CO2 separation with thermally rearranged polymer membranes: separation properties, aging and restoring

Thermally rearranged polymer membranes offer great advantages in gas separation applications owing to the presence of micropores, appropriately tuned cavities size and distribution, which make them highly permeable [1-5]. Their use in CO2 separation from flue gas type streams is an important field of application currently under development. In this work, the transport properties of a thermally rearranged hollow fibre membrane modules[2], also functionalized with carbon nanotubes [3], were evaluated feeding a gas mixture (CO2:N2:O2=15:80:5) having a typical flue gas composition, also in presence of water vapour. CO2 permeance measured with the dry mixture is quite similar to that measured feeding single gas owing to the high CO2 condensability in polymer matrix. On the contrary, a permeance reduction of N2 and O2 feeding the mixed -gas over their pure gas one was observed with a consequent mixed-gas CO2/N2 selectivity of 20 higher than 16 measured for single gas. The presence of water vapour in the mixture showed great influences on the permeation behaviour; a drop in permeance of ca. 48% was measured for CO2 and ca. 30 and ca. 40%, for N2 and O2, respectively. This was translated in a 20% reduction in CO2/N2 selectivity from 20 (dry condition) to 16 (wet condition) at 25°C. The membrane module, tested continuously for about six months, showed a certain drop in permances that can be attributed to aging effects induced by physical tendency to equilibrium state. To address this issue, a strategy for restoring initial mass transport properties was proposed. The permeance of CO2 and N2 were tracked for more than 500 days, divided in two cycles, under continuous exposure to the fed gas. After each cycle of observation lasted 190 and 285 days, the membrane module was restored by means of an appropriate procedure, foreseeing the injection of methanol. The separation properties, permeance and selectivity, of the membranes were fully recovered since both reached their earliest values (day 1). After the first restoring, the membrane monitoring, for further 285 days feeding single and mixed dry gases, again showed aging, similarly to that in the first cycle of observation. The second restoring carried out on the again aged module confirmed the repeatability of this procedure for recovering separation properties, option particularly attractive from an applicative point of view. Globally, after two restoring, the membrane module was still performant after 520 days.