In a groundbreaking advance that could revolutionize the way humanity produces one of its most essential agricultural chemicals, researchers at the University of New South Wales (UNSW) Sydney have harnessed artificial intelligence (AI) and machine learning to dramatically enhance the production of green ammonia. Ammonia, a nitrogen-rich compound critical for fertiliser production, underpins the global agricultural industry and has been credited with averting widespread famine during the 20th century. However, its traditional manufacture remains an energy-intensive process responsible for substantial carbon dioxide emissions, contributing approximately two percent of global greenhouse gases. This new development not only offers a sustainable alternative but also brings ammonia production into the modern era of efficient, low-carbon chemical synthesis.
The conventional Haber-Bosch process, developed over a century ago, requires extreme conditions—temperatures exceeding 400°C and pressures more than 200 times that of Earth’s atmosphere—to convert atmospheric nitrogen and hydrogen into ammonia. These harsh operational parameters demand enormous energy input, generally derived from fossil fuels, thereby entrenching ammonia production as a significant emitter of greenhouse gases. In an earlier breakthrough in 2021, the UNSW team demonstrated a novel method to synthesize ammonia using only air, water, and renewable energy sources, operating at ambient temperatures roughly equivalent to a warm summer day. While pioneering, this first proof-of-concept left ample room for process optimization and efficiency gains.
The central challenge that Dr. Ali Jalili and his colleagues faced was increasing the yield and energy efficiency of green ammonia production. Central to this was the identification of an optimal catalyst—a substance that accelerates the ammonia-forming chemical reaction without being consumed. Previous research suggested that 13 different metals possessed individual properties conducive to facets of the reaction, such as nitrogen or hydrogen absorption. Yet, the combination potential among these metals resulted in over 8,000 possible alloys, making experimental testing of each combination an impractical endeavor.
.adsslot_26ABiUhxFK{width:728px !important;height:90px !important;}
@media(max-width:1199px){ .adsslot_26ABiUhxFK{width:468px !important;height:60px !important;}
}
@media(max-width:767px){ .adsslot_26ABiUhxFK{width:320px !important;height:50px !important;}
}
ADVERTISEMENT
To circumvent this challenge, the UNSW team leveraged machine learning algorithms capable of analyzing the chemical behaviors of each metal and predicting synergistic combinations most likely to deliver superior catalytic performance. By training the AI with data derived from theoretical and experimental sources, the system shortlisted only 28 promising multi-metal catalysts for laboratory validation, thereby condensing thousands of potential experiments into a highly efficient and targeted testing regime. This approach drastically reduced both time and resource expenditure while maximizing the likelihood of discovering a superior catalyst.
The results exceeded all expectations. A novel five-metal alloy composed of iron, bismuth, nickel, tin, and zinc emerged as the most effective catalyst. This sophisticated high-entropy metal alloy facilitated a sevenfold increase in ammonia production rates relative to previous attempts. Moreover, the process exhibited nearly 100% Faradaic efficiency, a key metric indicating that virtually all electrical energy input was utilized to produce ammonia, with negligible wastage. Such efficiency gains herald a new era in which green ammonia production can be economically competitive with conventional Haber-Bosch methodologies.
Crucially, this green ammonia synthesis functions at an ambient temperature of approximately 25°C, less than one-tenth the temperature required by traditional industrial processes. The implications of this low-temperature operation are profound: reaction vessels and industrial infrastructure can be downsized, safety concerns related to high-pressure operation are mitigated, and the overall energy footprint is drastically reduced. These characteristics empower scalable and decentralized ammonia production, breaking away from the century-old paradigm of massive centralized industrial complexes.
Dr. Jalili envisions a near future where farmers no longer depend on large-scale manufacturing and complex supply chains to obtain ammonia fertilisers. Instead, modular, factory-built compact units—approximately the size of shipping containers—can be deployed directly on farms or in local communities. These plug-and-play systems integrate the AI-optimized catalyst with plasma generators and electrolysers, enabling onsite ammonia generation with minimal energy and capital investment. Such decentralization promises to eliminate transportation emissions, reduce costs, and bolster energy resilience within agricultural sectors worldwide.
Beyond fertiliser production, this innovation holds transformative potential for the burgeoning hydrogen economy. Ammonia, owing to its high hydrogen content and ease of liquefaction at ambient pressure, serves as a superior hydrogen carrier compared to liquid hydrogen itself. This property positions green ammonia as an ideal medium for renewable energy storage and transport, bridging current gaps in hydrogen infrastructure and economics. The ability to produce ammonia efficiently and sustainably thus opens new pathways for decarbonizing heavy industry, transportation, and energy storage systems.
The research team is actively deploying these AI-engineered catalysts within distributed ammonia modules, accelerating commercial uptake and cost-competitiveness. Their work, published in the prestigious journal Small, elucidates the catalyst’s molecular configuration and performance metrics, paving the way for further refinements and applications. Supported by the Australian Research Council and the ARC Discovery Early Career Research Award, the project exemplifies the convergence of artificial intelligence, materials science, and green chemistry to drive industrial sustainability.
As the world grapples with the imperative to reduce greenhouse gas emissions, this breakthrough signals a paradigm shift in one of the planet’s most carbon-intensive industries. By integrating cutting-edge computational tools with innovative chemistry, the UNSW Sydney researchers have provided a blueprint for transforming ammonia from a pollutant-intensive product into a pillar of sustainable agriculture and clean energy. The future of green ammonia promises to be not only more environmentally responsible but also more accessible, affordable, and adaptive to the dynamic needs of global food and energy systems.
Subject of Research: Not applicable
Article Title: Configuring a Liquid State High-Entropy Metal Alloy Electrocatalyst
News Publication Date: 17-Jun-2025
Web References:
UNSW news article on eco-friendly ammonia
Article DOI: 10.1002/smll.202504087
Haber-Bosch method – Wikipedia
References:
Ali Jalili et al., “Configuring a Liquid State High-Entropy Metal Alloy Electrocatalyst,” Small, 2025. DOI: 10.1002/smll.202504087
Image Credits: Not provided
Keywords: Ammonia, Green chemistry, Industrial chemistry, Sustainable agriculture, Renewable energy, Hydrogen fuel, Artificial intelligence, Catalysis
Tags: AI in sustainable agricultureenergy-efficient ammonia productionenvironmental impact of ammonia productiongreen ammonia production technologyinnovative research in agricultural chemicalsmachine learning in chemical engineeringmodernizing the Haber-Bosch processnitrogen-rich compounds in agriculturereducing carbon emissions in ammonia synthesisrenewable energy in chemical synthesissustainable fertilizer production methodsUniversity of New South Wales sustainability initiatives
