image: Using calculated infrared spectroscopy as input, the proposed machine learning framework, consisting of multiple blocks and a fully connected layer could accurately predict target structural and energetic information, along with critical reaction profiles. Art by Li Yang and Sheng Ye.
Credit: Art by Li Yang and Sheng Ye.
This study is led by Associate Prof. Li Yang, Prof. Sheng Ye (Anhui University), Prof. Thomas Heine (Technische Universität Dresden) and Prof. Jun Jiang (University of Science and Technology of China). In this work, infrared spectroscopy is utilized as a non-invasive method to probe molecular transformations, coupled with a machine-learned protocol to map the spectroscopic fingerprints to atomistic structures.
To intuitively illustrate this approach, they selected C–C coupling involving the interaction of two adjacent CO intermediates as a representative demonstration (see image below). For the machine learning architecture, they designed a convolutional neural network (CNN) consisting of multiple blocks and a fully connected layer. By investigating the multiple dynamic trajectories of C–C coupling on Cu surface, they demonstrated that this network can accurately predict local atomistic structures of the active surface evolution and the energetic variations by deciphering IR spectroscopy. This way, the structural rearrangements during C–C coupling could be recapitulated in detail, presenting a cost-effective strategy to monitor dynamic C–C coupling.
Furthermore, concerning the critical role of CO–CO dimerization reaction involved in C–C coupling, the authors applied the trained machine learning model specifically to the CO–CO dimerization reaction on the Cu surface. In which, they were able to derive crucial configurations and corresponding energy barriers from the IR spectroscopy data, setting a facile strategy to define the reaction profiles. Even on unfamiliar Cu-based surfaces, the reaction pathways could also be identified within the fine-tuning machine learning model. Specially, the predicted enhancement of CO–CO dimerization via metal dopant agrees well with previous experimental findings, confirming the reliability of machine learning model.
Dr. Li Yang and the collaborators integrate IR spectroscopy, first-principles calculations, and machine learning to provide a simple yet effective venue to track structural evolution in the C–C coupling process. The accessibility of spectroscopic fingerprints through both computational and experimental methods underscores the practicality and versatility of the proposed machine learning protocol for tracking complex structural evolution. Not only does it enable efficient model training using theoretical spectral data, but also holds promise for further refinement by incorporating experimental data to capture real-world environmental effects. The application of machine-learned spectroscopy offers a land of opportunity in chemical reaction research, exemplifying the recent paradigm shift towards “AI for Science”.
###
See the article:
Monitoring C–C coupling in catalytic reactions via machine-learned infrared spectroscopy https://doi.org/10.1093/nsr/nwae389
Journal
National Science Review