Analysis of the Pyrolysis of Methane Reaction over Molten Metals for CO2-Free Hydrogen Production: An Application of DFT and Machine Learning

The co-production of CO2 continues to remain the bane of several hydrogen production technologies, including the steam reforming of methane and the dry reforming of methane processes. Efficient utilization of abundant greenhouse gas in the form of methane provides opportunities for the design of an innovative system that will maximize the use of such a raw material in the most environmentally friendly manner. The study of the mechanism of the pyrolysis of methane reactions over molten metals provides promise for improved hydrogen yield and methane conversion with a greater turnover frequency. Catalyst electronic properties computed via Density Functional Theory using the Quantum Espresso code provided data that were built into a database. Using Bismuth as the base metal, active transition metals including Ni, Cu, Pd, Pt, Ag, and Au of different concentrations of 5, 10, 15, and 25% were placed on 96 atoms of the base metal and relaxed to obtain the optimized geometric structures for the catalytic reaction studies. The kinetics of the individual elementary steps of the pyrolysis reaction at preset temperatures over the bi-metals were calculated using the Car-Parinello (CP) method and Nudge Elastic Band (NEB) computations. The collated data of the various pyrolysis of methane reactions over the different bi-metals was used to train machine learning models for the prediction of reaction outcome, catalytic performance, and efficient operating conditions for the pyrolysis of methane over molten metals. The turnover frequency, which is determined using the transition state energies of the fundamental reaction cycles, will be used to simulate the stability of the catalyst.