Permeability, glassy polymer transport properties

Nonlinear, pressure-dependent sorption and transport of gases and vapors in glassy polymers have been observed frequently. The effect of pressure on the observable variables, solubility coefficient, permeability coefficient and diffusion timelag, is well documented (1, 2). Previous attempts to explain the pressure-dependent sorption and transport properties in glassy polymers can be classified as concentration-dependent and "dual-mode models. While the former deal mainly with vapor-polymer systems (1) the latter are unique for gas-glassy polymer systems (2). 

Although no commercial examples exist currently in the gas separation field, thin film composite membranes such as those pioneered by Cadotte and co-workers (10) may ultimately permit the use of novel materials with unique transport properties supported on standard porous membranes. Therefore, the focus in this paper will be on suggesting a basis for understanding differences in the permeability and selectivity properties of glassy polymers. Presumably, if such materials prove to be difficult to fabricate into conventional monolithic asymmetric structures, they could be produced in a composite form. Even if thin film composite structures are used, however, the chemical resistance of the material remains an important consideration. For this reason, a brief discussion of this topic will be offered. 

An overview of the physics of glassy polymers and the relationships between molecular mechanisms and macroscopic physical, mechanical and transport properties of polymer glasses is presented. The importance of local translational and/or rotational motions of molecular segments in the glass is discussed in terms of the implications for thermodynamic descriptions of the glass (configurational states and energy surfaces) as well as history dependent properties such as expansivity, refractive index, gas permeability, and viscoelastic mechanical behaviour. 

Transport Properties. Sorption and transport properties are highly dependent on the post-vitrification history of glassy polymers (77) hence one would expect parameters such as physical aging, antiplasticization and amorphous orientation to affect transport properties. The reduction in diffusivity and permeability due to aging, orientation, and antiplasticization can be modeled via entropy or fi ee volume arguments (77). In addition, diffusive jumps of penetrant molecules in glassy polymers can be affected by (facilitated by) the segmental mobility that is manifested in sub-Tg relaxations 78),... 

In contrast to the LCP results just presented, in glassy polymers used as gas separation membranes, free volume influences diffusion coefficients much more than solubility coefficients. Figure 6 provides an example of this effect. In this figure, the solubility, diffusivity, and permeability of methane in a series of glassy, aromatic, amorphous poly(isophthalamides) [PIPAs] are presented as a function of the fractional free volume in the polymer matrix. (More complete descriptions of the transport properties of this family of materials are available elsewhere (59, 40)). The fractional free volume is manipulated systematically in this family of glassy polymers by synthesizing polymers with different substituent and backbone elements as shown in... 

It is known that glassy polymer membranes can have a considerable size-sieving character, reflected mainly in the diffusive term of the transport equation. Many studies have therefore attempted to correlate the diffusion coefficient and the membrane permeability with the size of the penetrant molecules, for instance expressed in terms of the kinetic diameter, Lennard-Jones diameter or critical volume [40]. Since the transport takes place through the available free volume in the material, a correlation between the free volume fraction and transport properties should also exist. Through the years, authors have proposed different equations to correlate transport and FFV, starting with the historical model of Cohen and Turnbull for self diffusion [41], later adapted by Fujita for polymer systans [42]. Park and Paul adopted a somewhat simpler form of this equation to correlate the permeability coefficient with fractional free volume [43] ... 

Sample Dimension Dependent Effects. Recently, it has been noted that the permeability of glassy polymers decreases over time and the selectivity increases over time. Interestingly, the decreases in permeability and increases in selectivity are accelerated for thin films, on the order of a micrometer in thickness. The study of the gas transport properties in thin films is not only academic because film thicknesses of this order are often used in semiconductors, in adhesives, and in packaging. Furthermore, the skin thicknesses of asymmetric hollow fibers, the most common industrial membrane geometry, are usually 1 /u.m or less. 

In the past 25 years, relatively few attempts to increase gas separation membrane performance with dense film mixed matrices of zeolite and rubbery or glassy polymer have been reported. Table I summarizes practically all of the reported O2/N2 mixed matrix membranes. Permeabilities and permselectivities are specified as a range to encompass the various zeolite volume fractions studied. In general, an increase in permeability is observed with zeolite addition coupled with a slight increase in permselectivity. Despite the wide variety of combinations of zeolites with rubbery and glassy polymers, reported mixed matrix membranes fail to exhibit the desired O2/N2 performance increases. These failures have generally been attributed to defects between the matrix and molecular sieve domains. While this is certainly a possible practical source of failure, our work earlier 8) has addressed a more fundamental source caused by inattention to matching the transport properties of the molecular sieve and polymer matrix domains. This topic is discussed briefly prior to consideration of the practical defect issue noted above. 

---