Unlocking efficient CO2–to–methanol conversion on frustrated Lewis pair-functionalized UiO–67: A synergistic approach using DFT and SISSO

Abstract

Metal–organic framework-based catalysts demonstrate considerable promise for converting CO2 into valuable chemicals, particularly when combined with Frustrated Lewis Pairs (FLPs) to enhance H2 dissociation during hydrogenation reactions. This study employs density functional theory (DFT) calculations to investigate modified UiO–67 frameworks wherein FLPs are introduced via eight different functional groups (UiO–67–X) into the organic linker to facilitate H2 activation during CO2 hydrogenation to methanol (CH3OH). The reaction proceeds through three stages: (i) hydrogenation of CO2 to formic acid (HCOOH), (ii) conversion of HCOOH to formaldehyde (HCHO), and (iii) hydrogenation of HCHO to CH3OH. This study specifically focuses on steps (ii) and (iii), analyzing the detailed reaction mechanisms using optimized molecular structures and Gibbs free energy calculations to acquire insights into methanol formation on UiO–67–X. During HCOOH conversion to HCHO, adsorbed H2 undergoes heterolytic cleavage at the FLP sites, producing a proton (H+) and a hydride (H−) for subsequent HCOOH hydrogenation and dehydration. The energy barriers identified at this stage represent key kinetic limitations hindering efficient CO2-to-methanol conversion. Similarly, HCHO conversion to CH3OH proceeds via H2 dissociation, followed by concerted H+/H− transfer. Among the tested UiO–67–X catalysts, UiO–67–B(CH3)2 exhibits the highest catalytic activity for CO2 hydrogenation to methanol. Kinetic analyses are performed to assess reaction rates across a relevant temperature range, highlighting the notable influence of functional groups on catalytic performance. Additionally, the Sure Independence Screening and Sparsifying Operator (SISSO) machine-learning approach is used to identify optimal physical descriptors and derive a predictive model for the energetic span (δG), considerably lowering the computational cost associated with full reaction pathway calculations. Statistical validation confirms the robustness of these predictions. Overall, these findings underscore the vital role of FLP-assisted H2 dissociation in promoting CO2 hydrogenation to CH3OH, with UiO–67–B(CH3)2 serving as a promising catalyst.