Out of the box: The potential of AI in electrochemistry

In this series, we examine the potential of AI in material science across three areas of Topsoe’s research and development: batteries, electrochemistry, and heterogeneous catalysis. Drawing on insights from leading AI-for-science startups and Topsoe scientists, we explore AI’s potential to redefine energy – and the very materials it depends on.

What kind of challenges currently exist in material innovation, and how urgent is the need to address them?

By leveraging the rapidly expanding infrastructure of renewable power, electrochemical solutions play an essential role in diversifying our current energy systems as well as contributing to a more secure and resilient future of energy. The development of electrochemical solutions, such as electrolysis technologies, e-fuels and e-chemicals, has expanded rapidly over the past decade. In this field, efficiency gains, profitability, and reliable production are everything. While scaling these technologies is complex and dependent on many factors, the materials used within these technologies are the most fundamental building block that defines their success. The efficiency of these materials, their cost, abundance, and accessibility – all these factors are instrumental in successfully scaling these technologies.

Now, with the emerging promise of artificial intelligence-enabled material discovery, we may be on the cusp of dramatically expanding the material set that is used to build the next generation of technologies for the energy transition. With insights from Topsoe’s Head of Technology Scouting & Partnerships, Sune D. Ebbesen, we explore the current state of material innovation in electrochemistry, the challenges we face, and possibilities AI could unlock if applied to address them.

The complex world of electrochemistry

Electrochemistry is the study of how chemical reactions and electricity affect each other, covering processes where chemical reactions produce electrical energy, or electrical energy drives chemical change. E-SAF, e-ammonia, e-methanol and the wider family of e-fuels are all made through the technologies, catalysts and materials developed through this field of study.

Since 2007, Sune Ebbesen has been working in electrochemistry and solid oxide electrolysis cell research, witnessing first-hand how the field has evolved over decades of technological innovation and improvement. Electrochemistry requires a unique symbiosis of chemistry and engineering working in tandem, with experiments requiring parallel testing of chemical reactions and the physical structures where these reactions take place. “In classic catalysis testing, we are making and testing catalysts in a reactor that we do not necessarily design,” explains Ebbesen. “In electrocatalysis, we are not only testing the many variables that go into catalyst composition, but we also have to factor in – and test – the variables introduced by the extremely complex cell structures in which these reactions are happening.”

The state of material innovation in electrochemistry

Over the years, the design of cell containers and their different layers have been continuously optimized, allowing for lower temperatures, higher durability, and more cost-effective materials to be integrated. The electrochemical reactions that happen within these stacks, however, have not seen the same level of material innovation.

“We have used the same materials for the last 20-30 years,” explains Ebbesen. “We have done this because it is a tried, tested and a good system. However, it also means that the past 30 years of work in material science for electrochemistry – everything from in confronting challenges to optimizing performance – has been on this specific material set.”

While the material sets currently used in electrochemical solutions today are consistently able to deliver efficiency and high-performance, there is still massive potential to improve and ensure e- and green solutions can reach the scale needed for sizeable impact.

The materials problem of critical raw minerals

While Topsoe's electrolysis technology has been designed to minimize the need for rare-earth metals, many technologies central to a diversified future of energy heavily rely on non-renewable, high in demand materials. According to the International Energy Agency (IEA), the demand for critical raw minerals will expand by as much as 60% by 2040, with a large part of this demand attributed to the expansion of clean energy infrastructure. As demand rises, supply concentrations for rare earth minerals are seen as a key vulnerability in achieving decarbonization ambitions. Expanding the material set away from critical raw materials such as cobalt could avoid the risks of building energy infrastructure that relies on finite and non-renewable resources. Even with comparably more abundant materials like nickel, there is much potential that can be unlocked by finding alternative materials.

“Nickel, for example, is a material that is used in electrodes that reduces steam in electrolysers, can become quite mobile and moves around with increased output and production rates,” says Ebbesen. “It would be of course ideal if we could mitigate this challenge, but the reason we use nickel is because we simply have not found a higher-performing alternative. If we are really going to find something better, we have to think completely out of the box.”

It is in this out-of-the-box-thinking that AI systems have started to thrive. With the ability to rapidly simulate, test, and validate the vast number of combinations yet to be made in the periodic table, AI systems can expedite the discovery of new materials. Open-source data bases have already been discovering hundreds of new materials within months of operation, but with expert guidance and design, these systems may be able to discover alternatives that can support or even replace the role of critical raw materials in energy technologies.

This may take the form of:

1. Identifying substitution opportunities by discovering materials with similar composition and structure.
2. Optimizing material combinations so they require a lower percentage of critical raw materials in their composition.
3. Discover materials that have a higher degree of durability and reusability to create technologies with a higher degree of material circularity.

Now that we have systems that can start developing these new material candidates, it is the right time to start training them with the industrial knowledge needed to validate them. As Ebbesen explains: “It is the right time for electrochemistry to explore AI and vice versa. Both fields have matured to the point where they are complementary. It is time to see what they can unlock when combined.”

In our next article, we dive deeper into the potential of AI in material science by exploring the work of Entalpic, a start-up working with AI to design new materials and processes that drive the energy transition. Together with Topsoe, they are discovering what AI could unlock in the field of LNMO batteries.