In a closed-cycle piston engine, particularly a Stirling-type, the suggestion is that there could be an inherent thermodynamic advantage in having sequential reactions which are exothermic on formation and then endothermic milliseconds later, on the expansion stroke.
A resonance could then be engineered, especially if the decay was sharp and reliable and the engine ran at one speed only. However, this may not be what happens in practice with argon and hydrogen. If the lifetime of argonium happened with endotherm precisely at BDC, then that could present a bonus cooling effect in addition to the change in displacement. This would arguably increase the Carnot spread between the hot end and cold end of the Stirling. I have not been able to find evidence for this type of thermodynamic cycle in the literature. Jürg Wyttenbach wrote: ArH3+ is long time stable and Ar H3+ is the driving factor in Mills original SUNCELL reaction. In fact H3+ is the most abundant form of Hydrogen in deep space.