Standard Article

You have free access to this content

Microporous Polymers: Synthesis, Characterization, and Applications

  1. Jens Weber1,2,
  2. Qing Bo Meng1

Published Online: 15 APR 2014

DOI: 10.1002/0471440264.pst622

Encyclopedia of Polymer Science and Technology

Encyclopedia of Polymer Science and Technology

How to Cite

Weber, J. and Meng, Q. B. 2014. Microporous Polymers: Synthesis, Characterization, and Applications. Encyclopedia of Polymer Science and Technology. 1–49.

Author Information

  1. 1

    Max Planck Institute of Colloids and Interfaces, Department of Colloid Chemistry, Potsdam, Germany

  2. 2

    Currently on the leave to:Hochschule Zittau/Görlitz (University of Applied Science), Department of Chemistry, Zittau, Germany

Publication History

  1. Published Online: 15 APR 2014

Introduction

  1. Top of page
  2. Introduction
  3. Synthetic Methodologies
  4. Characterization of Microporous Polymers
  5. Applications
  6. Summary
  7. Bibliography

Nanoporous polymers, that is polymer materials having pores of sizes below 100 nm received increasing interest within the past few years. They offer new prospects in technologies such as gas separation/storage, separation science, and many more. Usually, nanoporous polymers are further discriminated by following IUPAC recommendations (1) by their pore size in the dry state into microporous polymers (pore sizes < 2 nm) and mesoporous polymers (2 < pore size < 50 nm). Materials having pore sizes larger than 50 nm are usually claimed to be macroporous. Purely meso- or macroporous polymers will not be discussed within this contribution in detail. Further information on these materials is, however, available from a variety of reviews (2-8). Please note that the term microporous polymer is used in accordance with the IUPAC recommendation on porous materials within this contribution. This usage differs from the membrane community, which often describe polymers that have pore sizes on the micrometer scale (ie, macroporous polymers) as microporous polymers.

The article will give an overview on synthetic methods toward microporous polymers, followed by a short overview on characterization methods used for the structural analysis of such polymers. Methods for the extraction of important parameters such as specific surface area S, porosity ϕ, or the pore size distribution (PSD) will be discussed critically. In the following, we point out potential applications, which have emerged during the past years. The article does not discuss the literature on the so-called covalent organic frameworks (COFs), which are crystalline materials in contrast to the amorphous polymers discussed herein. The interested reader can find information on COFs in a variety of reviews (9). Discrete organic microporous compound, for example, cages (10), will also be not considered within this contribution.

Synthetic Methodologies

  1. Top of page
  2. Introduction
  3. Synthetic Methodologies
  4. Characterization of Microporous Polymers
  5. Applications
  6. Summary
  7. Bibliography

We will provide a short overview about the main synthesis methods toward microporous polymers. More detailed information can be accessed from respective reviews and/or the original contributions cited therein (7, 8, 11-20). Recent developments, which were not yet covered within these reviews, will be discussed. The section will conclude with a concise overview on the possibilities to synthesize microporous polymers in well-defined morphologies (particles, films, monolithic) rather than in the common powder form. Shaping is of high importance for potential applications that require homogenous flow patterns, tight connections without short circuits etc.

Hyper-cross-linked Microporous Polymers

Microporous polymers can be obtained by various methods. The most simple routine is based on an extensive cross-linking of (weakly cross-linked) polymer chains in a solvent swollen state. If the resulting material is strong enough to withstand pore collapse upon solvent removal, a microporous material can be obtained. This approach works essentially very well for polystyrene-based systems and derivatives and was described by Davankov and co-workers in the 1970s. Several review articles by Davankov can serve as a valid introduction into the topic of this so-called hyper-cross-linked polymers (11, 15, 21, 22). The cross-linking is often based on Friedel–Crafts chemistry, such as iron chloride catalyzed alkylation reactions. It is, however, not restricted to Friedel–Crafts approaches, and other methodologies have been used indeed. Synthesis routines based on direct condensation/polymerization reactions have been developed as an alternative to the classic post-synthetic hyper-cross-linking methodology. A detailed review on these more recent developments is available (17). A schematic picture of the chemical structures and the pore structure of such hyper-cross-linked materials is given in Figure 1. The three-dimensional (3D) picture is based on molecular modeling of the respective porous structure and does show clearly the highly irregular porosity patterns.

figure

Figure 1. Left-hand side: chemical routes toward hyper-cross-linked polymers by either post-cross-linking of swollen polymers or by self-condensation of di/multifunctional monomers using Friedel–Crafts conditions (typically: FeCl3, dichloroethane); right-hand side, upper part: modeled 3D structure of a microporous polymer network obtained by self-condensation of p- and m-dichloroxylene; right-hand side, lower part: representative two-dimensional slice through the structure. Reprinted with permission from Ref. 23. Copyright (2007) American Chemical Society.

Soluble Microporous Polymers, PIMs

An alternative pathway toward the synthesis of microporous polymers is based on the concept of frustrated packing. The so-called polymers of intrinsic microporosity (PIM) principle is based on very rigid polymer chains, which contain a site of contortion (introducing an angle of 80–110°) within the repeat unit (24-26). The polymer chains cannot pack space efficiently as a consequence of the rigid backbone structures (typically ladder-type polymers or aromatic polyimides etc.) and the contortion site, which results in the formation of very large free volume. If this free volume is accessible from the outside, it can be regarded as some form of microporosity. Being solely based on the molecular structure of the polymer backbone, the porosity is claimed to be intrinsic. Figure 2 shows a 3D representation of PIM-1. PIM-1 is a benzodioxane-based polymer, which is formed from the reaction of 3,3,3′,3′-tetramethyl-1,1′-spirobiindane-5,5′,6,6′-tetraol and tetrafluoroterephthalonitrile in the presence of K2CO3 and can be considered as the most prominent soluble microporous polymer.

figure

Figure 2. Chemical structure and molecular model of PIM-1 (fragment) showing the contorted and open structure. Reproduced from Ref. 27 with permission of The Royal Society of Chemistry.

The concept of microporous but still soluble polymers is especially interesting from an application point of view. Soluble polymers can be easily processed into dense membranes, which have significant impact in modern technology and improvements of processes such as gas separation etc. might be anticipitated (see the section Applications) (28, 29). Hence, there was strong interest to synthesize other PIM-like polymers, which are also highly porous and soluble. Interestingly, almost 10 years after the first report on PIM-1, there are still only a rather limited number of soluble microporous polymers known. Systems based on benzodioxane formation reactions reported until 2010 are summarized in an excellent review by Budd and McKeown (16). Some new developments using the benzodioxane chemistry include the synthesis of a spirobifluorene-based PIM for improved gas permeation properties (30), as well as the use of spirobischromane monomers for PIM synthesis (31).

Besides the benzodioxane systems, there have been some reports on various microporous aromatic polyimides (32-38). A first report on a microporous spirobifluorene-based system appeared in 2007 (32), though microporous soluble polyimides have most probably been synthesized—but not classified as microporous—in former times within the area of gas separation membranes. The presence of a site of contortion such as spirocenters, binaphthalenes, or triptycenes, together with a rather rigid structure was however shown to be necessary to achieve a high degree of porosity (34). Microporous polyimides can also be synthesized using the benzodioxane formation reaction besides the common reaction between aromatic diamines, and dianhydrides (39). Alternatively, nonporous aromatic polyimides, which feature an ortho-positioned hydroxyl functionality can be transformed into microporous polymer using thermal rearrangement reactions (40, 41). In this case, it is possible to process the prepolymer into a desired form (membrane, fibers) before rearrangement into the microporous morphology. The process is somewhat similar to the well-known carbon molecular sieves, which are often also based on polyimide precursors but do not show comparably good mechanical stability (42).

Next to the formation of benzodioxane-based PIMs or microporous polyimides, there are only a few developments toward soluble microporous polymers, though some of them provide great potential for future applications. The formation of microporous polyester membranes is possible using rigid precursors with sites of contortion (43). The porosity of such systems is however rather low, probably as a consequence of the still rather high flexibility of the system. Indeed, molecular modeling could show that very high rigidity is beneficial for maximizing the porosity (44). A significant increase in the polymer rigidity and hence the gas permeation performance (while maintaining good film-forming properties) was made using tröger's base (TB) chemistry. Soluble polymers could be synthesized by McKeown and co-workers from aromatic diamines based on ethanoanthracene or spirobisindane motifs (see Fig. 3) (45). It can be envisaged that the use of the TB chemistry will lead to a number of new porous soluble polymers within the next years.

figure

Figure 3. Chemical structures and synthesis pathway of Trögers base (TB) based PIMs. Adapted from Ref. 45.

Finally, Cooper and co-workers introduced a method toward the fabrication of soluble, conjugated microporous polymers based on pyrene (46). A two-step Suzuki polycondensation approach was used, and fluorescent, fully conjugated polymers were formed. No freestanding films could be obtained so far; however, the coating of surfaces was easily possible. Given the large variety of application of conjugated (porous) polymers (see the section Applications), this work does also open up a lot of new possibilities as it allows an easy processing of the materials.

As stated previously, the number of known microporous-soluble polymers is still rather low, especially when compared to the large amount of known microporous polymer networks (see below). Polymer modification has been identified as an alternative way to tune and enhance the properties of microporous soluble polymers. Various modifications, which are most often based on the workhorse PIM-1, have been reported (see Fig. 4), mostly making use of the nitrile groups of PIM-1. Hydrolysis of the nitrile groups to carboxylic acid groups was reported by Guiver and co-workers (47). The obtained membranes showed tunable gas permeability and selectivity, and it was shown that the carboxylation leads to a reduction of the pore size and the overall porosity as a consequence of increased hydrogen bonding between the chain segments (48). Decarboxylation of the carboxylated PIM-1 was also shown as a subsequent possibility to enhance the gas permeation properties. The thermally induced decarboxylation leads finally to cross-linked membranes with higher resistance toward plasticization (49). Direct cross-linking of plain PIM-1 membranes by thermal treatment (∼250–300°C) using a vacuum furnace was reported as well as a way to improve the membrane performance (50). It was claimed that the nitrile groups undergo trimerization to form triazine rings, which serve as cross-linking nodes.

figure

Figure 4. Summary of the reported modification protocols of PIM-1.

Other reported modifications of the nitrile group of PIM-1 include formation of thioamides or cycloaddition reactions that lead to the formation of tetrazoles. The formation of thioamides by a reaction with P2S5 in the presence of sodium sulfite, as reported by Mason and co-workers, lead to membranes with increased gas selectivity (CO2/N2) but reduced porosity/permeability (51). This finding is in qualitative agreement with the results obtained on carboxylated PIM-1, which indicates that intermolecular interactions play again an important role. The Zn-catalyzed [2+3] cycloaddition of the nitrile group of PIM-1 with azides leads to the formation of tetrazole-PIM-1 (52). The modified PIM-1 shows again strong inter/intramolecular hydrogen bonding and a reduction of porosity as screened by cryogenic N2 adsorption, but an increased CO2–philicity and selectivity. Finally, the tetrazole unit can be modified by N-methylation (53), which lead to better solubility of the polymer, while maintaining good permeation properties.

In summary, it can be stated that modification of well-established microporous polymers (most extensively reported for PIM-1, but potentially also true for certain polyimides and TB-PIMs), can lead to a significant increase in their performance. Modification protocols could hence be more interesting for application development related to commercialization compared to the search for new systems (which requires still a trial-and-error methodology). New systems could however be identified more easily in the near future as some design principles have been derived during the last years (44). New developments such as soluble, conjugated microporous polymers provide, on the other hand, many fascinating promises. Within the next few years, they are certainly more interesting with regard to their fundamental properties rather than truly applicable in technology, while this might change upon gaining a broader knowledge base.

Microporous Polymer Networks

The preceding subsections introduced two major concepts for the creation of microporosity. A combination of the two concepts, that is frustrated packing and very rigid polymer strands, can indeed lead to a variety of microporous polymer network materials. The number of reported systems has consequently increased tremendously during the past years, including conjugated microporous polymers (CMPs), porous aromatic frameworks (PAFs), porous organic polymers (POPs), microporous organic polymers (MOPs), and many more. Additionally, routines for functionalization by postmodification of the networks as well as the direct use of functional moieties as building blocks have been developed, which results in a broad number of potential applications (see the section Applications). The major synthetic developments have been reviewed several times (13, 16, 18, 19, 54, 55). Figure 5 summarizes the chemical structures of some of the most common synthetic routes and building blocks.

figure

Figure 5. Schematic representation of common synthetic routes and monomers for microporous polymers.

The differentiation between networks based on the PIM principle versus networks based on the hyper-cross-linking approach is not always straightforward, and no clear line can be drawn. One example will be discussed to illustrate the phenomenon: While microporous CMPs have been reported even from monomers, which do not directly introduce a 3D contorted structure such as 1,3,5-substituted benzene derivatives, the situation is different for polyimides. A polyimide network based on a spirobifluorene core, that is a site of contortion, was shown to be highly microporous (SBET ∼ 980 m2 g−1). On the contrary, a polyimide network synthesized under the same conditions but based on 3,3′-diaminobenzidine, a monomer which does not directly imply high contortion, did not show pronounced microporosity (56).

The exact requirements that are necessary to achieve high surface areas are still under investigation. There are however some trends, which can be followed to reach high-surface area materials.

  1. Using tetrahedral monomers (such as tetraphenyl methane based monomers) does usually give rise to high surface areas, indeed on of the most porous amorphous polymeric material (PAF-1, SBET ∼ 5600 m2 g−1) is produced by Yamamoto coupling of tetrakis(4-bromophenyl)methane (57, 58). Comparably, a network prepared by Yamamoto coupling of 2,2′,7,7′-tetrabromospirobifluorene did show significantly lower surface areas (SBET ∼ 1300 m2 g−1) (59).
  2. The reaction should run to high conversions. This reduces the amount of unreacted end-groups, which are the least rigid components and might lead to pore blocking or collapse, respectively. In other words, the resulting network should feature a very high rigidity. For instance, the elemental analysis of PAF-1 indicated that there are almost no remaining end-groups; hence; a very high degree of cross-linking was achieved. This is also in line with earlier reports on spirobifluorene-based CMP networks by the Yamamoto reaction and reports on the synthesis of high molecular weight linear conjugated polymers. Such high conversions can be reached using both highly efficient chemistry and suitable solvents. As the polymer network grows, it should remain in a well-solvated state. As the reaction proceeds, the used solvent will become thermodynamically less good and phase separation will occur at some stage of the reaction. If this separation occurs in early stages of the reaction, it can be anticipated that the rate of conversion within the polymer-rich phase will drop and incomplete reaction will be observed. Furthermore, early phase separation will enhance the polymer–polymer contacts, which will ultimately also reduce the maximum surface areas. On the contrary, a late phase separation will allow higher conversions. Unfortunately, there are not many detailed studies available, which do clearly highlight this issue. An important study by Dawson and colleagues revealed that the choice of the reaction solvent does indeed have a significant influence on the porosity of various CMPs, which is due to the degree of conversion that can be reached (60). Likewise, Kiskan and Weber pointed out that the choice of the amine base, which is necessary for the synthesis of CMPs by the Sonogashira–Hagihara cross-coupling has a major impact on the observed specific surface areas (61, 62). Using diisopropylamine instead of triethylamine resulted in higher specific surface areas, which might be due to higher reactivity of the system.

Taking these general concepts into account, it is possible to synthesize high surface area polymer networks. To gain true benefits over common high-surface area materials such as activated carbon, functionalities must be included into the materials (63). This has been largely achieved within the past years, and a number of applications, ranging from catalysis, optical/sensing applications to gas separation, which benefit from well-defined porous polymers, will be discussed after some remarks on the micropore characterization. Before moving on, we however introduce important synthetic concepts, which allow the shaping of microporous networks into useful, macroscopic morphologies.

Shape Control of Microporous Polymers

Most technologies (eg, separation science, catalysis) require the use of porous materials in the form of particles, beads, membranes, or monolithic structures. Irregular powders are often not useful in large-scale applications as they tend to create high backpressures, inhomogeneous flow patterns, etc. Hence, there is considerable interest in shape control of porous materials, which is not only restricted to porous polymers but also to the field of zeolites of metal-organic frameworks. Synthetic shape control has serious advantages over postsynthetic shaping. It does not require the use of binder materials, which typically is inert mass (reducing the mass efficiency of the materials).

The preparation of polymer particles (latexes) and beads is a well-known procedure, especially for classic styrenic systems. Hence, it comes as no surprise that hyper-cross-linked polymers were prepared in the form of polymer beads already in the group of Davankov. Some recent advances were made in the synthesis of monodisperse microporous particles within the group of Sherrington and co-workers. Surfactant-free synthetic routes that lead to highly porous 420 nm sized particles were introduced (64, 65).

High-temperature miniemulsions based on a DMSO/isoparaffin/surfactant formulation were used by Crespy and co-workers to synthesize microporous melamine-based polymer networks (66). The resulting particles had diameters of 50 to 100 nm and showed surface areas of up to 300 m2 g−1. The porosity is lower compared to bulk materials, which could be due to the changed reaction conditions, which were necessary to achieve a stable miniemulsion. Alternatively, it can be imagined that surfactant might get trapped within micropores, from which it cannot be easily removed by solvent extraction (in analogy to nonane adsorption in microporous carbon).

The preparation of CMP nanoparticles was reported by various groups. Patra reported on the synthesis of fluorescent nanoparticles using Sonogashira cross-coupling polycondensation (67). The reaction was conducted in a miniemulsion (toluene/water/SDS) and particles of sizes between 30 and 60 nm were obtained. A comparable approach was used by Guo and co-workers, who synthesized colloidal CMPs by Sonogashira cross-coupling (68). They used however a different miniemulsion formulation than Patra (toluene/water/CTAB), resulting in slightly larger particles (d ∼ 100 nm). Specific surface areas up to 420 m2 g−1 were reported and the particles could be easily dispersed in various solvents. The surfactant-free preparation of porous polymer nanospheres for bioimaging applications was introduced by Wang and co-workers using a modified Suzuki-coupling protocol (69) Particles of 250–500 nm diameter were formed. It was suggested that these particles form by aggregation of smaller primary particles, which might be a consequence of the absence of surfactant.

The synthesis of highly porous polymer particles was initiated during the last few years and based on the rich knowledge of the handling and processing of polymer latexes, it can be expected that this field will grow continuously during the next years.

The preparation of monolithic micro/meso/macroporous materials is also of high interest. Multiscale porous systems combine high flow-through rates and very high surface areas, which makes them attractive replacements for packed beds of polymer particles, which have some drawbacks in terms of backpressure. The group of Kanatzidis introduced a monolithic multiscale porous aerogel based on polyreaction between terephthaldehyde and 1,5-dihydroxynaphthalene, see Figure 6 (70). The micro- and macroporous material has SBET ∼ 1200 m2 g−1 and its surface is decorated with hydroxyl groups, giving rise to functionality themselves or after modification by introduction of Ag nanoparticles. The material is made of agglomerated and cross-linked primary particles of a size of roughly 20–40 nm and possesses good mechanical strength.

figure

Figure 6. Polymerization reaction between terephthaladehyde and 1,5-dihydroxynaphthalene, Images of (a) as prepared and (b) dried Mon-POF, and (c) the cross section of the dried Mon-POF; and SEM image of Mon-POF (right-hand side). Reprinted with permission from Ref. 70. Copyright (2012) American Chemical Society.

Rose and co-workers reported monolithic micro/macroporous materials by cyclotrimerization of aromatic diacetyl compounds (71). The reaction in molten 4-toulene sulfonic acid yielded the formation of a monolithic material, which is again based on agglomerated primary particles of d ∼ 40–60 nm. Different reaction conditions, namely the use of SiCl4 as a source of HCl in an ethanol/toluene mixture, could not yield monolithic structures, highlighting the importance of the exact reaction conditions. Micro/macroporous polyurethane monoliths were reported by Jeromenok and co-workers (72). The monoliths were obtained by reaction of betulin (a natural diol dealing as stiff and rigid building block) and Desmodur RE in THF. Conventional drying leads however to significant shrinkage, which could potentially be reduced using scCO2 drying protocols. The usefulness of supercritical drying for preserving “extra” surface area provided by meso/macropores was indeed highlighted in the work of Farha and co-workers, who could also show that the catalytic activity of the materials benefitted significantly from the presence of the extra transport pores (190).

The preparation of aerogel-like structures as described above has some drawbacks, as the exact conditions that lead to the required phase separation toward small but well-connected particles or even continuous phases have to be identified in tedious screening experiments (73). A potential way around this issue is the use of the polyHIPE concept. The concept describes the use of a polymerizable high-internal phase emulsion (HIPE). HIPEs consist of a continuous phase (polymerizable), which is usually present in less than 26 vol% (74). Highly macroporous, interconnected monolithic materials are obtained after polymerization of the emulsion and removal of the internal phase. Schwaband co-workers prepared polyHIPEs based on styrenic monomers, which could be hyper-cross-linked using Friedel–Crafts chemistry in a second synthetic step (75). Highly micro- and macroporous materials with specific surface areas up to 1200 m2 g−1 could be prepared through this methodology. The polyHIPE concept was taken up by Zhang and co-workers, who prepared fully conjugated polyHIPEs based on photoactive CMP monomers using Suzuki coupling chemistry (76). The resulting monolithic macroporous materials did however feature only low specific surface areas (∼50 m2 g−1) in dry state. This could be due to the presence of alkylated monomers, which were needed for the formation of a stable emulsion. The presence of alkyl chains is known to reduce the overall porosity due to blocking or pore-filling effects (62, 77).

Finally, the preparation of cross-linked microporous films or coatings is of significant interest for potential applications. Wang and Zhang reported the synthesis of a microporous thermosetting film based on tetrakis(4-carboxyyphenyl)silane containing polyarylate precurosors. Smooth films could be coated and cured. Analysis of the films porosity indicated modest microporosity, which aroused as a consequence of the use of rather “soft” ester linkages. Those are however required as a precondition to obtain a robust, good film-forming coating. Moon and co-workers introduced a sol–gel method toward microporous urea-linked aromatic networks (78). The sol could be processed into free-standing films, which had some microporosity as screened by CO2 adsorption (see the section CO2 Adsorption for Micropore Analysis). Senkovska and Kiriy, who employed surface-initiated Ni-catalyzed Kumada catalyst-transfer polycondensation, presented an alternative method for the preparation of microporous coatings. Monodisperse SiO2 particles were surface modified to anchor the polymerization initiator. Subsequent polycondensation produced thiophene-based CMP coatings of ∼ 30 nm thickness. The porosity of the coating was analyzed (SBET ∼ 120 m2 g−1) and found to be lower compared to bulk materials prepared by the same methodology. It can be speculated that thin films do usually have a higher overall flexibility due to a larger amount of surface termini, which have lower rigidity, which might consequently lead to lower observable porosities.

A new route to cross-linked triazine framework based membranes was presented by Zhu and co-workers (79). Superacid catalyzed trimerization of aromatic dinitriles was used. The reaction was initiated at low temperature (−10°C), which resulted in a viscous, processable solution. Thermal treatment at 100°C of thin layers, which were obtained by casting, led to the formation of stable, free-standing membranes after removal of the superacid. The specific surface area was comparable to frameworks synthesized at high temperature using ZnCl2 as a catalyst (80).

A final example, which points into some different direction, is the use of X-ray lithography as a tool for simultaneous microfabrication and cross-linking of precursor polymer for the synthesis of thermally-rearranged polymers; see Figure 7, (81). The treatment of the films with hard X-rays does already lead to cross-linking and increased porosity/permeability. The method provides great potential for future implementation of microporous polymers, for example, in lab-on-chip applications. Such small-scale applications might be of high interest for sensing applications etc. and are in some contrast to large-scale adsorption applications, which would require the already discussed macroscopic shaping.

figure

Figure 7. Optical microscopic images of micropatterns and the corresponding Fourier transform infrared spectroscopy (FTIR) mapping spectra of the precursor polymer toward microporous films subjected to a dose of 25 kJ cm−3. Reproduced with permission from Ref. 81. Copyright © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany.

Characterization of Microporous Polymers

  1. Top of page
  2. Introduction
  3. Synthetic Methodologies
  4. Characterization of Microporous Polymers
  5. Applications
  6. Summary
  7. Bibliography

The characterization of nanoporous polymers, especially that of microporous ones, is not always trivial and is often more complicated compared to the analysis of inorganic microporous materials such as zeolites, activated carbon, or silica. The problems arise from the same issues that make porous polymers so attractive, namely their ductility, elasticity, and lightweight. The most important methods for the analysis of microporous polymers, ranging from gas adsorption to positron annihilation lifetime spectroscopy (PALS), scattering methods, and modeling approaches will be discussed in the following. A good overview on many of these methods (and some more specialized ones) can also be extracted from classic review papers on the estimation of free volume within polymeric membranes (82-84).

Gas Adsorption/Desorption Measurements

Gas adsorption/desorption measurements are probably the most widespread analytical method for the assessment of the porosity characteristics of microporous materials (1, 85, 86). It gives access to specific surface areas, the total accessible porosity ϕ of a system and allows usually the determination of PSD. Classically, N2 adsorption/desorption at 77.4 K is used for analytical purposes. The analysis of microporous materials using cryogenic N2 adsorption is however not trivial, and effects, which are not yet fully understood (such as low-pressure hysteresis), are observed. We will discuss some of the drawbacks associated with the interpretation of cryogenic N2 adsorption/desorption data after introduction of the main analytical models. CO2 adsorption at 273 K has been shown to be a versatile alternative for the analysis of microporous polymers and will be discussed afterward. The measurement of CO2 isotherms is also important with regard to a better understanding of the performance of microporous polymers in gas separation applications as will be discussed later.

Determination of Specific Surface Areas and PSD by Cryogenic N2 or Ar Adsorption

The specific surface area of materials is most commonly calculated on the basis of the Brunauer–Emmett–Teller (BET) method (see the relevant references in Ref. 85 for detailed information on the derivation). The method was developed for the analysis of planar surfaces by gas adsorption but can also be applied to curved surfaces as found in mesoporous materials. When it comes to the analysis of microporous materials, care has to be taken however. The BET method assumes formation of multilayers of adsorbate on the surface. This process does however overlap with micropore filling. The filling process happens already at very low pressures as a consequence of the overlap of the attractive potentials of the pore walls, which make condensation of the adsorbate energetically favorable. Hence, the classic BET model cannot work straight for microporous materials. Nevertheless, the BET model remains one of the most useful methods for assessment of specific surface areas, given it is used and interpreted correctly. Consistency criteria have been proposed (among them the right choice of the pressure range used for the BET analysis), which must be regarded, when microporous materials shall be analyzed by the BET method (87). A study of ultramicroporous metal-organic frameworks by Walton and Snurr gave evidence (based on experiments and molecular modeling) that the BET surface areas agree well with calculated geometrical surface areas, if attention is paid to the right analysis criteria (88).

Interestingly, the situation seems to be rather different for amorphous microporous polymers. A recent study by Hart and colleagues showed that BET surface areas of PIM-1 (experimental and modeled) do not agree well with the predicted geometric surface area, which was obtained by modeling (89). They suggested that the observed mismatch is a direct consequence of the pore topological characteristics and indicated that any amorphous adsorbent containing very small pores (so-called ultramicropores, d < 0.8 nm) is expected to show such mismatch. Results obtained from the analysis of other microporous systems support the finding that the determination of the specific surface area by the BET method has certain drawbacks. This seems to be especially true if a pronounced low pressure hysteresis is observed upon desorption at cryogenic temperatures (eg, N2 at 77.4 K). The origin of such hysteresis was frequently attributed to volume swelling (comparable to the dual-mode behavior of polymer at very high pressures) (90), and simulation studies were used to understand the potential swelling (91).

Recent studies indicate however that the reason could also be due to filling of micropores with limited accessibility and provide some evidence for the absence of swelling. Such so-called “restricted-access” micropores may arise as a consequence of the polymer flexibility and according to pore topology (92).

It was suggested that the filling of easily accessible pores could lead to contraction of narrow pore entrances toward other pores, thereby making them impassable. The thermodynamic origin of such contraction is the solvation pressure, which changes as the experiment runs on and allows relaxation of the contraction. This leads in turn to the filling of the restricted-access pores during the experiment. Although the low-pressure hysteresis can still be regarded as a not yet completely solved issue, two consequences can be drawn from the findings:

  1. The low-pressure hysteresis at cryogenic temperatures is not a simple kinetic restriction. This was indeed evidenced by various experiments using changed equilibration settings, which did not result in changed isotherms (92, 93).
  2. The findings raise the question, whether the desorption branch might be more suitable for the determination of the specific surface area and the micropore volume of polymeric samples, that exhibit a strong low-pressure hysteresis. Examples have been reported, where the use of the desorption branch can yield results, which are in agreement to results obtained by alternative measurements (such as CO2 adsorption, which does usually not suffer from hysteresis effects), whereas the adsorption branch analysis resulted in strong underestimation of the surface area and pore volume compared to other methods (92). This is part of an ongoing debate and as of autumn 2013 there is still no final picture on the exact determination of the specific surface area (and interpretation of its physical meaning) in microporous amorphous polymers.

An interesting side note and bridge to the earlier statement on the preconditions of obtaining polymers of ultra-high specific surface areas (ie, a high degree of condensation, high rigidity) is the fact that the amount of low-pressure hysteresis is most probably also related to the softness of the polymeric backbone. Likewise, the appearance of a significant hysteresis should not go along with ultrahigh surface areas (and does not so usually).

Nevertheless and as a consequence of a lack of comparably widespread methods, the BET model can still be regarded as a versatile tool for comparison of different polymer systems given that the above-mentioned consistency criteria are taken into account. If possible, alternative methods should be used additionally. Such methods include density-functional-theory (DFT) based methods, such as traditional non-linear DFT (NLDFT) and more recently developed quenched-solid DFT (QSDFT) (94). Basically, DFT methods take density fluctuations of the confined fluid into account, which are typically observed in micropores, where the adsorbate cannot be considered as a homogeneous phase (like in bulk state). Model adsorption/desorption isotherms can be calculated for various pore shapes and sizes based on the known density fluctuations and the intermolecular interaction parameters. The set of model isotherms can serve, simply speaking, as a pool (kernel) for a fitting process of the experimental isotherm, which in turn yields information on pore sizes and their distribution and specific surface areas. The strength of DFT methods lays in the wide variety of pore shape assumptions as well as its potential to bridge the length scales from micro- to mesopores. We refer the reader to appropriate review articles for more details (94, 95). QSDFT methods take additionally surface heterogeneities into account and may thus provide the closest-to-reality picture that is available currently. Again, attention has to be paid however to the low-pressure hysteresis phenomenon and analysis of both isotherm branches is recommended (see Fig. 8). To sum up, it can be stated that DFT-based methods have largely replaced classic (semiempirical) methods for the analysis of microporous materials, which are however still available and might find their use.

figure

Figure 8. Exemplary N2 isotherm (77.4 K) showing a pronounced hysteresis upon desorption (left) and schematic drawing of the possibility of formation of new pores (volume swelling) or of kinetically hindered filling of existing pores upon adsorption (right). Reprinted with permission from Ref. 92. Copyright (2013) American Chemical Society.

CO2 Adsorption for Micropore Analysis

CO2 adsorption at 273 K can be used as an additional tool for the analysis of micropores and the respective micropore surface area next to the classic cryogenic N2 or Ar adsorption/desorption that are conducted at 77.4 or 87.3 K, respectively. The measurements do usually not show significant adsorption/desorption hysteresis, which indicates that CO2 can probe the total micropore volume at the given conditions. CO2 adsorption/desorption poses less experimental demands compared to N2 adsorption. As a consequence of the high saturation pressure of CO2 at 273 K (p0 ∼ 26,140 mmHg), it is possible to reach very low relative pressures without the use of turbomolecular pumps. The high measurement temperature of 273.15 K does usually also allow faster measurements. A drawback of CO2 adsorption at 273.15 K is however the fact that relative pressures p/p0 > 0.03 can only be achieved using high-pressure adsorption. Hence, only micropores with sizes d < 1.2–1.5 nm can be analyzed under normal conditions. The use of CO2 adsorption as an additional tool for the analysis of microporous carbons has been known for quite some time and analytical models (including NLDFT and comparable methods) have been developed and commercialized (96-99). The advantage of combined analysis of CO2 and N2 adsorption/desorption isotherms is depicted in Figure 9, where it is shown that CO2 can give information on pore size regimes that can be hardly accessed by classic N2 adsorption. The mismatch between CO2 and N2 adsorption results can be attributed to the effect of low-pressure hysteresis in cryogenic gas adsorption and can be overcome by using the N2 desorption branch (see the discussion above) (92).

figure

Figure 9. Cumulative pore volume distribution of a microporous polyester network obtained from CO2 and N2 adsorption. Note that CO2 can give access to information, which could not be accessed by N2. The underestimation of the pore volume by N2 adsorption is obvious if the adsorption branch is used for analysis. The mismatch is a consequence of the low-pressure hysteresis and discussed within the article. Reprinted with permission from Ref. 92. Copyright (2013) American Chemical Society.

The use of CO2 adsorption for the analysis of microporous polymers has increased during the past years. It is especially noteworthy that there are various reported cases, in which CO2 could probe pores, that cryogenic N2 adsorption could not probe easily or even not at all (33, 34, 43, 48, 72, 78, 93, 100, 101). Furthermore, CO2 adsorption could also prove pore size changes in modified PIM-1 (48), that is details which are not easy to resolve using cryogenic N2 adsorption due to the above-mentioned problems associated with the low measurement temperatures and corresponding diffusional limitations. Despite the successful story, we remind the reader that there are some aspects that need always requirement before any finer details of the CO2 adsorption are discussed: If the polymer sample under analysis contains strongly interacting functionalities (eg, open metal centers, ions), the pore size analysis might be flawed. CO2 would preferentially adsorb at such positions first, leading to an overlay with the filling of the narrowest pores, which does also happen at very low relative pressures. Hence, it is advisable to measure the adsorption at another temperature (eg, 283 K) to calculate the isosteric heats of adsorption qst, which give information on the interactions strength. Reasonable methods to do so have been suggested within the zeolite/metal-organic framework community (102, 103). If no extraordinary high heats of adsorption are observed (typically qst < 35–40 kJ mol−1), we suggest to make use of the extra information that CO2 adsorption can provide.

Another issue, which might need more thorough investigation within the next years, came just recently to our awareness. Zukal and co-workers reported on microporous CMPs, which were analyzed by N2 and CO2 adsorption (93). It was observed that the CO2 uptake was strongly depended on the equilibration conditions, which were employed during the measurements (see Fig. 10). This behavior is typically only observed in strongly interaction systems, which show chemisorptive behavior (104, 105). Such behavior was however not expected for the CMP-like structure, and the effect was attributed to the softness of the materials. So far, the effect has only been reported for CO2 adsorption at ambient temperatures (293 and 333 K) and it remains speculative so far, whether the effect is also present in measurements at lower temperatures. This would have serious impact on the analysis of microporous polymers by CO2 adsorption but also on the screening of microporous polymers as CO2 capture materials.

figure

Figure 10. CO2 isotherms of microporous CMPs measured at 333 K at different equilibration conditions. Reproduced with permission from Ref. 93. © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany.

Finally, it should be stated that CO2 adsorption could also be conducted at 195 K, which typically results in isotherms that are comparable with classic cryogenic N2 or Ar adsorption isotherms and are analyzed comparably (92, 106).

Other Gases

The use of other gases for analytical purposes has been reported frequently as well. Hydrogen adsorption at 77.4 K was measured for many microporous polymers as a consequence of the search for hydrogen storage options throughout the past years. The use of hydrogen adsorption for analytical purposes is however less widespread. Generally, H2 can probe very small pores as a consequence of its small kinetic diameter and the rather high measurement temperature, which is above the critical temperature of H2 (no condensation within the pore space). This has been used in detailed studies on the pore accessibility, which focused especially on samples that did not show significant N2 adsorption at 77.4 K (33, 34, 56). Specific surface areas have been determined from such measurements using a Langmuir approach; it was, however, stated that the Langmuir model could not account for a good fit of the whole subatmospheric isotherm. Although a quantitative analysis (except the extraction of uptake capacities at high pressures) is hence not easily possible, the measurement of H2 adsorption remains nevertheless a good option to screen microporous materials for pore accessibility at low temperatures.

Xenon adsorption is reported as a tool for the analysis of microporous polymers from time to time. It can be used as a prescreening for the usability of 129Xe-NMR, which is a versatile method for the analysis of microporous materials on its own (see the section Spectroscopy). Xenon adsorption within PIMs at ambient and high (100°C) temperatures up to 3 bar was reported by Emmler and co-workers within a detailed characterization study (107). The data revealed the possibility of conditioning of PIM-1 films at high temperature and gas pressure and was analyzed using either classic methodologies such as Horvarth–Kawazoe or the dual-mode adsorption model.

Spectroscopy

Spectroscopic methods, especially nuclear magnetic resonance (NMR) and infrared (IR) techniques can give valuable information on the chemical state of adsorbates but also on porosity characteristics such as pore sizes. In the following, we discuss mainly the use of 129Xe-NMR, for reviews on the basics of this technique, please consult appropriate references (108, 109). The use of solid-state NMR as a tool for the characterization of the chemical nature of the porous material will not be discussed here. The use of pulsed field gradient NMR for the analysis of diffusion in microporous materials (such as zeolites or Metal-Organic Frameworks (MOFs)) will also not be discussed. For more information on this technique, which has not been applied yet to microporous polymers to the best of our knowledge, we refer the reader to Refs. (110) and (111).

129Xe has a good NMR sensitivity and rather high natural abundance of ∼ 26 %, which makes it to a good NMR probe. It is an inert gas with a large electron cloud, whose disturbance by its chemical environment is transmitted to the nucleus. Its chemical shift is hence highly sensitive to the environment of the Xe atoms. This can be used to analyze a number of materials. Indeed, xenon can be adsorbed in porous materials such as zeolites or activated carbons and used for the analysis of the pore wall chemistry and processes within the materials. Additionally, it can be pressed into amorphous soft matter such as common polymers (eg, polyethylene) (112), or even small proteins or organic cage-like molecules (113, 114), and dealt as a probe for the free volume of these materials (82).

The observed chemical shift of confined xenon, which can be well separated from free xenon, is usually pressure dependent and increases with increasing xenon pressure. A model developed by Fraissard and co-workers allows a quantification of the pore size based on the chemical shift of confined Xe atoms (109, 115). It is generally assumed that the observed chemical shift δ is a sum of various contributions:

  • mathml alt image(1)

where δ0 is the reference shift. δXe–XeρXe arises from the Xe–Xe collisions and is expected to vary linearly with the xenon density at low xenon loading and it shifts to higher values with increasing the xenon concentration. δS reflects the interactions of the xenon atoms with the surface. The larger these interactions are and/or the easier diffusion inside the pores is, the smaller the δS value (116). The term δE is the shift caused by the electric field that originates from the presence of cations in the porous material (eg, in zeolites), δM describes an extra term accounting for the presence of paramagnetic centers, and δSAS does finally relate to the shift due to strong adsorptions sites. The influence of the latter three terms can however be neglected in the analysis of most polymeric materials.

Several methods have been developed for the analysis of pressure-dependent 129Xe-NMR methods, which also include the differentiation of the signal in analogy to the dual-mode adsorption (especially important at high pressures) (107). An empirical relation, which is based on the results of Xe NMR investigations in zeolites (115), allows the determination of the mean free path λ (given in angström) of the confined Xe atoms, which is in turn related to the pore sizes:

  • mathml alt image(2)

Depending on the assumption of either spherical (Dsp = 2λ + DXe) or cylindrical (Dcy = λ + DXe) micropores, one obtains slightly different values for the pore radius. DXe is the diameter of the xenon atom. A generally accepted value is DXe = 4.4 Å, that is its van der Waals diameter, even though there is some discussion on the correct size of the xenon atom (82).

129Xe-NMR has been used for the analysis of microporous polymers as early as 1995 (117) and has continued to be an attractive method since then. An exemplary picture is given in Figure 11. Pressure-dependent 129Xe-NMR spectra of two different microporous polymer networks are depicted, and a clear increase in the chemical shift with pressure is observed. An additional signal (δ = 8 ppm) is visible at high pressures, which is related to xenon confined within the interparticulate void spaces. The analysis of the Xe data indicated comparable pore sizes for both materials (D ∼ 0.7 nm), which was not predicted from the gas adsorption data analysis, which suffered itself from the previously discussed low-pressure hysteresis phenomenon (118).

figure

Figure 11. Pressure dependent 129Xe NMR spectra of Xe confined within the micropores of two different spirobifluorene-based polymer networks. (Pressure increases from a to i). Free xenon can be observed at higher pressures (δ = 8 ppm). The characteristic anisotropy (as a consequence of Xe-wall interactions) of the signal of confined Xe can be observed as well. Reprinted with permission from Ref. 118. Copyright (2010) American Chemical Society.

129Xe-NMR can be performed at subatmospheric, but also at high pressures up to several atmospheres. In polymer samples, Xe gets molecularly dissolved at high pressures (Henry sites within the so-called dual mode adsorption model). This effect can impact the data analysis. Emmler and co-workers showed that the effect can however be ignored for common microporous polymers like PIM-1, if the Xe pressure does not exceed 3 bar. The error was shown to be within the expected experimental uncertainty, and pore sizes of D ∼ 0.8–0.9 nm were determined for PIM-1 (107).

In summary, xenon adsorption is a good addition to the analysis of microporous polymers. It seems to be rather sensitive to the somehow averaged pore size (eg, Xe-NMR indicated very small pore sizes of D ∼ 0.6 nm for an ultramicroporous polymer network, in accordance with very small pore sizes indicated by CO2 adsorption) (43), yet, there is now clear way how it can give an idea on the PSD. The development of modern NMR techniques and the continued research done also within the MOF field (119) lead however to the expectation that 129Xe-NMR will also continuously be developed and contribute strongly to the progress in the characterization of microporous polymers. Likewise, contributions from the analysis of flexible cage-like organic molecules such as clathrathes or microporous peptide-based materials by advanced 129Xe-NMR techniques will help further understanding of the particular effects observed in microporous polymers (113, 120).

13CO2 can also be adsorbed within micropores of polymer (networks) and analyzed by NMR techniques (72, 119). It is however less sensitive to its environment, and the chemical shift is only a few ppm different to that of free CO2 (δ = 126 ppm). It is however possible to detect chemisorbed CO2 (119, 121), which does show a strong change in the chemical shift. In the case of extremely narrow pores, in which CO2 cannot rotate anymore freely, a highly anisotropic and broad signal is observed (122), which can be used to identify such narrow pores.

IR spectroscopy of adsorbed CO2 can also be considered as a versatile technique and allows the extraction of finer details of adsorbed CO2, especially of chemisorbed CO2 (105, 123, 124). The interaction of various polymers with CO2 at medium to high gas pressures was also analyzed previously, and significant information on band shifts as a consequence of intermolecular interaction are available (125). More recent developments allow also the coupling of the adsorption process with in situ IR monitoring, which does also allow the extraction of kinetic information, which is for instance crucial for the performance of adsorbents under real conditions (126).

In summary, spectroscopic measurements are a powerful addition to other analytical methods and are expected to flourish within the next years. They are especially interesting with regard to their potential to monitor in situ processes, but also as an independent means of pore size determination. Most methods have been applied successfully to zeolites or metal-organic frameworks so far; however, there is no hindrance to apply them to microporous polymers as well and especially 129Xe-NMR has indeed been used extensively already.

Positron Annihilation Lifetime Spectroscopy

Another technique for the characterization of microporosity is PALS, and a number of overview articles are available (82, 83, 127-129). The technique is based on the formation and annihilation of positronium particles, which can be formed by a reaction of positrons and electrons. Positrons (p+), which are the antiparticles of electrons are created in a p+ source and allowed to interact with the sample to be analyzed. In rather infrequent events, they can react with electrons to form a positronium particle (Ps), which can differ in spin, giving rise to ortho-Ps and para-Ps. The Ps particles can interact with matter and annihilate under the emission of radiation, which allows the measurement of the o-Ps and p-Ps lifetimes. Detailed information on the necessary instrumentation (p+ source) are available from above-mentioned references.

Generally, the o-Ps lifetime τ3 is of interest for the analysis of free volume in polymers. It possesses longer lifetimes, which are however reduced upon confinement in porous media. The lifetime depends hence on the presence, size, and shape of voids, which enables the determination of pore sizes. This is illustrated by a semiempirical relation (Tao–Eldrup equation), which connects the lifetime to the pore radius:

  • mathml alt image(3)

where R3 is the radius of the spherical free-volume elements (FVE)/pore and R0 = R3 + 1.66 Å (an empirical term, determined using standardized pore sizes of zeolites). The above equation holds for spherical pores and is widely used for microporous polymer materials; modifications for different pore shapes have however been suggested (127).

An additional o-Ps lifetime (τ4) has been identified in the case of microporous polymers and other high free volume polymers, which is usually interpreted in terms of a bimodal or broad PSD (82, 130). A variety of data analysis packages have been developed, which can (next to the pore size) also give indications on the number density of pores (based on the relative intensity Ii of the corresponding lifetime i) and correspondingly also information on PSDs (see, eg, Refs. (84) and (127) and references therein).

Although PALS is a rather widespread method in the field of polymeric separation membranes, the situation for microporous polymers is still a matter of debate. The analysis of PIM-1 by PALS by different groups gives rather different results. While Cornelius and co-workers (130) and Yampolskii and co-workers, do report a bimodal FVE distribution for PIM-1 (131), only one lifetime and according pore size was reported by Fritsch and co-workers (107), which illustrates the problem of proper data interpretation. The presence of a bimodal PSD is not predicted by molecular modeling (132); however, a rather broad distribution can be imagined. A recent report of Sanchez and co-workers compared the PALS results obtained on thermally rearranged polymers with predictions obtained by atomistic modeling, which showed reasonable agreement in some cases, while the bimodality observed by PALS was indeed not predicted by modeling in all cases (see eg, Fig. 12) (133).

figure

Figure 12. Cavity size distributions of polyimides obtained from molecular modeling (CESA) and PALS. Please note that PALS does not predict cavities of ∼5 Å, which is in contrast to modeling methods. Reproduced from Ref. 133, Copyright (2011), with permission from Elsevier.

Despite some drawbacks of the finer data analysis, PALS is certainly a very valuable addition to the pool of available characterization methods. An advantage of the technique is surely its potential to sum over all pores, including nonaccessible pores (closed porosity), which can otherwise only be accessed by scattering methods or molecular modeling. Another clear advantage is the possibility to analyze the porosity (and its changes) also in response to external parameters such as temperature, which cannot be easily done by gas adsorption (107, 134). Although PALS has been used most often for the analysis of microporous/high free volume polymers, its use can also be expanded to the analysis of mesoporous materials (135, 136).

Molecular Modeling

Molecular modeling of microporous polymers has become a powerful technique not only for the predictive but also quantitative analysis of new materials. Most efforts spent so far targeted the modeling of hyper-cross-linked polymer networks (23, 137, 138), the modeling of soluble high-free volume polymers, including PIM-1 and its derivatives (89, 132, 139-145), and thermally rearranged polymers (133).

Abbott and Colina provided a new method for the in silico polymerization of hyper-cross-linked networks recently, which give also a good summary on previous work on such systems (138). They show that their simulation protocol results in structures, whose properties (such as gas storage, density) agree well with those of real samples, which provides an indirect way of validating their method. It can be expected that such methodologies can help to predict new microporous structures as important experimental conditions such as necessary cross-linking density and reactant concentration during the reaction can be analyzed before real-space laboratory action takes place.

PIM-1 and its derivatives represent a major direction in the search for new membrane materials for gas separation applications. Hence, their molecular simulation was put forward, as a better understanding is still required along with predictions, which derivatives might have the best potential for application as membranes. To the best of our knowledge, Heucheland co-workers were among the first groups, who worked on the simulation of PIM-1 (140, 145). They could already show that despite the rather stiff structure of PIM-1 as supposed from paper chemistry, PIM-1 can undergo significant elastic chain deformations. Later on, it was shown that the amount of chain deformation does also relate to the performance of membrane materials (45). The simulation of PIM-1 and also the introduction of new ways of structure generation in silico were also put forward by Colina and co-workers recently (89, 132, 141, 143). The models of PIM-1 were validated by comparison to experimental scattering data, which is facilitated by the especially well-developed scattering pattern of PIM-1 (141). Although the created models can explain some of the experimental gas adsorption (CH4, CO2) data fairly well (132, 145), they can however not correctly reproduce the experimental observation of the low-pressure hysteresis, which is observed for PIM-1, in a straightforward way, but only by indirect means (91). This questions somewhat the validity of some of the computer-generated models available so far, and additional contributions from the modeling community will be needed. One of the problems is related to the validity of the experimental data, which is sometimes used for model validation, such as densities. It is well known that, for instance, helium pycnometry can have serious experimental problems for the analysis of microporous materials. Nevertheless, there can be no doubt that these questions will be tackled and solved within the next years. This would make molecular modeling ultimately to some even more powerful characterization (and more importantly) predictive method for the analysis of microporous materials.

Finally, a short note shall be given on the “misuse” of modeling as a source of picture generation only. Good modeling work can provide much more information than only pictures. Trewin and Cooper showed recently that simple predictions based on the molecular geometry of the monomers only may not be sufficient to create a reasonable model of highly microporous structures (58). They highlighted the various possibilities provided by modeling to explain the extraordinary high specific surface area and porosity of PAF-1 (57) and showed that an amorphous model can explain the observed porosity more realistically than a simplistic highly crystalline order (as suggested originally but not proven by diffraction measurements). Simple models might hence be validated with appropriates example in comparison to the real world, especially those with regard to potential long-range order, i.e. crystallinity.

Scattering Methods

Scattering methods represent another option for the analysis of nanoporous materials. Solvent-filled or dry networks can be analyzed. Similarly to PALS, scattering methods can also probe closed porosity and allow the monitoring of processes (such as adsorption events) in situ (146, 147). Various methods can be used such as small-angle X-ray scattering (SAXS) experiments, which can give information on the nanometer-scale, whereas wide-angle X-ray scattering or diffraction (WAXS, WAXD) (also known as X-ray diffraction, XRD), gives information on the atomic (Å) scale. Neutron scattering can be used alternatively to X-ray scattering.

Small-angle scattering was used frequently for the analysis of microporous carbon materials (148-150). The analysis of hyper-cross-linked microporous polymeric resins by small-angle neutron scattering (SANS) was reported in 1996 by Hall and Sherrington. The scattering curves were analyzed assuming fractal behavior. Pore sizes and the amount of open/closed porosity could be determined based on contrast-match experiments using SANS (151, 152). SAXS was used by various groups to access information on the porosity of microporous polymers and is more widespread compared to SANS due to significantly lower experimental demands. Of most interest for the analysis of microporous polymers is the high q-range between 1 and 8 nm−1, corresponding to real space d values of 6–0.8 nm, where q = 4πλ−1 sin θ is the scattering vector (λ is the wavelength of the X-rays, θ is the scattering angle). Lee and co-workers reported the appearance of a peak in the SAXS patterns of thermally rearranged polymers, which were not found in the nonporous precursor materials (40). The peak could be attributed to FVEs (cavities) and was analyzed assuming weakly correlated, spherical FVEs. Cavity sizes of ∼ 6.3 Å were deduced from the scattering data analysis.

Comparably, SAXS measurements on carboxylated PIM-1 were conducted by Weber and co-workers (48). The observed halos were interpreted again based on the assumption of weakly correlated FVE. The SAXS patterns of carboxylated PIM-1 was shown to depend on the sample temperature. A change toward smaller q (ie, larger sizes) was observed, which was attributed to the relaxation of the polymer chains upon breakage of hydrogen bonds in between them. Molecular simulations could indeed show (see the section Positron Annihilation Lifetime Spectroscopy) that the polymer chains can be deformed to some extent. Such deformation could close pores, and this will happen if it is favorable on the total energetic balance (deformation penalty vs gain of reduced surface energy). Intermolecular interactions can stabilize such interactions but are thermally unstable. It is worth to point out at this stage that scattering techniques are especially useful for the monitoring of dynamic processes and might find more use in the future.

The situation is special for PIM-1, which possesses very well-defined scattering patterns with a variety of halos/peaks (which also deals for validation of atomistic models; see the section Positron Annihilation Lifetime Spectroscopy). The situation is however rather different for many other microporous polymers, which show only ill-defined scattering patterns. The analysis of various spirobifluorene-based polymers by SAXS revealed a two-phase behavior with the appearance of the so-called porod regime (56, 153, 154). The porod regime, which is characterized by a continuous decay of the scattering intensity I with I(q) ∝ q−4 or I(q) ∝ q−3 in the case of slit-smeared data, is characteristic for any porous system. It is intimately linked (via the porod length) to the surface area of the system and its porosity. It allows furthermore the extraction of model-independent number-averaged pore sizes. For more details, the reader is referred to the respective literature on small-angle scattering (149, 155, 156), while we discuss the results at this stage. Pore sizes and specific surface areas, which were determined by SAXS measurements, were in reasonable agreement with data obtained by cryogenic gas adsorption measurements. This was however not the case for materials where N2 adsorption at 77.4 K could not probe microporosity. For example, a spirobifluorene-based polyamide network did not show pronounced microporosity by cryogenic N2 adsorption (56). SAXS measurements indicated conversely that the sample was microporous and had a surface area of around 300 m2 g−1. This was supported by H2 adsorption at 77.4 K, which could indeed provide evidence for the presence of microporosity. Hence, SAXS can be a method to check potential microporosity, which is not always seen clearly by cryogenic N2 adsorption.

SAXS is also a rather powerful tool for the analysis of larger micropores in the transition regime to small mesopores. A comparative analysis of covalent-triazine frameworks (CTF)-derived materials could show a very good agreement between porosity characteristics derived by either N2 adsorption at 77.4 K or based on model-independent SAXS analysis (157).

The analysis of WAXS data can also be a source of additional information on microporosity. Again, PIM-1 is special as it shows a very well-defined scattering pattern, which was analyzed in detail; see Figure 13 (141). There is some discrepancy between the simulated pattern and the experimental curves in the low q range (SAXS area). The processing and aging/thermal history seems to have serious impact on these regions, which is somehow related to the overall porosity, though not yet clearly defined (91). The agreement is otherwise almost perfect.

figure

Figure 13. Simulated structure factor S(q) of PIM-1 compared to experimental WAXS intensity of PIM-1 samples with three different processing histories: a precipitated powder, a solution-cast film, and an identical film that was also treated with methanol. Reprinted with permission from Ref. 141. Copyright (2011) American Chemical Society.

Most of other microporous polymers show only one or two halos in the WAXS regime. One halo is often observed at real-space d values of 4.5–6 Å, which is mostly attributed to the average chain-to-chain distances. An additional halo at larger d spacing of ∼7.8 Å has been observed for microporous-soluble polymers based on spirobifluorene and was attributed to larger separation distances (due to the micropores; see Fig. 14) (32). It was shown that the relative intensity between the two halos varied depending on the processing history of the polymers. Precipitation from a weakly interacting solvent yielded a higher intensity of the peak corresponding to a higher amount of chains separated by larger distances, which was in accordance with a higher porosity of these samples. Similar results were observed for microporous polyimides based on binaphthalene monomers, for which the peak at lower scattering angle might however be due to the structure factor of the polymer rather than larger packing distances (33, 34).

figure

Figure 14. WAXS patterns of various spirobifluorene-based polymers precipitated from DMAc (a), resulting in nonoporous polymers and precipitated from THF or CHCl3 (b), resulting in microporous materials. Reproduced with permission from Ref. 32. Copyright © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany.

Comparative analysis of WAXS patterns can hence give some indication on the processing history, and first ideas on the packing density if two (independent) peaks are observed. Such methodology is however purely based on comparison of samples, which differ in their processing history and can hence be only regarded as an additional analytical tool. Finally, it was shown that the halo position (and the corresponding d value) does not necessarily correlate with the observed porosity. Polyimides that showed comparable d spacings of 6.9 Å did show completely different porosity characteristics, highlighting again that WAXS data require careful interpretation (34).

Summary

We have given a summary on the main characterization methods for microporous polymers with a special emphasis on gas adsorption. All methods suffer from more or less dramatic drawbacks up to date. Model assumptions have to be taken in almost all methods, which might have strong impact on the obtained data. A careful choice of ideally various methods is recommended, and the primary data (adsorption/desorption isotherms, complete scattering patterns, full PALS data) should be reported for better comparability alongside with results obtained from the usage of the widespread software packages and methods.

Nevertheless, the understanding of microporous polymers has improved significantly during the past years and might continue to develop during the next years. Given the importance of a thorough understanding of the porosity for any analysis of the materials performance in various technologies, we strongly encourage continued development of suitable characterization methods. Namely, direct imaging of porosity can be possible by means of AFM or STM methods (158-160); see Figure 15 for an example.

figure

Figure 15. Topographic study of CC3-R (an organic porous cage compound) crystal surfaces. (a) Modeled surface topography based on a van der Waals surface of the crystal face along Miller plane (1 1 1). Top 4 Å of the molecular lattice is superimposed as a guide for the eye. Indicated distances (in red) are measured between the most exposed atoms of the respective structures (based on X-ray crystallography). (b) Software zoom on contact mode topography AFM images of the (1 1 1) faces of CC3-R. (c) Profile plots along the pathways A, B, C, and D (in red, green, blue, and pink) for the measured contact mode topography AFM images of the (1 1 1) faces. Reproduced from Ref. 158 with permission of The Royal Society of Chemistry.

So far, gas adsorption and scattering will most probably remain the most important methods due to the fact that they are used in many laboratories. NMR methods could reach a comparable level of distribution. PALS is bound to a positron source, but also rather widespread, and opens many possibilities for researchers. Other exciting methods might however be underdevelopment right now, which will ultimately allow deeper insights.

Applications

  1. Top of page
  2. Introduction
  3. Synthetic Methodologies
  4. Characterization of Microporous Polymers
  5. Applications
  6. Summary
  7. Bibliography

Microporous polymers have been suggested for the application in a number of different technologies. Many reviews exist, which summarize most of the current state of the art methods, and we will refer the reader to them, whenever appropriate. We will furthermore try to critically highlight a few concepts and findings, which are worth of consideration, but also add information on the most recent trends and ideas. Some overview on the applications suggested so far can also be found in the more general reviews on microporous polymers (14, 16, 18, 19, 55). This section does mainly give an overview on academic research, which might open applications within the near/medium future. It can be expected that most applications might initially target small or niche applications, where only small amounts of the still rather expensive materials are required. Furthermore, it can be expected that microporous polymers are especially attractive if they combine high surface areas with pronounced chemical functionality. Applications, which do only require high specific surface areas, but not necessarily chemical functionality, might work in a more economic way using common materials like activated carbon. To the best of our knowledge, microporous polymers have been commercialized as membrane materials (eg, certain polyimides) and as hyper-cross-linked resins (Davankov resins, eg. Hypersol-Macronet® by Purolite) so far. The commercialization of materials developed lately might benefit from recent advances in cost-effective synthesis and shaping protocols.

Gas and Vapor Separation

Gases play an important role in various industries, and pure gases are needed often. Additionally, the treatment of waste streams, such as flue gases must be taken care of as a consequence of environmental demands. Gas/vapor separation or purification by porous materials has been a long-standing issue and was investigated using various materials such as zeolites or high-free volume polymers. Recently, the topic of gas adsorption in microporous polymers got considerable interest within the framework of the so-called hydrogen economy or carbon capture and sequestration techniques. The field of gas separation can generally be separated into membrane-based methods and adsorber bed–based methods, which use temperature or pressure swing adsorption/desorption cycles. We discuss mainly gas separation within this article. In principle, gases (such as hydrogen or methane) could be stored in porous polymers and reduce the necessary pressures; such technology is however much less developed (and has a uncertain future) and will hence not be discussed. Basic ideas can be found in the literature (12, 23, 161, 162).

Membranes

The use of polymeric membranes for gas or vapor separation is known for a long time, and several books and reviews are available (of which we refer only to the most recent ones) (29, 163-165), which give a concise overview on the potential of polymeric membranes and the underlying separation mechanisms. The use of high free volume polymers has been reported, and most microporous polymers are rather comparable to most of those, although several significant performance improvements have been reported. It was also recognized that the transition between the solution–diffusion mechanism, which is predominant in dense membranes, and pore–flow mechanism occurs in the pore size range of 5–10 Å, that is exactly at the pore size regime of microporous polymers. The delicate interplay between the stiffness of the polymer chains, the corresponding chain segment mobility (which is in an important parameter for gas diffusivity in the solution–diffusion regime), and the micropore size (which relates back to the potential molecular sieving properties) has been highlighted recently (166). PIMs based on TB (45) show a very good selectivity of O2 over N2 at high permeabilities, whereas the same polymers show a low CO2 over N2 selectivity. This trend can be explained using the Knudsen model, which relates the diffusivity to the molar mass of the gas. N2 is significantly lower in mass than CO2, which could give a much higher N2 permeation rate, resulting ultimately in a lowered selectivity. Such molar mass effect is less pronounced for O2/N2, resulting in much higher selectivities that exceed the 2008 Robeson bound (166). As stated initially, the field of gas separation membranes does flourish and many membranes have found industrial applications (eg, air separation, production of technical grade N2, CO2/CH4 separation, and many more) (29). In the corresponding recent perspective, Yampolskii also pointed out a variety of unsolved problems of gas separation (He separation from natural gas, hydrocarbon separation) and we refer the reader to those problems that are under investigation by many researcher units currently.

A significant amount of interest was spent recently on so-called mixed-matrix membranes (MMM), which combine membrane-forming polymers with fillers of permanent porosity (such as zeolites, MOFs, or organic cages) that have superior separation performance but cannot be easily processed into membranes. This approach can give highly permeable and selective membranes, which can surpass the classic limitations of polymeric membranes (known as Robeson′s upper bounds), and was shown to be more resistant against aging. MMMs combining PIM-1 and organic cages, silicalites, or zeolitic imidazolate framework (ZIF-8) have been reported and analyzed for various gas separation applications recently. Positive effects, such as reduced aging and improved performance, have been reported for a number of gas pairs but also for pervaporation (ethanol/water) (167-169).

More improvements can be expected in the near future, especially as the membrane market was growing much faster than expected during the past years according to some reports (29), and it can be expected that the research activities and the use of microporous polymers in various separation applications will continue to increase.

Adsorber Materials

Porous materials can also be used for gas separation in the form of adsorber columns. Doing so, pressure (vacuum)-swing adsorption or temperature swing adsorption processes can be designed (102, 170, 171). Such are well known from the zeolite, activated carbon, and MOF community and can also be modeled and implemented easily on industrial scales. Most attention was paid to the separation of CO2 (eg, from flue gases) during the past years, and many research units analyzed the potential of porous polymers as adsorbents and the results have been reviewed (172). At this stage, we cannot discuss all of the aspects in detail but just remind a few points. Porous polymer networks need to compete with highly porous and selective zeolites and activated carbon in the capture of CO2, and their performance should be benchmarked by comparison to such materials. Ideally, capacity, selectivity, which is usually determined by using ideal adsorbed solution theory (see, eg, Refs. (72) and (173)), and adsorption kinetics should be considered and might be combined into a figure of merit (173).

Generally, porous polymers might be effective adsorbents either by acting size selectively, that is allowing CO2 to pass easily, whereas N2 cannot pass straightaway or by presenting a strong CO2 affinity, which might be chemically induced, for example by the presence of amine or sulfonate ammonium groups (174, 175), although this topic is still somewhat under debate (176). Chemical interactions should however manifest in isosteric heats of adsorption, significantly higher than ∼35 kJ mol−1 (such values can be easily reached in small pores; cf the section CO2 Adsorption for Micropore Analysis). Chemically enhanced CO2 separation can be envisaged as promising, as it can combine high selectivity with high capacity. This is typically not possible for size-selective polymer networks, which do typically show a low capacity, while still possessing high selectivities. The low capacities do arise as a consequence of the amorphous nature of the polymers, which cannot allow the targeted synthesis of large pore volumes together with uniformly small entrances (though this would be possible with crystalline materials). Recent important work on chemically enhanced adsorption was reported by Zhou and co-workers (175), who also presented breakthrough studies on real gas mixtures, which do obviously present the necessary next step (see Fig. 16). Size selectivity (also including breakthrough studies) was reported by Unterlass and co-workers for a microporous polyimide system as well (177).

figure

Figure 16. CO2/N2 (15:85) breakthrough curves at 40°C for the PPN-6-SO3NH4 under dry conditions. The flow rate is 10 mL min−1. Reproduced from Ref. 175 with permission of The Royal Society of Chemistry.

While the preceding section discussed CO2/N2 separation, similar arguments (in comparison to classic materials; detailed separation studies, including the analysis of mixtures) also apply for other potentially interesting gas pairs (eg, CO2/CH4, C1-C3 alkanes (178), olefin-paraffin (179), or organic vapors), which have been targeted recently. Especially, the comparison with classic materials is necessary; see, eg, Ref. 180, as those can be used at (often) lower prices, which is an important economic argument for any real use. The demands on a pilot-scale CO2 capture unit from a real coal-fired power plant have been reported recently and might be useful for economic considerations (261 kg of zeolite 13X APG were used, adsorbent prices amount to ∼ 2 to 2.5 US$/kg as estimated from Alibaba.com, October 2013) (181). Other issues that need consideration include the water coadsorption effect (182), which could however be overcome on industrial scales. Designing of hydrophobic adsorbents might however help to reduce costs significantly. Additionally, the recent trend toward robust synthetic pathways that avoid catalysis by precious metals and provide “cheap” products (17, 177, 183) might support the design of new adsorbents within next years, which can seriously challenge traditional zeolites and carbons.

Microporous polymer networks can provide outstanding performance in some adsorption-based gas separation applications, which are often a direct consequence of their chemical fine-tuning rather than their high surface area alone. In comparison with classic zeolites or activated carbons, they have however not yet entered real applications on industrial scale. It can however be expected that they might do so with as their production costs decrease and attractive application (maybe in niche technologies) arise.

Catalysis

Microporous polymers are used in a number of applications related to catalysis. The state of the knowledge was recently summarized by Nguyen and co-workers (184) and Zhang and Riduan (185). Some new developments have however been reported since then, of which we present an outline of a few of them. We will discuss advances not only in metal-containing microporous polymers but also in heterogeneous organocatalysis.

Classically, microporous polymer networks were loaded with catalytically active metal species, such as Pd nanoparticles (154, 186). Recently, this methodology has been improved, for example, by incorporation of triphenylphosphine derivatives as building blocks of the microporous host. Doing so, high catalytic activities, for example for Suzuki reactions, were achieved (187).

Another way to combine catalytically active species into microporous networks is the use of specific ligands for metals as building blocks, which ultimately provide well-defined metal complexes (188). Classically, this has been achieved using phthtalocyanine or porphyrin derivatives and the progress has been summarized in 2008 by Budd and co-workers (189). More recent research in this direction was presented by Hupp and co-workers, who presented a microporous Al-porphyrin network, which was shown to be highly active in the catalytic degradation of nerve-agent stimulants by methanolysis (190).

Various groups have also incorporated salen-type binding motifs into microporous polymers. Deng and co-workers loaded Co or Al onto such polymers, and the resulting polymers have been proven to be active in capturing of CO2 and subsequent conversion to cyclic carbonates by a reaction with epoxides (191). A similar concept was also reported at almost the same time by the group of Son (192).

Catecholates are another binding motif for metals and have been employed recently. An iron catecholate system was reported by Hock and co-workers (193). The system is active for hydrosilylation reactions of aldehydes and ketones and represents an example for a heterogeneous catalyst, which has no homogeneous counterpart. A catalytically active TaV species, which is usually unstable, could be stabilized by single-site coordination on a microporous catecholate polymer (194). The complex was analyzed in detail by NMR and X-ray absorption spectroscopy and was shown to possess a higher catalytic activity for the hydrogenation of alkenes than its closest molecular analogue. This was attributed to the higher stability in confinement, which at the same time shows the potential of microporous polymers as catalysts (supports) and might direct future research.

Finally, N-heterocyclic carbenes (NHCs) were introduced as N-containing ligands next to the well-known bipyridine-type ligands and related species. Zhang and co-workers reported microporous polymer networks based on triptycene, which contain Pd coordinated by NHCs. The polymers are active for Suzuki reactions, and evidence was given that the reaction was indeed catalyzed by the metal complex but not by any potentially leached molecules/complexes (195). Microporous Cu-NHC complexes were finally reported by Thiel and co-workers as potential catalysts for the reaction of CO2 and epoxides toward cyclic carbonates (196).

In summary, metal-containing microporous polymers have been developed as catalysts for various applications and have left simple Pd nanoparticle fixation behind. It will be interesting to see how recent progress in catalysis such as the use of iron or other nonprecious metals will find its way into heterogeneous catalysis within next years.

A lot of recent interest was also spent on organocatalysis. Microporous polymers combine indeed the highly modular toolbox of polymer chemistry (with regard to functional groups) and the necessary high specific surface area. Recent examples include the incorporation of acid (–SO3H) and/or base (–NH2) functionalities in spirobifluorene-based networks and their use in cascade-type one-pot reactions (197). The intrinsic basicity of porous polybenzimidazoles was employed in the heterogeneoues catalysis of Knoevenagel reactions, which is an archetypical proof of concept of basic catalysis (198, 199). Other examples of organocatalysis include the use of microporous networks based on quaternary phosphonium moietites, which were shown to be active metal-free catalysts for the reaction between CO2 and epoxides (200). 4-(N,N-dimethylamino)pyridine is a well-known organocatalyst for various reactions. It was incorporated in CMP-like polymers by Wang and co-workers and shown to be an active catalyst for acylation reactions (201). The catalyst was also shown to be active under continuous flow conditions for at least 540 h without any loss in conversion, highlighting thereby the advantage of heterogeneous catalysts. Photoactive CMPs (see also the section Optoelectronic Properties) were also used in continuous flow reactions as photosensitizers for the generation of singlet oxygen, which is needed for the synthesis of ascaridole by oxidation of α-terpinene (202, 203). Photoactive CMPs were also shown to be effective heterogeneous photosensitizers for the visible light induced polymerization of methyl methacrylate, avoiding leakage of the sensitizer into the polymer (61).

Finally, highly entantioselective organocatalysts have been reported by the group of Thomas and co-workers, which are based on microporous 1,1'-Bi-2-naphthol (BINOL)-based polymer networks (see Fig. 17). The catalyst gave comparable yields, enantioselectivities, and reaction speeds as compared to their homogeneous counterparts in reactions such as asymmetric hydrogenations or Friedel–Crafts alkylations (204, 205).

figure

Figure 17. Chemical strategy toward chiral, catalytically active microporous networks. Reproduced with permission from Ref. 204. Copyright © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany.

These examples clearly show the versatility of porous polymers in a number of important reactions and also highlight a few possibilities, which cannot be easily offered by homogeneous catalysis so far. We remind the reader that (as for the other potential applications) this is the current state of the technology, which is subject to intense R&D activities.

Optoelectronic Properties

Most of the CMPs reported so far exhibited is due to the extended conjugation significant fluorescence (54, 55, 153), but fluorescence has also been reported and analyzed for PIM-1 (27, 206). This has, in combination with the large internal surface areas, created some interest from the scientific community. Most research was spent on CMP-like materials, though.

Synthetically, the tuning of band-gaps and emission colors was of interest and this problem was solved using a variety of methods. Dedicated monomers (eg, pyrene derivatives, triphenyl amine derivatives) could be used to obtain microporous polymers of a broad range of emission colors (202, 207-211). Other approaches include the use of copolymerization techniques (209, 212) and the synthesis of core–shell structures that provide an elegant way of color tuning from blue-violet to yellow emission; see Figure 18 (213).

figure

Figure 18. Photographs of core–shell CMPs of different architecture and composition under a hand-held UV light. (a) CMPs dispersed in THF; (b) CMPs dispersed in different solvents; (c) solid samples of the CMPs. Reproduced from Ref. 213 with permission of The Royal Society of Chemistry.

Potential applications of highly fluorescent microporous polymers have been outlined by Patra and Scherf (214), and initial reports on the use of these for sensing, bioimaging, etc. are already available. The energy transfer to dyes incorporated within the micropores was analyzed by various groups (67, 212, 215), and evidence for highly efficient and fast energy transfer was provided. Detailed photophysical studies revealed a cooperative process, which led to the description of the polymer as antenna system (215).

The energy-transfer properties can be used for sensing applications by fluorescence quenching or enhancement depending whether electron-rich or electron-poor arenes are adsorbed from the vapor phase into an carbazole-based microporous polymer (216). This allows the discrimination between nitrated aromatic compounds (potential explosives) and common arenes such as toluene. A comparative study by Novotney and Dichtel also showed the potential of CMPs for the detection of TNT from the vapor phase (217).

Finally, first bioimaging applications have been reported as well (69), showing that the application portfolio of fluorescent microporous polymers is still increasing and a number of more fascinating papers are expected to be published within the next few years. The analysis of significant differences of networks to linear conjugated polymers as a consequence of ring and strain effects are meanwhile under investigation as well (218).

The use of (micro)porous polymers as semiconductors as well as the use in the photocatalytic hydrogen production has also been largely analyzed; we refer the reader to Refs. (219-221) for more details on this topic.

Energy-Related Applications

New technologies are required for a sustainable energy concept, which can help to tackle or overcome some of the problems associated with the use of fossil fuels as energy sources. New materials are explored for the application in new battery types (Li-sulfur based) or in supercapacitors. Microporous polymers received some interest in this respect, and various reports on the use of microporous polymers in energy-related applications have appeared within the past few years (2011–2013).

An aza-fused CMP was suggested as a material for supercapacitive energy storage (222) and showed high energy and power densities along with good cycling stability in first laboratory tests. Composites of CMPs and iron or cobalt oxides were suggested as anode materials in lithium ion batteries (223, 224). An enhanced stability in cycling tests compared to plain nano-sized Co3O4 was attributed to some buffering effect of the CMP component.

An interesting study by Sakaushi and co-workers showed that microporous CTF derivatives, that is high-temperature stable aromatic C/N networks, can operate using p- and n-doping mechanism as a cathode material in lithium ion batteries due to its bipolarity; see Figure 19 (225), which open new possibilities. The use of this material in sodium-based batteries was also suggested (226).

figure

Figure 19. Electrical energy storage mechanism of amorphous CTF materials. Reproduced with permission from Ref. 225. Copyright © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany.

Most recent developments also target the immobilization of sulfur within microporous polymers for the use in Li-S batteries, which do often suffer from sulfur leakage into the electrolyte (227).

Again, this field is premature at the current stage (although basic principles are settled already) and might develop into something much larger within the next years as well.

Environmental Applications and Separation Technology

High surface area materials have found application in separation technology and environmental remediation/protection technologies since they are known, and the use of microporous polymers is hence a logical step. Typical uses mainly include the adsorption of organic molecules/contaminants from water or air as well as the adsorption of other low-molecular compounds such as iodine (relevant for treatment of nuclear waste; see, eg, Ref. 70) or heavy metal ions.

Hyper-cross-linked polymers have already been commercialized as mentioned earlier and used, for example for the adsorption of organic contaminants from (waste)water or the use thereof for purification purposes. The potential of commercial products for the adsorption of dye molecules was analyzed, and good performance (high capacity, fast kinetics) was reported, with a special emphasis on the possibility to achieve full regeneration by rinsing (228). This is different to most activated carbons, which can be used for adsorption but do often need thermal regeneration (burn-off) as the bulky dye molecules do not desorb easily. Pilot plants using this technology to purify wastewater streams have been reported (228). Dye removal by porous polymers has been reported a number of times; for exemple, CTFs have been proven to be very effective adsorbents for a variety of dyes due to their combination of extremely high surface areas with pore sizes ranging from micro- to mesopores (229). CTFs can also be combined with magnetic nanoparticles easily and be used for magnetic separations dye molecules from wastewater streams (230). Small organic solutes such as phenols or naphthalenes are other common (toxic) contaminants of wastewater streams in the chemical industry, which need to be removed effectively. Again, hyper-cross-linked polymer resins have been proven to be effective adsorbents, which had higher capacities than common macroporous resins (231, 232). Detailed adsorption thermodynamics and initial breakthrough studies have been performed as well. Resorcinol-modification of hyper-cross-linked polystyrene based polymers was used to enhance the adsorption properties toward p-hydroxybenzaldehyde, which is another common contaminant and base chemical with a rather high polarity and hydrophilicity (233). Finally, the use of common hyper-cross-linked polystyrene resins for the purification of plant extracts was presented by Huang and co-workers (234). Breakthrough curves of binary mixtures of berberine hydrochloride and 2-napthol were presented, and the effective purification and enrichment of the alkaloid was shown. Initial studies indicate that PIM-based microporous networks also show some potential for the adsorption of organic solutes (235, 236).

The use of hyper-cross-linked polymers based on polystyrene in chromatographic applications was also analyzed in great detail and summarized recently (237). Applications ranged from analytical separations in HPLC (238), to screening of the usefulness of the materials for their use in clinical blood cleaning (hemodialysis) (239), and the separation of inorganic salts and acids (eg, NaCl and HCl) on nominally neutral adsorbents (237). This effect was analyzed in more detail recently, and size-exclusion effects were suggested to be effective (240-242). Such studies can indeed of interest for practical applications for the separation of electrolytes. Davankov and co-workers also reported a reasonable strong tendency of hyper-cross-linked polymers for the adsorption of heavy metal ions and attributed this to the interaction between ions and the π-system of exposed benzene rings (237), also known as cation–π interaction (243). A further increase in the adsorption capacity can be achieved by sulfonation, which adds in some ion-exchange properties, which finally yield materials for environmental technologies, thus relating to the beginning of this section (244). Hyper-cross-linked sulfur containing polymers based on melamine has been prepared and screened for mercury removal. A reasonable high selectivity was observed, which was attributed to the high sulfur content, in agreement with reports on thiol functionalized metal-organic frameworks (245, 246). Such procedures receive more interest in recent years, especially as the use of mercury is banned by international protocols (247, 248).

Next to single-phase liquid separations, interest is also increasing for the treatment of oil spills or the adsorption of organic vapors from process air. Microporous polymers often posses a very high hydrophobicity, which allows efficient adsorption of organics. Oil-spill cleanup has been suggested by a number of researchers recently (249, 250), and macroscopic sponges could successfully be modified with hydrophobic, microporous CMPs to give high-performance materials (see Fig. 20) (250).

figure

Figure 20. (a) Camera image of the HCMP-1 treated sponge. (b) Optical microscopic image of the HCMP-1-treated sponge, showing the HCMP-1 microgel particles coated on the surface of sponge backbone. (c) When a piece of the sponge and the HCMP-1-treated sponge were placed in a water bath, the sponge absorbed water and sank to below the surface level, whereas the HCMP-1-treated sponge was floating on the water surface. (d) When the HCMP-1-treated sponge was pressed below the water surface under pressure, (e) it immediately floats on the surface of water after release of the pressure and no water uptake was observed during the procedure. (f–h) Snapshots showing the adsorption of a 12-cm2 red-colored octane film (dyed with red oil o) distributed on a water bath by a piece of the HCMP-1-treated sponge (0.6 × 0.6 × 0.6 cm). Reproduced from Ref. 250 with permission of The Royal Society of Chemistry.

The adsorption properties of organic vapors by microporous polymers has also been reported, and good performance has been reported generally (251-255). A reasonable good selectivity toward aromatic compounds over aliphatic compounds (eg, benzene vs cyclohexane) was reported (255), which was attributed to favorable π–π interactions between benzene and the aromatic frameworks. Generally, most microporous polymer networks are reasonably hydrophobic as a consequence of the mostly employed aromatic building blocks. The hydrophobicity was evidenced by water vapor physisorption experiments; see, eg, Refs. (252, 254), and (256). The field of organic vapor adsorption by porous polymers should take issues such as regeneration (257), dynamic adsorption behavior (breakthrough analysis) (258), and selectivity as well as comparison to classic activated carbons (259), into account within the next years. Following this path, the potential of microporous polymers could be highlighted and indeed will lead to industrial applications.

Summary

  1. Top of page
  2. Introduction
  3. Synthetic Methodologies
  4. Characterization of Microporous Polymers
  5. Applications
  6. Summary
  7. Bibliography

The field of microporous polymers has seen an explosive and highly dynamic development during the past few years. New synthetic procedures have been reported in detail, which also led to certain guidelines for the use in high specific surface areas, such as very high degree of condensation together with the use of the most rigid polymer chains. Within the next few years, it can be expected that major synthetic developments need to target shape control while maintaining high porosities. This is a classic strength of polymer science and might be a key benefit against zeolites or activated carbons. The controlled functionalization for enhancing the performance in various technologies is expected to be another major research topic, which is again a unique feature of organic chemistry and hence polymer chemistry.

The characterization of microporous polymers has also been developed significantly, and the understanding of the amorphous systems has been improved a lot, but does still lack explanation and straightforward extraction of key parameters such as specific surface areas or PSD. More fundamental insights are expected from molecular modeling within the next years, which could provide explanations, but also predictions on some of the still unsolved issues. That way, a positive feedback loop with experimental physical chemistry might be established and some of the open issues (such as reliable methods for the extraction of porosity parameters e.g. specific surface areas and PSD) could hopefully be answered in near future.

Despite the fact that not all features of microporous polymers have been completely understood, a variety of potential applications ranging from gas separation to light harvesting and environmental remediation have been showcased during the past years. Again, microporous polymers can lead to unique applications, which cannot be easily imagined without stabilization of unstable catalysts, certain semiconductor, and sensing applications. Surely, the development of various applications will benefit from interdisciplinary research and interaction with physicists and process engineers and that will become inevitable.

We hope that the present chapter could provide a good overview on the state-of-the art of microporous polymers and might stimulate further research activities on this fascinating class of materials.

Bibliography

  1. Top of page
  2. Introduction
  3. Synthetic Methodologies
  4. Characterization of Microporous Polymers
  5. Applications
  6. Summary
  7. Bibliography