The Joule–Thomson effect is discussed for a fluid composed of spherically symmetric Lennard–Jones(12,6) molecules (of “diameter” σ) confined between two planar, rigid, structureless solid substrates separated by sz=10 and 20σ. The effect of “strong” and “weak” of the substrate is studied by employing fluid-substrate potentials with and without attractive interactions, respectively. The focal point of this study is the confinement-induced depression of the inversion temperature Tinv with respect to the bulk value. It is defined such that during a Joule–Thomson expansion the temperature of a (confined or bulk) gas remains constant. In the limit of vanishing gas density, Tinv is computed from the second virial coefficient defined through a density expansion of the transverse stress T∥ in the gas. For higher densities Tinv is computed from the (transverse) expansion coefficient α∥ which is accessible through density and enthalpy fluctuations in mixed stress–strain ensemble Monte Carlo simulations. Results of these simulations are analyzed in terms of a mean-field theory which provides a qualitatively correct description of the Joule–Thomson effect in confined fluids. The smaller sz the more depressed (with respect to the bulk) is Tinv. The density dependence of Tinv is different for “strong” and “weak” substrates. Without attractive fluid–fluid interactions Tinv does not exist and the confined gas is always heated during a Joule–Thomson expansion. In this case α∥ is independent of the substrate material.
