Machine learning and catalyst engineering for sustainable hydrogen release from ammonia

The urgent demand for sustainable energy has directed attention toward ammonia decomposition as a promising route for carbon-free hydrogen production. This review explores the fundamental reaction mechanism of ammonia cracking, highlighting the sequential dehydrogenation of NH₃ and the critical role of noble metal catalysts, particularly ruthenium, in enabling high activity and stability. Strategies to enhance catalytic performance, including particle size control, promoter addition, support engineering, and external energy integration, are discussed in detail. Ruthenium-based systems remain the benchmark due to their superior low-temperature activity, though other noble metals and alloy catalysts also show potential. Machine learning emerges as a powerful tool to accelerate catalyst discovery and mechanistic understanding, with Gradient Boosting Regression models demonstrating the most accurate predictions for ammonia conversion and hydrogen formation. The integration of mechanistic insights, advanced synthesis strategies, and data-driven modelling provides a robust framework for designing efficient, scalable and cost-effective catalysts to advance the hydrogen economy.