Machine learning-assisted engineering of dual-metal chalcogenide-MXene based composite for enhanced supercapacitor application

Transition-metal chalcogenides (sulfides and phosphides) have been extensively explored with MXene as supercapacitor electrode materials. However, these studies have largely been limited to binary systems (sulfide/MXene or phosphide/MXene), most of which suffer from low conductivity, sluggish kinetics, limited capacitance, and poor stability. Thus, there is a need for a ternary composite integrating both sulfides and phosphides with MXene to overcome these limitations. Unfortunately, optimizing the composition of such multicomponent systems presents an additional challenge, requiring intelligent strategies beyond structural innovation. To address this, we fabricated a bimetallic VS 2 /Ni 2 P nanostructure uniformly dispersed within Ti 3 C 2 MXene nanosheets via hydrothermal synthesis. This hybrid suppresses MXene restacking and generates abundant heterointerfaces for enhanced charge transfer and redox activity. Furthermore, we employed machine learning (ML) to systematically optimize the composite composition. The ML-optimized VS 2 /Ni 2 P@MXene delivered a high specific capacitance of 1747.03 F g −1 (CV) and 1561.87 F g −1 (GCD), a low charge-transfer resistance of 1.38 Ω, and 92.73% capacitance retention after 5000 cycles, which significantly outperforming individual VS₂, individual Ni 2 P, binary VS 2 /Ni 2 P, and pristine MXene. An asymmetric device achieved 54 Wh kg −1 at 600 W kg −1 . This enhanced performance arises from synergistic interfacial effects among the dual-metal chalcogenides and MXene. To our knowledge, this is the first report of a bimetallic VS 2 /Ni 2 P@MXene ternary hybrid and the first application of ML-driven optimization to such a system, thus establishing a data-driven paradigm for next-generation supercapacitor electrodes.