Machine Learning\nBoosted Entropy-Engineered Synthesis\nof CuCo Nanometric Solid Solution Alloys for Near-100% Nitrate-to-Ammonia\nSelectivity

Nanometric solid solution alloys are utilized in a broad\nrange\nof fields, including catalysis, energy storage, medical application,\nand sensor technology. Unfortunately, the synthesis of these alloys\nbecomes increasingly challenging as the disparity between the metal\nelements grows, due to differences in atomic sizes, melting points,\nand chemical affinities. This study utilized a data-driven approach\nincorporating sample balancing enhancement techniques and multilayer\nperceptron (MLP) algorithms to improve the model’s ability\nto handle imbalanced data, significantly boosting the efficiency of\nexperimental parameter optimization. Building on this enhanced data\nprocessing framework, we developed an entropy-engineered synthesis\napproach specifically designed to produce stable, nanometric copper\nand cobalt (CuCo) solid solution alloys. Under conditions of −0.425\nV (vs RHE), the CuCo alloy exhibited nearly 100% Faraday efficiency\n(FE) and a high ammonia production rate of 232.17 mg h<sup>–1</sup> mg<sup>–1</sup>. Stability tests in a simulated industrial\nenvironment showed that the catalyst maintained over 80% FE and an\nammonia production rate exceeding 170 mg h<sup>–1</sup> mg<sup>–1</sup> over a testing period of 120 h, outperforming most\nreported catalysts. To delve deeper into the synergistic interaction\nmechanisms between Cu and Co, in situ Raman spectroscopy was utilized\nfor real-time monitoring, and density functional theory (DFT) calculations\nfurther substantiated our findings. These results not only highlight\nthe exceptional catalytic performance of the CuCo alloy but also reflect\nthe effective electronic and energy interactions between the two metals.