CATANA-2 : Cooperative modeling – experimental learning for the identification of low temperature catalysts for the synthesis and decomposition of NH3

This paradigm shift imposes new constraints on the temperature/pressure (T/P) operating conditions for synthesis, and consequently on the development of associated heterogeneous catalysts. Still with a view to using NH3 as a medium for storing H2, the cracking reaction must also be optimized from a catalytic point of view, as the process does not currently exist, yet the literature clearly shows that catalysts that are active for the cracking reaction are not necessarily the best for synthesis. The reverse is also true. The reverse is also true.

The objective of the CATANA-2 project is to identify new catalysts for the two stages of the process, NH3 synthesis and decomposition, by combining predictive approaches using DFT-machine learning (DFT-ML) with experimental approaches to catalyst synthesis and screening, and operando characterization. The combination of techniques will improve understanding of the reaction mechanisms, with the aim of optimizing the catalysts in relation to the T/P process conditions in order to minimize the energy requirements for each reaction. This process of accelerated discovery of more efficient catalysts, in a Power-to-NH3 scheme, should lead to the development of catalysts for (i) NH3 synthesis, compatible with H2 production by electrolysis (<50 bar, 300°C, with stability in stop-and-start mode), and (ii) the cracking of NH3 into H2 with a purity >99.7% under P<10 bar and T<550°C.

To achieve these objectives, a catalytic screening platform (REALCAT) will be used for catalyst synthesis and activity measurement (UCEIV-ULCO, UCCS-CNRS). Systematic characterizations will be performed to analyze their properties under reaction conditions to feed DFT-ML models, pre-established using evolutionary algorithms (ICGM-CNRS). Catalytic performance measurements will be performed on automated robots and the data will be implemented in the DFT simulation of the reaction mechanism, in relation to the characteristics of the catalysts. To refine the catalysts and better understand the role of each component of the system, operando experiments will be carried out using a wide variety of spectroscopic and diffraction techniques (SOLEIL (DIADEM platform) and CEA) combined with modeling of the recorded spectroscopic and diffraction data (UCCS-CNRS). All data will be analyzed in tandem to provide a clear picture of the catalytic mechanisms, the steps limiting the kinetics, the deactivation processes, and the appropriate regeneration approaches (if necessary).

Beyond the development of catalysts for NH3 synthesis and decomposition, this predictive approach will minimize experimental cycles by guiding formulations through the development of predictive algorithms (ICGM-CNRS). Ultimately, the data generated by this study will enable the development of scalable models with improved predictive capabilities, with the aim of facilitating further research in the field of energy.