Since their emergence, porous copolymer resins (PCR) have been used in various applications including sorption [1,2,3], solid supports in catalysis [4,5], chemical separations [6], ion-exchange [7,8], chromatography [7,8] and other applications. The usage of PCR materials is related to the unique physicochemical properties and the porous structural features of the polymer resins. Some examples of pore design include foaming (use of gaseous porogens), phase separation (use of solvent porogens), emulsions (high internal phase emulsions), and template synthesis (molecular imprinting) [9].

This review will focus on the phase separation technique where vinyl and acrylic PCR-based materials are synthesized from acrylates, styrene, divinylbenzene, and vinyl pyridine via suspension polymerization, in the presence of single and/or multi-component solvent porogens [2,4,6,7,10,11,12,13]. Suspension polymerization [14] is a technique where monomer(s), relatively insoluble in water is (are) dispersed as liquid droplets with a steric stabilizer that acts to hinder the coalescence of monomer droplets and the adhesion of partially polymerized particles during polymerization, using vigorous stirring to produce polymer particles as a dispersed phase. A porogen is either a low molecular weight compound [1,2,4,5,7,11,12,13,14] or a polymer [15,16,17] that is miscible with the monomers. The porogen is inert and can be readily removed, resulting in voids after the reaction with the formation of porous polymeric materials. The term “pore-forming agents” [9,10,14,15,16,17,18,19,20,21,22,23] is sometimes used instead of “porogens”.

The general mechanism of how porogens control the resin surface area, pore-size, pore-size distribution (PSD), pore-volume, and morphology is illustrated in Figure 1 [18]. The period of phase separation, shown in Figure 1, is the main rate determining step. As a general rule, phase separation will occur at a much later stage if the porogen is a “good” solvent or a solvating diluent [24], i.e., miscible with both the monomer and its resulting polymer. On the other hand, the “bad” solvent or non-solvating diluent has poor miscibility at the polymer level. If a non-solvating diluent is used (A in Figure 1), microgels [2,18] will form, fuse together, and aggregate into larger clusters. This results in an increased volume of the voids (pores) forming larger pores (i.e., macropores) and reduced polymer surface area. Unlike non-solvating diluents, a solvating diluent (B in Figure 1) delays phase separation; hence, ensuring that microgels retain their individuality even though they undergo fusion. This leads to a reduced distance between voids or clusters; hence, the pore-sizes are reduced (mesopores and micropores) and the polymer surface area increases. The mixture of A and B in Figure 1, offers two modes of phase separation, i.e., early and late. Thus, bimodal pores are formed as follows: micropores/mesopores or mesopores/macropores or micropores/macropores. It is noteworthy that many authors use the term “good thermodynamic solvent” instead of “solvating diluent”. In this review, the term solvating diluent is preferred because the thermodynamic parameters of these solvents are not adequately characterized in the context of PCR materials; therefore, the term “good thermodynamic solvent” was not used in view of the existing knowledge gaps (vide infra), and to avoid misleading the reader that a thermodynamic analysis of solvation was reviewed herein.

The allure towards this technique is demonstrated by the variation in the surface area (SA), pore-size, pore-size distribution (PSD), pore-volume, and morphology. The SA is found to vary from ~10¹–10³ m²/g, PSD ranges from microporous, mesoporous and macroporous with unimodal and bimodal distributions. Pore-volumes typically range from ~0–3 cm³/g [1,2,4,5,7,11,12,13,14,15,16,21,22] where the higher SA values are due to solvating diluents. The average pore diameter and PSD may also vary, i.e., increase, decrease, or both. Moreover, the morphology coarsens in a less subtle fashion when a non-solvating diluent is used, and does not coarsen or becomes more refined in the case of a solvating diluent [2,25,26].

Some common types of porogens investigated include organic solvents of variable molecular weight. Examples of low molecular weight solvent porogens include toluene, hexane, cyclohexanone, 2-ethylhexanol, p-xylene, n-heptane [2,4,15,16] and an example of a higher molecular weight solvent is poly(ethylene glycol) [15,23]. On the other hand, inorganic porogens include sodium chloride [24,25,26], silica [25], sodium hydrogen carbonate [27] and supercritical carbon dioxide (SCD) [25,26,27]. SCD is a suitable porogen substitute for other types of solvents that are otherwise challenging to remove from the polymer matrix at the purification stage and/or polymers which are sensitive to very small composition changes of the porogenic solvent mixtures. SCD offers fine control over average pore-sizes and PSD using temperature and pressure conditions to alter the density of SCD. The popular use of organic porogens is traced back to the 1950s [9]; whereas, the first example of SCD as a porogen for modifying PCR materials was more recently reported by Holmes and Cooper [28,29,30].

PCR tuning with porogens has not been extensively studied in a systematic manner and is the subject of this review. This paper describes the surface area, pore-size, PSD, pore-volume and changes in the morphological properties when polymers are tuned with porogenic “inert” solvents at various conditions. In addition, knowledge gaps related to this technique will be addressed and some recommendations for future research are suggested herein.

The criteria for choosing porogenic solvents is often based on their respective molecular size [24], alky chain length [31], mixtures of lipophilic and hydrophilic porogens [32], and solubility parameter (SP) [33,34,35,36,37]. The SP was identified as a quantitative criterion for predicting the pore structure properties of polymers; however, the SP was not adequately addressed in other studies [12,37]. There are three general types of SP values, i.e., a one-component SP is also referred to as the Hildebrand solubility parameter, a two-component solubility (physical-chemical) parameter, and the Hansen’s three-component solubility parameter [31,32,33,34,35,36].

The three approaches employing SP criteria have been applied in various studies. In view of its simplicity, the Hildebrand one-dimensional SP is used for solutions without molecular polarity or specific intermolecular interactions. The two-dimensional SP is preferred over the Hansen’s three-dimensional SP for solvents with lower molar volumes since they yield comparable results. Hansen’s three-dimensional SP is favored [33,36] for solvents with greater molar volumes, such as phthalates, and provides reasonably accurate results.

The tuning of PCR materials with the aforementioned porogens is determined by various factors [2,4,6,7,10,15,16,17,18,19,20,21,38], as follows: (i) the feed ratio of the porogen to monomer mixture (or porogen volume) yields optimal conditions when the ratio is maximized; (ii) porogens with variable SP; (iii) mixing multiple porogens, i.e., co-porogens; (iv) variable temperatures where higher temperatures produce smaller pores; (v) variable stirring speeds influence microemulsion formation and the particle size; (vi) crosslinking density affects the PSD; (vii) polymer composition where the relative monomer type and feed ratio affects the pore-size and PSD; and (viii) the nature of the polymerization technique. Baydal et al. [39] investigated the influence of different parameters and found that the type of porogen and the relative volume have a large influence on the copolymer resin properties. Based on their findings, the goal of this review focuses on tunable porogens of variable types, both as pure solvents and their mixtures (co-porogens). In addition, the influence of porogen volume is examined while the other factors are treated as secondary in their importance.

As mentioned above, this review covers the topic of suspension polymerization as the main synthetic approach for preparing PCR materials. A schematic representation of this polymerization technique was described previously (cf. Figure 8 in [40]). In general, there is an organic phase (also referred to as the discontinuous phase) comprised of predetermined amounts of monomer(s), initiator, and solvent(s) which are added into an aqueous phase, and is also referred to as the continuous phase. The mixture is comprised of a suspension stabilizer in water in a reactor vessel at ~60–70 °C with mechanical stirring at a suitable speed for a desired particle size. The reaction is maintained ~80 °C for 3–24 hours and the resulting spherical particles are washed with water, ethanol, or methanol, and extracted with a suitable solvent such as acetone for 24–48 hours [1,2,4,5,7,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,39,40]. A commonly used polymerization initiator agent is α-α’-azo-bis-isobutyronitrile (AIBN).

Research results on the influence of pure solvents as porogens for different resins are summarized in Table 1 [10,15,16,46]. Table 1 illustrates how the surface area varies in accordance with the use of different porogens. Additionally, the results show how a similar class of resins may be tuned to form micro-, meso- and macropores [7,12,44,45]. As mentioned above, solvents that act as solvating diluents may generate micropores; thereby, increasing the polymer surface area.

For example, the results from reference [16] in Table 1 show that toluene is a solvating diluent and cyclohexanone is a non-solvating diluent. 2-ethylhexanol is an intermediate case between solvating and non-solvating diluent. The results were explained by the trend for the SP values; toluene (18.2 MPa^1/2), 2-ethylhexanoic acid (19.4 MPa^1/2), and cyclohexanone (20.3 MPa^1/2); whereas, the resin is ~17–18 MPa^1/2. According to the SP values, toluene has a closer match to the resin, and this is based on the one-component SP using Hildebrand solubility parameters. However, this criterion has a fault which is illustrated in Table 1 (cf. [46]). The authors argue that 1-chlorodecane is a non-solvating diluent for the resin and it yields a greater surface area over the solvating diluents, demonstrating its use as a novel porogen. However, if one looks at the SP value of the solvents investigated, one observes a variation in the SP values; heptane (15.3), cyclohexane (16.8), 1-chlorodecane (17.0), toluene (18.2), and dibutyl phthalate (23.3 MPa^1/2). The use of a one-component SP criterion to explain the results, i.e., if |Δδ| ~ 0, the solvent is the solvating diluent and 1-chlorodecane should not have the greatest surface area. Cyclohexanol should have a surface area <25 m²/g when compared to the results from [2]. In contrast, Rabelo and Coutinho [34] have illustrated that a pure solvent can mislead the interpretation of results; hence, Hansen’s three-component SP value may be used. An example of this approach will be described further in Section 7.

The preparation of PCR materials using polar co-monomers that do not dissolve in the solvent are the most solvating diluents. The use of non-solvating diluent affects the nature of the products. In this case, the most solvating diluent is mixed with the non-solvating or an intermediary solvating diluent. The prediction of the SP value is shown in Equation (6) [46].

where φ₁ and φ₂ are the volume fractions of the two porogens, and δ₁and δ₂ are the respective SP values. Tuning is done by mixing the two solvents at different volume fractions until the desired properties are achieved. Table 2 shows some of the examples studied for such mixtures. Reference [16] shows that a non-solvating medium such as dimethylformamide (DMF; δ = 24.8 MPa^1/2) was used as a diluent and mixed with 2-ethylhexanoic acid to yield a higher surface area material with a unimodal size distribution of 20 nm. The product obtained from toluene mixtures has a decreased surface area relative to pure toluene (cf. Table 1). To this point, it is assumed that a one-component SP value may explain differences in the product characteristics; whereas, the SP values of mixtures are calculated using Equation (6). The respective values are for 2-ethylhexanol:DMF (21.2 MPa^1/2) and toluene/DMF (20.4 MPa^1/2). The difference in SP values (|Δδ| > 2) obtained using the high end of the SP scale for the resin indicates polymer formation is not favoured with high surface area and reduced pore size. The importance of using Hansens’s three-component SP is further illustrated by the results obtained for a toluene/DMF mixture. The result of adding two poor H-bonding diluents together describes why DMF contributes to an attenuation of the favourable SP of toluene, as evidenced by a decrease in the surface area of the resin. In the case of the 2-ethyhexanol mixtures, the authors conclude that the SP value increases, as compared to pure toluene and 2-ethyhexanol. The results indicate an increase in the SP value of DMF. This observation highlights one of the knowledge gaps outlined in Section 7 and Section 8. There is a pronounced effect on the matching of the SP value of the co-monomers rather than the resulting polymeric materials.

The volume of the porogenic solvent plays a major role in the structure and morphology of the resins. Previous research indicates that the copolymer resin prepared in the presence of larger amounts of porogen have a looser structure than those obtained with reduced levels of porogen [39,43,44,45]. In terms of pore formation, Figure 4 [42] illustrates three types of resins produced with higher porogen levels which result in a shift of the PSD toward larger pores (to the right) with greater pore abundance (higher peaks). The resin HS (poly(2-hydroxyethyl methacrylate-co-styrene)) has particles with smaller pores; whereas, HN (poly(2-hydroxyethyl methacrylate-co-N-vinyl-2-pyrrolidone)) particles possess larger pores. Therefore, pore size may be controlled by adjusting the relative porogen ratio. The authors conclude from their results that a lower porogen volume ratio generates fewer pores with a reduced pore size; whereas, a higher volume ratio of porogen contributes to greater numbers of pores with greater pore size. These results are further supported by Liu et al. [46] where greater concentration of 1-chlorodecane reveals that the precipitated phase exhibits limited swelling and the polymer aggregates behave as an immiscible “solid” polymer, similar to bulk microspheres prepared in the absence of porogen. The resulting final pore volume between aggregates should be close to the volume of the non-solvating porogen. In contrast, low levels of 1-chlorodecane yield a precipitated phase that undergoes swelling with a final pore volume less than that of the non-solvating porogen. The relative tuning of solvent porogen composition represents a good technique for tuning PSD.

There are examples of high molecular weight organic porogens and inorganic systems; however, this review is focused on low molecular weight organic solvents in view of the continued research interest in this area. Examples of inorganic porogens of interest include sodium chloride, sodium hydrogen carbonate, silica and supercritical carbon dioxide (SCD). SCD is treated as an inorganic porogen similar to the definition of other carbon halides, sulfides, and oxides as inorganic compounds. SCD is an interesting porogen with versatile tuning capability due to its unique thermodynamic characteristics. Thus, this review will focus on SCD relative to the other types of inorganic porogens.

The variation of the surface area, pore volume and pore sizes for different types of PCR materials due to variations in porogenic solvents was discussed in this review with limited commentary on PSD. The tuning of PCR materials can yield three categories of polymers; micro-, meso- and macroporous resins. Figure 2 in ref. [16] shows an example from non-solvating pure solvent (2-ethylhexanol) and a solvating diluent (toluene), respectively [16]. The PSD characteristics of PCR materials are clearly seen by the shift from a bimodal to a unimodal distribution. The general mechanism of pore formation speculates that larger pores are formed for non-solvating diluents over solvating diluents, but not for bimodal size distributions. One may speculate that the SP does not remain constant but varies continuously during polymerization when mixtures of solvents undergo phase separation, resulting in bimodal distributions. In contrast, the addition of a non-solvating diluent (DMF) to 2-ethylhexanol causes a shift from a bimodal to a unimodal distribution (cf. Figure 3 in ref. [16]). The conclusion is speculative since there is strong support that solvent properties apart from SP (although dispersive, dipolar interactions and H-bonding are included in Hansen’s model), are generally unaccounted for in many studies. Therefore, solvent/monomer and solvent/polymer interactions need to be further studied in a systematic fashion to understand the pore formation mechanism. Unimodal and bimodal behaviour was similarly observed in other studies for different PCR materials [27,43]. With the aid of tuning experimental variables, as discussed herein, tunable bimodal distribution may be achieved.

The morphological properties PCR materials have been reported in various ways. In one case, Ortiz-Palacios and coworkers [2] argued that solvating diluents retain their individuality as microgels (microspheres), according to the general pore formation mechanism of porogenic solvents. This phenomenon renders the appearance of the morphology of the resin as smooth from the outside and rough on the inside due to micropore formation. The opposite is true for non-solvating diluents where the interior appears dense due to the lack of pore structure. On the contrary, de Santa Maria and coworkers [48] argue that a solvating diluent will increase porosity; hence, the surface of the resins will appear rough in appearance. There are differences in the appearance of PCR-based materials on the outer surfaces according to the nature of the solvent. The apparent contradiction may be due to the chemical type of PCR material. Ortiz-Palacios and coworkers [2] studied the styrene-co-divinylbenzene resin while de Santa Maria et al. [48] studied 2-vinylpyridine-co-styrene-co-divinylbenzene. In support of this, Lin and coworkers [49] used a solvating diluent which resulted in a smooth and regular surface. The appearance became rough upon addition of another copolymer. Figure 9 illustrates an example of the rough (Group 1) and smooth surfaces (Group 2) of the PCR materials at variable porogen composition. This observation was recently explained by Scheler [50] where he concluded that Hansen’s SP value changed during polymerization due to changes in the composition of the porogenic solvent mixture. The affinity between the resins and the evolving solvent composition, decreases or increases, during polymerization. This leads to differences in the morphology of the PCR materials.