Insights into Reaction Mechanisms in Liquid Metals from Density Functional Theory: CH4 Pyrolysis in BiNiX (X = Cu, Al) Molten Metals as a Case Study

Pyrolysis of methane using molten metal catalysts in bubble column reactors presents a cleaner and coking-free alternative for hydrogen production from natural gas due to its lack of direct CO2 emissions. Bubble column reactors overcome coking as the byproduct carbon material floats to the top due to its lower density with respect to the molten metal and can be collected by physical means. Typically, the catalysts used in these systems consist of a low-melting-point solvent metal such as Sn or Bi, along with a higher melting point active component such as Ni. In this study, we present a first-principles investigation that explores the impact of a third component, a promoter, on the reactivity of a Bi-Ni molten alloy. Specifically, we examine Al and Cu as promoters and compare their effects. Unlike previous literature that mostly relies on static nudged elastic bands and free-energy-based calculations, we develop a comprehensive ab initio molecular dynamics-based protocol to provide a detailed account of the reaction mechanism. Our direct rate calculations reveal that, overall, including either promoter at 10 or 20% leads to a better performance than using an unpromoted Ni-Bi binary alloy. To overcome the intrinsic difficulties associated with rate calculations, we have performed a large number of ab initio molecular dynamics calculations and have analyzed the H dissociation times for each reaction step involved in the reaction mechanisms. Our findings reveal that overall Cu outperforms Al as a promoter under the simulation conditions employed. To better understand this outcome, we carefully examined each dissociation event and identified several intriguing trends, including the contrasting contributions of Al and Cu to reactivity. We have further rationalized our results with the help of simple descriptors such as binding energy and partial Bader charges in binary metal clusters and have found that the strength of the interaction of different metal species in the alloy with one another and the product fragments in the immediate neighborhood of the reaction as well as relative partial charges correlate very well with dissociation time data. The analysis methods developed in this study will be used in future research, enabling the screening of a broader range of promoters and facilitating the design of these promising energy production systems.