Introduction

Block copolymers (BCPs) are a versatile class of soft matter, which consist of at least two covalently linked polymer blocks[1]. Due to the inherent incompatibility of the blocks that arises from differences in their chemical and physical properties, BCPs find widespread use in research and applications, including nanomedicine[2,3,4,5,6], catalysis[7, 8], compatibilization[9, 10], formation of mesoporous nanostructures[11,12,13,14,15] or energy storage[16, 17], to name a few examples. Irrespective of the intended use, control over the self-assembly or microphase behavior of BCPs is of vital importance for the quality and homogeneity of the final structure. While BCPs were primarily assembled in solution[18,19,20,21,22,23] or in bulk and at interfaces[24,25,26,27,28,29,30,31,32], their self-assembly in the confinement of emulsion droplets only recently gained traction[33,34,35,36,37,38,39,40,41,42,43]. There, the BCP is typically dissolved in an organic solvent and emulsified with an aqueous surfactant solution to create a (3D) spherical confinement for the BCP. During evaporation of the organic solvent, the droplet shrinks, which concentrates the BCP, and ultimately triggers the nucleation and growth of the BCP morphology. The resulting solid multicompartment microparticles (MMs) thereby develop characteristic shapes and internal morphologies dictated by the BCP composition. In addition, parameters such as the rate of solvent evaporation[34, 44] and the affinity of the surfactant for certain blocks[36, 41, 45,46,47,48,49] can influence the orientation of the morphology and hence the shape of the MMs. This process allowed to produce e.g. hybrid particles[50], mesoporous structures[51,52,53], and photonic pigments[54,55,56].

While the morphological behavior of AB diblock copolymers in confinement is rather well understood, research on ABC triblock terpolymers is still comparably limited. The increased number of blocks and interaction parameters substantially increase the number of possible morphologies[1, 57]. The wealth of achievable morphologies can be mapped into ternary microphase diagrams, where subtle changes in block volume fraction can induce striking morphological transitions. Despite the appeal of largely increased morphological complexity, the introduction of a third block (and tuning its length) can be accompanied by time-consuming and complicated synthesis. Instead of synthesizing libraries of ABC triblock terpolymers, blending with homopolymers (hP) is an efficient and fast way to create large libraries of morphologies. If the hP has a comparatively low molecular weight, it swells the corresponding domain and increases its volume fraction (wet brush regime)[27, 58,59,60,61]. In contrast, blending with a hP of similar or higher molecular weight to the corresponding block prevents mixing and leads to phase separation of the hP instead (dry brush regime)[62,63,64]. The concept of blending was demonstrated for ABC triblock terpolymers in bulk as well as for AB diblock copolymers in confinement with the goal to obtain a certain structure or to study the microphase behavior[27, 60, 61, 63,64,65]. For blending of ABC triblock terpolymers in confinement, much less is known about targeting or altering specific morphologies.

In this work, we study the microphase behavior of polystyrene-block-polybutadiene-block-poly(methyl methacrylate) (SBM) in confinement for 6 block compositions, and systematically blend hPS, hPB, or hPM into the respective morphologies to study the effect of added volume on morphology. For emulsification, we use a Shirasu Porous Glass (SPG) membrane setup, leading to near-monodisperse MMs with controlled size. The overall shape and inner morphology of the MMs are analyzed by a combination of transmission and scanning electron microscopy (TEM, SEM). We show the possibility of replicating specific terpolymer morphologies, determine blending limits (e.g. excessive blending), and demonstrate the potential by transforming the morphology of one SBM into multiple other morphologies.

Materials and Methods

Materials. Analytical grade solvents and chemicals were used as received without further purification unless stated otherwise. Sodium dodecyl sulfate (SDS, > 99%), styrene, methyl methacrylate, 2-cyano-2-propyl dodecyl trithiocarbonate (CPDTC), 2,2’-azobis(2-methylpropionitrile) (azobisisobutyronitrile, AIBN) and polybutadiene homopolymer (hPB, Mn = 2.0 kg∙mol−1) were purchased from Sigma-Aldrich. The monomers were purified by running them over a silica column before use. OsO4 was obtained from Science Services (OsO4, 4 wt.% in H2O) and ultrapure water from a Milli-Q® Integral Water Purification System. Regenerated cellulose tubes with a molecular weight cut-off of 12–14 kDa and an average flat width of 33 mm (Sigma Aldrich) were used for dialysis. The polystyrene-block-polybutadiene-block-poly(methyl methacrylate) triblock terpolymers (PS-b-PB-b-PM or SBM) were synthesized by sequential anionic polymerization as described previously[66].

Synthesis of PS and PM homopolymers. Both hP were synthesized with reversible addition-fragmentation chain transfer (RAFT) polymerization to control molecular weight, which was kept below the block lengths of the corresponding SBMs. For the synthesis of hPM, methyl methacrylate (6.2 mL, 58.4 mmol, 39 eq.) and CPDTC (505 mg, 1.5 mmol, 1 eq.) were dissolved in dioxane (30 mL) and a stock solution of AIBN in dioxane (1 g∙L−1) was prepared. Both solutions were bubbled with Argon for 20 min after which 24 µL of the AIBN solution (0.1 eq.) were added to the reaction vessel. The reaction mixture was stirred under Argon at 80 °C for 2 h before it was rapidly cooled with liquid nitrogen and exposed to air to stop the reaction. The polymer was precipitated in n-hexane. Residual solvent was removed under reduced pressure to obtain 1.8 g of hPM (2.7 kg∙mol−1, Đ = 1.26). For the synthesis of hPS, styrene (16.8 mL, 146.9 mmol, 92 eq.) and CPDTC (642 mg, 1.6 mmol, 1 eq.) were dissolved in anisole (10 mL) and a stock solution of AIBN in anisole (1 g∙L−1) was prepared. Both solutions were bubbled with Argon for 20 min after which 80 µL of the AIBN solution (0.3 eq.) were added to the reaction vessel. The reaction mixture was stirred under Argon at 80 °C for 2.75 h and then rapidly cooled with liquid nitrogen and exposed to ambient air to stop the reaction. The polymer was precipitated in iso-propanol. Residual solvent was removed under reduced pressure to obtain 4.8 g of hPS (3.0 kg∙mol−1, Đ = 1.10).

Fabrication of (blended) polymer particles with SPG-setup. The hP and SBM copolymers were separately dissolved in chloroform (CHCl3) to prepare stock solutions (chP = 20 g∙L−1 and cSBM = 10 g∙L−1) that were mixed in varying ratios to achieve desired blend compositions. The specifics of SBM and SBM/hP are listed in Table 1 and a more detailed description on how these compositions were calculated can be found online in the Supplementary Information (SI). In a typical experiment, 1 mL of SBM or SBM/hP solution was emulsified in 20 mL of an aqueous SDS solution with a concentration of 5 g∙L−1. For that, the polymer solution was pushed through an SPG membrane with a pore diameter of 600 nm using pressurized Argon. The emulsion was kept stirring at 250 rpm to yield droplets with a homogeneous size distribution. After stirring for five additional days, the organic solvent had evaporated, which resulted in the formation of solid MMs with inner structure. The MM suspension (c = 0.5 g∙L−1) was dialyzed against ultrapure water to remove excess surfactant.

For SBM terpolymers, the volume fraction of the respective blocks is given in percent in the subscripts as φABC, while the superscripts provide the number average molecular weight, Mn, in kg∙mol−1, determined by size exclusion chromatography (SEC). Blend compositions follow a similar definition, i.e., subscripts refer to the final volume fraction φABC in percent as a combination of the original φABC plus added φhP. An overview of the SBMs and the SBM/hP used in this work is given in Table 1.

Table 1 SBM or SBM/hP used in this work
Full size table

Transmission electron microscopy (TEM). To prepare TEM samples, MMs were first stained with Os by placing 1.2 mL of MM suspension in an open vial, which was then put in a chamber together with another open vial containing OsO4 (4 wt.% in H2O). The chamber was closed, and the liquids were kept under stirring for 3 h after which the chamber was opened. To remove excessive Os, the MMs were cleaned by centrifugation and redispersion in ultrapure water. A drop of the MM suspension (c = 0.5 g∙L−1) was placed on a carbon-coated copper grid (400 mesh, Science Services) and excess liquid was blotted after 60 s using a filter paper. MMs were analyzed on a Talos L 120C (Thermo Fisher Scientific) with an acceleration voltage of 120 kV and an LaBF6-filament. Images were taken with a Ceta-F camera and Velox Software (Version 3.8.80). The ImageJ open-source software package (Version 1.53 k) was used for processing the data[67]. Compartment sizes were measured and averaged over at least 50 different locations.

Scanning electron microscopy (SEM). A cryo-field emission SEM (Zeiss Cross Beam 340) equipped with an energy-selective detector for 16-bit image series acquisition with up to 40,000 × 50,000-pixel resolution and in lens chamber was used for SEM imaging. Samples for SEM measurements were prepared by putting one drop of an approximately 0.5 g·L−1 MM dispersion on a silicon wafer and dried for at least 4 h. The samples were then sputtered with 4 nm Au using a Quorum PP3010T-Cryo chamber with integrated Q150T-Es high-end sputter coater.

Size exclusion chromatography (SEC). Information about number-average molecular weight (Mn) and dispersity (Đ = Mw/Mn) were obtained by SEC. The polymer was dissolved in THF at a concentration of 1.5 g·L−1 and the solution was then filtered through a PTFE syringe filter (pore size of 0.2 µm) prior to being measured on a 1260 Infinity Instrument (PSS/Agilent, Mainz). The device was equipped with an isocratic pump, SDV PSS columns with porosities ranging from 102 – 106 Å, a differential refractometer, and a UV–Vis multiwavelength detector. For the synthesized hPS, PS standards were used for calibration (PSS/Agilent, Mainz) with molecular weights ranging from 1000 to 1 000 000 g·mol–1 and narrow size distributions. For the synthesized hPM, PM standards were used in the same Mn-range.

Results and Discussion

Fabrication of MMs. All MMs (SBM or SBM/hP) were fabricated according to the following procedure. First, polymers were separately dissolved in chloroform (CHCl3) at concentrations of cSBM = 10 g∙L−1 or chP = 20 g∙L−1. Depending on the desired blend composition, the SBM solution was mixed with a predetermined amount of hP solution. The mixture was then emulsified with an aqueous SDS solution (cSDS = 5 g∙L−1) using the SPG membrane setup with a pore diameter of 600 nm. The organic solution was pushed through the membrane using pressurized Argon and the emulsion droplets were sheared off of the membrane by stirring. CHCl3 evaporated over the course of several days under continuous stirring, which led to the formation of solid MMs. The blend compositions were chosen to move between areas of the ternary microphase diagram with known morphologies (unblended SBMs) to verify whether hP blending is able to replicate the morphology of one SBM by blending another. Of the investigated SBM terpolymers or SBM/hP blends, we were able to induce morphology transitions from lamella-lamella (ll) to lamella-perforated lamella (lpl), from lpl to lamella-ring (lr), from lr to cylinder-on-cylinder (coc), and from coc to ll, lpl, and lr. The morphologies will be discussed in more detail in the individual sections below. An overview of the SBM and SBM/hP morphologies in confinement can be found in Fig. 1.

Fig. 1
figure 1

Ternary microphase diagram of SBM and SBM/hP blends. The different morphologies are color coded. SBM morphologies are represented by empty diamonds, while SBM/hP blends are shown as filled triangles. The color of the triangle represents the homopolymer that was used for blending: green = hPS, black = hPB and orange = hPM

Full size image

Morphological transition through hPS blending. Here and in the following, we will move through the ternary microphase diagram shown in Fig. 1, starting in the cyan area (ll-morphology) and increase the hPS content towards the red area (coc-morphology), i.e., we first transition from cyan to purple, then to dark blue and finally to red. We start by adding 55 vol% hPS relative to the PS block of S32B40M28125 equaling to an increase of φPS = 0.18 to reach a final composition of S50B29M21 (Fig. 2a). According to DLS, all produced MMs show a monomodal and narrow size distribution of around dh \(\approx\) 600 nm irrespective of blending (Figure S1). The addition of hPS (or any hP) does not negatively affect the stability of the emulsion droplets or disturb the microphase separation during MM solidification. The unblended S32B40M28125 originally formed spherical MMs with a concentric ll-morphology. The surface appears smooth (Fig. 2b) as PM forms the outermost shell, followed by a dark PB lamella (stained with OsO4) and a thicker, gray PS lamella. This pattern (MBSSBM) alternates towards the center of the MM. This structure is expected as all blocks have similar volume fractions. After blending with hPS to S50B29M21, we still find a concentric arrangement and a spherical MM (Fig. 2c), but the PB domain changed its appearance from a continuous lamella (dark line) to a discontinuous pattern (dark dotted line). The structural transition is induced as the hPS accumulates in the PS domains, resulting in a distinct swelling of the PS lamellae, as evidenced by the increase of lamella width (w) from wlam, PS ≈ 23 nm to wlam, PS ≈ 29 nm. The incorporation of the hPS also leads to a relative decrease of φPB from 0.40 to 0.30, which lies below the ll-lpl transition value of around φPB, ll↔lpl ≈ 0.36, so that PB forms a perforated lamella, yielding an lpl-morphology.[25, 60] Comparing this morphology with the unblended S50B30M20159 with similar volume fractions of the blocks, we also find the lpl-morphology (Fig. 2d)[60] with identical features as compared to the blended S50B29M21. The main differences are the larger domain sizes that originate from the larger molecular weight of the unblended S50B30M20159. Blending thus allows to migrate between different areas of the ternary microphase diagram.

Fig. 2
figure 2

Morphological transition of the ll-to lpl-morphology by blending of S32B40M28125 with 55 vol% hPS. a) Ternary microphase diagram with blending pathway indicated by the arrow. b) TEM image of S32B40M28125 MMs with ll-morphology (SEM in inset). c) TEM image of S50B29M21 MMs with lpl-morphology after blending (SEM in inset). d) The lpl-morphology of S50B30M20159. PB was stained with OsO4 and appears darkest. Scale bars are 200 nm

Full size image

Blending does, however, has its limits as demonstrated by the addition of the relatively large 150 vol% hPB relative to the PB block of S74B12M1489, equaling to an increase of φPB = 0.17 to obtain S60B29M11 (Figure S2a). There, we observed phase separation of the added hPB instead of blending. The morphology of S74B12M1489 can be ascribed to a cylinder-on-cylinder morphology within a PS matrix, as previously found for SBT triblock terpolymers with similar block volume fractions [43]. The blended S60B29M11 retains the morphology, despite the increase of φPB = 0.12 to φPB = 0.29, mainly because the hPB accumulates at the tip of the MMs (Figure S2b). The hPB compartment can be clearly seen in TEM as dark collapsed areas, but also in the SEM as flattened or dimpled areas. This behavior was unexpected, as the M ≈ 2 kg∙mol−1 of the homopolymer is 5 times smaller than the Mn, PB ≈ 10 kg∙mol−1 of the PB block, and we therefor should be working in the wet brush regime [61, 68]. One explanation might be that the PB block has an overall small φPB = 0.12 and the microdomain may not be able to take up larger amounts of additional hPB before being saturated, i.e. the PB blocks are stretched to a critical threshold and unable to accommodate further hPB.

Knowing that larger volume fractions of hP will lead to phase separation, we continue our path in the ternary microphase diagram from the lpl-morphology (purple area) into the lr-morphology (dark blue area) by feeding PS into an unblended SBM (Fig. 3a). S41B25M34143 develops an lpl-morphology in elliptic MMs with axially stacked orientation (Fig. 3b). In the axially stacked case, PS/PM lamellae are clearly visible as linear stripes (discs), whereas the perforated lamellae of PB can be identified by a combination of dark dots and diagonal stripes. Compared to the spherical shape that we usually find for the MMs, the ellipsoidal shape and the axial stacking of the lamellae are likely caused by an interplay of two competing contributions. While the preferred interaction of the surfactant (here: SDS) with one of the blocks (here: the PM block) typically favors the spherical arrangement [36], a relatively high Mn (as in S41B25M34143) entropically impedes bending of the lamellae and therefor favors a flat, axially stacked arrangement. Blending S41B25M34143 with 50 vol% hPS relative to the PS block is equal to an increase in φPS = 0.20 to reach a final composition of S61B16M23. This change in composition triggers a morphological transition of the PB microdomain into the lr-morphology. The thickness of the PS lamellae did not noticeably increase through blending (from 40 to 41 nm), but the lamella thickness is more homogeneous throughout the MM. The addition of low Mn hPS also appeared to cause a relaxation of the morphology, because the PS lamellae are more planar across the MM diameter instead of being bent, and the MM surface appears smoother than before in SEM and TEM.

Fig. 3
figure 3

Morphological transition of lpl- to lr-morphology by blending S41B25M34143 with 50 vol% hPS. a) Ternary microphase diagram with blending pathway indicated by the arrow. TEM and SEM (inset) images of b) unblended S41B25M34143 and c) S61B16M23 blended with 50 vol% hPS. PB was stained with OsO4 and appears darkest. Scale bars are 200 nm

Full size image

As we continue to enrich the PS domain, we observed MM with a PS matrix, which requires the PB and PM microdomains to change their shape as well (Fig. 4). Addition of 20 vol% hPS relative to the PS block of S59B16M25119 is equal to an increase of φPS = 0.11 to obtain S70B11M19. Since S59B16M25119 possess an lr-morphology (Fig. 4a), we expected the transition of at least two microdomains when φPS is increased [69]. Whereas the MMs still show a spherical structure with a smooth surface in SEM, TEM shows that PS now forms the matrix while PM changed into hexagonally packed cylinders (Fig. 4c). The PB domains maintain their cylindrical form and diameter, dcyl, PB ≈ 15 nm, resulting in an overall coc-morphology. The morphological change of PM is accompanied by an increase in the microdomain size from dcyl PM ≈ 27 nm in the blend compared to the thickness of the PM lamellae wlam, PM ≈ 20 nm before blending. Since the volume fraction φPM decreases during blending, an increase in PM domain size seems counterintuitive at first. However, the total volume of the PM block remains the same while the lamella microdomains separate into cylinders, which – at constant chain length – will result in a larger diameter.

Fig. 4
figure 4

Blending of S59B16M25119 with 20 vol% hPS. a) Ternary microphase diagram with blending pathway indicated by the arrow. b) TEM and SEM (inset) images of unblended S59B16M25119 and c) of S70B11M19 blended with 20 vol% hPS. PB was stained with OsO4 and appears darkest. Scale bars are 200 nm

Full size image

Morphological transitions by blending hPS, hPB, or hPM. We next utilize S33B23M44100, whose composition resides within the coc-morphology with PB cylinders on PS cylinders embedded in a PM matrix (Fig. 5a, b), but is likewise located near the lpl region (blue region in Fig. 5a) and the ll region (purple region in Fig. 5a). Thus, multiple transitions could be expected after blending. First, we blended S33B23M44100 with hPS to obtain S39B21M40 (20 vol%; φPS = 0.06) and S46B18M36 (40 vol%; φPS = 0.13). The morphology of S39B21M40 already shows signs of change towards an lpl-morphology but is still in a transition state so that the MM appears disordered (Figure S3)[70]. While merging of PS cylinders into lamellae takes place to some extent, the added hPS is not sufficient to induce a full transition. Blending a larger amount of hPS to reach S46B18M36 fully realized the transition and continuous PS lamellae are observed (Fig. 5c). During merging of cylinders, the PS domain decreases in thickness from cylinders with a diameter of dcyl, PS ≈ 40 nm to lamellae with a width of wlam, PS ≈ 25 nm. At the same time, the PM matrix adopts a lamella morphology as well, and PB now forms rings instead of straight cylinders due to the concentric PS/PM lamellae.

Fig. 5
figure 5

Structural transitions of blended S33B23M44100. a) Ternary microphase diagram with blending pathway indicated in the closeup. TEM and SEM (inset) images of b) unblended S33B23M44100, c) hPS blended S46B18M36, d) hPM blended S29B20M51, and e) hPB blended S28B35M37. PB was stained with OsO4 and appears darkest. Scale bars are 200 nm

Full size image

Next, we blended S33B23M44100 with hPM to obtain S29B20M51 (15 vol%; φPM = 0.07; Fig. 5d) and S26B18M56 (25 vol%; φPM = 0.12; Figure S3d). Both blends led to a coc morphology, but now the packing of PS domains is clearly hexagonal instead of tetragonal. The PB cylinders are at the interface between the PS cylinders and the PM matrix. Even though we did not observe a change in the microdomains, the results demonstrate that blending also allows to induce rather subtle changes in morphologies, making it a powerful tool for finetuning. Finally, we blended S33B23M44100 with hPB to S30B29M40 (30 vol%;φPB = 0.06; Figure S3e) and S28B35M37 (55 vol%; φPB = 0.12; Fig. 5e). In these blended compositions, all blocks have about equal volume fractions for which a transition to the ll-morphology is expected. For S30B29M40 (Figure S3e), we already partially see this trend, i.e., PS and PM clearly transition to lamellae, while PB mostly forms perforated lamellae, leading to the lpl-morphology. Again, the addition of more hPB led to continuous PB lamellae and a homogeneous axially stacked ll-morphology (Fig. 5e). Merging of PB cylinders towards lamellae is accompanied by merging of PS cylinders into lamellae with a thickness of wPS ≈ 18 nm. Since no PS was added to the system, the decrease in PS domain size is more pronounced than discussed above (S46B18M36100, Fig. 5c). In comparison, the decrease of the PB domain size is rather small, i.e., the width of the lamellae, wPB ≈ 17 nm, is only about 2 nm smaller than the diameter of the PB cylinders found for S33B23M44100.

Conclusions

In conclusion, we showed that blending linear SBM triblock terpolymers with homopolymers is an effective strategy for tuning the inner structure of terpolymer-based multicompartment microparticles. We induced and analyzed one-, two- or three-compartment-transitions in the particle morphologies and demonstrated the versatility and potential of this approach. In parallel, we investigated possible issues and limitations related to over-blending of the homopolymer, such as phase separation of the added homopolymers. Our results demonstrate that this blending approach successfully replicates the morphologies of pure SBM terpolymers, thereby avoiding the time- and cost-intensive synthesis of different terpolymer compositions. Looking ahead, we aim to further leverage the acquired knowledge and potential for the structural analysis of triblock terpolymer based microparticles and the establishment of a ternary microphase diagram as a comprehensive guideline for targeting specific morphologies. Future research will also investigate the simultaneous blending with two homopolymers to expand the scope of this approach.