Block Copolymers: Synthesis, Self-Assembly, and Applications

Abstract

Research on block copolymers (BCPs) has played a critical role in the development of polymer chemistry, with numerous pivotal contributions that have advanced our ability to prepare, characterize, theoretically model, and technologically exploit this class of materials in a myriad of ways in the fields of chemistry, physics, material sciences, and biological and medical sciences. The breathtaking progress has been driven by the advancement in experimental techniques enabling the synthesis and characterization of a wide range of block copolymers with tailored composition, architectures, and properties. In this review, we briefly discussed the recent progress in BCP synthesis, followed by a discussion of the fundamentals of self-assembly of BCPs along with their applications.

1. Introduction

Block copolymers (BCPs) are a specific class of copolymers, in which the chemically distinct monomer units are grouped in discrete blocks along the polymer chain [1]. Figure 1 illustrates a few of the many architectures of BCPs, which can be configured into linear, branched (graft and star), and cyclic molecular architectures. Thanks to the advancement of polymer synthetic strategies and techniques, e.g., controlled polymerization techniques along with facile post-polymerization functionalization, BCPs with precisely controlled molecular weights and defined macromolecular architectures can be prepared [2,3,4,5,6,7,8]. The extraordinary structural and compositional versatility of BCPs has facilitated an explosion in the discovery and implementation of innovative synthetic strategies capable of generating previously unattainable levels of man-made architectural complexity.

One of the ubiquitous features of BCPs is their ability to form a plethora of nanoscale ordered structures. By manipulating the molecular parameters such as the Flory-Huggins interaction parameter (χ), the degree of polymerization (N), and the volume fraction (f), various morphologies including spherical, cylindrical, lamellar, and others have been revealed both experimentally and theoretically [9,10]. Furthermore, recent studies have demonstrated that the macromolecular architecture is another key factor in controlling both the resulting morphologies and their extent of long range order [11]. Due to these fascinating features, research on BCPs has long been a popular topic worldwide. Figure 2 shows the number of publications with the topic “block copolymer” over the past five decades.

In addition to the academic interest in BCPs, the scope of applications for BCPs has been rapidly expanding to the fields of advanced materials (e.g., thermoplastic elastomers), drug delivery, patterning, porous materials, and many others over the last two decades [12]. Thermoplastic elastomers take advantage of the combination of rubbery segments and rigid segments within block copolymers. Drug encapsulation and delivery are facilitated by the amphiphilicity of block copolymers in solution. The application of block copolymers in both the soft lithography and synthesis of porous materials is based on the various nanoscale morphologies induced by self-assembly. Each of the applications mentioned above will be elaborated on in the main text.

This review briefly covers the recent breathtaking progress in the synthesis, self-assembly behavior, and applications of BCPs. Section 2 briefly describes the synthetic strategies for the preparation of BCPs. Section 3 discusses BCP self-assembly in bulk and in solution. Section 4 presents several major applications of BCPs in thermoplastic elastomer, drug delivery, soft lithography, and porous material applications. The conclusions and future perspectives of BCP research are presented in the last section. We hope that this review will attract emerging researchers to this field and advance the understanding and utilization of the fascinating BCP systems.

2. Synthesis of BCPs

Significant work has been focused on the synthesis of BCPs with structural and composition variety. The synthetic strategies mainly include: (1) the sequential addition of monomers via “living”/controlled polymerization techniques; and (2) coupling reactions exploiting the active chain-ends of different chain segments [13].

2.1. Sequential Addtion Polymerization

2.1.1. Controlled Polymerization

Controlled radical polymerization (CRP) techniques represent the most versatile and facile approach for BCP synthesis, mainly due to their compatibility with a wide spectrum of monomers, high tolerance of functional groups and impurities, and ease of experimental setup [14,15,16]. The preparation of BCPs using sequential CRP relies on the fact that the reactive polymer chain end is preserved by a reversible reaction between the active species and the dormant species [17,18,19]. Over the last two decades, numerous literature has described the synthesis of BCPs via controlled polymerization techniques including: atom transfer racial polymerization (ATRP), reversible addition-fragmentation chain transfer (RAFT) radical polymerization, nitroxide-mediated polymerization (NMP), and many others [20,21,22].

Polymer chain ends equipped with an alkyl halide enables ATRP to prepare BCPs through the addition of a second monomer. For example, a series of biocompatible, thermo-responsive ABA triblock copolymers comprising poly(N-isopropyl acrylamide) (PNIPAM) as the A block and poly(2-methacryloyloxyethyl phosphorylcholine) (PMPC) as the central B block was synthesized using ATRP initiated by a difunctional initiator, as illustrated in Scheme 1 [23]. An α,ω-Br terminated PMPC was first synthesized in methanol, purified, and used subsequently for the polymerization of NIPAM. This doubly amphiphilic BCP exhibits interesting thermosensitive properties, whereas above the lower critical solution temperature (LCST) of PNIPAM, the BCP self-assembles into a physical gel. The phosphoryl choline moiety is an important component of cell membranes. The cell viability experiments confirmed that these thermo-responsive gels are sufficiently biocompatible to act as a culture medium for hamster lung cells.

Matyjaszewski et al. reported the preparation of poly(n-butyl acrylate-b-methyl methacrylate) (PBA-b-PMMA) by activators regenerated by electron transfer atom transfer radical polymerization (ARGET ATRP) with ppm levels of Cu catalyst [24]. Controlled polymerization was realized using tris(2-pyridylmethyl)amine (TPMA) as a ligand because of its strong binding interaction to copper. This new ARGET system was also successfully applied to the efficient synthesis of styrene and n-butyl acrylate block copolymers [25]. Later on, miniemulsion ARGET ATRP was developed by the same group and used for the preparation of homopolymers and block copolymers [26,27]. However, ATRP still does not work for some monomers, though it has many advantages compared to other polymerization methods for block copolymerization. For example, some monomers contain functionality that can form complexes with transition metal catalyst and have a detrimental effect on polymerization, such as side reactions resulting in a broad polydispersity index (PDI) [15]. Tremendous efforts have been devoted to developing novel ligands to minimize such side reactions [28]. ATRP-based dispersion polymerizations have also been developed. Armes and coworkers synthesized zwitterionic PMPC-based block copolymers by a dispersion ATRP process [29]. PEG–Br was used as macroinitiator with CuBr/ 2,2′-bipyridine as the catalyst for the dispersion polymerization of PMPC in isopropanol/water (9:1, w/w). Using ethylene glycol dimethacrylate (EGDMA) as a crosslinker, hydrogel particles with controllable sizes were obtained. The particle size could be controlled by the block composition and initial MPC concentration. Using a similar strategy, the same group synthesized well-defined PEG-b-PDMAEMA-b-PMPC. Huang and coworkers synthesized a series of ABA triblock copolymers, consisting of double-bond-containing poly(phenoxyallene) (PPOA), PMMA, or PBA segments using sequential free radical polymerization and ATRP [30]. A bifunctional initiator bearing azo and halogen-containing ATRP initiating groups was first used to initiate the conventional free radical homopolymerization of phenoxyallene with a cumulated double bond to give a PPOA-based macroinitiator with ATRP-initiating groups at both ends. PMMA-b-PPOA-b-PMMA and PBA-b-PPOA-b-PBA triblock copolymers were synthesized by the ATRP of methyl methacrylate and n-butyl acrylate, as depicted in Scheme 2.

RAFT, on the other hand, has seen rapid growth due to its superior compatibility with a broader range of functionalities and high tolerance of impurities. The creative combination of RAFT polymerization with other polymerization techniques, such as ATRP or ring-opening polymerization, has extended the array of available architectures such as graft, star, hyperbranched, etc. [31,32]. The general method for the preparation of BCPs using the RAFT process is through sequential polymerization. For the synthesis of an AB diblock copolymer, the first block was synthesized via a RAFT process, followed by subsequent purification. The resulting end-reactive polymer acts as a macro-RAFT agent for a second polymerization step. To ensure complete end group functionalization, the polymerization yield of the first block is usually kept rather low (<30%). In contrast to this two-step process, Chaduc et al. reported a simple one-pot RAFT process for amphiphilic block copolymers in water [33]. Using 4-cyano-4-thiothiopropylsulfanyl pentanoic acid (CTPPA) as the RAFT agent, poly(acrylic acid), poly(methacrylic acid), or poly(methacrylic acid-co-poly(ethylene oxide) methyl ether methacrylate) was first formed in water. The resulting macro-RAFT agents were then directly used without further purification for the RAFT polymerization of styrene in water in the same reactor. This strategy leads to a very good control of the resulting amphiphilic block copolymers. Very recently, short poly(ethylene glycol) (PEG) was employed as the solvent in the macromolecular RAFT agent-mediated dispersion polymerization of BCPs such as PEG-b-PS, P4VP-b-PS, and PNIPAM-b-PS. A new formulation of polymerization-induced self-assembly of PEG named PEG-PISA to synthesize diblock copolymer nanoassemblies was reported. In PEG-PISA, the viscous PEG affords advantages including fast polymerization rate, good control over the synthesis of diblock copolymers, and in situ synthesis of both amphiphilic and doubly hydrophobic diblock copolymer nanoassemblies at a polymer concentration of up to 50%. Furthermore, two new self-assembled morphologies of ellipsoidal vesicles and nanotubes were formed via PEG-PISA. Nevertheless, RAFT dispersion polymerization often suffers from rather poor colloidal stability, leading to low polymerization rate, broad molecular weight distribution, and lack of control over molecular weight when using a small RAFT agent. The reason for this is believed to be the superswelling effect in the early nucleation stage [34]. Zhu et al. employed an amphiphilic oligomer, poly(acrylic acid-b-styrene) trithiocarbonate, as both surfactant and RAFT agent to polymerize styrene. This macro-RAFT agent was capable of enhancing the colloidal stability and, as a result, the polymerization was successful [34,35]. Well-controlled PS with a molecular weight up to 120 kg/mol and PS-b-PBA were readily synthesized using this method. Furthermore, the same group demonstrated that this method could be used as a promising approach to synthesize thermoplastic elastomer materials [36].

NMP is commonly used for the synthesis of block copolymers consisting of PS, PMMA, poly(vinyl acetate) (PVAc) and poly(dimethyl acrylamide) (PDMA). In a typical process, styrene is polymerized firstly using a bi-molecular initiator benzoyl peroxide and 2,2,6,6-tetramethylpiperidinoxy (TEMPO). The PS segment bearing the TEMPO end group is subsequently used as the macroinitiator for the second block. While the molecular weight distributions are usually broad, this may not limit their practical applications. The monomer conversion is also low when polymerized at high temperature. The deviation from the target molecular weight became more significant, especially when it became higher than 100,000 g/mol. Hawker et al. reported a modified NMP system for the controlled polymerization of dienes in the presence of alkoxyamine initiators based on a 2,2,5-trimethyl-4-phenyl-3-azahexane-3-oxy skeleton (Scheme 3). A detailed study revealed that this modified system was able to control the homopolymerization of isoprene to high conversion and molecular weights from 1000 to 100 000 g/mol with polydispersities of 1.06–1.15. Polyisoprene-containing BCPs, PI-b-PtBuA and PS-b-PI, were also prepared with similar control. In comparison with conventional TEMPO systems, these new systems exhibit a significant improvement in the ability to control the polymerization and further demonstrate the versatility of nitroxide-mediated living free radical procedures [37].

2.1.2. Living Anionic Polymerization

The pioneering work on BCP synthesis using living anionic polymerization (LAP) was reported by Szwarc et al. in 1956 [38,39]. This type of polymerization proceeds to quantitative conversion without chain transfer and/or chain termination. Although there might be an argument that the demanding experimental conditions is one of its limitations, LAP has proved itself to be the best polymerization technique for the preparation of well-defined BCPs based on vinyl monomers such as styrene, dienes, (meth)acrylates, vinyl pyridines, acrylonitriles, as well as cyclic monomers such as lactones, oxiranes, and siloxanes over the last 60 years [2,40,41]. Anionic polymerization is widely exploited in industry to create BCPs on a massive scale. Because of its superior control over the molecular weight, architecture, composition, and functionality as compared to controlled radical polymerization, almost all well-defined BCPs with complex architectures such as star, comb, graft, dendritic, etc. are achievable via the combination of anionic polymerization and linking chemistry [6,42,43,44,45,46,47]. There are several insightful reviews available that describe the recent progress in the synthesis of well-defined BCPs via LAP [48,49].

One key condition that must be satisfied is that the nucleophilicity of the macroanion must be sufficiently high to initiate the second monomer. Thus, the monomers must be added in the order of increasing electron affinity: styrene < butadiene ~ isoprene < vinyl pyridine < methyl acrylate < ethylene oxide. One classic example of the synthesis of multiblock copolymers by sequential polymerization was reported by Hadjichristidis et al., as shown in Scheme 4A [50]. The resulting BCPs exhibited a very low PDI of 1.04 and the compositions observed from ¹H-NMR were consistent with the feeding ratios.

Nevertheless, sequential LAP is restricted especially when the monomers exhibit different reactivities. For example, the living polymer of polydimethylsiloxane is not sufficiently nucleophilic to initiate the polymerization of t-butyl methacrylate. In order to prepare PDMS-b-PtBA, a linking reagent bearing with chlorosilane moiety was used due to its high reactivity (Scheme 4B) [51]. Hirao et al. prepared a wide variety of block copolymers with extremely low PDI using new linking methodology. Another advantage of this linking chemistry is that it opens endless possibilities of various molecular architectures [5,42,52,53], as demonstrated by the work of Mays, Hadjichristidis, Hirao, Quirk, and many others through chlorosilane and 1,1-diphenylethylene (DPE)-based linking chemistry (Figure 3). Well-defined BCPs with structural and compositional homogeneity prepared using LAP have served as templates for structure-property relationship studies and have also found many important industrial applications such as adhesives and sealants, in automotive, wire, paving, and footwear applications, etc. [54].

2.2. Combination of Different Polymerization Techniques

In some cases, sequential addition polymerization techniques are limited when the monomers are not suitable to be polymerized using the same polymerization mechanism. A second approach to BCPs is the coupling of two different polymerization techniques or block copolymer segments either using a “click” reaction such as Diels–Alder cycloaddition reaction, a thiol-ene reaction, or a copper-catalyzed azide–alkyne cycloaddition reaction (CuAAC). With appropriate care, this approach can yield BCPs without substantial homopolymers and comparable PDIs as compared with the sequential addition approach. Yagci et al. reported a one-pot synthetic approach to BCPs using the combination of ATRP and ring opening polymerization (ROP) methods simultaneously [56]. Two structurally different monomers were selected and the polymerization was initiated simultaneously using a difunctional initiator, as shown in Scheme 5. They found that the two polymerization mechanisms proceeded without affecting each other. The obtained products showed characteristic thermal transitions of both PMMA and polycaprolactone (PCL) blocks, indicating that BCPs were synthesized.

The synthesis of complex polymeric materials through post-polymerization coupling reactions has also attracted significant research interest. Van Hest et al. synthesized PMMA-b-PEG diblock copolymers via the combination of ATRP and alkyne-azide click reaction, as shown in Figure 4. Terminal alkyne and azide moieties were conveniently introduced via protected functionalized initiators [57].

Hawker et al. synthesized a series of cyclic PS-b-PEO BCPs using the ATRP and CuAAC of α,ω-azide-functionalized PS and α,ω-alkyne PEO homopolymers (Figure 5) [11]. The α,ω-azide-functionalized PS was polymerized using a difunctional ATRP initiator, ethylene bis-(2-bromoisobutyrate). The α,ω-Br-terminated PS was then conveniently converted to the azide moiety. The α,ω-alkyne PEO homopolymer was obtained by treating α,ω-hydroxyl PEO with propargyl bromide. The desired products were isolated using preparative gel permeation chromatography. Tezuka et al. reviewed interesting work about topological studies of BCPs [58].

RAFT polymers can be conveniently converted to thiols by treating the chain end with aliphatic amines. A block copolymer can therefore be obtained through related thiol chemistry (Figure 6).

Barner-Kowollik and Stenzel et al. first synthesized a series of star polymers by a combination of RAFT chemistry and hetero Diels-Alder reaction [60], in which the reaction occurs between thiocarbonyl group and a diene in a [4 + 2] cycloaddition process (Scheme 6).

The incorporation of supramolecular chemistry also provides a very interesting pathway to form BCPs. Schubert et al. demonstrated that terpyridine-terminated PS and PEO can form BCPs with RuCl₃ [61]. The coordination junction point was formed in two steps: the terpyridine-terminated PEO was selectively complexed with RuCl₃ yielding monocomplex of Ru(III); terpyridine-terminated PS was then reacted with the monocomplex under the reduction of Ru(III) to Ru(II), resulting in PS-b-PEO. One advantage of this approach is that the non-covalent junction point can be broken under external stimuli.

3. Self-Assembly of BCPs

3.1. Self-Assembly in Bulk

Self-assembly of BCPs with immiscible blocks has been extensively studied via experiments and simulations since the 1960s [62]. Several reviews are available on this topic [22,63,64]. The self-assembly process is driven by an unfavorable mixing enthalpy coupled with a small mixing entropy. Composition (f), the number of repeating units (N), and the Flory-Huggins interaction parameter (χ) are the important parameters that determine the morphologies, which include spheres (S), cylinders (C), gyroids (G), and lamellae (L) [65,66]. The self-assembly behavior may also be influenced by other external parameters such as mechanical or electric fields [67,68]. However, it is beyond the scope of this review to cover all these aspects. Therefore, we only address the basic concepts of microphase separation and present several recent examples of the morphological behavior of BCPs. Figure 7 shows that the different morphologies of a typical linear diblock BCP evolve from spherical to lamellar and can undergo disorder to order transitions, as a function of f and χN.

Furthermore, the self-assembly of triblock copolymers of poly(styrene-b-isoprene-b-styrene) (PS-b-PI-b-PS) has also been extensively investigated using small-angle X-ray scattering (SAXS) and transmission electron microscopy (TEM) by Storey et al. and Elabd et al. [70,71]. Unlike the conventional study on neutral BCPs, Mays et al. compared the morphologies of well-defined neutral PS-b-PI and its charged counterpart, sulfonated PS-b-fluorinated PI (sPS-b-fPI). Interestingly, Inversed morphologies were observed when the diblock BCP, 25sPS-b-75fPI (50% of PS block was sulfonated to sulfonic form), was casted from tetrahydrofuran compared with its neutral counterpart, as shown in Figure 8 [72].

Segalman et al. prepared a series of ionic conducting BCPs to investigate the relationship between ionic conductivity and domain spacing [73]. A well-defined BCP precursor, PS-b-PI, was synthesized followed by a thiol-ene reaction to attach the ionic moieties to PI segments. The domain spacing was controlled by keeping the volume fraction of ionic moieties constant and varying the BCP chain length. It was revealed that the ionic conductivity was independent of domain spacing. The insights gained by this work could facilitate the development of design rules for the next generation of high performance ion-conducting membranes. Balsara et al. studied the effect of salt on the morphology of electron conducting poly(3-(2′-ethylhexyl)thiophene)-b-ethylene oxide) (P3EHT-b-PEO) [74]. PEO was selected due to its good salt solvation capability. Their results showed that, in the melt state, the salt-free sample exhibits a gyroid morphology while the salt containing sample (r = 0.125) exhibits a lamellar morphology. Furthermore, quenching the salt-free sample to room temperature results in a lamellar morphology due to the breakout of the P3EHT crystals. In contrast, there is little change in the morphology of the salt-containing sample upon quenching to room temperature. This could be attributed to the increase in χN.

Owing to its structural versatility and precise tunability of morphology, dimensionality, and feature size, BCP is an ideal platform for studying periodically ordered functional materials on the mesoscale. Wiesner et al. demonstrated the utilization of triblock copolymer (PI-b-PS-b-PEO) self-assembly to direct the synthesis of a mesoporous niobium nitride (NbN) superconductor (Figure 9). The formation of a three-dimensionally continuous gyroidal mesoporous NbN superconductor exhibits a critical temperature (T_c) of about 7 to 8 K, a flux exclusion of about 5% compared to a dense NbN solid [75]. More examples will be given in Section 4.

The myriad morphologies that are formed by BCPs are usually obtained by slow cooling so that the polymer chains have enough time to reach thermodynamically preferred alignments. Rather than using slow cooling, Kim et al. rapidly quenched poly(isoprene-b-lactide) diblock copolymers from the disordered state and revealed an extraordinary thermal history dependence [76]. Whereas conventional cooling results in the formation of documented morphologies, rapidly cooled samples that are then heated from low temperature form the hexagonal C14 and cubic C15 Laves phases commonly found in metal alloys. This unusual discovery reinforces fundamental analogies between the way metals and self-assembled soft materials break symmetry when subjected to changes in thermodynamic state variables that drive phase transitions.

3.2. Self-Assembly in Solution

The self-assembly of amphiphilic BCPs has been a popular topic in fields such as nano-cargo delivery, biomedical/pharmaceutics, and nanotechnology [77,78,79]. Although the self-assembly of amphiphilic BCPs is also driven by the minimization of free energy in the system, self-assembly in solution is more complicated than BCP self-assembly in bulk. The morphologies are primarily determined by the so-called packing parameter (p = v/a₀l_c), where v is the volume of the hydrophobic chain, a₀ is the optimal area of the head group, and l_c is the length of the hydrophobic tail (Figure 10) [80].

By controlling the factors that are able to change the packing parameter such as BCP composition and concentration, water content, common solvent, and additives, a wide range of morphologies have been reported including spherical micelles, cylindrical, bi-continuous, lamella, vesicles, tubules, etc. [81] (Figure 11).

Manners and Winnik reported a series of interesting rod-like micelles using poly(ferrocenyldimethylsilane) (PFDMS). These micelles are able to grow epitaxially by the addition of more polymer, leading to extended micelles with a very narrow size distribution (Figure 12) [82].

Self-assembly of BCPs enables researchers to construct nanoscale structures with molecular level precision. Not only the chemical composition has significant effect on the properties of nanostructure; molecular architecture also has profound effect, as revealed by the recent studies. Tezuka and Yamamoto et al. explored the effect of topology on the thermal stability of self-assembled structures of BCPs: linear poly(butyl acrylate-b-ethylene oxide-b-butyl acrylate) (PBA-b-PEO-b-PBA) (1) and the cyclic counterpart (2) (Figure 13). Despite no distinctive change in the chemical composition or structure of the micelles, they found that the cloud point (T_c) was increased by more than 40 °C through the topological conversion. Moreover, the T_c can be tuned by the changing the ratio of 1 and 2 [83].

Using poly(1,2-butadiene-b-ehtylene oxide) (PB-b-PEO) in water media, Bates et al. observed the formation of “Y-junctions”, which further assembled to form a 3D network (Figure 14). This observation is in contrast to that of the low molecular weight counterpart, where only spherical micelles, wormlike micelles, and vesicles were obtained when mixed with water [84].

4. Applications

4.1. Applications as Thermoplastic Elastomers

One of the most important technical applications of BCPs is as thermoplastic elastomers (TPEs). This type of BCP typically contains physically crosslinked rigid glassy domains and a continuous soft rubbery domain. It offers the elasticity of conventional rubber and, since it is not chemically crosslinked, is suitable for typical plastic processes such as injection molding and melt extrusion. TPEs have found applications in adhesives, coatings, food packaging, and many other areas [85]. The most commonly studied TPEs are poly(styrene-b-isoprene-b-styrene) (SIS) and poly(styrene-b-butadiene-b-styrene) (SBS). These two types of BCPs were firstly developed by Holden and Milkovich through living anionic polymerization in the early 1960s [86]. The remarkable mechanical properties are attributed to the microphase separation of the two chemically incompatible blocks, PS and PI (or PB). SIS with high 3,4-isoprene exhibits a broad glass transition close to room temperature, providing a very good vibration damping material at room temperature [54]. Polyethylene (PE) or poly(ethylene-alt-propylene) can be obtained through the hydrogenation of the diene block; such products exhibit superior stability under light and oxygen. Hydrogenated SBS has been used in applications such as electrical wire and cable sheathing due to its cold temperature flexibility and higher extendibility to incorporate flame retardants over conventional poly(vinyl chloride) (PVC) [54]. Star-block copolymers of butadiene and styrene have found applications in adhesives and sealants [87].

Because the precise control over composition and architecture is possible through LAP, the resulting BCPs can be tailored to a wide variety of applications [88]. One research focus is to improve the upper service temperature of styrenic thermoplastic elastomers (S-TPEs), which is mainly controlled by the T_g of the hard domain. Considerable research has focused on developing high T_g glassy domains such as poly(α-methyl styrene) (T_g ~ 173 °C) [89], poly(α-methyl p-methyl styrene) (T_g ~ 183 °C) [90], and poly(tert-butyl styrene) (PtBS, T_g ~ 130 °C) [91]. In hydrocarbon solvent at room temperature, Mays et al. prepared a series of BCPs consisting of poly(benzofulvene (BF)-b-isoprene-b-benzofulvene) (FIF), in which FIF with 14 vol % of PBF exhibited a maximum stress of 14.3 ± 1.3 MPa and strain at break of 1390 ± 66% from tensile tests [92]. Dynamic mechanical analysis showed that the softening temperature of PBF in FIF was 145 °C, which is much higher than that of thermoplastic elastomers with PS hard blocks (Figure 15).

In comparison with diene-based TPEs, all acrylic-based TPEs showed higher oxidation stability and UV resistance. The same group synthesized a series of TP all acrylic-based high temperature thermoplastic elastomers containing poly(1-adamatyl acrylate) as a hard domain and poly(tetrahydrofurfuryl acrylate) as a soft domain by RAFT polymerization [93]. These TPEs exhibited superior stress-strain behavior compared to that of conventional all acrylic-based TPEs consisting of PMMA and PBA made by controlled radical processes [94], while the tensile strength was lower compared to that of similar products prepared via living anionic polymerization [95].

TPEs with non-linear architectures have also been synthesized and investigated over the last three decades. Multigraft copolymers of PI-g-PS are found to have high tensile strength, high strain at break, and low residue strain. Furthermore, Mays et al. synthesized a series of tetrafunctional multigraft copolymer (“centipede”) with varied branching points (Figure 16) [96]. The tetrafunctional copolymer with 10 branching points showed an exceptional elongation at break ~2100%, almost double that of commercial Kraton^® (1050% at similar composition), and only 40% residual strain on hysteresis experiments (elongated at 1400%) [97]. Detailed study revealed that the superelasticity was attributed to the distinct molecular architectures. Furthermore, the same group developed a low-cost emulsion polymerization to prepare comb multigraft copolymers such as PI-g-PS and PBA-g-PS as well as centipede PBA-g-PS for TPE applications [98,99].

4.2. Applications in Drug Delivery and Release

Advanced nanoscale systems created by self-assembly of BCPs have seen tremendous progress for drug delivery applications over the last decade [100,101]. The advances in BCP self-assembly offer effective control of morphology, surface chemistry, and the introduction of environmental responsiveness. Among these nanostructures, micelles and vesicles are the most studied morphologies. Block copolymer micelles are of interest for drug delivery applications for a number of reasons. First of all, hydrophobic drugs can be entrapped in the core and transported at concentrations that can exceed their intrinsic water solubility. Secondly, the hydrophilic blocks, which are often composed of PEO, can form hydrogen bonds with the aqueous surroundings and form a tight shell around the micellar core. Moreover, the PEO corona resists protein adsorption and cellular adhesion, protecting the hydrophobic drug against hydrolysis and enzymatic degradation. In addition, the PEO corona prevents recognition by the reticuloendothelial system and therefore increases the blood circulation times [100]. Thermo-responsive micelles are the most widely investigated, such as poly(N-isopropyl acrylamide-b-d,l-lactide), poly(N-isopropyl acrylamide-b-butyl methacrylate), and poly(N-isopropyl acrylamide-co-N,N-dimethylacrylamide)-b-poly(d,l-lactide-co-glycolide) [102,103]. These micelles undergo reversible structural changes that facilitate drug release once the temperature is elevated above the LCST of the polymers [104]. Other stimuli-responsive BCPs have also been extensively studied. Park et al. developed an amphiphilic BCP composed of PEG and poly(ε-(4-nitro)benzyloxycarbonyl-l-lysine) for hypoxia-sensitive drug delivery. The formed micelles encapsulated doxorubicin (DOX) in an aqueous condition and exhibited rapid intracellular release of DOX under the hypoxic condition [105].

BCPs used in drug delivery should have low toxicity or be non-toxic. BCP vesicles containing biocompatible PEO, PLA, and PCL have been evaluated [106,107]. Both PEO-b-PLA and PEO-b-PCL vesicles exhibit the controlled release of active dyes and anti-cancer drugs over periods of up to two weeks [108,109]. Discher et al. employed vesicles made from a mixture of biocompatible PB-b-PEO and PLA-b-PEO to simultaneously deliver two anti-cancer drugs, in which hydrophilic doxorubicin was encapsulated in the cavity and hydrophobic paclitaxel was embedded in the vesicle wall [106,110]. In vivo studies revealed that the vesicle-loaded drugs are more effective and sustainable in tumor shrinkage as compared with the direct injection of free drugs. Rates of encapsulant release from the hydrolyzable vesicles are accelerated with an increased proportion of PEG and the molar ratio of degradable copolymer. In particular, BCP vesicles modified with stimuli-responsive functionalities have been studied for smart drug delivery. Attaching the vesicle with antibodies allows the delivery of vesicles to targeted cells [111]. Other than their academic interest, thermosensitive liposomes (TSLs) have already been investigated for the treatment of breast cancer and colorectal liver metastasis in several clinical trials.

In spite of the advances in the design of peptide drugs and the emergence of gene therapy, there is still a need to further improve the performance of these systems that can precisely direct the drug to the desired site in the body and to accurately control the rate at which the drug is released.

4.3. Applications in Soft Lithography

One of the remarkable candidates for soft lithography is the directed self-assembly (DSA) of BCPs, because they can form ordered features at a length scale as low as a few nanometers, which is required for many of the most demanding next generation patterning applications, including the fabrication of bit patterned media (BPM) for hard disk drives as well as fin field effect transistor (FinFETs) and contact holes for microelectronics [112,113,114,115,116,117,118]. Poly(styrene-b-methyl methacrylate) (PS-b-PMMA) is the most studied BCP system because it can produce lamellar or cylinder morphologies, which are perpendicularly oriented to the substrate simply via thermal treatment (Figure 17) [119]. This phenomenon is attributed to the similarity of the interfacial energies between each block and the air interface [120]. Hawker et al. reported that the neutral surface (γPS-Air ≈ γPMMA-Air) could be achieved with thermally annealed PS-b-PMMA at 225 °C in air [121].

Unfortunately, the minimum feature size is limited to ~13 nm due to its low χ (~0.039 at 150 °C) [122]. To achieve very high resolution (i.e., small feature size), the chemical structure of the BCP must be carefully designed to maximize the chemical incompatibility between blocks (i.e., high-χ, low-N). Impressive work has been reported employing a variety of high-χ BCPs. Russell et al. were able to achieve a 3-nm cylindrical domain size using PS-b-PEO on a surface of a sapphire single-crystal (Figure 18) [123].

Hillmyer et al. synthesized poly(lactide-b-dimethylsiloxane-lactide) (PLA-PDMS-PLA) to achieve a 7-nm pitch size [124]. Some other BCPs that are capable of self-assembling into sub-10 nm domain sizes include PS-b-PDMS [125], PS-b-P2VP [126], and PS-b-PLA [118]. On the other hand, high-χ organic BCPs that can form sub-10 nm domains suffer from low etch contrast when sizes approach down to 10 nm. Since organic polymers that contain inorganic constituents, such as silicon, are inherently etch resistant, the incorporation of one such block in high-χ BCPs affords exceptional etch contrast. Cushen et al. synthesized a series of high-χ BCPs composed of oligosaccharides coupled to a silicon-containing polystyrene derivative. The BCPs exhibit hexagonally packed cylinders with a 5-nm feature size in thin film [127]. A polarity-switchable top layer was also applied to the surface of silicon-containing BCPs films to obtain perpendicular orientation of poly(styrene-b-trimethyl silyl styrene-b-styrene) (PS-b-PTMSS-b-PS) and poly(trimethyl silyl styrene-b-lactide) (PTMSS-b-PLA) after thermal annealing [128].

On the other hand, BCP architecture also plays a key role in controlling the feature size. Hawker compared linear PS-b-PEO and cyclic PS-b-PEO and found that the thin film self-assembly of the latter showed a ~30% decrease in domain spacing (Figure 19), which was attributed to the reduced hydrodynamic radii of the cyclic systems [11].

Limited synthetic access has largely restricted the application of cyclic BCPs in soft lithography. Other complex architecture such as star copolymers, etc. remains a largely unexplored regime. Nevertheless, the advancement of BCP design, synthesis, processing strategies, and morphological characterization tools (e.g., atomic force microscopy (AFM), tomographic TEM, grazing-incidence small-angle scattering (GISAXS), etc.) offers hope that this technology may impact industry and society in the coming years.

4.4. Applications in Mesoporous Materials

Research interest in ordered porous materials originated from the successful synthesis of porous silicates in the 1990s and soon spread to a variety of framework compositions including metal oxides, non-oxide inorganics, and carbon [129]. This group of materials exhibits periodically aligned structures and uniform cavities with sizes ranging from micro- (<2 nm), to meso- (2–50 nm), to macropores (>50 nm) [130], which lead to very high surface areas (up to 1500 m²/g) [131]. With these unique features, porous materials present great value for applications in energy conversion and storage [132,133,134,135], catalysis [136], drug delivery [137], gas capture [138,139], and water purification [140,141].

As mentioned in the previous section, the self-assembly of BCPs may induce a series of interesting 2D or 3D morphologies, which make them very useful as structure-directing “templates” for the fabrication of porous materials. Generally, the formation of porous materials through polymer templating involves the following steps: (1) Cooperative self-assembly of amphiphilic BCPs in the presence of a precursor for the framework (such as hydrolyzed silicate, metal oxides, or carbon source) where the precursor is directed to the hydrophilic domain through hydrogen bonding, ion pairing or hydrophilic interactions; (2) solidification/condensation/crosslinking/curing of the precursor to form the framework of the porous material; and (3) removal of the BCP template by solvent washing or thermal decomposition leaving spaces for pores. However, the specific procedures vary based on the final shape of the porous material, namely membranes, spheres, or monoliths, and detailed examples will be given below.

The most commonly used copolymer template are the commercial Pluronic surfactants, a group of poly(ethylene glycol-b-propylene glycol-b-ethylene glycol) (PEG-b-PPG-b-PEG) with limited options for lengths of polymer blocks. Alternatively, amphiphilic BCPs with wider choices for composition and molecular weight of the polymer segments are readily designed and synthesized in the laboratory through controlled/“living” polymerization techniques [142]. These synthesized BCPs usually contain poly(2-vinylpyridine) (P2VP), poly(4-vinylpyridine) (P4VP), or poly(ethylene oxide) (PEO) as the hydrophilic block, and PS, PMMA, PI, polyisobutylene (PIB), and polyacrylonitrile (PAN) as the hydrophobic block [143,144]. In addition to high versatility, BCPs synthesized in the laboratory show other merits as compared to commercial Pluronic copolymers. The contrast in hydrophilicity between the two selected blocks are usually higher than for the PEG and PPG pair in the Pluronics, which is beneficial for the self-assembly of polymer micelles in aqueous solution. Moreover, hydrophobic blocks such as PS and PMMA exhibit much higher glass transition temperatures than PPG, resulting in better stability of the self-assembled structures. In the case of copolymers containing PS and PAN, the hydrophobic segments can be converted to carbon residue in high yield in the carbonization step, which provides additional mechanical support to the mesoporous framework [145].

Evaporation-induced self-assembly (EISA) is widely used in the preparation of ordered mesoporous films and particles, which is especially important for oxides. In a typical EISA process, the cooperative self-assembly of a precursor with an amphiphilic template is induced by a concentration gradient of an organic solvent such as tetrahydrofuran (THF), chloroform, or dioxane. Different from a general solvent-annealing process, precursor molecules are crosslinked in EISA as the solvent evaporates, leaving no chance for structure refinement after it is formed [146,147]. Wiesner et al. synthesized mesostructured silicate/copolymer composite film through EISA with PS-b-PEO as a structure-directing agent in a mixture of THF and water [148]. A large variety of mesoporous metal oxides in the form of powder were synthesized through EISA using commercial Pluronics as a template [149]. In addition, mesoporous carbon films were prepared via EISA from a thermosetting precursor (phenol-formaldehyde system, resol, resorcinol, etc.) using a variety of BCP templates such as PS-b-P4VP, PEO-b-PS, PEO-b-PMMA, and PEO-b-PMMA-b-PS [143]. Recently, Wei and coworkers reported the synthesis of nitrogen-doped mesoporous carbon in the form of powder through EISA, using Pluoric copolymer as the template, with resol and dicyandiamide as sources for C and N, respectively [147]. For clarification, some research groups do not consider systems involving formaldehyde as an actual EISA process, since the crosslinking of precursors occurs after the cooperative self-assembly, instead of simultaneously, as in many EISA processes [146].

Following the traditional EISA method, an evaporation-induced aggregation assembly (EIAA) mechanism, was proposed [150]. This solution precipitation method can produce mesoporous materials with various shapes including spheres, fibers, and polyhedrons in the form of powder [151]. Moreover, the EIAA method tolerates a greater portion of water compared to EISA, which makes it more suitable for scaling up. In detail, the water-insoluble template PEO-b-PMMA and silica precursor were first dissolved in a mixture of THF and water. As THF evaporated, the BCP and silicate oligomers were driven to form composite micelles with silicates located at the shell. With further removal of the solvent, the micelles assembled into mesostructured particles at the liquid-liquid interface and precipitated from the solution. Subsequent calcination of the precipitants completely removed the PMMA block to form pores. A schematic illustration of EIAA is shown in Figure 20.

More recently, Tang et al. proposed a facile micelle method to synthesize N-doped mesoporous carbon spheres without the EISA process (Figure 21). The key difference of this research from EISA lies in the formation of stable micelles of dopamine (both C and N sources) using a high molecular weight PS-b-PEO as a template. After the micelle was formed, dopamine was polymerized using ammonia as an initiator. These “frozen” micelles joined together forming into larger spheres with sizes around 200 nm. These composite spheres were easily separated from the solution through centrifugation and allowed for further carbonization to remove the template and fabricate mesoporous spheres [152].

Many studies demonstrated the advantage of mesopores over micropores in the mass transport process in a variety of applications [135,153]. Larger pore size facilitates a more efficient diffusion of guest molecules within the pores [136], and also allows the loading of metal catalyst particles [154,155]. Thick pore walls are also favorable because of the possibility for further modification on the pore surface [151,155].

Since the pores are produced from the calcination of the hydrophobic segment of the BCP, high molecular weight of the sacrificial block is highly desired. Unfortunately, the mesoporous materials templated from commercial Pluronic copolymers show a pore size limited to 12 nm and a wall thickness no more than 6 nm [151]. BCPs with higher molecular weight in the sacrificial segment (higher hydrophobic volume in the copolymer) have been extensively utilized for larger pores in the resulting porous material. Using PEO-b-PS (M_w = 29,700 g/mol) as the copolymer template, adding a special hydrothermal treatment step, Deng et al. achieved mesoporous silica with pore diameters as large as 30.8 nm [156]. By varying the lengths of each segment of BCP templates, Tang et al. observed a strong correlation of the resulting pore size with the length of BCP, where PS₃₇-b-PEO₁₁₄, PS₁₇₈-b-PEO₈₈₆, and PS₁₇₃-b-PEO₁₇₀ resulted in pore sizes of 5.4 nm, 16 nm and 16 nm, respectively [152]. In another case, aluminum-organophosphonate films with tunable macropores (30–200 nm) were synthesized via ultra-high molecular weight PEO-b-PS (250,000 g/mol) templates, though the resulting material exhibited a relatively disordered mesostructure [157]. In addition to utilizing high molecular weight BCPs as templates to target material with larger pores, an additional hydrophobic component can be added to the mixture as a pore swelling agent. For example, with the aid of trimethylbenzene (TMB), the pore size of silica can be expanded to 30 nm [158]. Moreover, Deng et al. reported the use of PS homopolymer as the pore expanding agent in the synthesis of mesoporous carbon, when PS-b-PEO is selected as the template [159]. However, care must be taken when using the pore expanding agent to the composite in order to avoid the risk of unstable mesostructures [160].

The arrangement of a mesostructured material is closely related to the microphase separation of the BCP template. Interestingly, simply varying the amount of the precursor with respect to the copolymer would also lead to morphological changes in the composite system [148]. Garcia and coworkers fully explored the phase diagram of a polymer/alumina/silane composite system aimed for the synthesis of porous aluminosilicate using PEO-b-PI as a template. As illustrated in Figure 22, eight morphologies of the composite system were obtained by adding different amounts of the inorganic precursor into BCP templates [161]. Many interesting crystal symmetries could be achieved through fine tuning the template compositions, and examples are substantially reviewed [151].

5. Conclusions and Perspectives

In this review, we have discussed the robust synthetic approaches to well-defined block copolymers with controlled compositions and architectures, in which the sequential addition of controlled/living polymerization techniques and the combination of controlled polymerization techniques with facile coupling chemistry have expanded the spectrum of available BCPs. One of the most fascinating features of BCPs is their ability to self-assemble into nanoscale ordered structures with various morphologies such as spheres, cylinders, bicontinuous structures, lamellae, etc. Detailed study both from experiments and theoretical modeling has revealed that the nature of morphologies is determined by factors including interaction parameter (χ), degree of polymerization (N), and the volume fraction (f), as well as molecular architecture. Due to these unique features, BCPs have attracted massive attention as thermoplastic elastomers, as drug delivery systems, in soft lithography, as mesoporous materials, and in many other areas.

These fascinating BCPs with chemical and topological diversity bring new challenges for the establishment of structure-property relationships and novel applications. For instance, BCP materials including rod–coil diblocks and bottlebrush materials should be investigated to create more elaborate nanostructures. Recent advancements in BCP patterning have pushed the limits of feature sizes to both smaller (<5 nm) and larger scales (>50 nm). We expect that these new challenges can be coped with by precisely tunable interactions between blocks in terms of range and strength, which further relies on their precise synthesis and guidance from theory and simulation. Strategical placement of ionic moieties might offer a promising approach. Although significant work has been reported, more interdisciplinary collaborations are still needed in order to advance the synthesis, structural and dynamic characterization, theory and simulation, and to gain a clear picture of BCP structure-property relationships.