Development of Highly Active Catalysts for Low-Temperature CO₂ Hydrogenation to Methanol Using a Machine Learning...

Area of Science:

• Catalysis
• Materials Science
• Chemical Engineering

Background:

• Current industrial CO2 hydrogenation to methanol uses copper catalysts with low efficiency, requiring high temperatures and pressures.
• Developing effective catalysts is crucial for CO2 utilization and sustainable methanol production.

Purpose of the Study:

• To employ a machine learning (ML) approach to discover novel, low-temperature catalysts for CO2 hydrogenation to methanol.
• To screen a large number of potential catalysts and identify superior performers compared to existing benchmarks.

Main Methods:

• Iterative ML predictions combined with experimental validation in batch reactors.
• Screening of 580 distinct catalyst compositions.
• In situ/operando spectroscopic analysis to understand catalyst mechanisms.

Main Results:

• Identified 33 catalysts outperforming the benchmark Pt(3)/Mo(20)/TiO2.
• The optimal catalyst, Pt(5)/Mo(8)-Re(1)-W(0.7)/TiO2, achieved high methanol production rates (1.46 mmol g-1 h-1 in batch, 1.8 mmol g-1 h-1 in flow) at 150 °C and 4 MPa.
• Spectroscopic analysis revealed the roles of Pt (H2 dissociation), Mo oxides (oxygenated species), W (methanol desorption), and Re (formate conversion).

Conclusions:

• ML is a powerful tool for accelerating catalyst discovery in CO2 hydrogenation.
• The developed multi-component catalyst demonstrates significantly enhanced performance for low-temperature methanol synthesis.
• Understanding the synergistic effects of catalyst components provides insights for future catalyst design.