Leveraging Macromolecular Isomerism for Phase Complexity in Janus Nanograins

Abstract

It remains intriguing whether macromolecular isomerism, along with competing molecular interactions, could be leveraged to create unconventional phase structures and generate considerable phase complexity in soft matter. Herein, we report the synthesis, assembly, and phase behaviors of a series of precisely defined regioisomeric Janus nanograins with distinct core symmetry. They are named B₂DB₂ where B stands for iso-butyl-functionalized polyhedral oligomeric silsesquioxanes (POSS) and D stands for dihydroxyl-functionalized POSS. While BPOSS prefers crystallization with a flat interface, DPOSS prefers to phase-separate from BPOSS. In solution, they form 2D crystals owing to strong BPOSS crystallization. In bulk, the subtle competition between crystallization and phase separation is strongly influenced by the core symmetry, leading to distinct phase structures and transition behaviors. The phase complexity was understood based on their symmetry, molecular packing, and free energy profiles. The results demonstrate that regioisomerism could indeed generate profound phase complexity.

Short abstract

In a series of B₂AB₂-type Janus nanograins, macromolecular isomerism was combined with competing interactions between crystallization and phase separation to generate considerable phase complexity.

Introduction

Since its conception by Jacob Berzelius,¹ molecular isomerism has been recognized as a fundamental phenomenon in chemistry featuring drastic property changes with identical composition. While numerous examples have been documented for small molecules, macromolecular isomerism is a relatively recent subject.² Chain molecules are often polydispersed and rely on collective interactions among multiple regio- or stereocenters to exhibit distinct physical properties.^3,4 In the study of structure–property relationships in macromolecules, it remains elusive how significant changes could be introduced into macromolecular systems on variation of only a single regio- or stereocenter.^5,6 The question must be addressed within the framework of precision macromolecules.

Giant molecules are precision macromolecules built on molecular nanoparticles.⁷⁻¹⁰ Unlike synthetic polymers, they are not long-chain molecules, but conformationally rigid molecular nanograins with defined sizes and shapes. Hence, their assembly is sensitive to minute primary structural changes, providing an ideal platform for studying macromolecular isomerism.¹¹⁻¹⁶ Previously, a series of bifunctional polyhedral oligomeric silsesquioxane (POSS) regioisomers have been reported,^17,18 which promotes the investigation of regioconfiguration on the self-assembly of giant molecules, such as Janus star polymers,^18,19 double-chained giant surfactants,^20,21 quadruple-chained giant surfactants,^22,23 Janus amphiphilic particles,^24,25 giant polymeric chains,²⁶⁻²⁸ molecular patchy clusters,¹⁵ and particle-like smectic liquid crystals.²⁹ Not only does regiochemistry exert profound influences on the equilibrium structures and transition kinetics of their assembly (as in isomeric double-chained surfactants^20,21 and ABA-type Janus nanoparticles^24,25,29), but there is also a strong correlation between regiochemistry and the formation of Frank–Kasper phases (as in quadruple-chained giant surfactants,²² patchy particle isomers,¹⁵ and geometric isomers³⁰). While the profound role of isomerism in tuning assembly is unambiguously established, challenges remain on whether unconventional phases and considerable phase complexity could be introduced by varying only one regiochemistry in macromolecular isomers.

The emergence of unconventional phases in single-component molecular systems is a recent event that links molecular symmetry/design with molecular interactions/close-packing.^31,32 To date, such phases include 2D tilings,^33,34 3D networks,³⁵⁻⁴¹ and spherical-packing such as Frank–Kasper A15 phase and sigma phase,⁴²⁻⁵⁵ Z phase,⁵⁶ Laves C14 and C15 phases,⁵⁷⁻⁶³ etc. Recently, a deliberate pairing of giant molecules has led to peculiar self-sorting behavior which opens an avenue for constructing complex soft lattices, such as binary crystal phases⁶⁴⁻⁶⁷ and decagonal quasicrystals,⁶⁸ with intriguing dynamic properties.⁶⁹ Despite these successes, molecular isomerism has not yet been well exploited in generating unconventional phases.

Herein, we report the synthesis, assembly, and phase behaviors of a full set of regioisomeric Janus nanograins based on POSS with distinct core symmetry, namely, B₂DB₂ where B stands for iso-butyl-functionalized polyhedral oligomeric silsesquioxanes (POSS) and D stands for dihydroxyl-functionalized POSS. We envisioned that the interplay between BPOSS crystallization⁷⁰ and BPOSS/DPOSS phase separation⁴⁶ should lead to unconventional phases and complex phase behaviors as a sole consequence of macromolecular isomerism under the constraint of core symmetry and volume asymmetry.

Results and Discussion

Molecular Design and Precision Synthesis

Previously, we have shown that the reaction between octavinyl POSS and 1-thioglycerol gives a mixture of different adducts from which the three bifunctional POSS isomers can be isolated and fully identified.²² These ortho-, meta-, and para-isomers of bifunctional POSS are ideal scaffolds for regioisomer design.¹¹ By replacing the soft building blocks in previous examples (e.g., polystyrene chains and noncrystalline POSS) with a crystalline BPOSS, we intended to introduce crystallization as a strong competing force to phase separation. To reach spherical phases, a large volume asymmetry between hydrophilic DPOSS and hydrophobic BPOSS is required, which could be conveniently achieved by controlling a ratio of 1:4 for DPOSS:BPOSS for a final BPOSS volume fraction (f_BPOSS) of ∼0.82 (Scheme 1, Table 1).⁴⁶ Hence, we installed two hydroxyl groups on one vertex of bifunctional POSS for subsequent attachment of one BPOSS per hydroxyl group. A straightforward way to prepare such tetrafunctional POSS is via thiol–ene chemistry using thiolglycerol.²² The bifunctional regioisomers (V₆T₈-4OH) were converted to V₆T₈-4yne and linked with four BPOSS through a copper-catalyzed alkyne–azide cyclization reaction to give B₂VB₂ (Scheme 1, see Figures S1 and S2 for ¹H NMR and ¹³C NMR characterizations).²⁴ Then, the vinyl groups were converted to hydroxyl groups to yield the Janus nanograins B₂DB₂ by a second thiol–ene reaction (see Figures S3 and S4 for NMR characterizations).

Scheme 1. Synthetic Route of B₂DB₂ Giant Molecular Isomers.

Reagent and conditions: (i) 1-Thioglycerol; DMPA; UV (365 nm); 10 min; THF; separation yields for V₆T₈-4OH isomers, ∼8% for the para-isomer, ∼10% for the meta-isomer, and ∼10% for the ortho-isomer. (ii) 4-Pentynoic acid; DIPC; DMAP; DCM; 24 h; ∼80%. (iii) BPOSS-N₃; CuBr; PMDETA; THF; 12 h; ∼70%. (iv) 1-Thioglycerol; Irgacure 2959; UV (365 nm); 10 min; THF; ∼70%.

Table 1. Physical Property Summary of B₂DB₂ Regioisomers.

configuration	fBPOSSa	Tmb (°C)	Tcc (°C)	lattice 0d	d0e	lattice 1f	lattice dimensiong (nm)	lattice 2h	a2i (nm)	μj	A0,2k (nm2)
para	0.82	189	165	iLAM	3.4	hexagonal crystal	a = 5.76; c = 9.59	Colh	5.33	3.2	2.21
meta	0.82	186	168	iLAM	3.4	hexagonal crystal	a = 5.76; c = 9.59	iColh
ortho	0.82	186	165	iLAM	3.0	hexagonal crystal	a = 5.58; c = 9.38	A15	11.07	21.5	2.16

^aVolume fraction of BPOSS.

^bMelting temperatures are the extrapolated onset temperatures obtained from DSC.

^cCrystallization temperatures are peak values from DSC.

^dFrustrated lamellar packing of the samples.

^eLayer spacing of the iLAM morphology.

^fLattice 1 represents the structure developed after annealing at the temperature close to T_c.

^gLattice dimensions of lattice 1.

^hLattice 2 represents the phase-separated structure when BPOSS crystal melts.

ⁱLattice dimension of lattice 2.

^jAverage number of molecules within a 1 nm thick cross-section of the cylinders in the Col_h structure and the calculated average numbers of molecules per supramolecular sphere in the Frank–Kasper A15 structure.

^kAverage interfacial area per molecule calculated for lattice 2. The detailed calculation can be found in the Supporting Information.

The evidence for each isomer’s regioconfiguration is from their ²⁹Si NMR spectra (Figure 1A, bottom) where the resonances for the silicon atoms linked to the vinyl groups show up around −80 ppm and have only one peak (6 Si) for the para-isomer, three peaks (2:2:2) for the meta-isomer, and two peaks (2:4) for the ortho-isomer. After the final thiol–ene reaction, their chemical shifts move completely to −69 ppm with largely unchanged spectral patterns (Figure 1A, upper; and Figure S5).²⁵Figure 1B shows the MALDI-TOF mass spectra of the intermediates during the synthesis. Although these isomers have identical molecular weights consistent with theoretical values, they have distinct retention volumes in size exclusion chromatography (SEC) (Figure 1C) with the para-isomer being the smallest, the ortho-isomer the largest, and the meta-isomer in between. Hence, the para-isomer has the most extended conformation, and the ortho-isomer has the most compact one. We can conclude that three isomers with precisely defined primary structures have been successfully synthesized.

Figure 1.

(A) ²⁹Si NMR spectra show the conversion from B₂VB₂ (bottom) to B₂DB₂ (upper). (B) MALDI-TOF MS spectra show the monoisotopic peaks of B₂VB₂ (left) and B₂DB₂ (right). (C) SEC overlay of B₂VB₂ (upper) and B₂DB₂ (bottom). In all figures, black lines are used to denote the para-isomer, red lines the meta-isomer, and blue lines the ortho-isomer.

Characterization of Thermal Properties

To facilitate structure development, the as-prepared samples were slowly evaporated from a THF/MeCN mixed solution (v/v = 1/1) and then dried in vacuo overnight at 60 °C before further characterization.⁴⁴ To probe the phase behavior of these isomers, we evaluated their thermal stability using thermogravimetric analysis (TGA), which showed no obvious mass loss up to 200 °C (Figure 2A). In the differential scanning calorimetry (DSC) thermogram, the first cooling and second heating curves reveal similar melting temperatures (T_m’s) and crystallization temperatures (T_c’s) (Figure 2B–D). As summarized in Table 1, the para-B₂DB₂ has the highest T_m (189 °C), while meta- and ortho-B₂DB₂ have slightly lower T_m’s (186 °C). There is considerable supercooling for crystallization (∼20 °C). All three isomers have similar T_c’s (∼165 °C). Despite relatively simple DSC curves, we anticipated more complex phase behaviors for these samples. Hence, we followed the phase transitions of three isomers on heating using a combination of temperature-dependent wide-angle scattering (TD-WAXS), small-angle X-ray scattering (SAXS), and transmission electron microscopy (TEM) imaging (Figures 3–5), which indeed capture rich phase behaviors not obvious in the DSC.

Figure 2.

(A) TGA curves of the three isomers: black lines for para-, red lines for meta-, and blue lines for ortho-isomers. Second heating and first cooling differential scanning calorimetry traces of (B) para-, (C) meta-, and (D) ortho-isomers. The extrapolated onset temperatures are shown for the corresponding melting points, and the peak values are shown for crystallization.

Figure 3.

(A) TD-WAXS pattern of para-B₂DB₂ from 40 to 220 °C. TEM image (B) and SAXS curve (C) showing the ill-defined lamellar structure (iLAM) at 40 °C. TEM images (D–F) and SAXS curve (G) showing the crystal structure at 175 °C. TEM image (H) and SAXS curve (I) showing the microphase-separated Col_h phase at 185 °C. The top-right corner insets of TEM images (panels B, D, E, and H) are the FFT patterns. TEM images (F) and bottom-left corner insets of the TEM images (H) are images after the Fourier filtering. (J) Phase diagram corresponding to TD-WAXS. Iso: isotropic state.

Figure 5.

(A) TD-WAXS pattern of ortho-B₂DB₂ from 40 to 220 °C. TEM image (B) and SAXS curve (C) showing the ill-defined lamellae structure at 40 °C. TEM image (D–F) and SAXS curve (G) showing the superlattice after shearing and annealing. TEM image (H) and SAXS curve (I) showing the microphase-separated A15 phase after melting. The top-right corner insets of TEM images in panels B–E are the FFT pattern. The bottom-left corner insets of TEM images in panels H and F are images after the Fourier filter. (J) Phase diagram corresponding to TD-WAXS.

Phase Behaviors of the para-Isomer

For the para-isomer, the WAXS at 40 °C (Figure 3A) and the TEM images of the microtomed sample reveal a wavy lamellar morphology (Figure 3B). The layer spacing measured from the TEM image (∼3.4 nm) is consistent with the calculated value from SAXS (Figure 3C). This is assigned as an ill-defined, frustrated lamellae morphology (iLAM) whose formation is dominated by BPOSS crystallization.⁴⁴ On heating above 150 °C, the metastable iLAM gradually transforms into a much ordered, stable crystalline phase, as evidenced by multiple sharp peaks in WAXS (Figure 3A). The SAXS profile of the sample annealed at 175 °C (Figure 3G) reveals a diffraction pattern with three sharp peaks. The full spectrum shown in Figure S6 suggests a hexagonal crystal structure with a = 5.76 nm, c = 9.59 nm.⁷¹⁻⁷³ The corresponding index assignments are summarized in Table S1. To see the structure in real space, the crystalline sample was microtomed, stained, and imaged under TEM. As shown in Figure 3D–F, a large area of hexagonal honeycomb morphology (from ⟨001⟩ projection), as well as the layered morphology (from ⟨100⟩ projection), could be observed where the bright area is BPOSS, and the dark area is stained DPOSS. Selected area electron diffraction (SAED) shows the characteristic patterns of a hexagonal lattice (Figure S7) with dimensions matching those calculated from SAXS. The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) further shows a discernible contrast from the matrix to the core resembling the bright-field TEM image (Figure S8). Within this lattice, BPOSS crystallization is evidenced by the sharp diffraction peaks at ∼6.0 nm^–1 (Figure S6) corresponding to the characteristic size of BPOSS (∼1.1 nm).⁷⁴ We inferred that BPOSS would most likely crystallize into a matrix with a flat interface that wraps the core DPOSS cluster in a zigzag fashion.^44,75,76 When the crystal melts at ∼185 °C, only one sharp peak remains in the small-angle region, indicating mesophase formation. This structure is determined to be a Col_h showing the characteristic (10), (11), and (20) lattice plane with the q/q* of 1, √3, and √4 from the synchrotron SAXS profile (Figure 3I). The calculated lattice dimension is 5.33 nm, slightly smaller than that of a in the crystalline phase. The TEM image of the sample quenched from 185 °C also demonstrates 6-fold symmetry (Figure 3H). On further heating, there is an order–disorder transition at ∼220 °C.

Phase Behaviors of the meta-Isomer

For the meta-isomer, the WAXS at 40 °C (Figure 4A), the corresponding TEM image (Figure 4B), and SAXS profile (Figure 4C) reveal an iLAM morphology similar to that of the para-isomer with almost identical layer spacing (∼3.4 nm). On heating, crystallization occurs at ∼140 °C, leading to an ordered hexagonal crystal whose properties (see Figure 4D–F for TEM images, Figure 4G for the SAXS profile, and Table S2 for the index) are almost identical to those of the para-isomer. However, there is no mesophase formation on melting. The hexagonal crystal directly enters the isotropic state at ∼190 °C. We speculated that a mesophase similar to that in the para-isomer should also exist, albeit with limited stability, such that T_ODT is lower than the T_m of the hexagonal crystal. If so, since there is a ∼20 °C supercooling, it may be possible to develop such a mesophase during cooling before crystallization occurs. By quenching the sample from the isotropic state (200 °C) to 180 °C followed by annealing (Figure S9), an ordered phase was detected in SAXS, showing one sharp peak at q ∼ 1.3 nm^–1 (Figure 4I). To capture this mesophase, the sample was immediately quenched in liquid nitrogen, microtomed, and stained for imaging under TEM. A faint pattern with 6-fold symmetry could be discerned in Figure 4H. Hence, we tentatively assign this phase to be an ill-defined Col_h (iCol_h) mesophase with limited stability.

Figure 4.

(A) TD-WAXS pattern of meta-B₂DB₂ from 40 to 200 °C. TEM image (B) and SAXS curve (C) showing the iLAM at 40 °C. TEM image (D–F) and SAXS curve (G) showing the hexagonal crystal developed at 170 °C. TEM image (H) and SAXS curve (I) showing the microphase-separated iCol_h phase at ∼180 °C after quenching and annealing. The top-right corner insets of TEM images in panels B, D, E, and H are the FFT patterns. The bottom-left corner insets of the TEM images in panels H and F are images after the Fourier filter. (J) Phase diagram corresponding to TD-WAXS. Iso: isotropic state.

Phase Behaviors of the ortho-Isomer

For the ortho-isomer, the sample also shows an iLAM morphology at 40 °C as revealed by WAXS (Figure 5A), TEM image (Figure 5B), and SAXS (Figure 5C). However, it does not develop into an ordered crystalline phase on heating. With the disappearance of the peak at 1.8 nm^–1, only a broad peak at ∼1.3 nm^–1 appears. We hypothesized that the flat interface required by crystallization may be incompatible with the cone shape of the ortho-isomer, making it difficult to develop highly ordered crystalline phases.^44,77 On further heating above 170 °C where BPOSS crystallization no longer holds (Figure 5A), phase separation comes into play forming an ordered phase with q/q* equal to √2, √4, √5, √6, √8, , , , and in the SAXS profile (Figure 5I). It is assigned as the Frank–Kasper A15 phase with a space group of Pm3n. The indexing is shown in Table S4, and the calculated lattice dimension is 11.07 nm. The A15 lattice was further confirmed by TEM imaging on microtomed samples. The corresponding fast Fourier transform (FFT) pattern and local TEM image after Fourier filtering are shown as insets of Figure 5H. An unambiguous 4⁴ tiling pattern along the ⟨100⟩ direction is characteristic of the A15 phase.⁴⁴ The A15 phase is so stable that there is no order–disorder transition up to 220 °C.

Previously, it was found that mechanical shearing could help develop ordered structures.⁷⁸ Thus, the sample was sheared at 188 °C and naturally cooled to room temperature. By doing so, the ortho-isomer formed a similar hexagonal crystal as the para- and meta-isomers, as revealed by TEM (Figure 5E–F) and SAXS (Figure 5G), with minor differences in lattice dimension (Table S3). Intriguingly, the sheared crystalline sample can hardly develop into A15 phase on melting (Figure S10). This could be understood based on the geometric mismatch of the two ordered structures. The ortho-isomer is cone-shaped with four BPOSS motifs pointing to the same side. Crystallization requires a flat interface between BPOSS motifs, while A15 phase is formed based on the stacking of supramolecular spherical aggregates favoring curved interfaces. The geometric mismatch means that their mutual transitions are hindered by a large kinetic barrier. The hexagonally packed molecules in the crystal could hardly rearrange into spherical aggregates for A15 phase formation. Similarly, cooling the A15 phases below the T_m would just lead to local BPOSS crystallization disrupting the A15 structure without global rearrangement into the hexagonal lattice (Figure S11).

Phase Diagram of B₂DB₂ Isomers

Based on the above discussions, we constructed phase diagrams for the three regioisomers (Figure 6). When evaporated from solution, three isomers adopt a kinetically trapped iLAM structure dominated by the BPOSS crystallization with a flat interface. Domination of BPOSS crystallization also occurs in solution where phase separation between BPOSS and DPOSS becomes trivial. Single-crystalline 2D nanosheets could be obtained for all three isomers from highly diluted solutions in DMF/toluene (Figure S12). The SAED pattern reveals that these nanosheets consisted of two layers of crystalline BPOSS and one layer of DPOSS.^70,79 The formation of iLAM in bulk at low temperatures is thus not surprising. At elevated temperatures, molecules will reorganize to accommodate the BPOSS crystallization. When molecular symmetry allows it, as in para- and meta-isomers, formation of a globally ordered crystalline structure is spontaneous. Otherwise, it would require some assistance such as mechanical shearing to preorganize the molecules in the right orientation, as in the ortho-isomer. Nevertheless, they all form hexagonal crystals with minor differences in lattice dimension and molecular packing. Above T_m, phase separation dominates the mesophase formation. The para-isomer forms Col_h which becomes isotropic at 220 °C while the ortho-isomer adopts a Frank–Kasper A15 phase stable even at 220 °C. As for the meta-isomer, an ill-defined, metastable iCol_h phase could only be developed within the narrow window of supercooling. The kinetics from the development of mesophases was very fast (less than 30 min) for all three isomers (Figure S13). With the only difference in regioconfiguration, these isomers indeed show rich phase behaviors.

Figure 6.

Phase diagram of B₂DB₂ regioisomers. The brown shaded lines in meta-isomer represent the metastable iCol_h phase; the green shaded lines at lower temperatures represent the metastable iLAM phase.

Proposed Molecular Packing Scheme

To understand the phase diagram, we must understand the molecular packing in these structures. We assume that these structures may well be considered as “supramolecular crystals” possessing long-range translational order in terms of their basic self-assembled repeating units (motifs), but not necessarily in each atomic position within these motifs.⁸⁰ Although the current data do not support structural resolution at the atomic level, we proposed the following molecular packing for ordered structures (Figure 7). On the molecular level, the para-isomer probably takes on an expanded conformation with tethered BPOSS motifs positioned on almost opposite sides, and the ortho-isomer likely adopts a cone shape with all four BPOSS motifs pointing to the same side, while the meta-isomer is somewhere in between. Considering the volume fraction of BPOSS, it is not surprising to see that, on melting, they undergo microphase separation to form mesophases with curved interfaces. Under this circumstance, the para- and meta-isomer may adopt a fan-shape conformation with four BPOSS units bending toward one side of DPOSS and forming a columnar structure (Figure 7A). From the lattice dimension of the Col_h mesophases, we could determine that there are ∼3 molecules within a 1 nm thick column for both isomers. The DPOSS core is probably not highly ordered and, thus, should be uniformly stained and show up as dark regions under TEM (Figures 3H and 4H). By contrast, the cone-shaped geometry of ortho-B₂DB₂ leads to spherical aggregate formation which further packs into an equilibrium A15 phase (Figure 7). On average, there are ∼22 molecules per spherical aggregate. At lower temperatures, BPOSS crystallization dominates, leading to formation of hexagonal crystals. The densities were determined to be ∼1.24 g/cm³ for the para- and meta-isomers and ∼1.22 g/cm³ for the ortho-isomer. From their lattice dimensions, we deduce that there are ∼18 molecules per unit cell for the para- and meta-isomers and ∼16.5 molecules for the ortho-isomer. Considering that the hexagonal crystal structure and Col_h mesophase are closely related, they may share a very similar packing scheme. It is likely that three molecules form a layer with DPOSS in the center and BPOSS in the periphery and that there are about six layers in a unit cell whose thickness is roughly consistent with the c axis dimension (∼9.59 nm). The central DPOSS should be highly ordered, and the peripheral BPOSS motifs are crystallized. Thus, staining mainly occurs at the side chain and linker region, but less on the central POSS core and the peripheral BPOSS layer. This explains the pattern observed under TEM where the central white dot is the unstained core of DPOSS, the zigzag-like white matrix the crystallized BPOSS motifs, and the region in between the heavily stained DPOSS side chains and linkers. From the current data, we could not pin down how BPOSS motifs are arranged within the unit cell. We speculate that they probably wrap around the central DPOSS in a helical pattern, which is supported by the TEM images (Figures 3F, 4F, and 5F, bright lines) along the c-direction showing faint traces of a spiral structure. Therefore, the configuration of the para-isomer is most compatible with this packing, and the corresponding crystal is the easiest to form with the highest stability. The geometric mismatch between the meta-isomer configuration and the hexagonal crystal would introduce considerable tension on the linker region, which destabilizes the crystal as shown by its lower T_m. It is an even bigger challenge to accommodate the cone-shaped ortho-isomer into this packing scheme. As a result, the hexagonal crystal of the ortho-isomer is the most difficult one to form, requiring the assistance of mechanical shearing. Deformation of the unit cell also occurs which can accommodate fewer molecules per unit cell (∼16.5). Due to lattice symmetry mismatch, the transition between the hexagonal crystal and A15 mesophase is slow with a large kinetic barrier (Figure S10). The A15 phase could only be developed either by annealing a noncrystalline sample through an ill-defined lamellar intermediate (Figure 5A) or by quenching from the isotropic state (Figure S13C).

Figure 7.

(A) Cartoon illustration of the typical phase structures and transition behavior of para-B₂DB₂, meta-B₂DB₂, and ortho-B₂DB₂ regioisomers. (B) Corresponding free energy profiles; red arrows represent the heating process, green arrows cooling processes with different protocols, and orange arrows the solution process.

Understanding Phase Stability

The stabilities of the mesophase and crystal are evaluated by T_iso and T_m, respectively. The mesophase stability follows the order of A15 (ortho) > Col_h (para) > iCol_h (meta). Since the mesophase stability is critically related to the average cross-section area per molecule (A₀, see Table 1 and Table S5), we compared the mesophases and the corresponding A₀ values for the regioisomers of BDB,²⁴ B₂DB₂, DPOSS-2PS (or SDS),²¹ and S₂DS₂²² with similar volume fractions of the hydrophobic matrix. For consistency, the linker region was counted as part of the hydrophobic matrix, and thus, some of the data were recalculated based on the reported values. The calculated A₀ is ∼2.21 nm² for the Col_h mesophase formed by the para-isomer (eq S6) and ∼2.16 nm² on average for the A15 mesophase formed by the ortho-isomer (eq S7). The higher stability of the A15 phase is reflected in its small A₀ values, which is probably caused by a significantly larger number of molecules in one soft sphere that collectively stabilized the structure.

By comparing similar isomers in this series of samples (Table S5), we can see the following trends: (1) When the tethered part is the particulate BPOSS, the formation of ordered structures will be intricately dependent on the core geometry, which is easier in the cases of para- or ortho-isomers when the geometry matches but is rather difficult for the meta-isomer. (2) When the ordered structures can form in BDB and B₂DB₂, they tend to be more stable than the counterparts with tethered PS chains, which can be rationalized by the corresponding smaller entropic change in order–disorder transition. (3) For similar molecular geometries, compared to PS, tethering a BPOSS particle increases the conformational asymmetry at the interface and promotes the formation of curved morphologies and unconventional spherical phases at comparable or smaller volume fractions. (4) In all cases, the para-isomer seems to possess the highest A₀ values, and the ortho-isomer tends to have the smallest A₀ values, suggesting that the ortho-isomer has the most stable phase-separated structure among the three isomers since the clustered geometry is preorganized for phase separation. These results improve our understanding on the effect of macromolecular isomerism on the phase behaviors of giant molecule assemblies.

The introduction of the BPOSS motif makes crystallization a strong competing interaction to phase separation. When the two act in synergy, the structure stability is enhanced. For all samples, the hexagonal crystal is the thermodynamically stable phase, and the iLAM is metastable at lower temperatures. The para-isomer has the highest T_m for the crystal because its symmetry is most compatible with the proposed packing, whereas the other two have comparably smaller T_m’s. These competing interactions are also the origin of metastable phases and monotropic phase behaviors. To rationalize the phase behavior of these isomers, we demonstrate the transition pathways in the postulated free energy landscape in Figure 7B. As for para- and ortho-B₂DB₂, the Col_h and A15 are equilibrium phases. For meta-B₂DB₂, the iCol_h phase is metastable, with its isotropic temperature (T_iso) lower than its melting point T_m. Compared to previous results on the self-assembly of DPOSS-BPOSS molecular systems,^44,46 the reduced number of hydroxyl groups (from 14 to 12) leads to a subtle change in the free energy landscape. Together with the regioconfiguration, considerable phase complexity is generated. We envision that these kinds of Janus nanograins with tunable phase structures and sophisticated phase behaviors can serve as precursors in the fabrication of mesoporous nanosilica materials for catalysis, separation, etc.⁸¹⁻⁸³

Conclusion

In summary, a series of precisely defined regioisomeric Janus nanograins were synthesized and found to exhibit complex phase behaviors, forming distinct nanostructures as a combined result of symmetry/geometry, phase separation, and crystallization. When properly crystallized, three isomers form very similar hexagonal crystals. Notably, the ortho-isomer requires mechanical shearing to overcome the high nucleation barrier for ordered packing. On melting, phase separation dominates the mesophase formation. The Col_h phase and Frank–Kasper A15 phase were further detected in para- and ortho-isomers, while the meta-isomer exhibited a metastable and ill-defined Col_h phase due to the mismatched geometry. Also due to geometry mismatch, the transition between hexagonal crystal and Frank–Kasper A15 phase in the ortho-isomer was found to be very slow. Moreover, the stability of these phases was in general agreement with previous findings in counterparts, with the para-isomer having the most stable crystal structure and the ortho-isomer having the most stable mesophase. The results highlight a profound influence of minute structural changes on the assembly outcomes and suggest that regioconfiguration provides a new dimension for generating phase complexity and fine-tuning the structures and properties of materials. This affords enticing opportunities to comprehensively assess the role of macromolecular isomerism in the self-assembly of nanoscaled amphiphilic giant molecules and to leverage isomerism for controlled phase complexity and delicate function in soft matter.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acscentsci.2c01405.

• Chemicals and instrumentation, experimental procedures, density measurement, detailed characterizations, and calculations (PDF)

Author Contributions

^∥ Y.S. and D.H. contributed equally to this work.

The authors declare no competing financial interest.