R.M. Boom

Microstructures in phase-inversion membranes. Part I. Formation of macrovoids

A new mechanism for the formation of macrovoids in phase-inversion membranes is proposed. It is based on the observed difference in type of demixing of a thin film of a polymer solution when immersed in a nonsolvent bath: delayed or instantaneous demixing. The explanation for macrovoid formation assumes local conditions of delayed demixing in front of a certain layer of nuclei already formed, due to a change in the interfacial compositions at the advancing coagulation front, as compared to the original composition at the interface film-bath. Effects of variations in membrane formation conditions which can be calculated using the model of diffusive mass transport in thin films of polymeric solutions in combination with phase separation in phase-inversion membranes, completely support the mechanism as proposed.

Microstructures in phase inversion membranes. Part II. The role of a polymeric additive

Membranes were prepared from a casting solution of a water-soluble polymer, poly (vinyl pyrrolidone) (PVP), and a membrane forming polymer, poly (ether sulfone), in 1-methyl-2-pyrrolidone (NMP) as solvent by immersing them in mixtures of water and NMP. It was found that the addition of PVP to the ternary system suppresses the formation of macrovoids in the sub-layer, while the ultrafiltration-type top-layer consists of a closely packed layer of nodules. Using a model for pass transfer in this quaternary system, it is possible to explain the effects of the additive on macrovoid formation. Strong indications are found that the appearance of a nodular structure in the top-layer follows a mechanism of spinodal decomposition during the very early stages of the immersion step.

Mass transfer and thermodynamics during immersion precipitation for a two-polymer system: Evaluation with the system PES—PVP—NMP—water

An extended version of the mass transfer model by Reuvers et al. for a four-component system is evaluated, which is shown to be generally valid for short times. The thermodynamics under these circumstances are evaluated, together with the kinetics. Initial composition paths (concentration profiles) are calculated. It appears that delay of demixing is not possible when a polymeric additive is used, which is soluble in the nonsolvent, while the velocity of demixing decreases. The calculations are evaluated for the system poly(ether sulfone)-poly(vinylpyrrolidone)-N-methylpyrrolidone-water by means of light transmission measurements during immersion precipitation, for a wide range of compositions of the polymer solution and coagulation bath.

Preparation of asymmetric gas separation membranes with high selectivity by a dual-bath coagulation method

A new method for the preparation of gas separation membranes in a one-step procedure is presented, where common, non-volatile solvents can be used in the polymer solution. It concerns contacting of a polymer solution with two successive nonsolvent baths, whereby the first bath initiates the formation of a dense top layer and the second bath gives the actual polymer precipitation. Membranes made by this method will have high gas selectivity and do not need any additional coating. The new technique was used to make polyethersulfone (PES) hollow fibres from solutions consisting of 35% (w/w) polymer and 10% glycerol in N-methylpyrrolidone (NMP). High selectivities were obtained when using glycerol or 1-pentanol as the first nonsolvent and water as the second one. For a feed gas of 25 vol.% of CO2 in methane the intrinsic selectivity of PES [alpha (CO2/CH4) = 50] was easily obtained, without the necessity of an additinal coating step. By a step-wise, liquid exchange removal of residual fluids in the fibres, an improvement in flux could be obtained. This was accompanied by a somewhat lower selectivity compared to that of directly air-dried fibres.

Linearized cloudpoint curve correlation for ternary systems consisting of one polymer, one solvent and one non-solvent

A linear correlation function is found for cloudpoint composition curves of ternary systems consisting of one polymer, one solvent and one non-solvent. The conditions for validity of this correlation function appear to be that the polymer is strongly incompatible with the non-solvent, and that only liquid-liquid demixing occurs. The linearized cloudpoint (LCP) curve is interpreted in terms of the various parameters occurring in the Flory-Huggins theory. The slope of the LCP line appears to be only dependent on the molar volumes of the components. Information about the binary Flory-Huggins interaction parameters and their concentration dependence can be obtained from the intercept of the linearized curve. Cloudpoints induced by crystallization do not follow the correlation. This gives an opportunity to distinguish between crystallization and liquid-liquid demixing without any additional experiments.