Numerous studies have been already reported for the rigorous design of dense membrane separation units. These are classically based on the solution-diffusion framework, making use of the permeability concept, in order to compute transmembrane fluxes. While variable permeability conditions are expected to hold for a great number of situations (gas permeation through glassy polymers, vapor and liquid permeation), the incidence of this particular situation has been seldom addressed. It is tempting in that case to keep the solution-diffusion framework, based on a combination of permeability and driving force terms. It is shown in this work that such a strategy can potentially induce major inconsistencies in terms of fluxes computations, which may remain unidentified if complimentary tests are not performed. A new flux expression, consistent with the solution-diffusion approach, is proposed in order to circumvent this limitation and its application to a case study (acetone/toluene separation through a silicone membrane) is reported.
