Short communicationGas separation membranes from polymers of intrinsic microporosity
Introduction
Polymer membranes are used for a number of important gas separations, including O2/N2 (e.g., nitrogen generation) and CO2/CH4 (e.g., natural gas enrichment) [1]. The performance of a gas separation membrane is commonly characterised in terms of the permeability of a particular species (PA) and the selectivity towards one component of a gas pair (αA/B = PA/PB). Studies over the past two decades of a great number of membrane-forming polymers have revealed a trade-off relationship between selectivity and permeability, highly selective polymers generally exhibiting low permeability and vice versa [2], [3], [4], [5], [6]. The limits of performance in 1991 were demonstrated by Robeson [4], who drew upper bounds in double logarithmic plots of selectivity versus permeability for various gas pairs. He noted that polymers which perform well with one gas pair usually perform well with other gas pairs and that the slope of the upper bound line was related to the difference in the molecular diameters of the gases involved (although, of course, the behaviour of individual polymers may go against this trend). In recent years, there has been further development of polymers that are either highly selective or highly permeable, but it has proved difficult to achieve polymers which combine high permeability (e.g., oxygen permeability, Barrer) with a selectivity that breaks Robeson's upper bound (e.g., oxygen/nitrogen selectivity ). High selectivity is needed to achieve a high purity product, whilst high permeability is desired to minimize membrane area and thus capital cost.
The permeation of gases through polymer membranes is often understood in terms of a solution-diffusion model, for which the permeability coefficient is the product of a solubility coefficient (SA) and a diffusion coefficient (DA) (Eq. (1)).Selectivity may then arise either because of differences in the solubility coefficient (solubility selectivity, SA/SB) or because of differences in the diffusion coefficient (mobility selectivity, DA/DB) (Eq. (2)).Materials that perform close to the upper bound in the separation of simple gases are generally glassy polymers. These typically show mobility selectivity, with smaller gas molecules diffusing more rapidly than larger ones. In contrast, rubbery polymers are more often used to separate large organic molecules from smaller gas molecules on the basis of high solubility selectivity.
Freeman [7] has developed a theory to explain Robeson's empirical upper bound behaviour for gas separation by glassy polymers. His theory is based on four assumptions: (i) that the solution-diffusion model (Eq. (1)) is applicable, (ii) that the diffusion of small molecules through the polymer is an activated process for which the Arrhenius equation is obeyed, (iii) that there is a correlation between the pre-exponential term of the Arrhenius equation and the activation energy (a “linear free energy” relation) and (iv) that the activation energy is simply related to the kinetic diameter of the gas molecule. He indicates two ways in which polymer membranes might in principle be achieved that would perform beyond the upper bound in gas separations: (i) by improving the solubility selectivity and (ii) by increasing the stiffness of the polymer chain whilst simultaneously increasing interchain spacing. However, interchain separation should not be increased so much that mobility selectivity is lost, as may be the case for the ultrapermeable polymer poly(1-trimethylsilyl-1-propyne) (PTMSP), which has very low selectivities for permanent gases. Alentiev and Yampolskii [8] have sought to explain the trade-off between selectivity and gas permeability of glassy polymers on the basis of a free volume model. Although their theoretical approach differs from that of Freeman [7], they draw similar conclusions. In particular, they emphasize the potential for improving solubility selectivity and point out that this might be achieved either enthalpically (e.g. by enhancing specific interactions between polymer and penetrant) or entropically.
The present contribution concerns a new type of polymer—termed a polymer of intrinsic microporosity (PIM)—for which all rotational freedom is removed from the polymer backbone, but which has a randomly contorted molecular structure [9], [10], [11], [12], [13]. Gas permeation data are presented for membranes formed from two PIMs, denoted PIM-1 and PIM-7. The structures of these polymers and their preparation are indicated in Fig. 1. For these polymers there are no single bonds in the backbone about which rotation can occur. However, at intervals along the backbone there is a spiro-centre (i.e., a single tetrahedral C atom shared by two rings) which introduces a sharp bend into the chain. The result is an inflexible but contorted polymer molecule, as indicated for PIM-1 in Fig. 2. During film formation or precipitation from solution, sufficient free volume is trapped in the solid state that the polymers behave essentially like microporous materials (containing pores of effective size < 2 nm), as demonstrated by low-temperature N2 adsorption (discussed further below). Apparent surface areas from Brunauer–Emmett–Teller (BET) analysis are in the range 700–900 m2 g−1. These polymers are amorphous and remain glassy up to their decomposition temperatures (>350 °C).
Section snippets
Experimental
PIM-1 was prepared from 5,5′,6,6′-tetrahydroxy-3,3,3′,3′-tetramethyl-1,1′-spirobisindane (Monomer 1 in Fig. 1) and tetrafluoroterephthalonitrile (Monomer 2 in Fig. 1) as described previously [10]. PIM-7 was prepared by reacting monomer 1 with 7,7′,8,8′-tetrachloro-phenazyl-3,3,3′,3′-tetramethyl-1,1′-spirobisindane (Monomer 5 in Fig. 1). Full details of the synthetic procedure for PIM-7 will be published elsewhere.
Polymer membranes were prepared by casting from solution (2–5 wt.%) into a
Results and discussion
Pure gas permeation results are given in Table 1 for a PIM-1 membrane (46 μm thickness) and a PIM-7 membrane (28 μm thickness). For both PIMs, the order of permeability is CO2 > H2 > He > O2 > Ar > CH4 > N2 > Xe, unlike most glassy polymers, for which typically He > CO2. Only a few polymers, notably poly[1-(trimethylsilyl)-1-propyne] (PTMSP) [14], poly(4-methyl-2-pentyne) [15] and Teflon AF [16] exhibit similar or higher permeabilities. Robeson's 1991 upper bounds [4] for O2/N2 and CO2/CH4 are shown in Fig. 3a
Conclusions
Membranes formed from polymers PIM-1 and PIM-7 combine high gas permeability with high selectivity. Their performance is linked to microporous character that arises from a rigid, contorted molecular structure. PTMSP, which is more permeable but less selective than these PIMs, also exhibits microporous character, with larger apparent pore sizes. These results are consistent with Freeman's suggestion [7] that gas separation performance can be improved by increasing the stiffness of a polymer
Acknowledgements
We are grateful to EPSRC for funding. We thank K.-V. Peinemann (GKSS), K. Ohlrogge (GKSS), S. Shishatskiy (GKSS) and Y. Yampolskii (Institute of Petrochemical Synthesis, Moscow) for helpful discussion.
References (50)
- et al.
Relationship between gas separation properties and chemical structure in a series of aromatic polyimides
J. Membr. Sci.
(1988) - et al.
Effects of polymer structure on the gas permeability of silicone membranes
J. Membr. Sci.
(1988) Correlation of separation factor versus permeability for polymeric membranes
J. Membr. Sci.
(1991)- et al.
High performance polymers for membrane separation
Polymer
(1994) - et al.
A group contribution approach to predict permeability and permselectivity of aromatic polymers
J. Membr. Sci.
(1997) - et al.
Free volume model and tradeoff relations of gas permeability and selectivity in glassy polymers
J. Membr. Sci.
(2000) - et al.
Microporous polymeric materials
Mater. Today
(2004) - et al.
Poly[1-(trimethylsilyl)-1-propyne] and related polymers: synthesis, properties and functions
Prog. Polym. Sci.
(2001) - et al.
Synthesis and gas permeation properties of poly(4-methyl-2-pentyne)
J. Membr. Sci.
(1996) - et al.
Gas and vapor transport properties of amorphous perfluorinated copolymer membranes based on 2,2-bistrifluoromethyl-4,5-difluoro-1,3-dioxole/tetrafluoroethylene
J. Membr. Sci.
(1996)