The catalytic hydrogenation of CO₂ to methanol is a sustainable route to valorise waste carbon;(1) yet the development of homogeneous bifunctional catalysts (BFCs) remains constrained by low conversion rates, poor catalyst stability, and the complexity of the underlying catalytic mechanism.(2) We present a data driven workflow that integrates high-throughput chemical screening, quantum-mechanical (QM) calculations and machine learning (ML) predictions to accelerate the rational discovery of promising BFCs. Using the Ru-MACHO scaffold as a reference for this reaction, we systematically explore metal complex variations across four design axes: 11 metal centers, 2,597 pincer ligands, 76 trans-ligands, and 3 axial ligands, generating over 1.5·10^6 candidate structures. QM-derived data and ML models predict catalytic hydrogenation thermodynamics, enabling rapid identification of top-performing candidates for further mechanistic validation in targeted CO2 hydrogenation. This integrated approach delivers prioritized systems for experimental follow up, demonstrating how curated datasets, targeted quantum calculations and ML together streamline discovery of next generation homogeneous catalysts for efficient CO₂ to methanol conversion.
Dr. Lucía Morán