In the context of climate change, we need to reduce our CO2 emissions by using less energy. Another approach is to capture, store and use CO2, with the aim of moving towards a circular carbon economy and, ultimately, defossilization. With this in mind, the direct hydrogenation of CO2 enables it to be transformed into molecules of interest such as hydrocarbons, via the coupling of the reverse water gas shift (RWGS) reaction and Fischer-Tropsch synthesis (FTS).

Computational operando catalysis has recently emerged as a reasoned alternative to the development of new catalysts, thanks to a multi-scale approach from the atom down to the active particle, to model catalyst selectivity and activity. New tools combining ab initio simulations (DFT) and molecular dynamics (MD) via machine learning algorithms bridge the gap between the precision of DFT calculations and the power of atomistic simulations. Current bifunctional catalysts (active for RWGS, and FTS) for direct CO2 hydrogenation are based on doped iron oxides (metal promoters).

The aim of this project is the theoretical study of Na-FeOx and K-FeOx catalysts doped with Cu, Mn, Zn and Co, in 4 stages: DFT simulations (adsorption energies, density of states, energy barriers, transition states), microkinetic modeling (reaction constants, TOF), construction of interatomic potentials by DFT/machine learning coupling, simulation of whole particles (selectivity, activity, microscopic quantities).

This theoretical study will go hand in hand with the synthesis and experimental measurements of the studied catalysts, and optimized catalysts emerging from the computational results. All the accumulated data (DFT, MD, catalytic properties) will be fed into a database, which can eventually be exploited to identify descriptors of interest for CO2 hydrogenation.