Nanoporous ionic organic networks: from synthesis to materials applications

The past decade has witnessed the rapid progress in synthesizing nanoporous organic networks or polymer frameworks for various potential applications. Generally speaking, functionalization of porous networks to add extra properties and enhance materials performance could be achieved either during the pore formation (thus a concurrent approach) or post-synthetic modification (a sequential approach). Nanoporous organic networks which include ion pairs in a covalent manner are of special importance and possess extreme application profiles. Within these nanoporous ionic organic networks (NIONs), here with a pore size in the range from sub-1 nm to 100 nm, we observe a synergistic coupling of the electrostatic interaction of charges, the nanoconfinement within pores and the addressable functional units in soft matter resulting in a wide variety of functions and applications, above all catalysis, energy storage and conversion, as well as environmental operations. This review aims to highlight the recent progress in this area, and seeks to raise original perspectives that will stimulate future advancements at both the fundamental and applied level.


Introduction 1.1 General
Currently, there is a general consensus that novel materials are needed to address renewable energy storage and conversion technologies and environmental remediation processes for a more sustainable development of our planet earth. [1][2][3] To improve the efficiency and meet the ever-increasing energy requirements in various systems, nanoporous materials have been regarded as one of potential candidates due to the intrinsic characteristic of open channels and large specific surface area, coupled to the possible control of the accessibility, percolation as well as optimized mass transport for various device applications. [4][5][6][7][8][9][10][11] Among the developed porous materials, polymer-based species are attractive, as the easy processing and wide variability on monomers enabled advanced engineering within rather simple approaches. Typical features of those systems are high specific surface area, diverse pore dimensions, the use of lightweight elements only, strong covalent linkages, as well as addressable chemical functions. [12][13][14][15] Nanoporous materials refer to a class of porous materials having pore diameters from 100 nm down to around 1 nm. According to the international union of pure and applied chemistry (IUPAC) definition, micro-/meso-/macropores cover a pore size range of <2nm, 2-50 and >50 nm, respectively. Though "nanoporous materials" is not a strict term defined by the IUPAC nomenclature, it has been widely employed in the field of materials science and is used here accordingly. Traditional porous polymeric materials, derived from classical polymeric chemistry have a pore size mainly in the mesopore and macropore range, and are only partially considered in this review (<100 nm). These classical systems are fabricated commonly by the phase separation, 16 emulsion polymerization 17 or the hard template method. 18 On the other side, the recently developed microporous organic polymers (MOPs) include in fact several subclasses, such as polymer of intrinsic microporosity (PIMs), 15 covalent organic frameworks (COFs), 19 porous aromatic frameworks (PAFs), 20 and conjugated microporous polymers (CMPs). 14 They are conventionally synthesized by bottom-up approach from stiff molecular building blocks and usually possess a very high surface area as well as ordered pore architecture in some cases, e.g., COFs.
Functionalization of such porous polymers is usually target-motivated, that is key properties are adjusted and optimized to serve a specific purpose. Broadly speaking, the functions could be classified as chemical and physical ones. Examples of the former are acidity/basicity, [21][22] the ability to coordinate, or chemical activity under certain conditions and stimuli. [23][24][25][26] Typical examples of the latter are the electrical or optical properties. [27][28][29][30][31] Indeed, by means of versatility of organic chemistry, various functional moieties can be incorporated by direct synthesis from functional monomers or by the post-synthetic modifications (PSM) of the prefabricated porous skeleton. It should be noted that, for some porous matrices, the incorporation of additional chemical groups onto the pore wall will block the pores or at least restrict pore accessibility and lower pore volume and specific surface area. [32][33][34][35] This is a general phenomenon, e.g. in silica, zeolites, or metal-organic frameworks (MOFs). In this context, we emphasize that the incorporation of charged species into the porous skeleton could lower steric problems, because the dynamic ionic bonds between host skeleton and counterions add a dynamic component to host-guest interactions and the blocking effect. 36 materials. In addition to dynamic blocking or gating, as well as the ability to readjust the pore size by counterion exchange and counterion mixing, the existence of charge on the pore wall endows the pore skeleton with selective interactions with guest molecules due to the intrinsic charge repulsion/affinity effects, e.g. by Coulomb interaction, but also by ion-π-interactions or ion bridges. Electrostatic interaction therefore can play a critical role in amplifying separation and sorption efficiency, and the decorated ions in the porous networks could be chosen to increase the adsorbate-adsorbent interaction through the polarization effects/chemical bonding. 37 An enhanced separation efficiency can arise when the pore size drops below 10 nm so that the nanoconfinement effects allow a stronger coupling with the electrostatic interaction as a function of pore size. [42][43][44][45][46][47][48][49] For macroporous scaffolds, Coulombic charge plays a crucial role in maintaining appropriate hydrophilicity of scaffolds and affecting cell adhesion and recognition. For nanofiltration membranes, the surface charge as well as in many cases chemical specificity is important in controlling salt-rejection performances. Last but not the least, the incorporation of charged units into conjugated porous skeletons can promote or bias the electron/hole mobility in the skeleton due to the localization effects, [50][51][52] which leads to intriguing photoelectrochemical effects for energy applications.
The NIONs reviewed here involve the positioning of charges within the scaffold through top-down, bottom-up and post-synthetic methods. Partially or weakly charged porous networks as well as neutral skeletons with free ionic pairs or salts are out of the scope of the present review. We meanwhile try to select cases which feature how charge on the scaffold affects the properties of materials, while possibly analyzing the relationship between the charged character and pore size.

Synthesis of NIONs
The past several years have witnessed the rapid development of various synthetic strategies towards NIONs, which include the hard/soft templating method, the direct synthesis of microporous ionic organic network, the free radical polymerization approach, the ionic complexation method as well as the post-synthetic modification.
Each strategy demonstrates distinct advantages with respect to the pore structure control, processing of porous materials, or scalable production, as well as limitations from a synthetic perspective. A flexible balance of the porous structural parameters (pore shape, pore size, pore size distribution, pore distribution profile, etc.) is thus to be considered as a compromise to satisfy a specific research or application goal. Table   1 summarizes the strength and weakness of these strategies. Polymers bear tailored pore size usually in meso/macropore range; well-defined pore architecture can be obtained Control of interactions between the ionic organic precursor and copolymer template is tricky; the products lack micropores Direct synthesis of microporous ionic organic network Permanent porosity with high surface area is achievable; the pore structure/functional ionic sites can be tailored by designing monomers with targeted structures; ordered crystalline ionic network is achievable by judicious choice of building blocks and the reaction types Strict requirements on the monomer structures and synthetic routes are needed; fine design and control over meso-/macropores is unavailable; processing of materials for device applications is still a challenge Free radical Easy fabrication procedure without template, the Difficult to delicately tailor the pore polymerization product contains a broad range of pore size from micropores to meso/macropores; large scale production is possible. structure

Ionic complexation
Simple and large-scale synthetic method for meso/macroporous networks without template; pore size/structure can be controlled by tuning the species of polyelectrolytes; porous membrane synthesis is possible Not stable in highly concentrated ionic solution; pore structure lacks micropores, moderate BET surface area Post-synthesis Abundant porous network precursors are available for post-modification; porous structure can be processed with custom-designed ionic functionalities Homogeneous grafting of ionic moieties onto prefabricated skeleton is difficult in some cases; serious decrease of pore accessibility/structure distortion occurs if parent porous skeleton is instable in the post-synthesis condition 3.

Templated synthesis of NIONs
The template method is conceptionally straightforward and has been the most extensively employed for the preparation of porous networks ranging from inorganic to organic materials or their hybrids. Ideally, it is a kind of molding or casting technique for the direct replication of the inverse structure of a prefabricated template with shaped morphology. i.e. the template defines the pore. For the preparation of NIONs, compatibility between the ionic precursor, the host matrix and the template needs to be carefully considered. Aiming at a faithful template leading to a high quality of the porous polymer network, surface modifications are necessary in some cases to accommodate porous active sites and guide the growth of ionic polymers.
Besides, the template should be stable enough to sustain the conditions employed in the polymerization or polymer formation process. Last but not the least, a high degree of crosslinking is a "must" to trap the polymer structure against chain motion and pore collapse after template removal. monomer. 58 The Brunauer-Emmett-Teller (BET) specific surface area (S BET ) and average pore size were found to be 220 m 2 g −1 and 15 nm, respectively (note that though S BET is suitable for mesopores and not for micropores, it is nevertheless widely used to compare various materials.). The intrinsic affinity of the IL species to CO 2 and the stable mesopore transport structure made the material a fast and effective CO 2 sorbent.
It is worth mentioning that rational functionalization of pore walls facilitates infiltration of ionic species. One example is the copolymerization of an IL 3-benzyl-1-vinylimidazolium bromide with divinylbenzene as cross-linker in the presence of O-silylated SBA-15 (average pore size: 10.5 nm) as the hard template. 59 The surface hydroxyl groups in SBA-15 were protected by trimethylsilyl groups to assist diffusion of the hydrophobic IL monomer into the pores for nanocasting. The obtained material gave S BET and average pore size of 289 m 2 g -1 and 2.4 nm, respectively. In addition, rational adjustment of chain conformation can improve macromolecule infiltration. For example, as a weak polyelectrolyte poly(acrylic acid) (PAA) chains at a low pH or high ionic strength exhibited a coiled conformation and can infiltrate the nanoporous silica template, while they adopted an extended chain conformation at a high pH or low ionic strength, thus spatially excluded from the nanopores. 6 Currently the infiltration method has been only applied with silica and CaCO 3 template, but its principle can be extended to other inorganic ones, such as titania or alumnia.   3). 67 Their work employed LBL assembly with consecutive adsorption of cationic and anionic pillar [5]arenes. The resultant pore size was around 5 Å, which is the inherent cavity of pillar [5]arene molecules. The films featuring active pores allowed for shape-selective uptake of dinitrobenzene isomers: the film adsorbed para-dinitrobenzene but rejected ortho-and meta-dinitrobenzene. This selectivity was also bound to the surface electrostatic potential: para-dinitrobenzene was adsorbed into the films with a positive surface, but not the negative one.  68 The holes with an average size of 16 nm could be generated owing to the osmotic pressure difference in the core dissolution process.

Soft template synthesis of NIONs
The soft template method is equally important in the synthesis of porous materials. The method is based on the use of soft matters, majorly self-assembled block copolymers, as template. The phase separation occurs in the self-assembly process since the thermodynamic incompatibility of compositional heterogeneous segments prefers to minimize contact energy.
Because of segment chain connectivity, the separation is restricted to the nanometer scale to form ordered structural patterns. Generally, the resultant materials are dominantly meso-and macroporous, that is, the pore size is well above that of a single polymer chain. Especially materials with well-defined ordered mesopore structures were obtained on the basis of this method.
Although soft templating has been well explored for the synthesis of porous inorganic materials or neutral organic materials or their hybrids, 69-71 the exploration of this method towards ionic polymer frameworks is seldomly reported. Matching the electrostatic charge-density is important for enthalpic reasons. The interaction between the ionic organic precursor and copolymer template on one side must be strong enough to spatially guide the polymerization around the template, and on the other side should be weak enough that the template could be fully removed to reopen the porous structure.
Meanwhile, entropic interactions as a key player here restrict polymer chain conformations in a spatially confined reaction field.
Wang and co-authors reported an interesting example to synthesize NIONs by balancing the interaction between the soft template and the ionic polymers.
In their work, a hierarchical meso-/macroporous PIL monolith with tuneable pore structure was synthesized through free radical polymerization of an IL (1-allyl-3-vinylimidazolium chloride) by using the triblock copolymer P123 (EO 20 PO 70 EO 20 ) as the soft structure-directing template. 72 The dissolved P123 and the IL effectively interacted with each other through the S 0 H + X -I + mode (S 0 : nonionic surfactant P123, H + : hydrogen ions of ionization, X: IL anion, I + : IL cation), in which the H + concentration increases after introduction of the initiator ammonium persulfate. Protons were necessary to realize the interaction between the ionic polymer and the soft template P123, while the template can be removed by solvent extraction of the as-synthesized material in ethanol to leave behind the pores. The product possessed a S BET and average pore size of 143 m 2 g -1 and 28 nm, respectively. Similarly, Xiao and co-workers synthesized mesoporous sulfonated melamine-formaldehyde resin assisted by copolymer surfactant F127 as template. 73 After removal of the copolymer surfactant by ethanol extraction, the NION exhibited uniform mesopores with a S BET and average pore size of 256 m 2 g -1 and 10.2 nm, respectively.

Template-free synthesis of NIONs
From a synthetic point of view the template method is popular and straightforward, the process nevertheless inevitably involves the synthesis and removal of sacrificial components, which is time-/energy-consuming and in some cases comes with non-sustainable steps. Besides, the template method is powerful to produce meso-and macroporous NIONs, but significantly restricted when microporous materials are concerned. The introduction of both porosity and charge into organic "soft" materials without templates imposes strict demands on the resulting framework, as polymer chains normally pack space-efficiently to maximize intermolecular interactions, especially when electrostatic interactions between ionic species in the skeletons are involved. As a rule of thumb, the resulting polymers should be both very stiff and badly packing to enable such spontaneous porosity. Fortunately, diverse available functional groups and refined covalent-bond forming reactions in organic synthesis provide opportunities to tailor the pore structures of NIONs under such template-free conditions. Table 2 summarizes all the NIONs discussed in the following section prepared via the template-free approach. Abbreviations in Table 1

Synthesis of microporous ionic organic networks
In contrast to conventional synthesis of meso-and macroporous polymers, where long backbones formed by linear connection of polymerized monomers are interconnected by ditopic crosslinkers, microporous organic networks are typically constructed from monomer units that are multitopic (three or more connection points). It is known that smaller pores experience higher capillary pressure and higher surface energy, therefore the pores are instable and prone to collapse by bending and twisting of polymer chains to pack space efficiently. 6 In this regard, rigid and contorted polymer chains or high-degree connectivity/crosslinking enables polymers to maintain micropores and preclude/minimize swelling via adsorbing vapor or solvent molecules, which is advantageous over classically crosslinked polymers that are prone to pore swelling in contact with gas or solvent. However, it should not exaggerate the role of microporosity from an application viewpoint although it is a family of promising material that gains increased attention. Some excellent reviews on the scope of non-charged microporous polymers have been published, which are deserved to be consulted by readers who are interested in this area. 4,12,14,15,[74][75][76] Compared with neutral microporous networks, the synthesis of microporous ionic organic networks faces more challenges since the presence of ionic groups restricts the choice of solvents and available coupling chemistry. To achieve permanent porosity in microporous ionic organic networks, a crucial point is the choice of comparatively rigid monomers that, when crosslinked, yield pores with similarly rigid walls. Therefore, a rigid conjugated skeleton with reactive sites is a classic starting point. Similar to conventional neutral microporous materials, the principle for the synthesis of microporous ionic organic networks has drawn from an enormous number of modern bond-forming methodologies (e.g., molten salt induced polymerization, 77,78 metal-catalyzed cross-coupling reactions, 79 hypercrosslinking reactions, 80 polycondensation reactions, 81 etc.) to yield a wide range of structural frameworks. The advantage of slowed-down bond-forming reactions is in favor of pores that more closely match the dimensions of potential guest molecules. 76 It should be noted that due to the difficulty to delicately control the pore structure, a small fraction of mesopores frequently accompany the prevailing micropores in the final porous product. Commonly the as-synthesized materials could be either amorphous or in an ordered crystalline state (e.g., COFs), depending on the building blocks and the reaction mechanism. The geometry of the rigid building unit is highly relevant. Some typical building units of different configurations (such as planar triangle, octahedron, linear pattern, tetrahedron, and square, etc.) and symmetries (C2, C3, C4, and C6) and length developed for neutral microporous network formation are suitable as well for the construction of microporous ionic organic networks. Also for amorphous microporous ionic organic networks, although they are irregularly structured on larger scales, the local skeleton and nanoscale porosity of the materials can be still controlled by the initial monomers. In other words, the geometries of building units play vital roles in governing the structure of random networks. According to recent reports, microporous ionic organic networks could be synthesized from ionic monomers or through ionization reactions between neutral monomers during network formation. The two strategies will be discussed in detail next.

Microporous ionic organic networks from ionic building unit(s)
The synthesis of microporous ionic organic networks straightforward from the ionic monomers experienced rapid advance in the past few years. Many kinds of bond-formation reactions such as Sonogashira-Hagihara cross-coupling reaction,  4 ] with 1,3,5-triethynylbenzene via Sonogashira-Hagihara coupling yielded a microporous ionic organic network, and the chemical structure of the ionic material was proven by 13 C, 11 B, and 7 Li NMR spectroscopy. 40 The material showed a permanent porosity with S BET of 890 m 2 g -1 and a pore volume of 0.61 cm 3 g -1 . In comparison, the use of  incorporated phosphonium salts via Friedel-Crafts reaction (Fig. 4). 85 Benzene was used as the co-monomer for the synthesis to avoid the limitation of Friedel-Crafts reaction for electro-deficient aromatic units. These porous materials have high S BET up to 1168 m 2 g -1 . One of the prototypical approaches toward synthesis of quaternary ammonium containing microporous ionic organic networks was based on polycondensation reactions. By appending a nitrile functionality onto the imidazolium unit, Dai and co-workers successfully developed a series of microporous ionic organic networks based on cation-crosslinkable ionic liquids. 86 The key structural feature of the IL for synthesis of porous polymers is the presence of the functional nitrile groups that can trigger crosslinking reactions at elevated temperatures, which underwent cyclotrimerization reactions between nitrile groups on neighbouring IL cations to give a dynamic amorphous polytriazine networks. The negligible vapour pressure and the nitrile group of these ILs were the most essential features to obtain porous materials under ambient pressure without catalyst or external template. The resultant materials showed tuneable S BET in a large range from 2 to 814 m 2 g -1 when changing anions from Clto bis(pentafluoroethylsulfonyl)imide.
Another synthetic example of quaternary ammonium containing NIONs was reported by Son and co-workers (Fig. 5). 87 In their attempt, the tubular microporous organic network bearing imidazolium cations was prepared by Sonogashira coupling reaction of tetrakis(4-ethynylphenyl)methane and diiodoimidazolium salts. The resultant material offered S BET of 620 m 2 g -1 and pore volume of 0.36 cm 3 g -1 . Besides, microporous ionic organic networks with S BET up to 262 m 2 g -1 by incorporation of imidazolium linkers based on bifunctional aryl bromides via crosslinking with tetrafunctional boronic acids was reported by Kaskel, Glorius and co-workers. 88 In addition to imidazolium, a pyridinium unit is an equally interesting cation to be incorporated.  . 92 Using longer alkyne-bearing building blocks, larger surface areas in an Au-NHC functionalized porous network with S BET up to 798 m 2 g -1 were reported. 93 The combination of NHCmetal complex with hypercrosslinked polymer is an effective approach to prepare high surface area ionic network. For example, a Pd(II)-coordinated polyNHC hypercrosslinked network carried high surface area with S BET of 1229 Besides, some other anionic units such as carboxylate 95 and phosphate 96,97 groups have been incorporated into polymeric networks as well.

Microporous ionic organic networks from non-ionic building units
In addition to the direct use of ionic monomers, microporous ionic organic networks could also be synthesized via bond formation induced ionization from non-ionic monomers. One case is based on the nucleophilic substitution polycondensation. Chen This material gave a moderate surface area with S BET of 532 m 2 g −1 . Another approach to microporous ionic organic networks is via the formation of a spiroborate linkage, 101 a kind of ionic derivative of boronic acid that can be formed readily through the condensation of polyols with alkali tetraborate, or boric acid, or through the transesterification between borate and polyols in a thermodynamic equilibrium. One example of ionic COFs by introduction of spiroborate linkage into the network exhibited a high surface area with S BET up to 1259 m 2 g −1 (Fig. 7). 50 It is stable in water as no obvious decrease of pore surface area was observed after 2 days of storage in water.

Direct free radical polymerization
Apart from polycondensation, the direct polymerization based on other mechanisms such as free radical polymerization is able to yield NIONs. In a typical procedure, a homogeneous mixture solution containing monomer, cross-linker, porogen molecules, and initiator is in situ polymerized to induce the phase separation between the cross-linked polymer chains and the porogenic solvent. 6 To prepare a porous network, the porogenic solvent employed here should be a relatively good one for the monomers, but a thermodynamically poor one for the resulting polymer networks. ILs (e.g. a monomer prepared from 1-vinylimidazole and 1,4-dibromobutane) in the presence of the analogous nonpolymerizable IL 1-butyl-3-methylimidazolium bis(trifluoromethane sulfonyl)imide (Fig. 8). 102 In this case, the IL featuring a similar chemical nature as the PILs was regarded as an ideal solvent to obtain materials with small pores, as the IL and the in situ formed PILs were uniformly mixed rather than phase separated from each other during the polymerization. The IL used as porogenic solvent can be extracted easily after polymerization for reuse. Different monomer/porogen weight ratios (M/P = 1:0, 1:0.5, 1:1, 1:2, 1:3, 1:5, and 1:10) were systematically adopted to investigate their effect on the formation of the porous PILs.
The optimized material (M/P = 1:1) produced uniform mesopores with pore size ranging from 6.8 to 8.5 nm and a S BET of 86.21 m 2 g -1 .
Another example of NIONs with shaped macroscopic morphology could be prepared through the free radical homopolymerization of a series of rigid bis-vinylimidazolium salt monomers. 103 The resultant meso-/macroporous monoliths exhibited S BET and pore volume up to 224 m 2 g -1 and 0.57 cm 3 g -1 , respectively.  It is notable that NIONs obtained by free radical polymerization method in some cases give relatively low surface areas in comparison with other methods. This is perhaps attributed to the difficulty in controlling the radical polymerization kinetics, leading to heterogeneities and skin layers. To improve this, the combination of radical copolymerization with other polymerization methods can be applied to prepare NIONs with larger surface area. 107  Through modification of the fabrication procedure mentioned above, the products may include also nanoporous PIL membranes bearing a unique gradient of the degree of electrostatic complexation along the membrane cross-section. 43,[115][116][117][118] Commonly, the membrane fabrication was conducted according to the following procedure. 115 The PCMVImTf 2 N and PAA mixture in DMF solvent was cast onto a glass plate and dried to produce a yellowish sticky thin polymer blend film. Then the film on the glass was immersed in an aqueous NH 3 or NaOH solution to induce in situ ionic complexation between PAA and the surrounding PCMVImTf 2 N chains to build up the electrostatically crosslinked porous network (Fig. 11). Interestingly, the different gradient of crosslinking density can produce membranes containing different kinds of pores, and the resultant structure depended on the kind of organic acids. The chemical structure, such as the anion type and backbone architecture, of PILs can also modulate the nanopore size.

NIONs by post-synthesis
Besides the direct synthesis with or without template, post-synthetic modification  respectively. 21 The sulfonation of the samples was then carried out in a Teflon beaker placed in a Teflon-lined autoclave via a gas-solid reaction (Fig. 14).

Catalytic applications
The use of recyclable catalysts for organic synthesis to minimize waste production and optimize catalyst efficiency is one of current goals for the pursuit of greener, safer, and more environment-friendly technologies in chemical and pharmaceutical industries. [146][147][148][149][150] NIONs with charge decorated pore surface and intrinsic porosity in different ranges of sizes present new opportunities for heterogeneous catalysts. Their use in catalytic applications will be reviewed in the following subsections according to the types of functional catalytic species.

NIONs with functional organic group for catalysis
NIONs with intrinsic ionic units could be used as functional sites for catalytic application. One of important materials is the imidazolium-containing nanoporous polymers. In these catalysts, N-heterocyclic carbenes (NHCs) can be formed by in situ deprotonation of the C2 carbon in the cationic ring, thus constituting a class of very important organocatalysts. 151 The immobilization of a bifunctional imidazolium entity in a series of highly crosslinked organic framework materials was obtained by Suzuki coupling with tetrafunctional boronic acid linkers. 88 The resulting porous materials were applied as heterogeneous organocatalysts in the NHC-catalysed conjugated umpolung of ɑ, β-unsaturated cinnamaldehyde. The product yield and the stereoselectivity reach up to 86% and 59 %, respectively, which are comparable to that of the molecular analogues (85% and 65%, respectively). Given that NHCs are known to activate epoxides for the reaction with CO 2 to produce cyclic carbonates, the  Although a number of materials enabling the activation of epoxides for the reaction with CO 2 to cyclic carbonates have been reported, pursuing the high reaction activity at mild conditions, e.g., atmospheric pressure and low temperature, is still challenging. Recently, Wang and co-workers found that ionothermal-derived meso-/macroporous PILs from bis-vinylimidazolium salt precursors exhibited high activity toward efficient conversion of CO 2 at atmospheric pressure and low temperature among the reported metal-/solvent-/additive-free heterogeneous catalysts (Fig. 15). 103

NIONs with (organo)metallic group for catalysis
NIONs with an organic linker provide various opportunities for pre-synthetic or post-synthetic functionalization with metal ion as the active site for catalysis.
The advantages of these chemically well-defined moieties within an open porous framework include uniform dispersion and good accessibility, which in fact can promote the catalytic activity in hydrogenation, oxidation, and so on.
Such an advantage of a metal complex containing NION compared to the molecular derivative was proven by Lin et al. 24 In their work, NIONs with phosphorescent [Ru(bpy) 3 ] 2+ and [Ir(ppy) 2 (bpy)] + building blocks were analysed. The porous polymer-immobilized complexes were efficient catalysts for light-driven reactions, such as the aza-Henry reaction (yield up to 99%), the R-arylation of bromomalonate (yield up to 91%), and the oxyamination of aldehydes (yield up to 48%), while the yields were comparable with, in some cases even higher than those of the homogeneous analogues (Fig. 17).  Normally speaking, NIONs with transition metal complexes could be possibly transferred to metal nanoclusters in the catalytic environment due to the metal leaching into solution. Therefore, the stability of transition metal complexes should be carefully investigated before the catalytic experiments. A method of quantitative determination of the coordinated metal species is necessary to specify the catalytic sites. the presence of a hard template (silica particles ~12 nm in average size). 39 The S BET of resultant materials ranged from 107 to 132 m 2 g -1 with hierarchical meso-and macropores, apart from the micropores (Fig. 19). This kind of porous materials exhibited strong ability to confine the Au nanoparticles in an ultrasmall size (  Nanotechnology to tackle environmental issues plays a key role in enabling novel processes in environmental engineering and science, e.g. cost-effective technologies/materials for catalytic degradation, adsorptive removal and detection of contaminants. 163 As one of the key environmental issues of current research, a porous sorbent that can selectively capture CO 2 from a flue gas mixture to treat anthropogenic CO 2 emissions can be identified. [164][165][166] As compared to numerous kinds of porous materials that have been employed for CO 2 capture, NIONs with charge modified pore surface show a high selectivity toward CO 2 . 167 Ionic pore walls can contribute to the build-up of Coulombic fields required for the polarization and polarized binding of polar molecules. 168 This is especially true when the pore size drops below 10 nm so that nanoconfinement effects strongly couple with the electrostatic interaction. Zhu and co-workers combined Materials Studio (MS) with grand canonical Monte Carlo simulation and revealed that the values of binding energy followed an order of H 2 , O 2 , N 2 , CH 4 and CO 2 (small to large) on a series of quaternary pyridinium-type porous aromatic frameworks with tuneable channels. 37 This result is consistent with the observation in gas sorption experiment. The high value of CO 2 was ascribed to the distribution of a partial positive charge on the pyridinium groups, leading to a polarizing binding environment which enhances the affinity towards CO 2 owing to dipole-quadrupole interactions. 169 Besides the physical absorption process, CO 2 could also be absorbed through chemisorption processes in IL-based materials due to their strong interaction and the preferred formation of imidazolium-carboxylates, formally via a transient N-heterocyclic carbene intermediate. Especially materials with small pores and imidazolium species are predicted to be highly efficient for CO 2 capture. An interesting example of the coupling of physical and chemical absorption of CO 2 was studied in a recent report on mesoporous imidazolium-type PIL-based polyampholytes (Fig. 20). 111 These materials exhibited an intrinsic porosity with S BET of up to 260 m 2 g -1 . It was found that next to fast CO 2 adsorption to the surface of ionic network (which is the predominant mechanism in typical micro-/mesoporous organic polymers), additional volume uptake and swelling deep into the polymeric matrix occurs. This process is slow compared to the surface adsorption and comes with an energetic penalty.

Environment related applications
Moreover, CO 2 taken up as such could not be desorbed easily. Fourier transform infrared spectroscopy (FTIR) measurement provided evidence that the trapped CO 2 could partially be activated to form imidazolium-carboxylate zwitterions even at low temperature and CO 2 pressure. The mechanism was believed to be comparable with the previously reported formation of transient N-heterocyclic carbene intermediate within low-molecular ILs of the imidazolium-carboxylic anion type.
The NIONs with a more delicately modified pore surface could also be used for removal of ammonia gas from industrial air effluents which are basic in nature. For example, a NION was prepared by using PAF-1 20 with post-grafted -COOH groups, and it was tested for the selective removal of ammonia. 170 The material featuring a multiple interpenetrated structure dominated by micropores (< 6Å), exhibited an uptake of 17.7 mmol g -1 at 1 bar, which was the highest capacity in this application reported so far. Such exceptional performance could be attributed to multiple chemical interactions between the multiple acidic sites located in the host material and ammonia gas, which promoted the adsorption for pollutant capture.
Ionic exchange materials for adsorption of ionic pollutants or heavy metals are another class of relevant sorption materials, dominated by conventional ion-exchange resins. 171 Conventional ion exchange resins with large sized pores (in the range of several to tens of micrometers) face a number of drawbacks, such as inefficient accessibility of ion-exchange sites, limited kinetics and sometimes "outflow" of mobile phase under working conditions. 171 The NIONs with rigid skeleton as well as intrinsic nanosized pores already have shown applicability toward environmental pollution such as radioactive ions, 86 heavy metal ions, 95, 172 organic dyes, 110 and so on.
In this regard, Dai and co-workers showed that NIONs primarily operate by an ion exchange mode as demonstrated by imidazolium containing NIONs to capture perrhenate ions (ReO 4 -). 86 Energy-dispersive X-ray analysis (EDAX) results showed that in contact with a NaReO 4 solution, the porous network did not contain any Na + , reemphasizing that perrhenate is taken up through an ion-exchange process involving only the counteranions. Sometimes the micropore structure of the ionic networks also enhances the ability in ion capture. For example, high surface area of PAF-1 functionalized with a thiol group could work as a nano-trap for mercury and showed a record uptake capacity of mercury over 1000 mg g -1 . 172  were built up from a tetrahedral building unit, lithium tetrakis(4-iodophenyl)borate (LTIPB), and different alkyne monomers as linkers via a Sonogashira-Hagihara coupling reaction (Fig. 21). 83 The networks featured three effective sorption sites, i.e.
an ionic site, the phenyl ring, and triple bonds, and exhibited the highest reported iodine adsorption capability to date (2.71 g/g, 2.76 g/g, and 2.60 g/g of iodine for PAF-23, PAF-24, and PAF-25, respectively.). A control experiment was carried out by using neutral PAFs with similar topology for I 2 capture under the same conditions, which ended up with a much lower capacity. Membrane-based technologies for the production of pure water from seawater or brackish water are an energy-efficient, low environmental impact engineering solution.
NION-membranes are potentially a promising material for such application. 173,174 Zhang and co-workers reported a nanoporous membrane (pore size ranging from tens to hundreds of nanometers) through ionic complex of poly(acrylic acid-co-acrylonitrile)s and imidazolium-based polycations. Zeta potential measurements indicated that the obtained membranes were negatively charged at neutral pH conditions. The membranes with hierarchically structured nanopores exhibited moderate rejection to salts in the order of Na 2 SO 4 (59.5 %) > NaCl (6 %) > MgCl 2 (0.1 %), but high rejection to methyl orange (> 99.9 %). 173 Organic solvents are inevitably utilized in chemical industry, whilst quite some of them are not expected to show up near the consumer. Although various methods have been explored for solvent sensing, the exploration of smart actuating materials that are capable of adaptive motion, and/or reversible shape variation in response to solvent stimuli is an exotic option. 175,176 Currently, most reported polymer actuators toward solvent sensing applications suffered from a relatively low sensitivity for organic solvents due to the requirement of a substantial amount of secondary solvents to produce noticeable shape deformation or displacement. Our group recently reported a nanoporous (30-100 nm in pore size) PIL membrane prepared by ionic complexation between PCMVImTf 2 N and PAA, which carried a unique gradient of crosslinking density along the membrane corss-section. 116 The membrane in water readily bent upon adding as low as 0.25 mol% of acetone molecules (1 acetone per 400 water molecules). This is at least one order of magnitude more sensitive than other state-of-the-art solvent stimulus polymer actuators (SSPAs). Mechanism investigation revealed that the strong interaction between acetone and the ion pair in PCMVImTf 2 N together with the coexistent structural gradient led to a gradient absorption of acetone along the cross-section of membrane, resulting in a swelling gradient across the membrane to bend the membrane. Moreover, different from common nonporous SSPAs, the nanoporous channel not only accelerated mass transport of solvents into the membrane, but also weakened the overall bending rigidity by introducing a gas subphase.

Energy storage and conversion
As a new type of porous polymer, NIONs have been explored for various energy-related fields, e.g., carriers for enhanced gas storage, electrode materials for batteries, supercapacitors, and fuel cell membrane, over the past few years. In this section, we intend to catch this emerging field, illustrating the unique role of charge containing units combined with pores of different sizes on the performance of these materials. and resulted in stronger binding force. Similar trends have been observed as well in ion-functionalized nanoporous membrane system. 137 An enhanced performance could be realized both by increasing S BET as well as ionic density in the polymer network.

NIONs for photoelectrochemical energy storage and conversion
Porous organic networks decorated with charge units will lead to high mobility or biased transport of the electron/hole in the skeleton, especially when π-conjugated networks are considered. This kind of materials are photo/electrochemically active, [50][51][52]90 and show intriguing potential in certain energy related applications.
A tutorial and conceptual case of this kind of materials is ionic COFs with a highly ordered 2D open network structure, which has the capability to predictably organize redox-active groups. This fact makes them potential candidates for charge storage devices. For example, Dichtel and co-workers reported a β-ketonamine-linked 2D COF containing redox-active anthraquinone building unit. 179 The material exhibited a moderate capacitance of 48 ± 10 F g -1 at a current density of 0.1 A g -1 (1 M H 2 SO 4 ), which after 5000 cycles only dropped to 40 ± 9 F g -1 . It should be mentioned that the chemical and oxidative stability of COF linkages are mandatory for the direct use in electrochemical devices, and thus a careful design is advised.
Recently, Jiang and co-workers reported a facile and general strategy that converts a conventional COF into redox-active platform by delicate post-synthetic channel-wall functionalization with organic radicals (Fig. 24). 180 In their work, the conventional imine-linked COF ([HC≡C] X% -NiP-COF; X = 0, 50, and 100) as a scaffold with nickel porphyrin at the vertices and ethynyl unit on the channel wall was synthesized.
The ethynyl groups of this structure were then "clicked" to 4-azido-2,2,6,6-tetramethyl-1-piperidinyloxy in a smooth and clean manner to yield NION-based materials with enhanced ion conduction could also be important in fuel cell applications. 131,185 For example, mesoporous poly(benzimidazole) (PBI) membranes with pore sizes of ~10 nm were obtained by hard templating method. 131 The post-grafted phosphoric acid groups to yield a highly proton conducting material at zero humidity could be easily operated up to 180 o C. Moreover, the proton conductivity was one to two orders of magnitude higher than that of a non-porous PBI/H 3 PO 4 complex, as tested under similar conditions. Such excellent performance was related to the nanopores of PBI/H 3 PO 4 membrane. The interconnected pores provide the proton conduction highways, and meanwhile excessive membrane swelling was prohibited by a high crosslinking density of the membrane, thus the membrane was also mechanically stable. Another example of enhanced conductivity in a charged porous polymer network was given by using graphitic carbon nitride (g-C 3 N 4 ) 27 as a polymer precursor through post protonation with HCl. The resultant protonated material (g-C 3 N 4 -H + Cl -) showed at least 10 times enhancement in ionic conductivity as compared with parent g-C 3 N 4 (Fig. 23) .

Fig 23.
Nyquist impedance plots (scatters) for g-C 3 N 4 and g-C 3 N 4 -H + Cland simulation (lines). The frequency range is from 106 to 103 Hz, and the perturbation signal is 100 mV. Inset: dotted area in high magnification and equivalent circuit mold. The calculated resistances (R b ) before and after protonation were ca. 28  anion/cation-conducting behavior. 188 The suppressed proton conductivity in the neutral state was speculatively attributed to isolated amine sites and exceptional chain rigidity of the polymer, which was responsible for the generation of an ionic diode.
Electrode interlayer is a key structure between active layers and conducting electrodes that controls the transport of charge carriers in and out of cells. Jiang and co-workers employed a microporous network bearing polyborane carbazole unit as a new type of electrode interlayer. 189 They have found that the neutral network based thin-film exhibited extremely low work-function-selective electron flow; while upon ionic ligation and electro-oxidation, the charged network significantly increased the work function and turned into a hole conductor. Moreover, these charged thin films were compatible with various electrodes and offered outstanding functions in various types of devices, including solar cells and light-emitting diodes.
Salinity difference between sea water and river water creates an exploitable salinity gradient energy, so-called "blue energy", a clean energy source popularly discussed in the current energy crisis. 190 In principle, more than two terawatts of electricity can be potentially generated in the river estuary where the rivers flow into the sea. To capture this energy more efficiently, numerous efforts have been made to create mew materials enabling this transformation. Porous materials with high surface charge density and pore-sizes down to sub-10 nm are favorably discussed in this context. 190,191 Recently, by integrating a porous block copolymer membrane and precursor for carbonization (Fig. 25). 197   reported the use of non-noble metals incorporated COFs as templates and precursors for producing metal/carbon catalyst. 198 The resultant materials with rather uniform metal/nitrogen distribution showed efficient electrocatalytic activities toward 4 electron oxygen reduction reaction in both alkaline and acid media with an excellent stability as well as being free from methanol-crossover/CO-poisoning drawbacks.

Conclusion and perspectives
NIONs are emerging unique class of porous polymer networks that inherently combine high-density packing of ion pairs with porosity. The abundant synthetic routes and availability of building blocks allow to generate NIONs with rich structural diversity and functions. The rapid growth of this field in the past few years focuses on the design, synthesis and functional exploration of these materials. Especially, when charged character and pores in nanometer range work synergistically, intriguing applications in catalysis, energy storage and conversion, as well as environmental applications are obtained, outperforming the behavior of conventional nanoporous network with neutral skeleton.
As for the design and synthesis of porous structure, the porosity, pore environment and functionality can be tuned. Nevertheless, the reported values of S BET are currently still limited (< 1600 m 2 g -1 ) especially in comparison with that of MOFs and carbons. Enhanced surface area is therefore one of the future goals. Besides, delicate control of the pore structure (size and shape) and systematically tuning the synergistic effect of pore size and electrostatic interaction at a molecular scale are not fully understood and deserve more in-depth investigations.
Already now, NIONs by their size or shape selectivity, enhanced mass transport, and special pore environments, enable outstanding activities in heterogeneous catalysis. We believe that the combination of intrinsic ionic characteristic with additional functional groups/sites will offer new opportunities to extend catalytic applications of NIONs, for instance in tandem catalysis and for reaction cascades. It should be mentioned that the most of reported NIONs to date are amorphous structures: the pore structure could not be fully controlled due to the disorder-generating processes of polymerization and crosslinking. Such characteristics lead to a random distribution of catalytic sites in the porous network, which in turn makes the evaluation of catalytic sites and the control of the catalytic performance difficult. Recently, ionic COFs with ordered pore systems might generate new insight in this area, whilst their application in catalysis still needs to be expanded and evaluated. In addition to delicate choice of the raw materials and protocol of polymerization, seeking for appropriate conditions to facilitate weak interactions, such as π-π (cation-π, anion-π) stacking and hydrogen binding, is believed to be helpful to generate the materials with higher/improved pore control. As a final step, one might envision the synthesis of a material with densely packed "artificial enzyme pockets", with all secondary interactions being delicately placed to enable maximum reactivity and/or selectivity.
NIONs have found to be excellent candidates toward environmental related applications in terms of high absorption capacity and kinetics with regard to neutral porous materials. The enhanced performance with fast, efficient capture of pollutants, high selectivity and high recyclability of the absorbents has to be further explored. It is believed that the principle of designed materials with a monoatomic layer thick pore wall, open channels to avoid entrainment and a dense packing of ion exchange sites contributing to a high ion exchange capacity will inspire following research works. 199 In addition, to achieve meaningful adsorption capacity at lower concentrations or high selectivity, a high enthalpy of adsorption (ΔH ads ) is required.
The demanded enthalpies typically lie well beyond pure physical adsorption processes and will instead involve materials that interact multisite-physically or chemically with the analyte of interest. A few examples have already hinted to the benefits of ionic group triggered coordination/chemically bound guest species, whilst proper control of binding energy (high enough to boost selectivity, but low enough for a possible desorption process) is the key for optimization and recyclability of the absorbents without excessive energy demands. This could turn into a leading descriptor for customization design of NIONs.
NIONs based energy applications is indeed a burgeoning field in materials science, and the already proven practical applications are concerned with their structure stability with regard to mechanical load or chemically harsh conditions (e.g., strong acidic or basic conditions, high electrochemical stress). Although a few NIONs with superior stability have been explored, effective methods to address this issue remain an open race.
A parallel challenge in this area is concerned with the conductivity of framework.
Excellent exciton migration and charge carrier transport is necessary for many energy processes. Methodologies and molecular structures that can enhance the charge carrier mobility and electric conductivity are known from the development of OPV systems, but gain here a "third dimension", as packing and neighboring effects become addressable and are highly relevant. In this regard, systematic investigations are demanded to clarify the structure-property relationship of "3D-heterogeneous organic conductors". But even on a more primary level, the development of conjugated crosslinking porous network with built-in high density of ionic pairs is believed to bring new inspiration into materials chemistry of this area. Besides, device integration via NIONs designed for good processability is also of great importance. We envision that some novel techniques, such as ink-jet printing, wet-lithography or roll-to-roll processes, might be implemented to achieve these targets of integration and production.
Another emerging trend is downsizing the bulk NIONs to nanodimension, which holds great promise to extend the potential applications. Very recently, an intriguing example was described: guanidinium halide containing ionic network was self-exfoliated into ionic covalent organic nanosheets. 200 Intrinsic charges in the framework backbone and sandwiched anions in between the layers were believed to be responsive for such self-exfoliation process. Compared with conventional on-surface synthesis procedure for neutral COF-derived nanosheets, such method toward spontaneously generated nanosheets highlights the advantage of charge species which enable time-/energy-efficient procedures. The downsized ICONs with reduced dimensionality and well-defined in-plane charge transport may provide unique photoelectrochemical properties.
Currently, the investigation of NIONs is still in its infancy, and only materials of a limited scope have been designed and explored. Considering the versatility of ionic monomers as well as synthetic methods, there is however so much space to produce intriguing structures with improved physico-chemical properties. As a unique type of functional porous organic material with outstanding performances, NIONs will continue to draw interest and enquiry from both academia and industry. It will be exciting to witness the rapid development of this new field in the years to come.