations, we investigate a series of N₆-coordinated dual-atom catalysts (DACs) to identify efficient and cost-effective catalysts for the sulfur reduction reaction (SRR). Our results demonstrate that, compared with single-atom catalysts (SACs), DACs exhibit improved Li polysulfide adsorption and redox conversion through cooperative metal-sulfur interactions and frontier-orbital-mediated electronic coupling between adjacent metal centers. In particular, (N₃)Fe-Ni(N₃) and (N₃)Fe-Pt(N₃) show the most favorable SRR activity, with optimal adsorption energies (-1.0 to -2.3 eV), low free-energy changes (ΔG ≤0.5 eV) for the Li₂S₂ to Li₂S conversion, and facile Li₂S decomposition barriers (≤1.0 eV). Additionally, to accelerate catalyst screening, we introduce precise and accelerated configuration evaluation (PACE), a machine-learning-accelerated DFT workflow that integrates machine learning (ML) interatomic potentials with automated configuration evaluation. Furthermore, to rapidly predict the ΔG for unexplored DACs, we developed a regression model using physically interpretable descriptors. Finally, the electronic structure analyses reveal that the superior catalytic behavior arises from the unique nature of frontier orbitals that enables the dual metal centers to function as a "charge-transfer highway," optimal metal-metal covalent bonding that stabilizes reaction intermediates, and desirable d-band positioning that tunes adsorbate binding. This combined mechanistic and ML-DFT strategy offers general design principles based on the electronic and orbital fingerprints for identifying high-performance SRR catalysts for next-generation LSBs.