A unique model for studying the structure of heterogeneous catalysts was developed at the Zelinsky Institute

Homogeneous catalyst systems are generally considered to be "well-defined". The precursor of such a catalyst is often a single molecular structure that has been accurately characterized using various instrumental methods. Heterogeneous catalysis is more difficult to study. The available analytical approaches make it possible to characterize only a small area of the initial catalyst, but the structure of the metal centers of the supported catalysts is extremely inhomogeneous. Nevertheless, heterogeneous catalysts are of great importance in industry: they are easy to prepare and handle, stable, inexpensive, and their properties can be tuned, including by varying supports.

Scientists at the Laboratory of Metal Complex and Nanoscale Catalysts have taken a significant step forward to fully characterize individual particles of supported Pd/C catalyst. They used a nanomanipulator device that allows high-precision movement at the micro level with a field emission scanning electron microscope with minimal impact on the sample under study to directly select catalyst particles. A series of electron photos were taken to characterize the apparent surface area of the catalyst particles, which were further analyzed using artificial intelligence. By combining a neural network approach to nanocatalyst characterization and catalyst preparation through nanomanipulation, the researchers were able to increase the catalyst turnover rate (the number of reactant molecules converted per catalyst molecule per second) in the Suzuki cross-coupling reaction to 109 using only Pd/C catalyst particles with 1 µm size. This is an unprecedented value for heterogeneous catalysis. Thus, the methodology for studying supported catalysts developed at the ZIOC has a great potential for mechanistic studies and the development of highly efficient catalysts.

Source:

Dmitry B. Eremin, Alexey S. Galushko, Daniil A. Boiko, Evgeniy O. Pentsak, Igor V. Chistyakov, Valentine P. Ananikov Toward Totally Defined Nanocatalysis: Deep Learning Reveals the Extraordinary Activity of Single Pd/C Particles J. Am. Chem. Soc. 2022, 144, 6071–6079. DOI: 10.1021/jacs.2c01283.