What are catalysts and what are they for?

New diagnostic methods for catalysts are being developed at the International Research Institute for Intelligent Materials of the Southern Federal University (SFedU).

Catalysts are substances that accelerate chemical reactions and do not consume themselves. Almost every industrial process has its own catalyst, for example, the Phillips catalyst for the synthesis of polyethylene, the Ziegler—Natta catalyst for the production of polypropylene, rhodium catalysts for the production of alcohols in hydroformylation, and platinum catalysts for hydrogen energy.

Despite the enormous role of catalysts in modern technological civilization, for a long time very little was known about the processes that occur with them at the atomic level. To understand how the active centers of catalysts are arranged, why they degrade and how to improve their activity, scientists use X—ray absorption spectroscopy, one of the most accurate non-destructive methods for diagnosing atomic structure.

Experiments are carried out on synchrotrons, megaclass installations that literally "look" inside a substance during its operation, for example, at high temperatures and gas pressure.

Interpreting the obtained spectra is a difficult task. Even with instructions and the right direction, the analysis can take several hours. It may take a year from the moment of measurement to the publication of the results, even with the participation of specialized specialists from quantum chemistry, spectroscopy and computer modeling.

What solution was implemented

To simplify and speed up spectrum analysis, the interdisciplinary team of the SFedU Institute of Physics and Technology, namely Bogdan Protsenko, Mikhail Lifar, Georgy Asaturov, Daniil Kupriyanenko, Nazar Chubkov, Kirill Kulaev, Georgy Kochiev, Alexander Guda, Sergey Guda, under the guidance of Professor Alexander Soldatov, scientific director of the SFedU department, developed an intelligent agent researcher. It is based on a large language model and trained on specially assembled databases using the author's PyFitIt framework.

The architecture of the solution does not depend on the specific LLM, the agent can work with different models. Neural networks with a large number of parameters are well suited for scientific tasks, for example LLaMa 3.2 (40B)[* owned by Meta, which is recognized as extremist and banned in the Russian Federation] and DeepSeek R1: they provide higher accuracy of logical reasoning and allow you to form informed and meaningful conclusions.