Discovering CO Adsorption and Desorption Pathways from Chemical Reaction Neural Network Modeling of Transient...

Area of Science:

• Surface science
• Chemical kinetics
• Computational chemistry

Background:

• Interpreting complex surface reaction mechanisms from experimental data is challenging.
• Transient kinetic data offers rich mechanistic information but requires advanced analytical tools.
• Understanding CO adsorption/desorption on metal surfaces is crucial for catalysis.

Purpose of the Study:

• To develop and validate a data-driven framework for interpreting surface reaction mechanisms.
• To combine time-resolved infrared spectroscopy with chemical reaction neural networks (CRNNs).
• To elucidate the CO adsorption and desorption mechanism on Pd(111) using transient kinetic data.

Main Methods:

• Utilized time-resolved gas pulsing infrared spectroscopy to collect transient kinetic data.
• Employed chemical reaction neural networks (CRNNs) to model and analyze the experimental data.
• Systematically evaluated various reaction mechanisms, including different adsorption sites (hollow, bridge).

Main Results:

• Models with distinct adsorption site dynamics showed similar fits to experimental absorbance data.
• Analysis of spectral peak stability and predicted dynamics identified a preferred mechanism.
• The most physically consistent mechanism involves primary CO adsorption on bridge sites followed by conversion to hollow sites.

Conclusions:

• A data-driven approach using CRNNs can extract detailed surface reaction mechanisms from limited experimental data.
• Machine learning effectively bridges the gap between transient kinetic measurements and molecular-level understanding.
• This framework offers a powerful tool for mechanistic interpretation in surface science and catalysis.