In this study, 11 different hydrophobic materials made of high permeability glassy polymers such as PTMSP, PBTMST, PTMST, PTMGP, PMP, PVTMS, as well as polymeric blends based on PTMSP/PVTMS with varied fractional free volume, were studied by the extended method of hydrostatic weighing. Results clearly indicate the presence of interconnected pre-existing free volume elements (microcavities or micropores), which are accessible for the liquid molecules without polymer swelling. Depending on the polymer, the contribution of pre-existing microcavities to the free volume of polymer varies from 35% for PIM-1 to 85% for PTMSP/PVTMS (90110). Using the proposed method, it was possible to estimate total fractional accessible volume FAV(t): PTMSP (30%) > PTMSP/PVTMS (90110) (27%) > PTMSP/PVTMS (80/20) (25%) > PBTMST (24%) > PTMST (23%) > PTMGP (22%) > PTMSP/PVTMS (70/30) (21%) > PIM-1 (17%) > PMP (16%) > PTMSP/PVTMS (40/60) (11%) > PVTMS (4%). The applicability of the extended method of hydrostatic weighing for evaluation of the porous structure of the polymeric materials was confirmed by good agreement with the literature data on fractional free volume FFV estimated by PALS. FAV can be considered as a uniform parameter to describe the solvent transport regardless of the difference in the nature of high permeability glassy polymers. It was found that there is a threshold value of FAV(t), estimated as 12%, which is required for establishment of liquid permeability. FAV(t) values obtained for PVTMS (4%) and PTMSP/PVTMS (40/60) (11%) were not high enough to provide the formation of liquid percolation clusters, and, hence, liquid transport across the membrane at 20 bar. Investigation of water ethanol transport through PTMSP, PTMGP, PMP and PIM-1 showed that all polymers had two regions: (i) at lower concentration of ethanol, hydrophobic glassy polymers showed absence or hardly detectable liquid transport, (ii) the increase of ethanol concentration led to the establishment and further increase of liquid transport through the dense membranes. PTMSP, PTMGP, PMP and PIM-1 kept their barrier properties till the threshold value of FAV was equal to 26%, 17%, 15% and 12%, respectively. Such behavior was explained in terms of boundary conditions for formation of a percolation cluster within a specific glassy polymer with respect to its properties and chemical nature. Once such clusters were formed in the bulk material, no further noticeable increase in FAV(t) was required to enhance the liquid transport through the membrane. The different behavior of PIM-1 was attributed to the presence of a noticeable fraction of isolated holes. (C) 2014 Elsevier Ltd. All rights reserved.