Gas separation by dense polymer membranes is a very promising alternative to the cryogenic distillation or adsorption separation processes due to its much lower energy costs. This is critical for chemical industries, where the separation of mixtures accounts for over 50% of the energy costs. More energy-efficient methods should improve economical viability and lower greenhouse gas emissions. In addition, membrane devices are compact and fairly easy to operate. Unfortunately, polymers do not perform well under harsh conditions since they tend to lose their structural integrity at high temperatures and pressures. The University of Twente, The Netherlands, has recently developed new hybrid ultrathin membranes based on inorganic POSS cross-linked with organic imides in order to improve the thermomechanical resistance while maintaining the gas-sieving properties. The synthesis is suited to large-scale production and these hybrid polyPOSS-imide networks are indeed able to perform under tougher conditions than conventional polymers. However, the aliphatic arms of the POSS precursors used so far are too flexible and prone to thermal degradation, which prevents their use above ~300°C. In addition, the gas sieving abilities are strongly dependent on the precursors, the cross-linking densities, the temperatures and the pressures. Furthermore, experiments under harsh conditions are difficult to carry out and it has not been possible to characterize them over a large range of high temperatures and pressures. The aim of MOLHYB is to exploit the possibilities opened up by this innovative class of materials in order to develop new hybrid membranes capable of performing at very high temperatures and pressures, based on a combined molecular modelling and experimental approach. During an initial collaboration with Twente, the LEPMI at the University Savoie Mont Blanc, France (LEPMI-USMB), has developed realistic molecular models of two polyPOSS-imides at one cross-linking density. We intend to design more robust materials within MOLHYB. Molecular dynamics simulations will be used at LEPMI-USMB to pre-screen a novel set of candidate polyPOSS-imides for improved thermomechanical resistance, i.e. up to at least 400°C, without compromising their gas separation function. Their physical and mechanical properties will be characterized at the molecular-level as a function of the precursors, the cross-linking densities and the temperature. Only the most promising structures will then be synthesized and characterized experimentally by Twente. In parallel, single-gas sorption and transport in the selected model polyPOSS-imides will be studied at LEPMI-USMB for penetrants with different plasticizing capabilities, i.e. N2, CH4, CO2 and H2S, under a full set of both normal and harsh conditions. The latter are industrially-relevant conditions that are difficult to attain safely in the laboratory. Twente will carry out single-gas permeation experiments under a limited set of conditions to validate the model results. This will provide ideal selectivities for CO2/CH4, N2/CH4, CO2/N2, H2S/CH4 and CO2/H2S separations under a large range of conditions. To assess the influence of mixed-gas reservoirs, LEPMI-USMB will also consider CO2+CH4+H2S mixtures. The novel polyPOSS-imide membranes showing both improved thermoresistance and optimized gas separation properties will then be assembled at Twente atop inorganic porous hollow fibres in order to obtain supported materials that can be used for upscaling. Based on this combined approach, MOLHYB should lead to better materials for selective separations under harsh conditions, i.e. with mixed-gas reservoirs at high temperatures and pressures. Since The Netherlands are not part of the countries selected for PRCI, Twente will entirely provide its own funding. The ANR demand only concerns the French partner LEPMI-USMB and is mainly aimed at funding a Ph. D. student.