Principles and Characteristics of Polymerization-Induced Self-Assembly with Various Polymerization Techniques

Erik Jan Cornel,Jinhui Jiang,Shuai Chen,Jianzhong Du

DOI: https://doi.org/10.31635/ccschem.020.202000470

2021-01-01

CCS Chemistry

Abstract:Open AccessCCS ChemistryMINI REVIEW1 Apr 2021Principles and Characteristics of Polymerization-Induced Self-Assembly with Various Polymerization Techniques Erik Jan Cornel, Jinhui Jiang, Shuai Chen and Jianzhong Du Erik Jan Cornel *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Department of Polymeric Materials, School of Materials Science and Engineering, Key Laboratory of Advanced Civil Engineering Materials of Ministry of Education, Tongji University, Shanghai 201804 Google Scholar More articles by this author , Jinhui Jiang Department of Polymeric Materials, School of Materials Science and Engineering, Key Laboratory of Advanced Civil Engineering Materials of Ministry of Education, Tongji University, Shanghai 201804 Google Scholar More articles by this author , Shuai Chen Department of Polymeric Materials, School of Materials Science and Engineering, Key Laboratory of Advanced Civil Engineering Materials of Ministry of Education, Tongji University, Shanghai 201804 Google Scholar More articles by this author and Jianzhong Du *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Department of Polymeric Materials, School of Materials Science and Engineering, Key Laboratory of Advanced Civil Engineering Materials of Ministry of Education, Tongji University, Shanghai 201804 Department of Orthopedics, Shanghai Tenth People’s Hospital, Tongji University School of Medicine, Shanghai 200072 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.020.202000470 SectionsAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail A chemical reaction that drives a physical polymer self-assembly process, namely, polymerization-induced self-assembly (PISA), combines block copolymer synthesis and nanoparticle formation efficiently at high polymer concentrations. Various nanoparticle morphologies such as spheres, worms, and vesicles can be prepared readily in polar and nonpolar media. PISA has been well developed in combination with reversible addition-fragmentation chain transfer (RAFT) polymerization. Notably, developments with other polymerization methods are also achievable. In this report, first, we discuss the general principles of RAFT-PISA and the nanoparticles generated from this method. Specifically, new insights into polymer nucleation and subsequent morphological evolution are highlighted. Subsequently, PISA formulations that use other polymerization methods [atom transfer radical polymerization (ATRP), nitroxide-mediated polymerization (NMP), ring-opening metathesis polymerization (ROMP), and ring-opening polymerization (ROP) of N-carboxyanhydrides (NCAs)] are summarized in detail. Finally, more exotic PISA formulations are emphasized: these are based on organotellurium-mediated living radical polymerization (TERP), living anionic polymerization (LAP), addition-fragmentation chain transfer (AFCT) polymerization, reversible complexation-mediated polymerization (RCMP), and cobalt-mediated radical polymerization (CMRP), or utilization of a comonomer that undergoes radical ring-opening polymerization (rROP). This review is concluded with a perspective on the status and potential of PISA. Download figure Download PowerPoint Introduction As polymer nanoparticles become more applicable, the demand for their efficient synthesis grows. Polymer synthesis methods that induce nanoparticle formation are highly desired. Various polymerization techniques that enable in situ nanoparticle preparation are precipitation polymerization, dispersion polymerization, and emulsion polymerization.1,2 The former two utilize miscible monomers; the latter two utilize a polymer stabilizer or surfactant. These three techniques are based on a growing insoluble polymer chain or network that is stabilized by various means, and therefore, remain dispersed in the preparation media. An intrinsic limitation of these methods is that the resulting nanoparticle morphology is mainly spherical, lacking diversity in nanoparticle morphology. Furthermore, in the last two cases, stabilizer polymers or surfactants are not always desired in the final polymer product, and removal is often difficult. Other commonly used nanoparticle preparation methods are polymerization-induced self-assembly (PISA),3,4 and the more traditional postpolymerization solvent switch approach.5 Compared with the previously discussed approaches, these two methods utilize block copolymers consisting of a soluble block and an insoluble block. The resulting spherical particles can, therefore, also be seen as polymer micelles; higher order morphologies can also be achieved. To some extent, PISA is a type of dispersion (or emulsion) polymerization—an insoluble polymer block formed from a soluble (or immiscible) monomer. In contrast, conventional dispersion or emulsion polymerization utilizes a separate polymer stabilizer or surfactant for the preparation of latex particles. These stabilizers are, therefore, not chemically attached to the “core-forming” polymer.2 PISA is based on the chain extension of an initial soluble precursor block, which acts as a steric stabilizer, with a second insoluble polymer block that forms the nanoparticle core in situ. This polymerization reaction triggers diblock copolymer self-assembly once a sufficiently high degree of polymerization (DP) of the core-forming block is attained. PISA is efficient because polymer synthesis and assembly occur simultaneously. Furthermore, PISA can be performed at a range of final polymer concentrations (5–50% w/w).6 The most commonly observed nanoparticle morphologies obtained from PISA are spheres, worms, and vesicles. As far as we are aware, the first example of such a synthetic approach was reported in 2002,7 and its term “PISA” first appeared in the academic literature in 2009.8 The other frequently-used block copolymer self-assembly method, the solvent switch approach, is a well-established post-polymerization procedure that is based on the dropwise addition of a selective solvent to a dilute block copolymer solution.9 The addition of this selective solvent leads to the insolubility of one of the polymer blocks. This process triggers block copolymer self-assembly into various nanoparticle morphologies once a certain solvent ratio is attained. Subsequently, extensive dialysis is required to remove the suitable initial solvent. This approach is generally conducted at a polymer concentration of ∼1–3% w/w and might, therefore, seem inefficient compared with PISA. However, the solvent switch method allows a wider variety of polymers than PISA since this method is independent of the polymerization technique. In other words, the desired polymer is not always easily accessible with a polymerization technique that is well established for PISA. Hence, the solvent switch method is still widely used for the preparation of various application-oriented functional nanoparticles. Indeed, this also means that the development of new PISA protocols with different polymerization methods is highly desired. Here, the principles and characteristics of PISA are described with the focus on several polymerization methods that allow various types of monomers to be utilized. Reversible addition-fragmentation chain transfer (RAFT)-PISA is discussed first since this is unarguably the most well-established PISA method. General RAFT-PISA principles and nanoparticle characteristics are highlighted. Thereafter, the characteristics of other PISA formulations, based on other polymerization techniques, including atom transfer radical polymerization (ATRP), nitroxide-mediated polymerization (NMP), ring-opening metathesis polymerization (ROMP), and ring-opening polymerization (ROP) of N-carboxyanhydrides (NCAs), are discussed and compared. Subsequently, more exotic and novel PISA formulations are highlighted. These are based on organotellurium-mediated living radical polymerization (TERP), living anionic polymerization (LAP), addition-fragmentation chain transfer (AFCT), reversible complexation-mediated polymerization (RCMP), cobalt-mediated radical polymerization (CMRP), and radical ring-opening polymerization (rROP). This review concludes with a perspective on PISA, highlighting its current potential and recent scientific trends. We would guide the reader to other reviews for a more in-depth discussion of the recent trends in RAFT-PISA4,10,11 and ROMPISA,12 and an assessment of general PISA reaction conditions suitable for various monomer species.3 PISA can produce applicable nanoparticles that are functional or reactive, this is discussed in a recent review from Delaittre and coworkers.13 These reviews also provide comprehensive lists of reported PISA formulations; such a list is therefore not given in this review. RAFT-PISA RAFT polymerization is the most studied polymerization method for PISA. A range of morphologies can be obtained when using both RAFT dispersion14,15 and RAFT emulsion polymerization.16 Furthermore, this method allows nanoparticle cross-linking17 and can be performed in polar solvents, such as water14,18; various organic solvents and alcohols19–22; and in nonpolar solvents, such as n-alkanes15,23 and mineral oil.6,24 RAFT-PISA can also be performed in ionic liquids and critical CO2.25,26 A well-known example of an aqueous RAFT dispersion polymerization, and certainly, the most-studied, is the poly(glycerol monomethacrylate)-block-poly(2-hydroxypropyl methacrylate) (PGMA-b-PHPMA) composition (Figure 1a).14,27,28 Here, PGMA forms the soluble stabilizer block, and PHPMA forms the insoluble nanoparticle core. Nonpolar RAFT-PISA formulations are less well-known. A good example of such formulations is the poly(lauryl methacrylate)-block-poly(benzyl methacrylate) (PLMA-b-PBzMA) composition in various n-alkanes (Figure 2c),15 where the PLMA blocks act as an oil-soluble stabilizer and the PBzMA blocks from the insoluble nanoparticle core. Commonly observed nanoparticle morphologies are spheres, worms, and vesicles. All three morphologies are observed in polar and nonpolar media. Interestingly, these three morphologies can also be obtained via RAFT-PISA methods that are based on aqueous emulsion polymerizations.16,31 Figure 1 | PISA phase diagrams for PGMA-b-PHPMA nanoparticles in water, for the reproducible synthesis of spheres (S), worms (W), and vesicles (V) at high solids content. (a) Molecular structure of PGMA-b-PHPMA. (b) Master phase diagram where the PGMA DP and the PHPMA DP are varied at 20% w/w (and 10% w/w for PGMA DPs < 47). (c) Phase diagram where the PHPMA stabilizer block DP and the final copolymer concentration are varied. Reprinted with permission from ref 14. Copyright 2012 American Chemical Society; Reprinted with permission from ref 27. Copyright 2018 American Chemical Society. PISA, polymerization-induced self-assembly; PGMA-b-PHPMA, poly(glycerol monomethacrylate)-block-poly(2-hydroxypropyl methacrylate); DP, degree of polymerization. Download figure Download PowerPoint RAFT-PISA formulations are also well-known for the “PISA phase diagrams” (Figures 1b and 1c). Notably, these diagrams are not related to the traditional physical phase diagrams that consist of the three physical states of matter (solid, liquid, and gas). From these PISA phase diagrams it is clear that both the stabilizer and core-forming block DPs (or block volumes) can be altered to target the desired nanoparticle morphology reproducibly. This aspect is often referred to as the packing parameter. However, it is essential to realize that the same copolymer composition can form different nanoparticle morphologies when the solid content is varied (total polymer concentration). This can be explained by the morphological nanoparticle evolution mechanism during RAFT-PISA, which is generally limited to spherical nanoparticles when conducted at low concentrations (as described later in this review). Figure 2 | Temperature-responsive behavior of diblock copolymer worm gels prepared via RAFT-PISA in polar and nonpolar media. (a) Cooling an aqueous PGMA-b-PHPMA dispersion leads to a reversible worm-to-sphere transition. (b) Heating PLMA-b-PBzMA worms in n-dodecane induce a reversible worm-to-sphere transition. (c) Molecular structure of PLMA-b-PBzMA diblock copolymers. Reprinted with permission from refs 29 and 30. Copyright 2014 American Chemical Society. RAFT, reversible addition-fragmentation chain transfer; PISA, polymerization-induced self-assembly; PGMA-b-PHPMA, poly(glycerol monomethacrylate)-block-poly(2-hydroxypropyl methacrylate); PLMA-b-PBzMA, poly(lauryl methacrylate)-block-poly(benzyl methacrylate). Download figure Download PowerPoint There are essential differences between nanoparticles prepared via RAFT-PISA in water and nonpolar media: nanoparticles in water can have lower critical solution temperature (LCST) characteristics,29,32–35 while nanoparticles in nonpolar media tend to have upper critical solution temperature (UCST) characteristics.30,36,37 In other words, the insoluble core-forming block of PGMA-b-PHPMA nanoparticles in water and PLMA-b-PBzMA nanoparticles in oil becomes partially solvated at specific temperatures. The nanoparticle cores with LCST characteristics in water become progressively plasticized on cooling, while nanoparticle cores with UCST characteristics in nonpolar media become progressively plasticized on heating. Core plasticization is well documented for nanoparticle worms obtained from RAFT-PISA. Such worm dispersions form a viscous gel when prepared at sufficiently high polymer concentrations. It is believed that worm entanglements cause this viscous state. However, the percolation theory can also explain this phenomenon.38 Alterations in temperature can trigger core plasticization, this can lead to reversible morphological transitions to lower-order morphologies: cooling PGMA-b-PHPMA worms induce a morphological worm-to-sphere transition; the same transition is observed on heating PLMA-b-PBzMA worms (Figures 2a and 2b). Interestingly, this morphological change is characterized by a gel-to-liquid transition. The temperature at which the dispersion no longer forms a worm gel is defined as the critical gelation temperature (CGT).6 Such worm gels are also concentration-dependent: diluting these dispersions can lead to a free-flowing fluid once the critical gelation concentration (CGC) is attained, owing to reduced interworm contacts.38 The reversible nature of the thermally induced worm-to-sphere transition also disappears when the initial worms are sufficiently diluted.30 This suggests that the worm regeneration via sphere–sphere fusion is inhibited at low particle concentrations. The morphological worm-to-sphere transition can be prevented by cross-linking the initial worms.39 Indeed, this modification leads to a worm gel that is incapable of forming a free-flowing dispersion upon heat treatment at high solid contents. A reversible vesicle-to-worm transition can also be achieved according to this principle—resulting in an abrupt increase in viscosity.32,40 It is also noteworthy that pH-responsive nanoparticles have been prepared by aqueous RAFT-PISA that are capable of morphological transitions.41–44 These transitions are based on polymer end groups that become charged upon the addition of acid or base, leading to higher solvation of the nonionic polymer stabilizer block. Finally, thermally induced morphological transitions are also observed for more complex nanoparticles. The thermal behavior of nanoparticles prepared with an inherently different aqueous PISA protocol was reported recently by Cai and co-workers in 2020.45 In contrast to the previously discussed formulations, this aqueous PISA formulation is based on the chain extension of a macromolecular chain transfer agent (macro-CTA), poly[N-(2-hydroxypropyl)methacrylamide], with a charged core-forming monomer in the presence of an oppositely charged homopolymer. The resulting nanoparticles are polyion complexes. This approach is also known as polymerization-induced electrostatic self-assembly (PIESA). Conform to the previously discussed aqueous PGMA-b-PHPMA formulation, and a reversible morphological sphere-to-worm evolution could be realized on heating (Figure 2a). An irreversible worm-to-vesicle transition was also achieved by altering block DP of the charged core-forming block. This transition was attributed to the dehydration of the stabilizer block, as a result of the formation of hydrogen bonds between the core-forming block and the stabilizer block, instead of core plasticization. Morphological evolution during RAFT-PISA (from chains to spheres, to worms, to vesicles) Despite the differences in temperature-responsive behavior, polar and nonpolar RAFT-PISA formulations have similar characteristics in terms of in situ morphological evolution. Generally, this morphological evolution proceeds from dissolved polymer chains to spheres, to worms, to vesicles.46 Herein, each morphological transition is discussed separately. Polymer chain-to-sphere transition RAFT-PISA generally proceeds as follows: a soluble polymer precursor block (also known as macro-CTA) is prepared via solution polymerization. After purification, this macro-CTA is chain extended with a monomer to form a soluble diblock copolymer. The second polymer block becomes progressively insoluble as the polymerization proceeds. Polymer assembly (or nucleation) occurs once a critical core-forming block DP is achieved. Multiple dissolved diblock polymer chains aggregate during this process and form nascent spherical nanoparticles. Here, the second insoluble block forms the nanoparticle core, and the soluble block acts as a steric stabilizer for these initial nanostructures. One might consider that nucleation leads to premature termination of the RAFT polymerization. However, quite the opposite is observed. Enhanced first-order reaction kinetics is a well-established phenomenon that characterizes nucleation during RAFT-PISA.16,24,29,33,47–50 It is presumed that this increase in the reaction rate is caused by a local high monomer concentration within the nascent spherical nanoparticles. In other words, unreacted monomer is likely to diffuse into the nanoparticle cores and stimulate the polymerization reaction. Such small molecule diffusion processes are common. Solvent molecules are also capable of diffusing into a nanoparticle core.36,37,51,52 New insights into this nucleation-induced increase in the reaction rate were recently reported.37 Poly(stearyl methacrylate)-block-poly(2,2,2-trifluoroethyl methacrylate) (PSMA-b-PTFEMA) spheres were prepared in n-tetradecane (Figure 3a); a dispersion that can become highly transparent [refractive index (RI)-matched] depending on the applied temperature. This formulation is suitable for studying nucleation during RAFT-PISA since detailed reaction kinetics can be obtained by in situ 19F NMR (Figures 3b, 3c, and 3d). A 19F NMR spectrum comprises two distinct peaks that represent the fluorinated monomer and the growing fluorinated core-forming block. The polymer peak area in the 19F NMR spectrum continues to increase during the polymerization; however, the intensity (peak height) of this peak changes since it describes the polymer nucleation process. A reduction in peak intensity corresponds to the aggregation of polymer chains into nascent spherical nanoparticles. The obtained reaction kinetics for this formulation clearly shows three distinct kinetic reaction regimes. A very modest increase in reaction kinetics is observed after 60 min; an obvious increase is observed after 115 min. Each reaction regime is described as follows: growing dissolved diblock copolymer chains, spheres with nascent micelle cores, and spheres with compact micelle cores. Here, it is suggested that a fraction of the solvent is present within the growing nanoparticle cores, and that this solvent fraction reduces as polymerization proceeds, which would then delay the second larger increase in the reaction rate. Also, similar three-stage reaction kinetics are observed for an aqueous PISA formulation.53,54 However, more research is desired to explain this phenomenon further. Figure 3 | (a) Molecular structure of PSMA-b-PTFEMA, prepared via RAFT dispersion polymerization in n-tetradecane. (b) In situ 19F NMR studies during the PISA synthesis of PSMA-PTFEMA spheres. The TFEMA monomer peak area increases upon polymerization. The resulting polymer peak intensity decreases after diblock copolymer self-assembly (also known as nucleation). After that, polymerization occurs within the monomer (and solvent) swollen nanoparticles. (c) Three separate kinetics reaction regimes are observed during the chain extension of PSMA with TFEMA. (d) Normalized PTFEMA 19F NMR peak intensities plotted against TFEMA monomer conversion confirm three separate reaction stages. Reprinted with permission from ref 37. Copyright 2018 American Chemical Society. PSMA-b-PTFEMA, poly(stearyl methacrylate)-block-poly(2,2,2-trifluoroethyl methacrylate); RAFT, reversible addition-fragmentation chain transfer; PISA, polymerization-induced self-assembly. Download figure Download PowerPoint It is important to consider that the self-assembled spheres obtained from various RAFT-PISA formulations are not “frozen” nanostructures. A recently published paper from Armes and co-workers52 described a fast copolymer chain-exchange process between spherical nanoparticles prepared by PISA in nonpolar media. This work showed that a larger DP of the insoluble core-forming block and a higher temperature lead to a faster copolymer chain-exchange process between spherical nanoparticles. This study was conducted on the final nanoparticle spheres (post-PISA). However, it is likely that nanoparticle core-solvation by monomer and hot solvent aids this copolymer chain-exchange process during PISA. A different study showed that heating PLMA-b-PBzMA polymer spheres in n-dodecane cause a fraction of block copolymer chains to become molecularly dissolved.36 Indeed, this causes a reduction in the aggregation number (number of polymer chains per nanoparticle). This study also showed that diblock copolymer spheres with larger core-forming block DPs tended to become less solvated and do not produce molecularly dissolved block copolymers on heating. It is likely that copolymer chain-exchange can also be observed for nanoparticles obtained from aqueous PISA formulations.51 However, as far as we are aware, this has, unfortunately, not yet been established. Aqueous polymer nanoparticles are likely to have interesting dynamic characteristics since they can exhibit LCST-characteristics. Other aspects that need to be considered during the early stages of RAFT-PISA are that the aggregation number and spherical nanoparticle diameter increases as RAFT polymerization proceeds. This characteristic of RAFT-PISA is well established with in situ small-angle X-ray scattering (SAXS) studies on an aqueous emulsion polymerization,16 an aqueous dispersion polymerization,54 and a nonpolar dispersion polymerization.47 The increase in the number of polymer chains per nanoparticle cannot be explained merely by copolymer chain exchange between spherical nanoparticles because this process does not change the aggregation number when in equilibrium. There are three possible reasons for this phenomenon: (1) a fraction of diblock polymer chains is molecularly dissolved during the early stages of PISA. As the core-forming block DP increases, block copolymers become progressively more insoluble and insert in the nascent diblock copolymer nanoparticles. This theory can be further rationalized with the molecular weight distribution (MWD) of copolymer chains: chains with shorter DPs are more likely to be molecularly dissolved. (2) Sphere–sphere fusion causes the formation of larger spherical nanoparticles with a doubled number of copolymer chains. After fusion, the preservation of the spherical morphology might be aided by monomer and solvent fractions within the nanoparticle core. (3) Nanoparticles might be extremely dynamic during the early stages of PISA; thus all block copolymer chains might be capable of constant rearrangement, leading to the most energetically favorable arrangement. It is very likely that all three possible mechanisms are to some extent responsible for the increased aggregation number during RAFT-PISA. More research is required to further establish this characteristic of RAFT-PISA. Furthermore, it would be interesting to evaluate if this phenomena can be observed for PISA with other polymerization techniques. Sphere-to-worm transition Elongated structures (worms) can be observed when the core-forming block DP increases after the initial spherical particles are obtained. It is believed that this sphere-to-worm transition is caused by multiple sphere–sphere fusion processes. Indeed, it is difficult to imagine how such elongated structures would originate from a process other than sphere–sphere fusion such as copolymer chain exchange. This hypothesis is supported by the inability to attain this elongated structure when using a sufficiently long stabilizer block during RAFT-PISA.30 In this case, it is very likely that a thicker nanoparticle corona prevents sphere–sphere fusion. Instead of aggregating, spherical nanoparticles with a sufficiently long stabilizer block grow in diameter as the core-forming block becomes longer. Another argument that supports the sphere–sphere fusion mechanism is the concentration dependency of the copolymer on the sphere-to-worm evolution during RAFT-PISA. Worms are obtained only at sufficiently high polymer concentrations. A low concentration leads to the formation of solely spherical nanoparticles that grow in diameter when the core-forming block DP increases.14 Interestingly, fusion processes have also been observed at low polymer concentrations for nanoparticles prepared via RAFT-PISA and other techniques.36,55 However, these fusion processes are supported by specific repeat units, or by introduction of a second population diblock copolymer spheres with shorter core-forming block DP. Worms can also be obtained as a final temperature-responsive morphology, as discussed earlier (Figure 2a and 2b). It is noteworthy that this worm phase usually occupies a narrow region in the PISA phase diagram for various formulations. However, it is possible to increase the worm-shaped nanoparticles’ dimensions by preparing a polymer with a longer stabilizer and core-forming block DP for specific formulations.27 This leads to longer and thicker worms. Interestingly, the CGT is lower for the worm gels, which consist of longer total block DPs. Finally, it is important to consider that the sphere-to-worm evolution is the most crucial step in the formation of vesicular nanoparticles during RAFT-PISA. Indeed, it is challenging (if not impossible) to obtain the subsequent intermediate octopi, jellyfish, and vesicular nanostructures if worms are not initially formed. Worm-to-vesicle transition Targeting a sufficiently high core-forming block DP can lead to the formation of vesicular nanoparticles, provided that a sufficiently short stabilizer block DP is used (as discussed in the previous section). Transmission electron microscopy (TEM) studies of samples obtained during RAFT-PISA, quenched at intermediate monomer conversions, revealed various intermediate PGMA-b-PHPMA nanostructures in water (Figure 4). These data imply that the worm-to-vesicle transition is based on the coalescence of branched worms, to form octopi structures. These structures tend to curve and form nanoparticles with a jellyfish morphology, which then wrap up and create the final vesicle morphology. It is very challenging to obtain the intermediate octopi and jellyfish morphology as final structures, which are not plasticized by the remaining monomer, and thus, are generally not found in RAFT-PISA phase diagrams as a pure phase. However, they have been detected with TEM as intermediate morphology during RAFT-PISA in polar and nonpolar media.46,47 Figure 4 | The morphological worm-to-vesicle evolution during RAFT-PISA. TEM images correspond to the