This work describes vacuum pervaporation tests which were performed to identify the influence of five hydrocarbons typically presented in gasoline, i.e. n-hexane, n-heptane, n-octane, isooctane (2,2,4-trimethylpentane), and 1-octene on thiophene removal efficiency from hydrocarbon/thiophene binary mixtures. Commercial hydrophobic PEBA-based and PDMS-based composite membranes were applied. The impact of process parameters such as feed temperature and feed composition was examined. Experiments performed with various hydrocarbon confirmed, that the membrane performance and overall separation efficiency is influenced by the physiochemical properties of the mixture components (solute and solvent) and the differences in the affinities towards a given membrane. Permeate fluxes and corresponding separation factors presented typical trade-off regardless the type of membrane and solvent. The separation factors were higher in the case of the PEBA membrane which corresponded with the Hansen's Solubility Parameters Theory. The batch pervaporation removal of thiophene from isooctane was also modeled, proving that the PEBA membrane is more effective than the PDMS one. Solvents fluxes were 3 orders of magnitude higher than solute ones, therefore Pervaporation Separation Index (PSI) should be used only as a preliminary assessment of the membrane performance. (C) 2017 Elsevier B.V. All rights reserved.