Machine Learning Optimized Synthesis of Ag/Nitrodopamine-MWCNT Hybrid for Durable and High-Performance OER Catalysis

A wide range of multi-walled carbon nanotubes (MWCNTs)-based electrocatalysts have been reported for the oxygen evolution reaction (OER); however, their electrocatalytic performance is often limited by nanoparticle aggregation, poor structural integrity, and inadequate stability. To address these challenges, polymer-based linkers have been employed to mitigate nanoparticle aggregation, thereby enhancing structural robustness, interfacial binding, and durability. Despite this, maximizing electrocatalytic efficiency still requires precise and often labor-intensive optimization of the linker-catalyst interface. To overcome this challenge, herein we incorporated a machine learning (ML) assisted approach into the synthesis strategy, integrating nitrodopamine (NDA) as a molecular linker to enhance structural stability. Accordingly, we synthesized a composite material comprising MWCNTs functionalized with NDA-cross-linked silver nanoparticles (AgNPs). This strategy enabled precise tuning of NDA and Ag precursor concentrations, effectively reducing nanoparticle aggregation, enhancing structural integrity, and resulting in improved electrocatalytic performance and stability under oxidative conditions. Interestingly, the ML-optimized Ag/NDA/MWCNTs nanocomposite demonstrated enhanced electrocatalytic activity, achieving a low overpotential of 220 mV at 10 mA cm⁻² and a Tafel slope of 64.4 mV dec⁻¹, outperforming other electrocatalysts, including those with polydopamine linkers (Ag/PDA/MWCNTs) and without linkers (Ag/MWCNTs). The resulting electrocatalyst has shown high electrocatalytic activity and stability, highlighting the potential of ML-optimized design not only for advancing sustainable energy applications but also for guiding future, scalable developments in electrocatalyst design.