By Jeff Knox, Senior Associate
Each year the world produces more than 180 million metric tonnes of ammonia. An essential ingredient to modern fertilizer, it’s critical to the way we feed the world. Tomorrow, we have even greater hopes – ammonia as a stable means of storing energy could help ship renewable energy from the places where it’s made to the places where it’s used. It is surprising, then, that ammonia production is still largely done with a century-old process that requires tremendous temperature and pressure, yet still gets poor conversion efficiency – bad for business and bad for the planet. What if we could do better?
Copernic Catalysts is doing just that by bringing computational power to catalyst design to improve ammonia synthesis and other processes – making them more efficient, more sustainable, more economical. Chemical catalysts – ingredients to a chemical reaction that increase its speed without themselves being consumed – are time-tested science, but designing high-performance catalysts is challenging. Copernic’s approach to computational catalyst design, powered by proprietary data generated in-house, we believe will allow them to discover new and better catalysts much more quickly than has been traditionally possible. When it does founders Jacob Grose and Aruna Ramkrishnan, with their deep experience in the science and business of specialty chemicals, will be the right team to ensure those new catalysts move quickly into the hands that will put them to work making all ‘colors’ of ammonia more cheaply and efficiently. Under their leadership, Copernic embodies a sterling example of Innospark’s vision of AI for good.
The Problem
A simple, four-atom molecule, ammonia is a staple ingredient to modern fertilizer as a source of nitrogen. Without synthesized ammonia, we could not feed the world as we do today. The manufacture of ammonia was, and is, done via the Haber-Bosch process. This circa 1910 innovation combines elemental nitrogen and hydrogen under great temperature and pressure. Today it’s done in large quantity at huge plants. The product is ‘grey’ ammonia when the required energy comes from fossil fuel, ‘blue’ ammonia if fossil fuel energy is combined with carbon capture, and ‘green’ ammonia if the process is powered with renewable energy. However, in all cases the underlying process is the same and suffers from a relatively low per-pass conversion efficiency. Other critical chemicals leave similar room for better synthesis.
Enter catalysis. It’s long been known certain additives, often metals, help speed up chemical reactions by reducing the amount of energy required to drive them forward, their activation energy. Graphs like this one are the classic depiction. The action of a given catalyst is dictated in part by chemical properties – how well can we facilitate the rearrangement of atoms? – and in part by physical structure – how well do the reacting molecules physically maneuver against the catalytic ones? That combination means a catalyst’s exact composition, its macrostructure, and its microstructure all play a role in determining its overall performance.
The many possible combinations of those influencing factors create a wide variety of possible catalysts which, along with the complex process of catalyzed chemistry, makes the design of catalysts challenging even for experts. Compounding the problem, it’s difficult to check the performance of new designs in an efficient way. As a result, exploring the universe of possible catalysts for a given process has traditionally been slow and expensive.
The Copernic Solution
We’re excited about Copernic Catalysts because they’re showing the way to a faster, more efficient means of discovering novel catalysts that in turn allow for more efficient, more productive chemical manufacturing. The Copernic solution combines a custom experimental approach to generating novel data with a computational approach that leverages data to do better catalyst design.
Copernic’s custom experimental approach allows investigation of several candidate catalysts in parallel. Their approach allows for rapid generation of high-quality chemical data at scale. The proprietary data they generate fills what is otherwise a gap in the publicly available data in the field.
Leveraging the proprietary data generated by their experimental platform, Copernic partners with Schrödinger to create computational models of the catalytic chemistry they are seeking to improve. Using these models, and applying some expert filters around which classes of catalysts might satisfy commercial production requirements, the company can then choose the next generation of catalysts to design for experimental testing.
By iterating around this design-build-test-learn loop Copernic will produce best-in-class catalysts that drive efficiency in industrial processes. The catalysts they produce are drop-in solutions that will allow producers of all ammonia ‘colors’ to make more product with less energy. Better for the environment, better for business.
Into the Future of AI-enabled Catalyst Design
By applying computational methods to the challenging materials engineering problem of catalyst design, Copernic Catalysts is poised to both dramatically accelerate the pace of discovery and significantly improve the resulting catalyst. Their work will create products with strong, standalone business cases that also move us toward a lower carbon future. Jacob, Aruna, and team have developed the data and technology to make this computational approach possible and they are poised to quickly translate their work to industry impact. We’re excited to support Copernic Catalysts as they harness computational intelligence to revolutionize manufacturing and build a more sustainable future.
Source
- Annual ammonia production, Congressional Research Service; https://crsreports.congress.gov/product/pdf/IF/IF12273/1