Giant surfactants provide a versatile platform for sub-10-nm nanostructure engineering

If we consider MNPs as the polar heads, various giant surfactants can be constructed in analogy to their small-molecule surfactant counterparts, as schematically illustrated in Fig. 1A. Notably, the head groups can be different MNPs with patchy surface chemistry; whereas the tails can vary in chemical composition and chain topology, both of which greatly diversify the molecular design and provide endless possibilities in fine-tuning their structural formations. Extensive libraries of giant surfactants have been synthesized (41–44) and screened for their self-assembled morphologies in the bulk, thin film, and solution states, as outlined in SI Text. Following the “click” philosophy, the syntheses of these giant surfactants fulfill an assembly process that is modular, robust, and efficient (45). Not only can each of the materials be readily synthesized in gram quantities, but also the molecular parameters (such as the identity of periphery functional groups on MNPs and the length of polymer tails) can be individually tailored and systematically varied. The chemical structures of five exemplary giant surfactant libraries are shown in Fig. 1B, their synthetic approaches are shown in Schemes S1–S5, and their molecular characterizations are summarized in Table S1. In the present study, all of the polymer tails are hydrophobic polystyrene (PS). It should be noted that they can also vary from a wide range of selections (42). It is equally expected that the introduction of other chemically incompatible polymers with competing interactions would further enrich the palette and drive the formation of unique hierarchical structures.

In the bulk, these giant surfactants readily undergo microphase separation and self-assemble into various ordered morphologies at the nanometer scale. Owing to the high diffusion mobility of MNP heads and the lack of chain entanglement in the relatively low molecular weight (MW) region of polymer tails, they exhibit rapid self-assembly dynamics and achieve thermodynamically stable phases typically within minutes to hours upon thermal annealing. Fig. 2A shows a set of one-dimensional (1D) small-angle X-ray scattering (SAXS) profiles in reciprocal space obtained from a subset of samples in library 1, DPOSS-PS_n, where DPOSS represents hydroxyl-functionalized POSS and n denotes the average degree of polymerization of PS with different lengths. With increasing PS tail length (also the volume fraction of PS, V_PS, Table S1), the self-assembled structures change from lamellae (Lam), to double gyroids (DG), to hexagonally packed cylinders (Hex), and further to body-centered cubic spheres (BCC). This is in good agreement with the observation of microtomed thin sections of the bulk samples in real space under bright-field transmission electron microcopy (TEM). Even for the sample with the lowest MW of PS tail in this series (ca. 2.0 kg/mol), the microphase separated lamellar structure exhibits excellent orders and sharp interfaces, supported by the autocorrelation function analysis applied to the corresponding SAXS profile (Fig. S1). Notably all of these patterns were obtained without staining, indicating a high electron density contrast. The trend of this phase structure changes is also followed in the other two series of giant surfactants in library 1 with different head surface functional groups [carboxylic acids (APOSS-PS_n) and perfluorinated alkyl chains (FPOSS-PS_n)] and those in library 2 with a different MNP head group based on carboxylic acid-functionalized C₆₀. The feature sizes of all these samples are typically around 10 nm or smaller (Fig. 2 and Table S1).

Owing to the fixed and relatively small sizes of the head groups, only half of the structural phase diagram is observed in Fig. 2A. The lower limit of V_PS in libraries 1 and 2 is determined by the shortest PS with narrow dispersity (<1.10). At lower PS MWs, the phase structures are obscured by the increasing dispersity of the PS tails. A precise determination of the other half of the phase diagram would require well-defined oligomers. The structures could also be tuned by adding tails or heads to adjust the constraints on the interfacial curvatures as shown in Fig. 1A and more specifically in libraries 3–5. For example, a multiheaded giant surfactant with a low-MW PS tail, such as 3DPOSS-PS₁₉ in library 5, exhibits an inverse Hex phase shown in Fig. 2B, which has not yet been observed in the single-headed/single-tailed giant surfactant library.

On the other hand, using a different number of tails leads to topological isomers that have identical V_PS but distinct polymer topologies (e.g., one head/tail vs. multiple head/tails). An exceptional sensitivity of the resulting phase structure to molecular topology was observed, as demonstrated in two pairs of topological isomers, AC₆₀-PS₄₄ (a Lam) and AC₆₀-2PS₂₃ (a Hex), as well as DPOSS-PS₃₅ (a DG) and DPOSS-2PS₁₇ (a Hex) (Fig. 2C and Fig. S2). Therefore, an additional order parameter, the geometric cross-sectional area ratio between the head(s) and tail(s), should be introduced in determining the phase structures. This structural sensitivity is characteristic of small molecules but less significant in block copolymers. Further structural variations, such as using patchy heads or installing heads of different surface functionalities at each end of the polymer chain to create a giant bolaform surfactant, an analog of multiblock copolymers, offer unparalleled opportunities for nanostructure engineering at a sub-10-nm scale.

These self-assembled structures of giant surfactants in the bulk state immediately suggest their potential applications in thin-film nanopatterning. The understanding of thin-film physics has greatly advanced the block copolymer lithography over the past decade, but with feature sizes typically of 20–100 nm (4, 27, 46). The direct generation of laterally long-range ordered, densely packed, defect-free, device-oriented nanostructures (47) possessing sub-20-nm feature sizes and sharp interfacial line edges is a grand challenge in microelectronics to keep pace with the prediction of Moore’s law (48). Giant surfactants are perfect candidates to address these issues because, besides the proper feature sizes, the conformational rigidity and chemical monodispersity of the MNP heads would probably limit the line edge roughness, whereas the inorganic silicon–oxygen backbones of the POSS heads would increase the etch contrast between two phase domains. Moreover, the fast self-assembly dynamics of these materials could potentially facilitate the formation of long-range ordered nanostructures having low defect density via solvent or even simple thermal annealing within relatively short times. The latter thermal process is ideal for industrial applications. Furthermore, the shape and volume persistency of the heads might impose packing constraints under confined environments (such as thin films), thus providing access to unconventional nanopatterns.

DPOSS-PS₃₅ that exhibits a DG nanostructure in the bulk state was first studied in thin films. As shown in the TEM image of Fig. 3A, the thin film of DPOSS-PS₃₅ after solvent annealing displays a line pattern. The dark lines with width of ca. 2.5 nm correspond to the DPOSS domains, which are anticipated to be composed of two layers of DPOSS MNPs held head-to-head by collective hydrogen bonding. The grazing-incidence small-angle X-ray scattering (GISAXS) pattern (Fig. 3A, Inset) indicates a Hex structure in thin-film state, in which the cylinders preferentially orient along the {10} plane, parallel to the film substrate (49). The calculated spacing between cylinders based on the GISAXS result is ca. 11.0 nm, matching well with the TEM observations, from which the line spacing is determined to be ca. 6.0 nm due to the superposition of multiple layers of cylinders. The observed morphological transition from DG in the bulk to Hex in the thin film for DPOSS-PS₃₅ is due to the thin-film confinement. Using a similar procedure, the APOSS-PS₇₅ sample with a BCC structure in the bulk (Table S1) shows a rectangular dot pattern (Fig. 3B) under TEM. These dark spherical dots are APOSS domains with a diameter of ca. 4.0 nm. In-plane symmetry is analyzed on the basis of the ratio between the two in-plane vectors a₁ (the second-nearest neighbor) and a₂ (the first-nearest neighbor), following the method previously reported (50). The value of a₁/a₂ is determined to be 1.14 for the APOSS-PS₇₅ thin film on the basis of the GISAXS result shown in Fig. 3B. This value reveals that the POSS spherical domains pack into a face-centered orthorhombic (FCO) lattice, which possesses in-plane symmetry intermediate in between the hexagonal lattice (a₁/a₂ = 1) and BCC {110} plane (a₁/a₂ = 1.154) (50, 51). The assembled rectangular arrays conform to the industry-standard rectilinear coordinate systems. These observations promise giant surfactants as emerging nanomaterials for nanostructure fabrication in the field of microelectronics.

Whereas the self-assembly of giant surfactants in the bulk and thin-film states is similar to that of block copolymers, they behave like small-molecule surfactants in solution (30, 40, 52). Not only do they exhibit various micellar morphologies, but also the polymer tails are found to be highly stretched under certain circumstances, a feature reminiscent of small-molecule surfactants. In addition, unusual nanostructured colloidal particles have also been observed from compounds in libraries 2 and 4 by slow addition of a selective solvent, e.g., water, into their solution in 1,4-dioxane. A large difference exists between the colloidal particles obtained from a pair of topological isomers, AC₆₀-PS₄₄ and AC₆₀-2PS₂₃. Dynamic light scattering (DLS) shows a unimodal narrow size distribution for the formed particles with sizes centered at 220 nm for AC₆₀-PS₄₄ and 140 nm for AC₆₀-2PS₂₃, respectively (Fig. S3). Scanning electron microscopy (SEM) and TEM bright-field images indicate that spherical colloidal particles are formed from AC₆₀-PS₄₄ (Fig. 4 A–C) and double-truncated conical particles from AC₆₀-2PS₂₃ (Fig. 4 D–F). Tilting SEM experiments confirm the shape of particles (Fig. S4). A zoom-in view of the particles in the TEM further shows an onion-like inner structure in spherical particles (Fig. 4C) and a hexagonal inner structure for double-truncated conical particles (Fig. 4F). The inner structures are clearly similar to the corresponding bulk structures of these two samples (Fig. 2C). These colloidal particles are highly stable in solution, as revealed by their zeta potentials (−57.0 ± 5.5 mV for spherical particles and −60.4 ± 7.8 mV for double-truncated conical particles). It is thus rationalized that the surface of these particles is composed of mainly anionic AC₆₀ heads and the bulk portion undergoes further self-organization into various finer nanostructures via phase separation between hydrophilic AC₆₀ heads and hydrophobic PS tails. The formation of spherical colloidal particles in solution is probably due to the balance between a smaller bending energy caused by the curvature and a greater gain of surface free energy. In contrast, the formation of double-truncated conical particles is most likely dictated by the sixfold symmetry of the close-packed nanostructured cylinders. Although these colloidal particles are reminiscent of “cubosomes” (53), their formation is the result of strong collective hydrogen-bonding interactions and the conformational rigidity of the heads during the self-assembly process, which was not observed in traditional block copolymers.

In conclusion, we have demonstrated by the design, synthesis, and self-assembly of a broad class of giant surfactants that they provide a versatile platform for sub-10-nm nanostructure engineering. As unique additions to the traditional amphiphiles, this class of materials bridges the gap between small-molecule surfactants and block copolymers. They can form diverse structures in the bulk, thin-film, and solution states and exhibit a small-molecule surfactant–block copolymer duality. The self-assembly is found to be sensitive to primary chemical structures in terms of functional groups on MNPs, polymer topology, and chain composition, which suggests possibilities to fine-tune nanostructures through chemical structural variation of the giant surfactants. It should be noted that this class of giant surfactants belongs to an even broader concept of giant molecules and may also be constructed by combining a wide range of nanobuilding blocks and/or structural motifs together, such as MNPs, dendrimers, polymers, and globular proteins, etc. Through structural diversification, this class of materials opens up unique possibilities for macromolecular assemblies.

This work was supported by the National Science Foundation (DMR-0906898) and The Joint–Hope Education Foundation.